Additional farm work experience, on the other hand, is expected to increase wages since experience raises productivity at least initially. We use data from the U. S. Department of Labor’s National Agricultural Workers Survey to estimate the effects of worker characteristics on work histories and the relationship between current agricultural wages and work histories. The probabilities of farm work, non-farm work, unemployment. and time spent abroad are computed from the two year job history collected from each worker in the sample. The wage data and the contemporaneous variables that affect wages correspond to the jobs that workers held after the two-year survey period. The NAWS covers a nationally representative cross-section of workers from 72 counties in 25 states representing 12 distinct agricultural regions. Although only seasonal agricultural services workers are interviewed in the NAWS, SAS is defined broadly as most field work in perishable crop agriculture.5 For each of the interviewing cycles. 30 counties were selected randomly as interview sites. The number of interviews conducted during a given cycle is commensurate with the amount of SAS activity at that time of the year. Interviews are conducted every four months – in January, May, and September – to ensure as diverse a representation of workers as possible . The initial NAWS interviews were in 1988. Since then, the survey instrument has been revised several times. In our analysis. we use the five interview cycles that took place between 1989 and 1991, during which the survey instrument was not changed. Of the 4,718 interviews conducted during the five cycles, we use the 2,357 observations for which all relevant data are available.

Table I lists the means and standard deviations for all workers, those who only worked on farms,plastic planters wholesale those who worked on and off farms, those who had experienced unemployment, and those who spent some time abroad during the previous two years. The most common reasons for spending time abroad cited by farm workers are to visit relatives and to vacation . Only 1 percent eite unemployment as the reason to stay abroad . Of the 2.357 observations us.ed in this study, 1.669 workers or 72 percent engaged only in farm work when they were employed. Only 24 percent did both farm work and non-farm work. A majority of workers. 62 percent. experienced unemployment while 49 percent spent time abroad. A relatively small number of workers in the southeast and midwest experienced unemployment while a relatively large number of workers in the northwest experienced unemploy-ment. Relatively more farm-workers in the midwest, western plains, and northwest did non-farm work. In contrast, relatively few workers in the west performed non-farm work. White workers are 56 percent of sample, blacks are 2 percent of the sample, while the rest are native Americans, Asians, and Hispanies who did not indicate arace. Hispanies make up 92 percent of the sample. A relatively large number of them spent time abroad. Mexican-born workers are 79 percent of the sample, those born in the United States are 10 percent, and the rest were born in other countries. Those workers born in the United States are more likely to have done non-farm work and less likely to have spent time abroad than Mexican-born workers. Citizens are 12 percent of the sample. workers with arnnesty are 55 percent, legal permanent resident are 23 percent, and people who are unauthorized to work in the United States are II percent. A relatively large number of citizens did non-farm work while a relatively small number of LPR workers did so. The level of unemployment does not differ sub.stantial across legal-status groups. Relatively few citizens and LPR workers spent time abroad while a relatively large number of arnnesty workers did so. Only 12 percent of workers say they speak English well and about the same number claim to write well.

A relatively large number of workers with English skills performed non-farm work. However, they are no less likely than others to experience unemployment. They are much less likely than others to spend time abroad. Female workers are 22 percent of the sample. A relatively large number of female workers experienced unemployment while relatively few of them engaged in non-farm work. They were much less likely than men to spend time abroad. Workers with families are no more likely to do non-farm work and they are no Iess likely to experience unemployment than those without families. Workers who live with their families are less likely than others to spend time abroad. The skill level required for the most difficult task performed by a given worker during the two-year period determines the “skill level” . Workers were interviewed during the winter, spring, or fall seasons. Regardless of when interviewed, the worker’s history includes each season twice, so that the workers’ histories are comparable. Wages, which are only recorded at the time of interview, differ by season. Moreover, workers who are sampled in winter includes a relatively large share of those workers who enjoy year-round farm employment. In contrast, the spring sample contains a higher proportion of workers who do farm work only during the peak season. A relatively small number of the winter sample workers experienced unemployment, whereas a relatively large number of the spring sample workers did so. The average hourly wage for all workers in the sample is $6.30. There is little variation in wages among the different subsets of workers. Workers who specialized in farm work eam slightly higher average wages than those who performed non-farm work. The average worker had 5.6 years of education. The average number of years of farm work experience is 10.2 for the entire sample.

Those who performed non-farm work have slightly less experience in farm work than workers who specialized in farm work . Workers with a history of unemployment have as much experience in farm work as others.Farm work experience in the United States, age, and luck determine the probability of engaglng in non-farm work. At the sample mean of 33 years, a one percent increase in age raises the probability of doing non-farm work by one percent. On the other hand, at the sample mean of 10 years, a one percent increase in farm work experience reduces the probability of non-farm work by 0.9 percent. Workers who have friends or relatives in non-farm work are 5 percent more likely to do non-farm work than others. Compared to those interviewed in the fall, individuals interviewed in the spring are 2 percent more likely to do non-farm work. Compared to workers in the northeast, those in the western plains have a 17 percent greater chance, those in the northwest have a 18 percent greater chance, and those in the west have a 14 percent greater probability of unemployment. Women are 17 percent more likely to be unemployed than men. In addition,plastic plant pot workers sampled in the spring are 5 percent more likely than those sampled in the fall to be unemployed. Amnesty workers, on the other hand, are 4 per<:ent less likely than unauthorized workers to experience unemployment. It is interesting to note that neither citizens nor LPR workers are less likely to experience unemployment than unauthorized workers. Further, skill level, education, and English ability apparently have no effect on reducing one’s probability of unemployment. Thus, workers cannot reduce their odds of being unemployed through education or obtaining skills. As workers begin to establish themselves in the United States, they abstain from retuming horne for a while. But when their status becomes more solid, they spend more time in the horne country. Evaluated at the sample mean of 10 years, a one percent increase in years of U. S. farm work experience reduces the probability of staying abroad by 1.4 percent.As U. S. farm work experience rises above 25 years, however, workers increase the share of time they spend abroad. Legal status also affects the probability of being abroad. Amnesty workers are 15 percent less likely than unauthorized workers to spend time abroad. LPR workers, who are generally more established in the United States than arnnesty workers, are only 9 percent less likely than unauthorized workers to spend time abroad. Citizens, the most established legal status category, are not statistically significantly less likely than unauthorized workers to spend time abroad. Women and workers who live with their spouses are 6 percent less likely than others to spend time abroad. Workers who own a house in the United States have a 5 percent lower probability of spending time abroad than others. In addition, workers in the southeast are 8 percent less likely than workers in the northeast to spend time abroad. On the other hand, workers born in Mexico are 5 percent more likely than others to stay abroad.

Those sampled in the spring have a 2 percent greater chance of spending time abroad than those sampled in the fall. We want to consistently estimate the relationship between work histories and current agricultural wages. We can use ordinary least squares if the probabilities, Pij’ and the editor term in the wage equation are uncoITelated. If they are correlated, we ean use an instrumental variables technique, where the instruments are the fitted probabilities obtained by substituting the estimated coefficients, ~, from the multi-nomial-Iogit estimates into Equation 18.We can test the null hypothesis that these probabilities and the wage equation’s error term are uncorrelated using a Hausman test. Because the Hausman-test statistic, 7.69, is less than X~05, we fall 10 reject the null hypothesis of no correlation between the probabilities and the error term. As a result, we report the ordinary-least-squares estimates for the wage equation in Table 3. According to this wage equation, the agricultural wage is positively related to the probability of working on farms and negatively related to the probability of engaging in non-farm work. Neither of these relationships, however, is statistically significant at the 5 percent level. As predicted, the more time the worker spent unemployed in the last two years, the lower that worker’s current agricultural wage. We can reject the null-hypothesis that the coefficient on the share of time the worker was unemployed is not statistically significantly different from zero at the 5 percent level. A ten percentage point increase in the share of time unemployed leads to a 1 percent drop in the current agricultural ‘wage. Thus, unemployment reduces farm workers’ income in two ways. Initially, unemployment reduces a worker’s earnings by reducing the hours worked. Later, the worker may accept a relatively low-wage job because the worker lacks the resources to continue searching. We tested whether the relationship between work histories and current agricultural wages are different between the winter and other seasons. For example, we examined whether the relationship between the probability of unemployment and current agricultural wages differs for the winter sample from the rest. To examine this question, we included an interactive variable that is the product of the winter dummy and the probability of unemploy-ment. We then tested if the winter dummy and this interactive variable were jointly statistically significantly different from zero. All winter-probability combinations were tested. The test statistics ranged from 0.026 to 1.832 and all were smaller than the critical value F.os = 3.00. Therefore, we could not reject the null hypothesis that the relationship between work histories and current agricultural wages are the same for workers interviewed in all seasons. Wages vary substantially geographically. Compared to workers in the northeast, . workers in the southeast eam 16 percent 10wer wages while those in the western plains eam 19 percent lower wages. Skill levels also have substantial effects on wages. Supervisors eam 29 percent higher wages than unskilled workers who perform tasks other than harvesting. Neither unskilled harvesters nor semi-skilled workers make statistically significantly different wages than unskilled non-harvesters. Type of crop also creates some variation in wages. Specialty crop workers eam considerably less than all other workers. Relative to specialty crop workers, those who work on field crops eam 13 percent higher wages while individuals who work on nuts and fruits eam 14 percent more. Flower and nursery product workers and vegetable workers also eam 15 percent and 14 percent higher wages than specialty crop workers, respectively. Wages differ with legal status.