Skip to main content
. Author manuscript; available in PMC: 2025 Jun 1.
Published in final edited form as: World Dev. 2024 Feb 14;178:106575. doi: 10.1016/j.worlddev.2024.106575

Table 1:

Estimated Macroeconomic Return to Health

Dependent variable Growth rate of income per capita
Baseline controls Including working experience Including income inequality Including lagged controls Country growth trends Panel GMM
(1) (2) (3) (4) (5) (6)
Log income per capita (t − 1) −0.070***
(0.009)		 −0.071***
(0.009)		 −0.073***
(0.011)		 −0.158***
(0.014)		 −0.270***
(0.034)		 −0.060***
(0.023)
Growth of capital per worker 0.344***
(0.049)		 0.355***
(0.048)		 0.284***
(0.058)		 0.343***
(0.059)		 0.257***
(0.082)		 0.364***
(0.051)
Change in adult survival 0.698***
(0.258)		 0.899***
(0.238)		 0.792***
(0.301)		 0.694***
(0.253)		 0.394*
(0.220)		 0.653**
(0.271)
Change in secondary schooling 0.068***
(0.024)		 0.065***
(0.024)		 0.057**
(0.023)		 0.030
(0.023)		 0.049*
(0.026)		 0.062**
(0.028)
Growth of working-age population/total population 1.043***
(0.306)		 1.051***
(0.345)		 0.989**
(0.405)		 0.983***
(0.291)		 0.231
(0.415)		 0.968***
(0.361)
Countries 133 131 131 133 133 133
Observations 1020 948 731 1020 1020 1020
R 2 0.29 0.30 — 0.34 0.33 —
AR(2) p-value — — — — — 0.10
Hansen p-value — — — — — 0.15
Diff.-in-Hansen p-value — — — — — 0.97

Note: Estimation results for five-year panels of 133 countries over the period 1965–2015. Estimates are derived from ordinary least squares in specifications (1) to (5) and system GMM in (6). All specifications include time fixed effects, controls for lagged years of secondary schooling, and quality of economic institutions. Specification (2) includes controls for experience and experience squared; specification (3) includes the pre-tax, pre-transfer Gini coefficient; and specification (4) includes lagged controls for physical capital per worker, population health, and the size of the working-age population relative to the total population. Specifications in (5) and (6) account for country-specific growth trends by including country fixed effects. The panel GMM specification in (6) uses the first lag of the endogenous variables in the difference equation and the first difference of the endogenous variables in the level equation as instruments; standard errors in this specification are computed with the two-step procedure and corrected with respect to finite sample size (Windmeijer 2005). Standard errors are clustered at the country level and reported in parentheses. Asterisks indicate significance levels:

*

p < 0.1;

**

p < 0.05;

***

p < 0.01.