Bruce Bartlett: How Supply-Side Economics Trickled Down

Bruce Bartlett says we're all supply-siders now, but as explained in some detail below, I don't fully agree:

How Supply-Side Economics Trickled Down, by Bruce Bartlett, Commentary, NY Times: As one who was present at the creation of “supply-side economics” back in the 1970s, I think it is long past time that the phrase be put to rest. It did its job, creating a new consensus among economists on how to look at the national economy. But today it has become a frequently misleading and meaningless buzzword that gets in the way of good economic policy.

Today, supply-side economics has become associated with an obsession for cutting taxes under any and all circumstances. No longer do its advocates in Congress and elsewhere confine themselves to cutting marginal tax rates — the tax on each additional dollar earned — as the original supply-siders did. Rather, they support even the most gimmicky, economically dubious tax cuts with the same intensity.

The original supply-siders suggested that some tax cuts, under very special circumstances, might actually raise federal revenues. For example, cutting the capital gains tax rate might induce an unlocking effect that would cause more gains to be realized, thus causing more taxes to be paid on such gains even at a lower rate.

But today it is common to hear tax cutters claim, implausibly, that all tax cuts raise revenue. Last year, President Bush said, “You cut taxes and the tax revenues increase.” Senator John McCain told National Review magazine last month that “tax cuts, starting with Kennedy, as we all know, increase revenues.” Last week, Steve Forbes endorsed Rudolph Giuliani for the White House, saying, “He’s seen the results of supply-side economics firsthand — higher revenues from lower taxes.”

This is a simplification of what supply-side economics was all about, and it threatens to undermine the enormous gains that have been made in economic theory and policy over the last 30 years. Perhaps the best way of preventing that from happening is to kill the phrase “supply-side economics” and give it a decent burial.

It’s important to remember that at the time supply-side economics came into being, Keynesian economics dominated macroeconomic thinking and economic policy in Washington. Among the beliefs held by the Keynesians of that era were these: budget deficits stimulate economic growth; the means by which the government raises revenue is essentially irrelevant economically; government spending and tax cuts affect the economy in exactly the same way through their impact on aggregate spending; personal savings is bad for economic growth; monetary policy is impotent; and inflation is caused by low unemployment, among other things.

These beliefs led to many bad economic policies. In particular, they lay at the root of stagflation, that awful combination of high inflation and slow growth that bedeviled policy makers in the 1970s. Based on insights derived from the Nobel-winning economists Robert Mundell, Milton Friedman, James Buchanan and Friedrich Hayek, the supply-siders developed a new program based on tight money to stop inflation and cuts in marginal tax rates to stimulate growth.

As the staff economist for Representative Jack Kemp, a Republican of New York, I helped devise the tax plan he co-sponsored with Senator William Roth, a Delaware Republican. Kemp-Roth was intended to bring down the top statutory federal income tax rate to 50 percent from 70 percent and the bottom rate to 10 percent from 14 percent. We modeled this proposal on the Kennedy-Johnson tax cut of 1964, which lowered the top rate to 70 percent from 91 percent and the bottom rate to 14 percent from 20 percent.

We believed that our tax plan would stimulate the economy to such a degree that the federal government would not lose $1 of revenue for every $1 of tax cut. Studies of the 1964 tax cut showed that about a third of it was recouped, and we expected similar results. Thus, contrary to common belief, neither Jack Kemp nor William Roth nor Ronald Reagan ever said that there would be no revenue loss associated with an across-the-board cut in tax rates. We just thought it wouldn’t lose as much revenue as predicted by the standard revenue forecasting models, which were based on Keynesian principles.

Furthermore, our belief that we might get back a third of the revenue loss was always a long-run proposition. Even the most rabid supply-sider knew we would lose $1 of revenue for $1 of tax cut in the short term, because it took time for incentives to work and for people to change their behavior. ...

Moreover, we were adamant that only permanent cuts in marginal tax rates would stimulate the economy. We thought that temporary tax cuts, tax rebates, tax credits and such were economically worthless, and we strongly opposed them.

Today, hardly any economist believes what the Keynesians believed in the 1970s and most accept the basic ideas of supply-side economics — that incentives matter, that high tax rates are bad for growth, and that inflation is fundamentally a monetary phenomenon. Consequently, there is no longer any meaningful difference between supply-side economics and mainstream economics.

There is no question in my mind that we never could have overcome the stagflation of the 1970s as quickly or with as little pain as we did without the supply-side idea. But supply-side economics has done its job, just as Keynesian economics did in the 1930s. Those who campaign as its champions are fighting a fight long won — and it is time for supply-side rhetoric to go, with its essential truths embodied in mainstream economics and its perversions discarded for good.

As noted above, I don't fully agree, so let me cast this debate in a different framework where it's easier for me to highlight where we differ.

Let's start with the following fairly standard picture of the evolution of GDP. The red line shows actual output cycling over time, and the blue line shows that natural rate of output which also varies over time:

As depicted, the natural rate is subject to both permanent and temporary supply shocks causing growth to be uneven, but generally upward, and actual output is driven away from the natural rate by demand shocks. That is, the variation in the blue line is from supply-shocks, and the deviation of the red line from the blue line is from demand shocks.

In general, there are two types of policies to consider. The first is the use of monetary and fiscal policy to stabilize the economy. The goal here is to use changes in the money supply, government spending, and taxes to manage aggregate demand and minimize the deviations of actual output from the natural rate of output. This is shown by the dotted red line in the following diagram which is closer, on average to the natural rate than the no-policy outcome shown as the solid red line:

The second type of policy is growth policy. This is what many people mean when they use the term supply-side policy. The goal here is to increase the growth rate of output. This is shown by the upward rotation of the trajectory for the natural rate in the following diagram:

The natural rate of output is determined by the growth of technology, the growth of the capital stock, and the growth of the labor force so policies to increase the growth rate of output are directed at these factors. Examples of supply-side policies are (not all have proven to be equally effective, or effective at all in some cases, and this is far from exhaustive) tax breaks for research and development (to increase technology), tax breaks for IRAs (to increase saving, investment, and the capital stock), tax cuts on capital gains and dividend  (to make capital markets more efficient and increase the capital stock), spending on education (to make labor more productive), reductions in marginal tax rates (so people will increase work effort), accelerated depreciation (to make investment cheaper and increase he capital stock), and reductions in estate taxes (so people will work harder to leave more for their heirs).

Note also that it is the upward rotation in the supply-curve that generates the increase in taxes from supply side policies that you often hear about. The question, of course, is how much additional growth comes from a cut in taxes and here I agree to some extent with Bruce Bartlett. It depends upon the type of tax cuts that are enacted, some are more productive than others and hence some types of tax cuts generate more tax revenue than others. Whether the tax cut is permanent or temporary is also important.

We'd disagree over the magnitude however. While some types of tax cuts can affect growth, the effect is nowhere near large enough to generate a 33% tax revenue recovery rate, not even close, and, in any case, all the low-hanging fruit has already been plucked, something that is often overlooked.

That is not the end of our disagreement. Bartlett does not distinguish between various classes of models, between the short-run and long-run, or between stabilization and growth policy all of which are important distinctions so let me touch upon these issues.

There are two predominant views of the source of fluctuations in output, Real Business Cycle models and New Keynesian models.

Real Business Cycle (RBC) theorists believe that most if not all fluctuations in the economy are due to supply side shocks, aggregate demand shocks such as changes in the money supply, changes in taxes, and changes in government spending affect nominal variables such as prices but have little to do with changes in output over time (however,  government intervention does causes inefficiencies in these models so that less intervention is generally preferred to more). Thus, for RBC advocates, the red and blue lines lie nearly on top on one another because nearly all of the movement in output is due to supply shocks.

Obviously, then, in these models demand management - monetary and fiscal policy - can stabilize prices (and hence increase efficiency), but demand management has little effect on output.

Thus, since short-run demand management is ineffective in these models (and often counterproductive), all that's left is long-run growth policy and that's why people such as Bruce Bartlett, who have an RBC model in mind when thinking about policy, tend to focus on long-run, supply-side, growth enhancing policies.

Let me turn next to Keynesians. My focus is on the New Keynesian (NK) school, but I should note that I don't agree with all of the characteristics Bartlett associates with Keynesians of the 1960s and 1970s, and I certainly don't agree with the claim made in the next paragraph that after the policy failures of the 1970s "the supply-siders developed a new program based on tight money to stop inflation and cuts in marginal tax rates to stimulate growth." A standard expectations augmented Phillips curve story does a much better job of explaining these events, and the interest rate targeting rules used from the early 1980s onward are not based upon RBC models. Most RBC models don't even include money as it plays not role in either the short-run or long-run.

New Keynesians (NK) do not deny that shocks to aggregate supply can affect GDP nor that supply shocks can be large and important. However, New Keynesians also believe that aggregate demand shocks are important, i.e. that the difference between the blue and red lines is large.

New Keynesians attempt to stabilize actual output around the natural rate as shown above. Why does NK policy tend to focus on demand shocks rather than supply shocks? The answer is that although it would be ideal if we could use supply-side polices to smooth short-run fluctuations in output arising from supply shocks, the reality is that we cannot do this. As Bartlett notes, supply-side polices are very blunt, slow-acting policies that can affect output in the long-run, but they are all but useless in dealing with short-run fluctuations in the economy (thus, RBC theorists tend to focus mainly long-run growth).

Since supply cannot be managed in the short-run, that leaves demand management policies, i.e. monetary and fiscal policy. As we learned in the 1970s, demand side tools are not very effective instruments for offsetting supply-side shocks - trying to use demand side policy to offset supply shocks helped to generate the stagflation we saw at the time. We've learned since then, but practically we are still somewhat powerless to offset supply side shocks in the short-run - all we can do is manage demand to match changes in supply. That is, if a hurricane wipes out supply, we can use policy to reduce demand to match, but we can't do much to increase supply back to its initial level in the short-run.

What we can do much more effectively, if you believe in NK models, is stabilize demand shocks through demand management policy. These policies are relatively simple conceptually, the trick is to manage demand so that upward and downward demand shocks are offset by appropriate changes in policy, but that is easier said than done. Still, we do appear to be able to reduce variation over time through both monetary policy in particular, and also through fiscal policy (e.g. through automatic stabilizers).

The claim made by Bruce Bartlett that "there is no longer any meaningful difference between supply-side economics and mainstream economics" is not something I can agree with. There are big differences between the RBC and NK schools and big differences in the implications of the two schools for policy in the short-run. RBC advocates do not believe in short-run stabilization policy, their focus is solely on maximizing long-run output growth. NK theorists do not deny that robust economic growth is important, but they also believe that government can play a helpful role is smoothing short-run economic fluctuations.

Why do Republicans tend to endorse the RBC framework? I believe in many cases that belief in the RBC model arises from an honest view that the evidence is most supportive of this class of models. But in other cases I believe it is an ideological marriage. The RBC model has two features that make it attractive.

First, because it says short-run stabilization policy is ineffective, and that government intervention through either spending or taxes generates economic distortions, the RBC framework supports an approach where the role of government in the economy is minimized.

Second, because the RBC framework allows for tax cuts to produce higher growth by reducing inefficiencies, and because it is then possible to argue that tax revenues might increase, it gives two reasons for supporting tax cuts - higher growth and less than a full loss of tax revenue, i.e. a dollar tax cut does not cost a dollar (or, for serious ideologues, the tax-cuts even pay for themselves).

The NK model, on the other hand, supports active government intervention which is at odds with this ideology. In addition, because the focus in NK models is on stabilization of output around the natural rate, not on growth of the natural rate, tax-cuts do not have the dynamic long-run effects as in RBC models (though these can be added) and hence there is not as much ideological support for tax cuts in the NK framework.

This is much too long already, but a few more things. We don't we know which type of shock is most important? If demand shocks play a substantial role, we should pay attention to the NK policy prescriptions, but if aggregate supply shocks are the primary force behind business cycles, we should abandon short-run stabilization and focus solely on long-run growth. We don't we just look at the empirical evidence and figure this out?

The problem, essentially, is that we only have one time-series, GDP, and we want two things from it, supply shocks and demand shocks. Since we only have one piece of information and want two things from it, we must make an assumption of some sort. Under some assumptions, supply shocks appear predominant, but under others, demand shocks are the most important factor in business cycles. Because we have no way of knowing for sure which assumption is best, and because the econometric evidence changes as the assumptions change, we are left with uncertainty as to which type of shock matters most and hence which model we ought to prefer.

There is much more to say about all of this, I haven't even mentioned New Classical models, but that will have to do for now. Summarizing, contrary to what is implied in Bruce Bartlett's commentary, there are two distinct schools in economics, the RBC school and the NK school, and they have very different policy implications. Not everyone will agree with this, and that is the point I suppose, but I would argue that the mainstream view today is the NK model, though the RBC school has strong advocates and has made important contributions to our thinking (the long-run incentives Bruce Bartlett mentions are a good example).

So, here's where we agree. Both NK and RBC advocates see the long-run similarly. Both schools agree that demand side polices have little effect on long-run growth. Both agree that incentives matter, and that we should, of course, strive to enhance efficiency and long-run growth whenever possible. There is a difference in the two schools as to the strength of those incentives, but if that is all that is meant by supply-side polices, then fine, no problem, we're in agreement.

But there is a big disagreement over the short-run. RBC adherents take a hands-off, free market approach. Their model tells them that government interference causes inefficiencies, and that there is nothing to be gained in return in terms of enhanced stability. NK adherents believe government should take an active role in stabilizing the economy and that is something that, contrary to what is implied above, has not changed since the 1960s and 1970s. The model used by the NK school is very different from the models we used then - and our approach to policy is similarly different - but the basic idea that government intervention can help to stabilize output and employment in the short-run is unaltered.

Update: See comments by Bruce Bartlett and Paul Krugman.

Update: More at Angry Bear, Political Animal, Ezra Klein, Angry Bear again, Division of Labour, and Tim Worstall. Update: Brad DeLong too. . Update: One more from Angry Bear. Update: Lawrance Lux comments. Update: Angry Bear with more. Update: Talk Left also.

Update (4/11): Follow-up here.