Publications  Income Distribution


Hypothesis 3:

The Effect of Socio-political Systems


 In all regressions, the most consistently significant group of variables are the three that reflect socio-political systems. It is significant in spite of the fact that it enters the regression late. This result is primarily due to the two dummy variables which are individually significant: for East European countries and societies with socio-political dualism

(a) The East European countries for which we have data have an unusually egalitarian income distribution (Gini coefficient around .20 to .25) and a reasonably high per capita income (most are around $600—$800). They, therefore, do not fit well the Kuznets Curve, and in all regressions are significantly different from the rest of the observations. Most econometric analyses of the Kuznets Curve have taken account of this factor by excluding completely or through the use of a dummy variable the Communist countries. The dummy variable in most regressions shows a Gini for the few East European countries lower by .12, a major difference.

It is difficult to determine to what extent income distribution in these coun­tries is in fact more equal because most property income accrues to the State and to what extent recorded equality is due to different statistical conventions and the failure to take full account of income in kind and fringe benefits for the elite (better health services, special stores, cars, etc.).


(b) A dualistic socio-political structure, with an elite drawn from a different ethnic or racial group than the poor majority, makes for an unequal income distribu­tion. One would generally expect a “foreign” elite to be more forceful and blatant than one in a more homogeneous society in using its political power to ensure that it obtains a disproportionate share of income. These countries often also have a dualis­tic economic structure. Since economic dualism is usually defined in terms of a great gap between a high-productivity, high-income, modern sector and a low-productivity, low-income, traditional sector, one would expect to find a high correlation between dualism and inequality, but only because the same phenomenon is observed in both cases. Socio-political dualism differs analytically from economic dualism, It refers to countries dominated by a foreign elite, generally of European origin, with the majority of different racial or ethnic backgrounds. Examples include South Africa and Rhodesia; some countries of French West Africa before or shortly after indepen­dence; those countries of Latin America where the elite is of European origin, the majority Indian, Mestizo and Black; and a few other countries in Asia and Africa before or shortly after independence, before the indigenous elite took over real control. (See Statistical Appendix for full list.) The effect of socio-political dualism has been examined before (Papanek  and Bacha, ) but the variable has not been clearly defined or rigorously tested in a multi-factor analysis.

In all regressions of Tables 2 and 3, dualistic societies are significantly less egalitarian but the t-statistic ranges from 2.1 to 1.6. The latter, implying very weak significance, is in the Gini coefficient regression with regional dummies. Since dual­istic societies are concentrated in Sub-Saharan Africa and Latin America, the dual­istic variables lose statistical significance, but not necessarily real economic signifi­cance, when regions are introduced into the regressions: it may not be location in a region that causes inequality, but socio-political dualism which happens to be especi­ally pronounced in two regions.

Moreover, when the share of the poorest 40 percent is the dependent variable, the t-statistics for dualism are only 1.7 to 1.8, also indicating relatively weak signifi­cance. But this appears to be due in large part to the fact that the socio-political variables enter the regression quite late. In an earlier version of the work (Papanek and Kyn ) these variables entered earlier and the t-statistics ranged up to 2.6. They lost considerably in importance and significance (coefficients and t-statistics decline) when education was introduced in the regression and again when the structural variables were added, among which the importance of primary exports is the only significant one. A plausible explanation is that dualistic societies affect income distribution in part through indirect means, including education and the allocation of the concentrated resources generated by primary exports or primary production for the domestic market. Education is largely limited to the elite in dualistic societies. Concentrated (rental) income from primary exports can more easily be retained by the elite than more diffuse income from manufactured exports or primary production for the domestic market. Because of collinearity, the variable for dualistic societies loses statistical significance in the complete regression, but it probably does not lose real economic significance. The most reasonable explanation of the statistical results is that some of the other variables represent instruments through which dualistic societies achieve their implicit objectives.

It should be emphasized that these two variables for East European and socio-politically dualistic countries explain a great deal of the variance in income distribution. Since neither is causally related to per capita income, they have little predictive value for other countries’ income distribution as per capita income rises. As poor countries such as Tanzania, Bangladesh, India and Sudan reach the per capita income of the East European or dualistic countries, they will not as a result develop the more egalitarian income distribution of Eastern Europe or the unequal income of countries with an elite that is of different ethnic background from the majority. This is a case where cross-section data cannot be used to predict developments over time. Of course, the model has some predictive value in another sense: a change in socio­political systems is likely to lead to a change in income distribution. For instance, if a dualistic society changes to one in which the elite is drawn from the same ethnic group as the majority, we would expect income distribution to become more egalitarian.



(c)      The role of government in the economy, unlike the other two socio­political variables, often changes over a relatively short period of time. Other studies (Adelman and Morris , Papanek ) have concluded that economies are egali­tarian if government intervenes more extensively in the economy. Indeed, one of the primary justifications for government intervention is to improve equity. However, our results do not support the hypothesis that greater government intervention increases equality, at least when such intervention is measured by the share of public investment. The signs of the coefficients are almost universally in the right direc­tion negative for Gini, positive for the share of the poorest 40 percent but the coefficients are very low. As the percentage of public investment increases from around 10 percent or 20 percent (Philippines, U.S.A., Lebanon) to 50 percent (Pakistan, Gabon, Taiwan, Sweden, India) the Gini would drop by only .02 and the share of the poorest 40 percent would increase by less than .04 percent. Even a radical shift from 100 percent private investment to 100 percent public investment would increase the share by less than 0.1 percent. Moreover, the coefficients are universally not significant, as indicated by t-statistics (even when the variable enters the regression very early), although they are again quite stable.

How does one explain the observation that greater government intervention does not significantly affect income distribution? One possible explanation is that we have chosen an inappropriate proxy variable, that a substantial proportion of inter­ventionist governments do not control investment and vice versa, that a substantial proportion of laissez-faire governments actually control a good deal of investment. But an examination of the list of countries and their ranking on this variable a small sample was given above makes that an implausible explanation. A more likely explanation (see Papanek ) is that interventionist or popu­list governments often intervene on behalf of a different part of the elite, not of the poor. Their intervention benefits some of the political, bureaucratic and military leadership, often the workers in the public enterprises and the businessmen who receive government patronage. The costs are borne by the landless rural workers, the casual urban workers, and sometimes a landed or business elite out of political favor. While evidence for this hypothesis comes largely from Southern Asia, we suspect it applies more widely. Clearly, the effect of government intervention on growth and equity in mixed economies warrants further study.




OK Economics was designed and it is maintained by Oldrich Kyn.
To send me a message, please use one of the following addresses: ---

This website contains the following sections:

General  Economics:

Economic Systems:

Money and Banking:

Past students:

Czech Republic

Kyn’s Publications

 American education

free hit counters
Nutrisystem Diet Coupons