Who is Really Best for the Economy?

All I'm getting from TV and the blogosphere is spin. So, I figured I would do a little economic research. I picked a couple of leading economic factors and did cross correlation between Republican vs. Democrat, President vs. Senate vs. House. I have seen previous work published by Slate that looked at the President using averages, but in my opinion it was a little weak. Simple averages are not going to cut it. I want a little statistical inference to sway my vote. So here is my contribution to the signal to noise ratio.

Since there will be screams of bias from either side, here are my stances. I'm a Libertarian, also known as a Goldwater conservative. I am socially liberal and fiscally conservative. I have been a fan of John McCain and was going to vote for him in 2000 before he dropped out. However, I think Sarah Palin is an abomination and I'm not sure he is putting enough distance between himself and Bush's policies. I'm not convinced that Obama is the second coming, but I like his foreign policy positions. His protectionist stances and dislike of nuclear technology are big negatives in my book. Don't get me started on FISA. Thus I am torn and looking for a little guidance.

While the idea was good, there are a number of reasons to revisit the Slate article. First, the premise was flawed by assuming that economic factors are controlled by the president. Civics 101 teaches us that congress controls the purse strings, so their impact should be much more significant. Second, I assume a bias due to the source, a typically liberal magazine. It's not that I don't trust their numbers, it is just that data is sometimes cherry picked to support a conclusion. In defense of the article, the author admitted that the analysis was known to be limited.

Assumptions

I started by reproducing the same data that Slate had by going to government and financial sources. Also, I added historical Standard & Poor data for the time period in question, since that was also a debate that has raged in recent years. All the data here is directly from the Whitehouse, Census Bureau, Bureau of Labor and Statistics, and Bloomberg. I generated real GDP by calculating the annual change and applying an inflationary correction. Spending is shown as a percentage of nominal GDP. For the purposes of this analysis, lower spending across the board is desirable, including defense.

I used data from national elections and created three series, one for for the president, house and senate. For each year, a 1 is used represent either a republican president, a republican house majority or a republican senate majority and 0 for a democratic lead in each respective position. Next, I used statistical correlation with each of these series to determine if there is a relationship between an economic factor and the political majority of that branch of government.

Any correlation with a magnitude greater than .5 was considered high, while values less that .3 were assumed to be inconclusive. A positive correlation would infer a republican relationship and a negative correlation would indicate a democratic one. An inversion variable was added to account for the fact that some economic factors are desirable by their increase, such as GDP, and others are preferable by staying low, like inflation.

Cross correlation was performed with a 0 to 3 year lag to catch delayed effects. Any longer and the lag would appear in subsequent administrations thus further muddying analysis. A primary and secondary maximum were selected from each of the categories to highlight the magnitude of its effect.

Findings

Strong correlations were seen between democratic presidents and budget surpluses, republican presidents and low taxes and a republican senate and low defense spending relative to GDP. S&P dividends strongly followed a democratic house, however, the magnitude if their gains were greater under a republican control. The rest of the correlations are summarized below.

Limitations

The selection of 1960 as a starting point is arbitrary. I have good data on the stock market until then. If I find more, I might expand the scope. It could also include World War I/II and the Great Depression. This might change the conclusions, but I'm going with what I have. Also, correlation does not imply causation. This is not political fact, rather additional data to chew on while you decide.

Conclusions

With the house and senate available for comparison, it was apparent that there are much stronger correlations between the legislative branches than the executive for certain categories. The strongest correlations were between the executive branch in issues involving taxes and budgeting. The correlation between a republican house & senate and reduced defense expenditures relative to GDP was a surprising result and was further confirmed by inspection of the averages.

The performance of the stock market relative to the presidency was shown to have little to no correlation. However, there was a strong relationship between stock performance and a democratic legislature. The addition of the fact that average performance was higher under republican legislative control was somewhat contradictory. Given this information, this may lead to an inconclusive result.

There was an obvious, although not overwhelming, correlation between a democratic presence in government and positive economic factors. The exception being lower defense spending, taxes and inflation which are positive with respect to republican influence. While a causal conclusion cannot be drawn from this technique, there is enough evidence to peak interest and warrant further study.