Economists have long interpreted protectionist devices such as anti-dumping and countervailing duties as either practices facilitating collusion or simply as rent-extraction mechanisms. But these monopoly-oriented public choice interpretations do not explain why this protection is observed in some industries, but not others.
This research develops and tests a theory of selective protection designed to fit the characteristics of the industries to which that protection is applied. In particular, we argue that industries requiring large fixed investments in capacity and for which significant costs can be avoided by shutdowns are susceptible to empty cores during periods of slack demand.
Our hypothesis has an intuitive appeal to the U.S. steel industry because steel production needs large avoidable costs and the steel industry had been getting along with various protectionist devices through 1970 to 1990. To formalize the idea, we employ a model designed to fit the cost structure having avoidable fixed costs, and the finely divisible demand structure. The core status in two cooperative games is analyzed, given the trade policy, free trade or voluntary restraint agreements (VRAs). We determine that in the event of low demand, the core can be empty under free trade but can be restored by VRAs.
We test for empty cores in the U.S. steel industry by using time varying parameter estimation. This technique employs a Kalman filter in the context of a particular form of state space model. The main focus of the test is on the market structure parameter used to identify periods of empty cores by the time-varying estimation. The market structure parameter is identified from two reduced form equations of demand and supply by using monthly time-series data. These results show that while the steel industry suffered a major downturn in the early 1980's (1981-1982), the downturn appears to have generated large numbers of countervailing duty and anti-dumping filings and thus the core seemed to be prevented from being empty. Rather, this research found that the core of the market was empty during 1983 for several reasons.
We test the international collusion hypothesis by using the event study methodology (1984 VRAs event). We have three sample categories - U.S. steel producers, U.S. steel consumers, and Japanese steel producers. We have empirical results which are against the international collusion hypothesis, but favorable to our empty core hypothesis.