This is an archived article that was published on sltrib.com in 2017, and information in the article may be outdated. It is provided only for personal research purposes and may not be reprinted.
When you're wrong, you're wrong, no matter how famous and respected you might be as a scientist. Albert Einstein was wrong about quantum mechanics. Linus Pauling was wrong about the structure of DNA. And Milton Friedman was wrong about the permanent income hypothesis. But unlike with the first two examples, where scientists quickly realized the mistake, economists haven't yet come to grips with the reality.
Friedman's theory says that people's consumption isn't affected by how much they earn day-to-day. Instead, what they care about is how much they expect to earn during a lifetime. If they have a sudden, temporary loss of income — a spell of unemployment, for example — they borrow money to ride out the dip. If they get a windfall, like a government stimulus check, they stick it in the bank for a rainy day rather than use it to boost consumption. Only if people believe that their future earning power has changed do they respond by adjusting how much they spend.
This idea is important because it meant that we shouldn't expect fiscal stimulus to have much of an effect. Government checks are a temporary form of income, so Friedman's theory predicts that it won't change spending patterns, as advocates such as John Maynard Keynes believed. Even now, when economic models have become far more complex than anything in Friedman's time, economists still go back to Friedman's theory as a mental touchstone — a fundamental intuition that guides the way they make their models. My first macroeconomics professor believed in it deeply and instinctively, and would even bring it up in department seminars.
Unfortunately, intuition based on incorrect theories can lead us astray. Economists have known for a while that this theory doesn't fit the facts. When people get a windfall, they tend to spend some of it immediately. So economists have tried to patch up Friedman's theory, using a couple of plausible fixes. People might respond to temporary income changes because they're unable to borrow — if you want to spend more, but you've maxed out your credit cards and your home-equity credit line, a windfall from the government might free you from the tyranny of the bank. Lots of economists view credit constraints as a simple, minimally invasive way to save Friedman's basic idea.
But as economists get better and better data, they're finding that even this modification isn't enough. A new study by Peter Ganong and Pascal Noel shows that consumer behavior is more short term than almost any mainstream model predicts.
Ganong and Noel looked at data on the bank accounts of people who receive unemployment insurance (anonymous data, of course). They examined how spending tends to change when jobless benefits begin — usually soon after people lose their jobs. And they also look at how spending changes when the benefits run out. That kind of individual-level data was unheard of in Friedman's day, and it shows how computerization and empirics are revolutionizing the economics profession.
The authors' first finding isn't too surprising — when people lose their job, they start spending less. This is consistent with the credit-constraint model, since lots of people can't borrow enough to maintain the lifestyle they enjoyed when they had a job. After that initial drop, the authors find that spending continues to drift lower. That's no surprise either.
But it's when unemployment-insurance benefits end that the real mystery begins. That produces another big, instant drop in spending — one that's almost twice as big as the fall that happens when people are laid off! Here, from Ganong and Noel's paper, is a picture of their estimates of how spending and income fall, on average, when employment insurance ends:
The income drop is much bigger than the drop in spending, meaning that people aren't living completely hand-to-mouth. But it's much too big to be explained by any of the leading theories. Jobless-benefit exhaustion is hardly a surprise — people know exactly when the checks are going to stop arriving. And if borrowing limitations were the story, people would have saved more beforehand, knowing their benefits were going to run out.
So this behavior is a big puzzle for the leading theories. It means there's more going on here than banks' simple unwillingness to lend. For some reason, consumers are short-termist — when money stops coming in, they cut back, even if they know they'll probably get a new job in the relatively near future.
This means Friedman's theory doesn't just need a patch or two — it needs a major overhaul. No one knows what that is yet, although some economists have started to experiment with models of short-term thinking by consumers. But whatever the eventual explanation, it's likely that decades of believing in Friedman's idea have caused us to underrate the potential power of fiscal stimulus and other policies that boost short-term income.
Even the greatest scientists can be wrong. The measure of a science is how quickly it comes to grips with the mistakes its heroes make.
Noah Smith is a Bloomberg View columnist. He was an assistant professor of finance at Stony Brook University, and he blogs at Noahpinion.