Single post

The Sanders plan: interesting Keynesian questions remain

While we were away and then battling illness, New York Times columnist Justin Wolfers  called the debate on the effects of Sanders’s economic plans, which include expenditures on health care and other programs, along with efforts to redistribute income from rich to poor. Wolfers quotes a study by Christina and David Romer, who are not likely to be making a mathematical error. However, from the point of view of economists sympathetic with at least some of Sanders’s plans, there is plenty of room to further adjust the math proposed in the new study to be consistent with macroeconomic assumptions. In this an a future post, I will explain roughly in what directions I would have to modify the assumptions used by Friedman. The question is, are growth and output numbers beyond the range of recent experience achievable with an expansion of the stimulus provided by fiscal policy at the federal level? Or would limits on low-inflation growth be reached?

The conjectured mathematical error are now thought to involve an erroneous assumption of a permanent effect on economic growth. I say conjectured because Romer and Romer are unsure how to account for the error and guesses that source of it. Assuming a fiscal multiplier of between 1 and 1.6, Romer and Romer infer that the effects of an increase in spending in a certain raises output forever. In other words, a one-time increase in spending this year, followed by a decrease next year is apparently assumed to lead to a lasting increase in growth and output, accounting for oddly high projections in Friedman’s study. Also, the change in GDP induced by Sanders’s redistributive measure, which apparently is a one-time change, are assumed to be followed by similar increases in GDP throughout the period under study—about 10 years. Romer and Romer see the impacts of the stimulus measures and the redistributive measures as likely the result of similar mathematical errors and Wolfers reports that Friedman does not deny making this assumption–a purely mathematical issue, not a fundamental one about the potential of the economy for growth.

The error in accounting for the effects of ten years of stimulus result in a ten-percent increase in GDP in Friedman’s study by the end of the ten years—which is a bottom line Romer and Romer find implausible in any event. In principal, it is interesting to ask this question, in connection with Sanders’s plan or some other possible Keynesian plan. We learn however that the hypothesized increase in each of the ten years of temporary stimulus is 1.45 percent of GDP. Certainly, this amount alone would not bring the growth rates cited by Paul Krugman and others as unrealistically large. This would be a very modest amount of stimulus in an economy in which broad measures of unemployment are nearly 10 percent. Hypothetically, could such a thing be done? Clearly, the thought behind such a small increase is that the economy is alleged to be near full capacity, this being the end of an economic boom, however weak and slow–much in line with the thinking of the Fed and mainstream economists at this point. What about a different economic view that did not take standard assumptions for the sake of argument? Would the big growth numbers at certain points in past U.S. history cited by J. Galbraith in his letter for example be attainable with the right plan, perhaps including Sanders’s health care plan and other ideas? This question is rarely considered. The Romer and Romer study addresses the planned stimulus, redistibutional efforts, and the health care plan in combination, helping to account for the enormity of the effects bandied about in the debate, though Friedman’s study does not now appear to account for the multiplier effects foreseen as there as impacts of Sanders’s plan.

The numbers are still in question on a number of grounds:

1) Friedman’s choice of a multiplier of no more than one, which the Romers (and I) find somewhat conservative. Would multiplier effects exist at all in his view? A hypothetical increase of 100 billion in spending in 2017 would in my view lead to some additional rounds of increase in the initial and subsequent years. In other words, applying Friendman’s multiplier implies that newly hired workers would be drawn from people who were already employed, without the original employers rehiring. The assumption would appear to rest on a “devil’s advocate” argument. It is hard to imagine a reduction in household spending being the result of a big injection of new money into the private sector! So, the multiplier assumed up to this point leaves something to be desired. The multiplier of course depends on the use of the stimulus, and according to Keynesian estimates, some types of spending have multipliers commonly thought to exceed 2, for example. For example, an assumption of a marginal propensity to consume of 66.7 percent after taxes results in a multiplier of 1/(1 ‒ .667) = 3. It is hard to imagine for example how a food-stamp increase would have such a low multiplier effect. Tax cuts in the lower income brackets also can have higher multipliers than one.
2) The deficit has plunged from more than 10 percent of GDP at the height of the recession to about 2.6 percent now (according to January CBO numbers). Unemployment is hovering around 5 percent and just under 10 percent when near-unemployed categories are added. The New York Times reports over the weekend on recent signs that the labor force is rapidly increasing in size after a sudden tumble that was clearly at least related to the Great Recession.
As some analysts have anticipated, the return of discouraged workers is offsetting rising demand pressures in the labor market. Separately, the New York Times reports that some increased employment is reported in the health care industry, where the affordable health care act appears to have led to rising demand in the health care professions. Room would seem to exist for multiplier effects without a big run-up against capacity limits and inflation, given a sufficient fiscal stimulus of some kind.

In principal, could the required growth rates be achieved through an increase in the expansiveness of fiscal policy, and how would that happen? To be continued.


theme by teslathemes