David Maré (p)

This study estimates differences in productivity (mfp) across New Zealand urban areas, with a focus on the size of Auckland’s productivity premium. The estimates are based on analysis of firm-level data from Statistics New Zealand’s Longitudinal Business Database. The methods used in the paper overcome some of the biases that arise in standard approaches to spatial productivity estimation - biases arising from imperfect competition, spatial price variation, firm heterogeneity, and labour-sorting across cities. Ignoring these factors leads to biased estimates of the Auckland’s relative productivity performance. The study also investigates industry differences in spatial productivity patterns.

Hengky Kurniawan (p), Steven Poelhekke, Henri L.F. de Groot

This paper studies the effect of urban agglomeration on city productivity by using Indonesian data. The theory in urban economics says that worker productivity in a city will increase as an effect of agglomeration. However, the increase in productivity at some points in time will reach its peak and follows the inverted U-shape curve of the optimal level of agglomeration.
We test the relationship by using Indonesian data and employ the model developed by Au and Henderson (2006). We use non-agriculture output per worker as a measure for city productivities and number of worker as the measure for agglomeration. Some control variables are also used such as private capital stock per capita, foreign direct investment per worker and city’s market potential, and other variables that reflect the human capital and physical investment in a city. We use some instrumental variable to tackle endogeneity problem between city productivities and government policies, city governance and geographical location.
The preliminary results show that, for 95 cities, it follows the inverted U-shape but many of those interested variables are not statistically significant in predicting city productivities. The results show that about five percent of cities in Indonesia are too big, and more than half are considered too small. In the extreme cases, the cities with high concentration of manufacturing or service sector in the economy tend to be “too big”, for example city of Bontang with high manufacturing sector, and Jakarta as a service city. The results also show that capital per capita and foreign direct investment seems to have no significant effect in driving the worker productivity in Indonesian cities.
Moreover, we extend geographical unit by including districts (kabupaten) which is by formal definition is more rural than the city (kotamadya). In total, about 350 cities/districts are included in the estimation. However, including the districts in the sample does improve the significances of the interested variables but it does not follow the postulate of the inverted U-shape relation.
We conclude that the impact of urban agglomeration is high in determining the size of Indonesian cities, coupled with the high mobility of workers within the countries and less-strict policy in housing sector. Moreover, the attractiveness of urban life and the pressure from poverty in the less developed regions also play an important role in shaping the city size.

Fieke Visser (p), Philip McCann, Willem Jongman, Herman de Jong

see extended abstract