Regulatory Distortions and Capacity Investment: The Case of China’s Coal Power Industry (Job Market Paper)
China's electricity generation market is largely state planned, and planners allocate production to different power plants according to an extremely strict and complex regulatory structure. However, China also encourages private entry and investment in capacity in this market, which in turn very strictly determines the amount of electricity a plant can produce. Using a novel dataset on the industry, this paper estimates the effect of China's electricity planning policies on these investment incentives via a structural model. In a market where production is fully allocated by planning authorities, forward-looking plants must anticipate their expected stream of production (which they do not control) in order to make optimal investment decisions. My empirical strategy consists of two parts: First, I estimate the amount of misallocation each plant is subject to in production nonparametrically using a novel discrete choice model of planners' behavior in each market. I then develop a dynamic discrete choice model of capacity investment with a unique combination of traits: plants forecast their expected profits under these distortions in a nonstationary environment, with unobserved heterogeneity in investment costs, and market-level heterogeneity in returns to investment. All of these components are necessary to capture the complex and constantly evolving policy environment different plants face across China. I find that aggregate electricity demand could be met at a roughly 3\% lower cost per unit under the observed power infrastructure if plants were assigned production purely based on costs, which comes with significant consequences for carbon emissions. The dynamic model reveals that the investment response of a forward-looking plant following a change in (residual) production allocation is more than triple that of a comparable myopic plant. Additionally, persistent changes carry large absolute effects: a persistent reduction in a plant's residual allocation by one standard deviation decreases the chance of making major investments by 19-25\%, which comes with significant cost consequences for an affected plant. These are the first estimates of any kind of investment elasticity in this market, as well as the first to relate output misallocation to investment in a power market. Counterfactual simulations demonstrate that current policies are aggressively flattening this market's plant size distribution, which is consistent with planners concerned about market concentration in the event of electricity market deregulation. These findings help to illustrate the tradeoffs between efficiency and other considerations faced by planners in developing electricity markets.
Restructuring the Chinese Coal Power Market: Revenue vs. Physical Measures (draft available upon request)
This paper measures the effect of a major 2002 restructuring of the Chinese coal power market using newly available physical data on the industry. Two effects are investigated: individual firms growing more or less technically efficient on the intensive margin in response to the policy, and market allocation mechanisms allocating across firms more efficiently given efficiency distributions. To investigate the former I use a novel dataset and a difference-in-differences framework in the spirit of (Fabrizio et al., 2007) and (Gao and Van Biesebroeck, 2014). I find this measures are extremely sensitive to whether physical or revenue-based measures of efficiency are used, in some cases nullifying the positive results of (Gao and Van Biesebroeck, 2014) I find that the reforms also did little to change the optimality of input allocations across plants in the market by either measure.
Heterogeneous Technologies, Productivity and the State Sector in China (joint with Panle Jia Barwick, Shanjun Li, and Yifan Zhang)
Over the past 20 years, the manufacturing sector in China has seen both massive growth and deregulation. This has coincided with a few other well-established facts: First, this growth is due in large part both to an increase in the number of firms in the market, as well as an increase in average firm-level efficiency, as measured by total factor productivity. Second, firms controlled by the Chinese government (SOEs) generally have lower average TFP than private firms. And, third, the main distinction between SOEs and privately owned firms in the modern era is that SOEs are granted favorable access to capital markets. Analyses that incorporate these facts generally ignore that favorable access to capital markets may induce fundamentally different choices of technology across the two types of firms. This paper seeks to incorporate and estimate the differences in technology between SOEs and private firms using structural techniques, and assess the impact of incorporating this difference on standard TFP results in China. The evidence strongly favors that SOEs adopt significantly more labor-intensive technologies, and that this has implications for several key results on productivity, including (mis)allocation levels and TFP residual dispersion.