The Mythical Aggregate Demand Effect of Redistribution

By Casey B. Mulligan

Thursday, July 30, 2020

 

When you subsidize something, you get more of it. Unemployment is no exception. Subsidizing unemployment will result in more unemployment, both by increasing the number of layoffs and by increasing the duration of time that people remain unemployed.

 

The Congressional Budget Office has said that unemployment subsidies reduce unemployment in the short run because of an assumed increase in aggregate demand. The theory is that the income augmentation from enhanced unemployment insurance spurs consumption, thereby increasing the demand for workers. CBO is mistaken because aggregate demand consists of a lot more than consumption of the unemployed. Its report fails to keep track of all of the parties to a subsidy transaction, including taxpayers and lenders to the government (an oversight that could have been prevented by using an equilibrium model).

 

Instead, CBO analogizes unemployment checks to government purchases of goods and services, such as roads or military bases. But government purchases directly increase employment by paying people to work and produce something, whereas unemployment subsidies pay people not to work. Unemployment assistance and government purchases are antonyms when it comes to behavioral effects.

 

Research has shown that the unemployed “spend” their unemployment benefits, often on necessities. But this observation is at best a red herring when it comes to aggregate demand, for three reasons. First, an alternative to spending is saving, which creates demand for investment goods (it is sometimes said that consumer spending is much bigger than investment, but still, a dollar of consumption is worth no more or less than a dollar of investment!). Aggregate demand consists not only of consumption, but also of investment, so a subsidy that increases consumption may alter the composition of demand, but it does not increase it in the aggregate.

 

Second, the unemployed person would likely spend even more if he had a job, which the unemployment subsidy discourages. While an unemployed individual receiving a subsidy spends more than one not receiving a subsidy, neither spends more on average than an employed individual. The effect of incentivizing unemployment, therefore, reduces consumption. This is the channel through which unemployment benefits reduce aggregate demand.

 

Third, the taxpayers and lenders to our government finance these benefits and therefore have less to spend and save on other things. Even a foreign lender who decides to lend that extra $1 million to our government may well be lending less to U.S. households and companies. At best, redistribution from workers to the unemployed reallocates demand rather than increasing its total.

 

Finally, it is worth noting that if CBO is correct, it did the U.S. a disservice in March.  The March 2020 CARES Act unemployment benefits were partly justified as a means of reducing COVID-19 transmission by keeping people out of work. With CBO’s “model” saying that unemployment benefits actually increase work in the short run, the same benefits were increasing COVID-19 cases and ultimately deaths.