Lit in Review: Three on growth

Robert Barro, QJE, "Economic Growth in a Cross Section of Countries," ungated.
Method: cross-section, small multivariate, 98 countries, 1960-1985
Findings: Growth is unrelated to initial GDP/capita in simple regressions, but there is evidence of convergence once the schooling rate in 1960 is held constant. That is, a poor country grows faster than a richer one with the same level of human capital investment. This implies that poor countries will catch up (converge) to rich countries if they have high investment in human capital - ie, get their kids in school. Such countries also tend to have low fertility rates and higher levels of capital investment. There does not seem to be a correlation between growth and other public investment. Political instability is bad for growth and investment (surprise!). He notes there is still a lot of unexplained variation for Latin America and Africa.
Tie in for my research: Growth and investment are negatively related to the percent of GDP spent in government consumption. "An interpretation is that government consumption introduces distortions, such as high tax rates, but does not provide an offsetting stimulus to investment and growth."

Witold Henisz, 2000, Econ and Pol, "The Institutional Environment for Economic Growth." ungated.
Method: 1960-1995 in 5 year increments, OLS, GLS, GMM

Description: Henisz contends that the primary institutions investors care about involve a government's ability to commit to keeping in place policies favorable to investment. Once the investment is there, there is always the incentive to then tax it highly and investors lose. He identifies the number of branches of government with veto power, each having uniformly distributed preferences over a policy set. A country with a lot of veto points whose branches are at odds (controlled by different parties, e.g.) will not be able to change policies rapidly.
Findings: Reducing ICRG country risk by one standard deviation increases growth 1.3 percentage points. Increasing his measure of political constraints by one standard deviation would increase growth by between .5 and .9 percentage points, depending on regression technique.
"Odd" conclusion I draw: The reason the 1980s and 1990s were good for US growth was that Congress and the Presidency were in opposite hands. In the last decade, Reps or Dems controlled both and growth has been more tepid. As an "experiment," try electing Republicans to Congress in 2010 and see if growth improves.

Fiaschi and Lavezzi, 2007, J. Dev. Econ, "Nonlinear economic growth: Some theory and cross-country evidence."
Methods: Markov transition matrices and stochastic kernels, 120 countries 1950-1998, thresholds.
Findings: They break countries into four income groups and three growth groups [<0.5%, 0.5 to 2.5, >2.5%] to analyze what happens to growth in each segment. They estimate that as GDP/capita increases, the growth rate should decrease in the poorest segment, increase slowly in segment two, increase or remain high in segment there, and decrease again in segment four. They find evidence supporting convergence, but that it is a slow process: 110 years to converge, but there is only data for 48 years. While they do not find evidence favoring a poverty trap, poverty may be highly persistent.