From the August 10, 2015, issue of NR
As the presidential-election season kicks off, governors from all over America are seeking the Republican nomination. Historically, voters in presidential elections have tended to favor the executive experience of governors, and it seems more likely than not that the eventual Republican nominee will be one of these gentlemen.
A governor is different from other politicians in that he governs an identifiable state for an identifiable period of time, and thus can be empirically associated with an economic track record in a way that a senator or a heart surgeon cannot. Economics, however, remains an inexact science. Even if a million articles were written about the record of any single governor, consensus on whether the governor was good or bad for the economy would likely still prove elusive.
A hypothetical Governor Ponnuru could rescue a state from an abyss into which it threatened to fall yet still appear to have performed poorly when measured against a governor whose state had never come so close to catastrophe to begin with. Such a pattern might be especially relevant for Republicans who take over in historically blue states, where decades of Democratic malpractice may be difficult to undo in a term or two. Nonetheless, a governor who has a strong track record of job creation could well make the claim that he could deliver similar results to the nation as a whole. A governor without such a record would have to make a more difficult theoretical case that his policies have been conducive to job creation.
We attempted to identify the “governor effect” by comparing how state employment growth changed while a governor was in office with both national employment growth and the state’s employment growth under other governors. Those comparisons allow us to estimate the impact that a particular governor had on growth in the employment rate in his state, distinguishing that effect from the overall economy’s employment fluctuations and from the typical trajectory of employment fluctuations in the state.
Suppose, for example, that Texas employment typically goes up twice as fast as U.S. employment when the economy is expanding, and suppose further that we are considering a Texas governor who presided during an expansion. Our method would credit him with a positive effect on employment only if Texas employment went up more than twice as fast as U.S. employment while he was governor.
#related#As one can see, there is considerable variation in the “governor effect” on state employment levels. Not every governor generates results in the direction that he would have hoped; others enjoy bright track records. Employment growth when Governor Jeb Bush was in office, for instance, tended to be about 4 percent higher per year than one would have predicted if the average Florida governor had been in office. Governor Perry has a similarly impressive record. (An asterisk on the chart next to a candidate’s name means that our result has passed a standard test of statistical significance: We are 95 percent sure that it was not a matter of chance.)
Those with negative effects should not be too dismayed. No single statistical analysis can provide a complete picture of a candidate’s record as governor. New Jersey was a basket case long before Chris Christie took over, and was on a fiscal path similar to that of Greece. He probably did more for the state than this picture shows, because what would have happened without him would have been so terrible.
As with investments, past results do not guarantee future performance. But past results, analyzed well, certainly seem like a better guide to the future than mere speculation.