Notes: In column (1), the unit of observations is the leader–year. The sample is restricted to those who assume provincial leadership positions more than once. The definition of variables is the same as in column (4) of Table 2. Standard errors clustered at the province level are reported in parentheses. Reported in brackets is the |$p$|-value for the significance of the coefficient on Connection * Growth by wild cluster bootstrap-|$t$| (Cameron, Gelbach, and Miller 2008). Dummies of the number of years in office (two to twelve), office-by-province fixed effects, and office-by-year fixed effects are controlled for. In column (2), the leadership spell is the unit of observation. The sample is restricted to those who serve the second or third term of provincial leadership. The dependent variable is average annual growth during the tenure conditional on province and year fixed effects in a balanced province panel regression. Connection is an indicator of being connected in the last year of the previous term. Growth is the deviation of average conditional annual growth during the previous term from the sample average. No other variables are included as regressors. Robust standard errors are reported in parentheses.

Significant at 10%

significant at 5%.