Enjoyed the episode? Want to listen later? Subscribe by searching 80,000 Hours wherever you get your podcasts, or click one of the buttons below:

Anti-corruption in contexts where corruption is widespread only works if you can separate people who are violating rules for reasonable reasons from other people who are violating rules for unreasonable reasons. And if you can identify the people who are violating the rules for reasonable reasons and solve their problems so they can comply with rules, they become your allies.

Mushtaq Khan

If you’re living in the Niger Delta in Nigeria, your best bet at a high-paying career is probably ‘artisanal refining’ — or, in plain language, stealing oil from pipelines.

The resulting oil spills damage the environment and cause severe health problems, but the Nigerian government has continually failed in their attempts to stop this theft.

They send in the army, and the army gets corrupted. They send in enforcement agencies, and the enforcement agencies get corrupted. What’s happening here?

According to Mushtaq Khan, economics professor at SOAS University of London, this is a classic example of ‘networked corruption’. Everyone in the community is benefiting from the criminal enterprise — so much so that the locals would prefer civil war to following the law. It pays vastly better than other local jobs, hotels and restaurants have formed around it, and houses are even powered by the electricity generated from the oil.

In today’s episode, Mushtaq elaborates on the models he uses to understand these problems and make predictions he can test in the real world.

Some of the most important factors shaping the fate of nations are their structures of power: who is powerful, how they are organized, which interest groups can pull in favours with the government, and the constant push and pull between the country’s rulers and its ruled. While traditional economic theory has relatively little to say about these topics, institutional economists like Mushtaq have a lot to say, and participate in lively debates about which of their competing ideas best explain the world around us.

The issues at stake are nothing less than why some countries are rich and others are poor, why some countries are mostly law abiding while others are not, and why some government programmes improve public welfare while others just enrich the well connected.

Mushtaq’s specialties are anti-corruption and industrial policy, where he believes mainstream theory and practice are largely misguided. To root out fraud, aid agencies try to impose institutions and laws that work in countries like the U.K. today. Everyone nods their heads and appears to go along, but years later they find nothing has changed, or worse — the new anti-corruption laws are mostly just used to persecute anyone who challenges the country’s rulers.

As Mushtaq explains, to people who specialise in understanding why corruption is ubiquitous in some countries but not others, this is entirely predictable. Western agencies imagine a situation where most people are law abiding, but a handful of selfish fat cats are engaging in large-scale graft. In fact in the countries they’re trying to change everyone is breaking some rule or other, or participating in so-called ‘corruption’, because it’s the only way to get things done and always has been.

Mushtaq’s rule of thumb is that when the locals most concerned with a specific issue are invested in preserving a status quo they’re participating in, they almost always win out.

To actually reduce corruption, countries like his native Bangladesh have to follow the same gradual path the U.K. once did: find organizations that benefit from rule-abiding behaviour and are selfishly motivated to promote it, and help them police their peers.

Trying to impose a new way of doing things from the top down wasn’t how Europe modernised, and it won’t work elsewhere either.

In cases like oil theft in Nigeria, where no one wants to follow the rules, Mushtaq says corruption may be impossible to solve directly. Instead you have to play a long game, bringing in other employment opportunities, improving health services, and deploying alternative forms of energy — in the hope that one day this will give people a viable alternative to corruption.

In this extensive interview Rob and Mushtaq cover this and much more, including:

  • How does one test theories like this?
  • Why are companies in some poor countries so much less productive than their peers in rich countries?
  • Have rich countries just legalized the corruption in their societies?
  • What are the big live debates in institutional economics?
  • Should poor countries protect their industries from foreign competition?
  • Where has industrial policy worked, and why?
  • How can listeners use these theories to predict which policies will work in their own countries?

Get this episode by subscribing to our podcast on the world’s most pressing problems and how to solve them: type 80,000 Hours into your podcasting app. Or read the transcript below.

Producer: Keiran Harris
Audio mastering: Ben Cordell
Transcriptions: Sofia Davis-Fogel

Key points

Institutional economics

Mushtaq Khan: A political settlement describes the distribution of organizational power and capabilities. And I think that’s a very good and necessary map at a very rough-cut level to understand what is the society that you’re trying to analyze and describe. And I think a lot of social science analysis doesn’t do that. They just look at one side of the story, which is the rules. How much market should we have? How much competition should we have? What kinds of policies should the government have? Which policies work?

Mushtaq Khan: And then you find that actually a policy that worked very well in one country is disastrous in another country, and different countries which have successfully developed, have done so with quite different policies. And that’s a puzzle, but the puzzle becomes much more explicable when you look at this interaction: that actually the policies that work, the institutions that work, the rules that work depend on the capabilities, interests, and power of the players and how they implement and follow those policies. Because no policy can be enforced on people who don’t want to follow it.

Mushtaq Khan: The idea that a policy is a black box that the government just announces and everyone starts following it is a total mistake. A government is just one organization amongst many organizations in society. And there is an interplay going on between governments, political parties, opposition parties, trade unions, churches, mosques, the people, different kinds of agencies and forums… All of them are trying to influence this policy outcome, but also — and this is critical — implementing it on a daily basis. And if the vast majority of your organizations and society are happily violating the rules and not checking each other, then it’s not going to be an implementable policy.

Mushtaq Khan: Policies that work are policies where at least some players… And here I think we need to be very careful about what we mean by the powerful, because the powerful are not the top 1% of society. Who is powerful depends on the policy being discussed. For some policies, the powerful are people within a village. For other policies, the powerful are people who run big firms. But whoever is violating the rule needs to be checked by other people who are just as powerful as they are. And this is why, in common-law English tradition, there’s this concept of being judged by your peers. What does that mean? Judging by your peers means that at any level of society, there are other people and other organizations who are as powerful as you are. If your peers don’t check you, you will not get any rule-following behavior.

Mushtaq Khan: And we always make the mistake of thinking that enforcement always comes from above, that there is some kind of police or some kind of enforcement agency sitting and watching everyone. And if you violate rules, they come down on you. What we often forget is that this type of vertical enforcement only works if 99% — or at least 90% — of enforcement is happening without that, without what we call vertical enforcement. It is happening through horizontal enforcement, by peers at different levels of society. When we say powerful, we mean the relevant people at that level of activity who are as powerful as the people violating. Now, if all the people who are violating are interested in violating, and the only check is coming from people lower than them in terms of power, or you are only relying on enforcement from above, you’re not going to get anywhere.

Mushtaq Khan: What we are looking for always are these incremental processes where whatever policy you’re looking at, whichever level of society, you’re looking for rifts between the players, between the powerful at that level, and seeing if some of them might in their own interest support developmental outcomes in a rule-following way. And we promote that. And if you look at the history of how development happened in advanced countries and how a rule of law emerged, it was always through these incremental processes where people at different levels of society in their own interests started to follow rules and impose rules on their peers. And that horizontal process is for some strange reason not adequately examined by economists and institutional economists. We are often much more concerned about these vertical enforcement strategies, which then typically fail.

Organizational capabilities

Mushtaq Khan: In a lot of countries, and particularly in developing and emerging countries, it’s patently obvious that organizations linked to governments are extremely inefficient and corrupt. So when you have a policy which says let us try and develop infant industries by giving them subsidies or giving them access to cheaper land or industrial zones, or somehow supporting their development, these policy rents or policy resources immediately get captured by all kinds of people who are in a ‘political process.’

Mushtaq Khan: And so observers often come along and think wow, the politics here sucks, right? How can we get out of this politics? Let’s privatize, and let’s liberalize, and let’s take this away because the private sector is good and the private sector wants to make money. So the private sector will find market opportunities to do things. And what they ignore is that the private sector also has very low capabilities. The private sector also has firms that are not capable yet of really engaging at a global level in competition. And so it is not that politics is driving corruption. It’s often low capabilities driving a lot of these informal activities.

Mushtaq Khan: What happens with a lot of privatization and liberalization is you privatize these firms, and then you find that these private-sector firms are now borrowing money from banks which they are not repaying, the banks can’t collect this money from them. The banks collude with them. The private banks and the private firms collude in stealing money, which the public central bank has to then bail out, because you don’t want the system to collapse, and you’ve got another mess, but this time driven by private sector corruption.

Mushtaq Khan: This distinction is not good enough. What we’ve ignored is the distribution of capabilities and organizations, and how they’re interacting. Because the players are often the same. The players who are in politics are also in business. The players who are in business are often deeply connected to the army or in courts. And so what you have is a bunch of closely cooperating or colluding players with low capabilities. And here is the real problem: there is no quick way of raising capabilities. This is a very slow process. And the process of raising capabilities requires policy, and it requires support. The problem is that as soon as you say that our firms are actually not fit for purpose, we can’t open up trade and put them into global competition, and somehow foreign investors will invest in these companies and make them competitive… They will not.

Mushtaq Khan: What will happen is that these companies will go bust, and you will have a meltdown. So you don’t do that. But then if you say, okay, our starting point is very poor and we don’t have capabilities even to make a t-shirt, even to make a football, we can buy the machines, we can train people, but we don’t know how to put this machine and these skilled people together in an organization and actually churn out at the other end a football that can be marketed globally, that’s what’s missing. It’s very easy to actually buy the machines. It’s very easy even to train your people and have skills training. It’s much, much more difficult to put them together in an organization, which can actually produce on time a product of the right quality at the right price, which can compete against other countries that have those organizations.

Anti-corruption policies

Mushtaq Khan: A lot of [major anti-corruption strategies] sound like they’re very plausible and they should work, but most of them don’t work. And a lot of empirical studies have been done on their effectiveness. Most of it is really depressing reading. The standard anti-corruption approaches that are at play basically assume that most people in these countries are rule-following citizens. They’re going about their everyday business following the rules, and a few people who are powerful control information, or are able to violate rules for their own benefit. And so corruption is that misuse of power. That’s the definition everyone uses, that a few powerful people are misusing their power and appropriating resources from the public. If that is the case, then the obvious answer is transparency, accountability, and punishment.

Mushtaq Khan: Those three things describe the vast majority of anti-corruption measures that have been attempted. Transparency means you try to make transparent the processes through which decisions are made and the outcomes of those decisions, so that people can see who is misusing power. Accountability means you make it clear whose responsibility it was to make those decisions, so you can hold them to account. And then there is a process of holding them to account, whether through elections or through taking them to court or prosecuting them in different ways. And the final step of it is strengthening the enforcement agencies, the people who are going to do this prosecution.

Mushtaq Khan: A lot of money has been spent on anti-corruption commissions, on strengthening courts, on strengthening the police. A lot of money has been put into that. Transparency plus accountability plus enforcement capacity is describing, in a sense, the vast majority of anti-corruption efforts. And generally speaking, the efforts have not delivered results which are sustainable. So for a time, for six months or a few months, you might get an improvement, but then things revert back to where they were.

Mushtaq Khan: The question is why? And the question is very easy to answer once you understand this distribution of capabilities, interests, and power. Because what you have in developing countries is a set of capabilities across society that means that they’ve adopted the rules of advanced countries, company law, election rules, rules of property rights, rules of governance… But in fact, the vast majority of people in those countries don’t have the capability to adhere to or follow those rules. What you find is that 70–80% of society is informal. Most people have such low capabilities that their organizations do not even have the resources to register themselves as a registered organization. Forget about health and safety rules and employment rules and pension rules and other rules. They’re not even registered.

Mushtaq Khan: You have societies where 70–80% are unregistered. You also have big companies that are registered and all of them are also violating different rules. You have a political process which is not based on collecting taxes, and offering people services through the budget because the tax base is very low and the demands are very high. Politics is organized through patronage, through clientelism. All the politicians, from the prime minister downwards, are breaking rules. In other words, you have societies where almost everybody is violating some rule. When you come and say transparency, well it’s already transparent. If you go to a developing country, the taxi driver will tell you as you’re driving from the airport to your hotel, “This building belongs to that minister. That building was that minister’s, this prime minister made that much money.” It’s completely transparent, in the sense that people might not know the exact figures and so on, but they know it’s happening.

Mushtaq Khan: In terms of accountability, we have elections, we have all kinds of processes, but the powerful — remember that powerful means those who are relevant for that violation — remain where they are. The violation might be much lower down in society, but whoever is violating very rarely comes to trial and gets punished, because actually all of this is happening in networked ways. And each violator is linked with other violators, and everyone has something to lose. And so finally what happens is that your enforcement agencies are also corrupted. The anti-corruption commission becomes a tool in the hands of the government to chase the opposition and pick them up, or to threaten businesses who are not playing ball, saying that we will have an investigation of your tax. And then they start playing ball with the government. In other words, the police, the courts, the anti-corruption commission on whom a lot of money is being spent, themselves become a source of the problem.

Mushtaq Khan: When you look at that — and I think anyone who lives in a developing country will immediately recognize this — you can see that this approach from above, whether it’s transparency, whether it’s accountability, whether it’s enforcement, isn’t working. Now, it’s very important for me to make this clear: we are not saying that this should stop. I think that some of this needs to happen just to keep pressure on the system. But I think what needs to stop is an expectation that this is anywhere close to being sufficient to achieve anything. And we really need to work on the other aspects of anti-corruption, which I’m going to talk about later, which is where the action is. But it’s not that you should let these people off the hook. The point is that you might identify the corrupt prime minister. You might get rid of the corrupt prime minister, but since the system is a low-capability, informal system, the next prime minister will be just as corrupt. So we shouldn’t put our hopes on that process of enforcement from above for achieving either development or anti-corruption.

How to define capabilities

Mushtaq Khan: Capabilities can describe quite different things, depending on what kind of organization you’re talking about. Let’s take the simplest kind of productive organization in a developing country, which is a labor-intensive, very low-tech manufacturing of something of very little value and very little markup, something like a t-shirt or a football. Now, why is it that not every developing country is producing this? Why are there a few countries which produce it and most of them cannot? The answer is that to actually produce a t-shirt, you don’t just need the sewing machines, which are actually freely available, not under any kind of patent control, and you can buy them for a dime a dozen. And there are lots of people in developing countries with enough money to buy millions of these things.

Mushtaq Khan: It’s not even the question of skills, because the skills required are very basic. And really you could skill up people in a few weeks to operate these kinds of machines. And that includes electronics assembly processes. It can include making footballs and t-shirts and all kinds of basic manufacturing. Why doesn’t it happen? It doesn’t happen because the really big elephant in the room is, how do you organize this production in a factory? This includes things like, how do you deal with bottlenecks? How do you make sure that your processes are aligned so that the different inputs and outputs are aligned for the fastest throughput of materials? How do you reduce waste? How do you ensure quality control? How do you manage inventories? These things sound rather simple. And it sounds like, well, if I can hire some people with an MBA or a business admin degree, they should be able to tell me.

Mushtaq Khan: No, actually this has nothing much to do with what you learn in business school. It is to do with how you organize a whole team of people to operate seamlessly as an organic whole. And it sounds to us to be rather obvious, but this is an incredibly difficult thing to achieve. Take the example of a hospital in a developing country. Hospitals in developing countries have doctors who are very skilled. In fact, most of them would love to leave and take a job in an advanced country where they would perform perfectly well. They have all the machines that you require for a hospital. They have the drugs, or many or most of them. And yet their capacity to deliver good health services is very poor. The reason is not the quality of the people or the quality of the machines. It’s how it’s organized. Are you doing the cleaning properly? Are you managing the flow of tests so that the right tests go at the right time to the right doctor for the right patient?

Mushtaq Khan: Are you managing your entry so that the beds are kept just about full enough, but not overly full? Are you managing your quality control and your ordering of spare parts? And this is where it fails. This is where universities don’t work in some countries, hospitals don’t work in other countries. Not because they don’t have professors. I’m from a country where universities don’t work very well, but there are many people like me who are from that country who are world-class professors. But the problem is not the professor. The problem is not the machine or the desk or the whiteboard. The problem is the organization, and how all of this is put together. We really haven’t given much thought to that, from the simplest garment industry to the government itself, how the government is just an organization, the government doesn’t work because it is an organization.

Mushtaq Khan: Advanced countries are dealing with this on a daily basis. We are continuously restructuring how our universities work, how our governments work, our companies work, how factories work. It’s an endless process.

An example of effective horizontal checks

Mushtaq Khan: Bangladesh is a very vulnerable country. It suffers a lot from cyclones and rising sea levels. Hundreds of millions, billions of dollars have to be spent on mitigation activities. These include building embankments across rivers, on the sides of rivers to prevent flooding and building cyclone shelters. And Bangladesh has spent hundreds of millions, probably billions by now on that.

Mushtaq Khan: More has to be spent now. And I think some of the discussions which are going to happen now about supporting developing countries is about that. However, we also know that massive corruption happens in these investments. Transparency International Bangladesh, who is our partner in this study, estimated from a sample of studies that they did that something like 30% of the money going into climate investment projects disappears. 80% of them are badly constructed. Now this is huge. These are massive policy resources which are going into constructing these projects, which are big infrastructure projects. And a lot of money goes missing. How do we stop that? All of these have the same vertical enforcement reporting, community meetings, community boards, where all the information is put up. Transparency systems have no effect.

Mushtaq Khan: ACE looked at a comparison of embankment projects and cyclone shelter projects which had high corruption in some cases and low corruption in other cases. And this was as reported by key informants and Transparency International’s own monitoring. And our question was, why do some of these projects have high corruption and some of them have low corruption? Why are some better constructed and why are some constructed so badly when they all have the same vertical arrangements? And it turns out the difference is in the horizontal checks and balances. What we found is that in the projects that were well constructed, there was — by accident or design — a dual-use component of the investment. The embankments can be used as roads, and cyclone shelters can be used as shops, community centers, and schools.

Mushtaq Khan: Now, if you’ve designed the project properly, the dual-use benefit is quite important. And, as in a rich country, but even more so in a poor country, people are much more interested in the immediate benefits of the project than in the long-run climate mitigation which may or may not happen. When there are immediate benefits, like a road which connects up important bits of economic activity, or a school that is of benefit to the community, you suddenly find a significantly higher percentage of economically and politically more effective people in the local community becoming involved in monitoring. And so what we found is that people with above-average incomes become involved in a much greater percentage when there are immediate benefits. These are powerful at the local community level, but they’re actually, in absolute terms, extremely poor people.

Mushtaq Khan: This is another good example of how when we say powerful, it’s always in a relative sense of that community and of what is going on. In communities where lots of people are involved in monitoring, but they’re of lower-than-average power, you find highly corrupt projects. But when the more-than-average-power people in terms of landholding and economic income become involved — and remember, they are still very poor people — they are a very effective horizontal check, because the contractors and others who are making it are also from their group. They’re local people who are making these things. In a context where formal enforcement is very weak, the informal enforcement of meeting someone in a tea shop and saying, “Actually the cement you’re using, the sand you’re using is not really good, I can see that, the road will not be fit for purpose,” that threat is the really important threat. That someone in your group is saying, “This is not going to work.” And when that person starts making a fuss, then the political system, the enforcement system starts responding.

Mushtaq Khan: This is a critical aspect of reality in developing countries. That those informal networks and informal pressure are much more important. So how do you get the more effective people, the more capable people interested — and so capabilities come in, in this indirect sense — how to get more capable people interested in this? And we have, again, a very simple answer: When you’re designing your climate investments, make sure that there is some immediate benefit in terms of dual use which brings in the local community. Not just in an abstract sense, like oh we are interested in this in the long run, but in the immediate benefits. And our evidence shows that this makes a huge impact.

What happens when horizontal policing isn't possible

Mushtaq Khan: We have this fantastic project in the Niger Delta in Nigeria, where you have this problem of artisanal refinery, which is basically the local communities stealing oil from pipelines. And then they use that in local, very crude refining, which they then sell on. Now, this is a massive problem of theft, which has huge environmentally damaging consequences, and creates health problems which are very severe. You may even have seen pictures of the Niger Delta and the oil spills that are happening there as this oil is stolen. And the Nigerian government has tried year in and year out to stop this problem with enforcement. The army is sent in, and the army gets corrupted. Enforcement agencies are sent in, and they get corrupted. Sometimes they actually attack local communities, and civil war breaks out. The Niger Delta is close to or in civil war-type situations often because of these kinds of activities.

Mushtaq Khan: We went in and looked at what was going on here. And it was a really risky enterprise. We had local partners who were extremely well connected, and that’s why we could do this research. And we found that, actually, this is a classic example of networked corruption, where almost everybody is benefiting from this corruption and there is no rule-following behavior at all, because the local community has no alternative methods of employment which give you anything like the returns that they get from this artisanal refining — which is a very nice term for what is going on. And the returns that they get are often higher than even extremely well-paid jobs in the Nigerian army, for example. The local community also has lots of externalities. The service sector activities, which set up hotels and restaurants, different kinds of services run by women, you have trading activities…

Mushtaq Khan: Then the refined petroleum is sold on for generating electricity. And so it also is powering houses. It’s a complete economy, which is working in what we call ‘networked corruption.’ Now here you don’t find any insiders who are following rules. No one will report on anything. No one will support enforcement. Even though their benefits are unequal, the whole community is against any kind of external enforcement activity, and external enforcement generates violence and generates almost civil war-type issues. What do we do there? Our answer is, do not try vertical enforcement here, but try to find exit strategies from this problem. Here is a case where actually development is the answer. You have to find parallel employment opportunities for people. You have to improve health services and primary health clinics. You need to bring in solar power, so that the people who are buying this refined petroleum have some alternative sources of electricity, and so on.

Mushtaq Khan: And we could have a whole list of things which we think are feasible in that area which basically reduce the dependence of this community on this entrenched networked corruption. And so to us, that is the anti-corruption solution there, that alternative process of development, or just taking the steam out of this hothouse activity of artisanal refinery, which can be done, and is happening at the margin because some solar power is coming in, some alternatives. We want to accelerate that process. We’ll create the development, which will make it more possible for people to start exiting. This is an example. Now I think there are lots of such exit strategies which we identify as well, because a lot of activities in developing countries are like that, where the corruption is so entrenched that almost everybody in that activity is corrupt.

Mushtaq Khan: They might be corrupt for different reasons, but they’re all rule violating. And so when you come in to enforce the rules, you find no support. I think another way of thinking about it is that anti-corruption works when there are people who are violating rules for reasonable reasons, and there are other people who are violating rules for unreasonable reasons. And if you can identify the people who are violating the rules for reasonable reasons and solve those problems, then they become your allies, and then you minimize the number of people who are the true thieves, who are the small subset of people who are genuine thieves and free-riders. And when you can bring them down to a manageable number, like 10% or so, then your vertical enforcement starts working, right?

Mushtaq Khan: Because then the peers start saying, “Well actually, we are all following rules. Who do you think you are to go to the front of the queue? Go back and stand at the back of the queue.” Exactly like what happens when you break a queue in London. It’s not the police who stop you, it’s your peers who say “Hey, please go back to the back of the queue.”

Takeaway lessons

Mushtaq Khan: I think a lot of what I’m saying could be read by people in different ways. Let me boil it down to some very basic things which might focus people’s minds. I think now it’s commonplace in development policy that people will say, “Context matters. One size doesn’t fit all.” Everyone will say these things, from the World Bank, down to the Foreign, Commonwealth & Development Office, down to Oxfam. Everyone will say this.

Mushtaq Khan: The real question is, you have to ask yourself, “How does context matter? What am I supposed to do about this? What is it about it that matters?” Here is where I think all that I’m talking about has a direct relevance. Context matters in this way. You have a set of plans and policies, or you’re supporting some activities which you think will support human development. To me, that inevitably involves looking at capabilities, organizations, etc. Do you think these policies will be implemented? Do you think they can be enforced? This is where context matters. Context matters primarily in the sense that enforcement and implementation vary hugely depending on the interests and capabilities of the players themselves, and much less on the enforcement capability of some supra-agency, which is like the prime minister’s office, or the cabinet, or the police, or the anti-corruption commission, or anything like that.

Mushtaq Khan: The basic thing you should do, always, is look at the people you are working with, their capabilities, and then ask, “Here is a policy to help them do something better. Will they themselves actually support it?” Not are they saying they will, because everyone will say they will do whatever you ask them to do, because they want the money. But actually, from their past behavior and their activities, do I think this is something that they will actually be able to implement, enforce, monitor each other as peers, report back, and do the basic capabilities of policy implementation at a higher level exist to monitor and enforce this?

Mushtaq Khan: If not, you have several options. You can redesign the policy so you have greater confidence it works. You can say, “Actually, policies won’t work here, because there’s some network problem which is so integrated that I can’t break it.” Then you have to think out of the box about some exit strategy which meets the requirements of your anti-poverty or whatever that you are trying to achieve. Or you have to go back one step before, and ask, “Can I build those capabilities directly? Can I work with people to actually build those capabilities that I’m assuming exist with my policy? I shouldn’t assume they already exist. Maybe I need a policy that is one or two steps behind in terms of its sequencing.”

Mushtaq Khan: I think there are ways in which all of this makes a lot of practical sense. When you are asking yourself “Will this work here?” then of course you can refer to the work that I do and that others do. It’s all available. It’s free to access online. You can read some of the stuff on the SOAS website, or the work of Dani Rodrik, or North, Wallis, and Weingast, or Acemoglu, Johnson, and Robinson, and all the people we have discussed. Not all of it, but those bits of it that are relevant to the particular policy question that you’re addressing. I think that’s the starting point. That’s the starting point for you, as a development practitioner, beginning to refine your questions and getting down to better questions and then answers which you think might make sense there. Then, the final step is, don’t put all your eggs in one basket. Experiment with a trial, see how that goes, and only scale up once it’s working.

Articles, books, and other media discussed in the show

Mushtaq’s recent work

Videos with Mushtaq

Other links

Transcript

Rob’s intro [00:00:00]

Rob Wiblin: Hi listeners, this is the 80,000 Hours Podcast, where we have unusually in-depth conversations about the world’s most pressing problems, what you can do to solve them, and how to teach old warlords new tricks. I’m Rob Wiblin, Head of Research at 80,000 Hours.

As someone with an amateur passion for social science, today’s interview is a real delight and very different from what I expected when we booked Mushtaq as a guest.

Mushtaq Khan and I dive right into some of the most central questions in politics, economics, and development.

  • How do the powerful influence how a country is governed?
  • Why are some countries rich and others poor?
  • How can we forecast how a country’s institutions will evolve over time?
  • Why are some societies coordinated by legal systems, while others rely on informal mechanisms to coordinate activity?
  • Can you predict ahead of time which policy ideas are going to be corrupted?
  • How well does a laissez-faire approach to economic development work?

It’s not all theory though, because Mushtaq is very much an applied economist. He has spent years out in the field testing his ideas and seeing what good they can do in practice, and he’s only too happy to talk about the results of that research.

I basically put away all my notes and questions after the first five minutes and we just winged it, which I can’t remember ever doing before.

If you like this episode, note that Mushtaq heads the SOAS-ACE research consortium that is currently funded by the U.K.’s Foreign, Commonwealth and Development Office — but they are always looking for new sources of funding to continue or expand their research. If you’d like to learn more about their approach and opportunities to help out with their work, you can contact [email protected] or visit their website: ace.soas.ac.uk.

The last few years we’ve had fairly few episodes on global health and development, relative to how many of you out there in the audience are focused on those issues personally. There’s always many more pressing topics to cover than we have episodes to cover them, but our plan is to try to discuss those topics a bit more over the coming year.

Okay. Without further ado, here’s my interview with Professor Mushtaq Khan.

The interview begins [00:01:55]

Rob Wiblin: Today, I’m speaking with Mushtaq Khan. Mushtaq is a professor of economics at SOAS University of London, where he leads the Anti-Corruption Evidence Research Consortium, which has a grant from the U.K. government to figure out how to design anti-corruption projects so that they actually work. His main research interests are institutional economics and the dynamics of influence, rent-seeking, corruption, and clientelism. By reflecting on the selfish interests of the most powerful people in a particular political situation, he aims to anticipate the ways that government programs will ultimately be undermined and potentially find ways to get powerful groups to use their political influence for positive rather than destructive ends.

Rob Wiblin: Long ago, Mushtaq did his PhD in economics at the University of Cambridge, and over the course of a long career has, among many other things, advised the Department for International Development (DFID), the United Nations, and World Bank, taught in Thailand and Bangladesh, and done empirical research on the ground in Bangladesh, Thailand, Vietnam, Nepal, Palestine, Nigeria, India and Tanzania. Thanks so much for coming on the podcast, Mushtaq.

Mushtaq Khan: It’s a great pleasure.

Rob Wiblin: I’m particularly excited about this interview because this is the first in-person interview that I’ve done since episode 70 with Cassidy Nelson on the 12 best ways to stop the next pandemic and limit COVID-19. I guess that podcast didn’t achieve its goal entirely. But that one was recorded on the 11th of February 2020, which means that we’ve gone 527 days between in-person interviews. But here we are, both fully vaccinated with the windows open and negative lateral flow tests in hand. So fingers crossed this episode will kill neither of us.

Rob Wiblin: Alright. I hope we’ll get to talk about what sort of anti-corruption efforts are practical, and how one figures that out. But first, what are you working on at the moment and why do you think it’s important?

Mushtaq Khan: One of the important things we’re working on is actually policy implementation in developing countries. Policy is clearly extremely important for dealing with massive challenges that we face globally. And in developing countries in particular, but not just in developing countries, policy implementation is often thwarted by social forces working at cross purposes with the policy. And this is often ignored by policymakers. And whether that is anti-corruption, whether that’s industrial policy, whether that’s climate policy… If the policy cannot be implemented, it might look very good on paper, but it’s not a good policy. So a lot of my research is broadly to do with the general issue of policy implementation. And we are working on a number of global projects with multiple partners on different types of policy implementation issues.

Rob Wiblin: Hey listeners, I forgot to pause here and explain what industrial policy refers to. Basically industrial policy is a deliberate effort to encourage the growth of part of a country’s economy, most often some part of the manufacturing sector. Most governments do this to some degree, though some much more than others.

The classic questions economists study regarding industrial policy are whether it’s actually a good use of resources, how to best direct any subsidies, and how to prevent the relevant support being misappropriated or wasted propping up industries that are ultimately doomed regardless. Okay, back to the show.

Rob Wiblin: So your current main focus is this grant from the U.K.’s aid agency, where you’re trying to test out things on the ground in Bangladesh and other countries to collect evidence in favor of different theories and models and figure out what programs are likely to work? Is that right?

Mushtaq Khan: Yes. One of these is the Anti-Corruption Evidence Research Consortium, run out of SOAS, which is a research partnership consortium. It’s a partnership of 13 or so universities across the world and other partners, which are private-sector partners and agencies. And we are trying to figure out what might be feasible — which means implementable — anti-corruption strategies, which promote development. That’s one. And the other program is also funded by the Foreign, Commonwealth & Development Office (FCDO). That program is in Nepal, and we’re looking at policy implementation in Nepal’s transition. Nepal is setting up a federal structure of decentralized organizations after a long civil war. And so it’s a post-conflict society. And again, the issue of policy implementation is critical: What kinds of policies and institutions can actually be implemented? So it’s part of the same sort of question about implementing policies in difficult contexts.

Rob Wiblin: We’ll come back to some of those case studies later on, but I wanted to start by kind of setting the scene of what I’ve come to understand maybe as the Mushtaq institutional economics lens on things. And I’m curious to know whether perhaps I’ve understood things correctly, or maybe I’ve gotten the wrong end of things. So initially when I was preparing for this interview… I think on the surface, a lot of your work focuses on corruption and anti-corruption efforts, which is interesting and important in itself. And that’s what I originally planned to kind of build this conversation around. But as I was preparing more, I came to conceptualize your work a bit differently, as having a broader lens on things, rather than just being about corruption. It kind of feels like you bring a model of the effects of political influence on the development process as a whole, and on predicting how policy is actually going to play out in the real world as a whole.

Rob Wiblin: So on your view, it’s potentially desirable for the government to offer resources, to subsidize firms, to build up their organizational capacity through costly experimentation as a country is developing, but whenever those resources are made available, they become rents that organizations will try to use their influence to access and extract one way or another. And under some circumstances you’ll find that powerful people in organizations will turn that political influence to getting those resources in exchange for doing nothing useful in particular. And as a result, they might try to undermine anti-corruption efforts, because that is just what happens to serve their personal interests in that case. In rich countries you get potentially a similar phenomenon where organizations, powerful firms, might try to extract resources from society, but they’re more likely to do it through lobbying rather than illegally through corruption.

Rob Wiblin: But in a sense, corruption and legal rent-seeking through lobbying are two manifestations of related phenomena that play out just because countries are organized differently. But at other times you potentially find that powerful organizations and groups can line up behind the rule of law and strengthen policies that benefit a country as a whole, most importantly by policing the behavior of their peers, who they see free-riding on a system, and a system that they think that they benefit from, from existing. So in that situation, they potentially benefit personally from enforcing the kinds of laws and norms they would like to see.

Rob Wiblin: But anyway, all of these are just kind of particular expressions of this more general phenomenon that powerful organizations and individuals have strings to pull when they want to change what’s going on around them, and how a government is operating and what it’s doing. And if we want to predict how things are going to actually play out in rich countries and poor countries alike and do things like reduce corruption and improve peer monitoring, then we need to think about their motivations and incentives.

Rob Wiblin: How on point do you think that explanation of your research agenda is? And what else are key ideas from your work over the years that people should know about?

Mushtaq Khan: I think that’s quite a good summary, quite a good cut at it. Let me begin with some essential concepts that inform my thinking and how they’re interconnected, and then we’ll come back to the more specific questions you asked.

Mushtaq Khan: I think at a very general level, it’s informed by a very historical understanding of development. And if you ask me what is development, what is the main thing that distinguishes less-developed from more-developed countries, I think it’s something that most people don’t normally think about. It’s the distribution of organizational capabilities. Because whatever it is that a society produces — and I’m using ‘producing’ in the broader sense, it could be culture, it could be art, it could be services, it could be food — it requires organization. And the difference between a less-developed and a more-developed society is that the number of organizations that can be set up is much more complex and diverse in more-advanced countries compared to less-advanced countries. So organization is a critical part of my story.

Mushtaq Khan: Now, the debate within a lot of social science is, what’s stopping organizations developing in poor countries? Why don’t they just spontaneously grow like mushrooms? And here is this very complex interaction between organizations and the rules of the game, which economists call institutions. The institutions might include markets, but might also include states, might include other kinds of rules, which guide how organizations emerge and interact with each other. And I think here is where my position is different from some others. I think that the way in which institutions work is also affected by organizations, because organizations not only create the rules of the game, they’re also continuously trying to change and distort or affect how they’re implemented. And as you said, this is happening in advanced countries and in developing countries.

Mushtaq Khan: In advanced countries, it happens through lobbying. It happens through political processes. It happens through education and advertising and think tanks. In developing countries, it happens through all of that, but also a lot of informal activity of informal processes, which are sometimes outside the rule of law and sometimes completely contradictory with it — including, for example, corruption. So, that’s the fundamental difference. And I think that that difference itself reflects the distribution of organizational capabilities in society. Because the most powerful organizations in advanced countries happen to be powerful organizations which need rules for their own reproduction. They need rules for complex contracting. They need rules to raise finance in complex ways. They need to organize large numbers of people who are not known to them. They’re nameless, faceless people. So you need to have a contract-based rule of law system.

Mushtaq Khan: And so in advanced countries, generally speaking, most powerful organizations want a rule of law. And the difference in developing countries is that the powerful organizations are networks, which are informal patronage networks, kin networks, clientelist networks, tribal networks, religious networks, or even companies which are not that capable themselves, and their interactions with each other do not require a rule of law. And therefore a lot of their activity of lobbying, pressuring, and so on is informal. So I don’t think this is very much to do with culture or other kinds of things like that, although they do matter. It’s largely to do with the very nature of development, that developing countries have a preponderance of organizations that don’t need a rule of law. And yet, the fact that they don’t need a rule of law can stop high-capability organizations from developing.

Mushtaq Khan: Once you understand this complexity, you have to figure out how to nudge systems where you have a lot of players who don’t want a rule of law, but there might be pockets where in their own interests, some players want rules. And our job is to support that. Why? Because the aim of that is not a moral aim that we have to make people follow rules, or behave morally. The aim is to help the development of capabilities and organizations with capabilities that eventually results in higher levels of welfare for people, better lives. And eventually, as a byproduct, a rule of law emerges, because all these people eventually will need a general enforcement of rules for their own benefit.

Mushtaq Khan: I think, before we go on, I would like to define what a rule of law is, so that your listeners know the difference.

Rob Wiblin: Yeah, go for it.

Mushtaq Khan: A rule of law means that you have a system of rules, where if you are found to be violating that — and everybody might want to violate rules, right? Because there’s always a free-riding incentive. I want everybody to pay taxes, but if I can get away with not paying taxes, I will not. That is common to all systems where rules have a common benefit, which is much more than the private benefit. Everybody wants the rules in an advanced country, but also wants to violate the rules. The difference in an advanced country is that if you are caught violating the rules, then in general, however powerful you are, or whoever you are, you pay a penalty. In other words, the penalty for violation does not depend on who you are.

Mushtaq Khan: Now, what is different in countries where the rule of law is weaker or absent is that once you are caught and you are found to be violating a rule, your probability of evading punishment depends on who you are. It depends on who you know. It depends on what party you belong to. It depends on what organizations are supporting you. When that happens, technically speaking, you don’t have a rule of law. Because you have rules, and lots of people follow rules, but who follows the rules and how it will be enforced on them depends on who you are. As soon as that happens, I would say you don’t have a rule of law.

Mushtaq Khan: A rule of law requires a distribution of power in society, which is so broad-based that even the powerful follow rules, because there are lots of other powerful people who will not tolerate one of the powerful not following rules. But if the powerful are not very capable, they don’t need rules. And if they’re small in number, they can collude with each other and they will impose rules on everyone else, but they will violate the rules themselves. As soon as that happens, you don’t have a rule of law. And that’s why we describe the situation in most developing countries as rule by law.

Mushtaq Khan: Laws are enforced even when you don’t have a rule of law, but they’re enforced on those on whom they can be enforced. And those who can evade those, or bypass them, usually tend to get away. It may look like rules are enforced, but you might only have a very strong rule by law. And people often confuse that with a rule of law. A rule of law is a very, very specific thing. And only a very small handful of countries are close to that. And even in those countries, sometimes it’s—

Rob Wiblin: It’s touch and go.

Mushtaq Khan: It’s touch and go. But in most developing countries, it’s not touch and go. It’s clearly not there.

Institutional economics [00:15:37]

Rob Wiblin: So there’s a lot there that we’re going to unpack gradually over the next few hours. Let’s return to the first aspect of this that you were describing. So you’re broadly working within this sub-school of economics called ‘institutional economics,’ which studies governments and how they evolve, and how systems of organization and law evolve over time. The world’s a complex place, and you can conceptualize it lots of different ways, but it sounded like the framework that you’re trying to use is to think about how there’s institutions — like governments, legal systems, social norms, and so on — and then there’s also organizations — like firms, religious groups, social groups, villages — that those institutions are acting on. But it goes both ways.

Rob Wiblin: The institutions in society act on and influence the behavior of these firms, causing them to focus on improving their productivity or not. But then those firms also have a role in shaping what government policies there are and how they’re implemented, as do all other influential organizations in society. And so you have this cycle back and forth, and potentially that can move in a positive direction, where you get a virtuous circle, where rule of law improves. So you have more impartiality in the legal system. And then the firms are interested in supporting that, and fostering that, and increasing it, because that serves their interest, and so on. Or you can have a negative one where you just get stuck, where there isn’t a rule of law, the institutions are very corrupt, but also all of the influential organizations in society are kind of happy with that. Is that a reasonable summary?

Mushtaq Khan: I think that’s a very good summary. And I think that feedback between institutions and organizations is what I study in my concept of political settlements. A political settlement describes the distribution of organizational power and capabilities. And I think that’s a very good and necessary map at a very rough-cut level to understand what is the society that you’re trying to analyze and describe. And I think a lot of social science analysis doesn’t do that. They just look at one side of the story, which is the rules. How much market should we have? How much competition should we have? What kinds of policies should the government have? Which policies work?

Mushtaq Khan: And then you find that actually a policy that worked very well in one country is disastrous in another country, and different countries which have successfully developed, have done so with quite different policies. And that’s a puzzle, but the puzzle becomes much more explicable when you look at this interaction: that actually the policies that work, the institutions that work, the rules that work depend on the capabilities, interests, and power of the players and how they implement and follow those policies. Because no policy can be enforced on people who don’t want to follow it.

Mushtaq Khan: The idea that a policy is a black box that the government just announces and everyone starts following it is a total mistake. A government is just one organization amongst many organizations in society. And there is an interplay going on between governments, political parties, opposition parties, trade unions, churches, mosques, the people, different kinds of agencies and forums… All of them are trying to influence this policy outcome, but also — and this is critical — implementing it on a daily basis. And if the vast majority of your organizations and society are happily violating the rules and not checking each other, then it’s not going to be an implementable policy.

Mushtaq Khan: Policies that work are policies where at least some players… And here I think we need to be very careful about what we mean by the powerful, because the powerful are not the top 1% of society. Who is powerful depends on the policy being discussed. For some policies, the powerful are people within a village. For other policies, the powerful are people who run big firms. But whoever is violating the rule needs to be checked by other people who are just as powerful as they are. And this is why, in common-law English tradition, there’s this concept of being judged by your peers. What does that mean? Judging by your peers means that at any level of society, there are other people and other organizations who are as powerful as you are. If your peers don’t check you, you will not get any rule-following behavior.

Mushtaq Khan: And we always make the mistake of thinking that enforcement always comes from above, that there is some kind of police or some kind of enforcement agency sitting and watching everyone. And if you violate rules, they come down on you. What we often forget is that 99% — or at least 90% — of enforcement is happening without that, without what we call vertical enforcement. It is happening through horizontal enforcement, by peers at different levels of society. When we say powerful, we mean the relevant people at that level of activity who are as powerful as the people violating. Now, if all the people who are violating are interested in violating, and the only check is coming from people lower than them in terms of power, you’re not going to get anywhere. Or if you solely rely on external enforcement: the vertical enforcement.

Mushtaq Khan: What we are looking for always are these incremental processes where whatever policy you’re looking at, at whichever level of society, you’re looking for rifts between the players, between the powerful at that level, and seeing if some of them might in their own interest support developmental outcomes in a rule-following way. And we promote that. And if you look at the history of how development happened in advanced countries and how a rule of law emerged, it was always through these incremental processes where people at different levels of society in their own interests started to follow rules and impose rules on their peers. And that horizontal process is for some strange reason not adequately examined by economists and institutional economists. We are often much more concerned about these vertical enforcement strategies, which then typically fail.

Rob Wiblin: I’m very excited about this framework, and getting myself and listeners to use it a little bit more. I studied economics in undergrad, and we spent lots of time just thinking about what policies — in theory, or maybe even as a result of studies — would be optimal in principle, if you could get them enforced. But we spent almost no time on this institutional economics framework, where you think about what is actually going to happen in practice, given the interests and the capabilities of the organizations that exist in society.

Rob Wiblin: And I think that’s one reason why the policy advice that economists and people like me often offer comes across as extremely naive. It’s good in theory, but it won’t work in practice. It either gets rejected, or it’s implemented and then peters out and doesn’t achieve its desired goals, because we don’t use this framework of… And I guess it’s understandable, because it’s very hard to do. It’s a lot harder than just doing like economics 101 theory to figure out what policies to have. But we don’t often think about how the political process is going to play out, and how the implementation will actually play out.

Rob Wiblin: So this institutional economics framework is useful with anti-corruption, but it also potentially helps a lot with thinking about development policy as a whole, and with thinking even about what policies are actually likely to be implemented properly, or even get up out of the political process in rich countries as well. This process is going on in the U.K. and the U.S. and Australia and other places. And very often I think it can explain why particular policies get undermined, ultimately, and don’t achieve their goals, or why they’re just not going to get through the political process. Because they’re too contrary to the interests of the people who are most concerned. Is my optimism about the value of this framework justified, do you think? I would hope that you think it’s useful, but…

Mushtaq Khan: I think it’s useful, but it’s even worse than the way you put it. It’s actually even worse, because there isn’t this distinction between politics and markets, I think. And this is where I think it’s very useful to think about organizations and organizational capabilities. Let me give you an example. In a lot of countries, and particularly in developing and emerging countries, it’s patently obvious that organizations linked to governments are extremely inefficient and corrupt. So when you have a policy which says let us try and develop infant industries by giving them subsidies or giving them access to cheaper land or industrial zones, or somehow supporting their development, these policy rents or policy resources immediately get captured by all kinds of people who are in a ‘political process.’

Mushtaq Khan: And so observers often come along and think wow, the politics here sucks. How can we get out of this politics? Let’s privatize, and let’s liberalize, and let’s take this away because the private sector is good and the private sector wants to make money. So the private sector will find market opportunities to do things. And what they ignore is that the private sector also has very low capabilities. The private sector also has firms that are not capable yet of really engaging at a global level in competition. And so it is not that politics is driving corruption. It’s often low capabilities driving a lot of these informal activities.

Mushtaq Khan: What happens with a lot of privatization and liberalization is you privatize these firms, and then you find that these private-sector firms are now borrowing money from banks which they are not repaying, the banks can’t collect this money from them. The banks collude with them. The private banks and the private firms collude in stealing money, which the public central bank has to then bail out, because you don’t want the system to collapse, and you’ve got another mess, but this time driven by private sector corruption.

Mushtaq Khan: This distinction is not good enough. What we’ve ignored is the distribution of capabilities and organizations, and how they’re interacting. Because the players are often the same. The players who are in politics are also in business. The players who are in business are often deeply connected to the army or in courts. And so what you have is a bunch of closely cooperating or colluding players with low capabilities. And here is the real problem: there is no quick way of raising capabilities. This is a very slow process. And the process of raising capabilities requires policy, and it requires support. The problem is that as soon as you say that our firms are actually not fit for purpose, we can’t open up trade and put them into global competition, and somehow foreign investors will invest in these companies and make them competitive… They will not.

Mushtaq Khan: What will happen is that these companies will go bust, and you will have a meltdown. So you don’t do that. But then if you say, okay, our starting point is very poor and we don’t have capabilities even to make a t-shirt, even to make a football, we can buy the machines, we can train people, but we don’t know how to put this machine and these skilled people together in an organization and actually churn out at the other end a football that can be marketed globally, that’s what’s missing. It’s very easy to actually buy the machines. It’s very easy even to train your people and have skills training. It’s much, much more difficult to put them together in an organization, which can actually produce on time a product of the right quality at the right price, which can compete against other countries that have those organizations.

Mushtaq Khan: At this moment in time, for a lot of developing countries the competition is China. Earlier, it used to be South Korea. And before that, it used to be Germany and England. But whoever is the next country beyond you has an organizational capability which is much better than yours. And so just getting the machines and just getting the skilled labor and then saying, okay, we have a lower competence, but we can deal with that. Our productivity is much lower, but our wages are lower so we can compete. Actually, no. Your productivity is usually orders of magnitude worse than your wage advantage. And so even with the lowest wage in the world, you can’t compete.

Mushtaq Khan: This is where you need policies to actually build those organizations. And that’s where those policies then get captured. If you’ve designed them wrongly, because you haven’t aligned your policy with the interests and capabilities of your specific firms and specific sectors, then the firms say, okay, we’re getting the support, but actually this support will not take us to becoming a competitor with China. It’s better we use our political connections and our networks and our organizational pressure and our corruption to just capture this and make hay while the sun shines, because we’re never going to go from here to there. And that’s where it fails.

Mushtaq Khan: Actually getting a policy which is sufficient to take at least some of them to the point where they become actually competitive players, creates support for that policy, and then at least some of those players will say, actually, this policy is going to work for us, but those free-riders are stopping it. And then the horizontal pressure begins, and that horizontal pressure is what is often missing.

Mushtaq Khan: It’s actually very commonsensical once you understand it. It’s just amazing that people haven’t looked at who it is that they’re trying to help when they design their policy. And I can give you — and I’m sure we will talk about — examples of anti-corruption where a lot of the corruption is simply because the policy has no hope of succeeding, because you haven’t looked at all the other factors that are causing organizations to behave in those ways. Once you start addressing those things, you begin to have a policy which has a hope in hell of succeeding. And that is what we are trying to do.

Anti-corruption policies [00:28:45]

Rob Wiblin: Okay, so being a bit more empirical, what do we know about anti-corruption policies as they exist today? Is it the case that they sound plausible and get funded, but then ultimately don’t achieve their desired goals because of these institutional incentive issues that you’ve been talking about?

Mushtaq Khan: Well, I think that’s true for most of the major anti-corruption strategies that are at play right now. A lot of them sound like they’re very plausible and they should work, but most of them don’t work. And a lot of empirical studies have been done on their effectiveness (for instance, this one). Most of it is really depressing reading. The standard anti-corruption approaches that are at play basically assume that most people in these countries are rule-following citizens. They’re going about their everyday business following the rules, and a few people who are powerful control information, or are able to violate rules for their own benefit. And so corruption is that misuse of power. That’s the definition everyone uses, that a few powerful people are misusing their power and appropriating resources from the public. If that is the case, then the obvious answer is transparency, accountability, and punishment.

Mushtaq Khan: Those three things describe the vast majority of anti-corruption measures that have been attempted. Transparency means you try to make transparent the processes through which decisions are made and the outcomes of those decisions, so that people can see who is misusing power. Accountability means you make it clear whose responsibility it was to make those decisions, so you can hold them to account. And then there is a process of holding them to account, whether through elections or through taking them to court or prosecuting them in different ways. And the final step of it is strengthening the enforcement agencies, the people who are going to do this prosecution.

Mushtaq Khan: A lot of money has been spent on anti-corruption commissions, on strengthening courts, on strengthening the police. A lot of money has been put into that. Transparency plus accountability plus enforcement capacity is describing, in a sense, the vast majority of anti-corruption efforts. And generally speaking, the efforts have not delivered results which are sustainable. So for a time, for six months or a few months, you might get an improvement, but then things revert back to where they were.

Mushtaq Khan: The question is why? And the question is very easy to answer once you understand this distribution of capabilities, interests, and power. Because what you have in developing countries is a set of capabilities across society that means that they’ve adopted the rules of advanced countries, company law, election rules, rules of property rights, rules of governance… But in fact, the vast majority of people in those countries don’t have the capability to adhere to or follow those rules. What you find is that 70–80% of society is informal. Most people have such low capabilities that their organizations do not even have the resources to register themselves as a registered organization. Forget about health and safety rules and employment rules and pension rules and other rules. They’re not even registered.

Mushtaq Khan: You have societies where 70–80% are unregistered. You also have big companies that are registered and all of them are also violating different rules. You have a political process which is based not on collecting taxes, and offering people services through the budget because the tax base is very low and the demands are very high. Politics is organized through patronage, through clientelism. All the politicians, from the prime minister downwards, are breaking rules. In other words, you have societies where almost everybody is violating some rule. When you come and say transparency, well it’s already transparent. If you go to a developing country, the taxi driver will tell you as you’re driving from the airport to your hotel, “This building belongs to that minister. That building was that minister’s, this prime minister made that much money.” It’s completely transparent, in the sense that people might not know the exact figures and so on, but they know it’s happening.

Mushtaq Khan: In terms of accountability, we have elections, we have all kinds of processes, but the powerful — remember that powerful means those who are relevant for that violation — remain where they are. The violation might be much lower down in society, but whoever is violating very rarely comes to trial and gets punished, because actually all of this is happening in networked ways. And each violator is linked with other violators, and everyone has something to lose. And so finally what happens is that your enforcement agencies are also corrupted. The anti-corruption commission becomes a tool in the hands of the government to chase the opposition and pick them up, or to threaten businesses who are not playing ball, saying that we will have an investigation of your tax. And then they start playing ball with the government. In other words, the police, the courts, the anti-corruption commission on whom a lot of money is being spent, themselves become a source of the problem.

Mushtaq Khan: When you look at that — and I think anyone who lives in a developing country will immediately recognize this — you can see that this approach from above, whether it’s transparency, whether it’s accountability, whether it’s enforcement, isn’t working. Now, it’s very important for me to make this clear: we are not saying that this should stop. I think that some of this needs to happen just to keep pressure on the system. But I think what needs to stop is an expectation that this is anywhere close to being sufficient to achieve anything. And we really need to work on the other aspects of anti-corruption, which I’m going to talk about later, which is where the action is. But it’s not that you should let these people off the hook. The point is that you might identify the corrupt prime minister. You might get rid of the corrupt prime minister, but since the system is a low-capability, informal system, the next prime minister will be just as corrupt. So we shouldn’t put our hopes on that process of enforcement from above for achieving either development or anti-corruption.

Capabilities [00:34:51]

Rob Wiblin: So this concept of capabilities is super central in this model because the capabilities of people and organizations determine whether they want to compete globally or whether they prefer to take rents or use some other method to make money. How do you define or imagine what capabilities are, at a brass-tacks level, and maybe why does it take so long — decades, it sounds like — for a country to build up or organizations to build up the capabilities that they need?

Mushtaq Khan: Really big question there. I think at one level, the answer is quite simple, but it’s not a uniform thing. Capabilities can describe quite different things, depending on what kind of organization you’re talking about. Let’s take the simplest kind of productive organization in a developing country, which is a labor-intensive, very low-tech manufacturing of something of very little value and very little markup, something like a t-shirt or a football. Now, why is it that not every developing country is producing this? Why are there a few countries which produce it and most of them cannot? The answer is that to actually produce a t-shirt, you don’t just need the sewing machines, which are actually freely available, not under any kind of patent control, and you can buy them for a dime a dozen. And there are lots of people in developing countries with enough money to buy millions of these things.

Mushtaq Khan: It’s not even the question of skills, because the skills required are very basic. And really you could skill up people in a few weeks to operate these kinds of machines. And that includes electronics assembly processes. It can include making footballs and t-shirts and all kinds of basic manufacturing. Why doesn’t it happen? It doesn’t happen because the really big elephant in the room is, how do you organize this production in a factory? This includes things like, how do you deal with bottlenecks? How do you make sure that your processes are aligned so that the different inputs and outputs are aligned for the fastest throughput of materials? How do you reduce waste? How do you ensure quality control? How do you manage inventories? These things sound rather simple. And it sounds like, well, if I can hire some people with an MBA or a business admin degree, they should be able to tell me.

Mushtaq Khan: No, actually this has nothing much to do with what you learn in business school. It is to do with how you organize a whole team of people to operate seamlessly as an organic whole. And it sounds to us to be rather obvious, but this is an incredibly difficult thing to achieve. Take the example of a hospital in a developing country. Hospitals in developing countries have doctors who are very skilled. In fact, most of them would love to leave and take a job in an advanced country where they would perform perfectly well. They have all the machines that you require for a hospital. They have the drugs, or many or most of them. And yet their capacity to deliver good health services is very poor. The reason is not the quality of the people or the quality of the machines. It’s how it’s organized. Are you doing the cleaning properly? Are you managing the flow of tests so that the right tests go at the right time to the right doctor for the right patient?

Mushtaq Khan: Are you managing your entry so that the beds are kept just about full enough, but not overly full? Are you managing your quality control and your ordering of spare parts? And this is where it fails. This is where universities don’t work in some countries, hospitals don’t work in other countries. Not because they don’t have professors. I’m from a country where universities don’t work very well, but there are many people like me who are from that country who are quite good professors. But the problem is not the professor. The problem is not the machine or the desk or the whiteboard. The problem is the organization, and how all of this is put together. We really haven’t given much thought to that, from the simplest garment industry to the government itself, how the government is just an organization, the government doesn’t work because it is an organization.

Mushtaq Khan: Advanced countries are dealing with this on a daily basis. We are continuously restructuring how our universities work, how our governments work, our companies work, how factories work. It’s an endless process.

Rob Wiblin: And it’s difficult.

Mushtaq Khan: It’s difficult. But the point is that we are making incremental changes. What we are asking a developing country to do is make a massive leap from a society which is basically largely peasant agriculture and so on, where timekeeping and discipline and so on has a different meaning, to one where these things mean something quite different. And if you read European history or the history of Germany or Japan or advanced countries, that transition is a long and painful one. It’s about converting society into a lot of different organizations, which are working seamlessly. That’s not easy. It requires a lot of resources.

Mushtaq Khan: And here is another point that we forget. And that is the problem of late development. The problem of late development is that when England was the first industrializer, other countries in Europe were just a little bit behind. Germany could start catching up with Bismarckian policies and Friedrich List’s economic justification for supporting their infant industries. And it wasn’t such a big gap to catch up. Japan could catch up with Germany, but it required a lot of state assistance with the Meijis. By the time you come to today, the gap in productivity between China and Bangladesh, even in something as basic as garments production, is not just a few percentage points of productivity. It’s 7:1 or 8:1, which means that with one third or one fifth of the wages, Bangladesh finds it very hard to compete. And we’re talking about garments.

Mushtaq Khan: As soon as you start talking about something more sophisticated, that productivity gap might be 15, 20, or 30 times more in the next country. And no wage advantage will get you there. And this is where you need a lot of learning by doing. And that’s another key concept. The learning by doing is that you need to set up organizations, which are, when you set them up, completely uncompetitive. And then you have to find some way of giving them that knowledge, either through transfers from outside, or through their own learning, or through some combination. But none of this is going to happen purely through market incentives, because it’s not worth anyone’s while to do that, because there isn’t a lot of money you can make doing this. And this is why the learning that happens, which is very critical, doesn’t usually happen just through market incentives. It needs something else to drive it. And that is why capability development is so difficult. It’s basically that we have to have another way of thinking about capability development, as the problem of late development.

Mushtaq Khan: When the gap between your organizations and the next country is anything more than just marginal and incremental, then markets won’t solve the problem. If it’s an incremental small issue, then someone will come and invest, raise your competitiveness, and make lots of money, because your wages are lower. But if the gap is so big that they have to invest a lot and your wages…

Rob Wiblin: Even if they’re zero.

Mushtaq Khan: Even if they’re zero, it’s too risky and they’re not going to make money, they won’t come. And that’s actually what you see. Most developing countries have their gates open, waiting for investors, and they don’t come. The low wages don’t help. It’s your absent organizational capabilities that are really the binding constraint.

Why the market doesn’t solve the problem [00:42:29]

Rob Wiblin: So to think a little bit more about why is it that the market wouldn’t solve this problem, I suppose, so why is it that a firm in these countries wouldn’t say, “Well, we’re going to play a long game. We’re going to spend 10 years or 20 years developing our workforce so that we can organize ourselves properly, and then we’ll be able to compete. And then we’ll make a lot of money.” I suppose, one thing is you’ll have internal principal agent issues, where the people who are at the firm by the time it can ever make any money… It’s going to be so far down the road that the people might’ve changed. I guess you also have a problem that they’ll learn a whole bunch of stuff about how to make, say, garments manufacturing work in Bangladesh, but then that will bleed over into other firms. And there’ll be other people who mostly get the benefit from this, rather than the firm that’s investing all the money. I guess you also just have risk aversion.

Rob Wiblin: But yeah, it’s hard to pinpoint why it is that an investor or a firm itself wouldn’t even really try to make that leap towards becoming globally competitive in terms of their organization.

Mushtaq Khan: All of the things that you’ve mentioned. It’s to do with risk. It’s a huge risk if the gap is large. The return from taking that risk is a social benefit. Now that social benefit is that you develop organizations in your country that then create lots of jobs and so on, but the initial investor isn’t going to capture that. The point is that the social benefit is huge, but the private risk is also huge. You have to have some way of sharing that risk. And the risk is actually not just that you don’t know what will work. You might know what will work, but it will just take a long time to get there.

Rob Wiblin: And you’ll be making a loss the whole time.

Mushtaq Khan: And you’ll be making a loss the whole time. And you need to have many checks and balances for this thing to work. And so if it’s just one investor going in, in a very foolhardy way, into a society where everything is informal and lots of informal rents exist, you don’t know whether your investment will also be captured by these informal networks and be frittered away. And there’s no contract enforcement that you can do there, which says that I want my money back if you don’t succeed.

Mushtaq Khan: You are also talking about countries where, because of this informality and poor governance, you also have very weak contract enforcement. On top of everything else to do with learning, you have a contract enforcement problem. And people say, “Well, we could solve this problem by having better contract enforcement with good governance.” And that is this whole chicken-and-egg way of thinking.

Mushtaq Khan: You have to have your policy work in the world as it is now, not in the world you would like to be in. If you have a country where you have weak rule of law, where corruption is very high, where you have high levels of risk — in contracts, but also in the very process of delivery — you would have to be really foolish to go and invest lots of money there. So actually what people from outside invest in, or even people from within the country invest in, are things where you can make quick money easily by being close to power. By doing deals with powerful people and protecting your markets, you make your money quickly and you reduce your risk.

Mushtaq Khan: There’s nothing wrong with that, except that this is not developing those broad capabilities which are going to be essential for real development, as opposed to pockets of development where some investors and some politicians and so on are connected. And sometimes they do really good work. It’s not always that these kinds of networks and collusions are value destroying. They might be value adding, but they’re value adding in relatively easy areas, and they’re leaving the rest of the country behind. And I think that’s the problem. So we have to find ways of actually creating capabilities in areas where sensible, rational investors won’t automatically go.

Industrial policy [00:46:11]

Rob Wiblin: We’ve gotten into industrial policy earlier than I planned, but let’s just lead into it. It’s great material. The economist Dani Rodrik, who I think quite a lot of listeners will have heard from, has kind of a related explanation or defense of why industrial policy might be necessary for a country to develop around innovation externalities, where, for example, let’s say we go back to Bangladesh in the 70s or 80s, before it was doing lots of garment manufacturing at a productive level. You might have this theory, or a business person might have a theory that I can make garment manufacturing work really productively in Bangladesh. But if they go in there and risk all of their money trying to invest in this and trying to see whether it can work within this culture with the kind of people who they can hire, and they discover that it does work, they can immediately be copied by other firms that come in later on without having taken that risk or taking the initial losses, who will hire away their staff and use the lessons that they’ve learned.

Rob Wiblin: He focuses on — at least in some of the lectures that I’ve heard — he focuses on that kind of knowledge-externality issue, that you need to foster people to do experiments. It’s like we think government needs to foster scientific R&D because of the big externalities and the discoveries, but in a country like Bangladesh, maybe the discovery you want is what industries are actually feasible here? What industries can we make work? And then how do we make that work? Do you think that’s also an important explanation?

Mushtaq Khan: I think it’s a potentially plausible explanation, but I don’t find it plausible as it stands. Dani has many explanations, and I think he’s a great supporter of policy-based development, but I think this particular one doesn’t work too well. And I’ll quickly explain why. I don’t think he’s completely wrong, but I think that the reason why people don’t invest in these countries is not that they think we will discover something, and if we discover it there’s a lot of money to be made, but then others will copy us and this copying will take this money away. I think he’s making too close an analogy with innovation, because innovation is like that. So someone who innovates makes a rent, and then if that innovation is copied too quickly…

Rob Wiblin: It’s not enough rent.

Mushtaq Khan: Then the incentive to innovate goes away. Some rents have to be protected to create the incentive for the original investment. I think he was just trying to copy that idea into learning. But I don’t think that works, because the successful learner, regardless of whether anyone is copying you or not, isn’t going to make much rent. There isn’t rent in a garment factory. It’s not like discovering a new COVID medicine, or a new kind of telephone that does various magical things. Making a garment, you are competing with many other garment producers in the world. And as soon as buyers know that you have cracked the garment production in Bangladesh, they will give you a price that squeezes out most of the profit from you.

Mushtaq Khan: In fact, garment producers never make a rent in whichever country they are working in, because they’re part of very competitive global value chains. People don’t go into garments because they want to make rents. This analogy with innovation is completely wrong. This is the first mistake. People are not not investing in garments because they think their profits will go away. There was never any very high profit there. If you’re successful, you will make a normal profit. That’s all you can expect. Because as soon as you are profitable, the pricing and the quality control in the global value chain will not leave you with anything like an innovation rent. Absolutely impossible.

Rob Wiblin: I think the idea might be that you figure out how to take people whose salaries are extremely low, because their opportunity costs within this very poor country are very low. And you can earn a rent, in a sense, by producing garments at a productive rate while hiring people who you pay relatively little.

Mushtaq Khan: Yeah. But your buyer knows that. Your buyer knows your cost of production. They’re not going to give Bangladesh the same price they’re giving China. They know that it’s cheaper to produce it in Bangladesh. They will figure out exactly what the pricing is. And they will give you a price which is high enough to be competitive. And that’s what they will do. And that’s true for a lot of these low-tech commodity products. You do not make a lot of money simply because your wages are lower. And as I just said earlier, your productivity difference is still substantial. Figuring out how to make garments in Bangladesh doesn’t mean you’ve achieved Chinese productivity with Bangladeshi wages. It means you’ve raised productivity in Bangladesh just enough to survive with the initially very low wages, and any profit you have there will be intensely negotiated with the buyer chain, the value chain, that you will actually not make profit whatever happens. Because you’re starting off with very low productivity and making a product that many others can make.

Mushtaq Khan: The real question is a different one. How and why do people invest in these productivity-raising capability developments? And in a sense, the problem with Dani Rodrik’s — that particular explanation, he has many explanations, as I said — is that he doesn’t really spell out that process. It’s almost like a discovery issue. Bangladesh might be good at making garments, but no one knows that yet. So I go and discover that. Actually Bangladesh is not good or bad at making anything. And another country might not be good or bad. He has this example that some countries are good at making hats. And some countries are good at making bed sheets. And then the investor discovers what they are. I think that’s completely wrong. No country is good at making hats or good at making bed sheets.

Mushtaq Khan: Any of these poor countries could make hats or bedsheets. That’s not a discovery problem. It’s a capability development problem. And the capability development problem requires a lot of investment, which at the end will give you profits, but not massive profits. Can I give an example of this?

Rob Wiblin: Yeah, go for it. That sounds great.

Mushtaq Khan: If you look at what actually made Bangladesh succeed in the garment industry in the late 1970s, early 1980s, it had nothing to do with discovery. Actually, everyone was trying to get into garments. The Americans were trying to protect their own garments industry. And then they came up with this multi-fiber arrangement deal, which would protect their government industry from competition from the existing garments producers in South Korea and Turkey and so on by limiting imports from those countries with quotas. Then they had to make this deal acceptable to the world, to other people in GATT at that time, that was the precursor of the WTO. And so they came up with this deal that said, let’s get some votes from some countries by saying that the least developed countries (LDCs) would not have quotas. And the already competent garment countries will have quotas. So they got votes and they got this through.

Mushtaq Khan: The expectation was that these LDCs will not produce garments, they will have to raise their competitiveness, and it’ll take them too long to do it. Now, what happened was a very interesting deal, which created just the right incentives in Bangladesh. All LDCs had this (quota-free access), but very few LDCs broke into the American market. What happened in Bangladesh was a very chance event where the South Korean company Daewoo was looking to offload its fabric, because it was quota-constrained with garments it could no longer export. It was looking for a partner in any country in the developing world, but it was very risky.

Mushtaq Khan: What happened was in Bangladesh there was an investor, one of these politically connected, very rich, don’t-ask-where-the-money-came-from kind of people who was willing to buy all the machines, buy the land for the factory, set up this massive factory, but didn’t have the organizational capabilities. It wasn’t like someone came and discovered what needed to be done in Bangladesh. They knew it could be done, but they didn’t have the capabilities. So they did this deal. The deal was that this Bangladeshi company called Desh would send 150-odd mid-ranking staff to the Busan factory of Daewoo in South Korea. They would sit in the factory and learn the organization. Learning by doing. And that part of it, just the learning by doing part of it, would be financed by Daewoo. Everything else will be financed by the Bangladeshi investor, and if it was successful and they came back and Bangladesh started exporting, Daewoo would immediately deduct some, I forget the exact figures, something like 7% or 8% of the sales value from the sales. Which is massive, but nothing like an innovation rent.

Mushtaq Khan: 7% or 8% is, in the world of commodities, a massive amount of money. Now, why could they afford to do that? It’s because the quota rents in the U.S. had raised prices in the U.S. market higher than what the South Koreans were producing at. But that was the whole point of the quota. The quota raised prices in the American market to protect American producers who were less competitive than the South Koreans because the wages were very high. So the quota in the U.S. raised U.S. garment prices. And essentially what these people did was they raised the productivity of the Bangladeshi organization just enough, given the low wages in Bangladesh, that that quota rent would be sufficient to repay the Koreans. And that didn’t come through Bangladesh at all, so there was no problem that the ‘policy rents’ that were creating the incentives for the South Koreans to transfer the organizational knowledge to Desh could be captured by connected people in Bangladesh. That incentive went straight to the Koreans.

Mushtaq Khan: So they found a way of paying for the transferring of organizational capabilities without the rent being captured. And it worked brilliantly. They thought it would take them two years to learn how to make a garment, but it took them six months. Because the Bangladeshis who went there knew they couldn’t stay on in Korea. There was no way they could capture the rent. The Koreans knew that there’s no point drawing this out, because they’re going to get the same return anyway. They put an intensive amount of training into this organizational stuff for the Bangladeshis. They came back, and Desh grew at 100% a year from then on. And this story is now extremely well known. Out of that 150 people who knew now about what the organizational requirements were, 130 of them or so — some ridiculously large number — went and set up their own garment factories. And then the mushrooming began. Unlike what Dani said, this copying and imitation did not reduce the profits of Desh at all and they did not try to stop it.

Mushtaq Khan: It was a transfer of organizational knowledge with a specific set of rents, rather than a discovery problem with a rents problem, which is the one that Dani Rodrik talks about. They’re both rents and they’re both solving specific problems. But I think on this one, I think Dani was chasing the wrong target.

Rob Wiblin: So, on this view, in order to make this leap from being a very low organizational capacity who is not able to compete on the global market to actually being able to do so, you need some kind of subsidy, some sort of rent in the meantime, in order to foster that and get it going. I guess this sort of industrial policy is often viewed somewhat negatively, I think, within economics, because economists expect these rents to just be captured in exchange for doing nothing, which I guess you’re also worried about, because that is a very familiar story. How can one set up industrial policy and all these kinds of subsidies such that it’s not just captured in exchange for nothing, and the firms and people do face the incentives to make this leap in capabilities, rather than just taking the money and running?

Mushtaq Khan: So, there are two issues there. Some economists — I think they’re a small minority now, but they used to be quite the majority not so long ago — would argue that rents are not just captured, they’re actually a waste, they’re unnecessary. So, an efficient market would have no rents, because a rent means you’re paying something for something which is higher than the next best opportunity. Right? So, why should you pay more for a widget than some other country can be producing it at?

Mushtaq Khan: So, there’s that element of it. So, a lot of market economists, liberal economists, are against rents, because they just see it as inefficient. But then there’s an additional argument that there might be lots of market failures, there are places where people will not invest, because something has gone wrong. Either they can’t enforce contracts, or there is a capability problem, or there’s a risk problem, or there is, to use Dani Rodrik’s analysis, there’s a discovery problem. There’s some problem which is preventing private investment from rushing in, and then policy could support that. But this creates policy rents.

Mushtaq Khan: And then, the second tier of argument against that is that once you create that, it will be captured. So, you can create that policy support, but it will not achieve the objective you wanted, because people will just eat the subsidy and not deliver. So, here is where the point is well taken that a lot of support and subsidies are indeed captured. And I think those people who support industrial policy, who deny that or who reject that, are actually not doing us any service. I think we should be upfront and say actually, a lot of subsidies are wasted. The challenge is to understand why they are wasted, and how you design them so that the checks and balances and the enforcement capability of the country in question is appropriate for getting the result that you want to achieve.

Mushtaq Khan: Because if you can’t do that, then the naysayers will be proved right. And the real problem is that actually going back to the market is not a solution. To say that the policy rent gets captured so let’s liberalize doesn’t solve the problem, because we started off with that, the investment isn’t rushing in. So, I think that we are then trapped into a situation where either you have people who say, “No, no, the policy will work. It just needs to be better targeted or better designed.” And those who say, “No, it will be captured.” And I think both are wrong. Policies which work in one country don’t work in another country, and we really need to understand how to design the policy in that context, taking into account the capabilities and powers of the different organizations. Like the Bangladeshi garment industry story, that policy was accidentally designed because it was designed largely by the Americans and their Multifiber Arrangement.

Mushtaq Khan: They were not trying to help Bangladesh, but the policy was designed in a way that the politically networked companies in Bangladesh could not capture this rent, and that rent created the right incentives for a massive organizational knowledge transfer, which helped Bangladesh to become now the second-largest garment producer in the world. Just that one little experiment. So, I think the gains are huge in doing this right. But I think while analysis can tell you what is unlikely to work given what’s been captured in the past, and we can suggest interesting experiments of what might work, given the distribution of powers and capabilities, the aim of analysis is not to say, “This will definitely work.”

Mushtaq Khan: So, what we can give policymakers is a portfolio of things that might work, and you try them out in very small experiments and you scale up what works. And here I am entirely with Dani Rodrik. So, I think the experimental approach is at the heart of policy, and we need to try out different things and scale up what works. And the real tragedy is that countries like Bangladesh really haven’t got a good analysis of what actually worked in the Bangladeshi garment industry. Many people still think it was liberalization and low wages. And it had nothing to do with either, your wages can be as low as you like, you won’t develop. And I think they haven’t really learned the lesson of how to build organizational capabilities and to take it to step two, step three, step four, and so on.

South Korea [01:01:31]

Rob Wiblin: A famous case study in industrial policy that a lot of people will have heard a little bit about, even if they don’t know all the details, is South Korea. It seemed like the government did provide a whole bunch of subsidies to early industries, to early-stage firms. And yet, despite the risk that that kind of thing would just be captured unproductively and taken by rich and influential people, those subsidies actually did go into firms that ultimately went on to become very productive and very globally competitive. And I think you have an explanation for how it is that industrial policy worked in South Korea when it doesn’t in lots of other situations. Based around the incentives, who was powerful in that case, and how the incentives were set up for them. Do you mind laying that out?

Mushtaq Khan: So, this is a very interesting debate. And there were people who were saying South Korea had no industrial policy, it was just a free-market success story. And others were saying it had a very successful industrial policy, which was based on building capabilities in these big holding companies, the chaebol. Samsung today is an example of a chaebol. And these capabilities were built up by giving them the resources to do the learning by doing, to set up productive organizations and learn how to raise their productivity.

Mushtaq Khan: But very similar policies did not work anywhere near as well in India, in Pakistan, in many African countries, and so on. And my explanation — and this is how I started my work on political settlements and organizations — was to look not at the policy but at the capabilities and the power of the organizations that were being supported. And I think the distinctive part of the South Korean story is that the starting point of many of these organizations was already quite good, because the Japanese colonization of South Korea had created the knowledge of industrial processes, which was more widespread than in most developing countries at that stage of development.

Mushtaq Khan: The Japanese had run organizations in South Korea, and many of the managers of these organizations became the owners of the future South Korean chaebols. In the Japanese colonial period they were Japanese-owned, but the management had some access to that organizational knowledge. So, this is one difference, but the real difference was the nature of Japanese colonialism in breaking up many horizontal political networks in Korea. And that was because Japanese colonialism was a very aggressive and oppressive form of colonialism. It didn’t rule through intermediate classes in its colonies, it ruled directly, and it ruled through great force and great viciousness, which is why you will not find any Korean today or any Chinese today who has a good word to say about Japanese colonialism, because it was very rough.

Mushtaq Khan: One consequence of that was that those horizontal networks — which businesses have with politics and other groups and unions and so on — were actually decimated in South Korea. So, the business groups that emerged in the post-Japanese period did not have the networks to protect their rents, did not have the connections with politics. So, now in the 1960s Park Chung-hee comes on and he starts trying things which are, in a sense, quite obvious. We can’t produce these things, so why don’t we give some export subsidies? Why don’t we give some protection? Why don’t we give them some low-cost loans from the publicly owned banks? Things which every developing country has tried. It’s not rocket science. It’s obvious, you can’t produce these things, your productivity is low, let’s help our businesses.

Mushtaq Khan: The difference was not that the South Koreans had innovated something called industrial policy. Everybody and their dog was trying it at that time. In fact, the South Koreans learned a lot from Pakistan, which also had a military government at that time and was doing exactly the same things: export subsidies, import protection, low-cost loans to large business houses, et cetera. The difference was that in Pakistan, when those subsidies were given, firms discovered it takes a lot of effort to convert that into higher productivity, into global competitiveness. And here is a real problem, you kill yourself trying to achieve this productivity, and your prize at the end of it is you lose your subsidy, right? And now you’re floating in the ocean of global competition and surviving.

Mushtaq Khan: Nobody in their right mind would want to do it. You would only do it because you are compelled to do it, because you know that you only have the subsidy for a brief period of time. You either sink or swim. And if you swim, you will then be competitively swimming for the rest of your life. No one wants to do it unless they are compelled to do it. The Pakistani companies or the Indian companies, or the other companies in the developing world who got these subsidies figured this out quite quickly. They realized that with this protection to get to where I want to be competitive is a massive amount of effort and hard work, great risk. I will face a lot of opposition from everyone, and the prize is I lose my subsidy.

Mushtaq Khan: Instead, what I do is I use my networks. I’m connected to that politician. I’m connected to this trade union. I’m connected to this faction. I’m connected to this bureaucrat. I’m connected to this bank. Why don’t we share these rents and prevent anyone from taking it away? The South Koreans couldn’t do that, because these companies were not connected to the banks, to the politicians, and so on. And therefore, when the state gave these subsidies and they said to them, “You have to achieve these export targets,” there was no way they could protect their rents if they failed to achieve the targets.

Mushtaq Khan: These companies were quite happy to give kickbacks, by the way, to Park Chung-hee, to the top leaders. We know this now because there’s a lot of evidence about the corruption in the system at that time, just as we know the corruption in the Chinese system in the 1980s, when it was growing rapidly. The difference is this, if you’re Park Chung-hee and you know this company is not meeting its export target, but is willing to give me a kickback from my subsidy, do I want that, or do I want to give this subsidy to a company which will meet the export target and therefore will make lots of profits and therefore will be able to give me a kickback which is much bigger? Again, it’s not rocket science. If you’re Park Chung-hee you will say, “This is a failed company, it’s just giving me back some of my own money as a kickback. Why should I take that? I’ll close it down and I’ll shift that subsidy, that protection, to some other company.”

Mushtaq Khan: And very soon the South Korean companies realized this was what was going to happen. So actually, there’s no easy strategy of protecting your rent. The only strategy is to raise your productivity. And this is what they did. The reason why this didn’t happen in India and Pakistan is not because nobody realized this, they realized it. There are some very interesting studies back in the 1960s done by Indian bureaucrats. There is a committee report called The Dutt Committee Report, in the 1960s, where the Indian bureaucracy itself looked at the protection system, the licensing system, as it was slightly different in India from Pakistan and South Korea, but the objectives were the same.

Mushtaq Khan: They said, “Look, these licenses are all being captured by these politically connected companies and we can’t do anything about it.” It wasn’t ignorance, everybody knew what was going on. They just couldn’t take those licenses away from these connected companies, because the political costs would be too high. So, my explanation is that the South Korean industrial policy worked in South Korea, but it doesn’t tell us anything about what to do anywhere else. So, if you want to do industrial policy in India or Bangladesh or Nigeria, or wherever, you need to design it completely differently, in a way that you provide the resources to those companies to raise their capabilities, but not with these mechanisms which gave them subsidies upfront and then said, “If you don’t perform, we’ll take it away.”

Mushtaq Khan: It’s the taking it away which is not credible with connected companies, but taking it away was credible for the South Korean companies, for very specific political settlement reasons. The South Korean political settlement, the distribution of power — largely as a result of the very specific Japanese colonialism — was very different from the countries which had British or French colonialism, where during the colonial period the colonial powers worked through intermediate classes who became extremely powerful through that colonial experience and developed dense horizontal networks throughout society.

Mushtaq Khan: It was indirect rule. So, there’s a quip that India was ruled by Indians, and therefore, a lot of people colluded with the colonial power. I mean, we don’t like to talk about this anymore, but in fact, if you go back to India, or many countries with British colonialism, you will still find really old people who say, “The British weren’t that bad.” Why? Because the British actually ruled through local power elites, worked through them and shared resources with them in a way the Japanese did not. So, the British, actually, while they were nicer in one respect, left behind a much more toxic legacy in another respect, because you created a political settlement with all these informal networks horizontally between business and politics and bureaucrats and so on, which becomes really difficult if you are trying to do South Korean industrial policy. It’s not impossible to do other kinds of industrial policy, but South Korean industrial policy will not work.

Rob Wiblin: Just to recap, to check that I’ve understood and am going to remember the South Korean story properly. So, there’s I think, three different key aspects for why it happened to work in this case. One is, you have got the Japanese colonialism, which meant that South Koreans had already had exposure to productive industrial mechanisms and to the kind of capabilities and organizations or structures that you might need. So, that made it more imaginable for these firms to say, “Well, we actually are going to be able to compete globally.”

Rob Wiblin: Then you had — because the Japanese had also decimated all of the existing patronage networks — the ability to call in favors between firms and politicians, and so on, that had been swept away or much weakened. It meant that it was credible for the government to say, “We’re not going to continue these subsidies indefinitely and just keep providing you free money,” because these firms didn’t have existing favors that they could call in with the politicians.

Rob Wiblin: And then, the third one was that the system was set up via this corruption, such that it was in the interest of the politicians, and I guess, potentially the bureaucrats who were running the industrial policy, to fund firms that were becoming more productive or had good prospects of becoming globally competitive, because then they would get larger kickbacks, basically. So, it was in their interest as well to think about, how can we fund the firms that are most promising and defund the firms that are not performing and aren’t going to go anywhere? Is that kind of the right story?

Mushtaq Khan: You’ve put it more bluntly than I would, and I think in a more extreme way than I would, but I think that’s generally right. So, the last point, that the incentives were set up so that the corruption did not stop the politicians from disciplining inefficient firms, is absolutely right. You have to look at rent-seeking and corruption in the context of these organizational decisions and the bargaining between them.

Mushtaq Khan: I think it is really naive to think that in any system, even in an advanced country, that politicians, bureaucrats, and big businesses, and even small businesses, are operating in an altruistic way. They’re always operating to get some benefit for themselves. Now, the difference is that in advanced countries, that benefit has to be achieved in rule-following ways, but the rules allow that benefit to be captured. So, you find that politicians will support big companies because these big companies will give donations to their political party, or reemploy them later on as consultants, something that we’ve seen in the U.K. very recently.

Rob Wiblin: Or threaten them with campaigns against them.

Mushtaq Khan: Exactly. So, it’s naive to think that the powerful will not use their money and power to get what they want, except the difference is that now it’s happening a little bit more transparently. Society has the ability to countermobilize, and you can create political movements and parties which block that with new legislation. And so that’s better, but it’s not the case that in advanced countries, power and money and influence don’t matter, they matter massively. So, the issue really is that in South Korea, the powerful wanted to achieve development in their own interests. And I think that’s really important to understand that the powerful everywhere really want to achieve that.

Mushtaq Khan: It’s not that the rulers of Pakistan or the rulers of India were actually not trying to develop their economy. Pakistan and India were in military competition with each other, right? They needed to develop their economies to build their army. So, just like the South Koreans faced the threat from North Korea — and many people say that’s what focused their minds, actually — Pakistan and India were at loggerheads. So, if they could have, they would have used their capacity to drive the economy, because that benefits them. The reason why they didn’t do it is because the higher leaders were trapped by the lower-level networks that businesses, bureaucrats, banks, and politicians had, and they couldn’t override it.

Mushtaq Khan: And it wasn’t that they were stupid, it wasn’t that they hadn’t figured it out, their own reports and their own bureaucrats were telling them exactly what was going on. The newspapers in Pakistan were full of how the so-called 22 families had captured all the subsidies and were not paying back their loans. It was completely transparent. It’s not that no one knew, it’s just that you didn’t have the power to override them in the way that the South Korean president could override a chaebol who said, “I haven’t been able to meet my export targets.” And he would say, “Well, I want to give the money to somebody else.” And they would not have the power to create a huge fuss or create political costs or to mobilize other people against the president. And so, that is the difference. If there were no political constraints, it’s always beneficial for the powerful to drive growth, right?

Rob Wiblin: Because then one way or another, they get more power or more money.

Mushtaq Khan: They get more power. So, if they’re not doing it, it’s always useful to ask, “Why are they not doing this?” And there are only two possible answers. One is that they’re really abominably stupid, right? Or the other is there is some constraint politically, or in terms of organizational capabilities, or something that you need to identify. And I think there are very, very few leaders in the world — and certainly not Nehru or Ayub Khan, or the others who were leading these Asian countries in the 1960s — who were that stupid. They had political constraints which prevented them from taking on these people who were capturing rents.

Mushtaq Khan: And then the final nail in the coffin is that if everybody is capturing rents and taking them away, you just join them. Some of these kleptocratic leaders in developing countries simply become kleptocratic because everybody is capturing the rents, they can’t stop it, so they might as well take some money out to the Swiss banks themselves.

Chiang Kai-shek [01:16:01]

Mushtaq Khan: Can I give you another classic example of this?

Rob Wiblin: Yeah, go for it.

Mushtaq Khan: So, a classic example of this is Chiang Kai-shek, the leader of Taiwan who converted Taiwan from pretty much a backwater of China into a remarkable success story — in many ways more attractive even than South Korea, in terms of distribution of income and assets.

Rob Wiblin: It is still now the largest producer of semiconductors for computers; bigger than China.

Mushtaq Khan: Exactly.

Rob Wiblin: That’s incredible. Yeah. Bigger than the U.S.

Mushtaq Khan: Exactly. So, how do you explain this? Now, Chiang Kai-shek was in China in the 1940s, and he was one of the biggest warlords. He used to go around from place to place in a train full of guns and with his own personal army, and he was fighting other warlords. In his behavior and in his preferences, he was no different from your Afghan warlord of today. This chap loses a war with the Chinese communists, goes across the water to Taiwan and suddenly develops — through 49 or so years of martial law, one of the longest periods of authoritarian rule — one of the most developmental states in Asia. What happened? I mean, what happened across the water? And I think the answer is that in China, he was one of many warlords, he could enforce nothing. Everybody else was stealing. You start stealing as well.

Mushtaq Khan: In Taiwan, he had complete control. He had brought in an army which could completely dominate that society. Why would you steal from your own house? You then develop it, right? Now, I’m not saying that this is a recommendation for any other country, that you have this authoritarian rule. And I think this is a really important point, most authoritarian governments fail, because, while they are authoritarian, they actually cannot have that effective degree of control over their society that the Taiwanese and the South Koreans and others had. So without understanding that, authoritarianism is a really bad model to follow.

Mushtaq Khan: The most common form of authoritarianism is the one that we have seen in Pakistan and Nigeria and in the Democratic Republic of the Congo, and so on. You have an authoritarian ruler, he thinks — it’s usually a he — he thinks he is an authoritarian ruler who can do all these magical things. He soon discovers they have no capacity to enforce, because actually the society just below them is powerfully networked, and can block him on everything. And then, these authoritarian rulers become thieves themselves. They are the ones who are sending money out to the Swiss banks and so on. Again, just like industrial policy that works in some places and not others, authoritarianism works in some places and not others. And I think one of the real dangers in a lot of simplistic political economy analyses is to say that if you have authoritarianism and industrial policy, you will succeed. And actually, in nine cases out of 10, you will have a disaster.

The logic of political survival [01:18:43]

Rob Wiblin: This reminds me of the model in the book The Logic of Political Survival. Basically they have a model where the folk theory of power within politics is that the president or the prime minister can do whatever they want. But in reality, they’re accountable to this group of people just below them, who could chuck them out if they wanted to. And in order to maintain their position, they have to keep all of those people one rung below them happy so that they’ll get reelected, or at least so there won’t be a coup staged. And I guess you were describing that in Pakistan, the prime minister of Pakistan might really want to develop the country, but they’re constrained because they can’t annoy the people one level below them. Because if they do, they will be removed. And so the amount of discretion that they have is actually quite limited.

Rob Wiblin: On the other hand, Chiang Kai-shek in Taiwan in the 1950s had a lot of discretion, because they had such centralized control. There were so few political competitors, and no immediate military competitors — so he had a lot of discretion. And if you want to just force through policies that some people didn’t like, because they would lead to economic development, then he had the capacity to do that. And that was in his interest, because he was confident that he would still be in control in the 1960s. Because of just the overwhelming influence. It raises this uncomfortable issue where you might expect that countries that have stronger, more dominant, more centralized authoritarian control — like China — you might expect them to develop better, because they expect the party, the leader to be around for decades. They have longer time horizons.

Rob Wiblin: They’re not just trying to avoid getting kicked out next year or the year after that by appeasing the other power centers within the country. They have the discretion to think somewhat longer term and pursue these broad goals, which is an uncomfortable conclusion. Would you say that’s consistent with the evidence that unstable authoritarianism is less likely to lead to development than the more stable centralized authoritarianism like you see in China, or I guess Taiwan, South Korea, in the past?

Mushtaq Khan: It’s complicated. And I think there are a number of pieces to this puzzle. When we talk about authoritarianism, we are often talking in a very loose way. In my political settlements analysis, I cut it in different ways. It’s about how much horizontal competition the ruling coalition faces from people not in the ruling coalition, and how much implementation capacity they have down the ruling coalition vertically. Are other people lower down going to listen to instructions from above, and what are their incentives? And I think that from that perspective, to have a ruling coalition that both has a long time horizon because it doesn’t face very credible opposition horizontally and has the capacity to put instructions vertically down the chain which will be listened to — and people lower down the chain haven’t made their own coalitions with outsiders or with each other to block it — that’s extremely rare.

Mushtaq Khan: Now I think that in different ways, the East Asian countries have bits of that. And you would have to look at the history to understand this. But the other thing is, the other part of the political settlement story is that you don’t just look at the organization of political organizations. You have to look at the organization of economic organizations. And there again, as we were saying a little earlier, the South Korean organizations already had some capabilities. Now, this kind of top-down, policy-driven, capability development works when you have some business capabilities or productive capabilities in the business sector, and you can hothouse the further development of these organizational capabilities. But at some point your economic organizations become so complex that actually discretion becomes a constraint on them. When policy discretion becomes a constraint on them, they need to have a rule of law and contract, because they need to start making horizontal networks with other organizations, both internationally and nationally.

Mushtaq Khan: And that’s what happened in South Korea in the 2000s. When the chaebol who had been hothoused under the previous regimes began to say, “We want a rule of law. Because actually we can’t work anymore with your discretion, because you have no way of knowing what we need, it’s too complicated. And we can’t explain to you all the things we need. We need to make millions of horizontal contracts. And therefore we need a rule of law.” And because those organizations were now powerful and complex and had lots of money, you saw what happened in South Korea, the transition towards a rule of law began. And they’re pretty far along that way. And it’s pretty much irreversible. China is just beginning that journey, is maybe beginning to reach the point where the Huaweis of the world will find themselves highly constrained globally if it is perceived that they are not contract bound.

Mushtaq Khan: But there’s still not enough of them. And I mean, once you have thousands of Huaweis, and the number has to be large because of the size of the Chinese economy, you need a lot of them. But I’m sure that a time will come, unless China implodes for all kinds of other reasons before that, which is not very likely, although many people are hoping for that, but if these organizations grow in number, I’m sure we will see exactly the same pressures in China from below. Horizontal pressures from powerful organizations saying, actually we need a rule of law. And that’s when you will move there. I think the whole mistake in a lot of political economy analysis is examining the requirements of an advanced country, where actually you do need a rule of law to grow, and then acting as if the Chinas and South Koreas of the 1960s, 1980s, and 1990s have the same requirements. And they say, well, it’s a real puzzle how they grew, and maybe they’ll implode. And I think that’s confusing different stages of development, where the capabilities of your organizations require different kinds of support.

Rob Wiblin: It’s this interesting phenomenon that in some of the most interesting, striking case studies in economic development or in economics as a whole of all time, we see examples of countries with explosive growth that aren’t following the rules that you would think are necessary in the U.K. or the U.S. in order to produce lots of wealth. It’s quite contrary to the theory that we have, or the standard story that we tell about economic development and what’s necessary. So, you’ve got to have rule of law, you’ve got to get rid of corruption. Otherwise you can’t have firms, otherwise you can’t grow. It just flies in the face of that.

Rob Wiblin: And yet it’s not something that we talk about that much, because it’s so difficult to figure out what’s going on. It requires quite a different framework. But I guess it’s something that you and your colleagues are trying to provide. I’m not sure that this is the correct explanation, but it seems like one of the most important questions that we can possibly have in all of economics is, how does this happen? Because it shows that there’s something wrong with the way that we’ve been thinking about economic development — at least at that level of development — for quite a long time.

Mushtaq Khan: I think that’s absolutely right. And this is a real challenge. A lot of economics is driven by the experiences of England and the United States, which are just as much outliers in terms of global development as South Korea and China. If you just looked at South Korea and China, you would say that what you need for development is an authoritarian state with industrial policy. And if you looked at England and the U.S. you would say you need free markets and limited governments with their hands tied with lots of checks and balances from the outset. And I think both are wrong. And I think that in each case, what worked can be understood in terms of the distribution of organizational and political capabilities and powers. And I could give you a very nice narrative of why that worked in the U.S. and England, and why that worked in South Korea and China.

Mushtaq Khan: And I would then conclude by saying, this doesn’t tell you anything about what to do in Nigeria, right? Because then you have to ask yourself, given the distribution of power and capabilities across Nigerian economic political organizations, what would be the nudge, what would be that incremental change in policy which would get the more capable and developmental organizations in their own interest to start supporting further development? And neither the U.K. and U.S. model or the China and South Korea model have much to say. And so this is the real challenge. We really need to go back to understanding political economy 101 from the basics, and say these are all outliers. These are examples which are interesting, but we shouldn’t generalize from that. Just as you have a rather mindless industrial policy generalization coming from East Asian experiences, you have a rather mindless analysis of liberalization and checks and balances and good governance and rule of law coming from a misreading reading or a simplification of the U.S./U.K. experience. I think that’s the challenge.

Mushtaq Khan: And I think you’ve put your finger on it. I think part of the work we do as pluralist and heterodox economists is to disentangle and to demolish these models, which are built on hidden assumptions which are terribly unrealistic in the general case. They were true in those specific cases, but they don’t hold in the general case.

Rob Wiblin: These are super interesting case studies just to think about it at the upper level again. I think in this story, development is a mess, basically. Or it’s like, each country has to be considered individually. You need to look at South Korea, look at Nigeria, and think: “Who are the powerful people here, or the powerful groups? What are their interests? How can we nudge those interests in one way or the other? I think one reason why this lens isn’t often put on things, why people are often using these far more general approaches, thinking just free markets and trade is the only way, is that it’s just so much more work to do it this way. And you have to have so much local knowledge, familiarity with this specific country. And then you can’t just generalize and say, well, we should just copy what this other country did.

Rob Wiblin: How severe a floor is this in this school of thought? Like that you have to come up with a story or have some modeling in mind of the different power centers within a country, and then figure out what that implies, and that just seems very difficult to do. And so maybe it’s like something of a futile effort to predict things ahead of time?

Mushtaq Khan: I don’t think it’s so hopeless. I think that if you’re trying as an academic to make some analytical generalizations it’s tough. Because you have to go into a lot of the stuff that we’ve been talking about. And a lot of people who spend their lives saying something else will come and tell you you’re wrong. And you have to spend a lot of time fighting your colleagues who have invested their careers and their reputations in giving a slightly different story. I think that’s one part of the picture. But if you come to the real issue of how you do development on the ground, I think a lot of what I’m saying is really commonsensical. And you can do that with understanding what I’m saying in a very commonsensical way and not going into too many of the debates.

Mushtaq Khan: And the question at the ground level is simply this: You are in a country which is not working too well. You have a limited number of policy instruments to play with. You can’t smash everything and start again. There’s a history that explains where people are. And that history means that you have a certain number of players, organizations, politicians, political parties. You have to take that as given. You have no capacity to say, “I wish it were different.” This is your starting point. Then you say, these players are behaving in certain ways. Let me just sit here and observe how they’re behaving. And then I say, okay, I want to achieve the following. And then it comes back to what your normative goals of development are. And I’ve given you my normative goal. My normative goal is, I want to see a broader spread of organizational capabilities. Because I think that underpins almost everything else. Even if you wanted to fight poverty, I would ask how do I get poor people to have the capabilities to produce things for themselves and engage in activities for themselves, rather than just give handouts. Handouts might be essential for a bit, but the longer run solution is actually capabilities.

Mushtaq Khan: And I think there’s pretty wide agreement on that. So then the question is, this country is missing capabilities in every sector, in the following ways. And you pick some of the really vital sectors you want to work in. It could be health, it could be education, it could be manufacturing, it could be agriculture. Different countries have different mixes of priorities. And then you say, okay, capabilities are missing here. It’s not just that capital is missing and skills are missing, lots of resources are going into that. But the organizational capabilities are missing. Now, how do I support that? Who do I fund? Who do I give some resources to build that? And what are their likely responses going to be? Because I’m going to give resources to some group of people and they will have to convert that with a lot of effort into greater productivity and competitiveness.

Mushtaq Khan: And once they achieve these capabilities, my support will stop. I need to know — and they need to know — that it will stop; that they will actually achieve this so the support can stop. What are the credible checks and balances here? If you can’t figure that out in a commonsensical way, you shouldn’t go in. This is a really commonsensical way of thinking about it, which doesn’t require a massive amount of theory. It just requires massive amounts of common sense. And being there on the ground and seeing, okay, can I imagine this organization or this group behaving differently in five years’ time with my support? What would I do if I were in their place? If I was in their place and I had those networks and someone gave me the money, what would be my rational behavior? Would I try to take it and actually build the organization in the way I want?

Mushtaq Khan: Or is it rational for them to behave in a different way? That’s the 101 question. And you can do that without much analytical knowledge. And if you then get stuck, yes, you can go and read some of the stuff that people like me are writing, but I think you can go a long way without… And you know what, it’s incredible how people don’t even ask this basic question before they go in as aid agencies or as investors and plant tens of millions of dollars on a project, which if they had thought through for five seconds they would have realized it’s unlikely it will end up in the way they were thinking it will.

Rob Wiblin: It sounds like you’re saying that if you want to be a development practitioner who’s using this method to improve policy, make better suggestions to aid agencies or governments, to some extent it’s an art, but it’s like an art that you can learn and do a reasonable job at through studying lots of case studies, being familiar with examples from the past of how things have gone well and how they’ve gone badly, knowing who are likely to be the power centers and knowing how might you find out what their interests are and how influential they are. I guess knowing what kinds of questions to ask… You’re saying very similar questions show up in country after country. And I guess you’re saying you can go a long way just by asking really basic questions about things like, is it in the interest of someone who can sabotage this policy to sabotage it? Because in that case, you really don’t have a hope.

Mushtaq Khan: And it’s not just power centers. As I’ve said, it’s really even the organizations you are supporting who are powerful in terms of the activities they are doing. Powerful basically should be interpreted to mean any organization who has the capability to disrupt your policy or to deliver it. That could be, actually, on the face of it, some really not-very-powerful organizations, if those are the ones you are working with. Often policies don’t fail because someone at the top has an interest in subverting it. That does happen, but it’s really rare. Most of these policies that we use to deliver skills-training programs in developing countries for example, or health services and so on, the rents are not that big that the prime minister is concerned about capturing them. There are cases where policies have very significant rents, and then really powerful people — in the normal sense of the word, people at the very top of society — are involved.

Mushtaq Khan: Those are very tough to deal with.

Rob Wiblin: Oil, mining.

Mushtaq Khan: Oil, mining. But also big procurement programs and so on. But most programs and policies are at lower levels, where the players are powerful only in terms of that policy, but not in a national sense. Or certainly not in an international sense. It’s important to understand them as well. So, you’re funding these organizations, what are their capabilities and what do they really want? And will they actually monitor each other and report on each other if some of them are violating? Will they actually follow these rules in their own interest or not? I mean, to me, this sounds obvious. But it’s incredible how little work goes into doing that. People just listen to what people and organizations asking for support say and what their public statements are, but that’s not enough. You have to look at their behavior, not what people say.

Mushtaq Khan: Just because someone says I’m involved in these activities and I’m delivering these results… No, go and look at their behavior in the past. What have they got in the past? What have they delivered in the past? What networks do they have? How are they responding to different sorts of incentives? And it should be relatively… I mean, it’s not easy, but these are answerable questions. And if you sit and work at it for six months, you will have a good idea of the types of organizations you’re planning to work with and whether your policy is feasible or not.

Anti-corruption as a design of your policy [01:35:16]

Rob Wiblin: Okay so what kind of analysis do you want folks to actually do, in order to figure out whether the program they’re thinking about implementing is actually compatible with the incentives of the people involved, and so has a real prospect of working out?

Mushtaq Khan: So, I think we’re trying to do something quite radical here. We are trying to change what we mean by anti-corruption from that top-down, transparency/accountability/enforcement thing to something that is actually deliverable, feasible, and developmental. And that means we need to rethink what we are doing. And so what we are saying is that we are trying to achieve developmental outcomes with players who currently are not following the rules, and resources are being captured and wasted. Step one is to understand why this is happening in various sectors and activities and countries. And the reasons are quite different. Step two is to say, are some of these players behaving in unexpectedly rule-following ways? Can we see a difference in behavior? In the sector we are looking at, are some people developmental and rule following and others rule violating and not? Step three is to ask why. Why are some people following the rules better than others and providing better developmental outcomes?

Mushtaq Khan: Our hypothesis is that — and this has been borne out, and I’ll give you some examples of this — when you see this happening, it’s almost always because people who are following the rules better have the relevant capabilities and have some horizontal checks and balances going. That is they’re either dealing with other market players who are checking them, or they are supplying to other market players who are checking them, who have an interest in checking them to make sure they deliver the right things. And the ones who are corrupt do not have those horizontal checks and balances in the same sector, in the same activity. That’s the really interesting part of it. And then the next step for us to ask is, “Is there a feasible policy through which we can change the horizontal checks on the ones who are not following the rules so that they become like the ones who are following the rules?”

Mushtaq Khan: And this is a feasibility question. And then we look at the evidence of whether this can be done or not. And if it can be done, we say “Eureka, there is an anti-corruption strategy here.” And the anti-corruption strategy is you take as given the existing vertical enforcement transparency structures. Every country has courts, police and probably an anti-corruption agency. We take that as given, but they are not working because the players themselves aren’t reporting what’s going on. They’re not supporting enforcement, they’re colluding and hiding the facts. Once we get that sorted out, that system of enforcement will start working on more people than the few who were following the rules. That’s the anti-corruption strategy. Now, the next phase of it is beyond research, it’s implementation. All of our research is evidence based. We are pointing out the differences in behavior. We are providing an analysis. We are pointing out why this is important, and we’re pointing out if it is feasible to change those horizontal links.

Mushtaq Khan: What we can do as a research organization we are doing. The next step is — and this is where we are now — is to take these and sell these ideas to development agencies. And essentially what we are saying is that anti-corruption is not something you add on to your project. Anti-corruption is in the design of your project. Anti-corruption has to be in the design of your policy. And if the policy is designed in such a way that it takes into account the distribution of power and capabilities, and is resolving some of the issues forcing people to violate rules and helping the people who want to do business and make money to follow the rules, and ensuring it is profitable for them to do so, then you’re truly supporting them and supporting rule-following behavior. And these people and organizations now become your allies in reporting the ones who are not. This is a general principle, and I think we have got a lot of results, which I can describe to you in a second.

Mushtaq Khan: The second thing is that we also find lots of examples of sectors and activities where this is not the case. Where actually almost everybody is violating the rules. And we investigate the reasons for them violating the rules, and we may find they’re quite difficult to change. And then we say, actually trying to do anti-corruption on this activity right now is not going to be feasible. And I think we need to have the courage to say that, that you can pump as much money in here as you want, and it is not going to work. But then we also say, if this corrupt activity is having very negative effects on poor people, we need to find another solution. We call this an exit strategy. The first is a ‘voice strategy,’ where we are enhancing the voice of the horizontal players to fight corruption. But in some cases, there are no horizontal players who can play this function, and here we have an ‘exit strategy’ for addressing the effects of corruption in that activity.

Mushtaq Khan: Here, we need to have an out-of-the-box solution where we address some of the damaging effects of this corruption through an exit strategy, which will be mitigating the adverse effects on the poor. And both voice and exit strategies are part of our research offering. We help policymakers identify where they can actually reduce corruption feasibly, and where they need to exit from direct anti-corruption efforts and deliver the same thing or solve the problem in some other way. And finally we say there are areas which are really important, but where the horizontal checks are unlikely to happen at the moment and there may not even be a feasible exit strategy of working around the problem. The horizontal checks are unlikely to happen at the level of corruption by top political leaders in developing countries. Because here check and balance requires other political leaders who are equally powerful and networked who will check them. And these are the types of corruption that are hardest to attack until a broad distribution of power has emerged.

Mushtaq Khan: It’s not saying this corruption is okay, it’s certainly not saying that this is, in some sense, developmental. It’s just saying don’t think that by threatening prosecution on the ex-prime minister, this prime minister is going to behave differently. You should do it anyway. I think, as I said earlier, we have to have these transparency/accountability/enforcement threats going. But we shouldn’t get overly disappointed and demoralized when they don’t seem to have an effect. Because actually, there are good theoretical reasons why they shouldn’t have an effect. We should mainly focus on those feasible areas where there will be an effect. And by building on those incrementally, you are both nudging development and you are nudging these societies towards more and more rule-following behavior, which I think is an essential condition of a civil society. If you have a society where people can just violate rules all the time, it’s not a safe, happy place to live in. The rules might be good or bad, but we need to follow rules and then change the rules when they don’t work. And this is not the case in a lot of developing countries.

Rob Wiblin: I guess what you’re saying there isn’t that you’re just going to tolerate corruption at the high level forever, or expect that it can go on forever. But you’re taking this roundabout approach where rather than you just prosecute the previous prime minister and hope that that’s going to fix the issue, instead this is a many-decade process of improving capabilities of organizations, like changing the incentives of the actors such that they’re doing higher valued and complicated business, they’re doing complicated legal things. And so they want to have a rule of law properly. And then this will gradually filter out through the whole of society to the point where it’s no longer plausible that people would put up with people at the high level engaging in serious corruption.

Mushtaq Khan: And in the meantime, keep pressure on the corrupt prime minister. But the point is that you see a lot of anti-corruption activists get thoroughly demoralized when they’re told that one big push of transparency/accountability/enforcement will clean up the stables and you will become like Denmark, and then it doesn’t happen. It’s not going to happen. What actually happens through that false messaging is people get more demoralized. Then people opt out of any kind of political activity, and say, “There’s something wrong with our culture. There’s something wrong with our genetics. There’s something wrong with us. Or there is a global conspiracy. Because we’ve done all of these things, and we’re not making progress.” And I think that is really dangerous. When because of a false analysis you say this can be done, and you’re telling them to do something that cannot be done, it’s not just a waste of money. You are doing some real damage to the moral, ethical, and political fiber of this country, because people start losing hope. They become more desperate.

Mushtaq Khan: And I don’t think it’s stretching things too far to say that those people then become easy targets for extremism of different sorts who say “This is all a Western conspiracy, or some conspiracy to keep us like this. Our leaders are all corrupt because they are supported by outsiders. So we need to have this complete revolution, blow everything up and start again.” And by the way, some people even in advanced countries (e.g. Bo Rothstein) think that is needed in developing countries. They think that corruption is so embedded and collective and such an equilibrium in these countries that unless you blow it up in a ‘big bang,’ you can’t do incremental changes. And I think that’s a really dangerous idea.

Mushtaq Khan: If by incremental changes you mean those top-down enforcement efforts, we agree they don’t work. But the incremental changes that we are talking about — which are bottom up and playing on horizontal enforcement — do work. There already are players who are following rules in these countries. They need to be supported. We need to expand those spaces where rules are being followed. And saying that these incremental policies don’t work and you need to have a big bang, you need to overthrow this whole equilibrium—

Rob Wiblin: Is a misunderstanding.

Mushtaq Khan: Not only a misunderstanding. It’s deeply dangerous. Because the fallout of that will be even worse. Egypt and the Arab Spring is a very good example of this. The whole process — which was targeting the corruption of the Mubarak regime, which was a desperately unpleasant regime — didn’t achieve anything. And the poor Egyptians are back to a similar but more authoritarian and more illiberal regime. I think this is what we need to really avoid, and come up with a feasible, incremental pathway which makes things incrementally better. Not look for magic solutions that don’t work.

Examples of anti-corruption programs with good prospects [01:45:17]

Rob Wiblin: Yeah. What’s a specific example of an incremental anti-corruption program that you think will work, or has good prospects?

Mushtaq Khan: Well, so it’s not an anti-corruption program. It’s the kind of—

Rob Wiblin: The design of the incentives.

Mushtaq Khan: Exactly. Let me start with one which connects to what we were discussing about organizational capabilities directly. Developing countries are spending billions of dollars every year on skills-training programs, because we are told by employers that we don’t have skilled workers, and there’s a skill shortage. When these billions are spent on skills training, and people go and study the effects of the skills training, you find that they have almost no effect. The people you have trained are just as likely to get a job as people you haven’t trained. The treatment effect is close to zero. I mean, if you’re lucky, it’s 10%. Why is this? The first round of explanations people gave was that this is because the programs were badly designed. The trainers, who are usually private-sector companies who are paid to provide the training, don’t really take the time and effort to find out what the market wants.

Mushtaq Khan: Then countries have come up with incentive schemes, which say that if you train properly and your trainee actually gets a job, you get a bonus at the end. Or you get a part of your profit at the end. The last third of your payment is subject to your trainee getting a job. Now, you would have thought, great, good policy, it will solve the problem. But we went into… This was a fantastic piece of research we did in Bangladesh with a partner who was a delivery agency, working with U.K. Aid. And so we were all part of U.K. Aid, and so we had access to inside information. And what we found was that actually the training providers were falsifying how many of their workers actually got jobs, in order to release their money.

Mushtaq Khan: And then when we started looking outside this particular service provider, we found this problem was widespread, and people were openly talking about this as a widespread practice. We started digging deeper into why this was happening. Because the firms are saying we have a skill shortage, you are providing skills for those firms, and yet you’re falsifying numbers. We picked a sample of training providers who are doing exactly the same training. They were providing workers to the garment industry. They were all selected by the same supplier. The training providers were selected because their trainers had identical basic qualifications. They were supplying to the same industry, and the same entry-level workers, and they had the same incentive structures because they were paid by the same supplier.

Mushtaq Khan: In other words, we were controlling for all the things that could explain differences in behavior. And then we had access to the inside tracking information, and we found that some of these training providers were defrauding to the tune of 60% of the reported workers and others had all their invoices accurate. In some cases, 60% of the reported workers were not in jobs. They have to report the phone number and the ID number of these workers, so you can call them up and ask them, “Do you have a job?” And they say, “No.” But they’re reported as having a job.

Mushtaq Khan: Now, why is it that in the same industry, the same sector, some people are following the rules and some people are not? They have tried to address this problem with vertical enforcement, doing random checks. You do random checks, a few companies get caught. You cancel their contracts, but the new ones do exactly the same thing. It’s not that you can sort this out in that way. Then we went into why are these people being fraudulent? And it turns out there’s an interesting reason that nobody had looked at before, which is the demand side. Who are the firms you are supplying to? That is the horizontal link.

Mushtaq Khan: It turns out, to cut a long story short, the training providers who were supplying to clusters of firms that had high capabilities did not defraud at all, because all the trainees that they were producing were immediately sucked up and got jobs. Why? Because those firms have fast-moving production lines. Their internal factory organization is very good. If they take someone from the street, their production line slows down. So they will seek out someone who has got the basic training, because they can’t afford for the production lines to slow down. If you’re lucky enough as a training provider to be close to firms which have fast-moving production lines, all of the trainees get jobs, you’re fine. Those who are next to clusters of firms that have lower capabilities face a problem. They’re also looking for workers, but when you send them your skilled worker, their production line is moving slow anyway, so this worker doesn’t speed up their production line. They say, “Your worker is not properly trained, because I’m not making any more money. Nothing is happening faster. I would rather hire someone from the factory gate at a slightly lower wage.”

Mushtaq Khan: It took meticulous research but we found they were actually recruiting people at the factory gate who they thought were just as skilled as these slightly more expensive people, but the problem was actually their own production line. If your bad luck is you’re stuck next to these low-capability firms, whatever effort you put into training, you can’t get them jobs. These are not bad people, they’ve invested in the training. They then tell their workers, we’ve worked very hard on training you. If someone calls you, please tell them you’ve got a job. After we were doing this research for a time, people we called were saying, “No, no, I don’t want to speak to you,” because they had been told to do that by their trainers.

Mushtaq Khan: Now, what’s the solution? The solution is very simple. The solution is not that you have vertical enforcement and punishment. That’s not going to solve the problem. The problem is how do you raise the capabilities of these clusters of firms? And we know how that can be done, because it’s been done in Bangladesh. We found this example of a German development bank which combined skills training with a program which brought in consultants from Turkey, which is a more advanced garment producer, to restructure a factory. And the combination of those two things raised productivity by an astounding 30%.

Mushtaq Khan: That part of that investment, as the German development loan showed, could be a loan. It doesn’t have to be development funding in the form of a grant. It doesn’t have to be a subsidy. But you need to combine a commercial loan (or a development loan at a lower interest rate) to raise the capabilities of the firm with skills training, and then suddenly what you will find — and this is a prediction — you will find that now the firm, your horizontal partner, is going to monitor the training provider. Because if you send them badly trained workers, they will not take them. And if you send them well-trained workers, they will immediately employ them. And you have no incentive to lie, because all your trainees are getting jobs. And so the prediction is, based on step one, step two, and step three, we show evidence that some trainers are not corrupt. We show evidence that those trainers have horizontal checks and enforcement. We show the feasibility that you can raise the capabilities of those firms that are now low capability. And it is financially feasible, not through development aid, but through development lending.

Mushtaq Khan: And then fourth — and this is where it becomes a policy program — our prediction is if you did that, you would fight corruption and achieve development. Now the last bit of it is an offer to anyone who’s listening to this. If you’re an aid agency, please come up and try a small experiment with us in a place like Bangladesh, and let us do a proper study and show you, prove to you that this works. Because all we have done so far is looking at historical evidence, looking at how firms actually behave, and we’re predicting their behavior will change if we can change some of these horizontal links to specific types of horizontal organizations. Our feasibility analysis shows it can be done.

Mushtaq Khan: This is example one. And this example would not only save hundreds of millions of dollars if you could do it at a global scale, it would actually raise productivity. Because what is happening in skills training in developing countries is that countries are investing in skills, they’re not looking at the horizontal demand side of it, and these skilled people now with training are getting frustrated, and dare I say it, some of them are getting on boats and trying to leave the country. Doing the skills training badly not only doesn’t help the objective, which is to create jobs, it creates frustration, because you’re not looking at the other side of the equation. And the corruption in this case is driven by that missing analysis of the horizontal link.

Rob Wiblin: It’s interesting, in this example the anti-corruption effect is almost kind of by the by, because the whole thing is that this policy of training has been misconceived, perhaps from the beginning — at least for some of these organizations — because you’re training people with skills that are not actually needed, or can’t usefully be used by the firms that are available right there. So the corruption is kind of a symptom of this misconceived training policy. And you get rid of the corruption, but the most important thing is that by teaching these firms to make use of these higher skills, and to have much more productive and stronger organizational capability, then you’re actually going to develop, which was kind of the goal originally.

Mushtaq Khan: But it’s worse than that, because you created a policy without asking how people are likely to respond. The corruption hid the problem. Before we went into this, people were saying, “We are puzzled as to why this is happening. The incentive structure is there. The trainers are saying they have created employment. The firms are saying we are still short on workers. Productivity isn’t growing. When after a few years someone comes and does this tracer study, which is looking at the treatment effect, it turns out to be close to zero. What is happening here?” This puzzle existed because corruption hid the problem. So corruption has many effects. It’s not just that you are wasting resources, you don’t even know that you are wasting resources, because the corruption is covering up the problem. And you don’t know what the problem is.

Mushtaq Khan: It took dedicated anti-corruption evidence research to identify what was the cause of the corruption and how big it is, and that it can be solved. So, yes, you’re right that we are really going back to the development problem, but we should, because corruption is what makes us fail to achieve the development goal and hides the problem. Corruption is never in itself rather interesting. I’m not a moral crusader who is going after everyone who is corrupt. It’s the development that is being stopped by corruption that is really the critical issue.

The importance of getting overseas influences [​​01:56:05]

Rob Wiblin: So on your story of development, where it’s crucially about capabilities and building up organizations that can actually be productive and compete globally, it seems like having middle managers from Bangladesh visit other countries, like they did with South Korea in the garment industry, or having people who know how to run factories really efficiently in countries that are more developed visiting factories in Bangladesh or Nigeria or wherever, and explaining to them how they can organize themselves so they can accomplish so much more… That is like the key aspect of development, or that is like an absolutely central part of how you lift people out of poverty. Or I guess an alternative model will be having firms from more developed countries, like better organized firms, coming into a country and just scaling up the methods that they use overseas in this new location.

Rob Wiblin: It doesn’t seem like that is the key thing that people talk about as much with development, kind of this skills transfer, or managerial understanding transfer. Do you think that should be a much bigger part of the development conversation?

Mushtaq Khan: Yes, but it’s neither skills nor management. I think that, as I was just saying, you can skill people up and they still don’t have jobs. You can get good managers, but they still won’t achieve the outcome. It’s really organizational restructuring. That’s a really tough one, and it’s not discussed at all. But I think it’s not that everything is down to organizational capabilities. I think the point we are making in the anti-corruption evidence program is that it’s the horizontal checks that are important.

Mushtaq Khan: Let me give you another example, which is not so much about organization, but purely about horizontal checks. Now the horizontal checks depend upon organizational capabilities, but it’s not always reducible to just that. Here’s another example, which is really important in the context of global climate change and the massive investments that will be required to mitigate some of the effects of climate change. Bangladesh is a very vulnerable country. It suffers a lot from cyclones and rising sea levels. Hundreds of millions, billions of dollars have to be spent on mitigation activities. These include building embankments across rivers, on the sides of rivers to prevent flooding and building cyclone shelters. And Bangladesh has spent hundreds of millions, probably billions by now on that.

Mushtaq Khan: More has to be spent now. And I think some of the discussions which are going to happen now about supporting developing countries is about that. However, we also know that massive corruption happens in these investments. Transparency International Bangladesh, who is our partner in this study, estimated from a sample of studies that they did that something like 30% of the money going into climate investment projects disappears. 80% of them are badly constructed. Now this is huge. These are massive policy resources which are going into constructing these projects, which are big infrastructure projects. And a lot of money goes missing. How do we stop that? All of these have the same vertical enforcement reporting, community meetings, community boards, where all the information is put up. Transparency systems have no effect.

Mushtaq Khan: ACE looked at a comparison of embankment projects and cyclone shelter projects which had high corruption in some cases and low corruption in other cases. And this was as reported by key informants and Transparency International’s own monitoring. And our question was, why do some of these projects have high corruption and some of them have low corruption? Why are some better constructed and why are some constructed so badly when they all have the same vertical arrangements? And it turns out the difference is in the horizontal checks and balances. What we found is that in the projects that were well constructed, there was — by accident or design — a dual-use component of the investment. The embankments can be used as roads, and cyclone shelters can be used as shops, community centers, and schools.

Mushtaq Khan: Now, if you’ve designed the project properly, the dual-use benefit is quite important. And, as in a rich country, but even more so in a poor country, people are much more interested in the immediate benefits of the project than in the long-run climate mitigation which may or may not happen. When there are immediate benefits, like a road which connects up important bits of economic activity, or a school that is of benefit to the community, you suddenly find a significantly higher percentage of economically and politically more effective people in the local community becoming involved in monitoring. And so what we found is that people with above-average incomes become involved in a much greater percentage when there are immediate benefits. These people are powerful at the local community level, but they’re actually, in absolute terms, extremely poor people.

Mushtaq Khan: This is another good example of how when we say powerful, it’s always in a relative sense of that community and of what is going on. In communities where lots of people are involved, but they’re of lower-than-average power, you find highly corrupt projects. But when the more-than-average-power people in terms of landholding and economic income become involved — and remember, they are very poor people — they are a very effective horizontal check, because the contractors and others who are making it are also from their group. They’re local people who are making these things. In a context where formal enforcement is very weak, the informal enforcement of meeting someone in a tea shop and saying, “Actually the cement you’re using, the sand you’re using is not really good, I can see that, the road will not be fit for purpose,” that threat is the really important threat. That someone in your group is saying, “This is not going to work.” And when that person starts making a fuss, then the political system, the enforcement system starts responding.

Mushtaq Khan: This is a critical aspect of reality in developing countries. That those informal networks and informal pressure are much more important. So how do you get the more effective people, the more capable people — and so capabilities come in again, in this indirect sense — to get interested in this? And we have, again, a very simple answer: When you’re designing your climate investments, make sure that there is some immediate benefit in terms of dual use which brings in the local community. Not just in an abstract sense, like oh we are interested in this in the long run, but in the immediate benefits. And our evidence shows that this makes a huge impact.

Mushtaq Khan: Now, again, this is based on comparisons of differences in behavior. The policy suggestion is, let’s go and actually design some projects which deliberately build in high levels of dual-use benefits. And our prediction is that corruption will be much lower. Again, if anyone’s listening and interested in these kinds of ideas, please contact us, because we have some interesting ideas. So this is another example of horizontal checks and balances.

Rob Wiblin: The interesting thing there is that you get a lot more enforcement from people who live there, who can see the project, who care whether it’s going to actually succeed or not. And I suppose they have the clout to call the local political people or to report the corruption or the failure of the project. And so they’re in a much better position to hassle the people who they know already, who are working on the project, who are also aware that it’s less likely to be corrupt and so on. Whereas someone living in London working in an aid agency is in a much weaker position to try to enforce these things.

Actually capturing the primary effect [02:03:26]

Rob Wiblin: Both of these case studies seem very difficult to study, because there’s so many different moving pieces here. I suppose you have a team going in and doing interviews, like trying to understand what’s going on, but there might be like many different explanations that one could end up reaching, and there’s probably many contributing factors. It would be hard to know that you’ve cottoned on to precisely the right one.

Rob Wiblin: I suppose you’ve got this very plausible theory that you’ve got this local horizontal reinforcement, but you probably haven’t been able to actually listen to the phone calls where people were calling around and then making a fuss and preventing the corruption from occurring. How do you go about studying these and trying to make sure that you actually are capturing the primary effect rather than perhaps some red herring that isn’t as important?

Mushtaq Khan: That’s a really important question. Anti-corruption is difficult to study, because you have to work with a relatively small number of cases. You don’t have credible evidence on a very large scale, so these are all small-n studies — you have a small number of observations. But we can control — to some extent, not to the satisfaction of people who believe in randomized controlled trials and things like that — but we can control in some sense for other factors that might be affecting it. For example, in the embankments that we compared, we picked embankments which were almost the same in terms of value, in areas which were very similar, that are in any case quite far from the capital. We had them being implemented in the same period, so it was the same government in power. The agency that was implementing it was exactly the same. The funds came from the same source. The mechanisms for reporting were exactly the same. In other words, much of the other factors that could affect the outcomes—

Rob Wiblin: Were controlled for, to a surprising degree.

Mushtaq Khan: Yes, we always do that. We control to the highest possible extent, but we can’t do a large n where you would control by putting in variables to test in a multivariate regression, or in an RCT where you randomize all these confounding factors by picking individuals randomly. This is not feasible in anti-corruption studies, in the kinds of things that we are doing. We come up with our conclusions by saying, look, we have tried our best to control for all these things. It is still a hypothesis. Actually the proof of the pudding is in the actual testing, when we can say to an implementing agency that this is a much stronger basis on which to design policy than the basis on which we have been designing it so far, in which we have invested and wasted lots of money. So here is an idea of how you could design the next project.

Mushtaq Khan: And then if we could select a number of projects which have these characteristics, we could then compare them with the others. And then we would have an even stronger basis for saying this. But we will never get to a proper randomized controlled trial, because you can’t do it. If your project costs $100 million, you can’t do lots of them and then randomize across all the confounding variables and say, well, this particular thing, which is a dual use, really is significant. That you can’t do, because the scale of the trial would be worth more than you’re investing in the country, and so on.

Mushtaq Khan: Our evidence-based research does address your question of what are the other confounding factors, but it does it in a realistic way, given that we have small budgets, the data is very difficult to get, and therefore we need to build systems of accessing the data. And that means the number of cases we can study is small, but then we control for that, as I’ve described. In the same way, in the case of the skills providers, it was the same supplier, the same incentives, the same industry, the same qualifications of the trainers, then the same entry-level training. We have controlled for a lot of things which if you didn’t control for, you would need a much larger sample. But someone who doesn’t want to accept the story can always find reasons not to. I think the point is that we are giving policymakers who really are fed up with the poor results that they’re achieving, fed up with the poor outcomes, strong evidence-based analysis with a very plausible theory behind it and saying, you should really look at this very carefully.

Rob Wiblin: I guess it’s going to be very hard to get extremely confident that this is exactly the story, or that this is the most important phenomenon. But the comparison is not considering these things at all, or not having really an understanding of what is driving corruption. And so you’re building up more and more pieces of evidence until you can be confident enough that it’s worth trying this out, or at least worth considering these considerations in the design of the policy.

Rob Wiblin: I guess the hope is that by pursuing this research agenda over a long period of time, you can potentially learn more general lessons. One potentially more generalizable lesson from the first one might be that you might consistently expect to find that where you put in training programs that are teaching people things that actually the local industry can’t meaningfully use, then you might expect to get corruption resulting, or at least misreporting of the effects.

Rob Wiblin: And likewise, you might find that in many countries, it’s the case that if you’re building infrastructure projects that don’t benefit the people who are there, the people who can actually observe the corruption and potentially pressure people to stop it… If it doesn’t benefit them, then it’s more likely to have poor outcomes than one where the people who see it being built are directly concerned about it. It’s not as if in every case you’re going to have to go in uniquely and learn everything from scratch. You can hopefully build up a picture that will allow you to anticipate ahead of time what things are likely to go right and wrong.

Mushtaq Khan: I think there’s a simpler generalizable lesson here. The generalizable lesson here is that if you are trying to enhance developmental outcomes by making people follow rules and check each other and so on, don’t do that by relying solely on anti-corruption agencies and reporting/transparency/accountability. Ask yourself the following question: “In the area that I’m trying to work in, is there evidence of differences in behavior already existing? Can I already find projects which are well constructed and projects that are badly constructed? Can I already find training providers who are doing the training and jobs are created, and others who are not? Can I already find power stations that are built at low cost and producing greener power, and dirty power stations at high cost in the same country with the same systems?” If I can find them, then I’m on a winning ticket.

Mushtaq Khan: Then I go and ask, “What is going on here? Why are these people behaving differently?” We find that in a lot of cases, it’s something horizontal happening. It’s either that the people who are building it are being checked by the people who are using it, and they are of equal power, or it could be that people who are training are checked by the people who are interested in the trainees. That’s what is keeping the thing going, right? If you don’t find any differences in behavior, you’re in a tough situation to begin with. And if you do find them, we will say that this is the most policy-relevant way of thinking about it. Can we enhance the number of cases where this horizontal checking will be possible? And if that is feasible, and in the cases that I explained to you they are feasible, in our opinion, then you have a powerful tool on top of other tools you have. But without this tool, you won’t even discover the problem to begin with, and then you won’t have any way of solving it.

Mushtaq Khan: The other thing about this horizontal monitoring thing is that it’s not just that horizontal monitors can make a fuss. It’s also that they are best placed to reveal information. And the information that is revealed has to be of a usable form that identifies who is actually responsible. Now, in general, when information exists that there is corruption, often no one actually knows who is responsible. And everybody will deliberately create a lot of smoke and noise to prevent that from being clear. The only people who know, and the only people who can actually direct you to the source of the problem, are people on the ground there.

Mushtaq Khan: If you think of what makes crime prevention, and what makes fraud prevention and anti-corruption work in countries that we have experience of this, like the U.K., the evidence is that information coming from peers is important. This company says, “That company is not paying their value-added tax. And I know it. I can give the evidence.” But if I am not paying the tax and you’re not paying the tax — and nobody else is paying the tax — then we will all say, “Actually, we did,” or “We didn’t do any transactions,” or “The machines broke down,” or we will say something else to push away the investigation coming from above. And in fact, the investigation from above knows that everybody is violating, so they can’t catch them all. They’ll just pick up small amounts of money from all of them and go away. In other words, the agency is then corrupted.

Mushtaq Khan: If you don’t want any of that to happen, you have to find areas where it’s in the self-interest of the actors to say, “This is not right.” “You know, I am actually doing my work as a skills trainer. I’m getting the job done. And those companies are reporting that you’re not getting the job done.”

Mushtaq Khan: Now we have a huge coalition against that trainer. So when the enforcement agency comes and says, “There’s fraud reported here” — and the enforcement agency doesn’t have to be the government, it could be the development partner, right — they will actually have access to the information, because everyone will give it to them. In the same way that a local contractor who is using substandard cement or sand in making the embankment, if the local community is saying, “Actually, I know this is the case because I’m in the same social group as the contractor, I also build things, I know this is terrible stuff, it’s not going to work…” It’s very difficult for the enforcement agencies to say, “There is not enough evidence. This is rubbish.” It’s very difficult for that person to say, “It’s rubbish.” So the tendency for collusion is checked.

Mushtaq Khan: It doesn’t necessarily mean it’s going to work 100% of the time, but you have much more effective pressure now. Because information is out there. These people have an interest in following up on this information, making sure that it is followed up on. Because another common thing that happens is that something is reported and then the follow-up doesn’t happen. But now you have somebody who’s going to keep on saying, “What’s happening here?” And that’s extremely important. If you think of how things get done by the British police, if there was no one following up and saying, “What is happening here?” If no one was saying, “This is the evidence. This is the information. Why aren’t you acting on it?” The natural human tendency is to have an easy life. It’s the same.

Mushtaq Khan: On top of all the collusion and so on, a lot of it is this difficult investigation. You need someone who has an interest in chasing it. Another way of asking is, “Who is interested in your anti-corruption strategy? And why are they going to pursue it in their own self-interest?” And those people must have the same social, economic, political parity with the people you are chasing. And that’s what we mean by power, right? It’s not that they have to be the most powerful people in the country, but they have to be peers. If your peers are not checking you, it’s not going to work. And I think this is a really fundamental, common, generalizable lesson. Look for peers who are interested. If you can’t find them, you have a non-starter.

Rob Wiblin: Maybe do something else.

Mushtaq Khan: And if you can find them, ask them why some of them are monitoring and some are not. And then say, “Those who are not monitoring, can we change their incentives in a feasible way so they start monitoring?” And if you can, you have an anti-corruption strategy.

How less developed countries could successfully design subsidies [02:15:14]

Rob Wiblin: What have you learned about how a country like Bangladesh or India could design industrial policy, or subsidies for firms or industries, in such a way that it actually fosters increases in productivity rather than just being captured?

Mushtaq Khan: Many ways. I think one of them was the garment industry example, where a subsidy was given to a foreign company to come and transfer organizational knowledge to a local company. I think that this is a really different story from saying, “Let’s get foreign direct investment.” When you just attract foreign direct investment, the foreign investor is an island. And they might be producing things, but it’s very difficult for locals to copy what you are doing. The gap is often too big.

Mushtaq Khan: Successful industrial policy is not just about getting foreign investment. It’s getting a foreign company — which is possibly moving out of a certain market — to transfer their organizational know-how to you. And giving them sufficient incentives for them to do this in a way that can’t be captured. And I think in countries where you have political settlements where you have dense horizontal networks between politics and business, where the higher levels are politically weak because they face a lot of contestation and therefore have a short time horizon, the South Korean type of industrial policy where protection or subsidies are provided in advance and firms are then expected to achieve competitiveness is very unlikely to work.

Mushtaq Khan: The garment industry example in Bangladesh is a very good one. Another very good example of a strategy that may work in the more typical developing country is the Indian automobile industry in the 1980s. India tried to have a domestic automobile industry, and it produced this car called the Ambassador, which was a copy of an English car from the 1950s. But it was under very high tariff protection. It was globally uncompetitive, but it was a car. And the Indian consumer had to buy from between two or three brands. That’s all the choice they had. But this car was never going to become competitive. Industrial policy in India didn’t create the compulsion for the producers to continuously raise their productivity and improve quality and price to become globally competitive. Industrial policy achieved the production of a car, but then stopped there. And the subsidies keep coming in the form of 100% or so tariff barriers.

Mushtaq Khan: With that kind of tariff barrier, you’re effectively getting a massive subsidy from consumers, who are paying much more for this car than they needed to. Now, that system came unstuck gradually, because of a number of reasons, because subsidies started running out and consumers started complaining. But something really interesting happened in the auto industry in India which made them actually achieve a productive transformation. And that, again, was an accident. The accident was that in the 1980s, Prime Minister Indira Gandhi’s son Sanjay decided he wanted to make an automobile factory in India. And he set up a company called Maruti. But he had no idea of organization. There was a big empty factory. Nothing much was happening there. It was a real embarrassment for the Gandhi family. And then he died in a plane crash, and a large plant was left empty, with huge reputational loss for the Gandhi name.

Mushtaq Khan: His mother decided she would have to make it work, because people were laughing about it. This man set up this factory and nothing happened. It’s just an empty space, no organizational capabilities. She nationalized it. It became a public-sector company, and then she set her best bureaucrats on it to make it work. Get some foreign partnership. Get some foreign investment going. Nobody was interested. Because they realized that they were going to compete against very powerful domestic interests, and there was no guarantee that this investment would work. Except one small producer called Suzuki, which was mainly a motorcycle producer, had a very small line in automobiles. But they had worked in India for a long time, and they were trying to enter the motorcycle market. And they knew that the government was serious. And they knew that they really needed to make this work.

Mushtaq Khan: They sat with the government, and they came up with a deal. I have written papers on this, and many others have as well. The deal was broadly the following: Suzuki would have to come into partnership with Maruti, and would produce a car called the Maruti Suzuki. It would be sold in the Indian market. It would benefit from this massive tariff protection. But the quid pro quo was that 50% — or something of that order of magnitude, 60% — of the value added would have to be domestic. In other words, most of that production would have to be Indian, and they would have to achieve that in a set number of years. I think the number was five years. Now the price is ex post. Suzuki is getting this price after it produces the car, not before. You’re not giving the money to Indian producers and saying, “Here’s the money. Become competitive. If not, we’ll take it away.” That’s not a credible threat.

Mushtaq Khan: But here the money is being given not even to Indian producers, but to Suzuki. Now Suzuki is coming in… If you want to produce a car with high domestic value added, you have to use domestic tier 1 and tier 2 producers. So a car is actually produced, not in the final factory, but in what are called tier 1 and tier 2 producers. The tier 1 makes all the bits that go into the car, like carburetors and gearboxes and engine boxes. And then the tier 1 buys inputs from the tier 2, who make the bits which go into the tier 1. And then there’s a tier 3, tier 4, and so on. Most of the car is actually built in the tier 1, tier 2, and so on factories. The final assembly is just assembly. Suzuki realized it would have to work with the Indian-owned tier 1 and tier 2 companies, which already had some capacity because they knew what an Ambassador was. But they had never made parts of the quality that would need to go into a Suzuki-branded car.

Mushtaq Khan: And here’s the trick. Suzuki did not want to produce a rubbish car, because that would destroy their brand name, which would have a much bigger effect than any profit in the Indian market. So here you have this great combination of incentives. Suzuki has to build a high-quality car. It has to work with Indian producers, and it’s going to get a big prize at the end. That’s exactly what happened. They sent armies of people out, with Japanese just-in-time organizational technology. They worked with the tier 1 and tier 2. The tier 1 and tier 2 people were told, “Actually, you’re not getting any money upfront. It’s our investment. Our engineers are coming here. They’re going to tell you how to set up your factory, do the quality control, do the inventory management, meet the specs.” It’s not that they didn’t know the technology. They didn’t know how to do it with the specs of a modern high-quality car.

Mushtaq Khan: And the Japanese were saying, “We will teach you how to do it at our expense. The prize you get is that you get to sell to us. You become a Suzuki supplier.” Which is worth its weight in gold, because as a Suzuki supplier you have the international market to sell to. They all jumped at it. Because they knew that the effort would be converted into profits in a new market. Suzuki knew that its effort would be rewarded with the rent, and the rest is history. Suzuki monitored the effort of its tier 1 and 2 partners, and the latter made sure Suzuki wasn’t wasting their time. Maruti captured the entire Indian market almost in a couple of years. Just recently, a couple of years ago, the Ambassador went out of business. And then the Suzuki experiment was repeated with a number of other foreign investors who were coming in. Each of them were given the same terms.

Mushtaq Khan: And very soon, Indian tier 1 and tier 2 suppliers became the global standard. They started winning Deming Prizes and other international prizes for engineering. They became exporters in their own right. And then finally in the 1990s, India began to manufacture its own branded cars. And it could do that because it had Indian tier 1 and tier 2 producers. Because global tier 1 and tier 2 producers do not want to deal with small local national companies, because they have reputational loss and other kinds of issues and contracts with the big brands which prevent them from doing this business. India became one of the few countries in the world which has its own branded cars because it had built locally owned tier 1 and tier 2 producers.

Mushtaq Khan: Thailand produces the same number of cars, or roughly in the same ballpark. But it doesn’t have a Thai-branded car, because the Thai tier 1 and tier 2 are also multinationals. Whereas in India, these firms are Indian owned. This is a great example of how organizational capabilities can be transferred with appropriate incentives, and where the policies work given your distribution of power.

Mushtaq Khan: Because the prime minister was strongly interested in achieving this, she could take on the auto producers of India who were against this tooth and nail. And she did this by giving Suzuki an offer which looked really tough. “You have to achieve 60% domestic content in five years.” Which no one could say is anti-national, because you’re actually building… It’s just anti-national against those relatively uncompetitive auto firms which were eating up all the subsidies. By saying that we are actually building a national industry, and by doing it in this way, she could undermine that process of blocking. Now, here is another actually good lesson. If you want to do industrial policy in a country which is very clientelist, focus on one thing at a time. It’s industry by industry, problem by problem. And then focus all your attention on solving that.

Mushtaq Khan: In Bangladesh, it was the garment industry. Here it was the auto industry. And then they had a similar story in the pharma sector, which I’ve also written about. And these were all organization and technology transfer stories, which were not part of some big ex ante policy where we give companies lots of money in advance and hope they build their capabilities. And if they don’t perform, we threaten to take it back, which is a South Korean strategy. Here it was, “You will have opportunities of making money if you solve these capability problems, and we will help you connect with foreign suppliers. We will create the legal environment for it, but you’re not going to get any money upfront.” And I think that these are two extreme examples. I think a mix of the two would work in many countries. But I think figuring out how to build capabilities in contexts where you have very specific starting points in terms of capabilities and political settlements is important.

Mushtaq Khan: The Indian strategy, for example, would not work if you didn’t have tier 1 and tier 2 producers who had some capabilities already. It wouldn’t work in Ethiopia, or somewhere else which is trying to do industrial policy. You have to have a good understanding of your starting point, and you have to give a deal to a technology provider, an organizational knowledge provider. Suzuki knew the capabilities of the Indian firms, and knew that it was possible to raise them from point A to point B, and the prize it was getting was worth it. If you ask for someone to do something which is not profitable for them, they’re not going to do it. And you don’t have the capacity to enforce it yourself. So I think the lesson that comes from this is that feasible industrial policy requires solving these organizational capability problems, investment problems, technology transfer problems, etc. in ways that can be policed by the self-interest of the players, to a large extent, and then only marginally by external force. And that is just as true for industrial policy as it is for anti-corruption.

What happens when horizontal policing isn’t possible [02:26:34]

Rob Wiblin: If you’re trying to combat corruption but it’s just not practical to get horizontal policing by peers, are there any other approaches that might work a decent fraction of the time?

Mushtaq Khan: No. I think that’s a very important question. I think in a lot of areas you will not find — within the same activity — forums or organizations that are following rules and being developmental, and others that are not. And therefore this approach of building the horizontal links that make people follow rules might not be present in a strong way. And that I think is quite common in many areas in developing countries. And when that happens, I think our analysis tells us there are no simple solutions. We have an analysis of such sectors, and I’ll give you an example or two where that is not the case. And then we say that the standard vertical enforcement strategies from above will definitely not work, and might actually make things worse. And so there we have to find innovative ways of exiting from the problem by mitigating the damaging effects of corruption and finding solutions which are completely out of the box.

Mushtaq Khan: We have this fantastic project in the Niger Delta in Nigeria, where you have this problem of artisanal refining, which is basically the local communities stealing oil from pipelines. And then they use that in local, very crude refining, which they then sell on. Now, this is a massive problem of theft, which has huge environmentally damaging consequences, and creates health problems which are very severe. You may even have seen pictures of the Niger Delta and the oil spills that are happening there as this oil is stolen. And the Nigerian government has tried year in and year out to stop this problem with enforcement. The army is sent in, and the army gets corrupted. Enforcement agencies are sent in, and they get corrupted. Sometimes they actually attack local communities, and civil war breaks out. The Niger Delta is close to or in civil war-type situations often because of these kinds of activities.

Mushtaq Khan: We went in and looked at what was going on here. And it was a really risky enterprise. We had local partners who were extremely well connected, and that’s why we could do this research. And we found that, actually, this is a classic example of networked corruption, where almost everybody is benefiting from this corruption and there is no rule-following behavior at all, because the local community has no alternative methods of employment which give you anything like the returns that they get from this artisanal refining — which is a very nice term for what is going on. And the returns that they get are often higher than even extremely well-paid jobs in the Nigerian army, for example. The local community also has lots of externalities. The service sector activities, which set up hotels and restaurants, different kinds of services run by women, you have trading activities…

Mushtaq Khan: Then the refined petroleum is sold on for generating electricity. And so it also is powering houses. It’s a complete economy, which is working in what we call ‘networked corruption.’ Now here you don’t find any insiders who are following rules. No one will report on anything. No one will support enforcement. Even though their benefits are unequal, the whole community is against any kind of external enforcement activity, and external enforcement generates violence and generates almost civil war-type issues. What do we do there? Our answer is, do not try vertical enforcement here, but try to find exit strategies from this problem. Here is a case where actually development is the answer. You have to find parallel employment opportunities for people. You have to improve health services and primary health clinics. You need to bring in solar power, so that the people who are buying this refined petroleum have some alternative sources of electricity, and so on.

Mushtaq Khan: And we could have a whole list of things which we think are feasible in that area which basically reduce the dependence of this community on this entrenched networked corruption. And so to us, that is the anti-corruption solution there, that alternative process of development, or just taking the steam out of this hothouse activity of artisanal refining, which can be done, and is happening at the margin because some solar power is coming in, some alternatives. We want to accelerate that process. We should create development opportunities, which will make it more possible for people to start exiting. This is an example. Now I think there are lots of such exit strategies which we identify as well, because a lot of activities in developing countries are like that, where the corruption is so entrenched that almost everybody in that activity is corrupt.

Mushtaq Khan: They might be corrupt for different reasons, but they’re all rule violating. And so when you come in to enforce the rules, you find no support. I think another way of thinking about it is that anti-corruption works when there are people who are violating rules for reasonable reasons, and there are other people who are violating rules for unreasonable reasons. And if you can identify the people who are violating the rules for reasonable reasons and solve those problems, then they become your allies, and then you minimize the number of people who are the true thieves, who are the small subset of people who are genuine thieves and free-riders. And when you can bring them down to a manageable number, like 10% or so, then your vertical enforcement starts working, because you can begin to also rely on horizontal monitoring and enforcement.

Mushtaq Khan: Because then the peers start saying, “Well actually, we are all following rules. Who do you think you are that you can go to the front of the queue? Go back and stand at the back of the queue.” Exactly like what happens when you break a queue in London. It’s not the police who stop you, it’s your peers who say “Hey, please go back to the back of the queue.” If your peers are not saying anything, if you’re so powerful that you can go to the front of the queue and everybody else has violated something so they’re all going to keep quiet…

Mushtaq Khan: I think the mobilization of horizontal voice is therefore one strategy. In areas where you have reasonable people who are forced to violate rules, you can address that problem through all kinds of measures that make it possible for honest people to operate without violating rules. And they in turn become the horizontal pressure on the real violators.

Mushtaq Khan: But there are other areas where the situation is so bad that almost everybody is violating the rules, and you’re not going to get… There’s no one you can say is reasonable or unreasonable. They’re all equally unreasonable or equally reasonable. In the Niger Delta, they’re all actually reasonable thieves, because they have no other alternatives. You can’t say there’s a bunch who are free-riders. They’re actually all equally reasonable or equally unreasonable. And there, you say, there’s no immediate anti-corruption solution now. What you have to have is a development solution, an exit strategy that looks for out-of-the-box solutions. And I think it is really important to be able to distinguish between these different types of anti-corruption approaches, otherwise you lump everything in the same box and everything then fails.

Rule of law <--> economic development [02:33:40]

Rob Wiblin: There’s this general long-standing view that it’s rule of law and preventing corruption that causes economic development. But I guess you want to stand this on its head — and I guess this school of thought stands it on its head to some degree — and say it’s as true, maybe even more true, that development causes reductions in corruption and the development of rule of law.

Mushtaq Khan: The causality is in both directions, but the stronger causality is from development to the rule of law. But there are very important subsets of policy and development which are being blocked by the absence of a rule of law, which you need to address in these sectoral rule-following behavioral ways, even though you can’t achieve a full rule of law. What we are saying is in the absence of a rule of law, it doesn’t mean it’s fine to have corruption. It means that the corruption is blocking critical development in some areas which you’re not going to solve by a rule of law, but you can solve by sectoral rule-following behavior by getting the peers in that sector agreeing to, in their own self-interest, to police each other and to deliver some results. And so this is different from either saying, “Do the rule of law first, or forget about the rule of law and don’t do it.”

Mushtaq Khan: We aren’t saying either. We are saying, if you try to achieve a full rule of law, you will fail. But if you ignore corruption and rule-following behavior, you will also fail. You need to find these incremental intermediate solutions where you are nudging society sector by sector, activity by activity towards rule-following behavior in that activity, until such time that you have so many people who have such high capabilities that you have all of them collectively saying, “We have a general rule of law.” And anyone who violates even one of us on anything, we will not accept it. That’s a long way away.

Rob Wiblin: What would you say to someone in the audience who was skeptical of that view, that it’s actually primarily development that’s causing the rule of law to develop rather than the reverse? The idea that rule of law is a precursor for development is a very mainstream theory, it makes some intuitive sense, which is why a lot of people have bought it. Are there any key facts that you can point to that show why it’s actually more of the reverse?

Mushtaq Khan: So, if you take my definition of a rule of law, the rule of law means that you have a situation where anyone who is caught violating a law has roughly an equal probability of punishment. If you find me a country which is classified as underdeveloped or least developed where you have that, then I will say I’m wrong. But there’s not one. I think that what you find is that… And I think this is also at the heart of North, Wallis and Weingast’s Limited Access Orders analysis, I think that there are some commonalities here, because what we are saying is, look, it’s really ahistorical to say that you can have a rule of law anywhere you want.

Mushtaq Khan: If you have $8,000–$10,000 without counting oil or natural resource rents, it means you have a broad spread of capabilities in your society. Those societies are likely to end up — gradually, it’s not necessary, but they have the distribution of power in that society — to end up with a rule of law. The converse is not true. I can’t find a society, and maybe I’ve missed one or two, where you have very few capabilities, very few sectors that are globally competitive, and yet have a rule of law, which means that anyone who is breaking a rule has an equal probability of being reprimanded or punished. That tells me that there is something happening here which is quite systematic.

Rob Wiblin: I guess some people would say, “Well, you don’t find that because if they had a rule of law then they would have become rich.” But I suppose the idea is that if that was the main channel, then you would expect that there’ll be some countries that have advanced a lot in rule of law, but haven’t yet become economically developed. But in fact you don’t really see that. There’s at least substantial effects in both directions.

Mushtaq Khan: I think they improve in step. I think there’s a lot of evidence of that, but I think that the issue really is, where is the stronger direction of causality? And I think this is, ultimately, an unnecessary debate. Because I also agree that in many cases development is being stopped by rule-violating behavior. The only difference of opinion is whether we think that we can solve this problem by a general good governance reform that gives you rule-following behavior, or whether you have to do it sector by sector, activity by activity finding spaces where people have the self-interest to regulate themselves. And if that develops gradually into a rule of law, I think that’s the real debate, whether the transition to a rule of law is an event or a very slow process, not this “What comes first, rule of law or development?” Because I think everybody agrees that they tend to move together.

Violence [02:38:31]

Rob Wiblin: You mentioned Douglass North, John Wallis, and Barry Weingast, who are other people involved in institutional economics. I was very tempted to bring them in, because I think actually one of my favorite interviews or podcast episodes of all time was an interview from 2007 with Weingast about the book that eventually came out called Violence and Social Orders. I guess the way that they think about things has a lot in common, but it’s a little bit different. In the book they’re contrasting, I think it was open-access orders, like the kind of the thing that you might find in the U.K., with ‘natural states’ in less economically developed, less rule-of-law situations. Their model is that there’s different interest groups that have different abilities to do violence to others, to cause trouble, to potentially create a civil war and potentially win or oust people.

Rob Wiblin: And you end up with the rents of a society, the profits that can be extracted, getting split up between these different interest groups in some sort of proportion to their ability to cause trouble and cause violence and menace other people. And so you end up with a sort of equilibrium, where violence is the key issue that’s driving social structure. But you don’t necessarily have to have that violence used, because you end up at a settlement where people would rather just extract a particular fraction of rents and have a peace treaty between these different interest groups.

Rob Wiblin: In your work, at least as far as I’ve seen, violence isn’t a central theme. You talk about influence, and enforcement, and so on. But why is it that they focus so much on violence as a key driver in social structure, and you think about it a lot less, or talk about it less?

Mushtaq Khan: I really like their framework, although I disagree with it in important respects, and we have actually worked together. In their book called In The Shadow of Violence… My collaborator at SOAS Pallavi Roy and I have articles where we try to show how the political settlements framework that we use is actually not very dissimilar to the limited access order framework, but there are also important differences. The common feature is that we are both looking at the social orders of developing countries as organic things, which are in some sense in an equilibrium. It’s not a pathology, it’s not a disease that they don’t have good governance. It’s not some evil corrupt people who are driving that, and if we got rid of them everything would be fine. It’s a structural, systemic design feature of how these societies are put together, with a lot of informality and a lot of rents.

Mushtaq Khan: I think this is a common thing. And just like in the limited access order, the political settlement says something quite similar, which is that the distribution of rents has to be roughly proportionate to the distribution of power, otherwise the system goes into disequilibrium and then it adjusts back to some kind of equilibrium. So what’s the difference? The difference is that my approach puts a lot more emphasis on economic organizational capabilities and how they develop. It’s not just about violence potential, but it’s also about productivity potential. And I think that one limitation of their framework, as I see it, is that they don’t actually look very much at this interdependence between political power and economic capabilities. Economic capabilities develop in their model, but through some process that is not clearly specified.

Mushtaq Khan: Whereas for me, even more important than your violence potential is your productivity potential, your capabilities of producing, and where they come from. And those two are always interacting, right? That’s one difference. And I think that the process through which policies that are not aligned with the distribution of power are blocked is not always through the threat of violence, although that’s a good metaphor. Ultimately if things break down you have violence, which is often broken down by lots of prior steps, which are far from violence. It could just be, “I’m not going to follow these rules. I’m just going to violate them, and let’s see what you can do with it.” And if so many people are violating, you can’t do anything, even though they’re far from threatening violence. Or “I will push you and try and remove you from power. And this removing you from power might be a nonviolent process.”

Mushtaq Khan: It’s a whole load of oppositional tactics, which do not necessarily reduce to violence. But it’s to do with organizational capabilities. And your organizational capabilities might be political, and you might have very strong organizational capabilities and networks which block things. And then that’s your way of getting rents. Or your organizational capabilities might be productive, in which case you want to actually use that to produce things, and make money doing it that way. This is one big difference in a broader understanding of capabilities, beyond just violence potential. But I think violence potential in extremists is an important capability as well, but it’s not the only one.

Mushtaq Khan: The other difference is that, in their framework, the rents that are generated to keep the glue that keeps violence from emerging are created by limiting access, right? Hence the term ‘limited access order.’ They’re still thinking very much like neoclassical economists, where rents are created by restricting markets and creating monopolies. And in my view, we have a much more expansive understanding of rents. Rents are not just based on limiting access to markets. Rents are based on all types of policies that generate resources that would not have existed without the policy. Some rents can actually open up markets, like these learning rents. And the rents which attract investment are rents which actually are opening up markets which didn’t exist before. And so if you are only thinking of rents as extra incomes created by restrictions, you have a real tough time explaining the political reality of developing countries. Because actually, if you look at a typical developing country, it’s relatively very easy to set up organizations, including informal organizations.

Mushtaq Khan: No one is actually saying, “You can’t set up this organization; only those powerful people can set up organizations, and they will then corner the market and get rents.” In fact, it’s the reverse. The problem in a lot of developing countries is that entry is too free. Everyone and their dog can enter the political market. You have all kinds of crazy parties and organizations jostling for power. And in a sense, that is a source of instability. Actually, limited access orders, as they describe it, are not very limited access at all. It’s quite open access, and the rents are created in a variety of ways. Just as there are lots of open-access activities in developing countries creating rents, I think I also disagree with them that in advanced countries rents are always competed away by new organizations coming and challenging those rents. Advanced countries have long-lasting and persistent rents, and they keep being reproduced because the distribution of power reproduces them.

Mushtaq Khan: These are policy rents which range from agricultural subsidies in Europe to protection for different industries and protection for policies like patents and so on, which generate rents, and go on and on. The only difference between advanced countries and developing countries in my view is not that one is limited access and one is open access. One is rule of law and one is rule by law. In other words, to me, the difference in advanced countries is that the process of influencing and jostling for rents happens through the powerful organizations following rules and setting rules which are to their advantage, and those rules give them lots of rents. Like Google has massive rents because the rules allow them to get that. They’re not violating rules to get those rents, but anyone who says Google doesn’t have rents has a different definition of rents than the one I use.

Mushtaq Khan: Google’s rents don’t go away, because the influence they bring to bear on policy protects them. Those rents add to the U.S.’s prosperity, which benefits the political decision makers. But it’s all happening within the context of a rule of law. I think it’s misleading to say that the U.S. or advanced countries are open access in that all rents are continuously being competed away because everyone can set up an organization, just as it is wrong to say that in developing countries people are constrained from setting up organizations, and that is why some people have rents, and that is what keeps societies together. I think that’s misspecifying the problem. I think both advanced and developing countries have different types of persistent rents and rents are subject to different types of competition in both. In both, it’s technically quite easy to set up organizations, but not everybody can in either order.

Mushtaq Khan: And whether or not you set up an organization, you may not have the productive capabilities or the organizational capabilities to really make those organizations count. Anyone can set up a Google, but it won’t be a Google, because they don’t have the organizational capabilities to set up another Google. The issue is that the distribution of organizational capabilities is not uniform in any society, but it is perhaps more broad based in advanced countries, which is why there are a greater number of powerful organizations which check and balance each other, which is why this whole process of rent-seeking happens in a more rule-following way. Whereas in developing countries, that process is driven by organizations which have low capability, are less interested in production, and don’t need a rule of law. So a lot of these negotiations are happening in rule-violating ways.

Mushtaq Khan: To me, that’s the big difference in their political settlements, not that there are no or not many rents in advanced countries, or that they’re open access, and there are lots of rents in developing countries, and they’re limited access. I think that’s wrong framing. But where I agree with North, Wallis, and Weingast, and why we collaborated with them — and I think they’ve done a great service to political science and political economy — is that they pointed out, just like we’ve been pointing out, that the absence of good governance in developing countries is not a pathology. And it’s not a conspiracy of a few people. It’s a systemic feature of these societies. Just for that, I celebrate their work.

Rob Wiblin: It’s possible maybe that they’re focused a bit more on violence because I think perhaps a bit of a larger part of their agenda was explaining the pre-industrial world. They talk about that a lot, if you were trying to explain England in the 15th century, then I think direct violence was more likely to be used than these more subtle things like civil disobedience perhaps, or just trying to get an elected official removed and running a PR campaign against them or something like that.

Mushtaq Khan: Absolutely.

How this applies to developed countries [02:48:57]

Rob Wiblin: Let’s move on now to talking about developed countries. I’m quite interested in this because some listeners work in development and trying to reduce global poverty, but many work in policy areas, economics, and political science in rich countries. They want to provide good policy advice across a range of different things, like climate change or solving whatever other problems in the U.S. and U.K. If they want to get more out of this framework, understanding what questions they should ask about the policies that they support or are advocating for in a country like the U.K., how should they start developing the ability to apply this political science or political economy lens onto policy questions?

Mushtaq Khan: That’s a great question. I think it depends, really, both on one’s political economy frame of analysis, but also on one’s normative objectives. Why are we doing this analysis? Again, I come back to my own normative priors. I want to see a more egalitarian society where organizational productive capabilities are broad based. I think one of the problems that has happened in advanced countries, despite being rule of law and rule following, is that many economic policies in recent years were driven by people who were powerful, and who perhaps even thought that these policies would be of general benefit, but in fact, they were not.

Mushtaq Khan: Let’s think about how we ended up in the present situation, where across the developed countries you have the growth of populist — and in Europe, outright fascist — movements, which are gaining strength. You have the experience of Trump in the U.S., which almost brought about an end to the rule of law. It was really very close. And you have other such phenomena across the world. The reason for that, I think, is that liberal economic policies made a lot of sense on paper. And it didn’t discriminate between where the capabilities were being developed. By allowing a lot of globalized production, and not giving enough attention to what was happening to the capabilities, for example, in the North of England where the old industrial working-class areas were, it didn’t ask, “What will happen when we lose these productive capabilities and all the political power that comes with it?”

Mushtaq Khan: The political compact that described Western Europe and the U.K. in the 1960s was an implicit contract between the working classes and the capitalist classes based on the fact that the rich needed the poor for their production. And the quid pro quo was that there was going to be a welfare state. There would be a distribution of health and education which was fairer, and gradually, everyone would become middle class and prosperous. That process was undermined by a liberalization that was driven by the financial sector and by what is now crudely known as the ‘metropolitan elite,’ who saw liberalization as immediately beneficial even though production of a lot of things moved outside the country.

Mushtaq Khan: Now, I’m not saying that was wrong. I think that should have happened. But what should also have happened was a set of policies to protect and reenergize capabilities in the sectors that were losing jobs, and that didn’t happen.

Mushtaq Khan: The per-capita income of these countries increased, prosperity increased, but the distribution of incomes worsened dramatically. A lot of people were permanently unemployed, living off handouts. As that happened, the political settlement also changed. Because the rich began to realize that they no longer needed the poor of the rich countries for their own reproduction. The production was all happening in China and the Far East. They were no longer willing to pay such high taxes for health and education and public services. There was a gradual squeeze on the amount of redistribution that was happening. There was a double whammy. I think that the poor both lost their jobs and they also began to have the welfare state fraying at the edges, because that political quid pro quo that the rich of the rich countries needed the poor of the rich countries to be happy was no longer the case. You don’t care anymore. You don’t need them to not go on strike, because they’re not producing anything.

Mushtaq Khan: The price for that was heavy. The price for that was the growth of populist movements and so on. What we have is a very strange political space now, where the poor of the rich countries are mobilizing around parties and slogans which are not theirs historically. I mean, the Labour Party in the U.K. finds it extremely difficult to get working-class support, whereas the conservatives can, bizarrely. The Labour Party is now getting support from the metropolitan, highly educated people who want globalization, who like Europe, who like integration. This is an amazing paradox.

Mushtaq Khan: I think that the challenge is that we really didn’t understand this interdependence between power, capabilities and policies. The problem in advanced countries is different from in developing countries, but this close relationship between power capabilities and policy is important. I think that from a progressive perspective, if we want to go back to a situation where we have politically progressive inclusive policies, we have to find a way of connecting that. We can’t close ourselves off from the world and say, “We will put up tariff barriers. Everyone else goes to hell.” Which is what Trump began, but Biden seems to be continuing to some extent.

Mushtaq Khan: We have to be globalized and open. And just as we have to be concerned about developing capabilities in the developing world, we have to be concerned about developing or preserving capabilities in our own advanced countries, and building those capabilities in those areas which have been left behind. The old traditional mechanisms of industrial policy and so on may not work. We need to find much cleverer and smarter ways of building capabilities in areas which have been left behind so far that they’re not going to benefit from all the science and technology spending that these countries are thinking of. They really need more mid-level support for starting up mid-technology activities.

Mushtaq Khan: I’m not sure we’re thinking very actively about that, or what the politics behind that is, and how those resources will actually be delivered. But I think we are at a stage where we can start talking about these questions. I think those questions are now really important. And how you actually use policy to create or support the distribution of power that is progressive, and do it in a way that will not be blocked by those who are already powerful. And navigate those incremental steps towards that, that bit of it is no different from developing countries. I think that we have to find those paths in advanced countries too.

Policies to help left-behind groups [02:55:39]

Rob Wiblin: There’s a lot of this that I don’t know very much about, but I recently read about the long history of attempts by the U.K. government to try to foster development in the North of England. I guess I don’t know exactly how much was spent, but it’s been an issue since the 1980s, or maybe even before, that this area seems to be falling behind. Government after government has run programs, but it seems like most of them have not… Well, certainly, overall, the goal hasn’t been accomplished. I guess I’m not sure where to go with that. But do you have any ideas for what actual policies might work in practice for helping left-behind groups? I suppose it sounded like industrial policy, but with the modesty and understanding that, in many of these places, you need to be aiming at mid-level technology work where they might be able to be globally competitive.

Mushtaq Khan: I think policymakers still work on the assumption that if you build infrastructure, and a good rail link, and lots of roads, then businesses will go to these left-behind areas. The problem is exactly the same as in developing countries. Businesses will not necessarily go to those areas. The wages might be lower, the connections might be quite good, but the capabilities are missing. Despite the advantages and subsidies, you actually make less money. I think the same thing that would be true for developing countries is also true for left-behind parts of advanced countries. You have to ask, “What is the existing distribution of capabilities here, organizational capabilities? What kinds of firms can local people actually set up to benefit from this new road connection, or this new rail connection, or this new port that we are building?”

Mushtaq Khan: You might get a very depressing answer. You might find that actually those capabilities have been undermined by years of misuse and under-use, that people have forgotten what is a productive process. You have to then ask yourself what are the likely things that you could do. Those might not be the old manufacturing activities anymore. It might be completely new: service sector activities, or global chain activities, or whatever. But I think some thought has to be given to what might be those activities which are value generating, which are close to or reachable by where people are currently, in terms of their organizational capabilities, and then, what do we need to do in terms of policy support to build those organizations? They have to be local organizations which are driven by local people, not a few outsiders coming in for tax breaks and employing a few people and going away after a few years when the tax breaks disappear.

Mushtaq Khan: I’m not sure people are doing that. I think that they’re looking at ways of attracting investment, wherever it comes from, not looking at the distribution of capabilities locally, not looking at where the starting point is, and not asking how do we get from there to the next step, which might be a very different path from simply making the road and giving a tax break to some investor from somewhere else to come in and set up a firm. That’s better than nothing. But I think we could do a lot better. Because the real change in the distribution of power comes when lots of local organizations with local capabilities set up. They create the local environment, the ecology of the checks and balances of the local politics, of the local communities, and all of that, which can’t be replicated by a few big-ticket foreign investments.

What to study [02:58:50]

Rob Wiblin: Yeah, just stepping away from that specific issue, there’s an existing corpus of neoclassical economics, a school of thought that people get taught all over the place. You walk away with particular tools that you can use to analyze questions. Is there a similar corpus of political economy or institutional economics where someone can read these particular books or go through these particular lecture series and then walk away feeling like they’re equipped to analyze policy choices through this lens in a systematic way, rather than just arbitrarily noticing different effects that things might have? That’s a blocker for me, is that I feel… I could try to do analysis like this, but I feel like there’s so many things that are going on that I wouldn’t really know how to pick up, like what are the most important political economy implications of these questions? I’m not sure even what exactly I would study in order to get to the point where I would feel more confident that I was correctly analyzing a policy question and being able to predict what the implications might be.

Mushtaq Khan: That is a serious constraint. The advantage of neoclassical economics is that there is a set of techniques — which might be totally unrealistic in terms of a description of the world, but which most neoclassical economists agree are useful tools for thinking through things. Therefore, they can write textbooks about it, which you can buy and read, and one textbook will be quite similar to another. We haven’t got that for the heterodox school. Because there are differences in positions, and these positions can’t agree about those differences. For example, we’ve talked about North, Wallis, and Weingast. There’s a lot we agree about, but a lot of things we disagree about. We couldn’t sit together and write a textbook. I mean, when I talk to industrial policy people who are talking about how you need to develop capabilities in developing countries, there’s a lot we agree about until we come to the political settlement, and what will work and not work. Then, we find that there’s a lot we disagree about.

Mushtaq Khan: I think the answer to your question is that you have to do a lot of homework. It would be fun for me to say, “Just read my stuff, and you’ll get it all.” But, no. You have to make up your own mind. I have a particular position on things, and people I work with have a particular position on things. And the further out you get… I mean, if I think about Acemoglu, Johnson, and Robinson’s work on why nations fail, then there’s even less I agree with them about. But there is still stuff that I agree about even with them. I think there are these concentric circles. And I think each of these authors and approaches gives you something. But you have to read your own history and your own newspapers and say, “Actually, does this explain what is happening in country X, sector Y? If not, what commonsensical questions should I ask?” Then, you search for people who have addressed those questions. You will find different people who have addressed those questions. I think that’s the approach.

Mushtaq Khan: I don’t think it’s actually even feasible to have a political economy textbook, because it would be too big. There are too many different questions. Political economy is basically about how the world works. You can’t put that in a textbook. There are going to be different positions. Those different positions are not a problem. They are a good thing. Because nobody knows the truth. Nobody knows the right answer. But we each have our own angle, our own take. We can explain lots of things, but others can explain things in a slightly different way. And so I would say a good policymaker needs to have an awareness of different schools of thought, different methods, and so on. For some problems, the neoclassical approach might be actually quite good. For other approaches, it might be not only useless but dangerous. It might actually make things worse.

Mushtaq Khan: I can’t say don’t read any neoclassical economics, because it’s really useful to have those tools. In the same way, though, I think you should read all of these authors that we’ve discussed, and maybe also read some of my stuff, and then make up your own mind. Because ultimately, the best reality check is your own common sense.

Rob Wiblin: I’d like to get to a world where people have the traditional economy lens, but also a political economy lens. Then these things get integrated, and people go back and forth looking at the different angles on these problems. It seems like there’s a big barrier to getting the political economy lens applied systematically and regularly to all of these questions, because it’s hard to simplify as a set of questions that you ask one after another, or tools that you apply, like supply and demand curves, to all of these questions. I wonder whether there could be a lot of value in trying to boil it down to an actual toolkit that a policymaker… What if I’m a bureaucrat, and I’m trying to suggest something. I used to work in the Australian government, and we would suggest things about the electricity industry. But I wouldn’t know where to begin if I was trying to provide a political economy/institutional economics lens on things, because there’s so many different things you could do that it feels somewhat arbitrary which threads you decide to pull on.

Mushtaq Khan: I think that that’s a genuine description of a problem, but I think it’s not that neoclassical economics has an agreed-upon set of tools. Because if you look at how, even within neoclassical economics, people apply these tools, there are huge differences between them. The same with political economy. I think each framework has some basic questions that they’re asking, and then they’re answering them in some ways. If you look at the political settlements framework, it is asking a number of questions about policy implementation, and the interface between organizational power and institutions, and then asking you to locate your policy in a way that will make incremental changes effective. Then everything else follows from that. You need lots of different building blocks to make sense of it. As you rightly say, you can draw on neoclassical economics, you can draw on other political economy frameworks.

Mushtaq Khan: In the same way, if you look at Acemoglu, Johnson, and Robinson, their framing question is, “Are your institutions extractive or inclusive?” And then everything else is circling around that. Or is it a limited access order or an open access order? What are the different types of limited access orders, and what are the doorstep conditions?

Mushtaq Khan: It’s not that it’s just a confusing mess. Each political economy theory is giving you a set of organizing principles, and then asking you to say if your question can be fitted around these organizing principles. You’re asking why there isn’t one organizing principle. I’m saying—

Rob Wiblin: People don’t agree.

Mushtaq Khan: —reality is complex. People don’t agree. And maybe it’s not even desirable. Because this competition is a very good check and balance. I think neoclassical economics has had a free run for a long time. It’s got many, many, many things disastrously wrong, and is getting a challenge.

Mushtaq Khan: In neoclassical economics, the organizing principle is that people negotiate their own agreements at an individual level. The market is nothing but a set of contracts between people which are voluntarily made. You can explain quite a lot in terms of the voluntary contracts that people make, and the reasons why they can’t make those voluntary contracts. But at a deeper level, you’re not asking, “Why do people behave like that? Which contracts are enforced? Which contracts aren’t enforced?” As soon as you start asking that, you can’t just look at individuals. You’re looking at power structures. You’re looking at society. You’re looking at history.

Mushtaq Khan: In that sense, neoclassical economics is at the bottom of this food chain in asking about the individual contracting, which is really important and useful, but, actually, most of the interesting questions are about why those contracts aren’t enforced. That is how Douglass North began the journey of institutional economics in saying property rights and contracts may exist but they’re often not enforced.

Mushtaq Khan: Enforcement, I think, is the key. Enforcement takes you to all of those political economy questions which different people are cutting in different ways. But if you look at all the different political economy frameworks, just asking the same basic question, “How is the organization of society and of the collective organization affecting how individuals behave?” Whereas the neoclassical is starting from the other end, saying, “Let’s take preferences as given. Let’s take the constraints as given. This is how individuals behave.” And political economy is saying, “That’s trivial. You forgot the most important parts of the story of how we got there. How they finally make the contract is the most trivial part of the story.”

Mushtaq Khan: I think all of these things are connected. But I think that the really important questions about how social organization/social power affect how individuals behave, contract, the belief systems they have, how they enforce rules on each other, how they punish each other, these are historically specific questions to which you cannot have a general theory. Each of us are picking on aspects of this which we think are important. To some extent, all of them are important. It’s just that it would be impossible to put everything together. You have to pick the framework you think is the most parsimonious in giving you the tools that most likely give you something to work on without saying that everything else is wrong.

Rob Wiblin: What’s the relationship between your work and public choice theory? I guess it shares this factor of wanting to open up, say, the government, or institutions, and then look at it as a series of individuals who each have their own motives and their own ability to influence things. You could end up producing an outcome that wasn’t entirely originally intended. I get a slight feeling that public choice theory is institutional economics but for more free market-y people. Where everyone’s skeptical, but it’s quite similar in its analytical style.

Mushtaq Khan: I think the basic difference is that public choice theory is political economy, as you say, coming from neoclassical thinking. It’s saying, “Yes, structures matter. But we have to explain the structures also in terms of individual decisions.” In other words, individuals create the structure, and then the structure creates the individual, and the individuals also… But ultimately, you’re going back to individual incentives. The fundamental difference from that and almost all other kinds of political economy is that we take history seriously. To say that all structures are created by individuals, and therefore if the structure of society in India is different from the one in the United States, then we have to look at the individual incentives that created those structures, I think is a non-starter. It confuses the path dependence of history and the complexity of how structures are built up. Individuals today in India may not have any capacity of changing that structure to look like the one in the U.S. or Norway, not because they have some information deficit or anything like that, but because a structure itself has a reality and a meaning which affects the way individuals behave.

Mushtaq Khan: I think that my political economy takes both as important. We have to understand how those organizational structures are gradually evolving and how that is affecting individual decisions, which then feed back into those structures. But they’re not able to completely change those structures. I think this is summed up very well in a phrase by a bearded man from some time ago named Karl Marx, who said — he was using ‘man’ in the generic sense — he said, “Men make their own history, but they don’t make it exactly as they please.” It’s understanding what it is that they can’t do exactly as they please that is so important, and that has a lot to do with pre-existing power structures.

Mushtaq Khan: It’s not that political economists are saying individual agency doesn’t matter. It’s that individual agency matters a lot, but it’s constrained, and the constraints are not also explainable by individual agency. If you fall into that trap you are in this complete circularity. The constraints have some historical origin, and we are trying to map those historical origins. All the people we have discussed have different ways of mapping that history. History matters is, I think, where political economy begins and neoclassical economics stops. For neoclassical economics, there is no need for history. You can model everything from individuals by just changing their preferences and constraints. For us, the most important thing is, where did those enforcement capacities, those preferences, those beliefs, come from? That’s where history matters.

Takeaway lessons [03:11:10]

Rob Wiblin: Are there any takeaway lessons that you would want people in the audience to keep in mind if they’re working in global development, or I suppose working in offering policy advice in rich countries as well? What stuff should they keep in mind going forward?

Mushtaq Khan: I think a lot of what I’m saying could be read by people in different ways. Let me boil it down to some very basic things which might focus people’s minds. I think now it’s commonplace in development policy that people will say, “Context matters. One size doesn’t fit all.” Everyone will say these things, from the World Bank, down to the Foreign, Commonwealth & Development Office, down to Oxfam. Everyone will say this.

Rob Wiblin: I always slightly roll my eyes. It’s not because it’s not true, but because it feels just like a platitude. It’s a motherhood statement.

Mushtaq Khan: The real question is, you have to ask yourself, “How does context matter? What am I supposed to do about this platitude? What is it about context that matters?” Here is where I think all that I’m talking about has a direct relevance. Context matters in this way. You have a set of plans and policies, or you’re supporting some activities which you think will support human development. To me, that inevitably involves looking at capabilities, organizations, etc. Do you think these policies will be implemented? Do you think they can be enforced? This is where context matters. Context matters primarily in the sense that enforcement and implementation vary hugely depending on the interests and capabilities of the players themselves, and much less on the enforcement capability of some supra-agency, which is like the prime minister’s office, or the cabinet, or the police, or the anti-corruption commission, or anything like that.

Mushtaq Khan: The basic thing you should do, always, is look at the people you are working with, their capabilities, and then ask, “Here is a policy to help them do something better. Will they themselves actually support it and be able to implement it?” Not are they saying they will, because everyone will say they will do whatever you ask them to do, because they want the money. But actually, from their past behavior and their activities, do I think this is something that they will actually be able to implement, enforce, monitor each other as peers, report back; and do the basic capabilities of policy implementation at a higher level exist to monitor and enforce this?

Mushtaq Khan: If not, you have several options. You can redesign the policy so you have greater confidence it works. Or you can say, “Actually, these types of policies won’t work here, because there’s some network problem which is so integrated that I can’t break it.” Then you have to think out of the box about some exit strategy which meets the requirements of your anti-poverty or whatever that you are trying to achieve. Or you have to go back one step before, and ask, “Can I build those capabilities directly? Can I work with people to actually build those capabilities that I’m assuming exist to make my policy workable? I shouldn’t assume they do. Maybe I need a policy that is one or two steps behind in terms of its sequencing to develop the requisite capabilities.”

Mushtaq Khan: I think there are ways in which all of this makes a lot of practical sense. When you are asking yourself “Will this work here?” then of course you can refer to the work that I do and that others do. It’s all available. It’s free to access online. You can read some of the stuff on the SOAS website, or the work of Dani Rodrik, or North, Wallis, and Weingast, or Acemoglu, Johnson, and Robinson, and all the people we have discussed. Not all of it, but those bits of it that are relevant to the particular policy question that you’re addressing. I think that’s the starting point. That’s the starting point for you, as a development practitioner, beginning to refine your questions and getting down to better questions and then answers which you think might make sense there. Then, the final step is, don’t put all your eggs in one basket. Experiment with a trial, see how that goes, and only scale up once it’s working.

Rob Wiblin: Yeah, it’s interesting. I feel like in our personal lives, say if you’re working in an organization of 100 people and you have some problem, and someone suggests a policy that, in principle, would be very good for solving this problem to try to change people’s behavior, for example. In a context like that, where you potentially have a lot of local information, you can listen to this idea and be like, “No, that’s not going to work, because the marketing department will shut it down. They’ll make sure that it never gets implemented.” Or, “I know how this organization works. This will just go on the books, and everyone’s going to ignore it. No one’s going to take it seriously.”

Rob Wiblin: But we don’t ask that question as much in context that we don’t know, because we don’t understand enough about the people and the actors and their history of behavior to actually say anything. But it is just so essential. Someone coming into your organization without any knowledge of the people, or what they want, or how they work, and just adjusting reforms is probably going to have a very bad time — as, I guess, do we, when we go into other organizations and countries and just pull something out of a textbook.

Mushtaq Khan: Absolutely. You’ve hit the nail on the head. We think that fixing our little office of 10 people is really difficult, but we can go and give advice to a country of 160 million, and it will make a difference. That’s not going to happen. Development is a difficult process. It’s actually about, ultimately, exactly that: changing behavior, changing how organizations work. That’s a long-term process. If you are interested in development, you’ve got to be in there for the long haul, and you’ve got to be interested in actually understanding your market: the people who you are selling the ideas to. Then, figure out what’s the best way of selling this idea in a way that will be picked up and used. If you can’t think of why anyone should use this idea… If you wouldn’t do it if you were in their position, then don’t try it. It’s not going to work.

Rob Wiblin: One final question just before we finish. We’ve talked a whole lot about capabilities. I guess we’ve talked loosely as though the U.K. has fantastic capabilities in all kinds of ways. But I’m curious to know, are there any times when you’ve looked around in the U.K. and seen just extraordinarily… where you’ve seen organizations that are failing to have the capabilities that you might hope and expect, and just been like, “This is such a great demonstration of the problem of organizational difficulty.” I’m a connoisseur of complaints about big utilities firms and things like that, because sometimes their behavior is incredible.

Mushtaq Khan: All the time. You don’t even have to go to utility firms. The way in which universities are run sometimes is exasperating. Simple things take a long time to get done, because the systems aren’t set up properly.

Mushtaq Khan: But I think one has to be realistic about it. We’re never going to get a human organization which has zero transaction costs, which is so efficient that everything happens really quickly. This is one of the founding observations of institutional economics. Every organization has very high transaction costs. The real difference is between a small difference in that transaction cost. The issue is not that British universities are inefficient. They’re super inefficient, but they’re a lot more efficient than the universities in Bangladesh. I think that is where the difference is. I think the benchmark should not be that I want to get to a zero transaction cost/seamlessly working organization where I don’t have to go and explain things five times to 10 different people to get anything done. But it should go from I have to explain the same thing 100 times to 200 people who are all blocking me because they have no incentive to do anything, to a situation where, if I explain it to five people 10 times, they have the incentive to do something, and they will do something about it.

Rob Wiblin: My guest today has been Mushtaq Khan. Thanks so much for coming on the 80,000 Hours Podcast, Mushtaq.

Mushtaq Khan: Thank you, Rob.

Rob’s outro [03:19:10]

Rob Wiblin: As I mentioned in the intro, Mushtaq’s SOAS-ACE research consortium is currently funded by the U.K.’s FCDO, but they are always looking for new sources of funding to continue or expand their research. If you’d like to learn more about their approach and opportunities to help out with their work you could contact [email protected] or visit their website: https://ace.soas.ac.uk/.

Finally, just a personal recommendation. I’ve been enjoying a new podcast series called Real Dictators. It goes into detail on the roots of dictators like Muammar Gaddafi, Francisco Franco, and General Tojo, and how they successfully rose to power. I’d have thought I knew enough about that, but it turns out I knew only a tiny fraction of their backstory. The soundtrack for the show is also beautiful.

I’m recommending it for entertainment, not because it will help you have more social impact, though I suppose there’s more useless things one could learn about than times different societies went far off the rails.

Alright, the 80,000 Hours Podcast is produced by Keiran Harris.

Audio mastering by Ben Cordell.

Full transcripts are available on our site and made by Sofia Davis-Fogel.

Thanks for joining, talk to you again soon.

About the show

The 80,000 Hours Podcast features unusually in-depth conversations about the world’s most pressing problems and how you can use your career to solve them. We invite guests pursuing a wide range of career paths - from academics and activists to entrepreneurs and policymakers — to analyse the case for and against working on different issues and which approaches are best for solving them.

The 80,000 Hours Podcast is produced and edited by Keiran Harris. Get in touch with feedback or guest suggestions by emailing [email protected]

What should I listen to first?

We've carefully selected ten episodes we think it could make sense to listen to first, on a separate podcast feed:

Check out 'Effective Altruism: An Introduction'

Subscribe by searching for 80,000 Hours wherever you get podcasts, or click one of the buttons below:

If you're new, see the podcast homepage for ideas on where to start, or browse our full episode archive.