The USDA recently completed an analysis of potential economic and productivity impacts from changes to the availability of farm labor. Their results found that dairy would be one of the sectors most affected by changes. Although the study did not review one specific policy or any proposed legislation, the model does seem to compare the potential outcomes of mass deportation of unauthorized workers, the expansion of a guest-worker program for agriculture and doing nothing. Progressive Dairyman Editor Walt Cooley followed up with the study’s authors.

Cooley walt polo
Editor and Podcast Host / Progressive Dairy


Why did you complete this study? A legislative request? USDA priority? Please explain.


Hertz and Zahniser: We conducted the study for two reasons. First, we knew that the subject was important. Unauthorized workers – that is, people who lack the U.S. immigration status to work legally in this country – account for about half of hired farmworkers in U.S. crop agriculture, compared with 5 percent of the entire U.S. workforce. Thus, changes in the supply of foreign-born labor could have important effects on agriculture. Second, we felt confident, in advance, about our report’s economic simulation model because it had been used successfully by our co-authors Peter Dixon and Maureen Rimmer in a previous study of immigration policy.


The study tracked the potential outcomes of two scenarios compared to doing nothing. Explain both scenarios.


Hertz and Zahniser: We generate several long-term (15-year) projections for the U.S. economy – two scenarios in which the supply of foreign-born labor changes appreciably and a base forecast that serves as a benchmark for evaluating these scenarios.


In our first scenario, the annual number of temporary nonimmigrant foreign-born farmworkers employed in the U.S. expands over time and increases by about 156,000 by year 15 of the projections. These additional workers are assumed to be available to all agricultural sectors – including dairy farming, which has traditionally been excluded from the H-2A Temporary Agricultural Program.

In our second scenario, the unauthorized workforce in the entire economy – agricultural and nonagricultural – decreases by 2.1 million in absolute terms over the first five years of the scenario. By year five of this scenario, the unauthorized workforce in the U.S. economy as a whole is 4.0 million people smaller than in the base forecast.

Growth in the unauthorized workforce resumes thereafter – but at a slower pace than in the base forecast. By year 15, the projected size of the unauthorized workforce is 8.5 million, compared with 14.3 million in the base forecast, a difference of 5.8 million or 40 percent.


Describe the basic take-home messages of the outcomes of both scenarios.


Hertz and Zahniser: In our first scenario, a greater supply of temporary nonimmigrant farmworkers leads to their increased employment at lower wages. This, in turn, results in long-run increases in agricultural output and exports, above and beyond the growth projected by the base forecast. These increases are generally larger in labor-intensive sectors such as fruit, tree nuts, vegetables and nursery products.

By year 15 of the scenario, these four sectors experience a 1.1 percent to 2.0 percent increase in output and a 1.7 percent to 3.2 percent increase in exports, compared with the base forecast. Less labor-intensive sectors such as grains, oilseeds and livestock production tend to have smaller increases, ranging from 0.1 percent to 1.5 percent for output and from 0.2 percent to 2.6 percent for exports. Dairy farming experiences a 0.3 percent relative increase in output and a 2.3 percent relative increase in exports.

Accompanying this additional growth in agricultural output and employment, however, is a relative decrease of 5.7 percent in the number of U.S.-born and foreign-born permanent residents employed as farmworkers and a 3.4 percent relative decrease in their real wage rate. In the model, U.S.-born and other permanent-resident workers are assumed to compete with temporary nonimmigrant farmworkers in the labor market.

A 3.4 percent relative decrease in the wage rate does not mean that wages are projected to fall by 3.4 percent over the 15-year period of the projection. Rather, it means that the real wage rate in year 15 is projected to be 3.4 percent lower in the first scenario than in the base forecast.

In the second scenario, the effects on agricultural output and exports are opposite in sign, and larger in magnitude, than those in the first scenario. Fruit, tree nuts, vegetables and nursery production are again among the most affected sectors, with long-run relative declines of 2.0 percent to 5.4 percent in output and 2.5 percent to 9.3 percent in exports.

These effects tend to be smaller in other, less labor-intensive, sectors of agriculture – showing a 1.6 percent to 4.9 percent decrease in output and a 0.2 percent to 7.4 percent decrease in exports. Dairy farming experiences a 2.2 percent to 2.5 percent decrease in output and a 5.4 percent to 7.4 percent decrease in exports.

In the labor market, the number of unauthorized workers employed on farms falls by 34.1 percent to 38.8 percent, depending on modeling assumptions, relative to the base forecast for year 15. At the same time, the number of farmworkers who are either U.S.-born, or foreign-born permanent residents, increases by 2.4 percent to 4.0 percent in the long run, compared with the base forecast, and their wage rate increases by 3.3 percent to 7.5 percent.

However, the increased farm employment of U.S.-born and other permanent-resident workers is not sufficient to offset the decrease in unauthorized farmworkers. As a result, the total number of farmworkers decreases by 3.4 percent to 5.5 percent.


What would happen if a proposal allowed more than the scenario’s assumed increase of 156,000 workers?


Hertz and Zahniser: We wouldn’t know unless we actually used the model to estimate such a scenario and we haven’t estimated any scenarios that are based on specific legislative proposals. Within the context of our model, however, it’s likely that a larger increase in the number of temporary nonimmigrant workers would have effects that are similar in direction and larger in magnitude than the ones in our first scenario.


Why did you run several variants of your second scenario?


Hertz and Zahniser: In our model, the impact of changes in the labor supply depends on several parameters, and we conducted sensitivity analysis of two of those parameters to get a better idea of how the agricultural sector would adjust to a large, economy-wide reduction in the supply of unauthorized labor.

Because of this sensitivity analysis, we report the findings of our second scenario in terms of ranges of percentage changes rather than a single percentage change. One parameter governs how changes in the relative wage affect the willingness of workers to supply labor to different occupations.

By reducing the responsiveness of workers to relative wage changes, we could consider the possibility that U.S.-born and other permanent resident workers are generally not interested in performing farmwork. The other parameter governs the extent to which employers are able to substitute between authorized and unauthorized workers.

By reducing this substitutability, we could consider whether employers view authorized and unauthorized workers as more differentiated factors of production.


What did you acknowledge by incorporating the findings of previous studies on the supply of ag labor into this one?


Hertz and Zahniser: In economic research, it is common to conduct a literature review of related previous studies. We used our literature review to point out the similarities and differences between our approach and the methodologies of previous studies.

A key point that we made was that most research suggests a large reduction in the size of the foreign-born workforce would not be offset simply by an equally large increase in the employment of U.S.-born farmworkers. Instead, the reductions would lead to some combination of higher wages, increased employment of U.S.-born and other authorized workers, reduced farm output and advances in mechanization over the long run.


Reducing labor by no matter how much showed that agriculture would become more ‘capital-intensive.’ What might that mean in action?


Hertz and Zahniser: Within the context of our model, it means that the capital would account for a larger proportion of the inputs used in agricultural production. In practice, it might mean that labor-saving agricultural technologies would be developed and applied.

Previous ERS research emphasizes, however, that it is challenging to devise substitutes for labor in key steps of the production process of some agricultural commodities, especially among fruit and vegetables.

In their 2010 study of how specific U.S. fruit and vegetable sectors might respond to higher labor costs, Linda Calvin and Phil Martin found that efforts toward greater mechanization typically require “an integrated approach that includes changes in crop varieties, cultural practices and harvesting methods” and the development of machines that approximate the “judgment and dexterity of experienced farmworkers.” Examples of tasks that are particularly difficult to mechanize include scouting for insects and diseases, spraying, weeding, pruning, thinning and harvesting.


What is the main impact these scenarios would have on the broader U.S. economy?


Hertz and Zahniser: Our first scenario has little impact on the broader U.S. economy because the increase in the number of nonimmigrant foreign-born farmworkers is small compared with the size of the entire U.S. workforce.

In contrast, our second scenario has a more prominent impact because the reduction in the number of unauthorized workers – agricultural and nonagricultural – is much larger. Our results indicate that, relative to the base forecast, the real wages of U.S.-born and foreign-born permanent-resident workers would rise in some lower-paying occupations where unauthorized workers are common, decrease slightly in many higher-paying occupations and decrease slightly on average.

Several factors account for this slight decrease. First, the decrease in the supply of unauthorized labor leads to a long-run relative decrease in output – not just in agriculture but in all sectors of the economy. This, in turn, reduces incomes to many complementary factors of production, including capital, land and U.S.-born and other permanent-resident workers in higher-paying occupations.

Second, with fewer unauthorized workers, the occupational distribution of U.S.-born and other permanent resident workers necessarily shifts in the direction of more hired farmwork and other lower-paying occupations such as food service, child care and housekeeping and away from higher-paying occupations. The effect of this compositional change is to reduce the average real wage for U.S.-born and other permanent-resident workers, even as real wages rise in many lower-paying occupations.

In the long term, overall gross national product accruing to U.S.-born and other permanent-resident workers would fall by about 1 percent, compared with the base forecast. This finding may surprise those among your readers who believe that removing a large number of unauthorized workers would be a boon to the U.S. economy.

However, we should stress that our model applies to the long run, when the economy has returned to near full employment. Our results cannot directly be applied to the current economic situation in which 8 percent of the labor force is unemployed. PD

Both Steve Zahniser and Thomas Hertz are Agricultural Economists with the USDA's Economic Research Service.