A recent NYT article from this year’s economics Nobel Prize winners, Esther Duflo and Abhijit Banerjee, argues that “economists have somehow managed to hide in plain sight an enormously consequential finding from their research: Financial incentives are nowhere near as powerful as they are usually assumed to be.” Yet as we’ll see, Duflo and Banerjee cherry-pick studies that support their conclusion, while ignoring the body of literature that opposes it. Furthermore, in several cases even the studies they cite actually prove the opposite of what the new laureates lead the public to believe. Finally, Duflo and Banerjee only use this newfound principle—namely, the lack of financial incentives to alter behavior—when it serves to increase government power; they don’t use it, for example, to argue against carbon taxes or speeding tickets.
Duflo and Banerjee’s Evidence Against Incentives
Before proceeding to critique their case, let me generously quote from Duflo and Banerjee in their own words:
At least since Adam Smith and his famous B’s (“It is not from the benevolence of the butcher, the brewer, or the baker that we expect our dinner, but from their regard to their own self-interest.”), a fundamental premise of economics has been that financial incentives are the primary driver of human behavior…
This is unfortunate, because economists have somehow managed to hide in plain sight an enormously consequential finding from their research: Financial incentives are nowhere near as powerful as they are usually assumed to be.
We see it among the rich. No one seriously believes that salary caps lead top athletes to work less hard in the United States than they do in Europe, where there is no cap. Research shows that when top tax rates go up, tax evasion increases (and people try to move), but the rich don’t work less. The famous Reagan tax cuts did raise taxable income briefly, but only because people changed what they reported to tax authorities; once this was over, the effect disappeared.
We see it among the poor. Notwithstanding talk about “welfare queens,” 40 years of evidence shows that the poor do not stop working when welfare becomes more generous. In the famous negative income tax experiments of the 1970s, participants were guaranteed a minimum income that was taxed away as they earned more, effectively taxing extra earnings at rates ranging from 30 percent to 70 percent, and yet men’s labor hours went down by less than 10 percent. More recently, when members of the Cherokee tribe started getting dividends from the casino on their land, which made them 50 percent richer on average, there was no evidence that they worked less.
And it is true of everyone else as well — tax incentives do very little. For example, in famously “money-minded” Switzerland, when people got a two-year tax holiday because the tax code changed, there was absolutely no change in the labor supply. In the United States, economists have studied many temporary changes in the tax rate or in retirement incentives, and for the most part the impact of labor hours was minimal. Nor do people slack off if they are guaranteed an income: The Alaska Permanent Fund, which, since 1982, has handed out a yearly dividend of about $5,000 per household, has had no adverse impact on employment.
Wow! It seems as if financial incentives have no impact on labor markets, whether people are rich or poor. How could economists have been so blind to this, when after all—so Duflo and Banerjee tell us—it’s their own research that alerts us to the impotence of the tax code and welfare benefits to influence behavior?
The truth is that Duflo and Banerjee paint an extremely misleading picture with the above narrative. But before I dive into the details, let me first point out what the Nobel laureates don’t do with their observations.
Duflo and Banerjee Only Use Research to Justify A Bigger State
After summarizing the state of the economics research as I showed in the excerpt above, Duflo and Banerjee proceed to argue for more government social spending, to be paid for by higher taxes on the rich. The connection here is that the standard worries from conservatives and libertarians—namely, that welfare programs and high taxes will discourage economic output—are supposedly shown to be phantom menaces, in light of the research they had earlier summarized.
But why stop there? If it’s true that people don’t respond to tax incentives, then oops—there goes the case for a carbon tax, as well as “sin taxes” on alcohol and tobacco. And let’s get rid of speeding tickets too, since they must not deter dangerous driving.
And yet, that’s not how Duflo and Banerjee chose to take their article; they only used the ostensible misplaced faith in incentives to argue in favor of expanding government intervention, to justify policies that they endorse on other grounds. (Scott Alexander, Bryan Caplan, and David R. Henderson all had similar reactions to Duflo and Banerjee’s NYT piece.) If Duflo and Banerjee had coupled their calls for tax hikes with a simultaneous suggestion to end tax credits for electric vehicles I would be less suspicious.
Indeed, if we push this to the limit, economists should stop drawing upward sloping supply curves and downward sloping demand curves. We can no longer make sense of stores that slash candy prices after Halloween. Firms should stop offering bonuses to their employees for exemplary performance. The X-Prize competition is a farce. Do Duflo and Banerjee want to go this far? Do they really think financial incentives are a weak motivator of economic behavior?
Athletes Don’t Respond to Financial Incentives?
As shown in the long excerpt above, our Nobel laureates repeat a common objection to claims that people respond to incentives: Are we really supposed to believe that American athletes don’t work as hard because of salary caps? Ha ha!
For one thing (as David R. Henderson also pointed out), the salary cap applies to an entire team, so it’s actually not necessarily a hard constraint on the superstar’s salary. (In practice, it might mean that the highly paid superstar forces the team to pay less to the other players on the roster.) Furthermore, the salary cap is a voluntary mechanism that the teams in a league agree to, partially to ensure a distribution of talent so that the league is more competitive. If the teams based in NYC and Chicago won all the championships because they could just hire the best players, then the public as a whole might not watch the sport as much. So a salary cap isn’t at all analogous to high marginal tax rates.
Indeed, the salary cap only makes sense if players do respond to financial incentives! In other words, the point of the cap (besides the obvious goal of team owners to enrich themselves by engaging in cartel behavior) is to ensure that less wealthy franchises, based out of smaller cities, can also be competitive. The only way this mechanism works is if smaller cities can attract megastars by offering them a boatload of money.
Now it’s certainly true that a college superstar will likely go into professional sports for his career, whether there is a salary cap and regardless of top income tax rates. But that hardly proves “incentives don’t matter for star athletes.” It simply means that on the margin, it’s still more attractive to earn “only” a few million dollars as a famous athlete under a salary cap, rather than becoming an accountant.
But there are obvious examples where star athletes make decisions “for the money.” Would Conor McGregor and Floyd Mayweather have agreed to their recent match if their pay had been capped at $10,000 each, rather than the many millions they were each guaranteed? More generally, does anyone doubt that heavyweight boxing champs would defend their titles fewer times, if they weren’t allowed to keep significant amounts of money for each subsequent fight?
Sports fans are familiar with many examples of superstars making career decisions “for the money.” For example, in 1972 NHL legend Bobby Hull left the Chicago Blackhawks (which had come in 1st place in their conference the prior season) to join the new World Hockey Association (WHA), after he boasted that he would do so “for a million dollars” and they called his bluff.
For a similar example, Herschel Walker is arguably the greatest running back in college football history, and yet he initially played in the United States Football League (USFL)—rather than the more famous NFL—even though it was of course much less prestigious. The obvious reason? The USFL allowed him to turn pro a year earlier, and to pick his city, thereby maximizing his income (including endorsements).
For a final example, Alex Rodriguez (“A-Rod”) went from the very successful Mariners (who lost the AL pennant to the Yankees) in 2000 to the last-place Texas Rangers in 2001. Why would he make such a seemingly foolish decision? Because he signed a $252 million contract, which was at the time the highest in sports history.
So it’s true, star athletes in America don’t stop going to the gym just because of a salary cap—if they did, then the league owners would adjust the cap! But the idea that economic incentives stop mattering once we look at professional sports is absurd. Professional athletes respond to salaries just like professional lawyers and surgeons. Even though they are already rich, star athletes will abandon their fans and move across the country—or even across an ocean, if they were born in Europe—just to make more money.
Looking More Closely at Those Research Papers
When it comes to the research that Duflo and Banerjee cite, the irony is that these papers were published in a literature showing that incentives do matter; just skim the papers to see how their authors refer to pre-existing work.
But even on their own terms, some of the studies that Duflo and Banerjee cite, don’t really get the job done. In the interest of brevity, let me focus just on two.
First, consider the study of the Cherokee tribe that received a large windfall from casino revenue, which apparently had no effect on labor supply. Here is the relevant table reporting the results:
Now Duflo and Banerjee want us to focus on the fact that eligibility for casino cash transfers had no statistically significant influence on the labor supply decisions of either mothers or fathers in the tribe; that’s why there are no asterisks along the top row.
However, the table reports that a mother’s age and having young children also had no statistically significant impact on whether the mother worked. Does that sound right? Are we now saying that not only financial incentives, but family demographics too, have no impact on labor decisions? I suggest that something is wrong with the research design, perhaps just being a problem of inadequate sample size.
The Alaska Case Proves the Opposite of Claimed Outcome
Yet I’ve saved the best for last. Remember what Duflo and Banerjee said, regarding Alaska: “Nor do people slack off if they are guaranteed an income: The Alaska Permanent Fund, which, since 1982, has handed out a yearly dividend of about $5,000 per household, has had no adverse impact on employment.” And the NYT piece accompanies their claim with the following charts:
Now here’s what’s hilarious: You can use these very charts to conclude that the Alaska dividend reduced labor supply.
Specifically, the chart on the left is just showing how many people (as a percentage of the population) have a job, period. So that doesn’t change, with the introduction of the Permanent Fund. That means when people in Alaska starting getting the (modern equivalent) of about $5,000 per year in dividend checks, they didn’t quit their jobs altogether.
However, the chart on the right shows that there was a large increase in the proportion of workers in Alaska who worked part-time. Amazingly, the NYT commentary interprets this as evidence of Keynesian demand-side general equilibrium effects. But a more straightforward interpretation is: When people get $5,000 in annual checks from the government, some of them cut back on their hours worked.
And indeed, if you go to the original paper that was the source of these charts, you will find that yes indeed, the introduction of the Permanent Fund in 1982 went hand-in-hand (over the entire period) with a reduction in hours worked in Alaska, relative to the counterfactual control group (“synthetic Alaska”) constructed from weights placed on other states:
As the authors say, “Consistent with our results for the part-time rate, we estimate a reduction on intensive margin, albeit less than 1 hour per week.” (They also add, “Furthermore, we are not able to rule out a null effect on hours given our confidence intervals.”) For the layperson, let me explain: If the Alaska Fund had reduced labor supply on the extensive margin, it would mean people dropping out of the labor force altogether. But to reduce labor on the intensive margin merely means cutting back hours.
Stepping back, I ask: Is this really a crushing blow to the theory that financial incentives influence behavior? The government of Alaska started paying people the (modern) equivalent of some $5,000 per year, and in response some people cut back hours, though the checks didn’t induce people to stop working altogether. Isn’t this outcome what most people would have guessed?
Furthermore, whether this was predicted or not, isn’t it incredibly misleading to describe this to the NYT readers by saying that the Alaska program “has had no adverse impact on employment”? (To be clear, the authors of that paper themselves have a similar description; this summary wasn’t invented by Duflo and Banerjee.)
There is a growing drumbeat from some high-profile economists—including two of the recent Nobel laureates—to reassure Americans that large increases in income and wealth taxes won’t distort labor markets. Yet much of their arguments are very misleading to non-economists, who don’t make distinctions between employment on the intensive vs. extensive margins. Furthermore, the notion that “incentives don’t matter” has all sorts of implications, yet it seems only to be used to justify increases in government intervention.