The data are clear: The U.S. economy does better economically on a range of measures when a Democrat is president than when a Republican is. So what? The very question assumes a cause-effect relationship that ignores the complex interactions in any economy, particularly one as large and diverse as ours.
Even though based on a false premise, the question is getting attention in economic circles.
Alan Blinder, a distinguished Princeton economist, former Federal Reserve governor and adviser to Bill Clinton, together with colleague Mark Watson, published a paper examining alternative explanations of why the data reflect this trend.
That he spent any time at all on this lowers my respect for Blinder, whose introductory macro textbook I used for several years and with whom I conversed a few times in the 1990s when I worked at the Minneapolis Fed.
But one might make a case in his and Watson’s defense.
Yes, the assumption that which party holds the White House determines U.S. economic performance is, by itself, a classic post hoc fallacy. This is the erroneous assumption that just because something precedes another thing, the first necessarily causes the second.
But causation can exist. Thunderstorms were known to precede tornadoes before the precise atmospherics were understood, just as the apparent movement of the moon seemed related to tides well before gravitation was understood.
So there might be a causal relationship between political party of the president and national economic performance. Perhaps we just don’t understand it yet.
Blinder and Watson examined some possible links to see which could be rejected statistically and which could not. Nothing wrong with that, is there?
The answer is typical economist waffling: yes and no.
Step back and think about the measures considered. The primary one is growth of Gross Domestic Product, which is the monetary value of goods and services produced. It is a measure of output.
They also considered rates of growth of GDP per capita, industrial output, the number of jobs, employee hours worked, the S&P 500 stock index, corporate profits and productivity. In every one of the eight measures, the average rate of growth during Democratic administrations is higher than during Republican ones.
The differences are “statistically significant,” although the relevance of such “significance” to a study that can only examine the past 12 presidents and 65 years is dubious at best.
Coincidentally, Larry Bartels, a political scientist at Princeton, has demonstrated that income inequality rises during Republican administrations and falls during Democratic ones, with those at the very top doing best under the GOP and the bottom 20 percent of households doing best under Democrats. For a good summary of this, see economist Dani Rodrik’s March 31, 2008, blog post “American political economics in one picture.”
After slicing and dicing the data, Blinder and Watson conclude that discretionary economic policies, taxing, spending, regulation and social programs don’t show much effect in relating presidencies to economic performance.
Instead, factors outside the political realm such as oil price shocks, productivity growth, international economic and political conditions and public sentiment play a bigger role.
So they just looked for statistical relationships linking presidents to economic performance and didn’t find any.
Nothing wrong with that, is there?
Doesn’t that show they had no political agenda?
Well, yes. But wouldn’t economists and anyone able to think critically know all this already? Economies are complex.
Fundamental market forces of supply and demand operate regardless of what government does. Populations grow, new resources are discovered, and technology changes regardless of what government does.
Yes, government actions can foster or retard some things.
The Morrill Act establishing land grant colleges, the Homestead Act , the Mining Act of 1874 and the G.I. Bill all affected the availability and quality of natural and human resources.
The Manhattan Project and the 1960s space program gave powerful boosts to new technology.
But these effects became evident only over decades and cannot be attributed to individual presidents or conditions during specific administrations.
Indeed, any changes in government taxing and spending policies, or import restrictions, or regulation of any kind, or “income redistribution” like Social Security or food stamps, depend on action by Congress as well as the president.
Neither can take credit — or be blamed — alone.
And interest rates and the money supply are highly determined by the Federal Reserve, an institution carefully designed to be independent of control by the president or Congress. Everyone knows that events outside our country’s borders, including World Wars I and II and the Arab oil embargo of 1973, had enormous effects on the U.S. economy during administrations of both parties.
Finally, all economists know that, to the extent discretionary policies affect the nation’s economic performance, these effects occur only after long and variable lags.
To attribute outcomes to the years concurrent with any particular president is to ignore lessons taught in any introductory macro course.
Even if you arbitrarily lag the “results” by a year or two, you ignore the fact that such lags vary with different policies and that the lags for any one policy are not uniform over time.
In summary, our economy is so complex, with ever-evolving interactions, that ascribing economic outcomes to any single variable is a fool’s errand.
So no one with good sense needed two Ph.D.s to spend a lot of time applying econometric tests to large sets of data to determine that the party of any president has little directly to do with prosperity or recessions.
And the benighted multitudes who are convinced that this is true are not going to have their opinions changed by two academics.
Blinder and Watson could have spent their time more productively on any of several other questions, even though alternatives might not have had the same appeal to general audiences as that of presidential party affiliation.