Human Capital Accumulation in Pakistan in the Light of Debt, Military Expenditure and Politics

Abstract

We investigate factors responsible for low public sector human capital investment in Pakistan. The debt servicing burden coupled with a rising debt stock can impact on human capital expenditure. To account for this, a Debt Net Cost Index is developed to measure the evolving net cost of public debt starting in 1960. Political factors examined are regime type, frequency of elections, quality of democracy, international aid preferences, elite capture in terms of industrial concentration, military burden and external hostility indices. We find that for the period as a whole from 1960 onwards, political factors dominate economic explanations such as the burden of debt servicing in accounting for low levels of human development expenditure. Only during episodes of civilian rule do economic factors in the form of the debt servicing burden become salient. This is because civilian governments are often saddled with inherited debt from earlier military rulers. Pakistan's military regimes have been less resource constrained because of greater external flows, allowing them to spend more on everything, but the rise in human development expenditures was less than proportional to the rise in available resources.

Key words:

• Human capital
• Public debt
• Industrial elite
• Military expenditure
• Pakistan

Acknowledgements

We are grateful to an anonymous reviewer and an associate editor of this journal for helpful comments on a previous draft of this paper. Muhammad Saleh acknowledges the financial support of the Netherlands Fellowship Programme for financing his study at the International Institute of Social Studies where this research was conducted.

Notes

1 Such as the Human Development Indicator stated by the United Nations Development Programme, which is the unweighted average of per-capita income and composite education and health indicators.

2 See Bloom and Canning (2000) on the importance of health, and Baldacci et al. (2008) as an example of the copious literature on the importance of skills and education for growth.

3 [http://hdrstats.undp.org/en/countries/profiles/PAK.html], accessed 26 July 2012.

4 Two political parties have largely been heading the democratically elected civilian parliamentary system in Pakistan since 1972: Pakistan People's Party (PPP) and Pakistan Muslim League (PML).

5 Up to a point, military establishments offer employment and training to a selected segment of the military. Similarly, military research and development can have positive technological spillovers

6 One can imagine that in a true democracy the preferences of the median voter determine public policy (Downs, 1957), or that the social planner is utilitarian, maximizing the greatest good of the greatest number. This is, however, unlikely to be the case in factional and clientist states, or where mass poverty is endemic leading to the effective disenfranchisement of the poor.

7 Based on our DNI, we ran a simple regression of DNI on X_t, which shows how a percentage change in human capital spending can influence the net cost of public debt (see Table C2 in Appendix 3). A positive and significantly large coefficient justifies our use of predicted values of X_t as a proxy for measuring DNI. This also shows that any increase in X_t will cause a net welfare gain as a result of debt accumulation in Pakistan.

8 We found insignificant differences in our parameter estimates when, instead of DNI, all of the explanatory variables listed in Equation (8) were included directly into our VAR model with the whole sample. The results show that only domestic public debt positively affects human capital investment and ODA becomes an insignificant determinant of public sector human capital spending. The rest of the estimates including the political variables are not significantly different from our VAR estimates with the DNI. These results therefore justify the use of DNI as a proxy for measuring debt net cost first and then including it in the VAR model (see Appendix 3 for details of test results).