Tuesday, April 22, 2014

Tom Sargent: What's Between the Lines

For some obscure reason, Tom's Sargent's Berkeley graduation address from seven years ago is now getting attention in the blogosphere. It's innocuous stuff - basically core ideas in economics boiled down to life lessons for graduating seniors. Paul Krugman seems a little insecure about it though (here and here) and would like you to dismiss parts of it as some sort of right-wing "cockroach" ideas. I can see what Krugman is getting at - he might see these ideas as threatening, as they are at odds with his notion that economic policy is easy, or that large welfare gains from changes in policy are low-handing fruit. But I think you have to be a rather suspicious person to think that Sargent's synopsis of economic ideas represents a particular view on policy. Indeed, the ideas are entirely consistent with how Mike Woodford, or Ed Prescott, thinks about the world.

So, let's go through Sargent's address, so you can see what I mean:

I remember how happy I felt when I graduated from Berkeley many years ago. But I thought the graduation speeches were long. I will economize on words. Economics is organized common sense. Here is a short list of valuable lessons that our beautiful subject teaches.

[I had to think about whether I agree with the "organized common sense" view of economics. If common sense is supposed to be what the average layperson possesses, then economics is not common sense, as it's sometimes (if not often) counterintuitive. For example, Adam Smith told us that greed can be a good thing. That's certainly not part of collective wisdom, I think. But Sargent was trying to put the Berkeley graduates at ease. He's telling them that doing economics is just a matter of putting pieces of straightforward logic together to come up with a coherent set of ideas.]

1. Many things that are desirable are not feasible.
[This is typically one of the first things taught in Econ 101. As a society, and as individuals, we face constraints. But what are society's constraints? That's what economists argue about. Some people think, for example, that the U.S. economy is no more constrained in what it can achieve in 2014 than it was in 2007. Those people think that the decision-makers in the Fed and the U.S. Congress, are somehow not optimizing given their constraints - they think there are more desirable outcomes that are in fact feasible. Others don't agree.]

2. Individuals and communities face trade-offs.
[This is again a key part of any Econ 101 course - opportunity cost. This is about the nature of the constraints we face. Individuals face budget constraints. One can't have a new car without giving up something else - other goods and services in the present, or future consumption. Economic growth models tell us that, as a society, to attain higher productive capacity in the future, we have to give up consumption in the present. But, at the societal level, again, economists have key arguments about what the tradeoffs are. In the 1960s, Solow and Samuelson argued that a Phillips curve tradeoff existed - the central bank could achieve lower unemployment, but only at the expense of higher inflation; Milton Friedman said that this was only a short-run tradeoff, and in the long run no such tradeoff exists; the Phillips curve went out of fashion; the Phillips curve was resurrected in the 1990s by New Keynesians; many macroeconomists are still skeptical.]

3. Other people have more information about their abilities, their efforts, and their preferences than you do.
[I'm glad Sargent included this idea, which is not part of the typical Econ 101 curriculum. But this is at the heart of many contemporary economic issues. Why are there banks, and what do they do? What were the incentive problems that led up to the global financial crisis? What's a financial crisis about anyway? How do we explain unemployment and the process of labor market search? For all of these issues we have to think carefully about private information and how it matters for the functioning of markets and economic institutions. A whole array of economic machinery developed over the last 50 years allows us to come to grips with these things.]

4. Everyone responds to incentives, including people you want to help. That is why social safety nets don’t always end up working as intended.
[This is closely related to #3. Take unemployment insurance (UI), for example. A key problem in designing a UI system is that there is moral hazard - it's hard for the government to observe how much effort the unemployed are putting into looking for a job, and the more generous is UI, the less effort is expended in searching. So, this represents another societal tradeoff. Providing better UI is good, as it provides insurance against a bad outcome - being unemployed. But the better insurance UI provides, the lower is the search effort of the unemployed, on average, and the higher will be the average unemployment rate. The latter is bad for society, as we lose output.]

5. There are tradeoffs between equality and efficiency.
[Closely related to #5. For example, redistribution of income through the income tax system (progressive taxation by which we tax the rich at a higher rate than the poor) can be seen as a form of social insurance. Being born poor is no fault of an individual, and may disadvantage him or her for life. Just like UI, progressive taxation acts to insure people against bad outcomes. But redistribution changes incentives. If I am born poor, the insurance provided by the government discourages me from investing in my own human capital - acquiring education and skills that allow me to do better. Similarly, those born rich are also discouraged. Why acquire high-level human capital if the return on that human capital is low because it is taxed at a high rate? Again, society faces a tradeoff - we can redistribute the pie, but that tends to make the pie smaller.]

6. In an equilibrium of a game or an economy, people are satisfied with their choices. That is why it is difficult for well meaning outsiders to change things for better or worse.
[Note here that Sargent doesn't say it's impossible to make things better - it's just difficult. Here's an example. A typical Econ 101 example of this principle at work is the minimum wage law. Workers and firms engage in private contracts. Firms offer jobs at particular wages - the people who are employed in those jobs have accepted those terms of employment. The firms and workers did the best they could under the circumstances. So why should we intervene in these private contracts, for example by imposing a floor on the wage a firm can offer a worker? Well, we might see this as social insurance, in which case we are back to #5. What's the tradeoff? Further, an issue that comes up here is that the tradeoff might be better if we use one mechanism to redistribute income rather than another. In this case, Econ 101 tells us that we're better off as a society if we redistribute using the tax system, or lump sum payments to the poor rather than by using the minimum wage. One problem with the minimum wage is that we're taxing the firms we're asking to provide employment for the poor.]

7. In the future, you too will respond to incentives. That is why there are some promises that you’d like to make but can’t. No one will believe those promises because they know that later it will not be in your interest to deliver. The lesson here is this: before you make a promise, think about whether you will want to keep it if and when your circumstances change. This is how you earn a reputation.
[Anyone who has had to deal with 2-year-olds - or has been the chair of an economics department - understands this perfectly. The people whose behavior you are trying to shape with your current promises have excellent memories and will punish you forever if you break those promises. This is the basis for limited commitment models in economics. Limited commitment - e.g. the "time consistency" problem - has been highly influential in policy debates, from Kydland and Prescott's watershed work on. Limited commitment is also key to credit relationships, and the use of collateral - again crucial for figuring out the financial crisis, and the role of policy in the crisis.]

8. Governments and voters respond to incentives too. That is why governments sometimes default on loans and other promises that they have made.
[Again, this is critical for thinking about current policy issues - sovereign debt problems for example. Commitment and incentives were at the heart of the Fed's rationale for its recent forward guidance experiments. In New Keynesian macro models, which formed the theoretical backdrop for the policy, promises about the future are important for attaining good results in the present, in contexts where conventional policy doesn't work. But when the future arrives, the policymaker will want to go back on the promise, so commitment to the policy is required. Of course, this is a bad example of commitment at work, as the Fed botched it.]

9. It is feasible for one generation to shift costs to subsequent ones. That is what national government debts and the U.S. social security system do (but not the social security system of Singapore).
[Most of what you need to know about this is in Peter Diamond's 1965 paper. This is one of the neatest tricks in economics. The world consists of overlapping generations of individuals. If we can get each young generation to give a gift to each old generation, then (under some conditions) we can make everyone better off - forever. So there are some circumstances where - indeed - we can get something for nothing. In practice, the gifts from young to old can be accomplished through a social security system, through government debt (as Diamond showed), or through monetary exchange (as Samuelson showed). Very cool.]

10. When a government spends, its citizens eventually pay, either today or tomorrow, either through explicit taxes or implicit ones like inflation.
[This is closely related to #1 and #2. Running a government requires real resources - workers, buildings, fuel, etc. Those resources don't come for free - there are tradeoffs. Some people - Krugman for example - would argue that there are circumstances in which more government spending is effectively free. The argument is that, if there are unemployed workers willing to work, then employing them makes them better off, and gives the rest of us more stuff to boot. In the story, there can even be a multiplier - a free lunch - by which one unit of government spending produces more than one unit of additional GDP. The odd thing about this is the disconnect between modern Keynesians and Old Keynesians on the issue. Multipliers somehow seem convincing - they satisfy the "common sense" test of the person on the street. But the standard multiplier mechanism is nowhere to be seen in a mainstream New Keynesian model. Further, in episodes like World War II with massive increases in government spending, it's clear what the resource costs are - a small drop in current private consumption, and a very large drop in private investment spending (future consumption).]

11. Most people want other people to pay for public goods and government transfers (especially transfers to themselves).
[You can see this in current debates over health care. Some senior citizens receiving medicare benefits see those benefits as some kind of birth right, but object to government programs that attempt to extend health insurance to the previously uninsured. There are some goods and services for which public provision is appropriate - national defense being the leading example. But it's impossible to provide an appropriate amount of public goods if we are going to finance their provision through voluntary contributions (as we do in NPR pledge drives). We have to collectively agree to government confiscation (tax collection) to make this work.]

12. Because market prices aggregate traders’ information, it is difficult to forecast stock prices and interest rates and exchange rates.
[This is just simple financial arbitrage - there are no free lunches available to your average financial market participant. Fama called this the "efficient markets hypothesis." But that's a misnomer, as by now we understand that perfect arbitrage in financial markets can occur in economies which are highly inefficient in the standard economic sense.]

So, Sargent has been pretty terse - that's the way he is. But these 12 ideas are very much mainstream. The ideas are consistent with economic inefficiency, a role for government, and provide a basis for analyzing how governments and central banks can make our lives better. I'm not sure what Paul Krugman is afraid of.


  1. Stephen, I posted Sargent speech originally to make a point about about how economic analysis is free of political bias.

    Sargent is a life-long Democrat. All too many immediately smeared him as a right-winger and neo-liberal.

    Krugman was worse, saying that the speech appeared on the net with the "real goal here is to undermine the case for fighting unemployment in the here and now."

    I live in New Zealand. Not everything is about U.S. domestic politics.

    1. Jim,

      Thanks for posting it. That's very interesting. Krugman also made the point of linking to the guy from the American Enterprise Institute who re-posted it, to try to emphasize the point that this was some right-wing conspiracy.

      "Not everything is about U.S. domestic politics."


    2. This comment has been removed by the author.

    3. Spoken like a true Canadian.

    4. I live in the U.S. and feel at home here, but they do get a little weird sometimes.

  2. You spent all that time to say that Sargent was generally correct? It's Tom Sargent. Yes, of course he's generally correct.

    Do you really want to lob softballs that you're "not sure " what PK is afraid of? Um, you're the smartest economist out there and you can't understand his obvious point that a simplified graduation speech may not always correlate exactly to each real world situation and that people will misuse the simplified ideas in such a manner that will do more harm than good? I cannot understand how someone can spend all these massive intellectual resources of yours and still not understand PK's simple point, but whatever. Filters are strong. One way or the other, filters are really really strong. Why engage when you can just ignore and move the discussion to something else?

    1. "...that people will misuse the simplified ideas in such a manner that will do more harm than good?"

      Apparently you live in fear too. Get out there and control the message. Don't want the unwashed masses getting the wrong ideas.

  3. I think #2 has an ambiguity. It seems to say simply that tradeoffs are ever present, both for individuals and for communities, though their tradeoffs can differ.

    Another possible interpretation (though it does not literally say this) is that policy decisions involve tradeoffs between the benefits (or rights) of individuals and the benefit of the community. Now it is framed in terms of the individual vs collective political hot button. Whether Sargent meant this or not is open to debate, but I think this is the reading that concerned Krugman.

    1. Is that supposed to be constructive? Are you suggesting my comment is a crackpot theory?

      If you read this, it should be pretty clear that my comment is relevant: http://krugman.blogs.nytimes.com/2014/04/21/the-economy-is-not-like-a-household/?module=BlogPost-Title&version=Blog%20Main&contentCollection=Opinion&action=Click&pgtype=Blogs&region=Body

    2. I read Krugman's piece, which I linked to in the post. I think both of you are being quite silly.

  4. I expected the smearing campaign by PK, but I was disappointed to see our common friend Noah jump in. People read #3 and all they see is a case against government paternalism. But as you correctly state, #3 is relevant also for efficiency wages and credit market frictions, which create inefficiencies and therefore open the door to government intervention. How scared can someone be to keep seeing enemies at the gate all the time?

    1. Yes, I saw that and I thought "asymmetric information." Joe Stiglitz.

  5. I think the main quip readers have with #4 is that it misleads. Social safety net programs surely invite moral hazard, but the combination of ordeal mechanisms, positive externalities associated with higher income, time limits on the duration of benefits, the human desire to accomplish, the diminishing marginal utility of the dollar, and the second welfare theorem all make social safety net programs much more beneficial than costly. So, the comment looks like a cheap jab rather than a substantive remark on the grand lessons of intermediate micro.

    1. Your list of things contains only nonsense and falsehoods.

    2. Ok, let's look at a couple welfare programs, then discuss the extent to which they are "nonsense and falsehoods".

      TANF gives temporary financial aid, a 60 month maximum, to low-income families. TANF also requires 30 hours of work per week, and benefits can be cut or reduced if a woman is having a child outside marriage.

      Unemployment insurance you're probably aware of. It lasts 26 weeks in a typical year and helps provide labor market efficiency by allowing workers time to find jobs matching their skill. While unemployed, workers receive 40-50% their previous pay.

      The EITC is a refundable tax credit that essentially pays families for the work they contribute given that they make under a certain amount per year and meet a few other requirements.

      I'd like to know where, in the examples I gave, exactly where the "wrong incentives" are, and how those perverse incentives make the programs wholly ineffective or not worthwhile.

      In each welfare program, there are clear, limited durations for each benefit, work is required for each, and they each have social benefits. I've only touched on one of the topics I listed above, but suffice to say there are clear objections to the extent to which #4 is true. This is why, in my initial comment, I stated that it just comes across as a cheap shot, not a strong, substantive lesson that economics teaches us. If anything, economics teaches us that clever redistribution of resources results in a better society than one that do nothing to redistribute resources. See Welfare Theorem 2, Mirrlees.

    3. There is no positive externality associated with higher income. Limiting the duration of unemployment insurance reduces the moral hazard problem but it does not eliminate it. You are correct that the program may increaseefficiency by improving the quality of matches, but it is not clear that what we have in place is optimal (I think Steve has done some work on this). And do not forget that these programs are funded by distortionary taxation. This is not to say that they are bad or that they need to be eliminated (I support all three that you mentioned) but pretending they do not involve trade-offs is dishonest to say the least. Some liberals justify this by pointing out the dishonesty of the other side. But an eye for an eye leaves everybody blind.

    4. I want to have higher income, therefore I insist there is an important externality associated with me having low income. See how easy it is? And you've clearly misunderstood Mirrlees, who is talking about raising revenue in a world of incomplete insurance (insurance is not redistribution) and apparently don't understand the Second Welfare Theorem either. Perhaps a refresher course in principles would help?

  6. Steve,
    I have read the first paragraph of your blog post and Paul Krugman’s second post which is solely about Sargent’s second lesson, that there is no free lunch, neither for individuals nor for societies. Krugman disputes that this is true in general and in particular for aggregate demand management policies now. Reading you commentary on Sargent’s ninth lesson, that social security indeed provides a free lunch for society in dynamically inefficient overlapping generations economies, I would have thought that you agreed with Krugman in general, but not for his particular application. I suppose that Krugman sees more exceptions to the rule than either you or Sargent do, but that is a point one can debate. As for the ulterior reasons of those who posted Sargent’s remarks, that is a debate I can live without.
    Looking forward to your next blog post not involving Paul Krugman.
    Loosing Faith

    1. I think Sargent is honest. I think Krugman is not.

  7. Hello Mr. Wiliamson! I am reading your reccomendation-Kartik Arethya's new book and though I'm a skeptic of what you call 'modern macro' I am quite impressed-I particularly appreciate how accessible it is and his good faith effort to speak to the interested layperson in econ like myself.

    I had previously written of his book-saying what I just wrote here and he graciously responded to it over email. Here I wrote a bit about our encounter. To my mind this is the best of seen in proferssional economists-who don't snidely look down their nose at laypeople-like in my opinion a number of these economists do, Sumner is an example of someone who sometimes does this-but looks to educate and engage them.

    Happy Saturday my friends. My convo with Athreya http://diaryofarepublicanhater.blogspot.com/2014/04/i-exchange-emails-with-katrik-athreya.html

  8. "There are tradeoffs between equality and efficiency."
    Somebody has not read his Stiglitz.

    1. Why don't you enlighten us? (I have a feeling I am going to regret saying that)

    2. "Incentives and Risk Sharing in Sharecropping", Stiglitz 1974

    3. Yes, what about it?

  9. "It is feasible for one generation to shift costs to subsequent ones. That is what national government debts and the U.S. social security system do (but not the social security system of Singapore)."

    Nope. If GDP grows faster than public debt GDP/debt decreases. As Krugman has pointed out some time ago, the cross-sectional dimension of public debt is also important.

    1. I have no idea what you are trying to say.

    2. Government debts do only shift costs to following generations if GDP grows slower than public debt. Otherwise, i.e. when public debt / GDP is stable or declining the real distributional issue is one between the public and the people who hold public bonds.
      This conflict was clearly visible when Greece did not default on its debt (the normal thing to do when you are insolvent) but Germany et al rather decided to bail out the rentiers (mainly German banks holding Greek debt).

      We can learn more from what happened in Europe. Not using public debt for the sake of deficit spending in order to stabilize output lead to massive your unemployment. So when we are in a liquidity trap and monetary policy becomes fairly ineffective a temporal increase of public debt can actually beneficial for the next generation.

      This is why Krugman is correct, Sargent tells a right-wing story (doesn't mean he is a right-winger, most so called "moderate" centrists are nowadays economically right-wing) that a) obfuscates the real distributional issues of public debt.

      About social security systems, yes, pay-as-you-go is similar to a Ponzi scheme (unlike a Ponzi scheme it is stable) but the alternative of a capital-based system has disadvantages as well, namely financial crises and rapidly fluctuating asset prices. While the rate of return of the latter is usually larger than that of the latter (the interest rate is normal larger than the the rate of population growth plus the rate of wage growth, i.e. our economies are usually dynamic efficient) a significant part of the "gross" interest rate goes to financial intermediaries.
      So Sargent's argument about social security system is not really good either.

    3. "Government debts do only shift costs to following generations if GDP grows slower than public debt."

      The usual result is that a dynamic inefficiency exists (roughly) if the rate of interest exceeds the population growth rate, and that dynamic inefficiency can be corrected by a pay-as-you-go social security system. Or more precisely, see work by Yves Balasko and Karl Shell.

      Sargent is not obfuscating anything. If you think that's a right-wing story, you have a bizarre view of the world. He's not suggesting that fully-funded social security is superior to pay as you go. You're reading something into this that's not there.

    4. Anyone using the word "rentier" is just angry that someone else is smarter and richer than he is.