The Accidental Theorist: And Other Dispatches from the Dismal Science

by Paul Krugman

  During the seventies the Sweeneys were members of a baby-sitting co-op—a group of young couples, mostly working on Capitol Hill, who agreed to baby-sit for each others’ children on a rotating basis. Any such group, once it goes beyond a few members, requires some sort of system to make sure that each couple does its fair share. This particular co-op settled on a scrip system: Members were issued coupons worth one hour of baby-sitting time. When a couple went out for an evening, the baby-sittees would give the appropriate number of coupons to the baby-sitters, who would then be able to “spend” them on some other occasion. The system, as you can see, was self-policing: Over time each couple would necessarily give as many hours of baby-sitting as it got.

  Now if you think about it, a system like this requires that there be a fair bit of scrip in circulation. A couple might want to go out several times in close succession, and might find itself unable to take the time (or find the opportunity) to baby-sit and earn more coupons in between. Furthermore, a couple might be uncertain about its schedule, proving a further incentive to keep a reserve of scrip on hand. So on average, for the baby-sitting co-op to work properly, there needed to be quite a few coupons in circulation per couple.

  I don’t want to get into the fairly complicated details of how scrip was issued. Suffice it to say that after a while the co-op got into a situation in which there weren’t enough coupons out there. This had some peculiar effects. Couples became reluctant to go out for the evening—because the typical couple did not have much of a reserve of baby-sitting coupons, and was anxious to keep that reserve for an important occasion. In order to build up their reserves, couples tried to do extra baby-sitting—but one couple’s decision to go out was another’s opportunity to baby-sit, so chances to earn coupons by baby-sitting became hard to find; and so couples became even more reluctant to spend their reserves of coupons by going out. Eventually the co-op consisted largely of couples sitting glumly at home, unwilling to go out until they had more coupons, unable to earn more coupons because nobody else was going out either.

  In short, the baby-sitting co-op had managed to get itself into a recession.

  Most of the members of the co-op were lawyers, so it was hard for the economists in the group to convince them that the problem was essentially monetary. Instead, the initial reaction of the co-op’s officers was to treat the problem as something to be solved by regulation: For example, they tried to enforce a rule that required each couple to go out at least twice a month. Eventually, however, the economists prevailed, and more coupons were distributed.

  The results were astonishing: With larger reserves of scrip, couples became more willing to go out, which made it easier to get opportunities to baby-sit, which made people still more willing to go out, and so on. The GBP—the gross baby-sitting-co-op product, measured in units of babies sat—soared. Of course the co-op then went on to overdo it, issuing too many coupons. That led to new problems, and incipient signs of inflation…

  I warned you that there would be a quiz. So here it is—just one question: How did you feel about my telling you that story?

  If your answer was “Well, that was a cute story, but I don’t see what it can have to do with the U.S. economy,” you get a D. You have failed to understand the usefulness of simplified models in cutting through the complexity of the real world.

  If your answer was “What’s all this about? I want to talk about globalization and the new information economy, and he’s telling me about baby-sitting,” you get an F. Not only don’t you understand the uses of models, but you have fallen into the naive error of supposing that the way to be sophisticated about economics is to use big words and talk about big things.

  For the fact is that by studying our model—the baby-sitting co-op, which is like a miniature version of the real U.S. economy, in which Alan Greenspan controls the supply of coupons—we can gain some very important insights that many people who believe that they are knowledgeable about these things never do seem to grasp. Let me emphasize two insights in particular.

  First, we learn that there is a fundamental difference between the kind of growth associated with an increase in the money supply and the sources of longer-term growth in the economy. When GBP surged after the issuance of new coupons, it wasn’t because the couples in the co-op had installed new high-tech baby-sitting equipment, or because they had been retrained to be effective in the new global baby-sitting economy, or because they had been freed from the burden of government regulations and taxes that frustrated private sector baby-sitting initiative. All that happened was that a failure of coordination due to inadequate liquidity was cured by increasing the money supply—end of story.

  Second, as soon as we think about the baby-sitting story we realize that there are limits to monetary policy. Too little money is a bad thing; but while GBP could be expanded up to a point by printing more coupons, this process could only go so far. Indeed, issuing too many coupons actually hurt the co-op: an excessively expansionary monetary policy is counterproductive.

  Now surely these two insights apply to the full-scale, adult economy as well. Business-cycle recoveries like 1982–1989 or 1992–1994, in which the Fed pumps money into a depressed economy and brings idle capacity back into use, tell us very little about the kind of growth that the economy can achieve on a sustained basis. Anyone who says that the Reagan-era recovery is an indicator of the kind of long-term growth that our economy could achieve if only we would institute a flat tax has simply failed to learn one lesson of the baby-sitting co-op. Anyone who thinks that the Fed can arbitrarily choose a rate of growth as a target, and achieve it indefinitely, has failed to learn the other lesson.

  Can we be more specific? Indeed we can. In the U.S. economy, it is quite easy to separate cyclical fluctuations in output from the long-term growth in the economy’s potential. For the past twenty years it has been quite reliably true that whenever the economy grows faster than 2.5 percent, the unemployment rate falls (about half a point for every extra point of growth), while whenever the economy grows more slowly than 2.5 percent, the unemployment rate rises. This is a pretty solid indication that the economy’s potential output is growing at about 2.5 percent per year. And there is no sign whatsoever in recent data that this underlying rate of growth has accelerated—if anything, it has slipped a bit due to slower labor force growth.

  The limits to expansion are harder to pin down. But we know there must be a limit somewhere—the baby-sitting co-op tells us so. After thirty years of intense debate, and hundreds of statistical studies, most economists have come to agree that inflation will spiral upward if the Fed tries to push unemployment too low. How low is too low? Past experience suggests that the red line—the infamous NAIRU (non-accelerating-inflation rate of unemployment)—is an unemployment rate in the 5.5 to 6 percent range; but as I’ll explain in a second, the precise number is not crucial for the debate with the four-percenters.

  Now, armed with our model, let’s talk about what the Fed’s critics have been saying. In particular, what do we make of it when someone advocates a growth target of 4 percent per year for the next five years?

  Well, bear in mind that the rate of growth of potential output is reliably estimated to be less than 2.5 percent per year, and that every extra point of growth reduces the unemployment rate by half a percentage point. It immediately follows, then, that 4 percent growth for five years would mean targeting an eventual unemployment rate of something like 1.5 percent.

  Now I don’t know anyone who thinks that is a plausible goal. Maybe you think the NAIRU is 5 percent, not 5.5; maybe you even think that it’s 4.5, though that seems grossly inconsistent with the statistical evidence. But 1.5 percent? So how can smart people think that a 4 percent growth target is feasible? Well, I’ve had discussions with pro-growth types myself, and have had reports from other economists who have tried to have such discussions, so here is an outline of how the discussion goes.

  The first thing they say is that the old rules no longer
apply, because we have had a “productivity revolution.” The economy’s productive potential, we are told, is now growing much faster than in the past.

  What’s wrong with this claim? Well, for what it’s worth the numbers produced by the Bureau of Economic Analysis, which estimates productivity, do not show anything that looks like a productivity revolution. To be sure, many businessmen claim that the numbers are wrong (although there are some independent reasons to suspect that productivity growth remains fairly pedestrian). But in any case such claims don’t help the pro-growth argument. The reason is that the same numbers are used to estimate productivity growth and GDP growth. If you think that productivity is really growing at 3 percent, not the 1 percent the BEA reports, then you must also believe that GDP is really growing at 4 percent, not 2—in other words, the Fed is already giving us 4 percent growth, so what’s your problem? (By the way, I am told on unreliable hearsay that economists at the Fed have tried to explain this point to prominent four-percenters, and met a blank wall of incomprehension).

  The second argument that growth advocates usually produce is the claim that globalization—the new openness of the U.S. economy to imports—now prevents any resurgence of inflation. This can sound plausible—but only if you don’t know recent economic history, and don’t think too hard about it. How can anyone think that being an economy open to trade ensures against inflation, when they have the example of Britain to contemplate? In the late 1980s Britain—a nation with a share of imports in GDP almost three times that of the United States—allowed its monetary policy to be guided by wishful thinking about how much the economy could be expanded. The result was right out of the textbook: an explosion of inflation, which was brought under control only by a return to double-digit unemployment rates.

  Moreover, how can you discuss globalization without noticing that the U.S. has a floating exchange rate? If the Fed were to pursue a radically more expansionary monetary policy, one sure consequence would be a lower value of the dollar. If you really think that U.S. prices are basically limited by foreign competition, then you have to believe that a fall in the dollar will translate almost directly into higher inflation. In fact, traditional analyses of inflation in trading economies conclude that expansionary monetary policy has more, not less, effect on inflation in a country with a large import share and a floating exchange rate than it does in a relatively self-sufficient economy.

  So neither productivity growth nor globalization make sense as arguments for looser monetary policy. Maybe there are other arguments—but people like Felix Rohatyn have not produced them. In short, this is not a serious debate: Although the four-percenters command a lot of political and business support, intellectually they have failed to make even the ghost of a case for their views.

  Now you may think that what I am saying is that these guys are dumb—that as Bob Dole might put it, Paul Krugman thinks that he is smarter than Felix Rohatyn. But of course I’m not—after all, if I’m so smart, how come I’m not rich? No, the puzzle is why smart people say foolish things. Why haven’t the four-percenters managed to make a better case? What’s their model?

  The answer, of course, is that they have no model. What’s wrong with the kind of economics that Felix Rohatyn and many others practice is that they have failed to understand the principle. They think that you do economics the way a lawyer prepares a brief for a client—first you decide on your opinion, then you marshall as many plausible arguments as you can in support. And they imagine that the orthodoxies of economics—like the belief that the U.S. economy’s potential growth rate is only 2.5 percent, or that free trade is a good thing—were arrived at in the same way.

  But that’s not how serious economics is done. A real economist starts not with a policy view but with a story about how the world works. That story almost always takes the form of a model—a simplified representation of the world, which helps you cut through the complexities. Once you have a model, you can ask how well it fits the facts; if it fits them reasonably well, you can ask what sorts of magnitudes, what sort of tradeoffs, it implies. Your policy opinions then flow from the model, not the other way around. The reason economists at the Fed think that the economy can’t achieve 4 percent growth is not because they like slow growth, or because they are locked into a mindless orthodoxy: it is because they have a model of the U.S. economy that fits the facts very well and that tells them that 4 percent is a completely unrealistic target. They might be wrong—but to make a credible case for much faster growth you must counter the orthodox model with a better model, or you are engaged in an exercise in rhetoric rather than economics.

  Let me also say something else. Anyone who has ever made the effort to understand a really useful economic model (like the simple models on which economists base their argument for free trade) learns something important: The model is often smarter than you are. What I mean by that is that the act of putting your thoughts together into a coherent model often forces you into conclusions you never intended, forces you to give up fondly held beliefs. The result is that people who have understood even the simplest, most trivial-sounding economic models are often far more sophisticated than people who know thousands of facts and hundreds of anecdotes, who can use plenty of big words, but have no coherent framework to organize their thoughts. If you really understood my story about the baby-sitting co-op, congratulations: You now know more about the nature of monetary policy and the business cycle than 99 percent of the attendees at Renaissance Weekend. If you have taken the time to understand the story about England trading cloth for Portuguese wine that we teach to every freshman in Econ 1, I guarantee you that you know more about the nature of the global economy than the current U.S. Trade Representative (or most of his predecessors).

  I might as well raise another point. One thing that usually happens when I try to talk about the difference between serious economics and the kind of glib rhetoric that passes for sophistication is that people accuse me of being arrogant, of thinking that I know everything. I can’t imagine why. No, seriously—think about it. What someone like Felix Rohatyn is in effect saying is “I don’t need to make an effort to understand where the conventional views of economists come from; I don’t need to understand the stuff that’s in every undergraduate textbook; I’m such a smart guy that I can make up my own version of macroeconomics off the top of my head, and it will be much better than anything they have come up with.” Then along comes this irritating economist who points out a few gaping holes in his argument, basic errors that anyone who had bothered to understand the stuff in the undergraduate textbook would not have made. And people’s response is “That Krugman—he’s so arrogant.”

  Well, what can we do about this kind of thing? Let me be the first to admit that economists have not made it easy for smart people who don’t want to get too deep into the technicalities to understand the basics. Mathematics is a wonderful tool, but there are far too few attempts to explain the fundamental models of economics with a minimum of math; we need to make a real effort to write in English, and skip the differential topology. I’m trying, but the profession has a long way to go.

  But it’s also important for non-economists—people who want to be sophisticated about economic policy without getting Ph.D.s—to make an effort. As I said earlier, it’s not a matter of time, it’s a matter of attitude. The biggest problem with many businesspeople, political leaders, and others is that while they are willing to talk and read about economics ad nauseam, they are not willing to do anything that feels like going back to school. They would rather read five books by David Halberstam than one chapter in an undergraduate textbook; and they absolutely hate the idea that they need to work their way through whimsical stories about cloth and wine and baby-sitting rather than get right into pontificating about globalization and the new economy.

  But there is no way around it. If you want to be truly well-informed about economics (or anything else), you must go back to school—and keep going back, again and again. You must be
prepared to work through little models before you can use the big words—in fact, it is usually a good idea to try to avoid the big words altogether. If you balk at this task—if you think that you are too grown-up for this sort of thing—then you may sound impressive and sophisticated, but you will have no idea what you are talking about.

  A Good Word for Inflation

  Many years ago, Paul Samuelson memorably cautioned against basing economic policy on “shibboleths,” by which he meant slogans that take the place of hard thinking. Strictly speaking, this was an incorrect use of the word: The OED defines a shibboleth as “A catchword or formula adopted by a party or sect, by which their adherents or followers may be discerned, or those not their followers may be excluded.” But in a deeper sense Samuelson probably had it right: Simplistic ideas in economics often become badges of identity for groups of like-minded people, who repeat certain phrases to each other, and eventually mistake repetition for self-evident truth.

  Excerpted from “Fast Growth and Stable Prices: Just Say No,” Economist, August 1996.

  Public discussion of monetary policy is increasingly dominated by two such sects. The shibboleth of one sect is “growth” that of the other is “stable prices.” Those who belong to neither sect find it hard to get a hearing; indeed, journalists and politicians often seem baffled by economists who do not fit into these categories. Surely you must believe either that central banks should aim for zero inflation to the exclusion of all other goals (and that stable prices will bring huge economic benefits) or that central banks should stop worrying about inflation altogether and go for growth (and that by so doing they can bring back the growth rates of the 1960s).