Convergence or Divergence?

(This writing is an Abstract and Conclusion part of my conference paper, presented in the 8th International Conference of IRSA/Indonesian Regional Science Association and the 31st Annual Meeting and Conference of FAEA/Federation of Asian Economic Association. Both were in 2007).

Abstract

This paper studies convergence process across cities in East Java province of Indonesia. To do so, it particularly investigates the effects of human capital and public capital on convergence trend. It also explores the geographic dimension of knowledge spillovers that is perceived as a key mechanism for convergence and a particular pattern of economic (industrial) activities. This paper provides some evidence on the trend of economic divergence. During the study period, divergence trend was associated with both economic liberalisation and decentralisation, while convergence trend shortly existed during economic crisis. Further, the process of convergence is conditional. The effectiveness of policy interventions that may be able to enhance convergence is also discussed in this paper.

Main Points

This paper provides some evidence on the trend of economic divergence. During the study period, divergence trend was associated with both economic liberalisation and decentralisation, while convergence trend shortly existed during economic crisis. Further, the process of convergence is conditional. Conditional convergence is basically equivalent to sustained differences in the levels of city real output or income per capita, as evidenced by estimates with LSDV model, as supported with spatial regression model, and as implied with the pattern of concentration and specialisation. While it is true that there is some degree of (global) spatial autocorrelation, it is not strong enough to leak out some growth - enhancing spillover effects, such as knowledge spillovers. Thus, knowledge that is important to improve productivity (that eventually leads to higher growth rate) will only be contained in some clustering cities which are already rich and densely populated, providing higher degree of proximity among economic agents and facilitating knowledge externalities. Analysis on industrial concentration and cities’ specialization suggests that large geographical concentration of manufacturing employment has remained and convergence will therefore be conditional.

Is there any means to soften concentration to promote convergence? Improving infrastructure in East Java’s periphery cities is often suggested as one of policy measures. However, whether the outcome of improved infrastructure will promote spatial diffusion or, instead, enhance concentration is not clear, as there exists two conflicting forces at work. As infrastructure is improved, foreign firms which exercise a relatively large export share in comparison to domestic firms will more focus on cheap production possibilities. This behaviour will have a positive effect on spatial diffusion. Yet, even among foreign owned firms is it only in a few manufacturing sectors that the bulk of the production is exported. Thus, it seems reasonable to expect that the effect from improved infrastructure will be ambiguous and not necessarily reduce concentration (Sjöholm, (…), unpublished). In line with that notion, a variable of public capital used in this paper, utilised as a proxy for infrastructure, shows negative or non-significant relationship with convergence process (see table 2 for unconditional convergence and table 3 for conditional convergence). One interpretation can be derived from these results is that the spending of public capital (on infrastructure) is not cost effective, suffering from poor design and planning, mark-up, and corruption.

Though the role of human capital in this study proved to be convergence enhancing, I think that politicians and public officials tend to prefer infrastructure to human capital projects, as the former more fits their political or financial interests. As development strategy, from political perspective, infrastructure projects better fit five years of political cycle. Infrastructure projects take shorter time and are more visible, which are politically useful for boosting the chance of being (re)elected. Thus, more balanced development strategy between infrastructure and human capital projects is certainly a sensible policy option to reduce concentration and to promote convergence.

Under regional autonomy regime at city level that has been just implemented in Indonesia for about four years, city governments in East Java province are likely tempted to pursue some sorts of territorial competition policyto affect firms’ location decisions, while at the same time, less efforts are devoted in developing other local development strategies. Imbalanced economic performance will be augmented with imbalanced local development strategy by each city governments. Both literature and empirical works warn that such a policy option may be unnecessary or, worse, act as a strategy of waste (e.g. Cheshire and Gordon, 1996; Cheshire and Gordon, 1998; Rodríguez-Pose and Arbix, 2001). Considering that East Java province has long been one of the Indonesia’s main destinations of manufacturing activities (Mackie, 1993) and the spatial patterns as observed in this study, such a territorial competition policy, especially the poorly designed one, may be unnecessary for ‘core’ cities along Surabaya-Malang corridor. As a worse case, it may bring financial and technical burdens for ‘periphery’ cities. Thus, for city governments, this study should shed a light that a policy of ‘doing nothing’ with regard to East Java industrial agglomeration economies should be regarded as, if not the best policy, one of the sensible policy alternatives.

Some places benefit from structures created in the past, through a cumulative process, which has led to agglomeration economies. Hence, a pessimistic opinion about the relevance of policies attempting to reduce disparities between locations is a consequence of the analysis. If history plays such an important role, concentration of economic activity in places where agglomeration has been a product of a long time period seems inevitable. Furthermore, does there exist a cumulative process through history, policies cannot affect much density and thus productivity (Irawan and Allegria, 2006).

However, it does not mean that a policy of ‘doing something’ to promote human capital, local firms, and business environment are not encouraged. Discussing the paper by Ciccone and Hall (1996), Irawan and Allegria (2006) suggest that attempts of local and regional governments in order to increase productivity should be conducted to increase density. If concentration and proximity of workers is important, development of employment clusters or improvements of the city business centre are relevant to obtain productivity gains. In addition, policies satisfying the demand as well as reducing the price of commuting and the costs of doing business should contribute to the recreation of employment, reinforcing business and workers concentration, and agglomeration economies. At this point, the differences of underlying condition for convergence can be gradually reduced. Therefore, ‘the membership’ of convergence club can be expected to slowly expand.