Scott Sumner  

Don't apply AD to long run growth/distribution issues

Bio of Elinor Ostrom... A Waste of Paper?...

In the 1990s there was (or seemed to be) a consensus that you do not want to apply aggregate demand arguments to questions of long run growth and income distribution. Demand shocks have no long run real effects. In simple terms, you can juice an economy in the short run by printing money, but it won't affect long term growth, or the long run distribution of income.

As with any model this isn't quite right. One can imagine AD policies that are so inept that they even damage the long run growth of an economy, perhaps by slowing technological change, or by leading to poor institutional outcomes (think Nazi Germany). Nonetheless, the natural rate model is a reasonable approximation for most plausible public policies outside extreme monetary contraction.

Thus in the 1990s it was widely believed that a higher minimum wage would raise the natural rate of unemployment. More recent studies suggest that modest increases (say a dollar or two) might not have a significant effect on employment. My hunch is that these studies underestimate the long run impact, but I concede it's an open question.

Tim Worstall has a good post criticizing a recent study of the new $15 minimum wage in Los Angeles (which is not fully implemented until 2019.) He criticized the following claim, made in page 25 of the report:

Net spending increases because lower-wage workers spend higher proportions of their income, generating a greater multiplier than for households who would absorb the higher prices,
I actually have two problems with this sort of argument. First, it confuses consumption and aggregate demand. Aggregate demand is actually C + I + G + NX. And since S equals I, there is no obvious reason why more saving would reduce AD.

Some Keynesians will respond that a higher propensity to save will depress interest rates, which will depress velocity, and if M is held constant this will depress NGDP. They don't actually put it that way, but that's essentially what they are claiming, if translated into plain English.

The problem here is obvious---the Fed doesn't target M, it targets AD. So if Los Angeles residents did save more, the Fed will offset any contractionary effects with easier money. That's why real GDP growth sped up in 2013. And there's no reason to assume that AD in 2019 will be any lower than the Fed wishes it to be.

Again, all this was well known in the 1990s but seems to have recently been forgotten by a large segment of the economics profession. Interestingly, a similar mass amnesia occurred after the Great Depression, which also featured a long period of zero interest rates. If Paul Romer wants to spend his time exposing folly, a good place to start is looking at how the neoliberalism of the 1990s regressed to the vulgar Keynesianism of the 2010s.

But let's say I'm completely wrong, and that the minimum wage increase does boost AD. Is this a good argument for raising the minimum wage? No, because demand shocks have only temporary effects and the minimum wage is best viewed as a long run policy. There is no indication that the LA city council plans to repeal the law in the near future.

This is just one example of something I see all the time, AD arguments misapplied to long run structural economic problems. For instance there was a lot of debate about whether QE helps or hurts the poor. In fact, the impact of a particular monetary policy gesture is irrelevant, as any monetary regime is essentially neutral in the long run. Suppose monetary policy regime X will help the poor when policy is more expansionary than average, and vice versa. Since (by definition) policy is more expansionary than average about as often as it is less expansionary than average, the distributional effects will net out to roughly nothing. Even worse, people usually don't get even the short run effects right. For instance, in the short run QE probably helps the rich and poor more than the middle class, but so what?

Macro is simple: Use monetary policy to stabilize the path of NGDP, and use radical deregulation, privatization and tax reform to improve the supply side of the economy. Have a progressive consumption tax regime and pollution taxes, and also low wage subsidies to improve distribution. That's not far from the 1990s consensus, and it was correct. It's a pity that people on both the left and the right seem to have forgotten why the Clinton years were so good.

PS. The paper's authors were Michael Reich, Ken Jacobs, Annette Bernhardt and Ian Perry, three of whom are at Berkeley. Don't consider this post a personal criticism; when doing reports for governmental units one may be forced to look at a lot of factors, not all of which are important.

Comments and Sharing

COMMENTS (17 to date)
Alex Godofsky writes:
Some Keynesians will respond that a higher propensity to save will depress interest rates, which will depress velocity, and if M is held constant this will depress NGDP. They don't actually put it that way, but that's essentially what they are claiming, if translated into plain English.

That's some "plain English", Scott :P

Econymous writes:

How does this work with LA being local and Fed policy being nationwide? I know next to nothing about trade models, but I could see how, if "production in LA" were your outcome variable of interest, money going to people who never leave the city would boost that outcome variable more than money going to somebody who will put it in a Swiss bank account.

But I really ask because I don't know (and highly doubt my intuition on this topic)--how does this work?

D. F. Linton writes:

"Demand shocks have no long run real effects. In simply terms, you can juice an economy in the short run by printing money, but it won't affect long term growth, or the long run distribution of income."

Really? So if there are no long term effects on growth from short term "juice", why in your mind can't long term "juice" be obtained by a continuous sequence of short term effects?

You can fill your belly with the seed corn or mortgage your future for a new X-Box, but increased present consumption must have a future cost.

If not then we have all been fools, working, saving and investing when all that was needed was more money printing and infinite government deficits.

Dustin writes:

But does the Fed actually target AD in practice?

E. Harding writes:

@ Dustin
-It used to. Nobody knows what the Fed targets anymore, even after they declared their target in 2012. Probably not even the Fed itself.

E. Harding writes:

The mystery is how the low interest rates would depress velocity.

ThomasH writes:

Basic agreement but (oddly for you) you do not seem to recognize the damage done by seven years and counting that the Fed has let NDGP remain below it's pre-crisis trend. I think slow real growth since 2008 is mainly the fault of continuing tight money.

I'm especially surprised at your saying "the Fed doesn't target M, it targets AD." Well it is supposed to, but does it? So long a it is within it's self imposed inflation rate ceiling, it is not clear what rule the Fed is following.

Scott Sumner writes:

Alex, Consider the alternative. :)

Econymous, Good question. That's hard to say, and since it doesn't matter in any case, I decided not to spend time trying to figure it out.

DF. The economy goes back to the natural rate in the long run.

Dustin, AD, or something similar.

Thomas, I don't agree, I think it's mostly due to The Great Stagnation. But I concede that some of it is due to tight money. For instance, some portion of the increase in the disability rolls is due to tight money, and the capital stock is a bit smaller than otherwise.

Kevin Erdmann writes:

The collapse of housing capital, equal to about a year's worth of GDP, could have been largely avoided with Fed support.

Andrew_FL writes:

The High Wage Doctrine has risen from its grave to menace us once again.

The minimum wage is the mother of all sticky wages. In some sense, quite literally.

Maximum Liberty writes:
any monetary regime is essentially neutral in the long run

It's not quite the context that you meant, but monetary regimes are not neutral as between the financial and non-financial sectors. The presence of a central bank with the power to create money is associated over long periods of time with the financial sector being a larger portion of the economy.

Max L.

Bob Murphy writes:

Scott, good post, though I agree with the other commenter above that you wrote in a very unusual dialect of "plain English"...

More serious, though, I thought the same thing about your suggestion that the Fed would just loosen if AD were too low in LA. What if AD is too high in 15 other cities?

I would think an easier explanation is to say that prices are only sticky in the short run, and that eventually things--even at the city level--adjust to nominal levels.

E. Harding writes:

"Thomas, I don't agree, I think it's mostly due to The Great Stagnation."
-I don't think it's because of the Great Stagnation (1973-today), I think it's because of a much more recent First-World-wide productivity stagnation that began during the recession. This is especially visible in manufacturing:

Yu Guan writes:

I think the important point here is to find ways to increase consumer's spending power, instead of just debating about the increase of minimum wage. Based on my understanding, the middle and lower class people are the ones who spend most of their savings on essential daily items and food. Therefore, driving the economy. Without them, our economy would just arrive to a stand still. It means nothing to just raise the minimum wage, but at the same time, their buying power falls dramatically. I think this would only make things worse.

ThomasH writes:

@ Yu Guan

Increasing the consumption of low-consumption folks is an understandable objective but not the key to macroeconomic management. Whatever low consumption or high consumption people save or do not save, it the the job of monetary policy to make sure the financial system puts those savings into the hands of people who will invest them. Macroeconomic failings may, in general, hurt low income people the most, and avoiding them would help, but redistributing consumption is a separate problem that needs to be treated separately.

ThomasH writes:


I agree with your disagreement that 100% of the difference between real GDP today and what it would have been had NGDP (of even the price level) been held to its pre-crisis trend, but I was surprised you passed up the opportunity to point out the policy mistake. I think human capital costs -- the lower long-term prospects of the unemployed -- and the sharp decline in the participation rates should also be attributed in some significant part to monetary policy failures and to the expectation that they will continue. I just wish the Fed would not allow this kind of disagreement.

ThomasH writes:

Shouldn't your 2015 "Keynesian" say that the increase in savings will not reduce interest rates because they are at zero? But that means there is not way for the financial system to transfer savings to potential investors, investment will not rise to offset the increased savings and so NGDP will fall (as will velocity, if you keep track of those things).

The proper response is that even if that is true right now, that it is not wise to make long-run policy on the basis of (hopefully) temporary deviations of monetary policy from it's optimum. With luck, the Fed will have NGDP or the PL or whatever it is targeting in line by 2019. Therefore a better way to raise incomes of low-wage workers would be a higher EITC financed by an increase in the progressive consumption tax.

Comments for this entry have been closed
Return to top