The United States is holding a debate that is critical to its future: whether to tax, redistribute and consume income that would otherwise be invested.

Warren Buffett has weighed in, supporting higher taxes on wealthy taxpayers. Unfortunately, the evidence he uses to make his case is superficial and flawed. One can’t help but wonder: If he had stronger evidence, why wouldn’t he use it?

Buffett, the chairman of Berkshire Hathaway, is an iconic leader. The United States needs insightful analysis from him. His claim that taxing upper-income taxpayers doesn’t reduce investment runs counter to standard economic logic.

Federal Reserve surveys show the top 5 percent of households save and invest 40 percent of their income. Median- income households save very little, whereas the Buffett household probably invests 99 percent of its income.

If we tax, redistribute and consume income that otherwise would have been invested, the investable pool of savings declines. With a smaller pool of capital, less-attractive investment opportunities remain unfunded. Buffett tautologically claims investors will continue to invest in opportunities with expected returns above the cutoff point. Of course they will. Investment is lost at the margin.

Buffett points to the 1950s and 1960s, when marginal tax rates were higher, and claims that because the economy grew faster then, it can grow faster today with higher marginal tax rates.

What he fails to mention is that the advent of interstate highways and television knitted together the U.S economy in the 1950s. Large capital-intensive companies such as General Motors and Procter & Gamble raced to exploit previously unrealized economies of scale.

As a result, entrepreneurs and individual tax rates mattered much less to growth then than they do today. Growth accelerated independent of the tax rate.

The United States sent its workforce to college long before the rest of the world. That also opened new investment opportunities. Two decades of underinvestment in the private sector — first during the Great Depression and then during World War II — added further to the rebound. The cost of food dropped from more than 20 percent of gross domestic product to less than 10 percent, freeing resources to fuel the manufacturing boom.

A much smaller portion of GDP was taxed, redistributed and consumed. Federal, state and local government spending was 28 percent of GDP then versus close to 40 percent today. The 1950s and 1960s don’t provide evidence that increased government consumption, and the taxes needed to fund it, has no effect on growth. They show that investment matters.

Buffett also claims the commercialization of the Internet in the early 1990s created a huge tailwind that benefited the rich, as if investors did little to earn this success. Similarly, proponents of higher taxes and spending often claim that faster growth in the 1990s demonstrated that higher taxes on investors don’t hurt growth even though the commercialization of the Web would have accelerated growth independent of the tax rate.

Comparing the growth of the United States with Europe’s since the early 1990s removes the effect of the Internet. Both economies had access to the same technology and similarly educated workforces to capitalize on the web’s opportunities. Since then, the U.S. economy has grown 63 percent (in the period through the end of 2010); France and Germany’s together grew less than half as fast. U.S. productivity growth increased from 1.2 percent a year to 2 percent while France and Germany’s declined to less than 1.5 percent a year in the periods 1972-1995 versus 1995-2004. Without U.S. innovation, Europe’s growth would have been lower.