I don’t know much about economics, but this section doesn’t support the claim in the title. It simply has the title stating the claim and then states Open Phil funded policy.
But there’s many steps between funding and the outcomes the post is predicated on. For example, Open Phil’s funding might have been clumsy and ineffective, or on the other hand, it may have funded sophisticated thinkers who anticipated today’s “problems”.
While you do touch on the issues, it’s unclear whether the post has a useful model for what is going on. Then, in a deep sense, then it’s unclear how it can claim advocacy is good or bad.
I think understanding the 1) link between the funding and what was actually performed, and 2) getting a model of inflation from macroeconomics is the hard part.
This literacy and precision seems hard and important. Even glimpses of what might be going on when Open Phil funds a think tank is valuable.
Yes, I haven’t analyzed the effectiveness of these grants in depth and indeed postulate that these grants are relatively effective given OpenPhil’s track record in other areas and also their self-evaluation (they themselves say the grants are effective). But yes I do cite one paper that argues that ‘Technopopulism and Central Banks’ arguing that such advocacy is not very ineffective.
One thing I did come across during my research is an interaction on Twitter between Furman—Obama’s former economic advisor—and the CEO of EmployAmerica. Furman tweeted something against loose policy and then the EmployAmerica CEO jumped in and corrected him on some numbers and Furnam actually updated towards are more dovish position (now he seems to be back towards being a bit more hawkish). Policymakers listen to Furnam… so that’s already one plausible datapoint for policy influence of OpenPhil’s grantee.
As such, I think it’s prudent to assume that OpenPhil had some influence on macroeconomic policy, but the effect size might be small. I actually do think that the effect size will likely be small because overall funding for these think tanks was relatively modest in 2021. But especially given their commitment to hits-based giving, in expectations the effect size could have been higher.
However, ultimately, I see the first order effects of the advocacy on the macroeconomy are perhaps not as important as the second order effects of making suboptimal grants on OpenPhil and the wider EA community (including reputational effects).
Thanks for this thoughtful reply with specific examples and knowledge.
It’s clear that you know more than me and have deep thoughts.
Just to be direct, I wanted to just list out and “punch through” a few further critiques on your post. In the process of doing this, this comment comes off as pretty negative, but this isn’t the goal. It just seems good to push through this.
I found it easy to ungenerously interpret content in your post as rhetoric. For example, I don’t fully understand the top 3 or 4 subsections on inflation, which occupies a lot of your post. This may be a deep and principled examination of inflation, but it also seems to be simply enumerating the downsides of inflation. Yet, everyone agrees the recent, sudden and continuing unexpected inflation is a concern and wants to stop it.
As mentioned in my comment above (and also below in this comment), I didn’t understand how the realized inflation was tied to Open Phil’s advocacy, which seems like the point of your post. Because of this and how it precedes the section on claiming the advocacy is bad, I found it easy to pattern match this inflation to something like rhetoric. This affected my perception or the aesthetics of the rest of your post.
In a more direct way, I interpret the section in the end as containing rhetoric, “too little epistemic modesty”, “Technocratic blind spots”. I understand the point you are making, but someone who doesn’t agree that the case has been made, this weakens the post by giving an easy way to ascribe existing beliefs to the author (which may not be fair).
COVID-19 seems like a “table flip” that needs to be addressed. I don’t think anyone really knows how inflation works or how to start or stop it perfectly. But I think we can agree the pandemic was a wild event and maybe there was wild economic action taken. This wild action alone could both cause inflation, or interact with any sort of advocacy that you are criticizing. This could completely swamp out the effects of this advocacy itself and makes judgement difficult. I didn’t really see a lot of consideration of this (or the counterfactual of different policies during the pandemic). This is hard, but touching on this in a way that seems balanced seems important.
Why set the date 2021 to divide good and bad advocacy? I’m really uncertain about the “2021” date that you choose to divide advocacy between good or bad. I think you are really committed to this date, and it comes up again in the comments. But I think that, even setting aside or completely accepting every other idea or claim in the post, it’s very likely that actions taken from 2014 to 2020, also influence inflation today in 2022. In fact, if you’re claiming the current inflation is bad, it seems plausible that underlying processes started in early in 2021 or 2020. It seems possible the inflation, in a physical logical sense, precedes the effects of the funding that you say is good/benign.
The reason why I wanted to write the comment above is just to air out the thoughts in it. I’m not sure but my sentiment is that I think this could ultimately enable you, maybe by going past the particular topic of inflation.
In your post, there are a lot of interesting points, and it feels like going for this directly would make awesome writing in itself:
In particular, your view on “technocracy” seems interesting. Why not examine this directly?
What exactly is technocracy? What does this term even mean (like, isn’t criticism of technocracy self-contradictory)? What is the inherent problem/challenges/tradeoffs of “using technocracy”, or not using it? A well informed, fully principled dissection/take down seems interesting and could occupy many posts.
Also, it seems like you could be enabled directly much more. Like, maybe you should have access to pretty candid interviews with the think tanks grantees on what their policy is, as well as Open Phil’s team. This means that you can take on some of the “implications and recommendations” that you bring up.
Maybe you are entitled to enablement in a compelling sense: I speculate/have information that your post, “Growth and the case against randomista development” might also be the highest rated post in the EA forum 10 year contest. So, like, you literally have the best post ever here. That seems important and an achievement for you and John Halstead.
Thanks for the feedback- all taken in good spirit :) This is helpful as it helps me clarify.
I’ll try to explain a few assumptions implicit in the post and not assume very much macroeconomics knowledge, so sorry if the following is trivial or comes across as didactic, but maybe I’ve skipped a few steps / assumed too much knowledge. It’s quite a complex topic and it’s hard to wrap your head around and keep everything in mind as there are so many moving parts, this is why I kept things quite tight and terse. Ezra Klein also said a similar thing in a recent podcast with Furman. I also focused on the ‘case against higher inflation’ while trying to be balanced at points and mentioning the upsides. The ambition was not to provide a deep and principled examination of inflation as a whole, but rather point towards potential dangers, which on the margin, is neglected in this community. There are very clear drawbacks also of having low inflation and sluggish recovery after 2008, but that was beyond the scope of this post. Also it was beyond the scope to tie OpenPhil’s advocacy to higher inflation—I provided some evidence for that but mostly postulated that it is true because they say they think they’ve had an effect on this in the past.
How is inflation tied to Open Phil’s advocacy?
In macro, there’s a basic trade-off between unemployment and inflation (see more here). Politicians can create big stimulus packages and increase government debt, and that stimulates the economy and this creates jobs and leads to higher wages. More jobs lead to higher wages (c.f. Phillips curve—Wikipedia ), this leads to more inflation. If the stimulus is too big (above the output gap) of a recession, then you overshoot and the economy runs hot. Also, if the central bank prints too much money, keeps interest rates too low, or does quantitative easing, then this can of course also create inflation. There’s disagreement around the specifics, the time horizons, and how big the multipliers are and so on, but at the end of the day there’s economic consensus that at some point inflation will be too high, and you can only print so much money, and your stimulus can only be so big not to overheat the economy.
The basic causal model underlying this post is that OpenPhil’s grantees have advocated for bigger fiscal stimulus and also more expansionary monetary policy to create more jobs (e.g. funding ‘EmployAmerica’). This might have been correct after the sluggish recovery of 2008, where we should have been more expansionary, but now we’ve probably overshot.
everyone agrees the recent, sudden and continuing unexpected inflation is a concern and wants to stop it.
So there’s a lot to unpack here in this sentence.
Everyone agrees-
Only now, slowly, even the most dovish economists are coming around to thinking inflation is too high, and that we shouldn’t have been so loose in retrospect. Now the doves fade from view and even OpenPhil’s grantees write that now ‘it is more defensible to slow down the labor market for the sake of lower inflation’, after pushing a dovish stance for until recently.
recent, sudden and continuing unexpected inflation
It’s not very recent… inflation was already quite high last year, and also not very sudden: again already in May ’21, the futures markets predicted ~2.7% average inflation in the US over the next 5 years (which suggests it’s somewhat more persistent/continuing than those in team transitory might suggest) and an econometrics paper published then estimated a 32% chance of >6% inflation in 2022 and warned that the discussion was based on casual observation rather than econometrics. So the strong claim is that this wasn’t recent, sudden, unexpected and could have been predicted had OpenPhil put adequate resources towards analyzing this properly (which they don’t seem like they have). But the weak claim is that I’m just cherry picking the few people who correctly predicted that inflation is going to be bad (and there are many people who have ‘cried wolf’ on inflation and thus are not being listened to anymore… and many people joke that Larry Summers, who warned early about inflation last year, had to be right on something eventually). But if this turns out to be just a mistake ex-post, we might still perhaps learn from it.
The section contains rhetoric, “too little epistemic modesty”, “Technocratic blind spots”
I see what you’re pointing towards, and several reviewers indeed suggested taking out these sections, saying it weakens the piece, but I decided to leave them in, as I think it’s important. I could have just criticized the proximal cause of the mistake saying ‘We’ve caused too much inflation and this has bad effects and there were econometric models that predicted that, next time look at these models’. And then the response could have been ‘Oh well, mistakes were made, human error, hits-based giving, you win some, you lose some, we didn’t have such a big impact anyway at the size of the grants’. But I think there’s just something deeper and a bit crazy ‘big, if true’ stuff that’s amiss here, that I was trying to gesture at in those sections.
Like, if you really stop and think about it, that OpenPhil funded lobbying groups to influence foreign central banks, and then this was (if my analysis is correct, which might not be the case) not going in the correct direction, then that’s quite something. It’s not like they had funded something uncontroversially good (say human rights stuff abroad) and that had a bad side effect.
And I’d love to know what others think about that. More generally, on the margin, we have too much data in the EA community, and not enough theory.
COVID-19 seems like a “table flip” that needs to be addressed.
The pandemic was of course unprecedented and was a unique challenge and inflation is this environment is challenging. However, even though there’s a kernel of truth to what you say with ‘nobody really knows how inflation works and how to start or stop it perfectly’ this is a bit like ‘throwing out the baby with the bathwater’. We have some ideas and I cite some economists who created models and BOTECs that for instance suggested that the stimulus was based on atheoretical casual observation and so arbitrarily decided and thus somewhat arbitrarily large. To me this seems a bit crazy. And even if things were a bit wild as you say, I think we can still criticize bad decisions here, and it doesn’t mean we should fund ‘wild’ advocacy.
Why set the date 2021 to divide good and bad advocacy?
2021 is of course a somewhat arbitrary cutoff point. Sometimes there are lag times from policy changes to the economy reacting, and then on top of that you have more lag from advocacy affecting policy change etc. So it’s not inconceivable that pre 2021 grants also had negative effects.
However, there were reasons for this:
OpenPhil made two grants in mid-2021 to Dezernat Zukunft and EmployAmerica.
It also relates to my ‘Strong claim: Risk could have been predicted before the event and loose policy was a mistake ex-ante’ in the sense that in 2021 there were some voices and some analysis that predicted the current inflation overshoot. I haven’t found serious analysis that predicted current very high inflation in 2020.
your view on “technocracy” seems interesting. Why not examine this directly?
There’s something here that technocrats who lobby for higher inflation do not have to deal with consequences in a very relevant sense in that they are not as affected by high food prices etc. because they’re rich and thus might have technocratic blindspots.
maybe you should have access to pretty candid interviews with the think tanks grantees on what their policy is, as well as Open Phil’s team.
It’s all pretty transparent from OpenPhil’s grantees websites which I browsed, but I choose to not go into too much detail here, because I didn’t want to blame the funded orgs, but rather go more directly to the source (the founder).
Maybe you are entitled to enablement in a compelling sense
I think generally, it’s good to not privilege anyone’s arguments in any way (unless they have very specific technical expertise maybe… that’s why I try to cite the IGM booth economic expert survey as much as possible, and a broad range of experts from both the left and the right… even Nobelists can be very wrong)
I don’t know much about economics, but this section doesn’t support the claim in the title. It simply has the title stating the claim and then states Open Phil funded policy.
But there’s many steps between funding and the outcomes the post is predicated on. For example, Open Phil’s funding might have been clumsy and ineffective, or on the other hand, it may have funded sophisticated thinkers who anticipated today’s “problems”.
While you do touch on the issues, it’s unclear whether the post has a useful model for what is going on. Then, in a deep sense, then it’s unclear how it can claim advocacy is good or bad.
I think understanding the 1) link between the funding and what was actually performed, and 2) getting a model of inflation from macroeconomics is the hard part.
This literacy and precision seems hard and important. Even glimpses of what might be going on when Open Phil funds a think tank is valuable.
Yes, I haven’t analyzed the effectiveness of these grants in depth and indeed postulate that these grants are relatively effective given OpenPhil’s track record in other areas and also their self-evaluation (they themselves say the grants are effective). But yes I do cite one paper that argues that ‘Technopopulism and Central Banks’ arguing that such advocacy is not very ineffective.
One thing I did come across during my research is an interaction on Twitter between Furman—Obama’s former economic advisor—and the CEO of EmployAmerica. Furman tweeted something against loose policy and then the EmployAmerica CEO jumped in and corrected him on some numbers and Furnam actually updated towards are more dovish position (now he seems to be back towards being a bit more hawkish). Policymakers listen to Furnam… so that’s already one plausible datapoint for policy influence of OpenPhil’s grantee.
As such, I think it’s prudent to assume that OpenPhil had some influence on macroeconomic policy, but the effect size might be small. I actually do think that the effect size will likely be small because overall funding for these think tanks was relatively modest in 2021. But especially given their commitment to hits-based giving, in expectations the effect size could have been higher.
However, ultimately, I see the first order effects of the advocacy on the macroeconomy are perhaps not as important as the second order effects of making suboptimal grants on OpenPhil and the wider EA community (including reputational effects).
Thanks for this thoughtful reply with specific examples and knowledge.
It’s clear that you know more than me and have deep thoughts.
Just to be direct, I wanted to just list out and “punch through” a few further critiques on your post. In the process of doing this, this comment comes off as pretty negative, but this isn’t the goal. It just seems good to push through this.
I found it easy to ungenerously interpret content in your post as rhetoric. For example, I don’t fully understand the top 3 or 4 subsections on inflation, which occupies a lot of your post. This may be a deep and principled examination of inflation, but it also seems to be simply enumerating the downsides of inflation. Yet, everyone agrees the recent, sudden and continuing unexpected inflation is a concern and wants to stop it.
As mentioned in my comment above (and also below in this comment), I didn’t understand how the realized inflation was tied to Open Phil’s advocacy, which seems like the point of your post. Because of this and how it precedes the section on claiming the advocacy is bad, I found it easy to pattern match this inflation to something like rhetoric. This affected my perception or the aesthetics of the rest of your post.
In a more direct way, I interpret the section in the end as containing rhetoric, “too little epistemic modesty”, “Technocratic blind spots”. I understand the point you are making, but someone who doesn’t agree that the case has been made, this weakens the post by giving an easy way to ascribe existing beliefs to the author (which may not be fair).
COVID-19 seems like a “table flip” that needs to be addressed. I don’t think anyone really knows how inflation works or how to start or stop it perfectly. But I think we can agree the pandemic was a wild event and maybe there was wild economic action taken. This wild action alone could both cause inflation, or interact with any sort of advocacy that you are criticizing. This could completely swamp out the effects of this advocacy itself and makes judgement difficult. I didn’t really see a lot of consideration of this (or the counterfactual of different policies during the pandemic). This is hard, but touching on this in a way that seems balanced seems important.
Why set the date 2021 to divide good and bad advocacy? I’m really uncertain about the “2021” date that you choose to divide advocacy between good or bad. I think you are really committed to this date, and it comes up again in the comments. But I think that, even setting aside or completely accepting every other idea or claim in the post, it’s very likely that actions taken from 2014 to 2020, also influence inflation today in 2022. In fact, if you’re claiming the current inflation is bad, it seems plausible that underlying processes started in early in 2021 or 2020. It seems possible the inflation, in a physical logical sense, precedes the effects of the funding that you say is good/benign.
The reason why I wanted to write the comment above is just to air out the thoughts in it. I’m not sure but my sentiment is that I think this could ultimately enable you, maybe by going past the particular topic of inflation.
In your post, there are a lot of interesting points, and it feels like going for this directly would make awesome writing in itself:
In particular, your view on “technocracy” seems interesting. Why not examine this directly?
What exactly is technocracy? What does this term even mean (like, isn’t criticism of technocracy self-contradictory)? What is the inherent problem/challenges/tradeoffs of “using technocracy”, or not using it? A well informed, fully principled dissection/take down seems interesting and could occupy many posts.
Also, it seems like you could be enabled directly much more. Like, maybe you should have access to pretty candid interviews with the think tanks grantees on what their policy is, as well as Open Phil’s team. This means that you can take on some of the “implications and recommendations” that you bring up.
Maybe you are entitled to enablement in a compelling sense: I speculate/have information that your post, “Growth and the case against randomista development” might also be the highest rated post in the EA forum 10 year contest. So, like, you literally have the best post ever here. That seems important and an achievement for you and John Halstead.
Thanks for the feedback- all taken in good spirit :) This is helpful as it helps me clarify.
I’ll try to explain a few assumptions implicit in the post and not assume very much macroeconomics knowledge, so sorry if the following is trivial or comes across as didactic, but maybe I’ve skipped a few steps / assumed too much knowledge. It’s quite a complex topic and it’s hard to wrap your head around and keep everything in mind as there are so many moving parts, this is why I kept things quite tight and terse. Ezra Klein also said a similar thing in a recent podcast with Furman. I also focused on the ‘case against higher inflation’ while trying to be balanced at points and mentioning the upsides. The ambition was not to provide a deep and principled examination of inflation as a whole, but rather point towards potential dangers, which on the margin, is neglected in this community. There are very clear drawbacks also of having low inflation and sluggish recovery after 2008, but that was beyond the scope of this post. Also it was beyond the scope to tie OpenPhil’s advocacy to higher inflation—I provided some evidence for that but mostly postulated that it is true because they say they think they’ve had an effect on this in the past.
In macro, there’s a basic trade-off between unemployment and inflation (see more here). Politicians can create big stimulus packages and increase government debt, and that stimulates the economy and this creates jobs and leads to higher wages. More jobs lead to higher wages (c.f. Phillips curve—Wikipedia ), this leads to more inflation. If the stimulus is too big (above the output gap) of a recession, then you overshoot and the economy runs hot. Also, if the central bank prints too much money, keeps interest rates too low, or does quantitative easing, then this can of course also create inflation. There’s disagreement around the specifics, the time horizons, and how big the multipliers are and so on, but at the end of the day there’s economic consensus that at some point inflation will be too high, and you can only print so much money, and your stimulus can only be so big not to overheat the economy.
The basic causal model underlying this post is that OpenPhil’s grantees have advocated for bigger fiscal stimulus and also more expansionary monetary policy to create more jobs (e.g. funding ‘EmployAmerica’). This might have been correct after the sluggish recovery of 2008, where we should have been more expansionary, but now we’ve probably overshot.
So there’s a lot to unpack here in this sentence.
Only now, slowly, even the most dovish economists are coming around to thinking inflation is too high, and that we shouldn’t have been so loose in retrospect. Now the doves fade from view and even OpenPhil’s grantees write that now ‘it is more defensible to slow down the labor market for the sake of lower inflation’, after pushing a dovish stance for until recently.
It’s not very recent… inflation was already quite high last year, and also not very sudden: again already in May ’21, the futures markets predicted ~2.7% average inflation in the US over the next 5 years (which suggests it’s somewhat more persistent/continuing than those in team transitory might suggest) and an econometrics paper published then estimated a 32% chance of >6% inflation in 2022 and warned that the discussion was based on casual observation rather than econometrics. So the strong claim is that this wasn’t recent, sudden, unexpected and could have been predicted had OpenPhil put adequate resources towards analyzing this properly (which they don’t seem like they have). But the weak claim is that I’m just cherry picking the few people who correctly predicted that inflation is going to be bad (and there are many people who have ‘cried wolf’ on inflation and thus are not being listened to anymore… and many people joke that Larry Summers, who warned early about inflation last year, had to be right on something eventually). But if this turns out to be just a mistake ex-post, we might still perhaps learn from it.
I see what you’re pointing towards, and several reviewers indeed suggested taking out these sections, saying it weakens the piece, but I decided to leave them in, as I think it’s important. I could have just criticized the proximal cause of the mistake saying ‘We’ve caused too much inflation and this has bad effects and there were econometric models that predicted that, next time look at these models’. And then the response could have been ‘Oh well, mistakes were made, human error, hits-based giving, you win some, you lose some, we didn’t have such a big impact anyway at the size of the grants’. But I think there’s just something deeper and a bit crazy ‘big, if true’ stuff that’s amiss here, that I was trying to gesture at in those sections.
Like, if you really stop and think about it, that OpenPhil funded lobbying groups to influence foreign central banks, and then this was (if my analysis is correct, which might not be the case) not going in the correct direction, then that’s quite something. It’s not like they had funded something uncontroversially good (say human rights stuff abroad) and that had a bad side effect.
And I’d love to know what others think about that. More generally, on the margin, we have too much data in the EA community, and not enough theory.
The pandemic was of course unprecedented and was a unique challenge and inflation is this environment is challenging. However, even though there’s a kernel of truth to what you say with ‘nobody really knows how inflation works and how to start or stop it perfectly’ this is a bit like ‘throwing out the baby with the bathwater’. We have some ideas and I cite some economists who created models and BOTECs that for instance suggested that the stimulus was based on atheoretical casual observation and so arbitrarily decided and thus somewhat arbitrarily large. To me this seems a bit crazy. And even if things were a bit wild as you say, I think we can still criticize bad decisions here, and it doesn’t mean we should fund ‘wild’ advocacy.
2021 is of course a somewhat arbitrary cutoff point. Sometimes there are lag times from policy changes to the economy reacting, and then on top of that you have more lag from advocacy affecting policy change etc. So it’s not inconceivable that pre 2021 grants also had negative effects.
However, there were reasons for this:
OpenPhil made two grants in mid-2021 to Dezernat Zukunft and EmployAmerica.
It also relates to my ‘Strong claim: Risk could have been predicted before the event and loose policy was a mistake ex-ante’ in the sense that in 2021 there were some voices and some analysis that predicted the current inflation overshoot. I haven’t found serious analysis that predicted current very high inflation in 2020.
I think this would be beyond the scope of what I can take on currently. I refer the interested reader to the Weyl piece that I cite and also see how it contrasts with a recent post by Holden.
There’s something here that technocrats who lobby for higher inflation do not have to deal with consequences in a very relevant sense in that they are not as affected by high food prices etc. because they’re rich and thus might have technocratic blindspots.
It’s all pretty transparent from OpenPhil’s grantees websites which I browsed, but I choose to not go into too much detail here, because I didn’t want to blame the funded orgs, but rather go more directly to the source (the founder).
I think generally, it’s good to not privilege anyone’s arguments in any way (unless they have very specific technical expertise maybe… that’s why I try to cite the IGM booth economic expert survey as much as possible, and a broad range of experts from both the left and the right… even Nobelists can be very wrong)