These assumptions, in context with the earlier 

 discussion of fixed asset theory, imply a statis- 

 tical model wherein changes in resource quant- 

 ity (labor or vessels) are regressed on changes 

 in salmon landings, lagged by one year. Ideally, 

 the influence on resource use levels of acquisi- 

 tion prices and salvage values of the produc- 

 tive factors should also be taken into account. 

 In that these data were not available, the units 

 of observation were defined both cross-section- 

 ally and over time. Specifically, yearly data 

 between 1957 and 1966 for each of ten NMFS 

 statistical regions on the Pacific Coast were 

 used.'- This yielded a total of 80 observations 

 and allowed us to take into account, in a rough, 

 implicit fashion, cross-sectional differences 

 which might give rise to a variety of deviations 

 between acquisition prices and salvage values. 

 The statistical model is as follows: 



(1)X 



(' + 1) 



/ [^(o< '^'(" I )' ^u* I). 0\ 



where 



x' = index of fishing effort (number 

 of fishermen and net tonnage 

 of fishing vessels), 



L = index of salmon landings 

 (pounds), 



C = cyclical nature of the fishery 

 (dummy variable: 1 for all ex- 

 pected good runs, whether or 

 not they actually materialized, 

 and for all expected poor runs), 



U = unemployment rate of the civil- 

 ian labor force in the major 

 labor market, 



D = distance from the center of sal- 

 mon fishing activity in the 

 region to the nearest major 

 labor market. 



In order to test for symmetry between exit 

 and entry relationships with this model, the 



'- The regrions were Southeastern. Central, and West 

 em Alaska; Puget Sound and Coastal in Washington; 

 Columbia River in Washingrton and Oregon; Coastal 

 Oregon; and Northern, San Francisco and Monterey 

 in California (U.S. Department of the Interior, 1947- 

 1967). 



80 observations were divided into two subsets. 

 One subset, with 42 observations, consisted of 

 those years in which landings had increased 

 over the preceding year. Given our assump- 

 tions of an elastic product demand, fixed factor 

 supply (in the short run), and rent expecta- 

 tions, it follows that these observations repre- 

 sent years in which the MVP schedule of factors 

 had shifted to the right and was expected to 

 remain there, ceteris paribus. Similarly, the 

 35 observations''' in the second subset repre- 

 sented years in which MVP had shifted left- 

 ward. 



Fixed asset theory would suggest that aggre- 

 gate factor levels in an industry would either 

 increase or remain constant following years 

 of increased landings, and would either de- 

 crease or remain constant following years of 

 reduced landings. Table 5 indicates a definite 

 asymmetry between entry and exit relation- 

 ships. For example, in years of increased land- 

 ings, the index of vessel inputs in year t + 1 

 increased 0.32 per unit increase in the land- 

 ings index for year f. The coefficient for years 

 of decreased landings was very slightly nega- 

 tive, but not significantly different from zero. 

 Asymmetry is strongly suggested by the fact 

 that the Bi coefficients for the two subsets are 

 significantly different from each other in both 

 the labor and vessel equations. 



This ratchet mechanism is illustrated in 

 Figure 4. Net entry follows years of "good 

 catches," but net exit does not occur following 

 the "bad years". This is not hard to imagine 

 for specialized trolling vessels which may have 

 low salvage values outside of fishing or even 

 in other segments of the salmon fishery. It 

 is somewhat more difficult to rationalize, on 

 the other hand, for the labor resource, although 

 the human resource would no doubt be af- 

 fected by lack of mobility of the capital 

 resource. 



The relationships of resource use levels to 

 the other variables in the analysis are also 

 of interest, and are summarized here: 



(1) Expectations of cyclical runs in encour- 

 aging entry were more important follow- 

 ing years of declining landings than fol- 

 lowing years of increased landings. This 



'■' Three observations were not usable due to lack 

 of a "bench mark" year. 



143 



