David R. Henderson  

The Deaton Paradox

PRINT
Missing Links... They Scare Me...

Yesterday I was busy writing the piece on Nobel Prize winner Angus Deaton for the Wall Street Journal. Here's the link, although realize that it's gated. The good news: Hoover asked me to do a longer piece, which will be out soon. I'll post when it's out.

My word count for the WSJ piece was restricted to 550. A 550-word article is much harder to write than an 800-word article. The Hoover piece is just over 1500 words. But even in that, I didn't get a chance to handle one technical point about Angus Deaton's work. So here goes here.

Mike Bird, at the Business Insider, wrote a good article on Deaton, especially given how little time he wrote it in. He highlighted something that the Nobel Committee highlighted. Here's the relevant paragraph:

Deaton's 1980s studies, including one called "Why is consumption so smooth?" gave birth to a concept called the Deaton Paradox -- in short, sharp shocks to income didn't seem to cause similarly large shocks to consumption. That was an important development in understanding the actions of consumers, causing economists to rethink the "permanent income hypothesis" developed by Milton Friedman, which suggested that people spend based on their lifetime income.

Why do I find that interesting? Because even though those findings do undercut Friedman's permanent income hypothesis, they don't undercut the main thing that Friedman's PIH was used for. Indeed, they strengthen it.

Recall that one of the big issues in Keynesian economics was and is what percent of additional income people spend on consumption. If Keynesian fiscal policy gives people temporarily higher incomes and if the percent of that higher income spent on consumption is high, then the Keynesian multiplier is at work. People, including Friedman, used the PIH to say that when incomes go up temporarily people will spend only a small percent of that income on consumption. The smaller the percent they spend, the smaller is the Keynesian multiplier and the less effective is fiscal policy. Deaton's findings show that the percent of additional income that people spend when they get an "income shock" is even lower than Friedman thought. The result: the Keynesian multiplier is even lower and Keynesian fiscal policy is even less effective than Friedman and others thought.

HT to Harry Watson.


Comments and Sharing


CATEGORIES: Fiscal Policy




COMMENTS (12 to date)

What does "rethink" mean? Based upon my understanding from usage, I have come to expect "rethink" to be a synonymy for "challenge". I guess, David, that you are understanding Mike Bird's usage to mean "challenge"; you are reading Mike Bird to say that Deaton's work caused economists to doubt Uncle Milton.

But I've just looked up "rethink" in all the dictionaries readily at hand. None suggests the synonym "challenge". They only say: to think again. So that could also mean: Hey look! This is really true.

Is it possible that Deaton was agreeing with Uncle Milton? (I admit that I know nothing of Deaton, but what you've just posted.)

emerich writes:

I must be missing something because the Deaton paradox would seem to support MF's permanent income hypothesis. MF says a transitory shock to income will have little effect on consumption because it's not permanent, and Deaton's work supports that empirically. Where's the disconnect?

David R. Henderson writes:

@emerich,
You ARE missing something. Reread my post carefully.

Nathan W writes:

I think there's a very simple answer for this paradox.

We are creatures of habit, and habits take time to change/develop.

J Hanley writes:
A 550-word article is much harder to write than an 800-word article

Having just graded midterms, I think many of my students would disagree. But I suspect you're talking about serious writers. ;)

On the substance of your post, it's not quite clear to me whether Deaton's finding really does undercut the hypothesis. I could benefit from more development on that point.

I've always been a bit dubious about the hypothesis given the cognitive demands I think it requires. But I'm generally persuaded by empirical evidence, so Deaton's finding is interesting whether it undermines or actually supports the PI hypothesis.

William Bruntrager writes:

@emerich

I am confused by the same paragraph. It seems to me that Friedman's Permanent Income Hypothesis is perfectly consistent with the "Deaton Paradox," and thus I fail to see how Deaton's work presents a challenge to Friedman.

Basically I think there is a word missing in the description of Deaton's findings. Specifically, in the phrase "sharp shocks to income didn't seem to cause similarly large shocks to consumption."

I think you and I are thinking "sharp shocks to [current] income" when perhaps we ought to be thinking, "[lifetime] income."

What it comes down to is that the Deaton Paradox only makes sense as a paradox if people do not change their consumption based on *new information* about what their lifetime income will be, but rather maintain a smooth consumption from one period to the next, even if they find out they are going to be much richer than they previously believed.

None of this is spelled out in the article, of course, so I'm not sure how David Henderson's advice to read carefully could help either of us, but in any case it's the best I could come up with to make sense of the claim that Deaton presents a challenge to Friedman.

Andy Hallman writes:

Bryan, I've had the same experience with Mormons.

Most of the Mormons I came to know were on their mission and hoping to convert me. They were friendly without being pushy. I got to know others who had just finished their mission and were not as interested in converting. For my job, I interviewed students from Brigham Young University who came to my town for a Swing and Latin dance show. All of them were salt of the Earth students, although we are talking about an unrepresentative sample here.

My guess is that Mormons are trying hard to put on a happy face since their religion is considered an oddity by many in society, but not quite so odd as to be disqualifying for the Republican presidential nomination.

You raise some interesting points about wider applications of this phenomenon, about whether all groups would behave like the Mormons if they were as small, and whether the Mormons would behave as poorly as other groups if they were bigger.

Along similar lines, I am interested in reading more blog posts about social solidarity, when it's good and when it's bad. I know you've stated that it's bad, but perhaps you could delve into why a bit more.


David R. Henderson writes:

@emerich, J Hanley, and William Bruntrager,
According to Friedman’s PIH, a sharp shock to income in the short run should cause a small increase in consumption. Deaton found that it caused virtually no increase in consumption. That’s how his finding undercut Friedman’s finding. But, as I said in the post above, that finding actually strengthens the main conclusion that Friedman’s PIH was used for: to undercut the strength of the Keynesian fiscal policy multiplier.

David R. Henderson writes:

@Andy Hallman,
Good comment, wrong post.

Monami writes:

That's an interesting read. Mathematically, all that Deaton proposed with respect to rise in consumption due to rise in income, was limit tending to zero (steep slope) whereas MF proposed a graph with a relatively flat slope.

baconbacon writes:
Recall that one of the big issues in Keynesian economics was and is what percent of additional income people spend on consumption. If Keynesian fiscal policy gives people temporarily higher incomes and if the percent of that higher income spent on consumption is high, then the Keynesian multiplier is at work. People, including Friedman, used the PIH to say that when incomes go up temporarily people will spend only a small percent of that income on consumption.

I don't think this holds. The Keynesian framework is (allegedly) about underutilized resources due to a recession. This basic outline is that a recession causes a drop in income which causes a drop in spending leading to a further drop in income. The textbook response is a counter-cyclical spending response that prevents the drop in spending, not a temporary income boost above the baseline. To refute the Keynesian position with data of this type you would need to show that a decline in income didn't lead to a decline in spending.

Nathan W writes:

What's a better multiplier: investment or consumption?

If what is not consumed is saved, and what is saved is invested, then the fact that people are less responsive than expected in terms of consumption does not necessarily imply a less-than-expected multiplier.

Comments for this entry have been closed
Return to top