probability of remaining in the largest state for a 

 long period of time permit large firms to be more 

 secure and ultimately more stable than small 

 firms. Thus, large firms develop greater access to 

 raw supply sources which are currently scarce, and 

 greater knowledge of the national market accom- 

 panied by stability in supplying their customers. 



Small firms, being able to enter with relative 

 ease, find it very difficult to advance in size but 

 remain in their state without too much difficulty. 

 These firms are more likely to succeed if they 

 produce specialty products, sell in isolated mar- 

 kets, or develop forward integration from shrimp 

 fishing operations. 



Firms of medium size, neither displaying the 

 characteristics of large nor small firms, either exit 

 from the industry or make adjustments in their 

 size and/or product lines. Medium-sized firms tend 

 to be unstable initially because they apparently 

 are not organized to successfully enter shrimp 

 specialty markets yet are too small to compete in 

 the national major line shrimp markets. 



FORECASTING FIRM 



DISTRIBUTION AND PREDICTING 



A STRUCTUAL EQUILIBRIUM 



A forecast- for the 1961-71 period of the number 

 of shrimp processing firms in each state of nature 

 was conducted and compared with the actual 

 number of firms appearing in the data during the 

 same period (Table 3). The purpose of this 

 procedure was to evaluate the appropriateness of 

 the transition matrix for forecasting firm dis- 

 tribution within the industry.^ When comparing 

 actual firm numbers to predicted numbers in 

 states 2 through 6 for 1961 through 1971, 17 of the 

 30 predictions were accurate and in state 4, which 

 is least stable, 5 of the 6 predictions were accurate, 

 giving confidence that the dual equilibrium struc- 

 ture remains intact with the predicted numbers. 



-To forecast firm distribution in the Florida shrimp processing 

 industry over time requires that the transition matrix be 

 stationary; that is, the probabilities in the transition matrix do 

 not change over time. Although, the chi-square "goodness-of-fit" 

 test was conducted and the results show the transition matrix to 

 be stationary, predictions should be considered tentative due to 

 the small number of obsen'ations per cell caused by the low 

 number of firms in the industry. Forecasted distribution, 

 however, being very close to that found in the past, indicates that 

 the transition matrix remains useful for prediction. 



^Some of the differences may be due to the small number of 

 observations or to rounding procedures. 



FISHERY BULLETIN: VOL. 74. NO. 4 



Forecasting Firm Distribution Over Time 



The biannual forecasted distribution of firm size 

 for the Florida shrimp processing industry from 

 1973 to equilibrium appears in Table 3. Few 

 changes in the number of firms in each state are 

 observed. The smallest sizes (states 2 and 3) 

 experienced an increase of one firm each while the 

 remainder (states 4, 5, and 6) show no change. 

 Thus, there is an increase of two in the total 

 number of active firms. The number and size of 

 firms in the industry will attain a structural 

 equilibrium in a relatively short period of time. 



Equilibrium State Within the Market 



The equilibrium matrix for the Florida shrimp 

 processing industry was calculated to show the 

 final distribution of firms within the industry 

 under the assumption of a stationary transition 

 matrix (Derman et al. 1973). In equilibrium, firms 

 may still enter and exit but neither the number of 

 firms in each state of nature nor the total number 

 of firms in the industry changes once the equilib- 

 rium is reached. 



The distribution of firms in the equilibrium 

 state compared with the distribution of firms 



Table 3.-Actual number' of firms in each state of nature in the 

 Florida shrimp processing industry, compared with the corre- 

 sponding predicted numbers- using the transition matrix, 

 1959-71 and forecasting to 1985 and equilibrium. 



'Data from source (Florida State Chamber of Commerce). 



^Computed using the transition matrix. 



^The equilibrium probabilities of transition in column order for 

 the six states of nature were one (0.3881), two (0.1319), three 

 (0.2454), four (0.0685), five (0.0603), and six (0.1058) for each of 

 the six respective columns. 



882 



