ESSAYS ON SPATIAL DIFFERENTIATION AND IMPERFECT COMPETITION IN AGRICULTURAL PROCUREMENT MARKETS

2020-07-29T01:23:36Z (GMT) by Jinho Jung

First Essay: We study the effect of entry of ethanol plants on the spatial pattern of corn prices. We use pre- and post-entry data from corn elevators to implement a clean identification strategy that allows us to quantify how price effects vary with the size of the entrant (relative to local corn production) and with distance from the elevator to the entrant. We estimate Difference-In-Difference (DID) and DID-matching models with linear and non-linear distance specifications. We find that the average-sized entrant causes an increase in corn price that ranges from 10 to 15 cents per bushel at the plant’s location, depending on the model specification. We also find that, on average, the price effect dissipates 60 miles away from the plant. Our results indicate that the magnitude of the price effect as well as its spatial pattern vary substantially with the size of the entrant relative to local corn supply. Under our preferred model, the largest entrant in our sample causes an estimated price increase of 15 cents per bushel at the plant’s site and the price effect propagates over 100 miles away. In contrast, the smallest entrant causes a price increase of only 2 cents per bushel at the plant’s site and the price effect dissipates within 15 miles of the plant. Our results are qualitatively robust to the pre-treatment matching strategy, to whether spatial effects are assumed to be linear or nonlinear, and to placebo tests that falsify alternative explanations.


Second Essay: We estimate the cost of transporting corn and the resulting degree of spatial differentiation among downstream firms that buy corn from upstream farmers and examine whether such differentiation softens competition enabling buyers to exert market power (defined as the ability to pay a price for corn that is below its marginal value product net of processing cost). We estimate a structural model of spatial competition using corn procurement data from the US state of Indiana from 2004 to 2014. We adopt a strategy that allows us to estimate firm-level structural parameters while using aggregate data. Our results return a transportation cost of 0.12 cents per bushel per mile (3% of the corn price under average conditions), which provides evidence of spatial differentiation among buyers. The estimated average markdown is $0.80 per bushel (16% of the average corn price in the sample), of which $0.34 is explained by spatial differentiation and the rest by the fact that firms operated under binding capacity constraints. We also find that corn prices paid to farmers at the mill gate are independent of distance between the plant and the farm, providing evidence that firms do not engage in spatial price discrimination. Finally, we evaluate the effect of hypothetical mergers on input markets and farm surplus. A merger between nearby ethanol producers eases competition, increases markdowns by 20%, and triggers a sizable reduction in farm surplus. In contrast, a merger between distant buyers has little effect on competition and markdowns.


Third Essay: We study the dynamic response of local corn prices to entry of ethanol plants. We use spatially explicit panel data on elevator-level corn prices and ethanol plant entry and capacity to estimate an autoregressive distributed lag model with instrumental variables. We find that the average-sized entrant has no impact on local corn prices the year of entry. However, the price subsequently rises and stabilizes after two years at a level that is about 10 cents per bushel higher than the pre-entry level. This price effect dissipates as the distance between elevators and plants increase. Our results imply that long-run (2 years) supply elasticity is smaller than short-run (year of entry) supply elasticity. This may be due to rotation benefits that induce farmers to revert back to soybeans, after switching to corn due to price signals the year the plant enters. Furthermore, our results, in combination with findings in essay 2 of this dissertation, indicate that ethanol plants are likely to use pricing strategies consistent with a static rather than dynamic oligopsony competition.