24.07.2014 Views

The Long-Term Effects of Africa's Slave Trades - Social Sciences ...

The Long-Term Effects of Africa's Slave Trades - Social Sciences ...

The Long-Term Effects of Africa's Slave Trades - Social Sciences ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Table 10: Robustness and Sensitivity Checks. Dependent variable is ln real<br />

per capita GDP in 2000.<br />

Without colonizer fixed effects With colonizer fixed effects<br />

Omitted Observations coef s.e. N R 2 coef s.e. N R 2<br />

None −.108 ∗∗∗ (.025) 52 .27 −.108 ∗∗∗ (.024) 52 .49<br />

South Africa −.105 ∗∗∗ (.025) 51 .26 −.104 ∗∗∗ (.024) 51 .49<br />

North African countries −.097 ∗∗∗ (.027) 47 .23 −.096 ∗∗∗ (.025) 47 .47<br />

Island countries −.104 ∗∗∗ (.026) 47 .26 −.110 ∗∗∗ (.025) 47 .55<br />

All <strong>of</strong> the above −.078 ∗∗∗ (.028) 41 .17 −.079 ∗∗∗ (.027) 41 .53<br />

Zero export countries −.139 ∗∗∗ (.037) 39 .28 −.087 ∗∗ (.043) 39 .51<br />

|ê i| > 2.0 −.103 ∗∗∗ (.022) 47 .33 −.109 ∗∗∗ (.021) 48 .56<br />

|ê i| > 1.8 −.106 ∗∗∗ (.021) 46 .36 −.109 ∗∗∗ (.027) 48 .56<br />

|ê i| > 1.5 −.096 ∗∗∗ (.021) 45 .33 −.101 ∗∗∗ (.020) 46 .58<br />

|ê i| > 1.0 −.099 ∗∗∗ (.018) 35 .48 −.102 ∗∗∗ (.016) 35 .79<br />

dffits −.105 ∗∗∗ (.021) 44 .37 −.108 ∗∗∗ (.022) 49 .53<br />

Cook’s Distance −.108 ∗∗∗ (.023) 48 .32 −.108 ∗∗∗ (.024) 51 .47<br />

Welsch Distance −.088 ∗∗∗ (.024) 50 .22 −.108 ∗∗∗ (.024) 51 .47<br />

Notes: <strong>The</strong> dependent variable is the natural log <strong>of</strong> real per capita GDP in 2000. <strong>The</strong><br />

colonizer fixed effects are indicator variables for the identity <strong>of</strong> the colonizer at the time<br />

<strong>of</strong> independence. Coefficients are reported with standard errors in brackets. ∗∗∗ , ∗∗ , and ∗<br />

indicate significance at the 1, 5, and 10 percent levels. Influential variables were omitted<br />

using the following standard rules. dffits: omit if dffits i > 2 √ k/n. Cook’s distance:<br />

omit if Cook’s distance > 4/n. Welsch distance: omit if Welsch distance > 3 √ k. n is the<br />

number <strong>of</strong> observations and k is the number <strong>of</strong> independent variables.<br />

39

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!