Mary Robinette Kowal brought my attention to a blog post by author Scott Westerfeld called “Do the Math,” in which he claims:

Since World War II the Democrats have been overwhelmingly better at running the economy.

He runs some numbers on economic growth under Democratic presidencies vs. Republican presidencies, and comes up with the following numbers for percentage rate of growth:

Democrats get 4.326 growth per year on average
Republicans get 2.8 growth per year on average

That’s about 55% better for D’s than R’s. That’s huge.

I tried to duplicate his results by getting the numbers from the source he referred to, and came up with 4.236 for the Democrats and 2.8 for the Republicans, so obviously one of us has a tiny error in the D numbers.  But so far, so good. (At least as far as doing the math is concerned.  Things aren’t looking so good for the Republicans.)

But I noticed that Westerfeld started counting with 1948 rather than 1946, which was the first year after World War II.  Now, that can be explained by his only wanting to look at 60 years of data, but what happens to the numbers if we start with the first full year after the end of WWII?

Democrats get only 3.481 growth per year on average, compared to the Republicans’ 2.8.  Instead of 55% better, D’s are only 24% better.

Still, the Democrats do 24% better — that’s pretty significant, isn’t it?

Well, over the exact same post-war period, the economy grew at an average rate of 3.55% when American League teams won the World Series, and only 2.46% when National League teams won.  Therefore, an AL team winning is 44% better for the economy than an NL team.  (If we start with 1948, AL teams are still 23% better for the economy.)

Now, I didn’t do this analysis because I believe the league of the World Series winner has a major impact on the rate of economic growth (nor that the rate of economic growth somehow favors one league over the other in the World Series.)  I did it to illustrate that this simplistic form of analysis doesn’t really show a causal relationship.

That’s not the only problem with Westerfeld’s post.  He also does some analysis of economic grown based on party control of Congress, concluding:

Democrats in charge:
1948-52, 1955-80, 1987-2000, 2007
Total growth: 3.59
Republicans in charge:
1953-54, 1981-86 (control Senate, but not House), 2000-2006
Total growth: 2.83

Maybe Westerfeld was just really busy between 1995 and 2000, but he somehow missed that the Republicans took control of Congress after the 1994 elections.  (And in four pages of comments on the YA for Obama site, no one seems to have caught that mistake.)  The correct numbers are 3.57 for the D’s and 3.06 for the R’s.  If we use the true postwar period, going back to 1946, the D’s are reduced to 3.12, just barely ahead of the R’s.

But really, looking back to the 1940’s is a bit of a stretch, isn’t it?  The parties have changed a lot — the Democrats were still supporting Jim Crow laws back then, for example.  Lets look at the last 40 years of Congressional control: 2.96 for the D’s, 3.12 for the R’s.  Or the last 20 years: 2.67 for the D’s, 3.11 for the R’s.

Somehow, I doubt Westerfeld’s going to vote Republican for Congress based on doing the math.

Now lets look at the last set of numbers, with which Westerfeld tries to show that Democrats are the party of fiscal responsibility.

During R presidential terms, the average borrowing every year is 2.36% of the whole economy.
During D terms, the average is 0.21%.

In other words, Republican presidents borrow eleven times as much money as Democrats. I’ll say that again: eleven times as much.

First off, he’s done the math wrong, as far as I can tell.  I took the numbers from the document he linked to, and the number I come up with the the R’s is 2.38%.  Maybe that’s a typo on his part, but for the D’s I get 0.76%.  I don’t know where his mistake was, but if you’re going to “do the math,” you should do it right.  Saying “eleven times as much” twice doesn’t make it any truer; it’s only about three times as much.

For some reason, he only looks at the presidential numbers, not the Congressional ones, even though Congress, not the President, has control of the purse strings.  For the last 60 years, the Congressional numbers are 1.59 for the D’s, 2.01 for the R’s, which is much closer.

But then, you really have to look at history.  The modern welfare state in the U.S. began with Lyndon Johnson’s second term.  If we look at the numbers since 1965, when Democrats control Congress, the average borrowing is 2.48% of the whole economy.  When Republicans control at least one house, average borrowing is only 2.12%.  (When Republicans are in complete control of Congress, it’s only 1.08%.)

However, it’s important to emphasize that, tempting it may be to make partisan conclusions based on these correlations, the main driver of both the economic growth numbers and the deficit numbers is business boom-bust cycles over which the President and Congress actually have very little control.  (Do you really think the Internet boom would not have happened in the later 1990s with President Dole and a Democratic Congress instead of President Clinton and a Republican Congress?)

Just as one last reminder that correlation in these numbers really doesn’t mean a whole lot, I’ll point out that when an NFC team wins the Super Bowl, the deficit averages 2.99% of the economy that year, but when an AFC team wins, the deficit’s only 1.8%. 

Sometimes, doing the math is not enough.  You need to know the limitations of the math.  (And it helps to know a little history, too.)

P.S. I have nothing against Scott Westerfeld personally and I’ve heard his books are very good.  I’m merely pointing out flaws in his political analysis.