What Are the Effects of State Level Legislation Against the Hiring of Unauthorized Immigrants?

Very preliminary please do not cite What Are the Effects of State Level Legislation Against the Hiring of Unauthorized Immigrants? Sarah BohnPublic Policy Institute of Californiabohn@ppic.org Magnus LofstromPublic Policy Institute of California and IZALofstrom@ppic.org Steven RaphaelGoldman School of Public PolicyUniversity of California, Berkeley and IZAstevenraphael@berkeley.edu April 22, 2010 Abstract In spite of previous federal reforms, the undocumented immigrant population in this country has continued to grow, currently numbering about 12 million. While reform efforts appear to have stalled at the federal level, states have taken it upon themselves to establish laws against the employment of undocumented immigrants. In this paper we use this recent increase in state legislation, and in particular Arizona s Legal Arizona Workers Act (LAWA), to measure the specific effects of such legislation on labor market outcomes. Our focus on Arizona s legislation is merited by its comprehensiveness, the size of Arizona s immigrant population, and the fact that other states are looking to Arizona as a test case. In addition, the key components of LAWA mandatory use of E-Verify and employer sanctions are likely to be included in future federal reform efforts, and hence the study, once completed, will provide key insights into the intended and unintended consequences of both existing and likely future employer sanction policies. JEL Classification: J15, J18, J21, J23, J38 Keywords: immigration, undocumented, unauthorized, illegal, employer sanctions

1. Introduction In 1986 Congress enacted the Immigration Reform and Control Act (IRCA) with the intention of curtailing the inflow of unauthorized immigrant workers. The framers of IRCA pursued a three-pronged strategy for addressing unauthorized immigration. First, a general amnesty for those undocumented workers meeting certain residence or work requirements wiped the slate clean for the millions of undocumented immigrants already established in the country. Second, the legislation imposed sanctions on employers that hire unauthorized immigrants, attempting to remove the lure of higher-paying employment for would-be immigrants. Third, the legislation provided more funds for border enforcement, making it more difficult to enter the country without proper documentation. By all measures, IRCA failed to achieve its key objectives. Since IRCA s enactment, the undocumented immigrant population has grown from about 3 million to roughly 12 million (Passel and Cohn, 2009). Moreover, the verification and employer sanctions provisions of the law have failed to prevent unauthorized immigrants from working in the U.S. Despite several attempts at immigration reform, Congress has not been able to pass immigration legislation addressing the increase in the unauthorized immigrant worker population. Recently a number of states have taken up the issue of undocumented immigration, passing and implementing legislation that enhances the federal prohibitions against hiring unauthorized workers. Arguably the most restrictive of such state legislation is Arizona s Legal Arizona Workers Act (LAWA). LAWA was passed in July 2007 and implemented in January 2008. This paper aims to assess LAWA s impact on the most relevant groups of Arizona s labor force low-skilled workers. The labor market effects of state legislation aimed at reducing the employment of unauthorized immigrants are potentially widespread and within state possibly greater than those stemming from federal employer sanctions, due to the possible migration of workers and firms across state borders. However, although labor demand theory yields unambiguous predictions with regards to the labor market impacts for the undocumented, it provides ambiguous predictions for legal workers that depend on their degree of substitutability with undocumented labor. Moreover, the potential for statistical discrimination against authorized workers that may be perceived as unauthorized (for example, naturalized Hispanic immigrants) complicates a priori judgment as to the impacts of these laws. Although there is a significant literature evaluating the labor market impact of IRCA (e.g. Rivera- Batiz, 1999; Kossoudji and Cobb-Clark, 2002: Bansak and Raphael, 2001: and Amuedo-Dorantes, Bansak, and Rafael, 2007), to date there is no research assessing how these state laws are impacting the labor market outcomes of immigrants and the native born. Enforcement against the employment of undocumented immigrants is likely to play a key role in federal immigration reform. The results in this 1

paper can thus inform federal policymakers of the impacts of such legislation. Given the uncertain timing of federal reform, however, states continue to legislate on undocumented immigrants, and this research will inform state lawmakers who are considering initiating or enhancing their policies. In this paper we analyze the labor market impacts of state level legislation, specifically LAWA. In particular, we assess whether such legislation reduces employment levels, individual employment probabilities, and wages for foreign-born non-citizens most likely to be undocumented. We also analyze whether the labor market outcomes of observably documented workers, such as naturalized immigrants or the native-born, are impacted by these laws. We employ a series of quasi-experimental estimators using both within-state comparisons as well as comparisons across states to analyze the impact of the Arizona legislation. Our preliminary findings indicate that Arizona s 2007 legislation has its intended consequence of reducing employment of unauthorized workers, but also induced a significant population shift out of the state. LAWA may have had the unintended consequence of improving employment opportunities for low-skilled workers who remain in the state, even for unauthorized. The paper proceeds as follows. Section 2 describes the state legislative environment and features of LAWA. Sections 3 and 4 present theoretical labor market impacts of such legislation, and our empirical methodology for examining these impacts, respectively. Section 5 describes our data. Section 6 offers our preliminary findings, and 7 concludes. 2. State-Level Immigration Legislation Recently, there has been an unprecedented level of state legislative activity in the immigration policy domain. In 2009, state legislatures passed 333 immigration-related pieces of legislation, compared to only 38 during 2005. Although not all of these bills were designed to directly limit employment opportunities of undocumented immigrants, a number of states have passed legislation that does. In 2009, 14 states enacted 21 laws related to the employment of immigrants, up from only 5 laws enacted in 2005. Over 2005 to 2009, a total of 91 laws were enacted in 34 different states related to the employment of immigrants (National Conference of State Legislatures, 2009). The set of state legislative activity is broad and touches on several policy areas, including employment, health, education, identification, licenses, law enforcement, voting, and human trafficking. In this project we focus on legislation pertaining to employment, and in particular, those regulations that 2

target employers. A total of sixteen states have enacted immigration legislation intended to limit the hiring of undocumented workers. 1 The new state laws related to employment of undocumented immigrants vary greatly in terms of restrictiveness and implementation. Most of the laws pertain to specific features of employment law, such as licensure, minimum wage, day labor centers, and unemployment insurance. However, a number of the laws relate directly to overall employment of undocumented immigrants. A common feature of many of these new laws is increased use of E-Verify, a federal online work verification system, 2 and penalties for immigrants working illegally and for the employers who hire them. Colorado was the first of states to pass such legislation. Colorado s law requires any person or entity that has entered into a public contract with the state on or after August 2006 to certify that it has verified the legal status of all new hires using the federal government s E-Verify program. Similar laws or executive orders were enacted in Georgia in July 2007, Rhode Island in 2008, Minnesota in 2008, Missouri in 2009, and Utah in 2009. On the reverse side, Illinois passed a law in 2007 prohibiting employers from participating in the E-Verify system, although the statute was later struck down by a U.S. District Court. Penalties for violations of hiring undocumented immigrants in these laws vary, from fines in the order of $5,000 to business license suspension. South Carolina and Mississippi have legislation that phases in E-Verify use according to firm size. In South Carolina, employers of all sizes are phased in by July 2010. In Mississippi all are legislated to be phased in by July 2011. The penalties for hiring undocumented workers are more stringent under the Mississippi law, and include felonizing employment for undocumented workers and business license revocation. Oklahoma constitutes a special case. While the first phase of the legislation was scheduled to go into effect in November 2007, a court challenge has held up implementation. To date, Oklahoma has yet to implement the provisions of its bill. Arizona has enacted the most comprehensive legislation to date. The Legal Arizona Workers Act (LAWA) was signed into law in July 2007 and enacted on January 1, 2008. LAWA mandates the use of 1 Arizona, Arkansas, Colorado, Georgia, Idaho, Minnesota, Mississippi, Missouri, Nevada, Oklahoma, Rhode Island, South Carolina, Tennessee, Utah, Virginia and West Virginia. 2 The federal government has developed an online system called E-Verify, to verify workers I-9 information against Social Security and US Citizenship and Immigration Services (USCIS) databases. E-Verify completes two verification tasks: authentication of identity and verification of work authorization. The system functions for all workers, both citizens and non-citizens, and is generally intended for authorizing new hires, not existing employees. As of February 2009, over 100,000 businesses were registered with E-Verify. E-Verify provides fast results, if identity and work authorization are confirmed. When confirmation is not granted ( tentative non confirmation ), the employee may appeal and procedures are in place to protect workers pay during the appeal process. The main problems identified are delays in correcting tentative non confirmations, erroneous confirmations, and capacity to handle enrollment by more and more employers (Rosenblum, 2009; Westat, 2007; Meissner and Rosenblum, 2009). Use of the system is voluntary, except for firms with certain government contracts as of September 8, 2009. 3

E-Verify by all employers in Arizona to ensure the legal authorization to work for all new hires. The law imposes sanctions on employers who knowingly hire undocumented immigrants, including a business license suspension for the first offense and revocation upon a second. LAWA is distinctive among recent state legislation on the employment of undocumented immigrants in that it covers all firms, rather than only public firms or those with public contracts. Because of its broad range, LAWA is the only state-level legislation that currently reflects reform proposals at the federal level. Also, because all employers were subject to LAWA as of January 1, 2008, sufficient time has passed so that its effectiveness can be evaluated. In this study, we use the most recent data available to assess both the intended and unintended effects of LAWA. As mandates are rolled out in other states over the next two years, there may be potential for further estimation of the labor market impacts of similar legislation. 2.1 The Legal Arizona Workers Act The Legal Arizona Workers Act was signed into law in July 2007 and enacted on January 1, 2008. LAWA mandates the use of E-Verify by all employers in Arizona to ensure the legal authorization to work for all new hires. The law imposes sanctions on employers who knowingly hire undocumented immigrants, including a business license suspension for the first offense and revocation upon a second. LAWA is distinctive among recent state legislation on the employment of undocumented immigrants in that it covers all firms, rather than only public firms or those with public contracts. Our objective is to identify the labor market impacts of this state legislation. The empirical approach, presented below, hinges on the assertion that Arizona s LAWA can be viewed as a natural experiment. LAWA was enacted at a time of declining labor market conditions. However, a number of facts lead us to believe the policy change was exogenous to labor market conditions in the state and US. Arizona s legislature and populace had been confronting issues related to immigrants, and particularly the employment of undocumented immigrants, well before the turnaround in labor market conditions. Ballot measures aimed at undocumented immigrants were voted on in 2004 and 2006, and bills were introduced in the state legislature at least as early as 2004. In 2005 and 2006 then-governor Janet Napolitano vetoed two immigration-related bills, arguably for being too weak on employer sanctions for hiring undocumented workers. Finally in June 2007, the legislature passed LAWA, with tough sanctions on employers and the bill was signed into law by the Governor. These prior policy actions suggest that the final passage of LAWA was related to long-standing immigration policy concerns rather than current economic conditions. 4

In addition, there was uncertainty on whether LAWA would be enacted on January 1, 2008, given the legal challenges it faced. Federal lawsuits challenging the constitutionality of LAWA were brought by an alliance of civil rights advocates, business interests and immigrant rights groups. The challenge was dismissed, but not until early December. Anecdotal evidence suggests that those likely to be affected by LAWA were waiting on the outcome of the legal challenge before responding (see The Arizona Republic, October 8, 2007). Although this evidence suggests the enactment of LAWA was not endogenous with the widespread recession that hit in 2008, the concurrence of the two events complicates identification of LAWA s labor market effects. Our approach to this problem is detailed in section 4 below. Lastly, Arizona s enforcement measures operate in the context of federal enforcement actions. We are concerned about misidentifying impacts of changes in federal enforcement with changes due to LAWA. However, to-date all evidence suggests that LAWA was exogenous to federal enforcement policy changes. We examine two spheres of federal enforcement: worksite raids and other interior enforcement actions by Immigration and Customs Enforcement (ICE), and border enforcement controlled by Customs and Border Patrol (CBP), both agencies within the Department of Homeland Security. ICE has increased interior enforcement activity since 9/11, roughly, but there is no evidence from our preliminary evaluation of their press releases that these raids were (1) disproportionately focused on Arizona or (2) that increases occurred at the same time as LAWA s enactment. Similarly, there appear to be no significant changes in border enforcement in Arizona relative to other border states simultaneous with LAWA s enactment. 3. The Effect of State-Level Immigration Legislation on State Labor Markets The intention of Arizona s LAWA was to deter the hiring of unauthorized immigrant workers by making it more expensive for employers to do so while simultaneously making Arizona less attractive to both the existing and potential future unauthorized immigrants. A primary motivation behind such legislation is to improve labor market opportunities for those legally eligible to work in the U.S. Whether such legislation is effective in this regard depends in large part on the degree to which authorized and unauthorized workers compete with one another in the labor market. This in turn is likely to vary by skill, nativity, ethnicity, and the interaction of these dimensions. To the extent that employers can distinguish undocumented workers from documented workers, state laws punishing employers who hire workers targeted by the legislation can be modeled as increasing the relative costs of hiring unauthorized workers. Standard labor demand theory predicts that this would induce two reinforcing effects on the demand for undocumented immigrants and two offsetting effects on demand for other workers. Regarding the targeted workers, legislation increasing the relative costs of 5

hiring unauthorized workers should induce both employer substitution towards other workers as well as a general reduction in employment associated with the higher compliance costs (e.g., the costs of additional verification requirements, the expected value of the penalty if caught). These substitution and scale effects should unambiguously reduce demand for undocumented workers, lowering employment levels and perhaps average wages. (Assuming substantial out-flow of unauthorized workers, remaining immigrants lacking legal status would experience an upward pressure on wages). However, theory does not yield an unambiguous prediction for other groups of workers. To the extent that authorized immigrants and specific groups of the native-born are easily substitutable in production for undocumented immigrants, employer substitution will boost labor demand for such workers and ultimately impact their average employment and earnings. However, the size of this substitution effect will be smallest for those who are the least like the unauthorized. In fact, there are certain labor groups that are likely to be complementary in production with unauthorized labor, implying that irrespective of the scale of production in the economy, an increase in the cost of hiring unauthorized workers would actually harm these natives. For all workers, even those that are close substitutes for the unauthorized, higher compliance costs should result in a reduction in the overall scale of an organization s activities. Hence, offsetting substitution and scale effects do not permit prediction of the impact of these laws on documented immigrants and the native born. Theory does suggest however, that with perfect information on legal status those authorized workers most similar in skill to the unauthorized stand to gain the most. When employers cannot easily distinguish the documented from the undocumented, the potential impacts of such state legislation are complicated by the likelihood that employers form probabilistic assessments of the legal status of specific applicants and then act on those assessments. Specifically, if employers cannot tell with certainty who is and who is not authorized, employers may infer legal status through such visible signals as ethnicity, accent, or surname. Some employers wishing to reduce the probability of being fined or of having their business licenses revoked may avoid hiring applicants from specific groups altogether (for example, an outright ban against hiring Hispanic applicants). Given the imperfect signal of legal status, such statistical discrimination would likely impact both native and foreign-born applicants authorized to work in the U.S. Of course, such statistical discrimination is most likely to impact nativeborn and legal foreign-born Hispanics. Concern that sanctions targeted towards the hiring of unauthorized immigrants would result in employment discrimination against Hispanics is not new. In fact, this was a key point of contention in the legislative debate surrounding the 1986 IRCA. IRCA created the legal obligation for employers to verify the identity and work eligibility of all applicants, and introduced a series of graduated sanctions for 6

employers who failed to comply with the verification, record keeping, and hiring provisions of the law. Opponents of the legislation expressed concern that these provisions would lead to discrimination against all Hispanics. Indeed, there is some research finding that in the immediate aftermath of IRCA s passage, the hourly earnings of Latinos did decline (Bansak and Raphael, 2001). A key difference between the mid 1980s and the present time, however, is the past absence of a federal system for verifying identity and work eligibility. During the 1980s, employees were required to present one of a number of pieces of identification to comply with IRCA, many of which were little known documents issued by the then Immigration and Naturalization Service (INS). In the current period, the federal E-Verify system is available to employers, which ideally should provide unambiguous and accurate information regarding the work-eligibility of applicants. Several reports, however, suggest that the E-Verify system has its flaws. According to the American Immigration Lawyers Association (2008), the fundamental problem remains the fact that the Department of Homeland Security (DHS) must run the identity data it receives against the Social Security Administration (SSA) database. But the purpose and function of the SSA database was not meant for immigration enforcement. Moran (2007) argues that the SSA database has 17.8 million records that contain discrepancies that could result in a person (immigrant or U.S. citizen) being wrongly identified as not authorized for employment. According to a formal evaluation of the system by Westat (2007), less than 1 percent of natives but almost 10 percent of foreignborn U.S. citizens verified by the federal Web Basic Pilot between October 2006 and March 2007 received an erroneous non-confirmation of work authorization. The potential for statistical discrimination adds a third partial effect of these state laws that again should vary with the observable characteristics of job applicants. In the event that workers receive a socalled tentative non-confirmation of work eligibility, they are given a certain period to contest. Note that employers can verify work eligibility only after they have decided to hire the applicant in question. Furthermore, in the event of a tentative non-confirmation, the employer is not allowed to change the employment status or work conditions while the case is being reviewed. This element adds uncertainty and increases the expected hiring cost, and if relevant possible training cost. Employers wishing to avoid the risk of hiring undocumented workers may reduce their hiring of Latino applicants in particular and increase their hiring of applicants who are deemed lower risk for whatever reason. Since it is impossible to theoretically predict the net effects on employment of the channels discussed here, analyzing the impacts of these laws requires an empirical analysis. 4. Empirical Approach 7

Our theoretical discussion suggests that LAWA should differentially impact the employment and earnings of different groups of Arizona workers. First, we expect to see the largest declines in employment for sub-groups disproportionately comprised of unauthorized immigrants. For example, non-citizen Hispanic immigrants with low levels of educational attainment and limited English proficiency are the most likely to be unauthorized and thus are perhaps the most likely to be adversely affected by the law. Second, theory predicts that those who are most likely to be confused with the undocumented, such as naturalized foreign-born Hispanics or native-born Hispanics may suffer statistical discrimination by employers wishing to minimize the risk of being sanctioned. Such workers, however, are also likely to benefit from employer substitution away from unauthorized immigrants. Low-skilled, non-hispanic native-born workers (for example, African-American and non-hispanic white workers) should not be impacted by statistical discrimination and should benefit from substitution away from unauthorized immigrant labor. However, each of these groups would be adversely impacted by any effect of the legislation on the scale of operations of businesses in the state. With this theoretical ordering of likely impacts, one approach to estimating the impact of LAWA on labor market outcomes would be to compute the pre-post implementation changes in employment rates and wages for each group, and assess whether the relative changes conform to the predictions of theory. For example, our theoretical discussion predicts that the difference between the pre-post change in employment rates for foreign-born non-citizen Hispanics and native-born non-hispanic whites should be negative (that is to say, the employment rate of the first group should have declined relative to the second). We first examine whether there are any signals of an impact from LAWA in aggregate state-level data, which may suggest whether it is possible to identify impacts in individual-level models. Our data reveal evidence that the enactment of LAWA did affect the demographic composition of the working age population in Arizona. Tables 1 and 2 present population and employment estimates for US and foreignborn residents of Arizona and comparison states compiled from the 2000 Census and 2005-2008 ACS. After a period of continuous strong growth, Arizona s immigrant population decreased by 5.6 percent between 2007 and 2008, the period which covers enactment of LAWA (Table 1). In contrast, Arizona experienced continued growth in the US born working age population over the entire period from 2000 to 2008. The rate of the increase in this population decreased somewhat over the period 2000 to 2007 but increased, to some extent, from 2007 to 2008. The employment trends are very similar to the population trends (Table 1). However, when we focus on the employment of the least skilled workers those most likely to be undocumented immigrants and the native-born who compete with them we find that native employment continuously decreased from 2000 to 2007. This trend reversed in 2007, and by 2008 the 8

native employment level in this subgroup increased to its highest level in Arizona since 2000. Quite differently from natives, the employment of the least skilled immigrants increased substantially from 2000 to 2007 but declined in the post-lawa year of 2008. These summary statistics indicate that something occurred between 2007 and 2008 that appears to have differentially affected foreign-born, specifically those with lower levels of skill. Table 2 compares the same outcomes for Arizona s population and workforce to that in neighboring states. Native-born and immigrant populations grew at similar rates in Arizona and its neighboring states over the period 2000-2008. Both areas experienced net decreases in the foreign born population between 2007 and 2008. However, the observed greater decline in Arizona s immigrant population post-lawa (-5.6 percent compared to -1.9 percent in the surrounding states) is noteworthy given that the net increase in Arizona s foreign born population the year before was greater than in the comparison states (6.1 and 1.8 percent respectively). Our data also reveal that while native born employment grew more in Arizona between 2007 and 2008 than it did in the neighboring states, Arizona s foreign born employment numbers decreased while foreign born employment increased somewhat in the comparison states. Arizona s legislation intended to curtail the hiring of unauthorized immigrant workers appears to have made Arizona a noticeably less attractive option among the US undocumented immigrant population, and possibly to immigrants in general. This labor supply effect of the legislation, coupled with an impact on firms willingness to hire unauthorized workers, may also have affected labor market outcomes among its workforce, particularly low-skilled workers. These broad-level impacts do not control for differences human capital across native-born and immigrant populations, nor for differences in population and labor market composition across states. We proceed to our detailed empirical methodology utilizing individual level data. Our empirical strategy revolves around a series of difference-in-difference calculations using alternative groups of adults defined by education, race/ethnicity, nativity, and legal immigration status. We first contrast the pre-post change in employment outcomes for different groups within the state of Arizona, the idea being that the theory predicts the sign of the relative change in employment between groups. The main strength of this strategy is that the alternative groups in the comparison will be subjected to the same economic forces impacting the state s economy, excepting the differential impact of the legislation of course. The main weakness of this strategy is that each comparison group should be impacted by the legislation through the labor demand channels discussed in the previous section; that is to say, it is difficult to identify a within-state comparison group that serves to chart out the counterfactual employment path for the likely impacted group while not being impacted by the legislation itself. Hence, 9

it is impossible to identify the absolute effect of the legislation on the employment outcomes of each group with within-state comparison groups (i.e., we can identify relative effects only). To address this weakness, we also use a series of difference-in-difference estimates using comparison groups from neighboring states. Here, the main strength concerns the fact that the Arizona legislation is considerably less likely to have a direct impact on labor market outcomes for the comparison group. The key weakness of this alternative strategy is that the comparison groups, being residents in different states, may be subject to different temporal economic shocks when compared with workers in Arizona. In addition, if LAWA induces migration away from Arizona and into neighboring states, the across-state effects may be attenuated. Nonetheless, the results of the two strategies presented side-byside should allow us to triangulate whether LAWA measurably impacted the Arizona labor market. 4.1 Within-state Comparisons Our first empirical strategy is to calculate a series of difference-in-difference estimates making a host of within-state comparisons. Utilizing data on nativity, ethnicity/race, and education, we define the following groups of interest 3 : Foreign-born non-citizen Hispanics with a high school degree or less Foreign-born naturalized citizen Hispanics with a high school degree or less Native-born Hispanics with a high school degree or less Native-born non-hispanic whites with a high school degree or less The first group is what we may refer to as the treatment group. No representative data sets allow for identification of undocumented immigrants directly. However, we can refine the definition of this group to include a higher proportion of unauthorized workers, using established research on this population. Unauthorized immigrants are most likely to be younger, recently arrived, of Hispanic origin with less than a high school degree, and limited English proficiency (Passel and Cohn, 2009). We test the sensitivity of our estimates to changes in the group definition. Our principal within-state comparisons involve relative changes in employment rates and earnings among these four lesser-skilled sub-groups. The empirical specification for the within state difference-indifference estimator comparing the low-skilled group most likely to be adversely affected by LAWA 3 Unfortunately, the small sample size for African-Americans in Arizona precludes us from a reliable analysis of this important minority group. 10

(recently arrived foreign-born non-citizen Hispanics with limited English proficiency) to another lowskilled group can be expressed as: where, y j ' j j it = α + β Xit + γ0lsgi + γ1postit + γ2lsgi POSTit + ε it (1) j y it is the labor market outcome of interest (i.e., variables indicating employment or log hourly wages) for individual i in comparison group j, at time t, X it is a vector of socioeconomic characteristics, such as age, schooling level, marital status, family composition, as well as MSA fixed effects and possibly industry fixed effects, LSG is a less-skilled group indicator, POST is a pre-post dummy, ε it is an error term, and α, β, γ 0, γ 1 and γ 2 are parameters to be estimated. The key parameter in this equation, γ 2, measures the degree to which the pre-post change in the employment outcome observed in Arizona departs from that observed for otherwise similar individuals in the comparison group in Arizona. Theoretically, we expect this coefficient to be negative for those likely to be harmed by the law and positive for those likely to benefit. In most specifications we utilize data over the period 2005-2008, and include POST dummies for both 2007 and 2008. LAWA was implemented on January 1, 2008, but debate and passage of the law took place during much of 2007. Thus, including 2007 in the pre period would be inappropriate since workers or firms may have adjusted their behavior in anticipation of LAWA s enactment. By breaking out γ 1 and γ 2 into 2007 and 2008-specific effects, we may shed light on how LAWA s targeted population responded. 4.2 Cross-state comparisons None of the possible within-state comparisons are likely to identify comparison groups that are not impacted by LAWA. If the law is achieving its intended impacts, all low-skilled workers regardless of race, ethnicity, or legal status will be impacted. Among the higher-educated workers, employment outcomes may still be impacted by LAWA through establishment scale effects or through skill complementarities with low-skilled unauthorized workers. Also, low-skilled and high-skilled workers are concentrated in different occupations and industries and may not experience business cycle swings equally. Hence, the strategy based on within-state comparisons alone is limited. To address this weakness, we also compare the changes in employment outcomes for Arizona workers in each of the low-skilled groups identified above with the comparable groups of workers residing in Nevada, New Mexico, Utah as well as Southern and central inland metropolitan statistical areas of California. 4 Each of these southwestern areas has substantial foreign-born Hispanic populations 4 We include MSAs: Bakersfield, Fresno, Merced, Modesto, Riverside-San Bernardino, Stockton, Visalia-Tulare- Porterville. 11

and is likely to be exposed to similar regional economic forces. While migration from Arizona in response to the legislation may also drive down employment and earnings in these neighboring regions, we believe that such a response is likely to be small relative to the total impact in Arizona. Moreover, the direction of the impact of such a migratory response should unambiguously be negative. Hence, the estimated effects on earnings and the probability of employment of the Arizona legislation using these cross-state comparisons will be lower-bound estimates. Adopting equation (1) to the across-state comparisons is trivial. Calculating the regression adjusted relative change in employment across states requires replacing the comparison group dummy variable with an Arizona dummy and estimating the equation with the sample restricted to the group of interest only. In the following model, the dummy variable POST it for person i in month t is equal to one for observations from months following the implementation of the Arizona legislation and the variable AZ it is an indicator variable for observations in Arizona. The model is estimated separately for each of the four population groups defined previously. This yields the cross-state within population difference-indifference estimate of the impact of LAWA: y = α + β ' X + γ AZ + γpost + γ AZ POST + ε (2) it it 0 it 1 it 2 it it it The key parameter in equation (2), γ 2, measures the degree to which the pre-post change in the employment probability observed in Arizona departs from that observed for observations in the same group in the comparison states. Theoretically, we expect this coefficient to be negative for those likely to be harmed by the law and positive for those likely to benefit. 4.3 Triple difference estimations In addition to estimating all of the possible permutations of the difference-in-difference estimators in equations (1) and (2), we formally contrast the relative changes in outcomes occurring within Arizona to comparable relative changes in our chosen control states. For example, suppose we were interested in assessing whether the change in relative employment rates observed by contrasting foreign-born non-citizen Hispanics to native-born African-Americans in Arizona differs from the comparable relative change in the neighboring southwestern comparison states. Adopting this comparison of relative changes to a multivariate regression framework requires use of a difference-indifference-in-difference estimator. In doing so, we can combine the advantages of the within- and acrossestimators. Specifically, define the variable Hisp it as an indicator variable for our treatment population group. Restricting the data set to the treatment group and one of the other three comparison population groups, we will estimate the equation: 12

y = α + β' X + γ AZ + γ POST + γ Hisp + δ AZ POST + δ AZ Hisp + it it 0 it 1 it 2 it 0 it it 1 it it δ POST AZ + φaz POST Hisp + ε 2 it it it it it it (3) where all of the variables are defined as in (1) and (2). Here, the key parameter on the triple interaction term, φ, measures the extent to which the relative change in employment in Arizona (e.g. Hispanic noncitizen vs. Hispanic naturalized) differs from the relative change observed in the comparison states. Similar to the difference-in-difference estimators, this parameter should be negative for those groups harmed by LAWA and positive for those groups that benefit. The above equations form the crux of our identification strategy. Given the relative strengths and weaknesses of the alternative approaches we explore, we will pay particular attention to extensive robustness checking of our findings. We will establish that the estimated pre-post LAWA changes are not due to trends in differences across states or groups. This will include tests for observed pre-post changes in the outcomes before legislation was passed. Given the possibility of variation in enforcement within Arizona, including particularly vigilant enforcement in Maricopa County, we will also test for within state post-lawa changes. 5 In addition to the above efforts to validate the approach, in future work we will also explore empirical specifications allowing us to test for breaks in the linear trends of outcomes associated with LAWA. 5. Data We rely on individual level survey data from two sources: the 2006-2008 American Community Surveys (ACS) and the 2006-2009 Current Population Survey (CPS). Both the CPS and ACS data provide information on the employment and demographic characteristics of individuals with state-level geographic identifiers as well as information on race/ethnicity, education, age, other demographic characteristics, and most importantly legal immigration status (i.e. naturalized citizen or not). Summary statistics of individual characteristics for the four population groups defined in the previous section are 5 Our identification strategy could potentially be threatened by other immigration enforcement in the state happening concurrently to LAWA. The Minutemen organizations are active in Arizona, particularly in the border regions. In addition, the Sheriff of Maricopa County, Joseph Arpaio, has become nationally known as tough on undocumented immigration. Although these efforts are focused on illegal border crossers, human smuggling, and immigrants arrested for crimes, it is possible that such enforcement could affect the outcomes we are interested in measuring in this study. Our preliminary review of non-employment type enforcements does not suggest a change in the trend or level concurrent with the enactment of LAWA. These enforcements appear to have taken place at different levels in different parts of the state. Thus, as a robustness check to our identification strategy we can use variation in labor market outcomes for individuals across counties or metropolitan areas. 13

shown in appendix Table A1. The ACS is arguably the preferred source given its larger sample size as well as information on English ability used to more closely identify undocumented immigrants. A limitation with the currently available ACS data is that the most recent year for which ACS is available is 2008 and that earnings and wage information refers to the previous year. To best address these issues, we use the ACS to analyze employment, which is reported at the time of the survey, and we use the CPS to analyze LAWA s impact on wages. We rely, in particular, on the Merged Outgoing Rotation Groups (MORG) CPS data for our analysis of wage effects. The MORG is an extract of monthly CPS data for households that are asked about earnings and hours worked (in the fourth and eighth months), and is available through 2009. Though the sample size is considerably smaller than in the March CPS or ACS, the MORG affords higher frequency data and the most recent wage data. When pooled across 2005-2009, the sample size is large enough to conduct a reasonable analysis of LAWA s wage impacts. However, another drawback of using any CPS data is the absence of information on English ability, which is one key factor in defining our treatment group. We will test the sensitivity of our results to changes in this definition. Also by definition, the MORG contains repeated individuals (up to twice). We adjust our standard errors appropriately. For a better understanding of recent trends, we also use data from the 2000 Census. Lastly, all our analyses are restricted to men of working age, defined here to be individuals between 16 and 60. 6. Preliminary Empirical Results Before we turn to our difference-in-difference analysis of employment rates and hourly wages, we explore the possibility that the severe current housing related recession may have impacted unauthorized immigrants more than other groups and that Arizona may have been differentially affected. If so, this may challenge the reliability of our approach in providing estimates of LAWA s labor market effects. To address this issue we look at employment numbers by nativity/ethnic groups and industry. In particular, we focus on the most common industries in which unauthorized immigrants are employed: agriculture, construction, manufacturing, landscaping/building services, and restaurant/hotel services. Employment in Arizona of individuals in our treatment group decreased sharply between 2006 and 2008, by about 15 percent (Table 3). This is a larger drop than any of the low-skilled comparison groups; +2.4 percent for naturalized Hispanic immigrants, +18.8 percent for native born Hispanics and - 6.2 percent among native born non-hispanic whites. We observe decreases in post-pre employment for all four low-skilled groups in the comparison states. It is noteworthy that the employment decline of native born non-hispanic whites in these states, 6.0 percent, is nearly identical to those in Arizona. 14

Construction is the largest industry, both before and after LAWA, among Arizonans in the treatment group. The construction employment decline is substantial (35.3%) and the largest in levels but smaller in percent than the decline in agriculture (51.9%) and manufacturing (51.0%). The decline in construction employment of workers in the treatment group in Arizona is nearly identical to the employment decline of non-naturalized Hispanics in the comparison states. Employment of immigrants in this group increased in some industries in Arizona (landscaping/building services, restaurant/hotel services and in the remaining/other category). These were not sufficient to offset the decreases in the other industries. Employment of individuals in the treatment group residing in the comparison states also increased in these industries with the exception of landscaping/building services. Overall, the data suggest that the greater employment decline of non-naturalized recent immigrants in Arizona is not driven by the housing crisis and the related severe slowdown in construction, in this state. Compared to their counterparts in the surrounding states, the greater loss in employment in Arizona for this group was mostly due to the greater decline in employment in agriculture and manufacturing. Lastly, Table 3 provides information consistent with other Hispanic low-skilled workers in Arizona filling jobs post-lawa that were previously occupied by unauthorized immigrants. For example, construction employment of low-skilled native born Hispanics in Arizona increased by 14% post-lawa. Employment of workers in this same group, but residing in the comparison states, decreased by about two percent. 6.1 Employment Rates We begin by presenting raw employment rates for our four less-skilled population groups in Arizona and comparison states. The employment rate of low-skilled non-naturalized immigrant men in Arizona (our treatment group) continues to be very high after the enactment of LAWA (88.4 percent, Table 4, panel 1). However, unlike low-skilled naturalized immigrant and native born Hispanic men, nonnaturalized Hispanic immigrants experienced a decline in the employment rate between 2006 and 2008 (2.1, 4.2 and -2.9 percentage points respectively). The within Arizona across group differences in the post-pre LAWA changes in the employment rates provide us with raw estimates of the relative impact of LAWA. The estimated unadjusted LAWA effect on employment rates show that LAWA reduced the employment rates of individuals in the treatment group by 5 percentage points when compared to low-skilled naturalized Hispanic men and by 7 percentage points when compared to native born Hispanics. The estimated negative impact of LAWA is however only statistically significant when native born Hispanics are the comparison group. 15

When differences across groups in employment-relevant characteristics are accounted for by implementing equation (1), the estimated negative LAWA employment effect is somewhat smaller (Table 5). We estimate that the employment probability of our treatment group fell 6.6 percentage points relative to native-born Hispanics. This estimate is statistically significant. As in the unadjusted case, our regression-adjusted LAWA effects are smaller, and insignificant, when the treatment group is compared to naturalized Hispanic immigrants (-4.9 percentage points). Surprisingly, the data fail to reveal a relative LAWA effect when non-hispanic whites serve as the comparison group. The estimates of changes in employment rates from 2006 to 2007 in Table 5 are smaller and insignificant in all cases. The negative sign is consistent with the hypotheses suggesting LAWA may have a chilling impact on the labor market, even before the official enactment date. However, these estimates are too imprecise to draw any strong conclusions from. We find that these effects are not sensitive to the definition of the treatment group. Table 7 Panel 1 shows the main results from alternative definitions of the likely undocumented group. In parts B and E we expand the definition by including all levels of English proficiency, or all years since migration, respectively. Estimates increase slightly when including other English levels, and decrease slightly when we include older arrival immigrants. Test B is important, particularly, since when using CPS data for the wage regressions below, we are unable to define the group in exactly the same fashion due to data constraints. In test E, we restrict the undocumented group to be high school dropouts only, and find slightly smaller, but still insignificant results. We also find that the results are not sensitive to the inclusion of industry fixed effects, see part F. The statistically significant finding continues to be for the treatment group compared to Hispanic native-born, and this effect is slightly smaller (and marginally smaller standard error) than in our baseline results, but still large in economic impact. The unadjusted across states estimates suggest that the employment rates of individuals in the treatment group increased relative to those of individuals in the same group residing in the comparison states, 2.3% (Table 4). However, the data also show that the employment rate for all three Hispanic groups increased more in Arizona than in the comparison states, although only statistically significantly different from zero among native born Hispanics. Arizonans in this group experienced higher employment rates post-lawa by 9.5 percentage points compared to the individual in this group in surrounding states. We observe roughly the same post-pre change across states among non-hispanic whites. Similarly, the regression-adjusted across state difference-in-difference estimates show no significant difference between the treatment group in Arizona and the treatment group in the surrounding states (see Table 6). Native born Hispanics in Arizona, however, experienced a statistically significant 16

increase in employment probability relative to otherwise similar workers in comparison states, by 7.9 percentage points due to LAWA. Again, we find that these results are not sensitive to the definition of the likely undocumented immigrant group. We still find no statistically significant impacts for the treatment group in Arizona relative to comparison states when the group is defined more or less restrictively (see Table 7 panel 2). When older arrival immigrants are included or the group is restricted to high school dropouts, the estimates are slightly larger but not significant (in parts D and E). When we control for industry using fixed effects (part F), there is some indication of negative impacts for Hispanic naturalized immigrants in Arizona relative to other state, but the imprecision of these estimates allows for no strong conclusions. We have also conducted preliminary tests on the sensitivity of these results to definition of comparison states. From these tests, it appears our baseline estimates are likely to be a lower bound. Table 7 Panel 3 shows the results. In test B, we exclude California, in test C we add Texas, another important border state. In both cases, we find a mixture of positive and negative impacts in various population groups. Similarly, when we use all other states except our comparison group or throw in all states, no clear picture emerges. It appears that the likely undocumented group may have done worse in Arizona than those in comparison states, since these estimates are more likely negative. However, the imprecision and variability of the estimates suggests there is no clear employment impact for these population groups across states. By implementing equation 3, we essentially combine the findings from the within- and acrossstate models. Table 8 presents the triple-difference results for employment. We find that comparing the relative employment probability across states yields similar findings. The negative estimate for undocumented immigrants compared to native-born Hispanics persist, though results are all insignificant. The estimate suggests that the difference in employment probability for undocumented relative to nativeborn Hispanics is 4.5 percentage points lower in Arizona relative to the same difference in comparison states. Taken together, the full set of difference-in-difference estimates suggest that LAWA negatively impacted the employment probability of undocumented Hispanic workers relative to other Hispanic workers in Arizona, but had little, or small positive, impacts relative to similar workers in other states. One possible explanation for these results is that LAWA did not considerably reduce employment opportunities of unauthorized immigrants who stayed in Arizona after LAWA was enacted. Given that the mandate for use of E-verify applies only to new hires and not current employees, this explanation is plausible. At the same time, authorized low-skilled Hispanic workers the closest substitutes to 17