Table 4.
Robustness checks for alternative GDP per capita cut-offs for defining middle-income counties (Hypotheses 1.2, 2.2–2.4), OLS models.
| (1) | (2) | (3) | (4) | (5) | |
|---|---|---|---|---|---|
| ≥1000 | ≥2000 | ≥3000 | ≥4000 | ≥5000 | |
| Recession | 0.098*** | 0.051*** | 0.041*** | 0.046*** | 0.049*** |
| (0.005) | (0.003) | (0.003) | (0.004) | (0.004) | |
| RecessionINT | −0.159*** | −0.149*** | −0.172*** | −0.230*** | −0.162*** |
| (0.006) | (0.007) | (0.007) | (0.008) | (0.010) | |
| Service | 0.085*** | 0.087*** | 0.086*** | 0.093*** | 0.091*** |
| (0.003) | (0.002) | (0.002) | (0.002) | (0.002) | |
| ServiceINT | −0.005 | −0.006* | −0.067*** | −0.109*** | −0.109*** |
| (0.003) | (0.003) | (0.004) | (0.005) | (0.006) | |
| Cars | 0.161*** | 0.135*** | 0.138*** | 0.121*** | 0.098*** |
| (0.007) | (0.004) | (0.003) | (0.003) | (0.002) | |
| CarsINT | −0.086*** | −0.074*** | −0.129*** | −0.118*** | −0.083*** |
| (0.007) | (0.005) | (0.004) | (0.005) | (0.006) | |
| Working | −0.035*** | −0.018*** | −0.022*** | −0.031*** | −0.031*** |
| (0.002) | (0.001) | (0.001) | (0.001) | (0.001) | |
| WorkingINT | 0.002 | −0.014*** | −0.012*** | −0.013*** | −0.007** |
| (0.002) | (0.002) | (0.003) | (0.003) | (0.004) | |
| Total kcal (1000), pc | 0.001 | 0.043*** | 0.076*** | 0. 174*** | 0. 207*** |
| (0.006) | (0.004) | (0.039) | (0.004) | (0.003) | |
| Total kcal (1000), pcINT | 0.21*** | 0.223*** | 0.135*** | 0.083*** | −0.064*** |
| (0.007) | (0.006) | (0.005) | (0.006) | (0.008) | |
| TV sets | 0.145*** | 0.135*** | 0.138*** | 0.149*** | 0.150*** |
| (0.004) | (0.002) | (0.002) | (0.002) | (0.002) | |
| TV setsINT | 0.006 | 0.017*** | 0.011*** | 0.007 | −0.006 |
| (0.004) | (0.003) | (0.003) | (0.004) | (0.005) |
Notes: Cluster-robust standard errors in parentheses (based on DHS cluster identifiers). All specifications are OLS models. Note that OLS version was chosen over country fixed effects, as there was insufficient number of countries to estimate several parameters in columns 4 and 5 in CFE version.
p < 0.1.
**p < 0.05.
p < 0.01.
All specifications contain time dummies, as well as the same control variables as in Table 1.
Each column presents results for alternative cut-offs of middle income countries. Each row represents main parameter only, for different levels of GDP per capita (purchasing power parity, constant 2005 dollars) reflected in columns. The INT subscript refers to the interaction between the main parameter of interest and the dummy for the respective income threshold level. Total calories and recession are country-level covariates, the rest are individual level.
For example, consider the effect of working on the probability of being above normal weight, for the case when middle income countries are defined as having GDP per capita, PPP of 2000 dollars or more. The results are presented in column 2. For women living in low income countries, being employed reduces the probability of being above normal weight by about 1.8 p.p. In the middle income countries, the effect of being employed is also negative, and is (statistically significantly) stronger than in the low income countries, by about 1.5 p.p.
Source: DHS dataset, 1991–2009.