A recent study posted by the Congressional Budget Office, “Taxes and the Economy: An Economic Analysis of the Top Tax Rates Since 1945” purports to show that tax rates are uncorrelated with growth and savings. Some political commentators, such as Eliot Spitzer, have used the report to argue for tax increases, claiming they have no effect on growth. A closer look at the report, however, would leave any serious economist scratching their head in dismay.
The report simply plots each year’s U.S. per capita growth rate against the year’s U.S. marginal tax rate and finds no statistically significant correlation. There are two problems with this approach.
First, there are many factors that contribute to growth. In the 1950s for example, the United States knitted the nation’s economy together with TV and interstate highways. This created a window to capitalize newly accessible nationwide economies of scale. At the same time, the United States was the first economy to send its workforce to college. Again, this opened untapped growth opportunities. Two decades of underinvestment, first from the great Depression and then from WWII, added further to the growth opportunities. The cost of food fell from over 20% of GDP to less than 10% freeing resources needed to fuel growth. Federal, state and local government spending was low relative to GDP—28% then versus closer to 40% today. As a result, effective tax rates were lower then despite higher marginal rates. Tax shelters allowed investors to avoid high marginal rates. You can’t just compare the growth rate in the 1950s when marginal tax rates were higher (but effective tax rates were lower) to the early 2000s when marginal tax and growth rates were lower and claim there is no correlation, without taking all these differences into consideration.
That’s why they don’t award PhDs for not finding correlations. The world is complex. It’s easy not to find correlations and hard to find them. A legitimate analysis must establish the causation of growth and then show that taxes don’t contribute further to the statistical significant of the model’s prediction. But even then, because of the complexities in predicting growth, one has to remain very leery of any analysis that claims to prove anything from a lack of correlation.
One economist who tries this approach is Peter Lindert who argues “Why the Welfare State Looks Like a Free Lunch.” He uses multiple factors to predict the growth of high wage economies between 1978 and 1995. He claims to have a statistically significant model for predicting growth and that tax rates do not contribute significantly to the predictive power of his model. But after 1995, the U.S. economy grew faster than Europe and Japan’s despite all three economies having access to burgeoning internet technology that accelerated growth and similarly educated workforces to capitalize on these opportunities. Do the factors in his model predict this difference in growth? No.
Similarly, even the most sophisticated economic forecasting models failed to predict the growing risk of the financial crisis. It’s obvious these models overlook significant factors that effect growth independent of tax rates.
Some political commentators claim there is no proof whatsoever that tax rates effect growth. But the United States ran an experiment of lower government spending and taxes since the mid 1980s while Europe tried the opposite. And the United States grew faster after the commercialization of the internet. Since 1991, the U.S. grew 63% while France and Germany grew less than half as fast. Productivity growth rose from less than 1.5% per year to over 2% per year while France and Germany’s productive growth did the opposite—growth declined to less than 1.5% per year. Standards of living rose in the United States relative to France and Germany. Today America has 35 to 40% more hours of work per working age adult than France and Germany. No evidence? The differences are startling. And if anything, innovation produced by the U.S. and implemented by Europe economy pulled the growth rates of the rest of the world up.
A second problem with correlating tax rates to growth rates is the time frame. If Europe, for example, were to lower its spending and taxes, would it immediately grow like the United States? Of course not. Its workforce lacks the on-the-job training to quickly accelerate growth through innovation. Imagine trying to innovate in Greece today. Its workers simply don’t have the training and won’t for a long time, if ever.
The success of the United States is decades in the making. It can’t be recreated in a couple of years. It stems from years of positive feedback loops restructuring the U.S. economy. For example, success begets success. Relative success in the United States demands more work and risk-taking. Our most talented people work longer hours than their counterparts in Europe. They flock to business schools to get training that allows them to compete. And they start companies that grow the economy through innovation. That success creates companies like Google, Facebook, and countless others that give our workers far more valuable on-the-job training. That training increases their chances of success. Better chances increases risk-taking. The success of our risk-takers puts equity in to the hands of investors willing to underwrite the risks needed to produce innovation. No surprise, the America has much more equity per dollar of GDP than Europe and Japan, it underwrote more of the risks needed to produce innovation, and, as a result, it grew faster.
One recent state-of-the-art study, “Do Tax Structures Affect Aggregate Economic Growth? Empirical Evidence from a Panel of OECD Countries,” that looks at the effect of taxation on the growth of 21 high-wage economies over 35 years finds, “High top marginal tax rates of personal income tax reduce productivity growth by reducing entrepreneurial activity.” And that, “a strong reliance on income taxes seems to be associated with significantly lower levels of GDP per capita.”
This empirical evidence squares with economic theory. Higher payoffs for risk-taking should spur more risk-taking that’s needed to produce innovation. More innovation accelerates growth. A shortage of risk-taking and investment yields positive returns to risk-takers. That produces equity, which underwrites the risks that produce innovation. More equity to underwrite risk produces more risk-taking, which produces more innovation that increases growth. Taxing, redistributing and consuming income that would otherwise be invested does the opposite. It slows the accumulation of equity and lowers the expected return on risk-taking.
America chose one path lower government spending and taxes while Europe chose the opposite. After two decades of compounding feedback, the differences between the structures of these economies are stark. The U.S. economy can produce faster growth than Europe’s. A change in the tax rate is unlikely to change growth trajectories significantly in the short-run. Any effect should be expected to compound gradually.