What does an Economic Stimulus Do in the Long Run?

In the last article in this series, I talked about how economic stimulus is supposed to work.  The basic premise is that when the government spends more money on goods & services, they more than replace the spending that households would've done without the stimulus.  That is the marginal propensity to consume (MPC) of the government is higher than that of households.

Economist Robert Barro has argued for the past thirty years or so, that this simple story ignores important aspects of household reasoning.  Imagine for example, that the government decides to run a budget deficit by lowering taxes.  This should be stimulative, since it leaves more money in the hands of consumers, who will presumably spend at least some portion of it (based on their MPC).  Barro argues, however, that rational households will realize that cutting taxes today and leaving spending unchanged does not mean taxes can stay permanently lower.  Such a policy is unsustainable over a long period of time.  Forward looking households will realize that lowering taxes today means raising them in the future (holding government spending constant over time in our example).  As a result they have a lower tax burden today and higher taxes in the future.  Since most consumers prefer consumption that is smoothed out over time, rather than high consumption now and low consumption later, the best behavior is to save.  Barro showed that under certain circumstances households increase their savings by exactly the same amount as the government reduced their taxes.  This leaves consumption unchanged and hence gives not stimulus at all.

 Barro has termed this effect "Ricardian equivalence" based on work by David Ricardo in the 1800's. As mentioned, Ricardian equivalence holds only in certain circumstances.  First, all consumers need to be rational and forward looking.  If some consumers do not care about the future, they will view a drop in taxes today as an increase in spendable income and increase their consumption.  If some consumers do not expect to be alive when the tax increase occurs they will also be likely to increase their consumption.  In these cases, there will be a stimulative effect of cutting taxes, even when it must necessarily be a temporary cut. 

Similarly, if the government increases spending and does so by borrowing money, rational households will realize that even though taxes don' t rise today, they must eventually rise at some future date.  The further off into the future that date is expected to be, the greater the number of households that will expect to be dead when the increase hits.  And, hence, the greater the stimulus will be.

A perfect example of this effect occurred in 1992 when Presidents Bush used an executive order to reduce federal withholding of taxes from people's paychecks.  The order did not change their overall tax burden; the same amount of taxes was due on April 15^th.  However because withholding was lower the expected payment (refund) in April was larger (smaller) than before the change.  Since almost everyone in the economy expected to be around when the tax bill came due, this is an almost perfect implementation of a policy subject to Ricardian equivalence.  The response was exactly what Barro predicted, consumers saved most of their withholding and the policy had virtually no effect on the macroeconomy.

Very few economists believe that Barro's strict version of Ricardian equivalence where stimulus spending has zero effect is correct.  Nonetheless, his point is well taken.  That is, the effects of a stimulus are likely to be much smaller than those predicted by standard Keynesian models.  The effects are also likely to be larger when the future burden of paying for the stimulus is paid by future generations.  Stimulus spending that is paid back in the near future is likely to have very small effects.

As citizens and voters we need to ask ourselves if the short-run gain from a meaningful economic stimulus is worth the cost it imposes on future citizens and voters.