President Obama and congressional leaders meeting yesterday confronted calls for four key fiscal decisions: short-run fiscal stimulus, medium-term fiscal consolidation, and long-run tax and entitlement reform. Mr. Obama wants more spending, especially on infrastructure, and higher tax rates on income, capital gains and dividends (by allowing the lower Bush rates to expire). The intellectual and political left argues that the failed $814 billion stimulus in 2009 wasn't big enough, and that spending control any time soon will derail the economy.
But economic theory, history and statistical studies reveal that more taxes and spending are more likely to harm than help the economy. Those who demand spending control and oppose tax hikes hold the intellectual high ground.
Writing during the Great Depression, John Maynard Keynes argued that "sticky" wages and prices would not fall to clear the market when demand declines, so high unemployment would persist. Government spending produced a "multiplier" to output and income; as each dollar is spent, the recipient spends most of it, and so on. Ditto tax cuts and transfers, but the multiplier is assumed smaller.
Macroeconomics since Keynes has incorporated the effects of longer time horizons, expectations about future incomes and policies, and incentives (including marginal tax rates) on economic decisions.
Temporary small tax rebates, as in 2008 and 2009, result in only a few cents per dollar in spending. The bulk (according to economists such as Franco Modigliani and Milton Friedman) or all (according to Robert Barro of Harvard) is saved, as people spread any increased consumption over many years or anticipate future taxes necessary to finance the debt. Empirical studies (such as those by my colleague Robert Hall and Rick Mishkin of Columbia) conclude that most consumption is based on longer-term considerations.
In a dynamic economy, many parts are moving simultaneously and it is difficult to disentangle cause and effect. Taxes may be cut and spending increased at the same time and those may coincide with natural business cycle dynamics and monetary policy shifts.
Using powerful statistical methods to separate these effects in U.S. data, Andrew Mountford of the University of London and Harald Uhlig of the University of Chicago conclude that the small initial spending multiplier turns negative by the start of the second year. In a new cross-national time series study, Ethan Ilzetzki of the London School of Economics and Enrique Mendoza and Carlos Vegh of the University of Maryland conclude that in open economies with flexible exchange rates, "a fiscal expansion leads to no significant output gains."
My colleagues John Cogan and John Taylor, with Volker Wieland and Tobias Cwik, demonstrate that government purchases have a GDP impact far smaller in New Keynesian than Old Keynesian models and quickly crowd out the private sector. They estimate the effect of the February 2009 stimulus at a puny 0.2% of GDP by now.
By contrast, the last two major tax cuts—President Reagan's in 1981-83 and President George W. Bush's in 2003—boosted growth. They lowered marginal tax rates and were longer lasting, both keys to success. In a survey of fiscal policy changes in the OECD over the past four decades, Harvard's Albert Alesina and Silvia Ardagna conclude that tax cuts have been far more likely to increase growth than has more spending.
Former Obama adviser Christina Romer and David Romer of the University of California, Berkeley, estimate a tax-cut multiplier of 3.0, meaning $1 of lower taxes raises short-run output by $3. Messrs. Mountford and Uhlig show that substantial tax cuts had a far larger impact on output and employment than spending increases, with a multiplier up to 5.0.
Conversely, a tax increase is very damaging. Mr. Barro and Bain Capital's Charles Redlick estimate large negative effects of increased marginal tax rates on GDP. The best stimulus now is to stop the impending tax hikes. Mr. Alesina and Ms. Ardagna also conclude that spending cuts are more likely to reduce deficits and debt-to-GDP ratios, and less likely to cause recessions, than are tax increases.
These empirical studies leave many leading economists dubious about the ability of government spending to boost the economy in the short run. Worse, the large long-term costs of debt-financed spending are ignored in most studies of short-run fiscal stimulus and even more so in the political debate.
Mr. Uhlig estimates that a dollar of deficit-financed spending costs the economy a present value of $3.40. The spending would have to be remarkably productive, both in its own right and in generating jobs and income, for it to be worth even half that future cost. The University of Maryland's Carmen Reinhart, Harvard's Ken Rogoff and the International Monetary Fund all conclude that the high government debt-to-GDP ratios we are approaching damage growth severely.
The complexity of a dynamic market economy is not easily captured even by sophisticated modeling (an idea stressed by Friedrich Hayek and Robert Solow). But based on the best economic evidence, we should reject increased spending and increased taxes.v
If anything, we should lower marginal effective corporate and personal tax rates further (for example, along the lines suggested by the bipartisan deficit commission's Erskine Bowles and Alan Simpson). We should quickly enact an enforceable gradual phase-down of the spending explosion of recent years. That's what the president and congressional leaders should initiate. Then let the equally vital task of long-run tax and entitlement reform proceed.
Mr. Boskin is a professor of economics at Stanford University and a senior fellow at the Hoover Institution. He chaired the Council of Economic Advisers under President George H.W. Bush.