Grease or sand the wheel? The effects of individual bribe payments on aggregate productivity growth

While the Asian Paradox literature evidences a grease-the-wheel effect of corruption on individual firm productivity growth, the results call into question the aggregate effect of individual bribery. A large literature analyzes the drivers of productivity growth at the firm and sector levels, and suggests that corruption may affect these drivers through the resulting less favorable conditions for investment and technological progress, as well as distortions and misallocation in output and inputs markets. We are not aware of any literature that investigate the effects of micro-level corruption on aggregate productivity growth (APG) and its different components (intra-plant productivity growth; market shares reallocation between incumbents; and the effect of firm entry and exit in the industry). Filling this gap is important for understanding the dynamics by which corruption damages countries’ economic performance (Dal Bo and Rossi 2007) as measured by its main driver: productivity growth (Barro and Sala-i-Martin 2003). Using a unique panel dataset covering the population of Indonesian manufacturing establishments over the long period 1975-1995, we assess the cumulated effect of micro-bribe payments to aggregate productivity growth. Following Foster et al. (2001) methodology so as to model and decompose Aggregate Labour Productivity Growth (ALPG), and using establishment-level measures of productivity and bribe payments, we assess the contribution of three different corruption-classes of plants: plants paying zero, low, or high bribes. As a robustness check, and to account for the overestimation of the net entry effect, we use the Melitz and Polanec (2012) modeling approach that is a dynamic version of Olley and Pakes (1996).We demonstrate that the aggregate class of high-corruption plants contributes negatively to aggregate labour productivity growth, while the aggregate class of plants paying zero or low bribes contributes positively. Our results are robust to the use of two different measures of corruption. Thus, this paper shows that the two diverging effects of corruption, that is the grease- and the sand-the-wheel effects, coexist at different aggregation levels and for different classes of plants. While corruption has beneficial effects if kept low, the cost of high corruption in terms of aggregate productivity growth becomes evident.