Is it all about competence? The human capital of U.S. presidents and economic performance

A “Yellow Dog Democrat” is a person who would vote for a yellow dog (listed as a Democratic candidate), rather than a Republican candidate.

An early version of this paper was presented at the 2008 meetings of the Public Choice Society, where several useful comments were received. Subsequent versions of the paper have been posted on Social Science Research Network website since 2009.

This slogan was used by the Republican Party during Hoover’s 1928 campaign for office along with “vote for prosperity.”

Even in this case, presidential competence may influence policy choices at the margin. In their empirical estimations, Alesina and Rosenthal include the “components of competence” in the error term, because these components cannot be observed separately by the voters. The variance/covariance structure is used to examine whether “a change in party control of the White House makes a difference in aggregate economic growth in the year following the election.” This model of partisan administrative policy effectiveness is rejected by their analysis (Alesina and Rosenthal 1995: 216).

See Bordo and Rockoff (1996) for an overview of modifications to the Gold Standard during the nineteenth and early twentieth centuries and of the effects that such policies had on international capital flows. See Cleveland and Powell (1909) for a thorough overview of state and national subsidies to US railroads.

A second reason to exclude the post 2000 observations is an apparent data entry error in the Mitchell series. Mitchell (2003), suggests that the average growth rate of GNP person during George H. W. Bush’s term was as high as 9.5 %. In contrast, the averaged growth rate of GDP per capita was only 0.7 %; hence, it is obviously a typographical error in Mitchell (2003). (If we include the entry for George H. W. Bush, the correlation between the Mitchell and Maddison per capita GDP series decreases to 0.8707.)

A regression between a president’s average growth rate of real per capita GNP on the previous president’s average in the 1870–2009 period yields: dY_t = 0.02747 − 0.059dY_t−1 with a t-statistic of 4.89 for the constant and of −3.23 for the coefficient. The coefficient should have been zero (statistically) if the dY series were stationary. The R² for this regression is 0.31 and the DW is 2.14. Because it is unlikely that a “competent” president is always followed by an incompetent one, the most likely explanation is that business cycles occur for exogenous reasons but around a classical long-run steady-state growth path. Note that without shocks, the fact that the coefficient is less than one implies that, in the long run, the series is asymptotically stable (stationary) and would converge to the constant term, 0.02747, which can be interpreted as the long-run equilibrium growth rate of classical models (here, 2.7 %/year).

Military governors included Andrew Jackson and Andrew Johnson.

Martin van Buren.

After 1870, observations share the score of 0 on secretary of state, so we omit this variable from the estimates in Table 4.

When we distinguish between “Yrs House” and “Yrs Senate” and re-run years 4, we observe that both variables have positive but insignificant coefficients. The t statistic equals to 1.31 for “Yrs House,” and 0.76 for “Yrs Senate.”

Among the many pooling problems are assumptions that implicitly require institutions to be stable in the period of interest. Otherwise, country-fixed effects will not capture all relevant institutional differences. Political and economic institutions, however, changed at different rates during the 50 year period after 1875. Moreover, some of those changes may reflect the same leadership characteristics focused on in the international studies focusing on economic growth and political success (Hayo and Voigt 2013).