To the degree that work histories affect current wages, farm workers may try to mold their work histories so as to increase their agricultural wages. For example, if specializing in farm work raised their wages, they might turn down temporary non-farm work jobs. On the other hand, if obtaining non-agricultural employment allows them to hold out for higher paying farm jobs than otherwise, workers may benefit from seeking employment in both agricultural and non-agricultural sectors. Similarly, if learning English reduces a worker’s probability of being unemployed, workers can reduce their unemployment and increase future wages in this manner. In Section 2, we fist present the standard search theory explanation of the relationship between the reservation wage and the length of unemployment specils. We then modify the theory to reflect the conditions in the agricultural labor market. We use this model to show how agricultural workers’ wages are affected by their employment history. In Section 3, we use data from the National Agricultural Workers Survey to investigate how individual charaeteristies affect work history and whether work histories affect wages. We describe the data set in section 4. In Section 5, we use the estimates to simulate the effect of individual characteristics on work histories and wages. We draw conclusions in the final section.Cross-section studies in non*agricultural industries usually find a positive relationship between the wage and the length of employment or tenure .
Two explanations are commonly given for this relationship. First, as workers accumulate experience in a particular job, they become more productive. Thus, if they are paid the value of their marginal product, their wages increase over time. Second,plastic growers pots many jobs require training designed to raise workers’ productivity. Most of this training takes place when the worker is first employed. If the firm believes workers may unit before the firm captures the rectum on the training, employers make workers pay for part of the training cos18 through lower wages during the training period. Once the training is completed, however, workers’ wages rise reflecting their higher productivity. Employers pay higher wages once the training is completed because they do not want to lose productive workers . Wages may continue 10 rise if additional experience increases productivity. For such jobs, wages rise with job tenure. Agriculture employers employ few workers for long periods and rarely provide formal on-the-job training. Due to the seasonal nature of most agricultural work, it would be unprofitable to employ peak-season workers year round. Although there is an increased desire among agricultural employers in the post-IRCA years to retain workllrs, worker retention rates are not high in agriculture. In the NAWS sample, only forty percent of workers returned to the same employer a year later . Agricultural employers rnay not provide eostly formal training because they know that the tumover rate is high . Most job training in agriculture is informal and takes the form of leaming by doing. Workers gain on-the-job training and experience and increase productivity by working for a variety of farm employers over time, rather than working for only one employer. Thus, we expect to see a relationship between agricultural wages and the probability of being employed in agriculture rather than with the current jobtenure. If, however, workers retain their skills despite sporadie employment in farm work or skill requirements are low, even this relationship may be a tenuous one.
Some hired agricultural workers use non-farm employment to supplement their income and that the proportion of workers who do so has increased over time . Do farm workers who have been able to obtain non-farm employment enjoy more flexibility to seek higher paying agricultural jobs than those who did not find non-farm employment? Both Matta and Oliveira report that more educated workers had a higher probability of working in non-farm work. In addition, Gabbard and Perloff find that hired agricultural workers who had confidence in finding non-farm work received higher agricultural wages than those who did not. Gabbard and Perloff argue that workers who have confidence in finding jobs in the non-agricultural sector are likely to have higher reservation wages. and therefore higher agricultural wages, than others. There are two components to this argument. First, workers with greater search confidence – those with better education and English speaking ability – generally have higher reservation wages. Second, those workers who believe they can get jobs in the non-agricultural sector may have higher reservation wages than others because non-agricultural wages in general are higher than agricultural wages. Our question differs from that in Gabbard and Perloff. We want to know if hired agricultural workers who actually worked in non-farm work receive higher agricultural wages than others. There are two reasons why such workers may not receive higher agricultural wages than other workers.First, the wages in non-agricultural jobs available to hired farm workers may not be higher than agricultural wages. In the NAWS data, the average hourly wage for non-farm jobs held by farm workers was only $5.09 compared tu $5.13 for farm jobs . Further, Mines et al. repon that the non-farm jobs – mostly in construction and service industries – taken by hired agricultural workers are just as seasonal and insecure as agricultural jobs. A worker with relatively low-wage non-farm employment is less able to hold out for a good job than if that worker had earned a higher agricultural wage. Second, farm workers who take non-farm jobs may not be the ones with greater search confidence in the non-agricultural sector. The workers who have confidence in finding non-agricultural jobs are more likely tu be better educated and speak more English .
But, if available non-farm jobs are not superior to farm jobs in wages or job security, there is no incentive for these “high-quality” workers to choose non-farm jobs over farm jobs. Thus workers who have confidence in finding non-farm jobs and those who actually take them may differ. For example, those who have search confidence are better educated and have superior English skills to those workers who actually take non-farm jobs. Workers may have engaged in non-farm work because they bad better contacts or better information than other workers and not because they were more qualified. If so, there may be no qualitative difference between workers who worked in non-farm work and those who did not. Although both Matta and Oliveira find that educated workers had a higher probability of working in non-farm work, their December Current Population Survey data sets may not have been representative of all hired agricultural workers. That data set is not likely to yield a representative sample of hired farm workers because many immigrant workers return horne for the winter season. Only about a quarter of the CPS samples are minority workers, while nearly 88 percent of the NAWS sample are Hispanic workers. If non-farm jobs available to hired farm workers pay no more than farm jobs, and if there is no qualitative difference between workers who take non-farm jobs and those who do not, agricultural wages may not be positively related to the probability of working in the non-agricultural sector.The worker characteristics, Xl’ include race , ethnicity , place of birth , legal status , knowledge of English, gender, age, age squared, family background , years of education, farm work experience in the United States, skill level, whether the worker has friends or relatives in non-farm work, whether the worker owns a house in the United States, the region in which the worker was located,blueberry in pot and the cycle in which the worker was interviewed. Race and ethnicity may affect the probability of being employed in the non-farm sector due to racial or ethnic bias. In the farm sector, these variables are unlikely to affect the probability of being employed as most workers are minorities. If race and ethnicity do influence the probability of being hired in non-farm work, they may also affect the probabilityof being unemployed because farm-workers typically seck non-farm employment in the off season. Foreign-born workers are many times more likely 10 spend time abroad than those born in the United States. Legal status is likely to affect the probability of unemployment if unauthorized workers face greater difficulty in finding work than do legal workers.
In addition, unauthorized workers suffer from much shorter farm employment speils than others .Knowledge of English, though probably not a factor in finding farm work, may increase the probability of finding non-farm work where English ability is likely to be more important than it is in farm work. On the other hand, the type of non-farm work available to farm-workers may not require much English ability. Gender may influence the probability of being out of work. That female workers have shorter farm employment durations than men may be due to their heavier child-rearing and other family responsibilities. If so, the same responsibilities may force women 10 experience a higher probability of unemployment than men. Gender may also affect the probability of spending time abtoad. Indeed, within OUT sample, virtually the only workers who shuttle hetwecn their horne country and work sites in the United States are men . As workers age, they may become more established in a region, which may raise their probability of finding non-farm work. On the other hand, older workers may experience a lower probability of working in agriculture as farm work is more physically demanding than most non-farm work.Additional farm work experience should increase the probability of working in agriculture because experienced workers have better information and are more valuable to employers. Skil1 level, for the same reason, should raise the probability of working in agriculture. Workers who live with their families are likely to have a lower probability of spending time abroad. This is because these workers are more established in the United States, and are less inc1ined to visit their relatives in their horne countries or vacation abroad. Workers with families rnay be out of employment a smaller share of the time than other workers because they are under pressure to support their dependents.Geographical regions are expected to capture three different effects. First, workers in regions with long growing seasons are likely to experience a lower probability of unemployment than those in short growing-season regions. Second, individuals who live c10se to metropolitan areas are more likely to find non-farm work compared to those who do not Thus, workers in the regions that contain metropolitan areas, such as the northeast, rnay have a higher probability of non-farm employment than workers in predominantly rural regions such as the western plains. Finally, immigrant workers who live close to the Mexican border may return horne more frequently than others.The race and ethnicity variables are included to capture possible effects of discrimination. The gender dummy captures possible differences in wages between men and women. Although most previous studies fail to find gender wage gaps for a given task category, women may be concentrated in tasks that pay low wages. The effects of family and household composition on wages are unknown. On the one hand, workers who live with their families face greater search constraints than those who do not. On the other hand, the stability of workers who live with their children may be appealing 10 employers, or employers may prefer to hire related workers. The legal status dummies capture the return to being authorized to work in this country. Unauthorized workers could irnpose extra costs to employers in the form of fines and lost revenues if workers are apprehended , though such fines are few and far between. Wages should vary with tasks , crops , and skills . The geographical variables control for regional differences in macroeconomic conditions such as the demand for labor and cost of living. The seasonal dummies control for seasonal variation in wages. One might expect higher wages during the peak agricultural season. However, a fundamental difference between the summer sample and the winter sample may yield the opposite result. Those who are found working in agriculture during winter are not a typical group of hired farm workers since there is little agricultural demand for labor in winter. The winter sample must contain an unusually high proportion of long-term and skilled workers who may well be paid more than the average wage.The English 1anguage ability variables capture the possib1e effects of superior language ability in finding and getting better farm jobs. The importance of English skills is expected to be minimal given the widespread use of Spanish. the native 1anguage of most immigrants. in the agricultural industry.