Transcript

Rob’s intro [00:00:00]

Robert Wiblin: Hi listeners, this is the 80,000 Hours Podcast, where each week we have an unusually in-depth conversation about one of the world’s most pressing problems and how you can use your career to solve it. I’m Rob Wiblin, Director of Research at 80,000 Hours.

Today we’re going to discuss what might turn out to be one of the biggest ideas that comes out of the effective altruism community, with one of the smartest young researchers I know, Phil Trammell.

If I had to summarise the idea in one sentence, it would be that we can do much more good if, instead of trying to do good by, say, giving to charities today, we profitably invested our resources for very long periods of time, potentially centuries, and then use them to do good at a later, more auspicious, moment.

If that sounds counterintuitive to you, I totally understand.

But stick with us for a bit because the arguments for that conclusion are actually impressively strong, and the counterarguments not as decisive as you might think.

If this idea is right, it could be a crucial consideration, and many listeners should consider taking a different approach to improving the long-term than those we’ve highlighted so far.

In particular, movement building and then earning to give would look better than they currently seem to.

The ideas in this episode could be so important that we decided to go through them particularly thoroughly, so that we would have a resource people could use for some time.

Having said that, this episode was recorded several months ago, and if we were recording today there are plenty of small things Phil would say differently.

This is an area of active research, so we may have to get him or a colleague back again to give us an update on their latest thinking.

Just quickly before that, I should mention that as I’m recording this on 16 March 2020, the world is in full freakout mode over COVID-19, and I think it’s right to be. We’ve been spending a great deal of time reading about this emergency and will be trying to urgently bring you advice on how you can make a difference. Of course that advice is going to have to be a lot scrappier than what we usually aim for.

In the meantime you can get my unofficial personal opinions on what we ought to be doing, unchecked by anyone else, at twitter dot com slash robertwiblin.

Alright, without further ado, here’s the brilliant Phil Trammell.

The interview begins [00:02:23]

Robert Wiblin: Today, I’m speaking with Philip Trammell. Philip graduated from Brown University in 2015, where he majored in economics and mathematics, and was awarded the best economics thesis of the year. He’s published research on the theory of rational behavior under decision theoretic uncertainty and he’s spent the last year looking into other fundamental questions in effective altruism at the Global Priorities Institute in Oxford, whose research agenda he actually helped to coauthor.

One question he’s been thinking about in particular recently is the questions of when to give, which is pretty natural given that he’s taken the GWWC further pledge to give away all is income above $30k in 2017 dollars.

He is also a graduate student in economics at Oxford University and regularly writes what I think are very intriguing blog posts at philiptrammell.com. So thanks for coming on the podcast, Philip.

Philip Trammell: It’s an honor to be here.

Robert Wiblin: And today, I’m also joined by my colleague, Howie Lempel, who’s going to help me deal with what are occasionally some quite technical issues here. Welcome Howie.

Howie Lempel: Hey everybody.

Robert Wiblin: Today’s conversation is going to focus on the case for and against waiting before having a direct impact. And I guess one could do this either by saving money and earning investment returns and then donating a much larger sum at a later time, or by taking a role where you don’t have much immediate impact, but you can go about building your skills and then do something much more impressive later on. But first, Phil, why is this such an important question that people should be giving a significant amount of time to thinking about?

Philip Trammell: Right. First, I think it’s just extremely decision relevant. If we’re trying to figure out what the best use of a unit of resources is, there’s this huge space of possible options, right? A lot of causes we can give to, a lot of interventions within the causes. And, if we try narrowing down that space, there are a lot of questions we’ll have to ask and answer before we come anywhere close to an answer. But it just seems to me like there’s a more natural and a high information value way to divide up the space, which is to first ask whether we should be giving now or giving later or working now or working later? Second, I think a priori, there are pretty strong reasons to think that most of the time, one ought to be building up resources to use later. An analogy I like to use is that just as you would expect that the person in need you happen to be passing on the street is unlikely to be the person who can make the best use of the money in your wallet and you should try to find another location, another person to wire the money to.

Philip Trammell: We should typically think that the moment we find ourselves in is not the moment at which we can have the most impact and there’s going to be some more pivotal moment in the future for which we should wait. So it’s just very decision relevant. I think there are a lot of reasons to take the possibility of waiting to have impact more seriously.

Robert Wiblin: So why is this a counterintuitive idea? I guess if you look around, it seems most people are trying to have an impact right away and perhaps not doing the sorts of things that you might recommend if you thought, “Wow, maybe we should try to have an impact in hundreds of years time”.

Philip Trammell: Right. Yeah. Thanks for mentioning the hundreds of years point because so far everything we’ve said could have been interpreted as meaning you should just wait five years or 20 years. But the possibility I think is really under-considered is that we should wait for longer than a human lifetime. I think that’s counterintuitive because for one thing, most people just don’t care that much about what happens after they die. So most retirement planning decisions, political decisions and everything else that we develop our intuitions around are made on a timescale of months or years or decades. And the thoughts and the infrastructure that would be needed to actually explore the possibility of acting on a longer timescale therefore haven’t been explored necessarily. There just isn’t as much legal infrastructure or economic theory or whatever concerning the question of how to effectively spend one’s funds in centuries’ time.

Consequences for getting this question wrong [00:06:03]

Robert Wiblin: What could be the magnitude of the downside for people who are trying to have a big impact, and I guess the effective altruism community specifically, if we get this question of timing or the right discount rate to have about impact and are wrong in either direction where we either try to do it too soon or too late?

Philip Trammell: Yeah. In some sense, there’s no limit to the downside because it’s always possible that what you do is completely useless. But one fact I think to bring up front and center and highlight the importance of the question is that the US stock market has averaged over 7% annual returns over the past hundred years and maybe that won’t continue, but maybe it will or maybe it will be even higher. If those sorts of interest rates persist for another century, money invested now will roughly double in value every decade, which means it’ll multiply by a factor of two to the 10 after a hundred years. And obviously you might also expect it to get more expensive to do good over time, but whatever reasons there are to spend rather than invest for a hundred years, I have to overcome this like two to the 10 factor in favor of waiting. And that’s just after one century. So the stakes really are high.

What have people had to say about this question in the past? [00:07:22]

Robert Wiblin: What have people had to say about this question in the past? I guess it wasn’t completely made up by you?

Philip Trammell: Right, of course. So basically everyone thinking about how to do philanthropy has encountered this problem in some form and maybe five or so years ago there were a lot of blog posts in the EA community which tried to answer the question of whether to give now or later, but most of them were focused A), on the specific cause area of global poverty and B), on the timescale of a decade or so. Not really delving into what we’re going to be delving into: whether to consider giving on a much longer timescale. There are some exceptions. So Robin Hanson wrote a bit a few years ago about whether we should consider trying to save on a timescale of centuries or longer and recently has been talking about that idea in more depth. People can look up a lecture he’s given on “Long Legacies and Fights” is the title if they’re interested about his thoughts on this. But as far as I can tell, that’s basically the sum of it and no one’s really tried to create a reasonably formal model of the considerations at play and kind of what they have to say about whether we should try saving on a timescale of centuries.

Robert Wiblin: Yeah. People were talking about this quite a lot five years ago. Then the conversation died down a little bit until last year when it’s been very much back on the table again. Why do you think that is?

Philip Trammell: So I think different cause area camps within the effective altruism movement have all had different reasons for thinking that now is probably a good time to give. So people who think that global poverty is the most important cause will sometimes point out that the global poor are getting richer quite quickly and opportunities to help them are getting taken quite quickly and so maybe again, on a time scale of like a decade, you might think that the cost of helping the world’s poor is rising more quickly than 7% per year or whatever. Even that I’m not convinced of actually. So GiveWell’s top recommendation of the Against Malaria Foundation had an estimated cost per life saved of like $3,400 for a while and now they’re recommending Malaria Consortium at a cost per life saved of like $2,000 and you know, a lot of countries aren’t experiencing this catch-up growth, so even if on average that’s what we observe, then it’s still like–

Robert Wiblin: As long as someone isn’t.

Philip Trammell: Yeah, exactly. But anyway, that was roughly the line of reasoning. I think likewise in animal welfare, there’s been a lot of optimism about us maybe being on the verge of transformative technologies like clean meat which would kind of solve the problem. And so it’s just a matter of what we do to hasten it in the near term or to alleviate the suffering of animals that have to endure the short-term. And then among longtermist types, I think until recently that space was somewhat dominated by people who also thought not just that the most effective interventions were those broadly aimed at improving the long-term future, but in particular, that there were these quite substantial existential risks to the world or human civilization which were quite pressing in that they had to be dealt with very quickly or else we would suffer the catastrophe. So I think it’s a little suspicious that people across all these different cause areas and other areas of life outside of effective altruism so often seem to come to the conclusion that the best opportunity–

Robert Wiblin: Is uniquely pressing.

Philip Trammell: Yes, exactly. And I think as people have reflected on that, they’ve started taking the possibility of what you sometimes call ‘low discount rate longtermism’ more seriously. The idea that we should really try just to build up resources for use in the relatively distant future.

Howie Lempel: So I think another potential reason that, at least in public, there hasn’t been as much of this conversation is that five or 10 years ago the EA community was almost entirely composed of individual donors donating small amounts of money and then all sort of together, in public, figuring out how to approach a giving strategy and as it’s become the case that a very large portion of EAs financial resources are now housed at Open Phil and they sort of already have their endowment and don’t need to sort of convince others. It might be the case that even if a similar amount of conversation about this issue was going on, a lot of it could now be housed privately as Open Phil figures out how quickly to spend as opposed to publicly trying to convince tons of small donors.

Philip Trammell: Yeah, that seems like a possibility.

The case for saving [00:11:51]

Robert Wiblin: All right. So let’s maybe launch into this fully. I think to start off with maybe for the first hour or two even, we should mostly focus on the money case than the donating case because that’s easier to discuss investment returns. Things get a little bit trickier, there’s other things that come into play once we’re talking about people’s careers and timing of their career; should they try to have an impact now or in 20 years time? Maybe we can talk a little bit about careers as we go but maybe save that for a section towards the end.

Howie Lempel: So another assumption that we’re making in addition to starting the discussion by focusing on money as opposed to other resources you could use now or invest, is that we’re going to work with the simpler case of giving money directly to the global poor as sort of an example of what it would be to have a direct impact now and later on we may discuss the extent to which that changes if you’re doing other interventions.

Robert Wiblin: All right. With that out of the way, what do you think are the most compelling arguments for thinking that at least in an ordinary circumstance, philanthropists ought to try to be quite patient and wait before spending their funds? I guess we’ve slightly already alluded to perhaps what’s the main one, which is just that there’s pretty substantial investment returns if you wait a long time. Would you say that’s the key issue?

Philip Trammell: In some sense that’s the key issue when you’re thinking about doing something, you know, acting on an opportunity that’s basically always there, which is just to use the money to increase the consumption of people alive at the time that you’re spending it. Picking that apart a little more, as I mentioned earlier, what matters isn’t really the rate at which your resources grow, it’s how that compares to the rate at which it gets more expensive to help people. But, there’s this structural reason why you should think that the interest rate is almost always higher than the rate at which it gets more expensive to help people. And that’s that people are impatient. People discount their own wellbeing within their own lives. I mean, everyone’s just felt this. The prospect of immediate pain is weighed more heavily than the prospect of pain in decades.

Philip Trammell: And perhaps even more importantly, people discount what happens to their descendants relative to themselves. But a patient philanthropist might care about pleasures and pains whenever they occur and to whomever they occur. And so what we’re in is this world where interest rates are basically set by the impatient. If it gets 2% more expensive to help someone next year to avert a pain or to buy a pleasure, they’ll be indifferent between a dollar now and a dollar and 5 cents or something next year. And the difference between those two rates is their impatience. So by just waiting a year, you can help them a few percent more than you could have done or help them or their children or something than you could have by giving to them this year. And that logic just compounds. So as time goes on, you can just do more and more good by waiting.

Robert Wiblin: Yeah. I think this may not be completely obvious to the audience that economists think there’s going to be a pretty good relationship between how impatient people are in general or how impatient people with lots of wealth are in general and the kind of investment returns you would expect to get in the stock market or investing in housing and so on. Do you just want to explain that connection a little bit?

Philip Trammell: So let’s say right now it’s the case that everyone knows that if they put a dollar in the stock market, they’ll get a dollar and 6 cents next year. That’s not exactly right, of course, some of the rate is compensating people for risk and so on but whatever. Let’s say there’s some investment like that. What that means is that people on the margin are indifferent between a dollar now and a dollar and 6 cents next year, right?

Philip Trammell: I mean, anyone with a dollar could just buy a dollar and 6 cents next year with it. Now, if it’s the case that because people will be richer and so on, it’ll be 4% more expensive to help them next year to buy a util as it were. They’ll be indifferent between the dollar and the dollar and 6 cents because they discount their wellbeing at 2% a year. Now, if they discounted their wellbeing at 3% a year, then they’d be indifferent between a dollar and a dollar and 7 cents next year. And if they discounted it even more heavily, the interest rate would be even higher. So that’s the relationship between the two.

Howie Lempel: What do you mean by the interest rate when you talk about that?

Philip Trammell: Yeah, I mean every asset has its own expected rate of return and its own risk profile, and most of the discussion today I think will abstract away from that and just assume that there’s a single low risk rate at which people can invest resources today and earn more resources in the future. So if you can buy a share of some company today and you knew it was going to be worth 1.06 times what you paid for it next year, then that would mean an interest rate of 6% if there were no other dividends or anything involved.

Robert Wiblin: Yeah, and I guess this interest rate comes in all kinds of forms. So sometimes people buy houses and then the houses become more valuable and the rate of appreciation there is the interest rate there and you’ve got bonds with a kind of a coupon and they get back a principal at the end or you’d leave money in your bank account which is kind of similar and then we abstract away from all of the specific details of the specific kinds of investments and say, “What’s the background rate of investment returns that people expect from making an investment in the stock market”. So we don’t have to play around too much with the details. And of course the investment returns for these different asset classes tend to move up and down in tandem. Although for some riskier assets it tends to be higher. Is that a good summary?

Philip Trammell: Yeah, that all seems about right. I mean, we can always get ever more into the weeds of what we mean by a riskier asset. Maybe what matters more is whether an asset is correlated with other assets rather than its actual volatilities in itself.

Robert Wiblin: Okay. Yes. So with all of that out of the way, can you give any more concrete examples to give people a sense of how big this effect of investment returns is? Because I think when you just talk about two to the 10, it can be a little bit abstract. It’s like, we take like a hundred dollars, how much does that turn into in the future? And I guess given potential price increases, how much more stuff can you actually buy?

Philip Trammell: So first off, I would say that that 7% percent historical interest rate I mentioned was a real interest rate. So that was after inflation. So that means that someone investing a hundred years ago could really buy two to the 10 times more stuff today. Houses and all the rest of it. I mean maybe it’d help just to say that two to the 10 is 1024 and that waiting 110 years doubles that again. So that’s 2048 and so it goes.

Robert Wiblin: Yeah. So in a sense there’s a lot of potential power here. If you waited 200 years… So I just calculated this out. If you have a 7% return for 200 years then you have 750,000 times as much money, so maybe we should use these longer time scales if we want more striking examples. Is there anything else to say on this issue of just obviously you should potentially save if you’re getting these enormous investment returns and you’re going to have way more in future?

Philip Trammell: Yeah. We’ll talk about some counterarguments later. Of course, invested money can be lost and you might think that actually in some domains it’s getting more expensive to do good more quickly than that. But this seemed like a good place to start.

Robert Wiblin: Yeah. In the paper that we’ll link to which you’ve written up about this topic, you use these terms, I guess ‘r’ is the real interest rate. And then we talk about ‘g’, which is the global economic growth rate, right? I guess some people might recognize those terms from Piketty and his famous ‘r is greater than g’. I’m sure we’ve all read ‘Capital in the Twenty-First Century’. Yeah. And then in that paper you also run through this quite striking example where you talk about how even if your goal is to just transfer money to the world’s poorest people. And even if you think that severe poverty is basically gradually getting eliminated in a couple of hundred years time, there’ll be no one who’s poor by our standards today. It can still be better to save up this money and then give vastly larger amounts to people who in the future will be much richer than than we are now simply because the amounts of money that you can give is so much greater. Do you want to work through that case which I found quite counterintuitive when I read it.

Philip Trammell: Sure. I should say first that this is assuming utilitarianism. So if you have some what’s called a prioritarian moral view that says that the badly off get more moral weight, not just because it’s easier to help them or some other view along those lines, then this conclusion doesn’t hold. And also that it’s assuming that utility and consumption doesn’t literally just plateau at some point, but we can roughly extrapolate the curve, the logarithmic or another sort of reasonable looking curve out well beyond the observed range. But anyway, if it is the case that the logic I articulated earlier continues to hold, so interest rates continue to be a few percent higher than the rate at which you can make someone better off by spending on them or by giving them money to spend on themselves, then I think I worked out that in something like 279 years time, invested money just given to develop the world’s investors would create more welfare than that money given today to the world’s poorest. So even if you think that global poverty will be eliminated, and the fact that the developing world would completely catch up and so follow the same trajectory as the developed world is currently on, you’ll still have this opportunity, which is just to give it to the future rich who will be very, very rich. In fact, there’ll be like 250 times richer if they get 2% richer per year, but you’ll be like 165,000 times richer even if you just go at like 4.4% per year, right? So that’s lower than the historical stock market rate of return. And that 165,000 just barely swamps the fact that it gets harder to help them because they’ll all be 250 times richer. So that’s how the numbers seem to work out.

Robert Wiblin: I guess some listeners might be rolling their eyes right now and thinking, “Are people really going to be this much richer in 279 years time”? Of course they might not be in reality. But here we’re just imagining a toy example where economic growth just continues nicely every year and there’s good investment returns and people are still around. So it’s something like the next 279 years look a little bit like the previous 279 years.

Philip Trammell: Yeah. I think importantly, actually, even though that is the simplest way to set up the example, it doesn’t rely on growth continuing–

Robert Wiblin: It gets stronger if growth stops, right?

Philip Trammell: It depends why growth stops, but all that matters is for impatience to continue. As long as impatience continues, this gap between the interest rate and the rate at which it gets more expensive to help people will compound over time. So if everyone’s gotten poorer and you’ve gotten rich less quickly, right. Or if you even gotten a little poorer yourself in the meantime, there’ll still be this wedge between how much good each dollar does and how many dollars you have that’s compounded because of that impatience that was built into the interest rate.

Robert Wiblin: So to spell this out clearly. In this toy example to explain how this might work, we’re comparing giving, I guess, $1 to the world’s poorest people today or $165,000 to someone who is, I guess, 250 times as rich as the world’s poorest person today?

Philip Trammell: Not as the world’s poorest. As a typical investor in a developed world country.

Robert Wiblin: Oh right. Okay. Why not use the example of someone who’s very poor? I guess you think that there might be a lot of catch up growths with their income and so that complicates the issue a bit?

Philip Trammell: Yeah. A common counter argument to the idea of waiting to give is that the world’s poorest are currently getting richer more quickly than the interest rate due to some sort of market failure where they can’t borrow it to invest in their high return catch up projects at the moment. And this is a serious consideration. There are a lot of places that seem to be exhibiting some form of catch up growth and microfinance was, I mean my understanding is it hasn’t played out as well as it was initially hoped, but it was kind of motivated by this observation.

Philip Trammell: You have these poor regions that are growing more quickly than the interest rate and so you should actually either try to help them now rather than waiting just one or two years, or invest in those regions through some sort of set up that overcomes the costs of implementing loans in developed world or whatever. Anyway, the point is even if you think that there will be this full catch up growth, all that that means is that it would be better to give now than to give in like a few years time. But once they’ve caught up, they can’t grow faster than the world as a whole forever. Eventually their growth rate will just equal the the world growth rate because if they have their opportunities to grow that are higher than the interest rate, they’ll just borrow at the interest rate until all those are expended.

Philip Trammell: So we can just fully accept that argument and just ask, “Okay, well how long would it take before just giving to future developed world investors does more good than giving to the world’s poorest today”? I don’t think that’s the best thing to do. I just think it puts a bound on the argument.

Robert Wiblin: Yeah, you’re handicapping yourself here by having an intervention that doesn’t really make a ton of sense. But then, so your point is that after 279 years with these assumptions, the investment returns are enough to not only overcome the income growth that people will have over that time, but also the fact that you’re now targeting it at a person with a typical level of income at that time. But even given all of that, you might still have a bigger effect on welfare just because the amount of money that you have to give is growing 165,000 fold.

Philip Trammell: Yeah.

Robert Wiblin: I guess the argument looks stronger if we think about what would be the most effective interventions that you have available at that time? That could have a potentially pretty huge effect, right?

Philip Trammell: Yeah, I think so. I alluded to this at the beginning when I said that you should think that the moment at which you can have the most impact is probably not the present moment. And I mean that wouldn’t be an argument for waiting for any particular point in the future, but just that you should wait around for some moment when there’s a funding opportunity that seems really important and neglected and so on and you should expect those to come and go and like maybe there are some available now, but unless you think the most important ones ever are available right now you should probably wait around for a better one.

Robert Wiblin: Do you want to just respond to potential skeptics in the audience who might be thinking, “Oh, like when someone’s that rich, money makes absolutely no difference to their welfare”. I guess if it does actually make no difference, then that does get rid of this argument. But you’re just saying, “What if it has a tiny effect? Then it still helps”.

Philip Trammell: Right. I would say that it makes sense to have a lot of uncertainty about how to extrapolate the relationship between spending and welfare out beyond the range that we’ve observed. I could imagine someone thinking that consumption would make no difference to welfare to someone with our standard of living hundreds of years ago. But I do find that additional spending somewhat increases my wellbeing. But anyway, in the face of that uncertainty, all of the value of this strategy comes in the scenarios where it turns out that there is a lot that you can do with that much money even for people so rich. And, I think given the uncertainty, pretty much all the value comes from the unlikely case in which there’s a lot that we’ll learn how to do to create a lot of welfare. So there could be something like future technology or a completely harmless drug or something which is expensive to produce but which produces a lot of welfare and in the face of the uncertainty, it’s at least ambiguous which way the argument goes. But I think it actually strengthens the case for waiting because there is at least some possibility that we’ll be able to like discover ways to do a lot of good with a lot of resources.

Robert Wiblin: Okay. So another way of thinking about how your influence grows over time is not to think just about the absolute amount of money that you have in this ever-growing bank account, but maybe what share of all global wealth or what share of global income you command, which if you’re trying to have political influence or you know, buy things of which there’s a completely limited number, then it potentially matters more how much you have relative to other people rather than how much you have in absolute terms. Do you want to talk about how that instance looks different?

Philip Trammell: Yeah. I think that’s another scenario that’s worth considering. And you mentioned Piketty before who’s famous for popularizing the observation that ‘r’ tends to be greater than ‘g’; the interest rate tends to be higher than the rate of overall economic growth. The reason for that is related to what I was saying before about the interest rate tending to be higher than the rate at which it gets more expensive to help people. The people’s impatience plays into this. For those interested, you can look up Piketty’s reasoning or also Google the ‘Ramsey formula’. But anyway, the fact is that historically and theoretically it seems like as you grow, you’ll be growing faster than world wealth on average. And so even though your share of total wealth won’t be growing as fast as your absolute wealth, it’ll still be growing. And so in these contest-like scenarios, you’ll do better to invest than to try engaging in the contest at the moment.

Robert Wiblin: Yeah. Is there any way of maybe quantifying this? I guess I could potentially just throw some numbers into a calculator here where we imagine what if ‘r’ is 2% greater than ‘g’ and then think about how much your share of global wealth has increased over a 200 year time scale.

Philip Trammell: Sure.

Robert Wiblin: Okay. So as we mentioned earlier, if your absolute average, harmonic average, investment returns over the 200 year period are about 7%, then your absolute wealth would increase by about 750,000 over a 200 year period. Whereas your fraction of all global wealth would have increased about 52 fold, which is a lot less impressive than 750,000 fold. But nonetheless, I guess if you’re engaging in political activism, you’d rather have 50 times as much money relative to everyone else.

Philip Trammell: Exactly.

Hundred year leases [00:29:28]

Robert Wiblin: Yeah. So an example that I found quite neat from the paper was the example of the hundred year lease, which I think gives you a sense of how easy it is to increase the fraction of influence that you have over the future. So just to explain the general concept here, it’s possible to buy a piece of land and then lease it out to someone else for the next hundred years. So basically then what you’re doing is buying the right to that land from a hundred years onwards. Can you just explain how the price of that is effectively very different than buying the land right now?

*Philip Trammell: Sure. I mean it’s about 10% of the price. The effective price ends up being about 10%. So the hundred year lease costs about 90% of what the freehold costs. Yeah. I don’t think that land is necessarily the best long term investment or anything, but it’s a good visualization for the ways in which the patient can just bide their time and let the impatient do what they like with the near term and just come to have more say over what happens in the longer term. Obviously this reasoning compounds too, so at a 99% discount if this relationship persists and you can buy what happens to a patch of land from 200 years onward till you know, till the sun burns it up.

Robert Wiblin: Yeah. How come no one’s tried to buy a city this way?

Philip Trammell: I’m not sure why someone would want to buy a city in particular? I mean people do invest which is similar in some ways. Or at least lending is just this for assets in general. You tell some impatient person or whatever, “Well here’s some stuff and you can hold onto it for a while and do what you like with it and just give me back more stuff later”. It seems unlikely that the best portfolio will tend to be a single city. But basically every foundation that’s lasted for a long time of which there are quite a number, every sort of long lasting family trust. Maybe you could even count religious institutions. They’ve all engaged in long term investments and so I think there’s that.

Robert Wiblin: That’s a very polite way of saying “they don’t do that because it’s a bloody stupid idea, Rob.”

Are there any examples though of an organization saving money for a very long amount of time and then becoming much more influential a hundred years later than what it was to begin with? Or just for whatever practical reason, people haven’t really tried this strategy?

Philip Trammell: I think people have tried to implement patient financial behavior in various ways. Better and worse approximations of the literal model of setting up a fund that lasts a long time. I mentioned religious institutions before, right? They, in some sense, grow their assets through missionary activity and through the actual purchase of assets and the building of structures that are intended to last for a long time. There were a lot of foundations founded in the Gilded Age of the United States which have persisted to this day and have grown somewhat. Maybe not quite as much as they would’ve wanted to, because if they’re foundations, they have to disperse at least 5% of their assets every year in the US, though trusts don’t have to do that. So one could get around that if one thought to set up a trust. There are also some idiosyncratic examples of attempts to do this literally that have failed.

Philip Trammell: Well not failed, I should say. They haven’t all been complete failures, but no kind of amazing successes. So Benjamin Franklin put aside some money for the cities of Boston and Philadelphia to be invested for a few hundred years and then used to build schools there. And I think in one of the cities, I think in Philadelphia, the city ended up spending the money after like 100 years and then in Boston, they invested it in low return municipal bonds and things in the meantime, which ruined the plans a little bit. But I still think it’s possible that he ended up doing more good than he would have just by spending it at the time. It just wasn’t anything astronomical.

Robert Wiblin: Wasn’t there some Italian family that tried to do this? I guess they realized this strategy maybe very early on because they were involved in finance and they wanted to build a family trust that would just go in perpetuity and then they’d eventually end up as the most famous family.

Philip Trammell: Interesting. I’m not familiar with the Italian family. I know there’s one, ‘Holden’, someone with the last name, ‘Holden’ changed it to Holdeen with two E’s and wanted all his Holdeen descendants to be very rich. And I think that the trust was supposed to disperse to the Holdeens of the world in a thousand years. And his children ended up suing the trust after he died and the judge let them have the money. And one argument they made was that this was a risk to the world because it would take over the world. So that’s maybe something of a warning sign. On the other hand, I think the real outright failures have generally been family trusts. The economist Greg Mankiw has a paper, “Yes, r > g. So What?” in which he points out that family trusts tend not to last. But these other sorts of institutions: foundations and religious ones and maybe other examples haven’t been such failures and typically haven’t even been trying all that hard to really, really last a long time, but who’ve just been aiming at lasting on the scale of… I mean I don’t know, but they don’t seem to have put really intense thought into what happens in a thousand years time or anything like that. So I think we just have like small ‘n’.

Robert Wiblin: Okay. I’ll look up this example of this Italian family and see whether that’s a complete figment of my imagination, in which case I’ll probably edit it out and you won’t be hearing this listeners. But if it does exist, we’ll stick up a link to that in the show notes.

Should we be concerned about one group taking control of the world? [00:34:51]

Howie Lempel: To what extent should we be somewhat concerned about the idea that a set of people with a particular and maybe narrow set of values is going to be the most patient and then end up accumulating a really large amount of resources down the road and really sort of control a lot of the economy, a lot of the world.

Philip Trammell: Yeah. I think that is a possibility and it’s a risk from the perspective of someone not sharing the idiosyncratic values in question. And that’s why I think this is an important line of reasoning to think through whoever you are. Whatever your values are. Whatever you care about. Whatever your vision for how society should be. If you don’t think through the timing question right, then you’re just leaving the future on the table and whatever causes you care about or whatever moral theory you endorse says about how the world should be as long as it doesn’t incorporate some pure time preference. As long as you don’t explicitly reject the idea that you should care about what happens in a long time. You do have to contend with the fact that if you don’t contribute to saving up resources to do what you want in a long time, then other people will. So, I think there’s just no escaping this question and people of all kinds should just collaboratively think about how to spend their philanthropic resources over the course of the future.

Robert Wiblin: I guess some people also might find the idea feels a little bit cold or heartless in a way because you’re potentially just not contributing to solving the very salient injustices that are around today. I guess one thing that might help with that is just trying to make it more vivid and more salient, the potential injustices that could be around in hundreds’ of years time and the fact that you’d be able to potentially make a huge dent on those by having way more resources. Do you have any comments on that?

Philip Trammell: Yeah, I think that’s exactly right. I just want to emphasize that the example I raised at the beginning about how, from a utilitarian perspective, you can do more good by just saving up for amazing drugs or whatever for the super rich people in the 24th century really was what we said it was, which was this very handicapped argument about how, in some sense, the argument can go through even if you put a lot of restrictions on it. But, a big part of my own motivation for thinking through this problem as clearly as possible is the thought that there will be all kinds of injustices or immense tragedies that I hope philanthropy will have a role in solving in a long time, and that we shouldn’t ignore them in favor of the ones that are present today. Not to minimize those today, but just to put as much of a dent in them as we can regardless of the timescale.

Finding better interventions in the future [00:37:20]

Robert Wiblin: All right. Let’s push on to another kind of different consideration now, which is that potentially, over time, we’ll actually just find better things to do with the money and not even just buying bizarre luxury goods in 279 years’ time. Do you want to talk about this point?

Philip Trammell: When we look back over the last thousand years or whatever, and we wonder whether resources would be better put to philanthropic use by those lives at the time or invested for use by ourselves or other seemingly relatively thoughtful philanthropists today, we tend to think that the latter would do more good. So just by raw induction, by a sort of outside view, we should defer to our inheritors’ decisions on this front. We should expect that they have more information or have just reflected more on what is good. Also, we have been learning more about how to do good empirically in a variety of domains, right? So the whole field of development economics is relatively new, you know, just sort of a few decades old and it’s already undergone a number of revolutions of thinking of how much to prioritize randomized controlled trials and other methods and I think we’re still improving our ideas on that front. The whole idea of improving the welfare of farmed animals is a relatively new philanthropic concern and we’re still learning a lot there. And when it comes to how to influence the long-term, that’s also a field about which until recently, people hadn’t given a lot of attention to and we still seem to be learning a lot and there still seem to be a lot of open questions about very long run growth and long-term trajectories and population and moral values and risks to civilization and so on. So yeah, both my inside view of the state of the research and the evidence and the outside induction based view point me in the direction of thinking that people in awhile, maybe not a million years, but people in awhile should pretty robustly be expected to do more good with resources than we can today.

Robert Wiblin: Yeah. How confident are you that money spent today is better than what a thoughtful person who was able to put in as much time perhaps as Open Philanthropy is today. How they would have spent it in 1960 or 1900. I guess it’s slightly hard to know what they would have spent it on. It’s possible that they might’ve done something really impressive to do with nuclear weapons or doing research into what things are most important.

Philip Trammell: Yeah, I agree. It’s ambiguous. I guess I can just report my intuition on that. I think there might be a little bit of a trick going on when we look back and the list that you just gave, right, was the list that we have now about what seems best given the hindsight. And you can always cherry pick individuals who were actually working on those things at the time and say, “Well, if we’d given it to them, then they would have been able to keep it as opposed to investing it for our sake, they would have done more good. But the problem is that we don’t know who we are among the people today. Maybe we’re the people in 50 years’ time that people will look back and say–

Robert Wiblin: “Why did they think that was a good idea”?

Philip Trammell: Right. But obviously people 50 years ago, say, the people of 2069, could have done more good than us today in 2069 because Joe Shmoe who did the thing that now in hindsight seems great, knew what he was doing.

Howie Lempel: Just to make the counterargument a little bit. It does seem possible that in the fifties or sixties or seventies an altruistic philanthropist with a really global or long-term perspective, preventing nuclear war actually might’ve been among their top priorities and was something that foundations at the time worked on. The elimination of smallpox was another big philanthropic priority. Or at least an altruistic priority that might end up having some longer term effects. So there is some reason to think that people in the past may have made quite good use of money.

Philip Trammell: So yeah, I agree. There’s arguments on both sides of this question. It’s hard to put a number on it or anything like that.

Robert Wiblin: Yeah, so I guess in my mind at least it seems like we’ve come up with lots of really important crucial considerations in philanthropy over the last couple of decades. Not many people were thinking about existential risks, now we think about that. We think about the very long-term future, much more than people did before. I guess things like the simulation hypothesis from Nick Bostrom which I guess… It’s not clear that that actually changes what we ought to do very much, but it seems like it could potentially be a big deal. Are there any other kind of crucial considerations that stand out and maybe should we expect to have big revolutions in how we think about how to do good over the next couple of decades?

Philip Trammell: Yeah. It seems clear to me that we should expect about as many revolutions as you think we’ve had over the past few decades. I don’t think there is any big milestone. I think a big finish line that we crossed–

Robert Wiblin: I’ve become an adult now. It’s basically all over now. I think we figured it out. At least I won’t change my mind.

Philip Trammell: Right.

Howie Lempel: Something in favor of thinking that we’re likely to change our minds in the future is an experience that I had that I think a lot of other people in the effective altruist community have had of a decade ago having very different priorities than today. So there are, I think, a lot of people also who’ve the experience of giving mostly to global public health and antipoverty work 10 years ago and then heard arguments for working on interventions geared more towards the long-term future. Geared more towards x-risk reduction, recently. And so if you have a ton of people who made that change over the last decade, you should feel pretty unsure about what you might learn in the next few decades.

Philip Trammell: Yeah, exactly. Obviously this argument can’t run forever because there’s going to be some sort of stabilization of views after which you do get to feel confident that you’re doing something you won’t come to regret. Maybe if we haven’t come up with any crucial considerations after a hundred years or something, then we should go ahead and start giving. Formalizing what exactly that number is, that’s the domain of an academic field called optimal stopping. Someone wrote a popular book on this stuff called, “Algorithms to Live By”. You had the author on the podcast.

Robert Wiblin: Yeah. Brian Christian. One of our most popular episodes, actually.

Philip Trammell: Oh really? So I’d love to see how optimal stopping models apply to the question of timing philanthropy and learning. But in any event, just intuitively, it seems like we’ve been coming up with them quickly enough that we haven’t crossed any sort of finish line.

The hinge of history [00:43:46]

Robert Wiblin: So far we’ve got just the absolute amount of money you have is increasing. Then we’ve got your fraction of all global wealth is increasing pretty fast. And then we’ve got that we actually have very good reason to think that in the future we’ll be wiser and have better ideas about how to spend the money. I guess a fourth consideration is that there might be moments in the future that are just more important where, even setting aside what you know, there’s just going to be greater opportunities to have an influence. I guess you call this “hingeyness”, although some people are not huge fans of the term and maybe we should find something that sounds a little bit better.

Philip Trammell: Yeah. So Derek Parfit, the late great philosopher, made the claim in this book in 2011 that we were living at the hinge of history. And this is a claim that right now the actions that we take will determine whether we survive as a species. Whether we misuse nuclear weapons and other dangerous technologies or use them to last a very long time and perhaps even colonize space. So yeah, that’s an argument for now, plus or minus a few hundred years, being a really important episode. But it doesn’t seem to me like this binary thing where you’re either at the hinge or you’re not. It seems looking back through history, one can identify peaks and valleys to this property of being an important time. You know, some key battle in World War II was a very pivotal moment. So yeah, just out of deference to that phrase, I like to call this “hingeyness”. But that’s the idea. And likewise, looking ahead, one might wonder whether instead of spending now, we should be saving up for that key battle in World War III or something, or the constitutional convention after World War III that will, you know, changing one letter of which will be baked into the world government that lasts for however long.

Philip Trammell: And I think likewise, one might try to estimate how big those peaks and valleys are from looking back over the past few centuries or something. But it’s funny because other people have noticed that they were living at really, really important, really hingey times in the past as well. So there’s this great quote from John Adams, that he foresaw that as he was helping to write the US constitution, he was setting up institutions that would govern a large part of the world for thousands of years. And that would influence the institutions that followed them. And so far it seems like he was right in a big way. So I think it’s probably a bit of a stretch to say that 2011 or 2019 or something is like the hinge and so we’ve got to go ahead and spend right now and we should be more open to the possibility of unforeseen important moments in coming centuries or whatever.

Robert Wiblin: And I guess to be clear, when we’re thinking of hingey moments, pivotal moments in this context, we’re not just thinking of an election that might affect the next four or five years, but things like constitutions where it could potentially last hundreds of years or I suppose, conceivably, tracks that we could put humanity on that would just last indefinitely as long as humanity and its descendants continues. So we could be talking millions or billions of years here. So these could be really crucial moments.

Philip Trammell: Yeah, I think that’s right. It’s often hard to know just how long-lasting a given change will be. Because a constitution or a religion or something can, in subtle ways, shift the thoughts and behaviors of the people setting up the institutions and value systems and so on that will replace them and so on. So I don’t know if there’s a really obvious way to parse them, but when you come across different opportunities for a long-lasting impact or whatever but the aim, ultimately, is at doing as much good as possible over the very, very long-term and explicitly not about just winning an election that could very likely just be reversed in a few years’ time.

Robert Wiblin: Yeah. So we’ve got some examples there of, I guess, wars, writing constitutions. Are there other events that might be particularly hingey, that we should have in mind when we’re thinking of this concept?

Philip Trammell: Yeah. Perhaps the development of technologies that could have gone one way or another. Like nuclear weapons maybe could have. I’m not sure. I’m not an expert in any particular technology, but the introduction of… Or just like the development of ideas. So one might think that the axial age as it’s sometimes called, the age in which most of the world’s big religions were started sort of 2000 years ago plus or minus a thousand years or something was a really important time in shaping all of the moral intuitions of the people that followed, even if they’re no longer practicing members of the religions in question.

Robert Wiblin: So there was this era when I guess, it was Christianity, Islam, Judaism, Buddhism maybe, that all kind of got going around that era.

Philip Trammell: Yeah, toward the earlier parts of it you’ve got Zoroastrianism and various schools of Hinduism as well. Confucianism. All the Chinese philosophies.

Robert Wiblin: Interesting. Yeah. Do people have a theory for why? I guess maybe writing was kind of taking off then?

Philip Trammell: Yeah, I think it’s something like that. I would imagine that all of these guesses are rather speculative, but it’s something like once societies get very large scale, there’s a big cultural evolutionary advantage to having systems that can keep everyone following the same sort of sets of rules and coordinating and so on that there aren’t when societies are small and people can just sort of live by family trust and tribal intuitions or whatever. And then there was some period of time after the agricultural revolution when in various parts of the world societies got big enough that this became a big advantage. And so different ones caught on in different places. And so it’s something like that.

Robert Wiblin: So I guess we can envisage going to the early Christian conferences in the year 40 or the formation of the Catholic church–

Philip Trammell: Or the Council of Jerusalem.

Robert Wiblin: Right! Do you know anything about that? I don’t really know that much about the very early Christian history.

Philip Trammell: Sure. Questions about perhaps relatively small seeming things like whether Christians would have to be circumcised or whatever were decided at some of these councils and then affected billions of anatomies for the subsequent future.

Robert Wiblin: So a lot of this stuff just wasn’t revised later on. It really got stuck in?

Philip Trammell: Yeah. That’s an important point. So for a moment to really resemble a hinge it has to be really, really pivotal. It has to have some element of lock-in. So you make the change and it lasts for a long time. And I think an argument for this religious age being particularly moment is that religions do have sort of an endurance to them. Like once a rule gets set, it has some sanctity and so it’s not going to just be changed willy-nilly.

Robert Wiblin: I guess people also lose track of the reason why the decision was made in this case. So it becomes hard to revise because you can’t see all the drafts and see the comments thread on the Google doc and realize that it could have gone either way. There wasn’t really a good reason. That seems to be something with religions in particular.

Philip Trammell: Yeah, interesting. That might have something to do with it.

Howie Lempel: Although some religions seem to actually have done a pretty impressive job of tracking all the commentary or at least a big portion of it.

Robert Wiblin: Yeah. Interesting. So something that seems to be the case with all of these examples when we look back, is that we can be like, “Oh it was this potentially quite narrow period of time and this specific location which was very hingey”. Which sometimes differs I think from when people talk about, “Oh you know, this century is going to be especially hingey”. They’re talking about a very broad amount of time and a very broad amount of space and maybe what they mean to steelman it is that there’ll be important moments in this century but maybe there’ll be only one year in particular that’s especially important. But that does seem to potentially change it because if you think there’s a difference between, “It’s this century” and “It will be the year 2057” because then you’d be like, “Well, we definitely should be saving now if it’s going to be 2057”.

Philip Trammell: Yeah. I think that’s an important point. So there’s been some online discussion about whether we should really think that this is the most important century ever or something or the next few centuries are the most important time ever. And as far as I can tell, in a lot of people’s minds, this is an important crux of the question of whether to try having impact now or doing something really broad based like saving money for a long time or just growing a movement of reflective philanthropists that isn’t centered on any particular project, but just aims to grow over time. And I don’t think that’s quite right. So to go back to that Parfit quote, he says if we navigate through the next few centuries right, then things will go really well and if not, then we’ll blow ourselves up. And that’s what makes these few centuries the hinge of history. A few centuries is a long time. It’s much longer than most people are currently thinking about when to allocate their resources. I would still advocate investing if you thought that we were on this high plateau of hingeyness for the next 400 years because you would just have way more resources to spend in year 350 or something if you invest then and that would do way more good. So the argument for committing a lot of resources now really depends not on these few centuries being a hingey episode of history, but on this year or the next 10 years or something being extremely disproportionately influential over the future or especially rich with opportunities to do good.

Robert Wiblin: Yeah. So even if 2030 is somewhat more hingey than 2020, we may have been able to double our amount of resources just by investing it by then, and then also potentially get this benefit that it’s 2X as chaotic or 2X as hingey as this year and it looks like a pretty good deal. And in which case, save rather than give.

Philip Trammell: Yeah, exactly. I don’t know if chaotic is exactly the same as hingeyness, but anyway I know what you mean.

Robert Wiblin: What is the relationship between chaoticness and hingeyness? It does seem like you need it to be possible for things to go off in very different directions and sometimes things not being locked in, like normal institutions not functioning to just all push us towards one particular direction can make for hingeyness.

Philip Trammell: The distinction is just that the path has to be very sensitive to the initial conditions. So that’s what makes it seem like chaos, but it has to be predictable as well. So it can’t be a butterfly effect thing.

Robert Wiblin: So it has to be chaotic enough that things can be moved but not so chaotic that you can’t get any traction at all because it’s unforeseeable what effects actions will have.

Robert Wiblin: Yeah.

Robert Wiblin: Yeah. All right. So we’ll put up a link to this really neat blog post that Will MacAskill, who’s also been on the show, wrote about this question of whether the next century is likely to be among the most pivotal of all of the centuries that are yet to come. And he makes this argument that it’s very unlikely that this century is of special importance. One of the arguments he puts forward is humanity could continue for another million centuries. And so your prior, that is, before you’d even looked at the question, you might think, “Well maybe the odds are only one in a million”. And then if you try to collect evidence on, “Does this century look especially pivotal thinking about what other pivotal moments there might be far, far down the line in hundreds of thousands of years time. This doesn’t seem super compelling to think that this century is so much more compelling than say the point at which we start colonizing the solar system or leave the solar system or put together world government. That kind of thing. I guess, in particular, he seemed quite concerned that we would have this big bias towards thinking that the moment that we’re in is especially pivotal because it’s so clear to us the ways that it could be pivotal and maybe just also a bit of natural human narcissism. And so even if you look around and think, “Oh, this time looks especially pivotal”, well you should be really skeptical of that especially given that it’s such a hard question to answer, you should always worry, “Well actually this is just a perverse aspect and kind of my perspective on the world around me and so you’d never be able to update that far away from your base prior of it’s a one in a million chance. So maybe you could get up to ‘it’s a one in a thousand chance’ or maybe even ‘it’s a one in a hundred chance’. But we should never really be confident that this century is definitely the one.

Philip Trammell: Yeah, I’m sure there’s some amount of evidence that should lead one to conclude that this is the most important century ever. But it would have to be a lot. It would have to overcome that prior. And we should be careful to guard against the biases that might push us in the direction of thinking that we’re more important than we really are. That said, the argument for spending doesn’t really rely on this being the most important century because I think there is some background rate at which invested assets do run the risk of getting stolen or your inheritors drifting from the values that you would have wanted the resources to be put to. So, I think for practical purposes, you just need to say that this is the most important one out of the next 10 or something for it to be worth spending now rather than in a long time. But even so, I think it’s just not clear to me that this is the most important century out of the next 10.

Robert Wiblin: Yeah. Interesting. I have to admit it was a very nice blog post, by Will, and then the comments were also very good for internet comment sections. There were a few really good points made there by Toby Ord and Carl Shulman. Maybe Greg Lewis as well. Where I think they offered some pretty good counterarguments and reasons why we should attach a much higher prior probability to earlier centuries being pivotal than later ones, which would mean that maybe your prior on this century being the most important of all the ones to come might be more like 1% to begin with and then if you look around and think, “Oh wow, there’s such important things going on and there’s such great opportunities to affect the trajectory of things forever”. Maybe you could end up with a 10% probability all things considered of this century being an especially good opportunity. Or that there will be moments in this century that are especially good opportunities to have impact. And I guess that’s maybe where I come down right now, although it’s something that I really need to read more about and evaluate these arguments.

Philip Trammell: Yeah. I also need to think about this more, but I would just say again that those seem like good arguments for not attaching a totally flat prior. So you alluded to one pretty straightforward argument which is that earlier centuries can influence what happens in later ones. So there’s kind of no way that the very last one is going to have that much room for impact because everything’s about to end. But it doesn’t really seem to be a strong reason to think that there’s just way more for a patient philanthropist to do now than there will be in like a thousand years time.

Howie Lempel: Yeah. Also, if you have an existential catastrophe in any given century, then you know that means that future centuries can’t affect the long-term future at all. And so there’s some base rate at which at least one force is making century after century a little bit less hingey.

Philip Trammell: Yeah, I think that’s true. This might be a good time though to bring up perhaps a somewhat subtle point that’s central to all this, which is that at least some of the forces making a given year or century hingey, or making it seem more favorable to spend now than to wait a year, should be expected to be reflected in the interest rate. So for instance, if the risk of an existential catastrophe is very high and everyone knows that, say it’s 1% per year for awhile. Then, on the one hand, that’s a reason to spend now, but on the other hand, that’s a reason for everyone to demand 1% more or something resources next year in exchange for one this year. And if everyone’s seeing things the same way, then you should be back to being indifferent between spending now and spending next year except for that just impatience bit, which would still leave the interest rate even higher than what would be justified on the basis of the extinction risk.

Philip Trammell: And I would say that to maybe a lesser extent, the same things are true with respect to other sources of lock-in. So, if we’re on the cusp of a world government and, you know, everyone’s fighting over what will be written into its constitution, you’re going to have a lot of people trying to borrow so that they can hire their lobbyists and fling them at that effort and so the interest rate will be very high. And so it’s still not clear whether you can do more by spending now or by waiting.

Robert Wiblin: So that’s kind of what theory would say. But is that what we’ve seen historically, like during the Cuban missile crisis? Was there a sudden spike in interest rates and during World War II, did interest rates go up a whole lot? Or when the US constitution was being written?

Philip Trammell: I think they did in World War II. And there was all this effort to get people to buy war bonds and stuff. But, I think the key is, I mean, we’ll get to this a bit later when we talk about the timing of career efforts. But different kinds of resources will be able to exact different amounts of influence at different times. So even though we’ve been talking in broad strokes about hingeyness as if it’s like the single thing that fluctuates over time. That might be true sort of broadly speaking, but there’ll be particular assets that are hard to convert into others very quickly. And so from time to time, some might be temporarily much more influential than others. And during the Cuban missile crisis, and sometimes it was pivotal, but it doesn’t seem that money could have done a lot to change how it went. It was more like US presidents and other world leaders. And so yeah, there’s not really a market for those. But if there had been, I would certainly have expected the interest rate to have been very high around the Cuban missile crisis.

Robert Wiblin: Yeah, that makes sense. I guess there’s this other reason that you might expect the interest rate to go up, which is that people are worried about dying and so they’re like, “Well now I’m going to throw a party and cash out my investments”, because there may not be a tomorrow.

Philip Trammell: Yeah. So if it was known at the time during the Cuban missile crisis just how risky it was, then you should have expected interest rates to spike around then. I think it wasn’t known and so interest rates didn’t go up that much.

Robert Wiblin: I think even if people had known, even if people had thought there’s a 20% chance there all about to go up in smoke and my investments won’t be worth anything because they won’t be a stock market, I do wonder whether there really would have been such an increase or that asset prices would’ve changed all that much because you’re a rich person at that time. What are you really going to do with all the money? You’ll cash it all out and then there’s just no way to spend it fast enough over that period of a week. Like for an epic party, you don’t even have time to organize it. There’s nothing you can really do to blow your money that quickly. I think Alex Tabarrok has a nice blog post where he explains why it is that during seemingly very risky global moments, we don’t see that much of a shift in asset prices. One thing is that they kind of all become worthless simultaneously. So it’s like the relative prices of different assets don’t matter too much. And he’s like, “Well, but you could just go and turn it into consumption right now”, but then what are you really going to do? At least over a period of a week.

Philip Trammell: Yeah. That’s interesting. I haven’t seen that blog post, but I would say talking about the timescale of a week seems like it’s sort of pushing the theory a bit. But I’d be surprised if during a few years of really high risk or something, you didn’t see some shift toward present consumption.

Robert Wiblin: Yeah, and I guess his model would support that as well. Because there’s just much more you can do to spend money over a couple of years if you think you’re quite likely to die in a way that’s just not practical over just a period of days.

Philip Trammell: Yeah.

Does uncertainty lead us to wanting to wait? [01:01:52]

Robert Wiblin: All right. Let’s move on from this question of “What’s the most hingey moment?” to another general topic that I’m not quite sure what to make of. But, let’s just say that we are very uncertain about what will be the growth rate of the global economy, what will be the interest rate, when will the most hingey moment be? It seems we should have some diversification here. Have some kind of portfolio of different strategies. Are there arguments that if we’re uncertain, then that pushes us in favor of wanting to wait quite a long time. Have you thought about that side at all?

Philip Trammell: Sure. So if we’re uncertain about, for example, what investment returns will be, because the returns compound, most of the value is in that tail of optimistic possibilities where returns are surprisingly high. So it’s much better to invest when there’s a 50% chance that returns will be zero for a few hundred years and a 50% chance that there’ll be 10% per year for a few hundred years than to invest when you know they’re going to be 5% per year for a hundred years because 10% periods have just so much more impact. And likewise, if you’re not sure what the expropriation rate will be, what the value drift rate will be, what the rate at which we’ll be able to come up with the institutions and technologies to really stably fund causes over long time periods will be? Most of the value is in the cases where that rate is low.

Philip Trammell: And so if it turns out that the rate is close to zero, you’re able to set up that trust that does go over thousands of years time, and if it turns out that it’s high, then it just dies in a few years. But the expected lifespan there is still much higher than if you know that it’s a moderate number. So in the face of uncertainty about these parameters, it turns out that the expected value of investing is higher than it might intuitively seem.

Robert Wiblin: So inasmuch as you have uncertainty about these things, the higher variance increases the value of the savings strategy relative to the “have impact now” strategy.

Philip Trammell: Yeah.

Robert Wiblin: I suppose because the “have impact now” strategy doesn’t have this compounding aspect to it, which makes uncertainty appealing?

Philip Trammell: Yeah. In general, there might be some intervention you could enact now about which you’re not sure whether the effects will just peter out or whether they’ll compound or something and then there’d be some symmetry there.

Howie Lempel: When you think about diminishing returns and the fact that in some of those big win scenarios where the money isn’t expropriated and returns are really high, that those will be the same worlds where you have tons of resources and so maybe your marginal dollar is less valuable.

Philip Trammell: Oh, sure. Yeah. That’s certainly something that’s built into the model as I’m trying to formalize it. I would say that the asymmetry consideration holds there as well. So, if you’re not sure whether returns diminish sharply or only very gradually, then most of the value is in the scenario where they diminish gradually. But in general, you can be pretty confident that most efforts will have at least some degree of diminishing returns and that’s the reason to consider diversifying across strategies.

Philip Trammell: So if you think it just might be really important to have at least some resources going toward some cause. You want at least one lobbyist caring about the very long term at that post world war three constitutional convention and then having a second isn’t quite so important. Then, you’re going to want to put some resources toward some fund that aims to make use of these important moments as they come along. But then likewise, efforts in the present to reduce the risks of dangerous technologies or make use of the lowest hanging charitable fruit available right now. Those are going to have diminishing returns and so you’re going to probably want to diversify.

Howie Lempel: It also seems like diminishing returns in the future will only be a really big factor if you yourself or your community, is a big proportion of all of the work going into the areas that you’re working on. It’s possible that if the whole globe is focusing on the same thing that we think is crucial at the time, then diminishing returns on the scale of whatever trust you set up is not going to be very large.

Philip Trammell: Yeah, I think that’s right. And I think this hearkens back to another thing in terms of parameter uncertainty, which is that we don’t know to what extent future players, future governments and philanthropists and so on, will care about the sorts of things that are maybe a little idiosyncratic at the moment, but that we spend a disproportionate amount of time thinking about. Like the well-being of animals or people in the distant future. Whatever a utilitarian approach to philanthropy or something in that direction. So there is in some sense a risk to the project of the fund that these cause areas will get very popular after a while and then the fund will not be doing so much good because the returns will have diminished. But that seems far from certain. And so most of the value is in the scenario in which marginal returns are still very high.

Robert Wiblin: So we’ve been through five considerations here. Is there anything else that you want to add? Like arguments in favor of waiting to have an impact before we move on to counterarguments to these and reasons to go hastily.

Philip Trammell: Yeah. I think one last point is that just as one can make arguments for and against for now being an important time to spend, there’s also a sense in which we should perhaps think that now is a somewhat uniquely important time to save. And that’s that at the moment, as we discussed, most people are quite impatient and that means that interest rates are relatively high and one can buy that future patch of land. And one can really buy, in some sense, the right to determine what happens to a lot of the Earth’s resources well into the future from the current owners of those resources. You don’t care so much about what happened to them. You can just lend at a high interest rate. That can’t really persist forever. As time goes on, you would expect patient actors to take advantage of this opportunity and come to own more and more of the world’s resources. And if you just play that forward, ultimately, you might expect that everything or like something approaching 100% of the world’s assets to belong to the patient and when that happens, this opportunity will be gone forever. So that strikes me as a reason to really think that this too might be a fleeting opportunity.

Robert Wiblin: Yeah. So I guess you’re saying we have this brief moment in time when we might be able to take advantage of this opportunity and if we wait a couple of hundred years it might be gone. We might have forever given up the chance to save and have a much larger influence later on.

Philip Trammell: That’s right.

Robert Wiblin: I guess that seems slightly odd because interest rates in the past, at least since we started having lending, have always been high. So maybe you’re saying this is a special moment in that this is maybe the last window to get in on this opportunity. It’s been around for 5,000 years, so then it would be a bit of an interesting claim to say, “Well, but it’s going to disappear”.

Philip Trammell: Right? Yeah, that is a good point. I mean I would say that what needs to happen is institutions need to be stable enough for property rights to be secure for long enough for the patient really to be able to take advantage of this opportunity. And if it’s the case that someone in ancient Greece or whatever could have tried this, but then would’ve gotten pillaged by the Persians a few years later and so on, then it would still have been a high expected value strategy. But interest rates would’ve just been really, really high because you got pillaged every few years and the expected lifespan of these efforts would have been very short and so we just wouldn’t really have expected any of them to last. But now interest rates are lower, property rights are more secure, things are a lot more stable.

Philip Trammell: And it could be that they get much more stable. Like if nothing goes too far wrong with the US and China or whatever, we could be on the brink of an era in which foundations… Just like American foundations have managed to last for centuries. Efforts of this kind could just last for 500 years or something and then they really would start coming to own a substantial fraction of the world’s assets. So I wouldn’t say this is all that likely or obvious or whatever, but it just seems like a possibility worth considering. Maybe we are coming somewhat near the end of the era when this is possible.

Robert Wiblin: Yeah. Interesting. I guess this is kind of an aside, but I’ve seen papers arguing that people are inclined to massively overestimate the real rate of interest and the real return on capital because they often look at stock markets that still exist today and they tend to ignore like the Russian stock market, the Argentinian stock market and the German stock market. Like many of which went to zero. Yeah. So it was this survivorship bias that you get and if you include those, then the real return on investments goes much down. I suppose you might think, “Well the 21st century will be more stable because it seems like we’ve had fewer conflicts”. But I guess there’s always a risk of nuclear war that will put a lot of stock markets down to zero.

Philip Trammell: Yeah. So that is a good point. For the toy numbers I use in the paper that we discussed, I don’t use that 7% number from the US stock market because I think it is, to some extent, subject to the survivorship bias and so on. And, in fact, returns have been declining and growth has been declining at least in the developed world. But on the other hand, there are opportunities to earn above market returns which one can take advantage of if one is sufficiently patient or willing to tolerate volatility and so on, because one’s aiming to invest for a long time so you can invest in highly leveraged assets and–

Robert Wiblin: You can ride out the highs and lows much more than someone who’s saving for a time and can.

Philip Trammell: Yeah, exactly. And also just if you’re sufficiently big, you have certain investment opportunities with some fixed costs for getting into that smaller players don’t have access to. So if we’re thinking about what the community of patient philanthropists as a whole perhaps should do. What Open Phil should do or something. You might expect them to earn even higher than the stock market returns on average over the long run like US university endowments have historically done. So, I’m not sure how all these considerations push on balance, but I think the reasoning goes through, even with substantially less than 7% returns.

Counterarguments [01:11:36]

Robert Wiblin: All right. So I feel like we’ve done a pretty good job here for the last hour of laying out the reasons in favor of waiting to have an impact. So let’s turn now to arguments in favor of having a greater sense of urgency about having an impact. And it might be convenient to split this into two different sections. So one will be counterarguments to what we’ve been talking about so far and then maybe we can move onto independent free-standing arguments in favor of trying to do stuff right away. Sound good?

Philip Trammell: Sure.

Robert Wiblin: All right. So I guess a counterargument that probably a lot of people have been shouting into their iPhones or podcasting software has been that you could potentially have a lot of impact now by building a movement, or recruiting people, or spreading ideas, engaging in advocacy, that grows the amount of resources that are focused on the things that you’re interested in. And, of course, it seems better to start building a movement sooner because you get this compounding growth that might be well above the 7% or 5% or whatever that you can earn on the stock market. What do you have to say to people who are thinking that?

Philip Trammell: Yeah, I think that’s a really good point. In fact, in this write-up, I do try to make it clear that by investment, I really am explicitly including things like fundraising and at least certain kinds of movement building which have the same effect of turning resources now, not into good done now, but into more resources next year with which good will be done. I would be just a little careful to note that this has to be the sort of movement building advocacy work that really does look like fundraising in the sense that you’re not just putting more resources toward the cause next year, but toward the whole mindset of either giving to the cause or investing to give more in two years’ time to the cause. You might spend all your money and get all these recruits who are passionate about the cause that you’re trying to fund, but then they just do it all next year.

Robert Wiblin: The fools!

Philip Trammell: Right. And I don’t know exactly how high fidelity in this respect movement building tends to be or EA movement building in particular has been. So that’s one caveat. I guess another one is that when you’re actually investing, you’re generally creating new resources. You’re actually building the factories or whatever. Whereas when you’re just doing fundraising, you’re movement building, you’re just diverting resources from where they otherwise would have gone.

Robert Wiblin: You’re redistributing from some efforts to others.

Philip Trammell: Yeah. And so you have to think that what people otherwise would have done with the resources in question is of negligible value compared to what they’ll do after the funds had been put in your pot. And you might think that if you just look at what people are spending their money on, the world as a whole… I mean you might not, but you might. And if you do, it might seem like this is a safe assumption to make, but the sorts of people you’re most likely to recruit are the ones who probably were most inclined to do the sort of thing that you wanted anyway on their own. My intuition is that it’s easy to overestimate the real real returns to advocacy and movement building in this respect. But I haven’t actually looked through any detailed numbers on this. It’s just a caveat I would raise.

Howie Lempel: What do you think about changing the character or trajectory of a movement that you’re concerned about as opposed to just making it bigger?

Philip Trammell: I suppose I haven’t thought about that explicitly. A simple model of that might be that if you just change the direction it’s headed in, in such a way that multiplies everything they do by some factor. If you make everything that they do twice as effective or 1% more effective, then it multiplies all their resources by 1% times that diminishing returns number we discussed it earlier. And then it proceeds from there. I definitely feel the force of the intuition that this is a good thing to do. I think that to some extent, the effective altruism movement might’ve been overprioritizing spending rather than investment like movement building and so on. I hope that I’m improving people’s thoughts about that question and thereby multiplying the effectiveness of the movement as a whole to some extent.

Robert Wiblin: So, in a sense, you’re in furious agreement with the people who are saying, “Oh we should work on advocacy or movement building at least for a particular kind of building a movement or promoting ideas”. That is, building a movement that then just spends more time building itself even more, or focuses in large part on that to just keep on getting these compounding returns. But you can imagine what this looks like in 10 years’ time which is that you’re turning into some kind of multilevel marketing scheme or some kind of Ponzi scheme where just all that anyone does is promote the ideas and then when everyone’s like, “Ah, should we have an impact”? They’re like, “No, don’t have an impact. We just have to build the thing”. Wouldn’t we then need to have people actually doing useful things in order to promote the ideas. In order to just avoid looking like a ridiculous group of people and also to actually find out whether it’s an idea worth promoting at all or whether you should move on to something else?

Philip Trammell: Sure. I think the question is just about the rate at which one should be spending and picking that low-hanging fruit for impact or for learning.

Robert Wiblin: Yeah, so like, “What’s the right balance”?

Philip Trammell: Yeah, exactly. I’m not saying that it should be all about investing. I mean, there are cases in which it should literally just be investing if the returns are high enough and if enough other people are picking the fruit that you would’ve picked. If people are thoughtful about it, it’ll never really be a Ponzi scheme. The model does close. There comes a time when you’re big enough that the returns–

Robert Wiblin: I guess there are no easy ways to grow anymore: you’re plateauing out.

Philip Trammell: Well you could be plateauing, but it could also just be that because of the diminishing returns to spending, even if you could still grow it, like grow your assets 7% by waiting another year, you’ll be growing your impact some tiny fraction of that. And so it actually becomes best to start spending. So it’s just one of those things that unfortunately might look like a Ponzi scheme for a little while, but wouldn’t literally be one. That said, I guess I place some weight on the consideration you raised that it would look bad and that that counts for something. But, I don’t intuitively place a lot of weight on it, because I think part of the whole brand appeal of effective altruism in particular is that, “Oh, we’re the people willing to do the thing that the numbers said was best, rather than what kind of looked best on a thirty-second glance”. None of the reasoning’s going to be secret. I mean, if this is what a lot of people end up doing and people ask why, then there’ll be a link to a paper. There’ll be a link to an online little tool which we’ll link to and you can all play around with it if you like to clarify your own thoughts about the rate at which you think patient philanthropists should be investing versus spending. Maybe a 3D visual diagram. Maybe a virtual reality headset in 10 years. Once the idea is communicated, I think if it’s true, it’ll be compelling and it won’t seem fishy.

Robert Wiblin: Yeah. Well I think it definitely will seem fishy. I guess maybe I’m just a bad person in the effective altruism community because I feel like I’d be very concerned about a group that was dedicating almost all of its resources just to self promotion. You’d worry that it was going to be very impractical or that, I don’t know, you build a bunch of people who are very into this kind of theoretical model of when to have impact but wouldn’t actually really know how to accomplish stuff and it would at some point break down, like the right ideas wouldn’t be transmitted. I guess also the outside view is that groups that look like this, almost always in the past, have been scammers. So, yeah. I’d definitely have some concerns and I might want to try to create a thing that was similar but did invest something in trying to do good right away.

Philip Trammell: Yeah. So like I said, there will be that like low-hanging fruit for learning opportunities and so on which you should actually pick as you go along. And yeah, I think maybe I should put more weight on that outside view consideration as well, but it shouldn’t just be dictated by it.

Robert Wiblin: Yeah. PR concerns. To be honest, maybe what I’m saying isn’t true, because I remember back when I was first finding out about effective altruism and 80,000 Hours and all that kind of thing, I was very impressed with the fact that they were thinking about this timing issue and were open to the idea of saving money for extremely long periods of time and that was kind of a hot issue in the community at the time. Although yeah, I suppose I might have been disturbed if that was actually what was happening then. Just everything was going into the bank account. You’d really want to know who the trustees are of this bank account.

Howie Lempel: Yeah. Like your prior on some community that is just getting richer and richer and richer, and then just giving away all of their resources and using it altruistically is, for most people, probably incredibly low. So doing some work sooner just to demonstrate the fact that this actually is a community that cares about these things might be pretty important.

Robert Wiblin: I guess a huge endowment might also attract predators who would then try to take it over because they would see an opportunity to grab a whole lot of resources that are just sitting there.

Philip Trammell: Yeah. That is an important consideration which should just be built into the model I think. Like maybe the expropriation rate isn’t constant but grows with the size of the fund. On the other hand, I’m not super sure about that. I’ve heard a lot of people raise that point and point to some historical examples of large funds that have gotten expropriated or whatever.

Robert Wiblin: Seems like you could always fission it if that was the case. Just split it into two whenever it gets to a particular size.

Philip Trammell: Yeah, that’s one point. Another one is maybe you have more to spend on preventing getting expropriated. Or even just hiring mechanism design theorists to think about how to avoid value drift in some way or legal scholars to make it easier for people to sue the fund if they don’t abide by their original objectives and things like that. So this isn’t to sweep this under the rug, it’s very worth considering and I just don’t know exactly which direction ultimately points in.

Robert Wiblin: Yeah. We’ll talk about the risks involved in trying to actually build this in just one second, but first, research into crucial considerations and research into optimal timing and research into global priorities research. Does that all kind of count as an investment in the same way that movement building does because it’s an investment in knowledge in the future that you’ll be able to use in perpetuity?

Philip Trammell: I’m a little wary of just calling everything investment. This does seem like a common tendency, right? People will say, “Oh, I’m not spending now. I’m investing in our children by building schools or something”.

Robert Wiblin: Went to this fancy restaurant to invest in my image which maybe is how to get a job.

Philip Trammell: But yeah, I think there’s one very narrow case in which it does behave quite similarly to investment and that’s as I was alluding to before, when it increases the effectiveness of everything that follows by some constant percentage or whatever. So if you were literally doing a research project that made everything that whatever the community or the institution or whatever that was doing it, made everything they were doing 1% more effective or whatever, then it’s kind of like investing at a certain rate. But yeah, I mean you just have to make sure that the returns to doing that kind of work really are higher than the interest rate because otherwise you should just invest and then fund even more of that research in the future. And even though I am doing or trying to do this kind of work myself now, I haven’t actually done that calculation. So it’s not completely obvious to me that research is better than waiting now. I mean it certainly depends on the kind of research and so on.

Robert Wiblin: Except, of course, there’s one kind of research that can’t be delayed, Phil, which is research on optimal timing and optimal distribution of effort. Exactly the work that you do, coincidentally.

Philip Trammell: Yeah. I mean, as we mentioned earlier, there are pretty high stakes to getting this wrong. So if it were the case that everyone who spent their money and borrowed against all their future income and spent it all right this year if it weren’t for this one thought I had, then that would certainly multiply the effectiveness of EA activity down the line by quite a lot. So I do think it’s important. I didn’t pick the topic at random. I picked it because I think it’s important. But it’s just easy to say that what you’re doing is important without actually comparing it to this natural baseline of investment and that’s what I’d encourage doing, even though I haven’t actually literally done it in this case.

Robert Wiblin: I guess earlier we were saying that it seems like people have been coming up with important philosophical crucial considerations at a pretty good clip. I guess I haven’t thought about this that much, but it seems like it might be quite important to come up with those crucial considerations sooner rather than later. You know, even ones that aren’t about timing of resource expenditure. So do you consider that in favor of fundamental research being done now rather than saving and doing it in 20 years time or maybe just the investment returns are sufficiently large that it’s better to fund three times as much of it in 20 years’ time?

Philip Trammell: Yeah, I think it really depends on whether it would affect the donation timing because if these are just considerations that would affect whether we spend the money on A) or B) in a hundred years’ time, but it would only take 20 years to do the research or something, then you can do way more of that or way higher quality research into these crucial considerations if you start at year eighty. So I don’t think there’s any general rule saying that it’s better to come up with crucial considerations earlier rather than later. Only ones that could substantively affect the schedule.

Howie Lempel: I guess if you take a more extreme view and say there’s no incredibly strong reason to think that the hinge of history will be this year than any of the thousands of years going forward, it then becomes, I think, pretty unlikely that it’s important to do the crucial considerations research today as opposed to sometime in the future when you have more resources.

Robert Wiblin: Yeah, I guess inasmuch as you think there might be pivotal moments this century, then it makes a lot more intuitive sense if it’s important to rush forward with the knowledge that you have at that moment.

Philip Trammell: Yeah, I think that is a good point.

Robert Wiblin: Okay. Let’s move back to this issue of the risks involved with a fund that goes over many years. I guess, what kind of adjustments should we maybe make for the risk of the money being appropriated or I guess the people who are in charge of the fund just not sharing the values that we have now or the values that we would want them to have? Is that potentially a big factor in whether this was a good idea or not?

Philip Trammell: It’s certainly a big factor. I mean, it could always just be high enough that you should spend this second or this month or this decade or whatever. To try to get a better handle on it, ideally one would look back over the institutions of the past and try to estimate the rate at which they’ve drifted from their values. But there it’s very hard because it’s typically not very clear exactly what the values were. You know, they weren’t very well specified and even when they were, it’s not always clear how literally we should read them. So you know, this charity, ScotsCare that Will mentioned, I think, last time he was on this podcast.

Robert Wiblin: No, I don’t think he has.

Philip Trammell: But he mentions it. He likes to talk about it. This is like a classic dead hand charity that was set up to fund the poor Scottish people of London. And it’s hard to know whether the funder really cared about that or just saw that as an important problem of their own time. So yeah, I don’t have any kind of statistics on what the value drift rate has been. I can say that again, if we look at religions, at least some broad aspects of the world’s major religions which are very similar to the way they were thousands of years ago. Also these foundations like the Rockefeller foundation and so on from a few hundred years ago seemed to have stuck with their missions over time. And family trusts tend to have done quite badly. But thankfully the sort of thing that Open Phil, say, might give to, or might become or something if it took this line of reasoning seriously, would not really resemble a family trust.

Philip Trammell: So I think we can safely exclude them from the reference class. I should probably write this up more formally at some point and link to it. I don’t have it at the moment, but I did a cursory look at what seemed to me like the more relevant foundations and institutions that were set up over the past thousand years or something. I mean most of them in the past few hundred years. And, I came up with a very tentative value drift/expropriation rate of half a percent per year for ones that were explicitly aiming to last a long time with a relatively well defined set of values. So they tend to last something like 200 years. Before, I would say that they had drifted. One class there was medieval religious orders, which often seemed to exhibit this pattern of going back to the simplicity of the early church and really taking vows of poverty seriously and then having palaces 200 hundred years later and then there’d be like the resurgent Franciscans and Benedictines and so on which would be like a branch of the order that would kind of go back.

Robert Wiblin: So it’s a bit cyclical, potentially.

Philip Trammell: Yeah. I mean there it’s not really cyclical. I know what you mean. It’s just unfortunately, all the money that was given to the first Benedictines is still in the hands of… You know what I mean?

Robert Wiblin: Yeah, so you have breakaway groups, but they don’t get the money?

Philip Trammell: Right.

Robert Wiblin: Yeah. You mentioned ScotsCare. I guess I don’t know the details of that one, but that’s a foundation or a trust or or charity that was set up in London to take care of very poor Scottish people in London a long time ago which was apparently, at the time, a big problem but now the size of this charity perhaps has outpaced the number of Scottish people in poverty in London, tragically, at the moment.

Philip Trammell: That’s right.

Robert Wiblin: Do you know how long ago it was set up?

Philip Trammell: A few hundred years. I forget exactly.

Robert Wiblin: Yeah. I heard about an even more extreme case. I guess I’ll have to look this up and check that it’s not an apocryphal story, but it’s a foundation that was set up to take care of war widows in a particular village in England in the post world war II era, and tragically, Britain has not been in a major war over the last 60 years. So the number of war widows has been waning and I think the number of war widows in this specific village has become very small and is probably asymptoting towards zero. And I heard that this charity went to court in order to try to get their mandate expanded to take care of everyone in poverty in that village rather than merely war widows, but the court rejected it and said, “No, go and look harder for more war widows in this particular location”. Presumably at some point, when there’s literally zero, I would hope that the courts would reconsider. Well, maybe they shouldn’t because they should just hold onto this money, wait for World War III and take care of the many war widows in the post-nuclear apocalyptic world with an outstanding UK legal system.

Philip Trammell: Yeah. In any event, I think this highlights the general problem of figuring out whether an organization has really drifted or not.

Howie Lempel: How much do you think that this set of considerations affects your overall argument? I noticed you mentioning that the expected lifespan of a lot of these big, long reaching organizations is something like 200 years was your guess from the back of the envelope thing that you did and I’m curious about how much that magnitude matters?

Philip Trammell: Yeah. So again, interested listeners can do a little sensitivity analysis of their own by going to this discounting calculator and seeing what happens when you plug in a tenth of a percent per year drift rate or a 2% per year drift rate. But yes, it can matter a lot. So, in particular, if it gets higher than the impatience rate that people seem to build into their own decisions to consume or to invest, then that gets reflected in the interest rate. Then, that whole argument about investing for the rich people of, you know, 279 years from now doesn’t go through because the drift rate acts like an effective impatience rate for yourself and you should be indifferent between like a dollar now and a dollar and whatever the interest rate is next year. You should be indifferent between spending now or prefer if the rate is even higher. Then again I would point out that this does seem like a parameter about which it makes sense to have a lot of uncertainty. So, in the face of that uncertainty, there’ll still be a lot of expected value to be had in investing because it could be quite low.

Robert Wiblin: It’s actually been a slightly topical issue lately whether it’s the case that foundations might have too much fidelity, because it seems there were a lot of foundations that were set up in the 19th century which have goals that now we regard as not ideal. I guess a famous one is, I think, the Rhodes scholarship which was initially only available to white men and I think they needed an act of parliament I think in order to change that and open it up to broader groups. But there’s other foundations that have racist ideas or just what seem like antiquated ideas encoded into them and they’ve sometimes struggled to escape from that, although often courts will help them at some point. Yeah. So another risk is that there could be too low a level of value drift that, in fact, even if we give the foundation a pretty general goal of furthering welfare, perhaps in future we’ll realize that it’s a different conception of welfare that matters. Or maybe welfare isn’t the main thing and we’ll just have squandered our resources that way. Is that something you’ve considered at all?

Philip Trammell: Yeah, a little bit. I think this highlights the fact that what one’s theory of value is and how confident one is in it, is also an important input to the value of really setting up a fund with the targeted objective that really literally tries to act out the reasoning that we’ve been discussing. So on one end of the spectrum, if you’re just a hundred percent convinced of a very simple, easy to articulate theory, like hedonic utilitarianism, then you just write that into the fund’s charter and the trust’s charter and make sure that people can sue and that it’s easy for them to sue if they ever see it drifting from that. I think that’s maybe one reason why foundations have lasted a bit longer and religions even longer than that compared to family trusts and so on.

Philip Trammell: Because there’s a pretty clear objective and it won’t just sort of be reinterpreted and dissipate. But yeah, I’m not sure we should be so confident about that. And so, there’s going to be some right balance to strike between letting people do more reflection over time and change the goals of the fund in light of what seems valuable to them on more reflection and the risk that after a few generations it’ll end up just going toward whatever random winds it seems to have drifted to.

Robert Wiblin: I suppose you could give it a mandate of just, “Do the best thing by whatever people view the best thing to be in the future”. I’m not sure whether the law would permit such a vague mission statement at the moment, but you know, maybe in the future it would. Although I suppose then it’s so vague that it’s maybe problematic in the other direction because anyone can claim to be doing the right thing and then how do you litigate it to say that they’re misusing it when they spend it on themselves?

Philip Trammell: It’s a thorny problem for sure.

Robert Wiblin: Yeah. I guess this is a potentially interesting area for historical research or someone to look into this or legal research as well, to think about how ought this organization to be legally set up, so that it’s robust against these various different ways that it could fail and I guess, maybe, what is the best jurisdiction in which to operate now or the best legal setup that one could have if one felt that one needed to set this up relatively soon rather than waiting for a better legal infrastructure in decades’ time?

Philip Trammell: Yeah, that’s definitely a question for legal scholars. The best thing I’ve come across is that in at least the US and the UK and I think most of the English speaking world, you can set up charitable trusts which, as I mentioned earlier, don’t have that 5% per year disbursement requirement and which have this legally enforced mission. But how that interacts with all those philosophical questions about the extent to which we should allow the values to sort of change on reflection over time is kind of an interesting intersection. I’m not sure if any work has been done on that, but I’d be interested to see it.

Robert Wiblin: An advantage that this kind of foundation might have is that it seems like you’ve got a better shot at having the right level of drift if it’s aiming for a very general ideological goal where you expect that there will be people in the future who will share those ideas or at least share the natural evolution of those ideas inasmuch as there’s further thinking and further argument and further research and then they gradually change their minds. It could be that just by having people choose their successes, and as long as you’re part of a live, intellectual community that will exist in the future, then you’ve got a reasonable shot at neither totally drifting far away, but neither getting completely stuck in the ideas of 2019.

Philip Trammell: Yeah. That seems reasonable. With respect to just the details of the implementation, putting aside the question of the evolution of values or anything. One basic thought that I’ve heard mentioned… I think Will MacAskill mentioned this month, would be to have a committee where you swap out one member every however many years instead of having a single successor. And I’m sure a lot of thought could be put into how best to design the mechanism.

Howie Lempel: Just another factor that I could imagine being a big deal is if the risk that the funds were expropriated or sort of not allowed to continue, were pretty highly correlated with the probability that the fund was successful and just to make the stakes clear like we’ve talked about in the past, patient people sort of growing to own and control, like a decent percentage of the economy. And so that seems like the type of thing where you’d not want to rely on past examples. But I think that society’s going to have some strong reactions to that.

Philip Trammell: Yeah, that is a worry. So even now, I know Vox published an article a few months ago now; I think the title was “Silicon Valley’s Charity Stockpile Isn’t Getting Any Smaller” or something.

Robert Wiblin: It’s terrible!

Philip Trammell: I know, right? It’s about the emergence of donor-advised funds and these are these funds in which people can invest their money while thinking about where and when to give. It was all in very sinister tones and sort of suggesting that they should be taxed or more heavily regulated. Similarly, there was a proposal in the UK to start imposing the 5% per year disbursement requirement on charitable trusts in the UK at some point over the past 10 years, and it ended up not passing. But you know, it could happen and I think it might be more likely to happen if these things got much bigger to the point of owning substantial fractions of everything. So yeah, that does seem like a risk. On the other hand, again, maybe this is just a reason to think that the interest rate you could earn would be a bit less than it would otherwise be. If you think of it as a tax, that you have to spend a certain amount every year clearly communicating what you’re doing that this money really will all go to something that’s really valuable in the future and–

Robert Wiblin: But I’m not sure that that helps because it’s inasmuch as these piles of money get really big, then there’s this huge temptation for the present generation to raid them and spend it on themselves basically, which, to some degree I guess, is what’s happening right now potentially with the way that we think about these foundations or these trusts that are building up large amounts of money.

Philip Trammell: Right.

Robert Wiblin: People are uncomfortable about it because they see that they could do things that they want right now.

Philip Trammell: Yeah. Okay. So it might not just be about communication, but just an outright difference in preferences about what happens to the money.

Robert Wiblin: I suppose you can always engage in political advocacy with a large enough amount of money and wield influence in the back rooms.

Philip Trammell: Yeah, that seems like that would increase that delta term, right? That would increase that expropriation rate and would push in favor of giving. Yeah, I would say if it were just about a single large fund, then to some extent, this concern could be overcome by distributing it a bit. So you mentioned you could just fission the fund, right? And I mean if it were just a matter of turning a $2 billion fund into two $1 billion funds, then I don’t know how much that would help. But one way I’ve been envisioning this all along would be to just have it be a normal philanthropic strategy for people to have their own funds or their own trusts and a relatively informed community of philanthropists that’s aware of the arguments for a given disbursement schedule and then coordinate around that. So if a person thinks that we’re spending too fast, then they just save everything to move everything in the right direction. And if they think we’re spending too slowly, then they don’t save and they just spend everything and that would still result in the patient collectively owning more over time. But I think it has a better chance of like lasting than one fund with some rich person’s name on it coming to own big swathes of everything.

Robert Wiblin: Yeah. It’s interesting that the Catholic Church has enormous amounts of assets, but has managed to protect itself, I guess because many people share the ideology, but also, I guess through savvy political lobbying, they exercise influence and ensure that they don’t get raided.

Philip Trammell: Yeah, that’d be one example of a seeming success. Yeah. So another historical example that, if you’re interested you might look into, are the Islamic waqfs, or however they’re pronounced. So this is an institution allowed for in Islamic law, which is like a charitable trust in the English speaking world today. You can set up the fund with its mission and it has to stick to that mission indefinitely, long after the founder has died. And, over time, I think because of certain tax privileges, but also just because of investing disproportionately much of their assets, they came to control ever larger portions of the economies in which they resided. They were some striking numbers. I forget exactly what they were, but they were like into the teens and twenties percents of Egypt and the Ottoman empire and Persia belonged to these waqfs and they seem to have lasted and executed on their missions well enough. And you know, some were broad, providing for education in the Ottoman empire and some were narrow in like some soup kitchen in Palestine. And so you have the dead hand examples, but you also have the successes and they’re much less of a thing today and ultimately they succumbed just to expropriation or to world war one. But you know, a lot of small fortunes did as well, so I’m not sure. I’ve heard these used as examples of why we might expect there to be a positive relationship between the expropriation rate and the size of the fund, but I’m not sure what the data is on that. But anyway, they’re a fun example to look into.

Robert Wiblin: Yeah, that’s really interesting. I’ll look that up and maybe stick up a link to a paper or at least a Wikipedia article about that.

Philip Trammell: Yeah.

What about groups who have a particular sense of urgency? [01:40:46]

Robert Wiblin: All right. Let’s talk about some other more freestanding objections or reasons that people are trying to have a direct impact right away. I guess one group of people who have had a particular sense of urgency about doing things right now are people who are concerned about risks from advances in artificial intelligence. Some of whom think that there’s a pretty high chance that we could see real breakthroughs in AI that could cause it to have massive social effects over the next couple of decades, possibly even in this decade. I guess for them, the current moment seems particularly hingey, I suppose? And maybe also they’re concerned that if we don’t do things, then there won’t be a tomorrow to save. So it’s like they have to spend the money now. What do you have to say to that group?

Philip Trammell: Yeah. So if that probability is high enough, then we should spend a lot now. I think you might’ve even understated the case for caring about the effects of AI because it’s important to spend now if it imposes a big risk, but also if it could transform society in a way that renders future philanthropy kind of useless. If it just makes everyone so rich and the whole world a utopia, then there’s no sense saving for the problems of tomorrow. Those will also be gone. On the other hand though, something I think doesn’t quite pull all the way the other direction, but it’s also worth considering is that advances in AI could increase the value of saving. So, broadly speaking, it could increase the marginal productivity of the sorts of capital you’re saving up. The intelligence to turn the resources saved into even more good things in what would have been possible otherwise.

Robert Wiblin: So it increases ‘r’, potentially.

Philip Trammell: Well, yeah, it could increase the interest rate. Sure. Yeah. That’s one way to think about it.

Robert Wiblin: But you’re also thinking that AI might give us more opportunities later on to use the money really well.

Philip Trammell: Yeah, that’s right. So rather than just increasing the rate of return on resources, it would increase the extent to which those resources could actually be put to good use or something. It could also provide technological means for lowering that delta term, lowering that philanthropic discount rate at which resources could get stolen or values could drift. You could imagine very long lasting institutions which somehow managed to remain faithful to their original values for a very long time. Or property rights just being made much more secure by the kind of monitoring and governance that some advanced artificial intelligence could maintain.

Philip Trammell: And so then institutions set up now might actually really last for a very long time if AI has all the promise that people most excited about it think it could have. So yeah, you should just look at all the possibilities there. I mean this is all quite speculative of course. So I don’t have too much to say about it, but in principle it could go any way but on balance, I do agree that if you think that transformative AI is coming soon, that there is a very important and fleeting opportunity for spending that we shouldn’t pass.

Robert Wiblin: Yeah, I thought you might say that for people in that situation. It still could make sense to save up their money for five or 10 or 15 years to spend a moment somewhat in the future where it might be a little bit hingey or there might be somewhat better projects to fund or something like that.

Philip Trammell: Yeah, I think that’s also true. I was mostly directing that response to people who thought that there was transformative AI coming in the next decade or something. But if it’s longer than that, then it becomes ambiguous again about whether to go ahead and spend now or to wait for the possibility of an even more pivotal opportunity that arises over the coming years.

Robert Wiblin: Yeah. So I guess inasmuch as the question of how risky or how potentially massively beneficial is artificial intelligence in the next few decades, inasmuch as that’s the crux, we’ll probably just have to debate the substance of that question from a more technical perspective, which I guess isn’t the expertise of any of the three of us, so possibly we should just move on from this topic and bracket that and say, “Well, if that were true, then the argument potentially goes through”.

Philip Trammell: Yeah, I think that’s right.

Robert Wiblin: One thing that might be worth making explicit is even if you thought that every year there was a 1% probability of total human extinction, it seems like that does not have as much effect on your timing as what I would certainly intuitively think. Does that just increase this delta? This risk of total expropriation by one percentage point a year? Yeah. So that seems funny to me on some intuitive level. It feels like if we’re running such a high risk of extinction every year, it should just be super decisive and this would be such an important moment and we’d have to rush to do things.

Philip Trammell: I think that might be because if there’s a high risk of extinction per year, you probably also think that there are really impactful long lasting things that one could do each year by getting rid of the risks.

Robert Wiblin: Ah, so it’s increasing the hingeyness parameter potentially as well as–

Philip Trammell: And if it didn’t affect that and only was in the delta, then it wouldn’t make a big difference.

Robert Wiblin: Ah, so this is like, “What if there was a risk of some rays coming from space that would just wipe us out at one percentage point each year, with things that we could have no effect over”. We couldn’t build a defense against them. That does make a lot more intuitive sense. Okay. Yeah, I’m glad I clarified that. Yes. So the next argument which I think could potentially actually be really important and it’s maybe a bit of a shame that we’ve left it to this point to raise is the issue of kind of serial dependency of work that you’re doing to solve problems. So it often seems like you can’t get a hundred times as much work done in a given year just by throwing a hundred times as much money at it. Sometimes you have to spread your efforts over an amount of time because you have to finish earlier steps before you can move on to later steps. I guess people give this example of building a building. You can’t build the final floor at the same time as you’re building the first floor because they have to go in order. Could this potentially give us a pretty compelling reason to start work on object level issues quite a lot sooner than otherwise?

Philip Trammell: A lot of what is, in fact, serial dependency ends up just looking like diminishing returns, I think. So, as you said, you don’t do a hundred times as much work by hiring a hundred times as many workers and that’s a reason to spread your efforts. But, at least if you’re just talking in the abstract about the schedule on which you spend rather than who you hire when.

Robert Wiblin: So we should compare it to hiring a hundred as many workers for a hundred years or something like that. So having far fewer but spread out over time.

Philip Trammell: Right. I’ve already been accounting for the fact that there are diminishing returns and as you get very, very big then, as I mentioned, this is what gets us out of the Ponzi scheme conclusion, right? Once you get big enough, then you’re going to want to start spending, because if you got any bigger, returns would be diminishing by more than the interest rate, basically. In principle, this is an important complication, and for those interested in looking at the sketch of a model that I have so far, you’ll notice that it doesn’t allow for serial dependency. But I don’t think it too substantively affects most of the informal discussion that we’ve been having so far. Unless you’re talking about really precise sequences of actions that would take hundreds of years to execute. It’s like, “Yeah, there’s going to be some 10 year project which we could fund now, but there’s going to be a lot of them next year that could be started then. There’s going to be a lot in a hundred years time and you can find more of them if you wait.

Robert Wiblin: Okay, hold on. So you’re saying, “Well that’s true today, but it’s also true tomorrow”.

Philip Trammell: Yeah.

Robert Wiblin: Although, I suppose even those other things that you could start with more money in 50 years’ time, maybe you could have gotten even more juice by starting them now rather than 50 years’ time. Although I guess you’re saying we won’t know what they are at this point. So there’s things that will crop up that will be things that seem like they should be started then that we couldn’t start now.

Philip Trammell: That’s what I was thinking, yeah.

Robert Wiblin: Interesting. Though I suppose inasmuch as if you view all human endeavors as a series of things that often requires… you see this potentially with research where you need to make particular discoveries before it becomes apparent that there’s this new field that you need to start. Yeah, if it’s a single tree that’s growing upwards, then anything you do to push things forward now could potentially also push forward everything later on.

Philip Trammell: I see, yeah. I mean this is an interesting point. I’ll just say that the one point that you should make sure that doing the earlier stage of the project earlier really does speed up the accomplishment of your goals, broadly speaking, by faster than the interest rate in a sustainable way. So not just that you’ll get more than 7% of the way toward solving a problem by funding the researcher now then you could have done by waiting, but that the fruits of the research can be cashed in next year, and themselves produce at least 7% returns as well.

Robert Wiblin: Yeah. I guess with serial dependency, it means that if you think that the hingiest moment is going to be in 30 or 50 years time, then potentially then you do have to start doing work sometime ahead because you’re not going to be able to cram all of the work into that single year and then maybe that brings us back a bit more towards common sense. If you really think, “Well, if there’s something on this century that’s really crucial, then maybe we just have to spend some similar amount every single year in that century because we don’t know when it’s going to be, and even if we do, then we have to prepare ahead for it”.

Philip Trammell: Yeah, that seems right. But I think you can capture a lot of this intuition just by thinking about what the model says when there are diminishing returns to spending in each year. So let’s say impact is logarithmic in spending each year. So you want to spend at least something, because if you don’t spend anything you have negative infinite inflection; it’s really bad. But how well each stage of the 50 step process goes might depend on how fast you spend or something and then you want to spend something this year.

Robert Wiblin: Yeah, just a small fraction.

Philip Trammell: I was already saying that you probably want to spend at least a little bit this year. The question is just what fraction you should be spending. Yeah.

Robert Wiblin: So another argument that used to be really popular was pointing out that if you did things to reduce poverty in the developing world now, you know, building a factory, improving health, improving education, then that would mean that you know those people’s children or other people in the community would be richer. And there’d be all of these ripple on, flow through effects from the reduction of poverty and that these benefits might compound at a pretty high rate for quite a long time which gives you a reason to do it sooner because potentially they could be compounding at something approaching on some version of this argument. Yeah. What do you make of that idea of economic benefits compounding?

Philip Trammell: Well it’s not just approaching ‘r’. It would have to be that they compound faster than ‘r’. And that would require some sort of market failure. That’s preventing the beneficiaries from borrowing at ‘r’ to fund the project in question. If education really is, or health really is going to pay off faster than ‘r’, then it’s like, why don’t people just take out a loan to cure their disease and then earn those higher wages, whatever it is? So I think a priori, we should be somewhat skeptical that there are these opportunities for ways to help people that have compounding effects that are really sustainably higher than the interest rate. But we shouldn’t rule that out. There are probably some. And the real issue is that they would have to be able to compound at faster than the interest rate for a very long time.

Philip Trammell: Right, for as long as you might expect the fund to last. And that seems quite unlikely because, well that’s faster than the economic growth rate which would mean that like one school built hundreds of years ago, the flow through effects of that school would eventually come to dominate the whole economy of the region which was built or whatever. It will push up the growth rate to something higher than the interest rate. And it just seems a bit impossible that that sort of thing can really compound. It also seems impossible when you inspect what happens to the returns of the interventions in question, right? So you improve someone’s health, right? Their income is 20% higher or 50% higher in the next year than it otherwise would’ve been. But then what did they do with that money?

Philip Trammell: They don’t use it to cure that other disease they had that raises their income by another 50% and so on, year in, year out. They probably just do roughly what everyone else is doing with it, which is contributing to an economy that’s growing at the growth rate. They’re consuming most of it. They’re investing a little bit of it. That’s just not going to grow as fast as holding onto the money yourself and investing all of it. Eventually the benefits of that will outstrip this short term boost followed by a long-term slower growth rate.

Robert Wiblin: What about the question of some people who have suggested that the best thing to do with money for improving the world, I think somewhat facetiously, is just to save it because then that results in more capital accumulation, more business investment, more borrowing, I don’t know, and that that indirectly has a big effect on the economy or makes people a whole lot wealthier in the long term. Do you think there’s any kind of direct effects of the saving that are valuable?

Philip Trammell: Yeah, so there are some direct effects. It would increase wages a bit, because you have more capital around and since capital and labor are complements, there would be increasing marginal returns to labor, increasing wages. I am skeptical that increasing the economic growth rate just generically is the best use of philanthropic efforts. I know that some people have made the case that it is. So Tyler Cowen perhaps, most famously these days with his book, “Stubborn Attachments”, makes that case and others have as well. And there, yeah, maybe just perpetually investing it. Not necessarily in just a standard, not just putting it all in the stock market or whatever, but maybe there’s particular sorts of investments that have–

Robert Wiblin: Like a venture capital thing?

Philip Trammell: Yeah.

Robert Wiblin: Or like investments in emerging economies?

Philip Trammell: Yeah, maybe that. Maybe in a startup that you might expect to have a lot of positive externalities, like some biotech thing that couldn’t capture all the fruits of the research that it did.

Robert Wiblin: Something highly innovative.

Philip Trammell: Yeah, I think that is an effect. But I would expect that making use of really high leverage opportunities for impact as they come along and just making sure that you’re well funded enough to take advantage of them is just going to be most of the effect and whatever effect you have on the growth rate is going to be pretty small.

Robert Wiblin: Okay. So another argument that one could make, I think, that would attenuate the case for delaying impact is that, sure, you might be earning ‘r’ each year and so you’re accumulating more and more money, but the costs of the things that you’re going to want to buy in order to change the world later on are also going up at a pretty decent clip. So very often you might want to turn this foundation or use the resources in the foundation, to buy the time of people, of scientists, of policymakers, or whatever else, in order to actually have an impact. But their salaries are going up at 1 to 2 percent a year, though I guess for skilled workers, lately it’s been quite a bit more than that, so it could be that the cost of the inputs that you want to buy later on is going up and this is going to greatly weaken the apparent increase in your purchasing power that you thought you were going to have.

Philip Trammell: Yeah. I mean this just gets back to a lot of what we discussed earlier, I suppose. Their wages are going up at roughly the growth rate. So, well, ‘r’ is greater than ‘g’, so you’re going to be able to hire more of those workers in the future than you could hire now. I’m not sure why you would care about number of workers or number of worker hours per se.

Robert Wiblin: I thought you might say, “Well they’re getting more productive”. That’s one reason why the salaries are higher. I suppose in some industries productivity goes up a lot. So if you were using it to buy the time of manufacturing workers, then maybe this doesn’t bother you so much. But if you were doing it to hire masseuses whose productivity really hasn’t risen at all in a hundred years, then this doesn’t look so good.

Philip Trammell: Yeah. Okay. Well first off, if you really were just trying to hire as many people as possible, or as many worker hours as possible toward some end, what you care about is the rate at which wages are increasing. And that’s roughly at the economic growth rate. Wages as a whole has been a bit less than that actually because the share of output going to wages rather than capital has been declining.

Philip Trammell: But yeah, maybe for skilled workers it hasn’t been decreasing. Maybe it’s been increasing. But anyway, still quite a bit less than the interest rate. So you can just hire more of those workers in the future if you invest. And also, just to return to the original observation about impatience being baked into the interest rate. I would imagine that, in general, you wouldn’t care just about the number of workers that you hire, but about how much they produce and how much they contribute to–

Robert Wiblin: The actual project.

Philip Trammell: Yeah, right. Exactly. And if that actual project looks anything like the project of increasing human welfare, well in dollar terms, maybe they’re getting 2% more productive each year or whatever the economic growth rate is. But it’s getting a bit more than 2% more expensive to help people each year if that eta term mentioned way back when is greater than one. So, as time goes on, to do a given unit of good, you might expect to have to hire more than one worker or something. But eta times g still seems to be less than the interest rate because again, people are impatient and so when they’re deciding whether to hire one person this year or 1.07 next year, they’re indifferent and that’s because it’s actually only gotten 1.05 times as expensive to do a unit of good, for themselves. So it’s gotten 1.02 times more expensive to hire them, but then–

Robert Wiblin: Probably because they’re richer as well.

Philip Trammell: Yeah.

Robert Wiblin: Interesting.

Howie Lempel: I didn’t catch the response to the claim that it’s like service providers in particular that are disproportionately having their wages go up. And so, if it’s also the case that you disproportionately need to hire service workers to like do good, then it does seem like wages in one sector of the economy could go up faster than g. That could be a problem.

Philip Trammell: Yeah. I just think that’s true. Yeah, then you do have this fleeting opportunity to put this resource, a service worker, to use on this important project now, which in the future will, in some sense, be used up. It’ll be all bought up by all the future rich people for whom those workers are a really strong, complementary factor to the production of the things that they care about. Yeah. I mean sure. In that circumstance, then you should spend now while the work is cheap.

Robert Wiblin: Okay. So if you’re confident that the thing you’re going to want to buy in future, or that the best thing that you will be able to buy out of all the projects possible in the future, is this thing whose cost is increasing faster than g or faster than r even, then that is a pretty strong reason to do it now. I guess we might think that’s not terribly plausible at the cost of a given amount of scientific research or a given amount of any service work going up at that kind of rate per year. I guess this is at least not what we’ve seen so far.

Howie Lempel: Also, it would have to be going up and also literally the most important resource, right? So much that it doesn’t eventually just fall and become not the most cost-effective thing to do.

Robert Wiblin: Yeah, it does seem unlikely.

Howie Lempel: It typically has to be really necessary that you use that particular resource.

Robert Wiblin: All right, so we’ve spent quite a bit of time going through a whole bunch of arguments in favor of trying to have a direct impact right away and others in favor of waiting to have an impact. I guess, just to go over those again, in favor of delaying, you just have way more money basically and you’re a larger share a full global wealth and GDP, where like learning and finding better interventions over time and uncovering crucial considerations that might affect a lot how we want to spend it. Also just there could be even more crucial moments in which you want to act in future than the present day. And then you’ve also got this parameter or uncertainty point that we made that if you’re very uncertain about a lot of this stuff, then it seems that that can, for mathematical reasons, push you in favor of wanting to potentially save a little bit more. And then there’s also this point that this is potentially a fleeting opportunity during which interest rates are as high as they are and we could potentially build up such a large war chest. On the other side of the ledger, we’ve got movement building which potentially builds above market returns, which I guess you’re then categorizing as a form of investment and saving.

Philip Trammell: Certain kinds of movement building.

Robert Wiblin: Yeah. And some sorts of research as well. Then we’ve also got, obviously there’s a risk of the fund losing all of its money, potentially all of its investments going badly or the world being destroyed. So just like human extinction. Or the fund drifting a lot from its original goals in a way that you wouldn’t like and wouldn’t endorse. We’ve got the possibility that the current moment is super hingey and that we’re mistaken when we think that it’ll be more hingey later on. We’ve got this serial dependency issue that even if you knew a really important hingey moment was coming sometime down the line, you have to start working ahead of time because you can’t cram it all into a single moment. And then we’ve possibly got that economic benefits compound, so there’s some reasons to act earlier because of the flow through effects you get to harvest them for longer. And then we’ve also got this issue that even though r is quite high, to some extent, in some cases the cost of the inputs that you’re going to want to buy later on with all of the money that you’ve built up could have risen, and so it’s like your purchasing power hasn’t grown quite as much as you might have otherwise thought.

Philip Trammell: Right. Or it could have even fallen. That is a good point.

Robert Wiblin: Okay. So going through all that, which arguments in your mind are the most important or most decisive, do you think?

Philip Trammell: Yeah, I think just the straightforward interest rate point that you’ll have more to spend when you invest is pretty big. And relatedly, the impatience point that one should expect, systematically, interest rates to be high, to make it favorable to invest for someone who is patient. Learning also strikes me as an important reason in favor of waiting. I certainly feel like I’ve learned a lot about how to spend charity money and I expect to learn more in the future, and I expect us all to. And finally, the hingeyness thing. The fact that opportunities for really outsized impact come and go, and that one should spend much more when one finds oneself in a situation where they can take advantage of one of those opportunities would be the last. And structurally, it’s sort of ambiguous which direction that last one would point in, because it depends on whether we’re in one of those moments now. But my guess would be that we’re not at such an extreme version of one of those right now, that all things considered, pushes us in favor of wanting to spend a lot now.

How much should we actually save? [02:01:35]

Robert Wiblin: So it seems like we definitely shouldn’t be spending all of the resources that we have access to. Indeed, borrowing lots of money and then trying to cram it all into spending this year. That’d be crazy. Something like spending a hundred percent or more than a hundred percent. On the other side, it would seem pretty crazy if we did nothing to try to improve the world directly this year and instead just saved 100% of the resources. So there’s going to be a question of balance here. And inasmuch as the arguments in favor of delaying become stronger, then you save a larger fraction and inasmuch as the arguments are trying to do stuff right now become stronger, then you just save a smaller fraction into the next year. So it’s just a question of percentage and it’s just going to slide up and down depending on what kind of moment we’re in and how we view the balance here.

Philip Trammell: Yeah. So one sort of broad heuristic that I think is worth keeping in mind is the following. If you think impact is about logarithmic in the rate at which you’re spending, and if you think that you’re at a typical time, not a super important or super unimportant time from a philanthropic perspective, it turns out that the rate at which you should disperse, or patient philanthropists as a whole should disperse their assets each year, equals that philanthropic discount rate that is the rate of extinction plus the rate of value drift and fund expropriation and so on, all added together. So yeah, if you think the latter is 1% a year, we should all be dispersing 1% of our assets each year and investing the rest.

Philip Trammell: The intuition for that, broadly speaking, is that the less you put aside for the future, the higher the marginal returns to putting aside more for the future. But the higher the discount rate, the lower the marginal returns in terms of your impact to putting aside resources for the future. So those are the two effects and they precisely cancel out in the case of a logarithmic impact function. Anyway, I think that’s a nice heuristic. One red flag I just want to make clear is that this works out if, by assets, we’re referring to all of the resources which one could possibly get access to spend this year, including the discounted value of future income. So, at the moment, I think most of the EA community’s assets, unless you really, really stretch the term, are rather literal assets. They’re held by Open Philanthropy and various rich philanthropists and relatively little of it is in the form of future income.

Robert Wiblin: So the question of what fraction of resources is the effective altruism community, writ large, spending versus saving each year seems like it should be straightforward. But, in fact, it’s anything but straightforward. Potentially, if you’re just talking about a charitable foundation then you can look at the dispersal rate, but then what about if people are our greatest asset, Phil? I suppose if someone is going to live another 50 years, then would you say the next year is 2% of the resources that they have? I suppose then that means that as you get older you’re spending a larger fraction of the assets that you have. But then we should think about the people who come later on. Find the donations that might become aligned with those values that join later on I guess you can’t borrow against. But nonetheless, in a sense they’re part of the net value of all the net assets that you have. So you can see that this is a lot more complicated or it’d be trickier to think about than it might initially seem.

Philip Trammell: Yeah, it is a little trickier than it initially seems. But right now, EA wealth is concentrated enough that maybe it’s not all that tricky. So a relatively literal estimate of EA assets might be around $10 billion that Open Philanthropy has, just to think about it in round numbers, plus a few billion held by very rich, EA aligned philanthropists, plus the present discounted value of all of the Giving What We Can pledges and so on that are currently, in some sense, pledged to be committed to some philanthropic cause that we think is very important. You want to include the present discounted value because this is, in a sense, the budget available right now. This is the amount that we would all collectively have if we all borrowed against the money that was going to be put toward optimal philanthropy.

Philip Trammell: Okay. So if that’s the budget. If that’s the thing that we should spend 1% of say every year in this scenario where there’s a 1% discount rate and impact is logarithmic in the spending rate. That said, one might also want to include guesses about future EA movement growth or just future trends toward philanthropy and toward more thoughtful, reflective philanthropy that seem to be underway. There’s Bill Gates’ Giving Pledge which has encouraged a lot of the world’s wealthy to give more and to give more thoughtfully than they might otherwise have. And, even more abstractly, you might think about general trends in social values in one direction or another as part of this pool of assets being put to the very, very global public good of optimal philanthropy through the lens of some value system.

Philip Trammell: Now, I think the important point there is that most of those sort of abstract assets listed at the end there are very risky. We don’t have a very strong reason to think that everyone’s going to be on board with anything you or I may care about in the future. So most of the value of putting money aside explicitly for some purpose along those lines is going to be realized in the scenarios where you don’t have a thousand versions of Bill Gates and Dustin Moskovitz and so on in a hundred years’ time. And yeah, I prefer to just stick with more concrete estimates of EA assets and look at the dispersal rate of those.

Robert Wiblin: Because you want to focus particularly on the scenarios where there aren’t massive future assets that are coming on board.

Philip Trammell: Right. I guess for one thing, I just don’t think those scenarios are all that likely, but also, if you think that there’s a two thirds chance that that is the future we’re headed towards. And if you think that that influx of assets would just push the marginal value of a long-term fund like this to zero: it wouldn’t quite of course, but that’d be the extreme. So that just cuts the value of the fund down to a third of what it otherwise would’ve been. But we’re dealing with order of magnitude considerations at this point. And I think it’s unlikely that on its own, this would be the thing to swing the balance.

Robert Wiblin: Okay. And what do we do about the people, say, all of the people who read or listen to 80,000 Hours who might shift their career on the basis of thinking within this community. How do we think about the rate of dispersal of the assets that they have, which I guess is mostly human capital?

Philip Trammell: Yeah. So that’s also a bit abstract. I mean sometimes it’ll be concrete if someone’s deciding whether to spend money like investing in their education or whatever versus going ahead and doing direct work as they are. But by and large, I think I would just say that the qualitative considerations that we’ve been discussing apply to the timing of the dispersal of any asset including career capital. So if you think that now is a really important time and that transformative AI is coming in the next five years or whatever, then that applies to one’s career decisions as well. And if one thinks that perhaps one should care about the possibility that in 50 years you have this really pivotal moment, then likewise, one should maybe just get a whole ton of experience and PhDs and stuff that are directly aimed at being useful when the moment strikes.

Philip Trammell: One thing to mention is that we’ve been talking about the importance of a time. We’ve been talking about hingeyness as if it were this single quantity that you could just tag a moment with… 2019 is this important? But if what we really mean by importance is the ease with which resources can do good, it’ll depend on the resource. And you could imagine a scenario in which money can do a lot of good, but hours can’t or vice versa. Now over time, you would expect the two to move in tandem because to some extent one can be turned into the other. If there’s a glut of hours, people can earn to give. If there’s a lot of money, people can use it to fund 80,000 Hours to encourage people to switch careers or whatever. But there will be a lag.

Philip Trammell: And so it seems like there’s a bit of a lag right now in the EA community where there’s a lot of money with Open Phil but not a lot of people working on the sorts of things that might seem important from an EA perspective. And so, a lot of efforts being put into essentially converting money into people, but you don’t expect that gap to last for like 50 years. I would expect that, broadly speaking, you have this hingeyness thing which varies over time and then if you zoom in asset by asset, they could come apart depending on how easy it is to turn one asset into another.

Robert Wiblin: But then they tend to get pulled back together.

Philip Trammell: Right, exactly. So one shouldn’t be misled into thinking that because there’s all this demand for your particular skillset right now, that you should work rather than investing for the future because that’s just a passing circumstance.

Robert Wiblin: So in your mind, how much of all of the resources that you think are relevant for this discussion are we saving versus spending each year? And what do you think when we try to throw all of these considerations into the calculator that you’ve built, what would be the right level of saving versus spending? And I guess, does that imply that on, the margin, we should be saving more or less or we’re doing things about right?

Philip Trammell: I did a very ‘back of the envelope’ calculation on this a while back and estimated that, I think it was something like 5% of the relevant “EA assets” are being disbursed every year. And, so again, if you think impact is logarithmic in the spending rate, that would be right if you thought that there was like a 5% discount rate, which strikes me as quite high. But then again, I do to some extent feel the force of the arguments that now is an important time. So it might not be all that high, but I think it’s a little on the high side. But I also think that even if it is not too far wrong, it’s that way, kind of by accident. Sort of for the wrong reasons. So my understanding is that Open Phil is looking to spend more quickly and investing only because they haven’t yet built up the infrastructure to actually do that. They’re still hiring a lot of research analysts and so on. And that over the coming decades, we will collectively be spending more than that percentage per year.

Robert Wiblin: Unless I guess other donors come on board and then they’re inclined to say something like that. Yeah. It could shift in either direction, in theory. Okay. So you’d say we’re probably a little bit… I suppose it’s very uncertain because there’s so much that we don’t know and this has been obvious but you might say we’re maybe a little bit on the high side. So potentially if you know a new donor with $1 million who got on board and started listening to the 80,000 Hours podcast and was trying to decide what to do. You might think you’d be pretty happy to find them saving or putting it into a long-term foundation.

Philip Trammell: Well, yes, or funding a very high fidelity kind of movement building. I don’t want it to get glossed over because I do think that’s quite important.

Robert Wiblin: Some sort of investment, whether it’s financial investment or some higher value investment that involves advocacy and yeah.

Philip Trammell: But again, the kind of advocacy that actually does communicate the idea that perhaps what you should do is just put all your money into either more advocacy or just a fund.

Robert Wiblin: The Phil Trammell perpetual fund.

Philip Trammell: Right.

Robert Wiblin: Or perpetual movement. Do you know of any groups that are doing that kind of movement building at the moment that you’re more excited about because they have this kind of reinvestment aspect?

Philip Trammell: One thing that comes to mind is that Natalie Cargill at Effective Giving has been reaching out to, and making herself available to be reached out to by large philanthropists and talking with them in depth about all kinds of EA relevant questions including timing. I think that that’s promising. Both because when you manage to talk to someone with a lot of assets, that’s obviously a bigger opportunity, but also because spending more time per person means that the fidelity is going to be higher. Just sort of mass outreach seems much less promising in that respect.

Robert Wiblin: Yeah, you’re much less likely… So if you do a general media thing and then lots of people find out a little bit about you, the odds of them reinvesting in that kind of movement is just not going to happen. So it doesn’t have this quality that you’re looking for.

Philip Trammell: Yeah, exactly. If it’s just about something that’s relatively well known and easy to understand and implement, like becoming vegetarian or something and you’re just going to have your thirty second spot to most persuasively get that message out.

Howie Lempel: There also isn’t a currently, that I know of, of a really easy way for somebody to make one quick decision and decide that they’re investing in this way.

Philip Trammell: Yeah, exactly. That is a good point. Like it’s hard to tie your hands to long-term investment. You can put your money in a fund where it has to be given to some charitable purpose eventually. But yeah, that’s a good point.

Robert Wiblin: So I mentioned the idea that we should be investing more rather than trying to have a direct impact. Right now it’s kind of a tentative conclusion. I guess you don’t want people to go away with a memory that you definitely think that we should always be doing that. Do you have any sense of how likely it is that in five years’ time you’ll think that what you should’ve said now was actually, “Oh no, we should be saving less”?

Philip Trammell: Yeah, that’s a good question. Maybe a one in three chance or something like that. Yeah. Thanks for really asking that to make it clear that I’m still thinking through this and it’s quite tentative. Yeah.

Robert Wiblin: Yeah. I guess people always want bottom lines which is very understandable. That’s what I always want. But, at the same time, I guess you don’t want to lock-in bottom lines when you’re very early on a research project.

Philip Trammell: Right, right. I will say that it’s easier to spend once you’ve saved than to take back the money that you’ve spent.

Robert Wiblin: That’s a very good point that we actually haven’t raised, is that one of these things is reversible in a way that the other thing isn’t.

Philip Trammell: Yeah, I mean I think that’s true for most opportunities for direct impact that would be under consideration. Like if you have an opportunity to avert an existential risk and you let that go by and the risk happens, then that’s not reversible. And so it could be that looking back, I will be dead, but I could realize that there was a once in forever opportunity that was passed up and that won’t be reversible. But I think, yeah, in general, the reversibility consideration definitely pushes in favor of investment.

Robert Wiblin: Yeah. Another interesting thing with this is that inasmuch as you’re part of a community or a group of people that has a shared goal and is spending resources in order to have similar effects, then you have this issue that if you save more, someone else could choose to spend more to offset what you’re doing. And it’s a bit unclear if you do something, whether that is going to have the net effect that you were desiring, given that other people can try to undo what you’re doing with their resources. Do you have anything to say about that?

Philip Trammell: Yeah. I mean that’s not just about being in a community. If anything, that mitigates the problem. The problem is that what you’re providing is a public good that other people care about and the community is a way of resolving the problem that that raises. Namely, making it easier to coordinate around the provision. I think ultimately it just highlights the fact that donor coordination is an important problem that we should all, and I mean every philanthropist, should think harder about and develop more thorough and robust tools for dealing with. At the moment, I think it’s quite likely that most things that listeners to this podcast will want to fund, will end up just funging with Open Philanthropy, including saving.

Robert Wiblin: Yeah. Funging is this term for offsetting the behavior of other people in a particular way.

Philip Trammell: Right. So for every dollar you save, they spend a dollar more than they otherwise would have and vice versa. On average, of course. Not literally. But what one can try to do to be a bit more sophisticated before the optimal donor coordination mechanism has been worked out, is sort of subsidize the patient behavior of other people who care broadly about the same kind of things you care about, but are sort of acting on a higher discount rate. So you might somehow have a commitment to match spending on whatever the cause is. Whether it’s GiveDirectly cash transfers, or research into risk reduction or whatever it might be in 20 years’ time or in even longer than that. And thereby make it more appealing even for someone less patient to spend more slowly than they otherwise would.

Robert Wiblin: Yeah. So are there any other bottom lines or comments you’d want to make to people who are in the audience who are listening who are regularly donating money? Potentially Giving What We Can members who are giving 10% of their income each year. So like, what they should think in light of all of this research on optimal timing and maybe how it should affect them?

Philip Trammell: Sure. So one point is that again, if one’s behavior is just funging with that of other philanthropists, then giving more means that more goes toward good causes in general. But you know, thinking about the timing and the cause in particular doesn’t matter too much. But, if what you’re doing is actually making some counterfactual impact, I think the common practice of just giving a certain percentage of one’s income each year as one earns it, is almost certainly wrong. It’d be a really weird edge case for that to be the best schedule on which to give. Either you should probably think that we’re spending too slowly, in which case you should give as much as possible now, that is, borrow against yourself, right? So give 20% this year so that you have less to save for yourself for next year. You put aside more for your retirement next year. Or, you should just try to save up if you think that collectively, all the funders of whatever you’re putting your money to are spending more quickly than they should. So that’s one takeaway. The second one is that if you want to try saving money for a while, but don’t want to run the risk or want to minimize the risk of just getting more selfish and spending it all on yourself, something you can do is set up a donor advised fund. So this is a fund that you put the money in. You get to control what charitable cause it goes to. But it has to go to something charitable.

Philip Trammell: You can’t withdraw it for yourself. And the money grows tax free while it’s in there. So yeah, it grows at a faster rate. That’s a second takeaway. And a third one, which relates to the career point we discussed earlier, is that you might think that money has just more value in general. Philanthropic money has more value than it might otherwise have seemed. I think people in the EA community talked a lot more a few years ago about earning to give and about the good that can be done with one’s dollars. And, more recently, with that arrival of Open Philanthropy and other rich value aligned donors, the discussion shifted very heavily in the direction of what one can do with one’s career instead. And if one, I think, remembers that this temporary imbalance won’t last forever, right? That in time, the money will get turned into hours and money will become very useful on the margin again, then the value of earning to give starts seeming higher after all. And maybe there’s quite a lot at stake when one makes that extra hundred thousand or a million or whatever and puts it in that fund to just slightly sort of bend the course of history far down the line.

Implications for career choices [02:19:49]

Robert Wiblin: So that actually leads really perfectly into the next section that I was going to talk about. At the beginning we said we’re not going to talk about careers and we’ve mostly done that. They’ve come up a handful of times. But it would be interesting to focus now on what implications does this overall mindset, or this patient mindset, have for people’s specific career choices. I guess if you’re saying because of the ability of money to potentially compound for a very long time, it potentially weighs in favor of earning to give relative to other things and I suppose you’re also saying while we may not have so much money relative to the people involved forever, and so that also points in favor of earning to give. And of course one of the big differences here is that people die in a way that money doesn’t and that’s potentially one of the big disanalogies here, is that you can’t save up you as a person indefinitely and then spend yourself in 300 years. To some extent, you are compelled to spend a certain percentage of your time each year of the time that you have left. And your career capital, all of the skills that you’ve learned, die when you die as a result of aging. Are there any other kind of considerations we should highlight for how the person case is different than the money case?

Philip Trammell: No, I think those cover it.

Robert Wiblin: Yeah. Okay. So points in favor of earning to give might also point in favor of collecting career capital, or trying to build up your credentials and your skills very aggressively early on potentially if there was a trade off; if it’s the case that the job that allows you to have the biggest impact when you’re 22 doesn’t teach you that many skills and there’s some other job that doesn’t have almost any impact at all, but builds up your skills a great deal and will allow you to go further in the long run that seems to point in favor of career capital, like a patient role. Is that right?

Philip Trammell: Yeah, that does seem right. As you said, it’s not going to push as hard as in the money case because you’re not going to be able to just skill up for 300 years or whatever. But it would skew it a bit more. It would mean maybe you should spend a few more years at the beginning of your life building skills because you’ve discounted away the value that you’ll be able to produce as a highly skilled person later in life less.

Robert Wiblin: Yeah. So I guess that to some extent, that’s relying on this view that probably the present year or the present decade isn’t especially hingey relative to times that might come later on. I guess if you ever do hit a moment where it seems like, “Yes, this is the hingiest year of my lifetime, or at least of my career”, then maybe you want to switch out of building career capital and just go for impact in that period.

Philip Trammell: Yeah, that’s right. I mean, this was true in the fund case as well, which is why I thought the question of whether this century is probably the most important ever was kind of a red herring. It just depends on the life expectancy of the thing under consideration. And is this going to be the best time to spend the asset for as long as the asset can be expected to last is roughly the idea; is this a really important time in my life or is there going to be an important one later on.

Robert Wiblin: Are there any other implications? I guess one that I wasn’t quite so sure about is maybe this perspective pushes in favor of being a generalist rather than a specialist because you might think, “Well the things that we think we want to specialize in right now, maybe we’re mistaken about that. We’re going to have new considerations in future and you want to keep your options open”. I suppose it’s a general thing in favor of flexibility, potentially, inasmuch as you think it’s not necessarily the case that we can see what the hingiest moment is of your career today. Is there anything in that direction that you buy into?

Philip Trammell: Yeah, I think that seems roughly right to me. I mean, you’re 80,000 Hours. You’ve probably thought about this in more detail. So I don’t know exactly what the best generalist careers are or anything like that. I like to tell myself that an econ PhD is sort of up there. But I would say it’s not as simple as pushing in the direction of generalist over specialist skills. This line of thinking could also recommend developing the specialist skills that you think would be or could be most valuable in a really important moment if it happened to arise. So like becoming a constitution writer, however you train for that. Maybe it’s kind of a cartoonish example. But yeah, hopefully that conveys the idea.

Howie Lempel: Seems like another thing pushing in the direction of being a generalist is just that if you’re earning to give, it doesn’t really inherently matter what you’re specializing in.

New Speaker:
So indirectly through that path it pushes in favor of generalism.

Howie Lempel: Right, seems like it.

Philip Trammell: Earning to give is kind of like being a generalist. It’s just you run into the problem that a time may come when a certain skill is in particularly high demand and there isn’t the time to convert money into people with this skill in question.

Robert Wiblin: When we’re zeroing in on this one consideration which, even if it’s kind of an important consideration, it’s only one among many, we’ve got to make sure we don’t lose sight of each individual’s personal fit and what jobs are actually available to them. I guess also coordinating with other people who are trying to solve the same problem. So potentially if you know other people who are building career capital, there might be a reason why they want to go and have direct impact right away because of these diminishing return reasons that we were talking about. And I guess also just your comparative advantage relative to what other people can do is going to be… There’s a whole lot of other stuff going on here, but we shouldn’t just think about discount rates in isolation.

Philip Trammell: Sure, sure. Of course.

Robert Wiblin: I guess another consideration for people’s careers that potentially spills out of this is that it seems like global priorities research or at least research into optimal timing specifically could be really important. Do you have any advice for people in the audience who are considering, I guess, possibly becoming academics, studying economics like you are, or potentially working on global priorities research questions specifically. Do you think more people should be doing this? How might they go about it? What kind of people are suitable?

Philip Trammell: Yeah, I definitely think more people should be doing it. I think there are a lot of undergrads involved in effective altruism or just recent graduates who majored in economics and have a very suitable mindset about all this. But there are surprisingly few who are in grad school for economics or like seriously considering going into grad school specifically to study these sorts of big picture longtermist questions that are relevant to patient philanthropy as I like to call it. I’m not sure exactly why that is. I guess one possibility is that it just doesn’t occur to people that the tools of economics are very well suited to answering or at least shedding light on these sorts of questions. I know that a few years back, some people in effective altruism, including 80,000 Hours, were recommending that people study economics to do development economics and I mean there’s a lot of valuable work to be done in development economics, but it’s just, you know, diminishing returns again, I think it’s just not quite as neglected and to my mind, like high stakes as some of the longtermist questions that I’m trying to at least start working on. So I would just say there’s this really high impact thing that people can do if they find that they have some skill at economics and they should really go into it because it’s very important and very neglected.

Howie Lempel: Do you have a strong take on what subfields of economics people ought to focus on?

Philip Trammell: Yeah, that’s another good question. So first I’ll say this. There’s some debate in economics about the relative value of theoretical and empirical work, right? And there has been a big shift over the past few decades from theory to empirics. As you know, we’ve gotten computers and just the ability to do better in empirical work and something like half of econ papers used to be theory and now it’s like 5%. Anyway, regardless of what background beliefs you are bringing to the table about the relative value of theoretical and empirical work, I think it’s pretty clear that the further out you’re looking into the future, the more quickly the value of empirical insights decay is than the value of theoretical insights. So it should kind of shift you toward theory. Within theory, there’s plenty of interesting questions in microeconomic theory, some of which we’ve even discussed I guess on the game theory of how people should interact when they have different discount rates and want to leverage each other to be more patient or less patient, I suppose, than they otherwise would be. The mechanism design for keeping institutions faithful to their original values. Yeah. There’s some questions along those lines.

Philip Trammell: Then in macro theory, there’s lots of questions about long run growth and about the interaction between growth and other sort of macro variables, like maybe just the rate of technological risk. There’s a really great paper that I had the privilege of helping to oversee this summer, which was written by this guy, Leopold, who looks at what a tweaked version of Jones’ growth model has to say about the relationship between economic growth and existential risk, and I think more work along those lines would be really valuable. Then there’s just a lot of stuff to be learned from the economics of discounting and optimal timing and the economics of catastrophe as it’s sometimes called. Environmental economists have thought a lot more about the optimal mitigation of catastrophes than most other economists or indeed most other people, and there’s a lot to be learned from them for existential risk more generally, I think.

Robert Wiblin: Yeah. So just before that, you were going to talk a little bit about how people can get involved in global priorities research.

Philip Trammell: Yeah, so just for a little bit of background, global priorities research is this term that’s been coined by some folks in the EA community over the past few years for foundational academic research, broadly speaking, into the question, or even just how to think through the question, of how to do the most good with limited resources. So far it has sort of seen itself as somewhere at the intersection of philosophy and economics and it’s just been dominated by philosophers because all the big scholars in EA, so far, pretty much have been philosophers. So if you want to get into it as a philosopher, it’s just much easier. You can come to Oxford, be advised by one of the philosophers at GPI or elsewhere in the community there.

Philip Trammell: As an economist, I would say there isn’t really that much infrastructure at this point. So the path would just be to be a good economics student. Get very familiar with the EA literature. I feel like that’s almost another major at this point. Or like at least a minor. Like having read the 500 blog posts or whatever and knowing the jargon.

Howie Lempel: Is there actually a list somewhere?

Philip Trammell: Well there is a list at the bottom of the GPI research agenda actually.

Howie Lempel: Oh yeah. I’ve seen that. That was really good.

Philip Trammell: And just scattered throughout it in the references.

Robert Wiblin: As you’re saying, it’s important potentially to get up to speed on lots of the concepts that people involved are familiar with.

Philip Trammell: Right. Indeed, part of the motivation for GPI and for making an academic research field out of this is to convert this informal literature into actual textbooks and academic papers that will be taught in classrooms and be just a mainstream established field of knowledge. But until that happens, it’s really valuable to be familiar with this sort of body of thought. And I guess just given the ways in which it can interact with philosophy, or even just given the contingent fact of all the philosophers that will be around if you try to work at GPI or talk with other people who are calling themselves global priorities researchers. Knowing at least some of the relevant philosophy, I think, turns out to be rather important. So that’s, in particular, parts of ethics, decision theory and epistemology would be the big three normative domains in philosophy. And I think that actually can be quite important for understanding why the questions that seem most important from an EA perspective are often neglected in mainstream economics.

Philip Trammell: So we’ve been talking a lot about discount rates, right? It’s very common in economics to confuse welfare with preferences. Whereas if people right now just care more about what happens in the near-term than what happens in the long-term, economists will casually say that it increases welfare more to discount or something. And that’s just the kind of Philosophy 101 rookie mistake that would not fly in a place full of philosophers, so I think those would probably be the big three prerequisites to getting into global priorities research as an economist.

Robert Wiblin: Yeah. Do you have any comments on the question of what sorts of people are really well suited to going into global priorities research?

Philip Trammell: Yeah. So at least on the economics side, it would just be someone who is good at economics. At least decent at philosophy and good at crystallizing nebulous problems. Not just the ability to execute on a tried and true methodology but to formalize problems that haven’t yet quite been well posed.

Howie Lempel: Do you have a strong take on…. let’s say I’m starting my senior year in university. I’m really talented at economics. I want to do global priorities research. Would you tell this person to go straight into a PhD? Do you think it’s important that they spend some time doing a bunch of global priorities research before they start grad school? Do you have a strong take there?

Philip Trammell: Yeah, that’s a good question. So I’ve spent the last year at GPI working somewhat independently and on some projects, and I definitely think that that year has let me clarify my own thoughts about what I want to get out of grad school and what questions I’ll be researching. That option is open to others now. There’s a GPI pre-doc program that we’ve just launched. The first pair have just started and there might be applications for that for next year opening in a few months’ time. There might not be as well. I’m afraid I’m not sure yet. But at least there will be the year after that. People can also spend the summer as a summer researcher at CEA/FHI which are other organizations in Oxford also doing EA research. So yeah, I think that those would be the most directly relevant opportunities for people interested in taking some time off and I think that can be useful. I also think it can be useful to explore a given subfield in more depth. So if you’re not sure whether you want to do theory work or empirical work or something and you can get an RA job doing one of those things, probably empirical as there are many more of them, then yeah, you should try to get an RA job for a year or something.

Philip Trammell: But you have to be careful with that because there’s a lot of grunt work jobs out there and you could end up wasting two years doing work that doesn’t actually give you much information about the field that you’re considering. So it’s a good opportunity if you can find it, but just be careful.

Robert Wiblin: Nice. All right. We’ve been going for ages and should probably wrap up. I guess, as a final question… As I mentioned at the start, you’ve taken this pledge to give away what’s probably going to end up being a pretty substantial fraction of your income. How has all of this research affected your own giving plans, if at all?

Philip Trammell: Yeah, so interestingly the research hasn’t affected them at all because this all started in high school when I inherited a bunch of money and decided I would set up a donor advised fund, one of those things I mentioned earlier, where you put the money in and invest it tax free. And I’ve continued to put all my excess money, all my charity money into the fund while I continue to think about when, and ultimately perhaps where, to give. So I’ve always had the intuition that there’s so much more to learn and that there’s so many opportunities that could open up, that it might be better to invest, and this research… I have learned a lot doing it. This isn’t just some sort of propaganda where I’m trying to justify my existing decision. I mean further research could definitely overturn the conclusion, but at least in my case, it hasn’t actually changed what I’ve been doing.

Robert Wiblin: Well, our guest today has been Philip Trammell. Thanks so much for coming on the 80,000 Hours Podcast, Phil.

Philip Trammell: Thanks for having me, Rob.

Robert Wiblin: And thanks for joining, Howie.

Rob’s outro [02:34:52]

Robert Wiblin: I look forward to returning to this issue once the research Phil and others are doing on it has advanced further.

This episode being a few months old, I’m sure he has had a few further insights in the meantime.

It’s such a wonderful topic, with so many compelling considerations on both sides, I can’t wait to see where things stand in a few years’ time.

The 80,000 Hours Podcast is produced by Keiran Harris. Audio mastering by Ben Cordell. Transcripts by Zakee Ulhaq.

Thanks for joining, talk to you next week.