Three features of the data stand out.
First, it is not reasonable to use the history of GDP growth unadjusted for demographics in judging what is a reasonable forecast going forward. Over the 1961-2000 period the adult population grew at an average rate of 1.4 percent, compared to a forecasted growth rate of 0.2 percent over 2021-2027. So the future will lack the substantial population growth tailwind that the American economy has enjoyed historically.
Moreover GDP growth in the earlier period was supported by a tailwind from changing societal attitudes toward's women's work - particularly married women. The labor force participation rate for married women rose from 33 percent in 1960 to 62 percent in 2000 and has been roughly constant since then. So the arithmetic of demography implies that growth should be substantially slower than has been observed historically.
Second, the Trump administration forecast of growth of 3 percent for seven years between 2021 and 2027 is the equivalent of 2.8 percent growth per US adult. In the last 56 years US growth has never exceeded this figure for more than 3 consecutive years. Even if the economy's underlying growth rate increased to 3 percent, is it really plausible that there will be no recession between now and 2027?
Third, there are occasions during the 1960s, 1980s and 1990s when growth averaged 2.8 percent or a little bit more over a seven-year period. But each of these eras involved substantial cyclical recovery with the unemployment rate falling by 2.0, 4.1 and 2.9 percent over the respective periods. In contrast, the Trump administration predicts that unemployment in 2021 will be 4.8 percent. A decline of 2 plus percent from 4.8 percent seems highly implausible. There have been no periods with growth remotely comparable to Trump's forecasts that did not involve substantial cyclical tailwinds.
The Trump economic team has not engaged in serious analysis or been in dialogue with those who are capable of it so they have had nothing to say in defense of their forecast except extravagant claims for their policies. Taking their supply-side perspective, do they really believe that that through tax cuts and deregulation they are going to accomplish more than Ronald Reagan, who after all reduced the top tax rate from 70 to 28 percent?
Between 1981 and 1988, GDP per adult grew by an average of 2.5 percent, distinctly slower than what they are forecasting. Even this figure reflects a substantial cyclical tailwind from the decline in unemployment from 7.6 percent to 5.5 percent (which from Okun's law implies adding about half a percent to GDP growth) - something unavailable in the present context.
OMB Director Mulvaney, in a flailing defense against my criticism of his budget arithmetic, notes correctly that the Obama administration's early forecasts overestimated the pace of recovery. True. We made the same error that the consensus of professional forecasters, the Fed and the IMF all made. We did not offer our own original view of our programs' impact. There is a reason why businesses are audited and a reason why heretofore government economic forecasts have followed the consensus. No one is credible evaluating their own work.
A business trying to sell stock on the basis of a document half as hype-filled as the Trump budget would be a joke. No reputable investment bank would underwrite their offering. A great mystery here is why the experienced investment bankers in senior positions in the Trump administration hold the Budget of the United States to so much lower standards of integrity than they applied in their earlier lives.
Commentary by Larry Summers, the former Treasury Secretary and currently the Charles W. Eliot University Professor and President Emeritus at Harvard University. Follow him on Twitter @
With thanks to Anna Stansbury for data assistance.
For more insight from CNBC contributors, follow