ABSTRACT
This paper uses industrial level data from 21 developing and emerging economies over the period of 1995–2013, to analyze the impact of globalization, in particular, trade orientation of industries onto female employment share. The fractional probit estimation reveals that taking cumulative measures of export and import share often camouflages the impact of trade on female employment. The disintegration of export and import share according to their trading partners reveals that exports and imports from the developed world alone contribute to higher female employment. Moreover, it is the low-tech exports to developed countries and high-tech imports from developed countries which results in an increase in female employment. These findings call for the strengthening of trade ties with the developed world, especially when it comes to promoting low-tech exports and high-tech imports. Our results also reveal that the trading links with the developed world can further enhance female employment if developing country possesses a greater pool of educated female labor force.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 The unemployment differential across gender for year 2017 is computed by subtracting male unemployment rate (column [5] of Table ) from female unemployment rate (column [6] of Table ).
2 Although the concept of globalization is best applied at a country or regional level but this paper conceptualizes globalization at industry level whereby higher export and import orientation of industries is taken as a sign of being more globalized. In our dataset the correlation between industrial export and import share is +0.39 therefore negating the convention that industries with higher export share tend to have a lower import share or vice versa. This may be explained by the fact that in our dataset, industrial level imports not only include final products but also intermediary goods and services. Hence, theoretically a higher import share can exist alongside, either with higher or lower industrial exports. In short, our paper assumes that greater industrial ties with international market in terms of higher exports, imports or both are a sign of being more globalized. Another aspect of globalization is taken to be foreign direct investment (FDI). Since measure of FDI is not available at 4-digit industrial classification this variable is taken at country level therefore the regression results corresponding this variable need to be interpreted keeping this caveat in mind.
3 Usually firm-level data does not contain information about trading partners whereas country level data presents this information at a very aggregated level. The use of industrial level data is thus well suited to analyze whether trading partners matter in influencing the relationship between trade and female employment shares as information pertaining to trading partners is usually available at this level of aggregation.
4 No theoretical or conceptual difficulty arises if is alternatively taken to be a logistic distribution (Papke and Wooldridge Citation2008). However, probit function is preferred as it is computationally simple especially when handling unobserved heterogeneity and endogenous explanatory variables.
5 For detailed account of UNIDO statistical database refer to: http://stat.unido.org/
6 The detailed results for sector-wise female employment share is available on request.
7 As part of robustness test two alternative measures of female force participation rate (FLFP) are used, the results for which are presented in (refer to Appendix 3). These modified dependent variables are closer to the formal definition of FLFP which equals the number of female employed as a proportion of female population aged above 15 years. In order to construct FLFP at industry level, the first measure of FLFP equals number of female employed in industry i, located in country c at time period t divided by the total labor force measured at country level whereas the second measure uses country-wise measure of total female force in the denominator term. The coefficients of EXshare_developed i.e. 0.07 and 0.06 in columns (2) and (4) respectively, of support the notion that higher industrial exports to developed countries is related to higher FLPR whereas the contribution of imports to developed countries onto FLFP is insignificant. Surprisingly, the coefficient of imports from developing countries in columns (2) and (4) are significant at 10% and 5% respectively indicating that increased imports from developing countries tend to substitute local production thereby decreasing FLFP.
8 For each country and time period, industries are ranked according to their export intensity. The industries possessing the top 50 quantiles of export share are classified as comparative advantage industries.
9 The system-GMM approach simultaneously solves a system of two equations, one is the differenced equation and other is the level equation (Roodman Citation2009). The former employs instruments in level form whereas the latter uses instruments in difference form. The efficiency in the estimation is gained by generating more instruments for the endogenized variables.
10 Given the unavailability of appropriate external instruments, the lagged variables of the endogenous variables are used to instrument for the endogenous variables.
11 The strategy adopted by Fatima (Citation2017) has been used to classify the industries into high and low-technology sectors. Refer to Appendix 4 for classification of sectors into low- and high-tech.
12 The stata command ‘xtgee’ fits population-averaged panel data models by using GEE.