Examining the Impact of the Real ID Act on U.S. State-Level Agricultural Exports

Yunzhe Zhu

Department of Agricultural Economics, University of , USA

Email：[email protected]

Selected Paper prepared for presentation at the Southern Agricultural Economics Association (SAEA) Annual Meeting, Birmingham, , February 2-5, 2019

Copyright 2019 by Yunzhe Zhu. All rights reserved. Readers may make verbatim copies of this document for non-commercial purposes by any means, provided that this copyright notice appears on all such copies. Examining the Impact of the Real ID Act on U.S. State-Level Agricultural Exports

Yunzhe Zhu

Department of Agricultural Economics, University of Kentucky, USA

______

Abstract

This study employs difference-in-differences (DID) method to estimate impacts of the implementation of the Real ID Act on U.S. exports for seven categories of agricultural commodities. Empirical results show that compliance with the Act has a statistically negative impact on exports of agricultural commodities produced in the labor intensive sectors of fruits and vegetables, where most illegal1 immigrant farmworkers are hired. The Act implementation leads to decreases in exports for “fruits and tree nuts” and “fresh vegetables” by 25.0% and

12.8%, respectively. Such a strong immigrant – trade link suggests the need for a deceleration of current immigration law enforcement and a tacit tolerance of unauthorized employment.

Keywords: the Real ID Act; U.S. agricultural exports; difference-in-differences; illegal immigrants; labor intensive sectors

______

Introduction

The Real ID Act, enacted on May 11th, 2005, is an that modifies U.S. federal

1 Unauthorized, undocumented and illegal will be used interchangeably. These terms are used to mean a person who resides in the United States but who is not a U.S. citizen, has not been admitted for permanent residence, and is not in a set of specific authorized temporary statues permitting longer-term residence and work (Passel et al., 2004).

1

3

2.5 Millions

2

1.5

1

0.5

0 1990 1995 2000 2005 2010 2015 2020

California

Figure 1 – Unauthorized immigrant population trends for top states, 1990-2016

Source: Pew Research Center estimates based on augmented U.S. Census Bureau data.

law pertaining to new authentication standards for state-issued driver’s licenses and non-driver

identification cards, as well as various immigration policies pertaining to terrorism. While the

passage of the Real ID Act is aimed to strengthen U.S. national security, several portions of the

Act have imposed stricter standards of proof for individuals applying for the Real ID credentials

– thereby restricting illegal immigrants, who are unable to prove their legal status or who lack social security numbers, from working, causing many immigrants and foreign nationals to lose their jobs. Figure 1 provides unauthorized immigrant population trends for top U.S. states. As

figure 1 shows, the numbers of unauthorized immigrants, for most states, rose sharply in the

1990s and reached their peaks in 2007 when the recession began. They declined through the end

of the recession in 2009 and then stabilized with a slightly further decline until ticking down in

2016. Given the large number of unauthorized immigrants in the U.S., the impact caused by the

Real ID Act is expected to be huge.

2

The history behind the Real ID Act started with the terrorist attacks, September 11, 2001, after which the whole nation was shocked by the sheer horror. U.S. states then accelerated their effort to counter issues with counterfeit driver’s licenses and identification cards in order to prevent terrorists from gaining immigration status in the U.S.. In July 2002, the first “National Strategy for Homeland Security” produced by the Office of Homeland Security was released. It outlined major state initiatives, including driver’s licenses: states with assistance from the federal government should craft solutions to curtail the future abuse of driver’s licenses by terrorist organizations. In July 2004, the 9/11 Commission issued a 585 page report on how to reform the

U.S. Intelligence community and to implement other security measures to prevent future terrorist attacks against the U.S.. On page 390, under the heading Immigration Law and Enforcement, minimum standards for identification document was developed. In December 2004, President

Bush signed into law the “National Intelligence Reform Act of 2004”. The law required the U.S.

Secretary of Transportation to establish a negotiated rule making process to establish minimum standards for state-issued driver's licenses and identification cards. On May 11, 2005, President

Bush signed into law the “Emergency Supplemental Appropriation for Defense, the Global War on Terror, and Tsunami Relief, 2005”, which included the “Real ID Act of 2005”.

Although people whose ID cards do not meet the new federal standards will not be able to fly on a domestic commercial flight or enter a federal building after October 1, 2020, the Act participation by states is voluntary. Some states have adopted the Act, while some others refused to implement it. As of October 2018, 37 states and territories have been certified as compliant, and 19, that provide adequate justification for noncompliance, have granted extensions of time to meet the Real ID requirements. It is expected that all but four of the 56 U.S. states/territories will be issuing Real ID compliant licenses/IDs by early 2019 (U.S. Department of Homeland

3

Security). Regan and Deering (2009) found that relatively less populous and less wealthy states,

which are likely to be more impacted by unfunded mandates, and more conservative states,

which are more likely to be concerned about retaining state control, were more likely to oppose

the Real ID Act. To get more insight into the geographic differences, Figure 2 shows the status

of states regarding the Real ID Act by the end of 2016. In figure 2, the 23 states in green are considered in the “compliant group” and the 27 states in yellow are considered in the “non-

compliant group”. In addition, there is tremendous variation in the timing of the Act

implementation across U.S. states. As table 1 shows, the earliest state implementation was in

South Dakota in December 2009 and the latest one was in in August 2017, leaving 23

states not (yet) compliant, so the Real ID Act was not mandatorily imposed at the national level

and not all states treated it identically. Hoynes and Schanzenbach (2009), in their study of food

stamp program, emphasized the shortcomings of the absence of such variation in the timing for a

research design. The variation of the Real ID Act starting date precludes boiling down the study

to a simple before-after DID analysis.

When it comes to the agricultural sector, the Pew Research Center (2016) estimated that

325,000 unauthorized immigrants work in the U.S. agricultural sector. Passel and Cohn (2015)

reported that it is farming in which unauthorized immigrant employees are the highest share of

the workforce in most states. The farming sector has grown increasingly dependent on a steady

supply of workers who have entered the U.S. illegally and this has created a situation where

presently half of all crop farm workers are unauthorized (Ruark and Moinuddin, 2011). Table 2

presents 2014 estimated unauthorized immigrant population by state. These tabulations show that

the numbers of both unauthorized immigrants and unauthorized immigrant farmworkers differ

significantly among states. For the number of unauthorized immigrants, , New York,

4

Figure 2 – The status of states regarding the Real ID Act by the end of 2016

Source: The author’s tabulation of data from DMV.ORG (Table 1).

5 Table 1 – Dates of issuing Real ID compliant credentials among U.S. states by December 15, 2017

State Dates of issuing Real ID compliant credentials Alabama February 2012 April 2016 October 2016 July 2013 October 2011 July 2010 Florida January 2010 July 2012 March 2012 January 2010 January 2013 Kansas August 2017 June 2016 June 2017 September 2013 November 2014 New November 2016 May 2017 May 2011 December 2009 June 2017 Texas October 2016 January 2010 January 2014 West January 2012 January 2013 June 2011 Not (yet) compliant California Not (yet) compliant Not (yet) compliant Illinois Not (yet) compliant Kentucky Not (yet) compliant Not (yet) compliant Not (yet) compliant Not (yet) compliant Not (yet) compliant Not (yet) compliant Not (yet) compliant New Jersey Not (yet) compliant Not (yet) compliant Not (yet) compliant Not (yet) compliant Not (yet) compliant Not (yet) compliant Not (yet) compliant Virginia Not (yet) compliant Not (yet) compliant Under review Under review New York Under review Source: DMV.ORG

6

Table 2 – Estimated unauthorized immigrant population, by state, 2014

Unauthorized Immigrant Population Unauthorized Immigrant Population in farming, fishing and State forestry occupations Population Rank Population Rank Alabama 65,000 32 665 20 Alaska 10,000 44 12 39 Arizona 325,000 9 5360 5 Arkansas 70,000 30 888 17 California 2,350,000 1 73728 1 Colorado 200,000 15 1037 16 Connecticut 120,000 19 142 35 Delaware 25,000 40 119 37 Florida 850,000 3 7979 2 Georgia 375,000 7 2525 8 Hawaii 45,000 37 315 30 Idaho 45,000 36 2507 9 Illinois 450,000 6 * Indiana 110,000 21 320 29 Iowa 40,000 38 279 31 Kansas 75,000 29 650 21 Kentucky 50,000 34 1241 10 Louisiana 70,000 31 791 19 Maine < 5,000 47 * Maryland 250,000 12 442 26 Massachusetts 210,000 13 * Michigan 130,000 17 1110 14 Minnesota 100,000 23 464 24 Mississippi 25,000 42 139 36 Missouri 55,000 33 274 32 Montana < 5,000 50 * Nebraska 45,000 35 177 34 Nevada 210,000 14 189 33 New Hampshire 10,000 43 * New Jersey 500,000 5 874 18 85,000 27 1169 11 New York 775,000 4 * North Carolina 350,000 8 3116 7 North Dakota < 5,000 49 * Ohio 95,000 24 454 25 Oklahoma 95,000 25 478 23 Oregon 130,000 18 5705 4 Pennsylvania 180,000 16 1062 15 Rhode Island 30,000 39 34 38 South Carolina 85,000 26 1161 12 South Dakota 5,000 46 * Tennessee 120,000 20 366 28 Texas 1,650,000 2 4001 6 Utah 100,000 22 406 27 Vermont < 5,000 51 * Virginia 300,000 10 1116 13 Washington 250,000 11 7866 3 < 5,000 48 * Wisconsin 80,000 28 628 22 Wyoming 5,000 45 * Note: The numbers in column 4 are products of estimated total employments in farming, fishing and forestry occupations and shares of unauthorized immigrant workers in these occupations, and have been rounded up to the nearest digits. “*” indicates either “data unavailable” or “data unreported due to less than 5,000 unauthorized immigrants in the civilian labor force in 2014”. Source: Pew Research Center estimates for 2014 based on augmented American Community Survey. 2014 State Occupational Employment Estimates, Bureau of Labor Statistics, United States Department of Labor.

7

Texas and Florida are ranked top 4. The former two states are located respectively in the western

and eastern coast, where there are more job opportunities and friendlier immigration policies.

California even made history by officially becoming the first sanctuary state in the nation in

2018. These can partially explain the large number of unauthorized immigrants there. The latter

two states have tighter policies towards illegals, but they draw a large number of them due to the

fact that they are located near the U.S. border with Latin America. These four states plus New

Jersey and Illinois account for 59% of unauthorized immigrants in the U.S. (Pew Research

Center 2014). For the number of unauthorized immigrant farmworkers, it is much smaller than

the number of the total population, because only 4% of unauthorized immigrant workers held

farming jobs, and they were more likely to be employed (48% of them) in service and

construction occupations (Pew Research Center 2014). California and Florida are still ranked on

top. Farming in the two states has the highest shares of unauthorized immigrant workers (35%

and 33%, respectively). New York and Illinois’s percentages of civilian workforce in farming

that consists of unauthorized immigrants fail to get into the top three in these two states

(percentages less than that in the occupations of construction, service and production) and thus

are not reported. Generally, unauthorized immigrant farmworkers are widely distributed in both

“compliant” and “non-compliant” states, which is valid for data analysis.

The Real ID Act affects those immigrant farmworkers in two ways. First, being deprived of ID cards causes losing jobs. U.S. immigrants will have to go through a stricter background check for employment. They will have to present proof of lawful immigration status. Most illegal immigrants are unable to present the required documents and they thus will fail to pass such an employee verification. In this case, their driver’s licenses or non-driver identification cards would not be renewed or would be retroactively canceled, and employers in the agricultural

8

sector might be unable or unwilling to hire individuals who cannot provide ID cards because of

fearing sanctions under federal laws. Orrenius and Zavodny (2009) found strong evidence of a

decline in employment among male Latin American immigrants under the stricter immigration-

related law enforcement including the implementation of the Real ID Act in the post-9/11 period.

Second, weaker workforce participation takes place. Studies show that individuals who lack the ability to obtain a driver’s license have more difficulties in maintaining steady employment

(Pawasarat and Stetzer 1998; Sandradanziger et al. 2000). Lack of public transportation is especially prominent in rural areas where many immigrants settle. Driving in these regions is not a privilege, but rather, a necessity to perform daily activities associated with living such as working or conducting regular business transactions (García, 2006). Those illegal immigrant farmworkers do not always have access to public transportation or other transit opportunities, and they thus will become less flexible in response to advanced planning for work shifts and overall have weaker ability to work more frequently. To farm employers’ perspective, they will find it less attractive to hire individuals, who often show up late or miss work shifts. If the illegals are allowed to lawfully drive to and from work, they would be better workers by arriving to work consistently and on time.

Which sectors in the field of U.S. agriculture does the Real ID Act affect most? For decades, the farm labor supply in the sectors of fruits and vegetables has depended heavily on federal immigration policy and enforcement. Labor intensive fruit-and-vegetable farming attracts a relatively large illegal workforce. Carroll et al. (2011) documented the findings of the National

Agricultural Workers Survey (1989 – 2009) that most of the 54,000 farmworkers interviewed work in the sectors of “Fruits and Nuts” and “Vegetables”. Martin and Calvin (2010) reported

that farms producing fruits and nuts, vegetables and melons, and horticultural specialties, such as

9

greenhouse and nursery crops, accounted for $ 13.6 billion, or over half of the $26.4 billion in

U.S. farm labor expenditures in 2007. On the other hand, relatively low wages, hard physical

labor, and seasonal work reduce the appeal of farm work to most U.S. citizens, so an orchardist,

nursery operator, packer or processor, who hires seasonal farmworkers, may have no alternatives

but choose illegal immigrants to work for them. By contrast, land-intensive crops (e.g., wheat, corn, cotton, soybeans and sorghum) are largely automated and do not depend heavily on illegal immigrant workers. Hence, the Real ID Act is expected to have a significant impact on the labor intensive sectors, such as the sectors of fruits and vegetables.

The Real ID Act implementation will result in two economic effects towards labor intensive agricultural sectors. First, putting upward pressure on farm wages. Unauthorized immigrant farmworkers have annual incomes that are $5,600 less than that of authorized workers working in the same farming sector (Ruark and Moinuddin, 2011). Although they are underpaid compared to their legal counterparts, agriculture industry groups in many states have consistently complained of a of agricultural workers and the Act implementation will deteriorate this situation. Holding the labor demand curve unchanged, the reduced labor availability will cause a substantial leftward shift of the agricultural labor supply curve – thereby increasing farm wages and, in turn, raising the cost of farm labor for employers in the “compliant” states. Employers in the “noncompliant” states, on the other hand, will gain from lower labor costs and the ability to use their land, capital and technology more productively. Second, disrupting the flow of

“pickers”. Fruits and vegetables become ripe at a fixed time and must be picked quickly before they rot. During harvest seasons, both migrant and seasonal farmworkers2 may move from farm

2 In general, migrant farmworkers are individuals who travel a greater distance to farm sites and cannot return on a daily basis to their permanent residences. Seasonal farmworkers, on the other hand, temporarily work in agriculture without having to leave their residence (Lewis et al., 2017).

10

to farm to remain employed. If farmers cannot find workers when they need them, their crops

may be ruined. Following the growing season, those illegal immigrant farmworkers often travel a

set route (e.g., Florida and its way north), so disrupting this flow can be devastating to the fruit

and vegetable sectors. Given the two economic effects, those non-compliant states will gain a comparative advantage in producing labor intensive products, by which a state’s trade pattern is determined. States with a comparative disadvantage, on the other hand, will experience a decrease in the exports.

As discussed above, the following hypothesis is made:

H. The implementation of the U.S. Real ID Act has a significantly negative impact on exports for the labor intensive agricultural sectors of fruits and vegetables in states compliant with the

Act. The non-labor intensive sectors would be influenced as well, but in a statistically insignificant way.

Literature review

The reliance of labor intensive sectors on unauthorized immigrants has made the question,

“How to address their unauthorized status?”, one of the most challenging policy issues these

days. Much has been written about the connection between immigration policies and

international trade. Previous studies found today’s labor-scarce economies having liberalizing

trade and restrictive immigration policies. Fundamentals behind this policy paradox as well as its

validity have become a big academic concern. Peters (2014) argued that increasing trade

openness allows U.S. firms to move production overseas – thereby decreasing their need for

labor at home and leading them to spend their political capital on issues other than immigration,

and that their lack of support for open immigration allows policymakers to restrict immigration.

11

Zahniser et al. (2011) conducted an immigration policy simulation and found that for the most labor intensive agricultural sectors, the policy expansion scenario (looser application of immigration controls) resulted in a long-run 1-2% increase in output and 0.2-3.2% increase in exports, while the policy enforcement scenario (tighter application of immigration controls) resulted in a 2-4% decrease in output and 0.8-6.3% decrease in exports. Genc et al. (2012) performed a meta-analysis and found that an increase in the number of immigrants by 10 percent may be expected to increase the volume of trade on average by about 1.5 percent and that the impact is lower for trade in homogeneous goods. Faustino and Leitão (2008) used a static and dynamic panel data analysis and found that the stock of immigrants has a positive effect on all types of Portuguese trades. As cited above, previous studies tend to point out the pernicious effects of accelerated immigration enforcement in developed countries on their international trades and suggest a tacit tolerance of unauthorized employment.

When it comes to the impacts of the Real ID Act, most existing literature focused on examining the case law upon which some of the provisions are based and offering interpretations for unclear provisions. Cianciarulo (2006) argued that several portions of the Act may result in the denial of bona fide asylum applications and provided concrete guidance for policymakers to protect victims of persecution. Fletcher (2006) predicted that the Real ID Act will perpetuate gender bias and widen the gap between accesses to protection of asylum seekers in general and that of asylum seekers escaping gender-related persecution in particular, and discussed legal strategies for minimizing the negative consequences of the Act through both strategic representation of asylum seekers and broader advocacy efforts. García (2006) concluded that 1) under the anti-immigrant provisions of the Real ID Act, it will be more difficult for immigrants to obtain asylum in the U.S.; 2) immigrants, legal and illegal, will have difficulty obtaining an

12

acceptable driver’s license or identification card which are necessary to live and conduct everyday transactions in the U.S.; 3) immigrant lives will be put at risk with the construction of physical barriers along the U.S. borders.

Few of existing literature have analyzed the impact of anti-immigration policies, including the

Real ID Act, on U.S. economy. Previous studies related to the connection between anti- immigration policies and U.S. economy have stated the negative correlation based only on surveys. Michigan Law Revision Commission, 47th annual report (2015-2016) presented that

Michigan agriculture industry groups have consistently complained of a shortage of agricultural

workers in the state since 2008 when Michigan began to require applicants for driver’s licenses

and state identification to provide proof of U.S. citizenship or immigration status. Michigan’s

Migrant and Seasonal Farmworkers Workgroup (MSFW) (2013) issued a report that

recommended improving the system in which migrant and seasonal farmworkers go about

applying for licenses and that made it clear – access to driver’s licenses is extremely important

for regular seasonal and migrant farmworkers’ ability to participate in the workforce. Fitz (2012)

reported that the Georgia Agribusiness Council estimated that the state could lose up to $1

billion in produce from a lack of immigrant labor after the passage of Georgia’s anti-immigrant

law, H.B.87 and that a survey of farmers conducted by the Georgia Department of Agriculture

found 56 percent of those surveyed were experiencing difficulty finding workers. So far, no

study has addressed the issue of causal effect of those anti-immigration policies on U.S.

economy, not to mention on exports, by using econometric methods (e.g., DID, regression

discontinuity, propensity score and synthetic control). This article contributes to the literature by

applying DID method to quantifying the expected negative impact and provides empirical

evidence for further studies.

13

Data

The website of United States Department of Agriculture (USDA) Economic Research Service provides the calendar-year (January to December) state agricultural export estimates using the new U.S. farm-receipts-based method starting from 2000 to 2016 by commodity. The commodity coverage for exports includes 24 categories, as well as aggregate estimates for plant products, animal products and total agricultural exports. Nine categories (fresh fruits, processed fruits, tree nuts, fresh vegetables, wheat, corn, soybeans, beef and veal, pork) are included for the analysis. “Fresh fruits”, “processed fruits” and “tree nuts” are merged into one category “fruits and tree nuts” as many classification methods use, and estimates for each of the agricultural products are weighted by the corresponding, 2010 state-level production measured in U.S. dollars. The following lists the data source for those weights:

(1) Fruits and tree nuts: fruit & tree nut totals, utilized – production, measured in $, 2010 survey data, USDA Quick Stats.

(2) Fresh vegetables: vegetable totals, fresh market – production, measured in $, 2010 survey data, USDA Quick Stats.

(3) Wheat: 861 – Wheat – Acreage, Production, and Value by Leading States, Crops, Section 17.

Agriculture, 2012 Statistical Abstract of the United States, U.S. Census Bureau.

(4) Corn: 859 – Corn – Acreage, Production, and Value by Leading States, Crops, Section 17.

Agriculture, 2012 Statistical Abstract of the United States, U.S. Census Bureau.

(5) Soybeans: 860 – Soybeans – Acreage, Production, and Value by Leading States, Crops,

Section 17. Agriculture, 2012 Statistical Abstract of the United States, U.S. Census Bureau.

14

(6) Beef and veal: 873 – Cattle and Calves – Number, Production, and Value by State, Meat And

Livestock, Section 17. Agriculture, 2012 Statistical Abstract of the United States, U.S. Census

Bureau.

(7) Pork: 872 – Hogs and Pigs – Number, Production, and Value by State, Meat And Livestock,

Section 17. Agriculture, 2012 Statistical Abstract of the United States, U.S. Census Bureau.

Other categories are excluded from the analysis due to lack of the corresponding weights or the large number of states with zero export value of those categories of commodities. Also excluded, for each of the included categories, are those states that have zero export value in any one of the

17 years. Two reasons account for this: 1) log-linear models require nonzero data for dependent variables; 2) the total export values of states with zero export value in at least one year are all very small – thereby making themselves inconsequential for the analysis. Data about the status of states regarding the Real ID Act (i.e., compliant or not, dates of issuing Real ID credentials) are obtained from DMV.ORG, a privately owned website. The website of the Bureau of Economic

Analysis (BEA) provides 2000 – 2016 state GDP in current dollars for the industry of “Farms” that includes both crop and animal production. The website of “Bureau of Labor Statistics, U.S.

Department of Labor” provides 2000 – 2016 estimated total employment in “Farming, Fishing, and Forestry” occupations by state. 13 missing data of the ag-occupations are filled in with state average values. The final data set contains seven categories of commodities, the dummy for state status, the “Farm” GDP and the ag-occupation estimates. There are separate regressions for each of the commodities on the 17-year balanced panel data. Table 3 presents summary statistics for the variables used in this study.

15

Table 3 – Variables and summary statistics Description (millions of $) Mean S.D. Min. Max.

Dependent Variable

Ft Exports for fruits and tree nuts (N=595) 308.63 1377.68 2.34 12948.32

Fv Exports for fresh vegetables (N=578) 51.32 132.46 0.08 1091.60

W Wheat exports (N=714) 151.40 247.04 0.39 1689.25

C Corn exports (N=697) 195.15 370.55 0.81 2544.73

S Soybean exports (N=527) 444.26 644.84 0.38 3572.05

Bv Exports for beef and veal (N=850) 80.21 152.32 0.02 1129.27

P Pork exports (N=816) 82.45 229.27 0.02 2319.23

Independent Variable

CState 1 if issuing the Real ID compliant credentials; 0 0.12 0.32 0 1

otherwise (N=850)

FGDP GDP for the industry of “Farms” (billions of current $) 2.44 3.34 0.009 28.9

(N=850)

AgOccup Estimated total employment in “Farming, Fishing, and 0.88 2.63 0.01 21.54

Forestry” occupations (ten thousand people) (N=850)

Note: Data of GDP for the industry of “Farms” (as originally measured in millions of current $) and data of ag-occupation (as originally measured in one person) have been scaled in units of billions of current $ and ten thousand people, respectively.

Econometric Model

Following the methodology from Hoynes and Schanzenbach’s (2009) study of the food stamp program, a DID model is employed, with controls for state and year fixed effects. For state- specific, time-variant characteristics, I add GDP for the industry of “Farm” into the regression equation, because theories in the field of international trade predict that larger states (as measured by their GDP) tend to trade more. Another control variable is the estimated total employment in “Farming, Fishing, and Forestry” occupations. Pew Research Center (2016) reported that unauthorized immigrants are nearly a quarter of the workforce in farming, fishing and forestry occupations in 2016, which mainly consist of agriculture work and that these

16 occupations also have the highest share of unauthorized immigrants in at least 26 states. Despite the lack of the state-level data about the number of unauthorized immigrant farmworkers, the ag- occupation estimates can be considered as a proxy that indicates each state’s unauthorized immigrant population working in the agricultural sector. In particular, the following model is estimated:

ln = + + + + + + (1)

𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑖𝑖𝑖𝑖𝑖𝑖 𝛼𝛼 𝛽𝛽𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑠𝑠𝑠𝑠 𝛾𝛾1𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑠𝑠𝑠𝑠 𝛾𝛾2𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝑠𝑠𝑠𝑠 𝜂𝜂𝑠𝑠 𝛿𝛿𝑡𝑡 𝜀𝜀𝑖𝑖𝑖𝑖𝑖𝑖 where ln is log of the export for commodity i exported from state s in year t.

𝑖𝑖𝑖𝑖𝑖𝑖 𝑠𝑠𝑠𝑠 is an dummy𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 variable equal to one if state s in year t issues the Real ID compliant credentials.𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝐶𝐶

is the “Farms” GDP for state s in year t. is the estimated total employment

𝑠𝑠𝑠𝑠 𝑠𝑠𝑠𝑠 in𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 “Farming, Fishing, and Forestry” occupations 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴for state s in year t. are state fixed effects,

𝑠𝑠 𝑡𝑡 are year fixed effects, and are the error terms. Standard errors are 𝜂𝜂clustered on state to 𝛿𝛿

𝑖𝑖𝑖𝑖𝑖𝑖 control unobservable characteristics𝜀𝜀 that will lead to similar effects in different years for the same state. All estimates are weighted by the corresponding, 2010 state-level production.

Previous studies raised one issue that may undermine the validity of the research design. The

DID approach depends heavily on the assumption that the treatment and control groups experience a common trend in the absence of a policy change, with the control group continuing on the same trajectory but the treatment group changing after the policy change. However, some studies have provided evidence that could render the counterfactual trend in the present case untenable. Bohn et al. (2014) summarized that the 2007 Legal Arizona Workers Act (LAWA), a restrictive state law concerning unauthorized immigrants, may cause those planning to migrate illegally to Arizona to migrate elsewhere, because LAWA has made it more difficult for them to find work there. They also documented a notable and statistically significant reduction in the proportion of the Hispanic noncitizen population in Arizona that matches the timing of LAWA’s

17

implementation. Watson (2013) found that one type of 287(g) agreement3 nearly doubles the

propensity for the foreign-born to relocate within the United States. Pena (2009) found that

personal and community networks are primary determinants of locational choices for Mexican

agricultural workers in the U.S.. The literature above generally showed that illegal immigrant

farmworkers are more willing to work in states with more lenient immigration policies and the

newly arrived or prospective ones are more likely to follow their friends or relatives. In other

words, tightening immigration policies would have impacts on illegal immigrant’s location

choices and force them to migrate. Similarly, when a state implements the Real ID Act, it may

experience a decrease in exports due to a loss of undocumented workers and those workers likely

relocate to states in which the Act has not yet been implemented, possibly increasing exports.

This “spillover” or “leakage” effect biases the DID approach in favor of the counterfactual

assumption. However, such a “spillover” or “leakage” actually is not as expected. Figure 3

shows labor mobility of estimated unauthorized immigrant workforce among U.S. states from

2009 to 2014. As Figure 3 shows, while the number of unauthorized immigrant workforce

generally remained unchanged for most states, four “noncompliant” states – California, Kansas,

Illinois, South Carolina experienced a decrease rather than an increase in the number of unauthorized immigrants in their workforces. It seems that the implementation of the Real ID

Act and the movement of labor do not have the expected relationship and that moving from one

state to another must have been much more difficult than what common sense tells us for those

immigrants. Therefore, the present study assumes the counterfactual trend – the number of

unauthorized immigrant workers for each state remains stable after the Act implementation.

3 287(g) agreement aims to expand the federal government’s enforcement capacities while enabling state, county, and local law enforcement agencies to respond directly to popular concerns regarding illegal immigration (Rodriguez et al., 2010)

18

Figure 3 – Labor mobility of estimated unauthorized immigrant workforce from 2009 to 2014

Source: Pew Research Center estimates for 2009-2014 based on augmented American Community Survey.

19 Empirical results

Estimation results are presented in table 4. As row 1 of table 4 shows, the coefficients on the state-status dummy variables have the expected negative signs for “fruits and tree nuts” and

“fresh vegetables”, and they are statistically significant at the 1% and 5% significance levels, respectively. The coefficients in row 1 for other agricultural products are all statistically insignificant, which is in accord with the hypothesis that there is not a statistically significant relationship between the implementation of the Real ID Act and exports for non-labor intensive sectors that do not depend heavily on illegal immigrant workers. In row 2 of table 4, the magnitudes of the coefficients on the “Farms” GDP are much smaller than that in row 1. Given that the “Farms” GDP have been scaled in units of billions of U.S. dollars, the effect caused by the GDP at the aggregated level on agricultural exports for individual commodities is very small, despite the different significances and signs of the coefficients. As row 3 of table 4 shows, only the coefficients on the ag-occupation for “fruits and tree nuts” and “fresh vegetables” are statistically significant and they both are positive – a remarkable sign that these two commodities

are produced in labor intensive sectors. Holding other factors constant, an ten thousand increase

in the total employment in ag-occupations leads to increases in exports for “fruits and tree nuts”

and “fresh vegetables” by 5.2% and 4.4%, respectively. Overall, the values of R2 are all above

90%, indicating a well-fitted functional form.

20

Table 4 – Estimation results Fruits and tree nuts Fresh vegetables Wheat Corn Soybeans Beef and veal Pork

(lnFt) (lnFv) (lnW) (lnC) (lnS) (lnBv) (lnP)

CState - 0.250*** - 0.128** - 0.085 - 0.040 - 0.055 0.037 0.069

(0.077) (0.056) (0.062) (0.070) (0.053) (0.040) (0.048)

FGDP 0.023*** 0.003 - 0.021 - 0.010 - 0.042*** 0.025*** 0.043***

(0.006) (0.004) (0.017) (0.015) (0.011) (0.009) (0.012)

AgOccup 0.052*** 0.044*** - 0.037 - 0.077 0.160 0.025 - 0.118

(0.011) (0.008) (0.034) (0.080) (0.156) (0.024) (0.085)

Constant 2.230*** - 1.059*** - 0.250** 0.340*** - 0.611 - 2.572*** - 3.556***

(0.064) (0.032) (0.107) (0.067) (0.520) (0.019) (0.035)

R2 0.996 0.992 0.970 0.987 0.986 0.994 0.994

Number of observations 595 578 714 697 527 850 816

States excluded AK, AR, DE, IA, AK, CT, HI, IA, AK, CT, HI, AK, CT, HI, AK, AZ, CA, ID, WA KS, KY, MT, NE, KS, KY, LA, ME, MA, NH, ME, MA, NV, CO, CT, HI, NV, NH, ND, RI, MT, NE, NH, RI, VT NH, RI, VT ID, ME, MA, SD, TN, WY ND, OK, SD, MT, NV, NH, VT, WV, WY NM, OR, RI, UT, VT, WA, WY Note: Each parameter is from a separate regression of the outcome variable with state fixed effects and year fixed effects, and the standard errors are clustered on state. Standard errors are in parentheses. All estimates are weighted by 2010 commodity-specific production. * p < 0.1, ** p < 0.05, *** p < 0.01

21

Conclusions and discussions

The basic purpose of this study is to quantify the economic impact of the implementation of

the Real ID Act on U.S. agricultural exports. The results are generally consistent with expectations. The hypothesis was that compliance with the Act impacts significantly and negatively on exports of agricultural commodities produced in the labor intensive sectors of fruits and vegetables, where most illegal immigrant farmworkers are hired, and that it does not have a statistically significant impact on agricultural exports for non-labor intensive sectors. In particular, the Act enforcement is estimated to reduce exports for “fruits and tree nuts” and

“fresh vegetables” by 25.0% and 12.8%, respectively. As expected, exports for crops that are largely automated and exports for animal products do not change in a statistically significant manner, suggesting weak dependence on illegal immigrant workers.

The implications of these results suggest a way of legalizing these illegal immigrant farmworkers. For decades, the shortage of labor force has become a critical problem that haunts certain sectors of U.S. agriculture. Unsteady federal immigration policy and enforcement put

U.S. farmers in a dilemma – whether or not hire an immigrant worker when realizing that his or her documentation may not be valid, and those immigrant workers’ working conditions and benefits cannot be improved. The saving grace for U.S. agriculture has been the H-2A . It provides a temporary visa to foreign workers. When the job is over they return to their homes. However, this program is limited to temporary workers and it is too costly for some U.S. farmers. Given the strong connection between farm labor supply and agricultural economy, creating a process through which those unauthorized immigrants can work legally, at least giving them basic rights for living, such as issuing identification card or driver’s license, would stabilize the agricultural workforce and enhance U.S. food security.

22

The present study has its own limitation. The analysis is limited only to the agricultural sector.

However, most illegal immigrant workers are hired in the industries of construction and services.

If further studies could compare results to that in the sectors of construction, hotels and restaurants, like Fan et al.’s (2016) recession-farmworker study, it would provide a better and

more comprehensive understanding of the consequences caused by the implementation of the

Real ID Act and the conclusions drawn here would be reinforced.

Reference

Bohn, S., Lofstrom, M., & Raphael, S. (2014). Did the 2007 Legal Arizona Workers Act reduce the state's unauthorized immigrant population? Review of Economics and Statistics, 96(2), 258-269. Carroll, D., Georges, A., & Saltz, R. (2011). Changing characteristics of US farm workers: 21 years of findings from the National Agricultural Workers Survey. Paper presented at the Immigration reform and agriculture conference: Implications for farmers, farm workers, and communities. Cianciarulo, M. S. (2006). Terrorism and asylum seekers: Why the real ID act is a false promise. Harv. J. on Legis., 43, 101. Fan, M., Pena, A. A., & Perloff, J. M. (2016). Effects of the great recession on the US agricultural labor market. American Journal of Agricultural Economics, 98(4), 1146- 1157. Faustino, H. C., & Leitão, N. C. (2008). Immigration and trade in Portugal: a static and dynamic panel data analysis. Fletcher, A. (2006). The REAL ID Act: Furthering gender bias in US asylum law. Berkeley J. Gender L. & Just., 21, 111. García, A. (2006). The Real Id Act and the Negative Impact on Latino Immigrants. Scholar, 9, 275. Genc, M., Gheasi, M., Nijkamp, P., & Poot, J. (2012). The impact of immigration on international trade: a meta-analysis. Migration impact assessment: New horizons, 301. Hoynes, H. W., & Schanzenbach, D. W. (2009). Consumption responses to in-kind transfers: Evidence from the introduction of the food stamp program. American Economic Journal: Applied Economics, 1(4), 109-139. Lewis, B., Martinez, R., & Coronado, J. (2017). Farmworkers in Michigan. JSRI Report Report No. 29. East Lansing, MI: The Julian Samora Research Institute, Michigan State University. Martin, P., & Calvin, L. (2010). Immigration reform: what does it mean for agriculture and rural America? Applied Economic Perspectives and Policy, 32(2), 232-253. Orrenius, P. M., & Zavodny, M. (2009). The effects of tougher enforcement on the job prospects of recent Latin American immigrants. Journal of Policy Analysis and Management, 28(2), 239-257.

23

Passel, J. S., & Cohn, D. (2015). Share of unauthorized immigrant workers in production, construction jobs falls since 2007. Pew Research Center Hispanic Trends. Passel, J.S., J. Van Hook, and F.D. Bean. 2004. "Estimates of the legal and unauthorized foreign- born population for the United States and selected states, based on Census 2000." US Bureau of the Census, Washington, DC. Pawasarat, J., and F. Stetzer. 1998. "Removing transportation barriers to employment: Assessing driver's license and vehicle ownership patterns of low-income populations." Pena, A. A. (2009). Locational Choices of the Legal and Illegal: The Case of Mexican Agricultural Workers in the US 1. International Migration Review, 43(4), 850-880. Peters, M. E. (2014). Trade, foreign direct investment, and immigration policy making in the United States. International Organization, 68(4), 811-844. Regan, P. M., & Deering, C. J. (2009). State opposition to REAL ID. Publius: The Journal of Federalism, 39(3), 476-505. Rodriguez, C., et al. 2010. "A program in flux: New priorities and implementation challenges for 287 (g)." Washington, DC: Migration Policy Institute. Ruark, E. A., & Moinuddin, A. (2011). Illegal immigration and Agribusiness: the effect on the agriculture industry of converting to a legal workforce. Washington, DC: Federation for American Immigration Reform. SANDRADANZIGER, M.C., S. DANZIGER, and C. Heflin. 2000. "Barriers to the employment of welfare recipients." Prosperity for all?: The economic boom and African Americans:245. See Homeland Security – Q. How long will an extension last? https://www.dhs.gov/state- extensions See Jeffrey S. Passel and D’Vera Cohn, Size of U.S. Unauthorized Immigrant Workforce Stable After the Great Recession, Pew Research Center, November 3, 2016, available at: http://www.pewhispanic.org/2016/11/03/size-of-u-s-unauthorized-immigrant-workforce- stable-after-the-great-recession/ See Jeffrey S. Passel and D’Vera Cohn, U.S. Unauthorized Immigrant Total Dips to Lowest Level in a Decade, Pew Research Center, November 27, 2018, available at: http://www.pewhispanic.org/2018/11/27/u-s-unauthorized-immigrant-total-dips-to-lowest-level- in-a-decade/ See Marshall Fitz, Putting Undocumented Immigrants on the Road to Citizenship Will Help Congress Overcome a Host of Other Policy Roadblocks (2012), available at: https://www.americanprogress.org/issues/immigration/reports/2012/11/14/44885/time-to- legalize-our-11-million-undocumented-immigrants/ See RICHARD D. MCLELLAN, et al., Michigan Law Revision Commission, 47th annual report (2015-2016), available at: https://council.legislature.mi.gov/Content/Files/mlrc/2015- 2016MLRCAnnualReport_FINAL.pdf See The Migrant and Seasonal Farmworkers’ Workgroup report, March 25, 2013, available at: www.michigan.gov/documents/mdcr/MSFW_Progress_Report_415366_7.pdf Watson, T. (2013). Enforcement and immigrant location choice, Working Papers, No. 13-10, Federal Reserve Bank of Boston, Boston, MA. Zahniser, S., Hertz, T., Dixon, P., & Rimmer, M. (2011). Immigration policy and its possible effects on US agriculture and the market for hired farm labor: a simulation analysis. American Journal of Agricultural Economics, 94(2), 477-482.

24

Appendix: Data Sources

State export data available at: https://www.ers.usda.gov/data-products/state-export-data/

Data about the start dates of issuing Real ID credentials by state available at:

https://www.dmv.org/articles/which-states-are-real-id-compliant

Annual GDP data by state available at:

https://apps.bea.gov/itable/iTable.cfm?ReqID=70&step=1#reqid=70&step=1&isuri=1

Data about the estimated total employments in farming, fishing and forestry occupations

available at: https://www.bls.gov/oes/tables.htm

Data about occupations with highest shares of unauthorized immigrant workers available at: http://www.pewhispanic.org/2016/11/03/appendix-d-detailed-tables/

Unauthorized immigrant population trends for states available at:

http://www.pewhispanic.org/interactives/unauthorized-trends/

2010 state-level production for “fruits and tree nuts” and “fresh vegetables” available at USDA

Quick Stats: https://quickstats.nass.usda.gov/

2010 state-level production for “wheat”, “corn”, “soybeans” and “beef and veal” available at

Section 17. Agriculture, 2012 Statistical Abstract of the United States:

https://www.census.gov/library/publications/2011/compendia/statab/131ed/agriculture.ht

ml

25