Bryan Caplan  

The Dining Room Table Responds

Hayek on Knowledge... Correcting a Favorite Textbook...
Penny: He's a really good-looking guy, and I thought he was kind of cheesy at first...

Billy: [quietly] Trust your instincts.

Penny: But, he turned out to be totally sweet. Sometimes people are layered like that. There's something totally different underneath than what's on the surface.

Billy: And sometimes there's a third even deeper level and that one is the same as the top surface one.

                  --Dr. Horrible's Sing-Along Blog
The story so far:

Common sense - and the Paul Krugman of 1999 - say that cutting wages reduces unemployment. 

The Paul Krugman of 2009 objects that cutting wages reduces Aggregate Demand, so cutting wages won't help. 

I reply that cutting wages actually increases Aggregate Demand, so it will help. 

Then Krugman gets frustrated, repeats that I'm "simply assuming that a rise or fall in nominal wages is equivalent to a rise or fall in real wages," and says that he feels like he's "arguing with a dining room table."

My reply to his reply: I'm not simply assuming that nominal and real wages rise and fall together.  What I'm saying, rather, is that the secondary effect on Aggregate Demand that Paul emphasizes is, in all probability, positive rather than negative.  If labor demand is elastic, then lower wages imply higher labor income, and higher AD.  And even if labor income falls, employer income goes up when wages fall, so AD could still easily rise as a result of lower wages.

I'll admit that counter-examples are conceivable.  Consider the polar case where labor demand is perfectly inelastic.  If wages go down by 10%, employment stays the same, so labor income goes down by 10%.  Employer income goes up by an equal amount, so the net effect on AD depends on the marginal propensity to stuff income under one's mattress.  If that propensity is higher for employers than workers, then AD goes down; otherwise, it still goes up.  The key thing to notice about this example: You need bizarre assumptions to get the result that Krugman treats as normal.  

If he had a pile of empirical evidence on his side, I could at least understand Krugman's doubts about the common sense position.  But all he's got is a theoretical curiosity in search of an empirical anomaly; straightforward historical and contemporary tests support the standard view that wage cuts are the cure for unemployment. 

Bottom line: I may be a dining room table, but Krugman doesn't have a leg to stand on.

Comments and Sharing

COMMENTS (18 to date)
Steve M. writes:

Bottom line: Caplan believes in smaller gov't, Krugman believes in bigger gov't

Anthony writes:

I think Krugman's point is that you're right in normal times but wrong when interest rates are at the zero bound.

Anthony writes:

I think Krugman's point is that you're right in normal times but wrong when interest rates are at the zero bound.

Ryan Vann writes:

This argument seems to lack context. Wages don't just randomly fall for no reason. It is important to examine the conditions that lead to the decrease before talking about the aggregate effects.

So, Prof Caplan what is the assumption behind the decrease in wages? Also, I would like to point out that your example of wage cuts being a cure for unemployment is a bit misleading. What you link to more specifically demonstrates how wage fixing above a market clearing price reduces unemployment. In that specific case, wage cuts are indeed the solution. I would say that, to me, you are far more consistent and coherent than Krugman on the matter. However a little expounding would be appreciated.

david writes:

At some point during this thread of argument, the liquidity trap element vanished. Anyone know why?

Steve Roth writes:

Whatever piece of household furniture you may be (the library floorlamp?), I don't think your (implicit) syllogism from Scott's table stands up. It's really not the silver bullet in this argument. (Though it is bloody persuasive in its own way.)

1. A mandated 20% increase in wages for a large part of the population resulted in a rapid decline in industrial production.


2. An allowed low-percentage decrease in wages for a small percentage of population will result in (a rapid?) increase in aggregate demand.

There are differences in:

1. Number of people affected
2. Size of the percent change
3. Mandated versus allowed
4. Expected results

That before considering threshold effects, or for that matter the interest-rate/inflation climate.

This does not mean you're wrong--just that this syllogism doesn't hold water.

Also, speaking of questionable assumptions, is labor demand elastic? Particularly in a time of stagnant demand for goods and services?

And I mentioned already another weak assumption: aren't owners--being wealthier--in fact less likely to spend newfound income?

ryan vann writes:

Probably because even liquidity trap pushers don't believe we are in one.

Mercure writes:

Bryan, your argument holds only if employers can cut wages while keeping the prices of their products constants. It's possible if only one employer cut it's wages. If my understanding is correct, Krugman argue that if wages are cut for all employers, the competition will push the products prices lower too. So the initial wage cut will lead to deflation.
You can disagree with Krugman, but I am surprised that you manage to answer him without even mentioning the word "price".

Ryan Vann writes:

I'll restate again that I think our conclusions about the effects of wage decreases hinge on the assumptions we make about why the wages are decreasing. With that said, I would like to address Mercure's criticism.

I don't believe that prices need decrease simply because input costs decreased in aggregate. For one, firms could decide to use the surplus in production capacity over labor wages to fund expansions (a point I believe Prof. Caplan is trying to articulate). Likewise, firms observed a previous price level which consumers found acceptable. They might just decide to do some rent seeking, and make some economic surplus. There could be inelastic demand (not a far reaching assumption in a recession, as people tighten belts and only spend on the most necessary/desired stuff, regardless of price).

However, where I do believe negative growth pressures are pretty much guaranteed with wage cuts are in the debt default risk. It is assumed our current recession is due to insolvency of an abnormal percentage debtors. This is multiplied by the prevalence of securitized debt obligations. Basically, as nominal wages decrease but nominal debt payment remain constant, risk of default increases, causing a vicious cycle.

RL writes:

What is this thing Bryan has with attacking Nobel Prize winners...?

"If labor demand is elastic" -- In a steep recession, w/ interest rates at zero?

OK, so prices would tend to fall. We're in a situation w/ nominal interest rates pinned at zero, and we more in a deflationary direction. So real interest rates up.

You are arguing that Aggregate Demand shifts which way?

Thorstein Veblen writes:

Hey Bryan, can you share with your blog readers on what page Paul Krugman says that, in depression-like liquidity trap territory, cutting wages reduces unemployment?

[Comment edited.--Econlib Ed.]

Rick Schaut writes:

Let's see, you take Krugman's 1999 statement out of context, and still refuse to discuss the implications of being pegged at the zero short-run interest rate bound. Krugman's right. You are a dining room table, your attempts to adorn a new coat of varnish notwithstanding.

jonm writes:

[Comment removed for supplying false email address. Email the to request restoring this comment. A valid email address is required to post comments on EconLog.--Econlib Ed.]

VR writes:

Based on this post, we are to believe that is reasonable to assume that, in a steep recession and zero interest rate environment, labour demand is elastic and that consumers have a lower marginal propensity to consume than businesses.

I'm not sure how those ended up being the more reasonable set of assumptions in this discussion.

Scott Sumner writes:

Bryan, I agree with you. I make a slightly different argument in a new post, which I think addresses the liquidity trap problem raised by Krugman:

I wrote this before I saw your new post.

mattmc writes:

I feel that these arguments are limited in that the unemployment rate being measured is US, but the forces affecting it are global.

It is reasonable to note that one driver of unemployment, say in the textile, software, or accounting industry, is that workers at lower wages are available in other countries. At some point, lowering wages brings some of that work back, and some kinds of deflation (such as the deflation of a bubble) are good.

Jim Vernon writes:

"... the net effect on AD depends on the marginal propensity to stuff income under one's mattress ..."

If this is sensible (and it seems so to me), isn't the only "bizarre" assumption needed for Krugman's outcome, in fact, merely the opposite of your assumption about MPC (or MPSIUOM)? I'm not sure it's bizarre to assume one way or the other that producers' MPC is greater (or less) that workers' MPC. Did I miss the point of that? In other words, this seems to boil down to a disagreement about the magnitude of the multiplier effect under two different assumptions.

More to the point, isn't the MPC assumption a testable hypothesis? (This saves me from having to disagree with anyone smarter than I, which could include everyone in this room.)

Best regards,

Comments for this entry have been closed
Return to top