Thomas Kuhn is most famous for two concepts: revolutions and paradigms. Looking at the history of science, Kuhn argued that it does not show a steady accumulation of knowledge, but shows long periods of relative conceptual stasis (what he called “normal science”) punctuated by “revolutions” where the conceptual foundations of a particular scientific field are overturned and replaced. During periods of normal science, scientists generally see the world and interpret observations in the same way, but, according to Kuhn, during revolutionary periods competing scientific camps see the world differently. In some sense, they inhabit different worlds.
Macroeconomics has gone through multiple “revolutions” in the twentieth century, most notably as economists largely came to consensus in the 1920s and 30s over John Maynard Keynes’s view of the economy after the Great Depression, then largely switched to Friedman’s monetarist view in the 1960s.1 Currently macroeconomists appear to be divided between the two camps: neither Keynesian nor monetarist macroeconomics is clearly dominant. Given Kuhn’s account of paradigms, this suggests that Keynesians and monetarists will see the world in different ways. This seems particularly plausible in the extremely complex domain of economics. However, I think the extent of this divide is easily overstated, for macroeconomics in particular and science in general. This post looks at an example where the disagreement between Keynesians and monetarists seems to be more rhetorical than conceptual, more about emphasis than perception.
Matthew Yglesias is a correspondent for Slate and has previously worked for such liberal bastions as ThinkProgress and The Atlantic. He recently posted about two talks he attended at the American Economics Associate annual meeting that both found evidence of “big fiscal policy multipliers“. Fiscal multiplies measure the change in GDP that will result from each dollar increase or decrease in government spending or taxation. For instance, if there is a fiscal multiplier of 2, then for each $1 of government spending the GDP will increase by $2. The higher fiscal multipliers are, the more effective policies such as fiscal stimulus will be in boosting the economy, and the more damaging cuts will be. Therefore when Yglesias posts about “big” fiscal policy multipliers, he is making an argument for government intervention in the economy through fiscal stimulus.
Yglesias cited two papers in his post.2 The first, by Daniel Shoag, found a fiscal multiplier of 1.43, while the second, by Blanchard and Leigh, suggested a fiscal multiplier between 0.9 and 1.7. Do these authors consider these numbers to be “big”? Do these papers support stimulus spending? On first look it seems so. A report authored by Blanchard and based on the research in the above paper argues:
If the multipliers underlying the growth forecasts were about 0.5, as this informal evidence suggests, our results indicate that multipliers have actually been in the 0.9 to 1.7 range since the Great Recession. This finding is consistent with research suggesting that in today’s environment of substantial economic slack, monetary policy constrained by the zero lower bound, and synchronized fiscal adjustment across numerous economies, multipliers may be well above 1.3
And Shoag’s paper concludes:
This study, like much of the literature cited above, finds that local, windfall-financed government spending has large employment and income effects…. While these results should be interpreted cautiously, the mounting evidence from a number of different studies on local, windfall multipliers suggests a growing consensus on this issue, both generally and post-2008. 4
Shoag suggests that his multiplier of 1.43 indicates “large employment and income effects” while Blanchard suggests that his range of 0.9 – 1.7 is “well above 1″.
But now look at what Garret Jones, an economist at the conservative George Mason University and blogger for the Library of Economics and Liberty has to say on this:
My view: 1.5 and 1.4 are not big multipliers. If a dollar of government spending or tax cuts genuinely boosted the short run economy by two or three dollars, I’d call that big: At that point, even if the government spending was purely wasteful, you’d be substantially growing the private sector in a big, obvious way. A prominent freshman economics textbook by Case/Fair/Oster says that “in reality, the multiplier is about 2,” but reality has been disagreeing with that assessment lately.
A multiplier of 1.5 or 1.4 is unimpressive as a grand argument for stimulus.5
Jones isn’t disputing the findings of either of these papers. Rather, he is disputing what they suggest about economic policy and how they effect economists’ beliefs about fiscal multipliers. He is arguing about what counts as a “big” multiplier. While the previous papers and Yglesias’s post suggested that these results were surprisingly large, Jones suggests they are surprisingly small. It isn’t that they are living in different worlds, or interpreting observations differently. Jones isn’t seeing ducks while Yglesias is seeing rabbits. Instead, they are using these results to different rhetorical purposes and juxtaposing them with different previous findings. It isn’t so much psychology as strategy.
- See: Harry G. Johnson, “The Keynesian Revolution and the Monetarist Counter-Revolution”, The American Economic Review, Vol. 61, No. 2 (May 1971) ↩
- Daniel Shoag, “Using State Pension Shocks to Estimate Fiscal Multipliers since the Great Depression“; Olivier Blanchard and Daneil Leigh “Growth Forecast Errors and Fiscal Multipliers“ International Monetary Fund. ↩
- International Monetary Fund “World Economic Outlook October 2012: Coping with High Debt and Sluggish Growth“, p. 43 ↩
- Shoag, p. 5 ↩
- Garrett Jones “When is a Spending Multiplier ‘Big’?” ↩
How does this example differ from some of the examples that Kuhn draws from natural science? Consider an example straight from the "X-rated Chapter X" of "Structure," where Kuhn did his best to show that scientists in different paradigms "live in different worlds." The example is the re-interpretati...
How does this example differ from some of the examples that Kuhn draws from natural science? Consider an example straight from the “X-rated Chapter X” of “Structure,” where Kuhn did his best to show that scientists in different paradigms “live in different worlds.” The example is the re-interpretation by John Dalton of data about the ratios in which different substances combine in different reactions. This sort of data had been collected well before Dalton suggested that substances combine in fixed proportions by weight in chemical reactions. As Kuhn tells the story, Dalton took this data, tweaked it a bit, and drew attention to apparent whole-number ratios in that data. There were plenty of cases where the data did not quite fit his theory, but he finessed this data and focused on the cases that worked.
Could this be described as “using [existing] results to different rhetorical purposes and juxtaposing them with different previous findings”?
Another thing that is present in the chemistry case is a difference in the definitions of key terms that results in a kind of circularity. To simplify, whenever Dalton came across a reaction in which the reactants did not combine in fixed proportions, he denied that it was a “chemical” reaction. This ran roughshod over his rivals’ definition of a “chemical” reaction, but from Dalton’s point of view it was perfectly coherent.
Is there anything similar going on in the economics case? I might be wrong, but it looks like there is a kind of circularity in the following argument from Jones:
“If a dollar of government spending or tax cuts genuinely boosted the short run economy by two or three dollars, I’d call that big: At that point, even if the government spending was purely wasteful, you’d be substantially growing the private sector in a big, obvious way.”
On one reading of this quote, Jones is skeptical about the value of the dollars that the government initially injects into the economy, so that in order to “break even” on government spending it is necessary to have a multiplier of at least 2. But Jones’ opponents are unlikely to share his skepticism about the initial value of government dollars, so they are unlikely to require such a high multiplier.
(I’m not an economist, so what I’ve just written might be jibberish, but I don’t know how else to interpret Jones’ argument).
What other sorts of argument do economists adduce when trying to decide what counts as a “big” multiplier? Or is it fairly arbitrary, like the decision to take 0.05 as a threshold of statistical significance in experimental trials?
To "break even" you need a multiplier of at least 1, though even for multipliers of less than one there can apparently be a short term boost in GDP. Jones is saying that for stimulus to be a good idea there would have to be a multiplier of at least 2. This is because he thinks there are other proble...
To “break even” you need a multiplier of at least 1, though even for multipliers of less than one there can apparently be a short term boost in GDP. Jones is saying that for stimulus to be a good idea there would have to be a multiplier of at least 2. This is because he thinks there are other problems caused by government spending—he doesn’t think that *just* boosting GDP is reason enough to justify spending.
Where I see my disagreement with Kuhn is, crudely, that Kuhn thinks his scientists believe their own rhetoric—they *see* the world differently—where I suspect the explanation is much simpler: they see the world in essentially the same way, but (out of self-interest or due to differing values) choose to describe it differently. People are of course capable of self-deception, and probably reality lies somewhere in between the extremes between Kuhn and my position, but I think it lies much closer to the same vision, different rhetoric pole than the different worlds pole.
I tend to view incommensurability of perspective as an analytical tactic of last resort. There exist a number of different approaches to economics, which we might productively regard as different paradigms. These would include the physics-like calculus of aggregates one gets with Marshall or Keyne...
I tend to view incommensurability of perspective as an analytical tactic of last resort. There exist a number of different approaches to economics, which we might productively regard as different paradigms. These would include the physics-like calculus of aggregates one gets with Marshall or Keynes; and then there are two more combinatorial-type approaches, one of which is “structural” economics, which tends to study things like the allocation of capital between sectors; a second is “micro-foundational” economics, which grounds aggregate phenomena in choices made by typical producers and consumers (“agents” in the jargon). In principle, all these approaches are supposed to be compatible with each other, but in practice, as I understand it, they don’t fit together very cleanly.
Now, “Keynesian” and “monetarist” approaches need not fall across any sort of paradigmatic divide, as the divide mainly has to do with policy preferences. Friedman’s key monetarist insight is that fiscal policy tends to be offset by expectations that taxes, now or later, will offset whatever growth government spending creates. Thus, government spending is merely a “redistribution” of an otherwise fixed money supply; since a central planner cannot do this more efficiently than a market, redistribution will necessarily be destructive to the economy (this is the preoccupation of another crew, the so-called Austrian School followers of von Mises and Hayek—this stuff is huge in libertarian circles).
Now, generally speaking, modern “Keynesians” (Paul Krugman and Larry Summers are good examples) actually acknowledge the monetarist insight, except in depressed economies (when monetary policy is ineffective), when resources are otherwise idle and can be liberated through fiscal policy. A conservative monetarist, however, will assume that expectations of future taxes still applies. The Keynesian responds that any expectations should be discounted, particularly since expectations of continued reduced demand will dominate business investment decisions. This may represent a clash of worldviews, but it is not a clash of paradigms, and is, in principle, resolvable through empirical observation. (In reality, the issue is highly ideological.)
This, I think also explains the multiplier dispute. Something like 1.5 would seem high to a Keynesian, because it is higher than the lack of multiplier, which results from money sitting idle; it would seem low to a monetarist who could only countenance fiscal policy if it could exceed whatever multiplier can be achieved by funnelling that same money through the market—so, a monetarist would be apt to posit an unachievably high multiplier as their threshold for accepting fiscal policy as sound policy.
But it’s not quite so simple as that! In fact, Friedman’s insight is often cited as resulting from the micro-foundational approach, that only by considering the choices of the individual consumer will their expectations of future taxes become visible. But the real doyen of micro-foundations is not Friedman but, as I understand it, Robert Lucas, which leads us to “real business cycle theory”.
Real business cycle theory holds that business cycles do not result from shortfalls in demand, but, rather, are a rational market response to misallocations of resources in an economy; unemployment will exist until resources are reallocated into full efficiency. Misallocations can result from changing circumstances (say, other countries becoming more competitive), but conservatives are apt to view them as resulting from government intervention in natural market functions (say, through protectionism, or “picking winners”).
Now, this actually does look something like Kuhnian incommensurability, as a real business cycle theorist will look at high unemployment and call it “structural”, where a Keynesian will look at that same high unemployment and call it a shortfall in demand. Ostensibly, as Paul Krugman will tell you, this issue can also be settled empirically by looking to see how unemployment is distributed throughout the economy. But, again, economists are unlikely to agree, and this makes a difference, since fiscal policy, to the real business cycle theorist, will interfere in the natural process of reallocation of resources (an activity which would have a higher multiplier), where a Keynesian would regard it as employing idle resources (idleness having a lower multiplier than fiscal policy).
I should say I’m not an economic expert myself, and it’s likely I’ve used bad terminology, or have missed nuances, but that’s my understanding of the situation. It’s also an issue which could use systematic scrutiny from historians of economics.