Table 6.

Robustness to Income per Capita Data Quality Concerns

	(1)	(2)	(3)	(4)	(5)	(6)
	Weighted OLS	Weighted OLS	Weighted OLS	Weighted OLS	Weighted OLS	Weighted OLS
	Observations Weighted According to:
	Income Data Frequency			Total Population Size
	Dependent Variable is Log Income per Capita in:
	1500 CE	1000 CE	1 CE	1500 CE	1000 CE	1 CE
Log Years since Neolithic Transition	0.173 (0.162)	0.122* (0.063)	0.189 (0.121)	0.278 (0.171)	0.143* (0.068)	0.289 (0.175)
Log Land Productivity	0.039 (0.023)	−0.045* (0.022)	0.008 (0.031)	−0.005 (0.026)	−0.062* (0.030)	−0.011 (0.027)
Log Absolute Latitude	−0.042 (0.080)	0.205* (0.108)	−0.442 (0.362)	−0.089 (0.052)	0.298*** (0.031)	0.080 (0.089)
Mean Distance to Nearest Coast or River	0.219 (0.202)	−0.370** (0.148)	0.139 (0.298)	0.332** (0.148)	−0.592*** (0.108)	−0.180 (0.189)
Percentage of Land within 100 km of Coast or River	0.153 (0.169)	−0.228 (0.137)	0.159 (0.257)	0.329 (0.227)	−0.477*** (0.122)	0.003 (0.277)
Continent Dummies	Yes	Yes	Yes	Yes	Yes	Yes
Observations	31	26	29	31	26	29
R-squared	0.54	0.79	0.29	0.74	0.83	0.45

Summary – This table demonstrates that the relatively small effects of land productivity and the level of technological advancement, as proxied by the timing of the Neolithic Revolution, on income per capita in the years 1500 CE, 1000 CE and 1 CE remain robust under two different weighted regression methodologies, designed to dispel concerns regarding the quality of the historical income per capita data series.

Notes – (i) log land productivity is the first principal component of the log of the percentage of arable land and the log of an agricultural suitability index; (ii) the weight of country i in regressions (1)–(3) is inversely proportional to the frequency with which i’s income per capita occurs in the corresponding samples, i.e., $w_i=n_i^{-1}/{\sum }_i{n}_i^{-1}$, where n_i is the number of countries with income per capita identical to i; (iii) the weight of country i in regressions (4)–(6) is directly proportional to the population size of i in the corresponding samples, i.e., w_i = p_i/Σ_i p_i, where p_i is the size of the population of i; (iv) a single continent dummy is used to represent the Americas, which is natural given the historical period examined; (v) regressions (2)–(3) and (5)–(6) do not employ the Oceania dummy due to a single observation for this continent in the corresponding regression samples, restricted by the availability of income per capita data; (vi) robust standard error estimates are reported in parentheses; (vii) *** denotes statistical significance at the 1 percent level, ** at the 5 percent level, and * at the 10 percent level, all for two-sided hypothesis tests.