Hostname: page-component-77c89778f8-fv566 Total loading time: 0 Render date: 2024-07-21T07:57:31.277Z Has data issue: false hasContentIssue false

The politics of immigrant policy in the 50 US states, 2005-2011

Published online by Cambridge University Press:  26 February 2013

James E. Monogan III*
Affiliation:
Department of Political Science, University of Georgia, USA
*
James E. Monogan IIIAssistant Professor Department of Political Science University of Georgia 413 Baldwin Hall Athens, GA 30602 USA Email: monogan@uga.edu
Rights & Permissions [Opens in a new window]

Abstract

This article asks what shaped immigrant policy in the 50 states between 2005 and 2011. Theoretically, politicians are influenced by electoral considerations as they craft laws. Law-makers consider both current public opinion and how the electorate is likely to change, at least in the near future. Empirically, the article analyses an original dataset on immigrant-related laws enacted by the states with a Bayesian spatial conditionally autoregressive model. The analysis shows that state immigrant policy is affected primarily by legislative professionalism, electoral ideology, state wealth and change in the foreign-born population.

Type
Research Article
Copyright
Copyright © Cambridge University Press 2013

Introduction

From 2005 to 2011, immigration held a prominent position on state political agendas, with 1,536 new laws being adopted by the states. During this time frame, new admissions of immigrants rose to some of the highest levels in history, and many of the states these immigrants settled in are new destinations that have not seen substantial immigration in recent history. With the Senate's failure to pass the McCain-Kennedy bill in 2007, an issue that substantially lends itself to resolution at the federal level remained largely unresolved. The states therefore had to continue to cope with a large population of undocumented residents, providing social services to immigrants, and incorporating newcomers into the workforce.

One of the more prominent state policies enacted during this time was Arizona's 2010 enactment of Senate Bill 1070. Specifically, the law includes a “show me your papers” provision that makes it a misdemeanour to not carry immigration papers and allows police to conduct immigration status checks (Archibold Reference Archibold2010). The law also makes it a crime for undocumented immigrants to seek or maintain employment, and it allows police to arrest those suspected of committing deportable offences without warrants.

In the case of Arizona v. United States (Reference Arizona2012), the US Supreme Court considered the constitutionality of the 2010 state law. The court upheld the “show me your papers” portion of the law, but declared the provisions that criminalised employment and authorised warrantless arrests to be unconstitutional. Writing for the majority, Justice Anthony M. Kennedy stated:

The National Government has significant power to regulate immigration. With power comes responsibility, and the sound exercise of national power over immigration depends on the Nation's meeting its responsibility to base its laws on a political will informed by searching, thoughtful, rational civic discourse. Arizona may have understandable frustrations with the problems caused by illegal immigration while that process continues, but the State may not pursue policies that undermine federal law.

Although there has been a trend in recent years for more authority on immigrant policy to be shifted to the states, this ruling serves to curb that trend with a reiteration of federal supremacy on immigration policy when the two conflict. Despite this legal development, however, the states still retain and exert considerable influence over immigrant policy. On such an important issue and in such a critical context, what shaped state-level policy at this time?

In answering this question, it is key to observe that state-level immigrant policy has several unique features that distinguish it from other policy areas. Firstly, the politics of the issue are changing over time. Immigration is an issue for which hostile policy positions may be popular in the present, but unpopular in the future. Hence, unlike other policy areas, the degree to which law-makers are forward-looking is important to consider. Secondly, with immigrant policy many of the people directly affected by the laws cannot vote. Hence, law-makers may choose to adopt policy that is hostile towards immigrants with little regard for how immigrants would vote in response. Thirdly, the issue is likely to have a substantial geographic component to it. Neighbouring states are likely to adopt similar immigrant policy to their neighbours, even in the context of a full policy model.

To explain what shapes immigrant policy in the context of these unique factors, this article proceeds, firstly, by discussing background information on immigrant policy. Secondly, it lays out a theory of immigrant policy-making at the state level. It then describes the data gathered on legislative outcomes and the method used to analyse policy tone. Fourthly, it presents the results of the empirical analysis and several descriptive cases of policy adoption to elucidate the results. Finally, the article discusses the implications of this study.

Background on immigrant policy

The role of the states in policy toward immigrants has risen and fallen at different times in American history. For example, until the end of the nineteenth century, immigration was principally the subject of state regulation in the face of a federal vacuum. However, the Constitution prescribes to Congress the power to “establish a uniform Rule of Naturalization”. Based on this constitutional authority, the courts declared state authority on admission and exclusion of immigrants to be unconstitutional once the federal government started taking a more active role in immigration law (Spiro Reference Spiro1997, 1,628).

In the 1990s, though, states again were given the opportunity to shape laws related to immigrants. Specifically, the Personal Responsibility and Work Opportunity Act of 1996 gave individual states considerable leeway in how they chose to administer the Temporary Assistance for Needy Families (TANF) programme. Under this law, states were given the authority to regulate the distribution of benefits, including how generous the states choose to be with immigrants. The Urban Institute has completed several studies on state welfare eligibility requirements and reports that the states differ widely in how stringent they are in deciding eligibility rules with respect to foreign-born individuals (Rowe Reference Rowe2000; Rowe and Versteeg Reference Rowe and Versteeg2005). Hero and Preuhs (Reference Hero and Preuhs2007) further studied this and found that liberal states respond to immigration issues by adopting immigrant-friendly TANF laws; however, a trade-off in policy must be achieved by lowering maximum cash benefits.Footnote 1

A major upshot of legislation such as the Personal Responsibility Act has been that states have been invited back into the fold to enact more laws related to immigrants (Spiro Reference Spiro1997). However, crafting regulations on foreign-born eligibility for welfare benefits differs distinctly from enacting laws regarding legal entry of immigrants and the allocation of visas. Fix and Passel (Reference Fix and Passel1994, 3–16) review the policy context of immigration in the United States and observe that the federal government regulates immigration policy alone, providing a single coherent code on who may enter the United States and in what numbers. However, immigrant policy, or laws related to how immigrants are treated and integrated into society after entry, is shaped by federal, state and local governments. Thus, the renewed attention to immigration by the states is important in how it shapes the lives of both legal and illegal immigrants who are stateside, but cannot directly influence individuals’ ability to enter the nation.

Past studies have shown that the national economic and political context affects whether new immigration measures will be adopted. Daniels (Reference Daniels2005), reviewing the history of American immigration policy, observes that reforms usually arise in response to surges in nativism. Such surges resulted in Chinese exclusion, excluding Haitian immigrants, and a desire to keep Somalians out by the Reagan administration. Further, whenever policy action is taken, the law must satisfy a wide array of concerns, from nativist concerns to a desire to have labour available for farmers. A particularly good example of how this wide range of concerns can delay and then shape policy is the Immigration Reform and Control Act of 1986, or Simpson-Mazzoli Act, which took years to push through Congress (Daniels Reference Daniels2005, 222–224). Further, Citrin et al. (Reference Citrin, Green, Muste and Wong1997) shows that citizens’ views on immigration – in both good times and bad – depends largely on retrospective evaluations of the national economy and anxiety over taxes. Opinion also depends on feelings towards Hispanics and Asians, thus fitting with the idea that nativism prompts salience of the immigration issue.

Beyond these considerations of the economy and nativist concerns shaping policy, state-level laws could be important because they might influence which states individuals choose to migrate to. Borjas (Reference Borjas1999, 1,756) points out that there have been many accounting-oriented studies that offer differing conclusions as to whether immigrants produce a net surplus or deficit. However, changing the rules of the welfare state adjusts the incentives of whether to migrate or not. Therefore, it is important to develop a theory of how the unique context the states face shapes the laws they enact.

State policy-makers and the immigration issue

Compared with other policy areas, the immigration issue is unique because states have particular limitations in what they can do. As has been mentioned, states can do very little to directly regulate border control or levels of new admissions (e.g. immigration policy), as the federal government controls these rules. State laws are principally limited to how legal and illegal immigrants who already reside in a state are treated (e.g. immigrant policy). Despite this, both the Personal Responsibility and Work Opportunity Act of 1996 and the Illegal Immigration Reform and Immigrant Responsibility Act of 1996 created substantial niches of immigrant policy that the states could exert authority over. These federal acts of devolution on immigrant policy, combined with federal inaction at the peak of an immigration wave, seems to have invited states to take matters into their own hands on this issue. Again, however, the case of Arizona v. United States (Reference Arizona2012) clearly limits the states from undermining federal law. Given, though, that the states still have a fair deal of leeway within this legal framework, what motivates policy-makers as they craft legislation in this distinct policy area?

One feature of the politics of immigration that is critical to policy-making is that the political context of immigration issues clearly is changing over time. As legislators create policy, they are ever mindful of how electoral rivals could exploit their voting record during the next campaign (Arnold Reference Arnold1990). Many immigrants are ineligible to vote because they have not completed the naturalisation and citizenship process. Further, many immigrants entered the country illegally, precluding them from eligibility for citizenship. Therefore, if a legislator votes, for example, to exclude illegal immigrants from public benefits, his or her rival in the next election does not have the option of mobilising those who lost benefits. On the other hand, if the legislator voted to instate or maintain programmes that helped illegal immigrants, then an opponent in the next election could gain support by arguing that the incumbent supported policies that imposed unfair costs on taxpayers. Hence, the legislator who is solely concerned about the next election ought to have a slight preference for policy that is more hostile to immigrants, all else being equal.

In the longer run, however, the policy-maker's electoral calculus changes. Anti-immigrant policies easily can be painted as anti-Latino, and the proportion of the population that is Latino is growing rapidly. The larger this proportion of the population gets, the more candidates for office can gain electorally by appealing to Latino Americans. Therefore, years down the road as state legislators and governors seek re-election or even higher political offices (e.g. statewide offices, congressional seats or even the presidency), their opponents will increasingly find it profitable to argue that candidates who supported anti-immigrant policies in the past are anti-Latino. The growth rate of the Latino population suggests that such an anti-Latino portrayal can pose an increasing problem for the prospects of winning an election. Therefore, the more a state policy-maker values winning elections in the long term, the more he or she should be supportive of policy that is friendly to immigrants.

Monogan (Reference Monogan2012) presents a formal model of this type of dilemma. Specifically, in this formalisation, parties must choose a position on an issue that stands for both a current and a future election. Voters choose which party to vote for based on spatial utility (Downs Reference Downs1957; Hotelling Reference Hotelling1929). From the first to the second election, however, the preferences of the electorate shift such that the median voter's issue preference moves to a new location, but the parties are stuck at their old positions. In this game, a key factor in how parties place themselves is how much value they place on the future election relative to the more valuable current election. The result is that the more highly the parties value winning in the future, the closer to the future median voter parties will place themselves.

The politics of immigration fit this scenario. The policies that state legislators vote to adopt now always can be brought up in a future election, be it the very next election or decades into the future. In the present, many immigrants cannot vote and many Latino Americans are not registered to vote. Hence, taking a hard line on immigration may resonate with voters who feel threatened by the growing presence of Latino Americans with little drawback at present. In the longer term, though, having a history of taking a hard stand on immigration could be a losing proposition.

All of this implies that states with legislators who tend to take a longer view of their career should enact more welcoming immigrant policy than states where legislators take a more short-sighted approach. While immigrant policy also may respond to the factors that are known to generally influence state policy, immigration is unique in that legislators’ preferences could change simply based on how concerned they are about their future prospects. Because the focus of this work is at the state level, any assessment of the importance of time perspective must compare states where member legislators are more likely to be short-sighted to states where legislators are more likely to be far-sighted.

Legislative professionalism is one predictor that ought to separate states based on future electoral incentives. Indeed, several studies have shown that legislative professionalism leads politicians to become more career-oriented: Berry et al. (Reference Berry, Berkman and Schneiderman2000) argue that legislative professionalism provides incentives and capacity for members to pursue a long legislative career. For this reason, members of professional legislatures win re-election more frequently than members of non-professional legislatures, and their chances of success are less sensitive to external factors, such as coat-tail effects.

Additionally, Maestas (Reference Maestas2000) argues that professional legislatures attract more members with progressive ambition, or the desire to advance to higher office in a long political career. Finally, Maestas et al. (Reference Maestas, Maisel and Stone2005) show that members of professional state legislatures are significantly more likely to be recruited as candidates for the US Congress than those of less professional legislatures. All of which suggests that the time horizon for policy-makers in states with professional legislatures is longer, so policy in these states ought to reflect longer-term electoral considerations.

Finally, whether a state has term limits for its legislature also speaks to the issue of its members’ time horizon. In states with term limits, members of the legislature still would like to win in the next few elections, but they are precluded from competing in legislative elections in the distant future. Hence, the prospects of winning in the long term would seem to be less important than in states without term limits. While nothing is to prevent members of term-limited legislatures from seeking higher offices, elected officials’ long-term options are certainly more narrow in term-limited states, which ought to generally reduce the value of winning in the future. Overall, then, legislative professionalism and term limits are two means of separating states based on the degree to which legislators are forward-looking.

The current state electoral context

While long-term thinking can be critical to state policy directed towards immigrants, the current electoral context also ought to shape policy in ways that resemble other policy areas. Mayhew (Reference Mayhew1974) makes one of the most prominent cases that legislators’ behaviour is based on the electoral connection, maintaining that elected officials constantly take actions that will raise their favourability with constituents and increase re-election prospects. In a complementary study, Arnold (Reference Arnold1990) argues that law-makers even consider the preferences of an inattentive public when casting votes on policy matters. For example, if an elected official fears that an opponent could blame voters’ problems on a particular policy, then the office-holder will vote against the policy even if the public is largely unaware of the issue at the time. For immigrant policy, several factors – including public opinion, the state economy and change in the foreign-born population – set the stage for how legislators’ votes could factor into the next election.

Firstly, in crafting laws that are electorally viable, state legislators and governors are likely to consider public sentiment. Research on state politics offers substantial evidence that public policy is responsive to public sentiment. Erikson et al. (Reference Erikson, Wright and McIver1993), the seminal work on state opinion-policy congruence, demonstrates that general policy liberalism is responsive to the overall ideology of the state's electorate. Even more deeply, Lax and Phillips (Reference Lax and Phillips2012) examine 39 policies across eight issue areas and find that these very specific policies are responsive to public opinion regarding the policy itself. This study does, however, observe that, despite the relationship between opinion and policy in a general model, policy is only congruent with the majority opinion in a state half the time.

Prior research on immigrant policy specifically suggests that policies align more closely with symbolic ideology than issue-specific ideology. For example, Citrin et al. (Reference Citrin, Reingold, Walters and Green1990) demonstrate that citizens’ votes on California Proposition 63 were primarily a function of political ideology and partisanship, ostensibly because this English-only policy is primarily symbolic. Further, Lax and Phillips (Reference Lax and Phillips2012) analyse four potential immigrant policies that states might adopt: issuance of driving licences to illegal immigrants, prohibition of bilingual education, providing in-state tuition for children of illegal immigrants, and requiring the state government to verify citizenship status before making hiring decisions. The issue of driving licences for illegal immigrants was the only one of these four for which a majority of states (84 per cent) had a policy congruent with public opinion on the issue. For the other three issues, very few states (6 per cent, 18 per cent and 22 per cent, respectively) had a policy consistent with issue-specific public opinion (Lax and Phillips Reference Lax and Phillips2012, 156). Finally, Ramakrishnan and Wong (Reference Ramakrishnan and Wong2010) show that local-level ordinances are strongly responsive to the partisan make-up of a community. While distinct from ideology, this partisanship result further suggests that the electorate's general disposition, rather than issue-specific views, are key to immigrant policy.

Given this prior research that suggests state policy generally, and immigration policy specifically, is responsive to ideological identification, I would expect that electorally-sensitive politicians ought to respond to public ideology in the laws they create. The more conservative a state's electorate is, the easier it is for an opponent to mobilizse opposition to an incumbent who voted for softer immigrant policies. In a liberal state, by contrast, it would be harder to create a coalition of voters opposed to immigrant-friendly policies, so legislators would feel less pressure against voting for welcoming laws or against hostile immigrant laws. Therefore, it is hypothesised that as the public ideology of a state becomes more liberal, immigrant policy should be more welcoming in nature and less hostile towards new immigrants.

While congruence between public opinion and policy outcomes has been observed for many issues, not all scholars believe that this pattern arises because policy-makers are responsive to public preferences. Some, such as Jacobs and Shapiro (Reference Jacobs and Shapiro2000), argue that politicians instead try to lead the public to favour the policies they wish to implement. While this explanation of opinion-policy congruence is possible in general, it seems unlikely in this case because this article operationalises public sentiment with symbolic ideology, which is remarkably stable over time. This operationalisation therefore makes it reasonable to assume that the ideology of the public is exogenous to the policy-making process.Footnote 2

It also should be noted that immigration can be a cross-cutting issue: Tichenor makes this argument by classifying political elites on immigration according to two dimensions – whether immigrant rights should be expanded or limited and whether admissions of new immigrants should be expanded or restricted (Tichenor Reference Tichenor2002, 36). Although this classification can better distinguish immigration preferences than ideology alone, generally liberals are more favourable to expansive immigrant rights and expansive immigrant admissions. Therefore, general public liberalism should effectively explain welcoming policy toward immigrants, be it related to rights or residence. Overall, given prior research on the importance of public opinion for policy-making, ideology is strongly expected to have an important influence on immigrant policy.

The economy and demographics

If elected officials are concerned not just about current opinion, but how voters could be mobilised by an electoral rival, then there are two additional factors that should be important to immigrant policy: the economy and the growth rate of the immigrant population. The economy of a state may influence immigrant policy because law-makers might respond to the economic concerns of their constituents when adopting policy. In fact, literature has shown that the economy is a crucial consideration in public opinion and policy about immigration (Alvarez and Butterfield Reference Alvarez and Butterfield2000; Hopkins Reference Hopkins2010; Martin Reference Martin1990; Yohai Reference Yohai2008). For this reason, I believe the larger a state's per capita gross state product (GSP) is, the more welcoming state policy ought to be.

State wealth captures several factors that may be of concern to election-seeking legislators: Firstly, in wealthier states, the state's median voter is less likely to hold a job that might be threatened by unskilled labourers. Secondly, wealthier states may need to use new unskilled labourers for jobs that residents are unwilling to do. Thirdly, and perhaps most importantly, wealthier states can more easily afford the welfare-related costs of new residents. Though various reports offer conflicting conclusions on whether illegal immigrants produce a net fiscal gain (Smith and Edmonston Reference Smith and Edmonston1997) or loss (Clark et al. Reference Clark, Passel, Zimmermann and Fix1994) for a state, public debates on immigration seem to assume that the public costs illegal immigrants create for health care, incarceration and education outweigh any new revenues they create for the state. In poorer states, it is more likely that voters can be mobilised by claims that welcoming policies produce fiscal losses for the state. All of this suggests that welcoming immigrant policy ought to be easier for law-makers to sell to wealthier constituents. By the same token, in poorer states it is more likely that challengers in elections can raise concerns over welcoming immigrant policies that resonate with voters. Hence, policy should be more welcoming in wealthy states on average.

The second factor worth considering is how the growth rate in the number of immigrants may influence policy. From 2000 to 2008, the number of foreign-born residents in the United States increased 22.0 per cent.Footnote 3 Hence, several states have seen a marked rise in the number of immigrants both in sheer numbers and as a percentage of the population. Historically, immigrants have been received, year after year, by the same states, such as California, Florida and New York (Borjas Reference Borjas1999, 1,738–1,739). However, the states that have seen the largest increases in new arrivals of late have been new destinations – distinct from the places immigrants have moved to in the past. Singer et al. (Reference Singer, Hardwick and Brettell2008) and Zúñiga and Hernández-León (Reference Zúñiga and Hernández-León2005) offer examples of how places like Colorado, Iowa, Nebraska and North Carolina have drawn newcomers with labour needs in agriculture and meat-packing, yet often have then responded to the immigrants with social and political hostility.

Whenever minority groups grow at this rate, it becomes salient to the public, which comes to interact socially and economically with the growing group and also sees new media reports on the changing demographic situation. While Hispanics have formed the largest group of newcomers in many places, there are also areas with large non-Hispanic foreign-born populations that are going to be more visible to the public, such as the Somali refugees in Minneapolis or the large Asian populations in several west coast cities. Considering the growth rate of a state's foreign-born population offers a chance to understand how the size of large, newcoming minority populations influences the way states address the immigration issue.

The public attention that a growing foreign-born population attracts is likely to elicit a legislative response: electorally, members may be inclined to enact new legislation that will be comforting to those who are threatened with the rise of the new group. A rapidly growing foreign-born population with a different racial or ethnic background from a group that forms an electoral majority allows politicians the opportunity to use the foreign-born group as scapegoats for a variety of social problems. By blaming such problems on the out-group (in the current immigration context, likely Latino or Asian Americans) and proposing policies to curb immigration or otherwise diminish the status of the out-group, politicians may garner support from the in-group (white voters, in most constituencies).Footnote 4 This strategy would resemble the politics of racial threat that Key (Reference Key1949) explains was used frequently in the South. Several have argued that this group-based strategy was embraced by California Governor Pete Wilson in promoting Proposition 187 in 1994 (Diamond Reference Diamond1996, Nicholson Reference Nicholson2005, Whalen Reference Whalen2002). Even if state legislators are hesitant to use the politics of racial threat themselves, they certainly may recognise that a growing immigrant population allows electoral opponents the opportunity to use this strategy. At the very least then, policy should be more hostile in states with high immigrant growth because law-makers are afraid of voting for welcoming policies.

As a final point, a state's policy towards immigrants is likely to be similar to the policy that a state's neighbours adopt. A variety of unobserved and unmeasured factors are likely to shape immigrant policy. For example, even beyond what any economic measure could capture, the industries and labour needs of a state are likely to be similar to its neighbours. If two neighbouring states have a large need for work in slaughterhouses, they both might adopt policies that do not require private employers to verify the legal status of their employees. Alternatively, the constituents in neighbouring states may be similarly xenophobic (which is difficult to measure accurately). If state legislators in both states recognise that their constituents have these attitudes, then they are more likely to adopt a hostile immigrant policy in each state. As a final example, there is a good chance that immigrants settle in similar patterns in neighbouring states. If two bordering states each witness newcomers settling into large cities, then their policy response may be different from two bordering states whose immigrants settle in rural towns or become migrant workers. Overall, immigration is an issue for which it is highly likely that the unmodelled variance in policy will be spatially correlated, with neighbours adopting similar policies.

Data and method

With this theoretical understanding of what ought to shape immigrant policy tone, I measured legislative output in the 50 states using summaries of immigrant-related laws gathered by the National Conference of State Legislatures (NCSL).Footnote 5 I coded each legislative action based on two factors: firstly, whether the action had a welcoming or hostile tone towards immigrants and, secondly, on the scope of the action. The scope was coded onto a four-point scale: (1) symbolic, (2) affecting a small group of immigrants, (3) affecting many immigrants in a substantial way and (4) directly affecting immigrants’ ability to reside in a state.Footnote 6 The coding scheme on the scope of the laws was constructed with the assistance of legal scholars and members of immigrant policy think tanks. The scheme combines elements of how the laws are commonly categorised in the substantive study of immigration law with the need to rank the laws by importance.

The following examples illustrate how legislative output is distributed into each of the four scope categories: firstly, an example legislative act receiving the lowest significance score due to its symbolic nature is a 2007 joint resolution by the Illinois legislature (HJR 27). The legislature expressed opposition to a federal law, the REAL ID Act of 2005, that required states to follow federal guidelines when issuing driving licences and ID cards. Since the REAL ID Act was unfriendly towards immigrants, this oppositional resolution is welcoming to immigrants. At the same time, the resolution was merely a position-taking opportunity having no direct bearing on policy, so it is coded as a strictly symbolic act (rated 1 on the four-point scale). Secondly, a law that has a substantive impact, but only for a limited number of immigrants, is a 2007 Michigan law (HB 4207, Act 19) that provides licensure to applicant nurses who are licensed in Canada. This act is coded as welcoming to immigrants and rated 2 on the four-point scale. Thirdly, a law having a substantial impact on a large number of immigrants is a 2008 Kansas law (SB 81) that requires persons to be US citizens or legal immigrants and provide documentation of their status in order to qualify for Kansas's discretionary SCHIP programme, thereby excluding illegal foreign-born children from this health-care programme. This law is coded as hostile to immigrants and rated 3 on the four-point scale. Lastly, an example of a law scored as having the highest substantive significance, as it directly affects immigrants’ ability to reside in the state, is Colorado's 2006 law (SB 90) that requires all state and local law enforcement agencies to report suspected illegal immigrant arrestees to US Immigration and Customs Enforcement (ICE). This law is coded as hostile to immigrants and rated 4 on the four-point scope scale.Footnote 7

Again, it is key to note that only the federal government has formal power to make immigration policy in the areas of levels of admissions, border control and naturalisation standards. Hence, the state-level laws of these four categories are best thought of as immigrant policy, as they relate to immigrants without affecting several important immigration parameters. Nevertheless, the laws of the highest category come as close as state law can to manipulating numbers of immigrants in a state. The aforementioned Colorado law is an example of a 287(g) agreement, which has been available as an option to states since the adoption of the Illegal Immigration Reform and Immigrant Responsibility Act of 1996. Yet, states also may go out of their way to make legal entry as easy as possible: this is illustrated with a 2008 New York law (AB 2019) which establishes that, provided a few criteria are met, adoption judgements from foreign countries shall hold the same weight as a New York court ruling. Hence, several states do all they are permitted in order to shape the size of their foreign-born populations.

An overall distribution of state immigrant policies, by strength and tone, appears in Table 1. There is a greater number of welcoming laws than hostile, but welcoming laws are primarily symbolic or have a small impact while stronger laws tend to be more hostile. As can be seen, the most common laws are symbolic. These laws are principally resolutions that affirm some principle or honour an immigrant-related cause, so they are easy to pass and fairly uncontroversial. However, nearly as many laws fall into the small-scale category. All of these laws either relate to a small immigrant group or a feature of life that is relatively unimportant. Hence, the stakes are relatively low among small-scale laws, so they are easier to pass and less likely to serve as campaign material in the next election. The top two categories are laws that have broad-reaching effects, though, so they are more likely to cause prolonged debates and are more prone to emerge as issues in elections.

Table 1 Strength and tone of state immigration laws 2005–2011

Source: Author coding of NCSL records.

Cell entries are percentage of total laws. Raw counts in parentheses.

Outcome variable

The primary outcome that is modelled in subsequent analysis is based on the direction and strength of adopted laws in each state. Specifically, the measure is a ratio of welcoming to hostile laws from all seven years, with each law weighted by its rating on the four-point scope scale. The formula is presented in Equation 1:

$${\rm{immigrant\ policy}} = {\rm{log}}\left( {\frac{{\mathop{\sum}{{\rm{welcoming}}\ {\rm{laws}}\:\times \:{\rm{scope}}\: + \:1} }}{{\mathop{\sum}{{\rm{hostile\ laws}}\:\times \:{\rm{scope}}\: + \:1} }}} \right)$$

The logged ratio has the benefit of showing where the preponderance of laws are. Positive values represent welcomeness to immigrants; negative values represent overall hostility to immigrants; and when the laws are perfectly balanced, the ratio is 1 and its log is 0.Footnote 8

Figure 1 shows a map of this variable in the 50 states. In the map, darker-shaded states are more welcoming toward immigrants, taking positive values of the measure, while lighter-shaded states are more hostile, taking negative values. The map shows a few interesting features. Surprisingly, the four states that share a border with Mexico display great diversity in the tone of immigrant policy, suggesting this is not a factor in itself. Despite this diversity, though, the map does show substantial clustering of states. Though there is not a clear north-south or east-west trend in policy, many states do have a similar policy tone to their neighbours. This serves to affirm the previously mentioned idea that neighbour-based geographic correlation should be considered in the model.

Figure 1 Map of immigrant policy tone 2005–2011

Explanatory model

To model the overall tone of state immigrant policy, I use a Bayesian conditionally autoregressive model for spatial data, which offers two major advantages. Firstly, this data set includes all 50 states, so this study analyses a complete population rather than a random sample. Bayesian statistics allow the researcher to account for uncertainty in population analysis more easily, in contrast to classical methods, which are designed to draw population inferences from a sample. Hence, this is a more natural approach for state-level analysis (Gill Reference Gill2001).

Secondly, in state politics generally, geography ought to matter because neighbouring states are more likely to share unmeasured traits than distant states. These traits can be expected to influence outcomes such as policy. Many models of state politics assume that the data are a sample of independent observations, but conditionally autoregressive (CAR) models include a component in the error term that accounts for similarity to neighbours’ outcomes.Footnote 9 Additionally, as was mentioned, many of the lurking unobserved variables that are particularly likely to shape immigrant policy are going to be similar among neighbouring states. To ignore the spatial correlation that is likely to be present in the error terms would be similar to ignoring serial correlation in time series data. The model would underestimate the level of uncertainty in the results, potentially leading to false inferences. Hence, a Bayesian CAR model makes more accurate assumptions about the data and therefore should lead to more accurate inferences.

For the structural specification of the model, the input variables are as follows: legislative professionalism is captured by Squire's (2012) index, which is a composite of legislator salary and benefits, time demands of service, and staff and resources that a state legislature has relative to Congress.Footnote 10 The square root of professionalism is included based on model diagnostics of functional form. Term limits is simply an indicator variable coded 1 if the state limits the number of terms legislators may serve, and 0 otherwise. Public sentiment is captured through citizen ideology, a pervasive measure of which is to aggregate survey respondents’ self-reported ideology by state. Erikson et al. (Reference Erikson, Wright and McIver1993) created this measure using data from CBS/NYT polls, which they have updated through 2003. Since these data substantially precede the time frame under consideration, this paper measures self-reported ideology in the same way, but using the 2006, 2007, 2008 and 2010 Cooperative Congressional Election Study (CCES).Footnote 11 To control for partisanship of elected officials, variables for unified Republican and unified Democratic control are included in the model, each of which is a count of the number of years a party had complete control of the policy-making institutions (Klarner Reference Klarner2003). The change in the foreign-born population is measured with the percentage increase in the number of foreign-born residents from 2000 to 2008. The per capita gross state product was measured for each state in 2005 to 2010 and averaged across all years.

To formalise the described model, the Bayesian CAR model is specified as follows:

$$ \matrix{ {{\rm Y}_i}\: = \:{{{\bf x}_{i}^{\prime}}}\brbeta \: + \:{{\theta }_i}\: + \:{{\phi }_i} \cr {{\beta }_j} \sim {\tf="Times-i"N}\left( {0,100} \right) \cr {{\theta }_i} \sim {\tf="Times-i"N}\left( {0,\sigma _{h}^{2} } \right) \cr {{\phi }_i} \sim CAR\left( {1/{{\sigma }_c}} \right) \cr 1/{{\sigma }_h} \sim {\rm{G}}\left( {{\rm{0}}{\rm{.001,}}\,{\rm{0}}{\rm{.001}}} \right) \cr 1/{{\sigma }_c} \sim {\rm{G}}\left( {{\rm{0}}{\rm{.1,}}\,{\rm{0}}{\rm{.1}}} \right) $$

Here Yi is the measure of immigrant policy presented in Equation 1 for each state, xi is each state's vector of covariates (ideology, unified Democratic control, unified Republican control, per capita gross state product, term limits, change in foreign-born, square root of legislative professionalism, and a constant), β is the vector of coefficients, θi is an independent disturbance, φi is the spatially autoregressive disturbance, βj for j ∈ {0,…,7} refers to each specific regression coefficient, $$)(-->$<> \sigma _{h}^{2} <$> <!--$$ is the variance of the stochastic disturbances of policy θi, and $$)(-->$<> \sigma _{c}^{2} <$> <!--$$ is the variance of the conditionally autoregressive (CAR) disturbances φi.Footnote 12

Results

The results from the model support the primary expectations: legislative professionalism, citizen ideology, state wealth and change in foreign-born population all have robust effects on state immigrant policy.Footnote 13Table 2 presents the results of this model as summaries of the posterior distributions of all parameters.

Table 2 Model of net immigrant policy tone (posterior summaries)

N = 50, DIC = 84.617. Estimates computed with WinBUGS 1.4.3 (Lunn et al. Reference Lunn, Thomas, Best and Spiegelhalter2000). Results based on a post-burn-in MCMC sample of 270,000. The burn-in period was 10,000 iterations each on three chains. σc represents the standard deviation of clustering effects. σh represents the standard deviation of heterogeneous error.

There is a robust effect of legislative professionalism on policy. More professional legislatures are more likely to enact welcoming laws, which fits with the story that professional legislatures attract members with a longer career horizon. Based on the posterior mean in Table 2, if a state increased from the mean level of legislative professionalism by one percentage point, such an increase would result in a 1.72 per cent expected increase in the ratio of welcoming to unwelcoming weighted laws, all else being equal.Footnote 14 Whether a state has term limits does not have as robust an effect on immigrant policy as legislative professionalism does; however, the posterior summary for the coefficient on term limits indicates that there is a 0.8839 probability that term-limited states adopt more hostile policies than non-term limited states. On balance, then, the evidence is suggestive that legislators’ career timeline is influential to immigrant policy.

There is also a strong and robust effect of citizen ideology on immigrant policy. States with liberal electorates are more likely to adopt a welcoming tone toward immigrants. The mean of the posterior presented in Table 2 implies that a percentage point increase on this citizen ideology scale leads to a 4.1 per cent expected increase in the ratio of welcoming to unwelcoming laws, holding all else constant.Footnote 15 It appears, then, that policy-makers are responsive to symbolic ideology in the state.

Immigrant policy also is responsive to state wealth and demographics. A $1,000 increase in per capita gross state product results in an expected increase of 2.7 per cent in the ratio, ceteris paribus. Meanwhile, a percentage point increase in the change in foreign-born population leads to an expected decrease of 0.3 per cent in the welcomingness of laws, ceteris paribus. As was suggested, then, in wealthier states it appears that legislators can more easily enact welcoming policy. Meanwhile, as the change in the foreign-born population increases, policy becomes more hostile.

Lastly, Table 2 presents the components of the unexplained variance in modelling state immigration policy. At their posterior means, the standard deviation of spatially dependent errors is 0.266 while the standard deviation of heterogeneous errors (or independent errors) is 0.529. These values imply that almost a third of the unexplained variance in immigrant policy is due to clustering, or similarity to the policy activity of a state's neighbours.Footnote 16 This result, firstly, reaffirms that states tend to adopt policies similar to their neighbours. Secondly, it suggests that the previously stated concern over spatial correlation due to unobserved state similarities was well founded, even after immigrant policy was modelled. In terms of inference, ignoring such substantial spatial correlation would have wrongly underestimated the variance in the posteriors, posing threats to valid inference.

Cases

Among the results from the empirical model, there is an absence of a clear effect for institutional partisanship. Yet this non-effect can be explained elegantly by considering the effect of partisanship in the context of public ideology. In particular, public ideology has a strong influence on who office-holders are, so party control of government and the ideology of the state's electorate will typically be consistent. Among the instances in which institutional partisanship and public ideology are incongruent, however, examining a few cases can be useful for illustrating the potential tension and how it is resolved. For example, during this time frame, several states enacted omnibus laws that were designed to comprehensively address immigration. Georgia was first in 2006 with a bill that was loudly protested by immigrant rights groups on account of harsh provisions, such as denying tax-supported benefits to illegal immigrants and penalising employers who hired illegal immigrants (Odem Reference Odem2008, 130–132). Yet, the Georgia law also had welcoming facets, such as increasing the penalties for human trafficking and penalising non-attorneys who claimed to be immigration assistance providers. Oklahoma in 2007 adopted a roundly anti-immigrant omnibus law that was harsher even than Georgia's. It makes sense that conservative states like Georgia and Oklahoma might take the lead in advancing such thorough anti-immigrant policy, and that Georgia's law was adopted by a unified Republican government. However, it was a Democratic governor, Brad Henry, who signed Oklahoma's especially harsh bill into law.

As another case, Janet Napolitano, as a Democratic governor of Arizona, vetoed several laws that were hostile to immigrants, such as legislation to eliminate in-state tuition to illegal students and a bill to require local police to enforce immigration laws by arresting illegal immigrants. However, she also signed one of the toughest employment laws, which threatens to revoke the business licences of companies hiring illegal workers (Savage Reference Savage2008). On the one hand, many Democrats, including Barack Obama, have favoured employer penalties, which may be a plank offered to the Democratic constituency of low-wage workers. Alternatively, Napolitano may have taken this clearly anti-immigrant stance as a concession to the many conservative voters in Arizona after making pro-immigrant gestures with her prior vetoes.

This explanation seems even clearer in the context of policies adopted in Arizona since the end of Napolitano's term. In 2010, under unified Republican control, Governor Jan Brewer signed the aforementioned Senate Bill 1070 into law, which is similar to one of the hostile bills Napolitano vetoed. While parts of this bill have been overturned by the US Supreme Court, the “show me your papers” provision remains in place. Further, the legislative enactment of the law fits conservative Arizona's pattern of passing hostile immigrant policy with law-makers of both parties. These case studies as a whole offer further evidence that electoral concerns over public ideology outweigh party in state policy-making when the two conflict, which offers a simple explanation for the non-effect of partisanship in the empirical analysis.

More uniquely to immigrant policy, however, is the idea that the office-holder's time horizon can shape laws. This idea is supported by the strong effect that legislative professionalism has on immigrant policy. While it is impossible to determine the precise cause of this state-level result, the finding is consistent with the argument that professional legislatures’ policy is made by politicians who are looking into the political future and trying to curry favour among growing minority groups, such as Latino or Asian Americans. Further support for this theory emerges when considering George W. Bush's tone on immigration when he was governor of Texas. As a Republican governor in a conservative state, it might have been easy for him to push harsh immigration laws to please voters in his party and his state. However, Bush took a friendly approach towards Hispanics generally and the immigration issue specifically. Seeking immigrant support was a key part of the Bush-Rove strategy, which was oriented to long-term Republican dominance (Hamburger and Wallsten Reference Hamburger and Wallsten2006).

Further, this trend continued with his Republican successor in Texas, Rick Perry, who also has run for president. Specifically, Perry chose to snub several members of his own party by publicly opposing immigration policies, such as building a wall along the Mexican border and restricting birthright citizenship (Brooks Reference Brooks2006). Only when he faced a stiff primary challenge from Senator Kay Bailey Hutchison for the 2010 governor's race did Perry's time horizon shorten enough for him to send Texas Rangers to the border with Mexico and talk tough on the issue.

Conclusion

This article offers an understanding of state action towards immigrants, especially in the face of federal inaction. It has focused on the overall tone of policy, be it welcoming or hostile to newcomers. The immigration issue has proven unique in several regards. Firstly, a substantial portion of the unexplained variance in immigrant policy is spatially correlated. In other words, even after modelling immigrant policy as a function of a variety of factors, neighbouring states still have a tendency to be similar to one another. Given that many unmeasured and unmodelled factors – such as labour needs by local employers, residents’ regard for immigrants, and settlement patterns of newcomers – are likely to be similar in more proximate locales, it makes sense that a major share of unexplained variance would spatially correlate. Ignoring this feature would have created artificially small credible intervals and misleading results. Hence, any policy model that is likely to be spatially correlated could benefit from entertaining a spatial conditionally autoregressive model.

Secondly, several results imply that state legislators are mindful of how opponents could garner votes in the next election and consider this when voting on policy. This is apparent in part because public ideology strongly predicts how welcoming policy is, which is to a large degree consistent with the literature not only on immigration policy but general state policy. More intriguingly, and more unique to immigration, state wealth is a key predictor of immigrant policy. This implies that the poorer a state is, the more policy-makers fear being accused of offering select benefits to immigrants at taxpayers’ expense. Further, the larger the growth rate in the foreign-born population, the more hostile policy is. This implies that incumbents are willing to adopt policies that are harmful to a fast-growing group of individuals who cannot vote if it will be more favourably viewed by eligible voters who are concerned by this growing population. Therefore, thoughts on how an electoral opponent may behave lead economic and demographic considerations to distinctly factor into immigrant policy.

Finally, but perhaps most importantly, the empirical results fit with the theory that the longer legislators’ career time horizons are, the more welcoming a state's immigrant policy is. This is most clearly supported by the robust effect of legislative professionalism on immigrant policy. Though less robust, there is also a large probability that the presence of term limits in a state reduces how welcoming immigrant policy is. The findings for professionalism and term limits are important because they suggest that forward-looking legislators behave differently from myopic legislators. This suggests that, on any issue for which public preferences are changing over time, policy-makers may act in a manner consistent with their long-run concerns.

One issue area where a similar pattern would be expected to emerge is gay rights. Younger people, more than older generations, tend to favour gay rights. Hence, generational replacement implies that, in the future, politicians who are tied to anti-gay rights positions may be at an electoral disadvantage. Therefore, more forward-looking politicians ought to be more likely to vote for policies that protect gay rights.

Another issue area for which this dynamic is possible is environmental policy. Firstly, younger people are more pro-environment than older voters. Secondly, the scientific evidence is clear on issues such as global warming and scarcity of potable water. Hence, both generational replacement and the flow of expert information to the mass public suggest that politicians who are tied to anti-environmental positions in the future may be at an electoral disadvantage. Again, policy-makers with a long career horizon ought to be more likely to support environment-friendly laws.

Prior research has shown that the electoral concerns of policy-makers are important for understanding policy outcomes. This research reaffirms this, as public ideology shapes immigrant policy much like it shapes other policy areas. However, in many important policy areas – immigration, gay rights and environmental policy, to name a few – how the electorate behaves is changing in a clear way. Elected office-holders can be expected to adjust their behaviour based on future public preferences if they aspire to higher offices and longer careers. Hence, for these dynamic issues, it is important to consider not only the current electoral context but the future of the electorate.

Acknowledgements

All replication data and code are available at my Dataverse: http://hdl.handle.net/1902.1/16471.

For helpful commentary, I would like to thank George Rabinowitz, Virginia Gray, Tom Carsey, Jeff Gill, Brian Fogarty, Mark Ramirez, Dominik Hangartner, Andrew Whitford, Jamie Carson, Marco Steenbergen, Jim Stimson, Georg Vanberg, the participants in UNC's state politics working group and several anonymous reviewers.

For coding data for reliability purposes, I thank Cameron Morgan.

For providing data, I would like to thank Tom Carsey, Peverill Squire, Justin H. Phillips, Jeffrey R. Lax, Melody Crowder-Meyer, Jeffrey Passel and those cited in the data appendices.

For help in constructing the measure of strength of immigration laws, I also thank Jonathan Blazer, Kevin Johnson, Kris Kobach, Jenny Levy, Jack Martin and Jessica Vaughn.

Previous versions of this paper have been presented at the Annual Conference on State Politics and Policy, May 2009, Chapel Hill, NC, the Annual Summer Meeting of the Society of Political Methodology, July 2009, New Haven, CT, and the Department of Political Science at the University of Missouri – St. Louis.

Appendix 1: Data and analysis notes

Table A.1 offers the descriptive statistics of the variables used in the analysis. Table A.2 lists the multi-provision laws adopted in this time frame and enumerates how many separate provisions were coded. Table A.3 displays the number of laws in each tone and scope category adopted by each state. Table A.4 is an alternative specification of the model from the paper that replaces institutional partisanship with Shor and McCarty's (Reference Shor and McCarty2011) measure of state legislative ideology, averaged from 2000 to 2009. Figure A.1 shows a density plot of the outcome measure defined in Equation 1 from the article, and a Kolmogorov-Smirnov test of the outcome measure did not reject the null hypothesis that policy tone is normally distributed (D = 0.0805, p = 0.9023). The sources of all of these variables are listed as follows:

Table A.1 Descriptive statistics of continuous variables

Table A.2 Omnibus laws enacted by states 2005–2011

Table A.3 Counts of immigration laws by tone and scope 2005–2011

Table A.4 Model of net immigrant policy tone, substituting legislative ideology for partisan control (posterior summaries)

Figure A.1 Density plot of immigrant policy tone 2005–2011

Immigration laws enacted by states: The National Conference of State Legislatures:

2005: www.ncsl.org/programs/immig/IMMIGStateLegisJuly06.htm (Accessed 28 September 2007)

2006: www.ncsl.org/programs/immig/6ImmigEnactedLegis3.htm (Accessed 27 March 2007)

2007: www.ncsl.org/programs/immig/2007immigrationfinal.htm (Accessed 1 October 2008)

2008: www.ncsl.org/programs/immig/2008StateLegislationImmigration.htm (Accessed 1 February 2009)

2009: www.ncsl.org/documents/immig/2009ImmigLaws.pdf (Accessed 22 June 2010)

2010: www.ncsl.org/default.aspx?TabId=21857 (Accessed 30 May 2011)

2011: www.ncsl.org/issues-research/immigration/state-immigration-legislation-report-dec-2011.aspx (Accessed 24 January 2012)

Change in state foreign-born population from 2000 to 2008: US Census Bureau (2007, Table 40).

State population for 2005 to 2010: US Census Bureau, www.census.gov/popest/data/ (accessed 7 May 2012).

GDP by state for 2005 to 2010 (2011 dollars, per capita calculated, then averaged): Bureau of Economic Analysis, US Department of Commerce, www.bea.gov/regional/(accessed 7 May 2012).

Squire's index of legislative professionalism in 2009: Squire (Reference Squire2012).

Term limits for state legislators: National Conference of State Legislatures, www.ncsl.org/legislatures-elections/legisdata/chart-of-term-limits-states.aspx (accessed 8 May 2012).

Party control of state government for 2005 to 2011: An update of Klarner (Reference Klarner2003), www.indstate.edu/polisci/klarnerpolitics.htm (accessed 8 May 2012).

Self-reported citizen ideology in 2006, 2007, 2008 and 2010:

Notes: The electoral ideology measure replicates the measure from Erikson, Wright and McIver (1993), which subtracts the percentage of self-identified conservatives from the percentage of self-identified liberals across both the 2006 and 2008 surveys. I use the survey weights of the CCES in constructing this measure.

Appendix 2: Coding rules for significance of laws

(4) Impacts residence: Laws designed to directly affect the number of foreign-born residents in a state, typically illegal immigrants. This category includes laws that either commission state and local authorities to enforce federal immigration law or specifically snub federal law by refusing to report immigration status to federal authorities. Also, laws that open or close a choke point, such as eligibility for driving licences or employability. Should driving licences be granted regardless of immigration status or should these be restricted? Can a worker or employer be severely punished, via jail or revocation of business licence, if an illegal immigrant is hired? Is the state recruiting outside workers?

(3) Large-scale effect: Laws that create general incentives or disincentives for any immigrant who may enter a state. These include providing or restricting benefits for legal or illegal immigrants, including legislation regarding naturalisation programmes, worker's compensation coverage, retirement, higher education funding or bilingual provisions. This also includes smaller provisions in deportation, employment or licensing laws. Such smaller provisions may include requiring or restricting immigration status verification by employers, making small changes in ease of getting a driving licence, and screening arrested persons for immigration status.

(2) Small-scale effect: Laws that create incentives or disincentives, but which are likely to apply only to a small subgroup of potential immigrants, such as professionals from a specific field, those who may work for a public contractor, asylees or trafficking victims. These laws might speak to job eligibility or benefit eligibility for the people in these small groups, or may penalise non-immigrants whose behaviour on behalf of these groups is outlawed (i.e. employers of illegal immigrants, traffickers or smugglers). Also, laws related to matters less central to immigrants’ lives, such as voting, professional licences, gun licences, property rights and specified immigrant protection (such as regulating matchmaking services or notarios) fit here. Implementing laws also belong here (i.e. delivering federal funds or developing protocols to deliver services).

(1) Symbolic: Symbolic laws that make an issue statement to Congress, request another branch of government to take action, launch a study or task force, or affirm a principle (such as a commitment to cultural heritage, requesting that employers hire legal persons or declaring English as a state's official language). Many of these symbolic measures are joint resolutions.

Appendix 3: WinBUGS replication code

The following code was used to estimate the model reported in Table 2 in the article:

#MODEL

model {

for (i in 1:N) {

immig0511[i]∼dnorm(muY[i], tau.h)

muY[i]<-constant+beta[1]*pubIdeolCCES[i]+

beta[2]*repUnif[i]+beta[3]*demUnif[i]+

beta[4]*pcgsp1000[i]+beta[5]*termLimits[i]+ beta[6]*changeForeign[i]+

beta[7]*sqrtProf[i]+phi[i]

sqrtProf[i]<-sqrt(squireProfess[i])

unused[i]<-ID[i]+multicultural[i]+

demhouse0511[i]+demsen0511[i]+

demgov0511[i]+immig0508[i]+ unemployment[i]+pctForeign[i]+bilingual[i]+ immlicense[i]+tuition[i]+verify[i]+immOpin[i]+

pubIdeolEWM[i]+legIdeol[i]

}

phi[1:N]∼car.normal(adj[], weights[], num[], tau.c)

for (i in 1:7) {beta[i]∼dnorm(0.0,0.001)}

constant∼dnorm (0.0,0.001)

tau.h∼dgamma(1.0E-3,1.0E-3)

tau.c∼dgamma(1.0E-1,1.0E-1)

sd.h <-1/sqrt(tau.h) #marginal SD of heterogeneity effects

sd.c <-sd(phi[]) #marginal SD of clustering (spatial) effects

alpha<-sd.c/(sd.h + sd.c) #relative clustering }

#INITIAL VALUES AND DATA ARE AVAILABLE ON THE AUTHOR'S DATAVERSE

Footnotes

N = 50. †Dichotomous. *In thousands. Estimates computed with R 2.14.0.

N = 50, DIC = 85.511. Estimates computed with WinBUGS 1.4.3. Results based on a post-burn-in MCMC sample of 270,000. The burn-in period was 10,000 iterations each on three chains. σc represents the standard deviation of clustering effects. σh represents the standard deviation of heterogeneous error.

References

1 Intriguingly, Hero and Preuhs (2007) find that government ideology is a strong predictor of TANF laws, but citizen ideology is not. By contrast, the appendix re-estimates the model of Table 2 including both legislative and citizen ideology, but finds that citizen ideology is a strong predictor but legislative ideology is not. Perhaps TANF policy was sufficiently technical that legislators could more easily adopt their own preferences relative to other aspects of immigrant policy.

2 As a robustness check in case of endogeneity bias, I re-estimated the model of Table 2 using two-stage least squares. In this model, I introduced Erikson, Wright and McIver's measure of public ideology from 1995 to 1999 as an instrument for ideology during the time frame of interest. Presumably, the old values of ideology could not have been affected by current policy, but they could affect later values of public ideology. Public opinion ideology continued to show a strong positive effect in this alternative model, with a coefficient of 0.056 and a 90 per cent confidence interval of [0.027,0.084].

3 Source: Table on percentage change in foreign-born by state generated by Terrazas and Batalova of the MPI Data Hub, http://migrationinformation.org/DataHub/, accessed 22 June 2010.

4 Of new legal admissions of immigrants in 2008, 43.1 per cent came from Latin America and 34.6 per cent came from Asia (United States Department of Homeland Security 2009, Table 3).

5 See the appendix on data and analysis for a more complete source reference. Also, the example law synopses of the next paragraph are all based on the cited reports by NCSL. I focused on laws that were enacted by each state between the years 2005 and 2011.

6 As a check on the coding scheme, a second coder independently coded all of the 2008 laws (36.7 per cent of the total) for comparison with the author's coding. For the binary tone of the law, Krippendorf's α = 0.789, and for the ordinal scale of scope, α = 0.723. Both exceed 0.7, indicating acceptable inter-coder reliability between the author and the second coder.

7 The appendices describe more thoroughly the coding rules for placing each law into a category and shows the number of laws every state adopted in each of the scope and tone categories. Omnibus laws were broken up such that each provision was coded as a separate law. The appendix includes a table listing the laws that were split up.

8 The appendices list descriptive statistics of this and all other measures in Table A.1, scores for each state with their components in Table A.3, and a density plot of the measure in Figure A.1. All estimates were computed in R 2.14.0 (R Development Core Team 2009). Regarding Equation 1, the addition of one extra welcoming and one extra hostile law prevents any undefined ratios or logs. The main findings of the analysis remain intact even if this measure is constructed with raw counts of laws (rather than weighted counts), using percentages (rather than ratios), and by adding 2 or 3 to the numerator and denominator (rather than 1). The data were pooled across all years to prevent unrepresentative scores due to micronumerosity.

9 For more detail, Banerjee et al. (2004, Chapter 3) describe methods for areal data, or data defined by geographic boundaries, that account for spatial correlation. This discussion includes a technical description of the CAR model.

10 More specifically, Squire captures salary and benefits through the base legislative salary reported by each state, time demands through the number of days a legislature meets per year, and staff and resources through legislative staff figures gathered by the National Conference of State Legislatures (specifically total staff during the session, including permanent and session-only staff). A complete description of the measure is available in Squire (2007).

11 For more details on the data sources for ideology and the other input variables, see the appendix. The Pearson correlation between this CCES ideology measure and a 1995–1999 aggregation of Erikson 1993's data is 0.832. As an alternative specification of the model in Table 2, public ideology was measured with a factor analysis of self-reported ideology, Berry et al. (1998) measure of citizen ideology, and presidential voting in 2004. The results were similar to those reported.

12 The hyperpriors on the precision of the two random effects terms ($$)(-->$<> 1/{{\sigma }_h} <$> <!--$$ and $$)(-->$<> 1/{{\sigma }_c} <$> <!--$$) are based on the recommendation of Best et al. (1999), who studied the effect of changing the distributional form of the smoothing prior. The goal is to assume beforehand that the unexplained variance is split evenly between the clustering term and the heterogeneous term, thereby preventing the hyperpriors from shaping how the variance is distributed (Carlin and Pérez 2000; Banerjee et al. 2004). I also considered the recommendation of Bernardinelli et al. (1995) and recovered a similar posterior distribution for the two variance terms. Also, for modelling purposes, Alaska is treated as a neighbour of Washington, and Hawaii is treated as a neighbour of California. All other neighbours are defined by whether the two states share a border.

13 Specifically, the probability that these variables have an effect in the expected direction exceeds 0.95. Also, the Gelman-Rubin (Gelman and Rubin 1992) statistic converged to 1 for all coefficients, indicating the absence of evidence of nonconvergence.

14 All results are interpreted as if the mean value of a parameter holds. Interpreting effects in terms of percentage increases is possible because the dependent variable is logged, so taking the exponential of an unstandardised coefficient provides a multiplicative factor by which the ratio of welcoming to unwelcoming laws increases. For this particular coefficient, Squire's professionalism measure is based on proportions of legislative capacities compared to the US Congress, so an increase of 0.01 in this variable could be interpreted as a percentage point increase in professionalism. Since the square root of the measure is taken, the marginal effect of the variable changes based on the level of professionalism. Moving from the mean of professionalism (0.19) to a percentage point higher (0.20) amounts to an increase of 0.011324 in the square root measure. exp(1.506 × 0.011324) = 1.0172.

15 As an alternative specification, I also estimated this model with an additional control for public opinion on immigration issues. To do this, I incorporated Lax and Phillips's (2012) measures of state public opinion on the issues of issuance of driving licences to illegal immigrants, prohibition of bilingual education, providing in-state tuition for children of illegal immigrants, and requiring the state government to verify citizenship status before making hiring decisions. These measures of public sentiment were constructed using the novel method described in Lax and Phillips (2009). I combined these four variables into one measure of immigration opinion using principal components analysis and included this measure in a model otherwise identical to the one reported in Table 2. In this alternative model, symbolic ideology maintained its robust and positive effect with a posterior mean of 0.037 and a 90 per cent credible interval of [0.013,0.061]. The posterior coefficient for issue-specific public opinion had a positive mean of 0.071, but the effect was not very robust as the 80 per cent credible interval was [−0.105,0.247]. This implies that legislators are more concerned about symbolic ideology than issue-specific immigration opinion.

16 More formally, Best et al. (1999) define $$)(-->$<>\psi \: = \:\frac{{sd\left( \phi \right)}}{{sd\left( \theta \right)\: + \:sd\left( \phi \right)}} <$> <!--$$ as a measure of the share of random effects variability due to clustering. Table 2 shows that, for this model, the posterior mean of ψ is 0.332.

Figure 0

Table 1 Strength and tone of state immigration laws 2005–2011

Figure 1

Figure 1 Map of immigrant policy tone 2005–2011

Figure 2

Table 2 Model of net immigrant policy tone (posterior summaries)

Figure 3

Table A.1 Descriptive statistics of continuous variables

Figure 4

Table A.2 Omnibus laws enacted by states 2005–2011

Figure 5

Table A.3 Counts of immigration laws by tone and scope 2005–2011

Figure 6

Table A.4 Model of net immigrant policy tone, substituting legislative ideology for partisan control (posterior summaries)

Figure 7

Figure A.1 Density plot of immigrant policy tone 2005–2011