Authors: Maureen Stobb, Banks Miller, and Joshua Kennedy
Publication: Policy Studies Journal 51: 667-684
Original source: psj_final.pdf
Conversion note: this Markdown was generated by text extraction from the hosted file. Review equations, tables, notes, and citations before treating it as a final web copy.
1Department of Political Science and International Studies, Georgia Southern University, Statesboro, Georgia, USA 2School of Economic, Political and Policy Sciences, University of Texas at Dallas, Richardson, Texas, USA Correspondence Maureen Stobb, Georgia Southern University, Statesboro, GA, USA. Email: mstobb@georgiasouthern.edu
| Policies concerning undocumented immigrants are inevitably ambivalent, creating uncertainty and confusion in the implementation process. We identify a clear example of this ambivalence —U.S. law setting standards for determining the credibility of asylum seekers—that resulted in an increase in asylum grants despite policymakers’ intention to make it harder for individuals to obtain the status. We argue that this law, The REAL ID Act of 2005, sent mixed messages to immigration judges (IJs), street-level bureaucrats who implement immigration policy. It increased IJ discretion, but set vague limits. We theorize that IJs, behaving in a bounded rationality framework, use their professional legal training as a short-cut and look primarily to the courts for guidance. Our evidence supports our argument. After the passage of the REAL ID Act, IJ decision-making is more closely aligned with the preferences of their political and legal principals, and, in the final score, the federal circuit courts are the winners. KEYWORDS asylum, frontline implementation, immigrants, policy ambiguity, REAL ID Act ORIGINAL ARTICLE Mixed messages & bounded rationality: The perverse consequences of real ID for immigration policy Maureen Stobb1 | Banks Miller2 | Joshua Kennedy1 DOI: 10.1111/psj.12486 Received: 31 May 2022 Revised: 22 July 2022 Accepted: 22 August 2022 Policy Stud J. 2022;00:1–18. © 2022 Policy Studies Organization.wileyonlinelibrary.com/journal/psj 1 |
The REAL ID Act is characterized by many as restrictive immigration policy, explicitly passed in response to the 9/11 terrorist attacks and a wave of public opinion hostile to foreigners. One express goal of the law was to tighten refugee admissions by giving immigration judges more discretion in denying bogus asylum claims. In short, immigration judges were no longer required to point to an inconsistency undermining a key aspect of the persecution claim in order to find that the applicant was not credible, a clear rejection of the “heart of the claim” rule embraced by the majority of federal judiciary and the administrative courts. At the same time, Congress was not clear about the criteria that would replace the prevailing credibility standard. Immigration judges were required to consider the “totality of the circumstances,” language positioned by some legislators as reincorporating central components of the heart of the claim rule. 1 This new, vague, and conflicting credibility standard became law on May 11, 2005.
STOBB eT al .2 The passage of the Real ID Act was expected by many of its proponents and a number of scholars to make asylum claims more difficult to win or at least to not meaningfully affect the rates of asylum grants (e.g., Anker, 2016 ; Cianciarulo, 2006 ; Fletcher, 2006; Galloni, 2008 ; Kanstroom, 2006 /07; Keith et al., 2015; White, 2006). It was not expected at all to be followed by an increase in grants, and yet this is precisely what happened (e.g., Keith et al., 2015). Indeed, our own data show that before REAL ID was enacted the average asylum grant rate was 29%, but after REAL ID it is 40%. We investigate this gap, arguing that the REAL ID Act placed considerable discretion in the hands of IJs (e.g., Kanstroom, 2006). In turn, IJs, behaving in a bounded rationality framework (e.g., Simon, 1957 ), rely on their professional training as a shortcut for decision-making: being trained lawyers, IJs are considerably more deferential to their superiors in the circuit courts. Thus, although the explicit purpose of the Act was to take power from the U.S. Courts of Appeals, in the final score, the federal circuit courts continued to have the most influence on whether an applicant is found credible and, ultimately, if they receive asylum. We test our theory with a quantitative analysis of over 900,000 IJ decisions made before and after the REAL ID was passed, providing evidence of how this change occurs through the legal standards set by each principal. Our findings support our argument. The REAL ID Act altered the decision-making environment for Immigration Judges, allowing for tighter control of IJ decision making by those above them, primarily those in the legal hierarchy. This study makes an important contribution to the immigration policy gap debate. Numerous scholars across disciplines have asked why immigration policy outcomes tend not to match the intent of the public or the government actors supporting policy adoption, and have called for in-depth examinations of the role of the judiciary in particular (e.g., Bonjour, 2011 ; Cornelius & Tsuda, 2004 ; Freeman, 1992, 1994). In addition, immigration policy studies scholars note that removal as a means of immigration control has been largely ignored by the literature (Ellerman, 2007; Giuraudon & Lahav, 2007). We contribute to this conversation by developing a theory concerning the specific mechanisms by which significant and persistent gaps between immigration policies and outcomes develop, focusing on the response of street-level bureaucrats to vague and ambivalent statutes in removal proceedings. This research provides valuable insights for policy studies scholars, as immigration policy studies is an emergent area of public policy, “both typical and underutilized,” that can help scholars tackle vital questions (Filindra & Goodman, 2019, p. 500). Specifically, our findings speak to the growing literature building on the work of scholars in several policy domains who have demonstrated that understanding policymaking requires attention to local implementation processes in addition to policy formulation (Giuraudon & Lahav, 2007; Lipsky, 1980; Pressman & Wildavsky, 1984; Van der Leun, 2007). We proceed as follows. First, we describe the adoption of the new credibility standard by Congress, the executive and the judiciary over time. Next, we detail our theory concerning the mechanism by which the gap in implementation occurred. We then explain our methodology for testing our expectations, and present our results. We end by discussing the implications of our findings and suggesting further avenues of research. ALTERING THE CREDIBILITY LANDSCAPE Congress altered credibility standards through the passage of the REAL ID Act, which supporters labeled anti-terrorism legislation (Cianciarulo, 2006). Its sponsor, House Judiciary Committee Chairman James Sensenbrenner (R-Wis.), argued that the changes would give immigration judges greater discretion to deny “bogus asylum applications by aliens who are clearly lying.” 2 According to Sensenbrenner, “liberal, activist judges” on the U.S. Courts of Appeals overturned “clearly established precedent” to make “asylum laws vulnerable to fraud and abuse.”2 Numerous Democrats in the House and Senate criticized the bill as creating unnecessary hurdles for legitimate asylum seekers (Miller Keith & Holmes, 2015). The new law applied to all cases in which asylum applications were filed after its effective date, May 11, 2005. 3 We contend that the REAL ID Act imposed two competing legal standards for IJs to employ when deciding asylum cases—the rejection of the “heart of the claim rule” and the promotion of the “totality
MIXeD MeSSa GeS & BOUNDeD Ra TIONalITY 3 of the circumstances standard.” As a result, the Act appears to both encourage IJs to base their adverse credibility findings on minor inconsistencies, while implying that doing so would be unreasonable. We argue that some circuits emphasized the first standard, rendering more conservative outcomes, while others adopted the second rendering more liberal outcomes. We detail our argument further below. The statute explicitly rejected the heart of the claim rule, the approach of the majority of the circuits. Under this rule, an immigration judge must find a connection between any inconsistencies and the “heart of the claim” to justify a determining that the applicant was not credible (Anker, 2016). Examples of inconsistencies going to the heart of an applicant’s asylum claim include details about the claimed persecution, such as the date and the number of attacks. The REAL ID Act allows immigration judges to base an adverse credibility determination on any inconsistency, without regard as to whether it went to the heart of the applicant’s claim of persecution. For example, an IJ could find an applicant lacks credibility because of such minor inconsistencies in his or her testimony as birthdays and wedding dates that are not connected to the claimed basis of the persecution, This policy change increases immigration judges’ discretion in determining applicants’ credibility determinations by making it harder for court of appeals’ to reverse them. However, the Act also calls on IJs to consider a second approach requiring that they look at the “totality of circumstances” when examining any potential inconsistencies. Cianciarulo (2006, p. 135) dismisses the fear of some that the new standard would allow for “immaterial discrepancies” to sink asylum applications by noting that Congress clearly specified that these discrepancies may not possess “controlling weight” on the decision. Under the law, immigration judges may consider a list of factors and any other relevant considerations, without clarifying the weight given to each. REAL ID does not include an explicit reasonableness requirement. 4 In ideological terms, the heart of the claim rule is best categorized as a liberal standard because, by narrowing the permissible bases for an adverse credibility determination, it increases the probability that an immigration judge will find an asylum applicant credible and therefore grant asylum (Stobb, 2019 ). For issues pertaining to criminal procedure, civil rights, First Amendment, due process, and privacy, a liberal outcome is pro-underdog, pro–civil liberties or civil rights claimant, especially those exercising less protected civil rights (Segal & Spaeth, 1993 , p. 243). An asylum applicant can be characterized as an underdog, likened to a civil rights claimant, because he claims his government violated his fundamental rights. The REAL ID standard is more conservative than the heart of the claim rule because it explicitly rejected the requirement of a nexus between the inconsistency and the persecution claim, making it easier to reject a claim on the basis of a minor inconsistency and making a conservative outcome (the denial of an asylum claim) more likely. Any attempt by a court of appeal to take advantage of the totality of the circumstances language in the statute to craft a standard maintaining key elements of the heart of the claim rule by another name, as discussed below, would also be a liberal standard, in the sense that it is more liberal than the REAL ID Act. The Board of Immigration Appeals (BIA)—an executive agency that is part of the Department of Justice—responded to Congress’ directive with a precedential opinion in 2007 emphasizing language rejecting the heart of the claim rule and affirming an adverse IJ credibility determination not related to the heart of the claim. 5 The BIA also asserted that the REAL ID Act directed Immigration Judges to follow a “commonsense” approach in considering the circumstances of the applicant and witnesses ( Matter of J-Y-C-, 2007, 262). Ultimately, the BIA sent mixed messages in interpreting the REAL ID Act. On the one hand, it discarded the heart of the claim rule, but on the other it cited “totality of the circumstances” and “reasonableness” language evoking principals underlying the heart of the claim rule. The circuit courts of appeal have also recognized Congress’ clear directive to reject the heart of the claim standard, concluding it intended to give IJs more discretion in determining credibility. 6 At the same time, they have noted the inclusion of the “totality of the circumstances language,” varying in the emphasis placed upon the latter, and considered the extent to which Congress intended to incorporate a reasonableness requirement despite the absence of the word from the language of the bill. The First Circuit noted the totality of the circumstances language but refused to allow legislative history to trump the statute’s text, suggesting future panels should consider the extent to which a reasonableness requirement
STOBB eT al .4 limits IJ discretion. 7 The Second and Fifth Circuit went a little further, concluding that they must defer to an IJs adverse credibility determination unless it is unreasonable. 8 Similarly, the Third Circuit recently cautioned that the immigration judge must consider context, and not “cherry pick” inconsistencies. 9 Two circuits seem to have sidestepped the REAL ID’s clear rejection of the heart of the claim rule. The Seventh and the Ninth argue that immigration judges cannot rely on trivial inconsistencies in making an adverse credibility determination, and a rule of reasonableness must be followed. 10 The remaining circuits have avoided adopting a binding post-REAL ID credibility standard by not addressing the Act in a precedential decision. 11 Table 1 below summarizes the timing and ideological directionality of these changes across the circuits. It shows that 6 circuits changed standards as suggested by REAL ID, while six circuits made no changes. Also note, the 2nd Circuit switched back to a more liberal standard in 2018. Thus, the REAL ID Act presented IJs with conflicting or at least confusing messages—in Congress’ language, and the interpretations of the executive and the courts. We detail our theory explaining the effect of these actors’ adoption of the new REAL ID credibility standards on street-level implementation of asylum policy in the next section. STREET-LEVEL IMPLEMENTATION To begin, a number of scholars studying immigration policy have spoken of the “policy gap perspective” (Miller et al., 2015 , p. 150), or the gap between what the law was intended to accomplish and the actual outcome (see also Bonjour, 2011; Boswell, 2007; Cornelius & Tsuda, 2004; Freeman, 1992, 1994; Johannesson, 2018). The intent has been described as that of the governmental actors drafting the policy and proclaiming its objectives, and that of the public, as reflected in “the generally protectionist bent of public opinion in democratic states” (Boswell, 2007, p. 75; Johannesson, 2018, p. 1163). Such a perspective shares much in common with principal-agent theory, which essentially posits that, because of informational asymmetries and monitoring difficulties, agents may not always be faithful to the principal’s wishes (e.g., Miller, 1992 , 2005; Spence & Zeckhauser, 1971 ). Because the principal cannot see all that the agent is doing, and because the agent has access to information that the principal does not, it is theoretically possible (and expected in some circumstances) that agents will act directly in their own interests rather than those of the principal. This “gap” between what the principal wants and what they receive from the agent can be referred to as agency loss (e.g., Miller, 2005). Circuit Year Change 1st 2008 Liberal → Conservative 2nd 2008 Liberal → Conservative 2nd 2018 Conservative → Liberal 3rd 2021 Liberal → Conservative 4th – None 5th – None 6th 2009 Liberal → Conservative 7th – None 8th 2014 Liberal → Conservative 9th – None 10th – None 11th 2006 Liberal → Conservative D.C – None
MIXeD MeSSa GeS & BOUNDeD Ra TIONalITY 5 Framed in this manner, we might think of the policy gap in immigration policy as another example of agency loss. These losses can result either because the policy has unintended consequences, which, previous scholarship has argued, was the case with the Real ID Act (e.g., Miller et al., 2015), or because agents do not adequately enforce or implement the policy (e.g., Cornelius & Tsuda, 2004 ). This can occur for a variety of reasons. Congress, for example, is likelier to grant wider discretion to the executive branch in cases where the branches theoretically share policy goals (e.g., Epstein & O’Halloran, 1999; Huber & Shipan, 2002 ; Huber et al., 2001). Such a scenario existed when the Real ID Act was initially passed: Republicans controlled both Houses of Congress and the presidency, and they largely intended the act to lessen the number of undocumented immigrants in the United States, and limit the possibility of potential terrorists gaming the asylum system to gain entry (e.g., Miller et al., 2015). Congress may have felt comfortable in such a scenario granting more discretion to the executive branch, and the Real ID Act as written certainly afforded a considerable amount of decision-making authority to IJs (e.g., Kanstroom, 2006). The difficulty with granting discretion is straightforward: it affords the agent considerably more room to act against the principal’s wishes. But IJs serve more than one master; it is somewhat difficult to speak of Congress’s intent at all given that Congress is not an “it” but a collection of individuals with a variety of preferences. What one principal, say, a committee chair, desires may not be what another principal, like the leader of the majority party, prefers. It is therefore difficult to speak of legislative intent, because whose intent we are speaking of can vary a great deal (see, e.g., Shepsle, 1992 ). There are also the preferences of the president (and the president’s immediate subordinates with authority over IJs, such as the Attorney General) and the courts (particularly the circuit courts, which have jurisdiction to hear appeals of IJ decisions) to consider. Because a multiplicity of principals theoretically induces a loss of control (e.g., Clinton et al., 2014; Gailmard, 2009), the possibility of a policy gap is significant; the agents, after all, may not be clear on what they are supposed to do precisely because their principals are not speaking with a unified voice. But even in the rare circumstance that respective principals do speak clearly, there is no guarantee that agency loss or a policy gap will not result. The difficulties with monitoring and informational asymmetries remain. To get a fuller picture of what is occurring, it is important to understand things from the bureaucrat’s perspective. The question of what drives bureaucratic behavior is well-trodden ground (e.g., Carpenter, 2001; Golden, 2000; Huber, 2007; Niskanen, 1975; Simon, 1957, 1978; Wilson, 1989/2000). While Niskanen ( 1975) famously argued that bureaucrats are largely motivated by budget maximization, other work contends that what drives bureaucrats is the desire for autonomy, or the ability to do their work without undue interference from the “political” branches (e.g., Carpenter, 2001; Huber, 2007; Potter, 2019). Further, scholarship demonstrates that bureaucrats are motivated by a sincere desire to do good work, regardless of their own feelings about the priorities set by their principals in Congress, the White House and, to a lesser extent, the courts (e.g., Golden, 2000 ; Wilson, 1989/2000). This is not to say that their personal ideologies are never relevant, as recent scholarship has shown how bureaucrats can strategically manipulate the regulatory process to achieve outcomes they prefer and avoid stringent oversight (e.g., Potter, 2019). Of course, IJs are not “regulators” per se, but rather adjudicators, and while it is not clear that this is the case in all adjudicative agencies, previous scholarship has shown that the preferences of IJs matter as it pertains to outcomes in asylum cases (e.g., Miller et al., 2015). In general, we expect IJs to operate under a bounded rationality framework of the kind elucidated by Herbert Simon (1957). In brief, and in contrast to standard rational choice theory, bounded rationality recognizes bureaucrats as “satisficers” who, rather than considering all available alternatives and maximizing their utility, behave with the intent to be rational but choose from a more limited and conveniently available set of alternatives (see also Jones [1999 ] for more discussion of bounded rationality). Simon does not speak of any particular kind of bureaucrat in developing this framework, and though we believe there are reasons to expect that IJs, being adjudicative rather than regulatory in nature, are unique in some aspects of their behavior, we do not expect that they are any less likely to satisfice. What, for an IJ, does satisficing look like? We expect that the preferences of their multiple principals will have an effect, but in general, a boundedly rational IJ will be considerably more deferential to the pref-
STOBB eT al .6 erences of the circuit in which they sit. Part of the reason for this is due to the professional background of IJs; as Wilson (1989/2000) notably argues, the norms of the profession in which a bureaucrat finds him- or herself (assuming that the bureaucrat in question is a professional) have considerable predictive power regarding bureaucratic behavior. IJs are attorneys by profession and are acting in a functionally judicial capacity. The reputational costs of being reversed by a court may well be more intimidating to an IJ than the potential of incurring the wrath of Congress or the president. Besides, the likelihood of a single ruling on the part of an IJ attracting even the attention of Congress or the president is remote, whereas appeal of a decision to a circuit guarantees such attention. As is the case with many other policies, decisions in asylum adjudication are complex and, we suspect, a function of a number of factors: the facts of the case, the language of the statute, the preferences of the IJs, and the preferences of the IJs’ respective principals (Congress, the president, and the courts). The size of the policy gap associated with the Real ID Act is considerable enough that one must wonder what accounts for this. Shifts in party control of Congress, the transfer of authority from one president to another, and changing court composition are likely to explain at least part of this gap. James Sensenbrenner, considered the Real ID Act’s principal author, is no longer in Congress. The president who signed the law has been out of office for a decade and a half. Presidents since have had their share of opportunities to populate the courts with judges more in line with the president’s ideological preferences. But the language of the Real ID Act plays a role as well. As noted above, the rejection of the heart of the claim rule was paired with the inclusion of “totality of the circumstances” language that at least one senator characterized as adding a reasonableness requirement akin to the discarded rule. In the face of such uncertainty, we argue that IJs, as boundedly rational agents, consider the alternative interpretations of Congress, the executive and the courts, and defer most to the latter. We would therefore expect that, as the dominant ideology of each branch shifts, so too will IJ grant rates. Because the branch with the greatest influence, the courts, is fragmented and varies widely by circuit in dominant ideology, implementation of the new credibility standard opened the door to an increase in grant rates, perhaps in contrast to what Sensenbrenner and many of the law’s supporters in Congress and the public anticipated. Our theory builds on prior research recognizing the key role of courts in the immigration policy gap, curbing government capacity to control immigration primarily through their defense of migrant rights (Cornelius et al., 1994; Guiraudon & Lahav, 2000; Hollifield, 2000; Joppke, 1998, 1999). At the same time, we follow recent work suggesting their role should be further delineated because they may not always be a force for liberalization of immigration policy, and their impact may be indirect (Bonjour, 2016; Johannesson, 2018; Mascia, 2020). Our argument is distinct from that of prior qualitative work suggesting that the REAL ID Act decreased IJ discretion in evidentiary matters, shifting cases to the courts of appeals (Hamlin, 2014). We argue that the REAL ID boosted the power of all three principals by increasing IJ discretion in determining applicant credibility, and that the judiciary is not the sole winner in the power game. In sum, the adoption of the REAL ID credibility standard should be read as presenting IJs with several conflicting cues. The explicit rejection of the heart of the claim rule, a liberal standard, broadened their discretion, making it harder for the BIA and circuits to remand a denial. This new approach was more conservative than the heart of the claim rule. At the same time, the Board and circuit judges could focus on the totality of the circumstances requirement, subsuming the minor inconsistencies standard within it, and read it as placing limits on this increased discretion. Indeed, as noted above, circuits have cited this language to craft a standard rejecting trivial inconsistencies, essentially holding onto the materiality requirement of the liberal heart of the claim rule, thereby maintaining a liberal standard. To interpret these mixed messages, IJs will look to their three principals for guidance, in addition to relying on their personal policy proclivities. To test this theory, we develop several hypotheses concerning the impact of the three principal’s preferences with regard to asylum outcomes on the likelihood of an asylum grant before and after the REAL ID Act changed credibility standards. Hypothesis 1 After the adoption of the credibility standard in a circuit, as the percentage of a circuit that is Democratic increases so too will the asylum grant rate of IJs.
MIXeD MeSSa GeS & BOUNDeD Ra TIONalITY 7 Hypothesis 2 After the adoption of the credibility standard by Congress, control by Democrats of Congress will increase the asylum grant rate of IJs. Hypothesis 3 After the adoption of the credibility standard by the executive, control of the White House by Democrats will increase the asylum grant rate of IJs. DATA AND METHODOLOGY Data & variables To examine the effects of Real ID, we utilize data from EOIR on the decisions of IJs to grant asylum that spans from 1995 through 2021. In total we have 941,375 observations. We detail the approach used for cleaning the administrative dataset in the appendix. Real ID went into effect in 2005, so this timespan gives 11 years before implementation and 15 years after to observe the effects of the law on asylum decision making by IJs. The dependent variable in our dataset is whether an IJ grants asylum or not (coded 1 if yes; 0 otherwise). 12 We create three separate indicator variables for the decision making phase—the adoption or selection of the best policy option with regard to credibility standards, across each of the branches of the federal government. Our focus is on the variable for the adoption of the REAL ID standard in case law from the federal circuit courts. We refer to this variable as circuit credibility and code it as one in circuit-years for which there is a precedent requiring the enforcement of the Real ID credibility standards and zero otherwise (see Table 1 above). For Congress we code adoption of the Real ID credibility standard as occurring in 2005 (when the law is passed) and this variable is referred to as congressional credibility. It is equal to one after 2005 and zero before. For the executive branch we code the adoption of Real ID’s credibility standards as occurring in 2007 and we refer to this variable as executive credibility—it is coded one from 2007 on and zero before. As noted above, the BIA adopted the REAL ID’s credibility standard in a precedential decision, Matter of J-Y-C-, in 2007. Our key measures of ideology are straightforward. For each circuit, for each year, we code the percentage of active judges who are Democrats. This variable ranges from 8% to 83% and is called percent Democratic. For the executive branch we simply code partisan control and it is equal to one if Republicans control and zero otherwise (executive GOP). For Congress we code whether there is unified Democratic control (equal to negative one), mixed control (equal to zero), or unified Republican control (equal to one). In our dataset unified Republican control is the most common of the three conditions in Congress and we refer to this variable Congress GOP . In addition to these variables, we include a control for the ideology of the IJ that is coded based on the approach first articulated by Keith et al. (2013) and Miller et al. (2015) and is a factor score of background characteristics with higher scores indicating greater IJ liberalism (i.e., an inclination to grant asylum). Note, we estimate models with additional controls for the characteristics of the individual case (including whether the asylee has an attorney and characteristics of the country they are fleeing) in the model with controls. We provide additional details on these control variables in the appendix. Results Given the dichotomous nature of our dependent variable we estimate logit models with standard errors simultaneously clustered on the IJ and on the year to account for sources of non-independence in the errors. Table 2 includes results for two regressions, one without the controls discussed in the appendix and one including them—our results do not vary across the specifications. 13 In discussing the results we focus on Model 1. Both models fit the data well, but in what follows we will focus on interpretation of the interaction terms in Model 1 graphically. We will revisit the importance of IJ ideology subsequently, although it is worth noting that it predicts IJ votes in Model 1, but not in Model 2.
STOBB eT al .8 Our hypotheses are focused on how the adoption of the REAL ID Act’s new credibility standards by each branch altered the decision making environment and allowed for tighter control of IJ decision making by those above them in the political and legal hierarchy. Figure 1 illustrates the relationship between the partisanship of the bench in the federal appellate courts and the likelihood that an IJ grants asylum in that circuit. The dashed-line represents the effects after the REAL ID credibility standards are put in to effect in a circuit, while the solid-line represents the effects before the credibility stand ards are put in place. Confirming Hypothesis 1, before REAL ID there is very little relationship between the presumed aggregate preferences in the appellate courts and the decision making of IJs in that circuit; while after the REAL ID Act is implemented by the circuit, as the percentage of the circuit that is Democratic increases so too will the likelihood on of an IJ granting asylum. In post-REAL ID era within each circuit (which as a reminder occurs in different years in each circuit) the model predicts that a circuit dominated by Republicans (20% Democratic) will correlate with a likelihood of granting asylum of just 18%. Alternatively, in a circuit heavily dominated by Democrats (80% Democratic) the likelihood of asylum being granted jumps to 70%. The change over a similar spread of partisan dominance is about 3 percentage points in the pre-REAL ID period. Figure 1 offers strong support for the notion that REAL ID engendered greater control of IJ decision making by the circuit courts. Hypotheses 2 suggest that REAL ID also engendered greater ideological control by the legislature. As a reminder, we code the adoption of the REAL ID credibility standard by Congress as the enactment of the statute itself. Figure 2 displays the results of an interaction of partisan control of Congress with the implementation of REAL ID on the likelihood an IJ grants asylum. The solid line represents the relationship between asylum rates and congressional control before REAL ID is enacted. There is a slight Model 1 Model 2 Coeff. S.E. Coeff. S.E. Credibility standards Circuit Cred. −1.82* 0.29 −1.51* 0.29 Congressional Cred. 0.71* 0.21 0.81* 0.20 Executive Cred. 0.15 0.25 0.91* 0.24 Political control Circuit Perc. Democratic 0.79* 0.28 1.06* 0.29 Congress GOP 0.05 0.06 0.15* 0.04 Executive GOP 0.24* 0.09 0.46* 0.06 Interactions Percent Dem * Circ. Cred. 3.66* 0.75 2.93* 0.73 Cong. GOP * Cong. Cred. −0.42* 0.10 −0.50* 0.10 Exec. GOP * Exec Cred. −0.72* 0.18 −1.27* 0.19 IJ Liberalism 0.09* 0.04 0.07 0.04 Constant −1.31 0.18 −1.80 0.24 N 941,375 902,601 Years 27 26 Immigration Judges 716 687 Controls No Yes Wald χ 2 440.47 (p = 0.000) 1291.10 (p = 0.000) PRE 0.07 0.11 *p < 0.05.
MIXeD MeSSa GeS & BOUNDeD Ra TIONalITY 9 upward slope, which runs counter to the expected ideological proclivities of a conservative Congress (e.g., Miller et al., 2020). A Congress controlled by Republicans should not result in an increase in grant rates over one controlled by Democrats. The dashed-line represents the effect of control of Congress on asylum grant rates after the enactment of REAL ID. Here we see a significant downward slope, as grant rates decline from 52% under Democratic control to 39% under Republican control. This is strongly suggestive evidence in support of hypothesis 2—REAL ID seemed to have aided the translation of the preferences of Congress into IJ decision making. Hypothesis 3 is focused on the role of REAL ID in enhancing the correspondence between the preferences of the executive branch and the decision making of IJs. Once again, the solid line shows the difference in grant rates before the enactment of REAL ID. The effect here is the opposite of what we would expect, given conservative preferences over asylum granting. The grant seems to be slightly higher in GOP administrations than in Democratic ones. Alternatively, focusing on the dashed-line representing effects after the enactment of the credibility standard in the BIA in 2007 we see the relationship we would expect—grant rates are lower in Republican administrations than in Democratic ones. Note, however, that the 95% confidence intervals overlap in this scenario, so we cannot be sure that there is a real differ-
STOBB eT al .10 ence between administrations in the post-REAL ID environment. Figure 3, then, offers some tentative support for hypothesis 3. Taking the evidence across all three branches into account, there is strong evidence that the adoption of the REAL ID’s credibility standard enabled close ideological control of the IJs by political or legal superiors. Within this larger conclusion, it is also the case that change in the partisanship of the circuits is the most important predictor of how IJs will decide asylum cases in the post-REAL ID environment. It is slightly hard to parse out whether to refer to the effects of circuit partisanship that we see here as partisan/ideological effects or as legal effects. Technically speaking, what we have coded are changes in the legal standards used to assess the credibility of asylum applicants, and interacted them with changes in circuit ideology. Yet it is also clear that the language of the law enacted is vague enough to allow for considerable discretion to come into IJ assessments of applicant credibility. If IJs rationally anticipate review of their decision, as we expect they do, then it is logical that they also anticipate how the ideological proclivities of the reviewing circuit will affect review . Robustness checks Our theory is that IJs act as boundedly rational agents attempting to respond to conflicting signals from multiple principals. One empirical implication of that story of IJ behavior is that an IJ’s own ideology should not be any more predictive of their decision to grant asylum based on the enactment of REAL ID, since the change induced by REAL ID is responsiveness to principals. The most potent of the three principals we have examined are the circuit courts and so we test this expectation with respect to REAL ID and IJ policy preferences in the context of the enactment of REAL ID standards by the circuit courts.
Figure 4 below displays this result with the underlying model presented in the appendix.Figure 4 clearly
shows no statistically discernable change in how IJ ideology affects IJ decision making, bolstering support for our bounded rationality approach. One argument with respect to our findings of responsiveness to the partisanship of the president is that it is driven mostly by responsiveness to Trump, who utilized immigration as a campaign issue to a greater extent than previous presidents (Gimpel, 2017). Therefore, it is possible that our results for post-REAL ID responsiveness, which encompass only three presidents: George W . Bush, Obama, and Trump, are driven by the unusual character of Trump’s issue focus. To test this we re-estimated our model excluding Trump and found results that closely mirror those presented above and we are confident that
MIXeD MeSSa GeS & BOUNDeD Ra TIONalITY 11 our results for responsiveness to the White House are not induced by the uniqueness of the Trump presidency. The model used to estimate this result is presented in the appendix. DISCUSSION AND CONCLUSION Our results indicate that, rather than freeing immigration judges from the control of circuit court judges, as at least some members of Congress intended, the adoption of the REAL ID Act’s credibility standard enabled significantly closer ideological control by the courts. At the same time, the legislature and the executive also gained influence. The latter may or may not have been planned. Comparing the changes in the impact of the three principals before and after the adoption of the REAL ID standards by each, we demonstrate that the increase in the influence of circuit preferences is the greatest. Change in the partisanship of the circuits is the most important predictor of IJ decision making in asylum cases, and the power is in large part due to the REAL ID Act’s policy change. Therefore, although the Act benefited all three principals, the circuit courts benefited the most, making them the winners in the competition for influence. Ultimately, the increase in discretion led to a reining in of IJs, as their decisions moved into alignment with the ideological preferences of their superiors, and did not significantly alter the impact of their personal preferences. We argue several key factors can explain this outcome: ambivalent statutory language, a multiplicity of principals, close monitoring by circuit courts, and both the professionalism of immigration judges along with the professional norms associated with their station. Researchers studying immigration policy, and policy studies scholars in general, can build on these findings by examining the effects of these elements in other contexts. For example, prior research examining the implementation of immigration policy found that a high degree of professionalism (among teachers and health care professionals, for example) tends to lead to the use of policy ambivalence to help rather than exclude undocumented immigrants from social benefits, according to the bureaucrats’ preferences and contra to the express intent of at least some supervisors (Giuraudon & Lahav, 2007; Van der Leun, 2007). Those lower on the professionalism scale seem to apply the rules more legalistically (Giuraudon & Lahav, 2007 ; Van der Leun, 2007). We find that for administrative adjudicators who, as lawyers, are high on the professionalism scale, ambivalence and increases in discretion enhance hierarchical control, particularly by the judiciary. Further research is needed to understand frontline implementation by these key bureaucratic actors.
STOBB eT al .12 Our results also have important implications for the area of immigration policy studies specifically. Scholars recognize the inevitable ambivalence of policies concerning undocumented immigrants (Giuraudon & Lahav, 2007; Van der Leun, 2003, 2007). Future research should examine whether bureaucrats, particularly adjudicative administrators, respond to other statutes or policy commands by adhering more closely to their political and legal superiors. Such studies could look to the U.S. context, and compare our findings with the responses of similarly positioned adjudicators abroad. As the use of administrative adjudicators, and legalistic procedures, is on the rise in numerous policy areas (Koch, 2005), the implications are wide-reaching. The results speak as well to issues of bureaucratic control. As scholars continue to debate the relative influence of Congress, the president, and the courts on the behavior of federal agencies, our research can shed some light on the nature of this relationship between principal(s) and agent(s). The work here has focused on IJs, and future research would benefit from expanding to studying a broader array of agencies and bureaucrats, both those engaging primarily in adjudicative activities, as IJs do, and those engaging principally in regulatory activities. While we have found evidence that all three principals can have a measurable effect on agency outcomes, the relative influence of the judiciary is worth considerable attention. Does this level of influence carry across all agencies? We have argued that IJs, being lawyers by training, satisfice in part by leaning on professional norms and are more deferential to judges as a result. It is not clear that, say, economists in the Federal Trade Commission would be as deferential to judges; perhaps the courts retain primary influence over agencies engaging in adjudicative behavior, while Congress and the president are more influential in regulatory contexts. These salient questions merit additional scholarly attention. ORCID Maureen Stobb https://orcid.org/0000-0003-2865-3815 ENDNOTES 1 Comment by Senator Brownback, 151 Cong. Rec. S4838 (daily ed. May 10, 2005). Senator Brownback argued these changes require that credibility determinations be reasonable, and defined an unreasonable finding as one that is based on inconsistencies not going to the heart of the claim, lacking other evidence of the asylum applicant’s deception, and failing to take into account the applicant’s individual circumstances, such as gender, cultural or educational background, and history of trauma. 2 Comment by Representative Sensenbrenner, the sponsor of the bill. 151 Cong. Rec. at H453 (daily ed. February 9, 2005). 3 REAL ID Act § 101(a)(3), 119 Stat. at 303 (to be codified at section 208(b)(1)(B)(iii) of the Immigration and Nationality Act , 8 U.S.C. § 1158(b)(1)(B)(iii)). 4 REAL ID text: Considering the totality of the circumstances, and all relevant factors, a trier of fact may base a credibility determination on the demeanor, candor, or responsiveness of the applicant or witness, the inherent plausibility of the applicant’s or witness’s account, the consistency between the applicant’s or witness’s written and oral statements (whenever made and whether or not under oath, and considering the circumstances under which the statements were made), the internal consistency of each such statement, the consistency of such statements with other evidence of record (including the reports of the Department of State on country conditions), and any inaccuracies or falsehoods in such statements, without regard to whether an inconsistency, inaccuracy, or falsehood goes to the heart of the applicant’s claim, or any other relevant factor. REAL ID Act § 101(a)(3), 119 Stat. at 303 (to be codified at section 208(b)(1)(B)(iii) of the Act, 8 U.S.C. § 1158(b)(1)(B)(iii)). 5 The Board found permissible the immigration judge’s reliance upon the “rapid” and “agitated” manner of the applicant’s testimony, conflicts between the applicant’s testimony and his prior statement, and those of a witness, and the applicant’s failure to produce corroborating evidence that was reasonably available (Matter of J-Y-C-, 2007, 264). 6 See Lin v. Mukasey ( 2008a) (1st Circuit); Lin v. Mukasey ( 2008b) (2nd Circuit); Wang v. Holder (2009) (5th Circuit); El-Moussa v. Holder (2009) (6th Circuit); Limbeya v. Holder (2014) (8th Circuit); Chen v. U.S. Atty. Gen. (2006) (11th Circuit). Note, IJs are subject to the jurisdiction of the circuit court that covers the geographical area in which they make a decision. So, an IJ deciding cases in California will be subject to the rules of the 9th Circuit, while one deciding cases in Florida will be subject to the rules of the 11th Circuit. 7 Lin v. Mukasey 2008a, 27 FN3. 8 Lin v. Mukasey 2008b; Wang v. Holder (2009). 9 B.C. v. Attorney General United States (2021). 10 Hasan v Holder (2009); Krishnapillai v. Holder (2009); Shrestha v. Holder (2010).
MIXeD MeSSa GeS & BOUNDeD Ra TIONalITY 13 11 See, for example, Djadjou v. Holder (2011), Htun v. Lynch (2016). 12 The appendix contains descriptive information for all of the included variables. 13 Note that the N varies between the two models because we do not have complete data across all years for the variables included in the model with controls.
Anker, Deborah E. 2016. Law of Asylum in the United States. Eagan: Thomas Reuters. Bonjour, Saskia. 2011. “The Power and Morals of Policy Makers: Reassessing the Control Gap Debate.” International Migration Review 45(1): 89–122. Bonjour, Saskia. 2016. “Speaking of Rights: The Influence of Law and Courts on the Making of Family Migration Policies in Germany.” Law & Policy 38(4): 328–48. Boswell, C. 2007. “Theorizing Migration Policy: Is There a Third Way?” International Migration Review 41(1): 75–100. Carpenter, Daniel P . 2001. The Forging of Bureaucratic Autonomy: Reputations, Networks, and Policy Innovation in Executive Agencies, 1862–1928. Princeton: Princeton University Press. Cianciarulo, Marisa Silenzi. 2006. “Terrorism and Asylum Seekers: Why the Real ID Act is a False Promise.” Harvard Journal on Legislation 43(1): 101–43. Clinton, Joshua D., David E. Lewis, and Jennifer L. Selin. 2014. “Influencing the Bureaucracy: The Irony of Congressional Oversight.” American Journal of Political Science 58(2): 387–401. Cornelius, Wayne A., Philip L. Martin, and James F. Hollifield. 1994. “Introduction: The Ambivalent Quest for Immigration Control.” In Controlling Immigration: A Global Perspective, edited by Wayne A. Cornelius, Philip L. Martin, and James F. Hollifield, 3–41. Stanford: Stanford University Press. Cornelius, Wayne A., and Takeyuki Tsuda. 2004. Controlling Immigration: The Limits of Government Intervention. Stanford, CA: Stanford University Press. Ellerman, Antje. 2007. “Street-Level Democracy: How Immigration Bureaucrats Manage Public Opposition.” In Immigration Policy in Europe, edited by Virginie Guiraudon and Gallya Lahav, 93–109. New York: Routledge. Epstein, David, and Sharyn O’Halloran. 1999. Delegating Powers: A Transaction Cost Politics Approach to Policy Making Under Separate Powers. New York: Cambridge University Press. Filindra, Alexandra, and Sara Wallace Goodman. 2019. “Studying Public Policy through Immigration Policy: Advances in Theory and Measurement.” Policy Studies Journal 47(3): 498–516. Fletcher, Aubra. 2006. “The REAL ID Act: Furthering Gender Bias in U.S. Asylum Law.” Berkeley Journal of Gender, Law & Justice 21: 111–31. Freeman, Gary P . 1992. “Migration Policy and Politics in the Receiving States.” The International Migration Review 26(4): 1144–67. Freeman, Gary P . 1994. “Can Liberal States Control Unwanted Migration?” The Annals of the American Academy of Political and Social Science 534: 17–30. Gailmard, Sean. 2009. “Multiple Principals and Oversight of Bureaucratic Policy-Making.” Journal of Theoretical Politics 21(2): 161–86. Galloni, Tania. 2008. “Keeping it Real: Judicial Review of Asylum Credibility Determinations in the Eleventh Circuit after the REAL ID Act.” University of Miami Law Review 62(4): 1037–62. Gimpel, James G. 2017. Immigration Policy Opinion and the 2016 Presidential Vote . Washington, D.C.: Center for Immigration Studies https://cis.org/Report/Immigration-Policy-Opinion-and-2016-Presidential-Vote. Golden, Marissa Martino. 2000. What Motivates Bureaucrats? Politics and Administration During the Reagan Years . New York: Columbia University Press. Guiraudon, Virginie, and Gallya Lahav. 2000. “A Reappraisal of the State Sovereignty Debate: The Case of Migration Control.” Comparative Political Studies 33(2): 163–95. Guiraudon, Virginie, and Gallya Lahav, eds. 2007. Immigration Policy in Europe. New York: Routledge. Hamlin, Rebecca. 2014. Let Me Be a Refugee: Administrative Justice and the Politics of Asylum in the United States, Canada, and Australia. Oxford: Oxford University Press. Hollifield, James F. 2000. “The Politics of International Migration: How Can We ‘Bring the State Back In’?” In Migration Theory: Talking across Disciplines, edited by Caroline Brettell and James F. Hollifield, 137–86. New York: Routledge. Huber, Gregory A. 2007. The Craft of Bureaucratic Neutrality: Interests and Influence in Governmental Regulation of Occupational Safety. New York: Cambridge University Press. Huber, John D., and Charles R. Shipan. 2002. Deliberate Discretion? The Institutional Foundations of Bureaucratic Autonomy. New York: Cambridge University Press. Huber, John D., Charles R. Shipan, and Madelaine Pfahler. 2001. “Legislatures and Statutory Control of Bureaucracy.” American Journal of Political Science 45(2): 330–45. Johannesson, Livia. 2018. “Exploring the “Liberal Paradox” from the Inside: Evidence from the Swedish Migration Courts.” International Migration Review 52(4): 1162–85. Jones, Bryan D. 1999. “Bounded Rationality.” Annual Review of Political Science 2: 297–321. Joppke, Christian. 1998. “Why Liberal States Accept Unwanted Migration.” World Politics 50: 266–93. Joppke, Christian. 1999. Immigration and the Nation-State: The United States, Germany, and Great Britain. Oxford: Oxford University Press.
STOBB eT al .14 Kanstroom, Daniel. 2006–2007. “The Better Part of Valor: The REAL ID Act, Direction, and the Rule of Immigration Law.” New York Law School Law Review 51(1):161–208. Keith, Linda Camp, Jennifer Holmes, and Banks Miller. 2013. “Explaining Divergence in Asylum Grant Rates among Immigration Judges: An Attitudinal and Cognitive Approach.” Law & Policy 35: 261–89. Keith, Linda Camp, Banks P . Miller, and Jennifer S. Holmes. 2015. “How Draconian Are the Changes to US Asylum Law? A Monthly Time Series Analysis (1990-2010).” Human Rights Quarterly 37(1): 153–87. Koch, Charles. 2005. “Policymaking by the Administrative Judiciary.” Alabama Law Review 56: 693–740. Lipsky, Michael. 1980. Street-level Bureaucracy: Dilemmas of the Individual in Public Services. New York: Russell Sage Foundation. Mascia, Carla. 2020. “How bureaucracies shape access to rights: The implementation of family reunification in Belgium.” Journal of Ethnic and Migration Studies 47: 2127–43. https://doi.org/10.1080/1369183X.2020.1726734. Miller, Banks, Linda Camp Keith, and Jennifer S. Holmes. 2015. Immigration Judges and U.S. Asylum Policy. Philadelphia: University of Pennsylvania Press. Miller, Banks, Jennifer S. Holmes, and Linda Camp Keith. 2020. “The Preferences of Political Elites and Humanitarian Immigration to the United States.” RSF: The Russell Sage Foundation Journal of the Social Sciences 6(3): 150–71. Miller, Gary J. 1992. Managerial Dilemmas: The Political Economy of Hierarchy. New York: Cambridge University Press. Miller, Gary J. 2005. “The Political Evolution of Principal-Agent Models.” Annual Review of Political Science 8: 203–25. Niskanen, William A. 1975. “Bureaucrats and Politicians.” Journal of Law and Economics 18: 617–43. Pressman, Jeffrey L. and Aaron Wildavsky. 1984. Implementation (3rd expanded ed.). Berkeley: Oakland University Press (orig. 1973). Potter, Rachael Augustine. 2019. Bending the Rules: Procedural Politicking in the Bureaucracy. Chicago: University of Chicago Press. Shepsle, Kenneth A. 1992. “Congress Is a ‘They,’ Not an ‘It’: Legislative Intent as Oxymoron.” International Review of Law and Economics 12(2): 239–56. Segal, Jeffrey A., and Harold J. Spaeth. 1993. The Supreme Court and the Attitudinal Model. New York: Cambridge University Press. Simon, Herbert A. 1957. Administrative Behavior: A Study of Decision Making Processes in Administrative Organization, 2nd ed. New York: The Free Press. Simon, Herbert A. 1978. “Rationality as Process and as Product of Thought.” The American Economic Review 68(2): 1–16. Spence, Michael, and Richard Zeckhauser. 1971. “Insurance, Information, and Individual Action.” The American Economic Review 61(2): 380–7. Stobb, Maureen. 2019. “Power of the Pen or the Gavel? Determining Asylum Standards on the Courts of Appeals.” Justice System Journal 40(3): 196–220. Van der Leun, Joanne. 2003. Looking for Loopholes: Processes of Incorporation of Illegal Immigrants in The Netherlands. Amsterdam: Amsterdam University Press. Van der Leun, Joanne. 2007. “Excluding Illegal Migrants in The Netherlands: Between National Policies and Local Implementation.” In Immigration Policy in Europe, edited by Virginie Guiraudon and Gallya Lahav, 93–109. New York: Routledge. White, Victor P . 2006. “U.S. Asylum Law out of Sync with International Obligations: REAL ID Act.” San Diego International Law Journal 8(1): 209–58. Wilson, James Q. 1989/2000. Bureaucracy: What Government Agencies Do and Why They Do It. New York: Basic Books. CASES CITED B.C. v. Attorney General United States, 12 F.4th 306 (3 rd Cir. 2021). Chen v. U.S. Atty. Gen. 2006 (11 th Cir. 2006). Djadjou v. Holder, 662 F.3d 265 (4 th Cir. 2011). El-Moussa v. Holder, 569 F.3d 250 (6 th Cir. 2009). Hasan v Holder, 571 F.3d 631 (7 th Cir. 2009). Htun v. Lynch, 818 F.3d 1111 (10 th Cir. 2016). Krishnapillai v. Holder, 563 F.3d 606 (7 th Cir. 2009). Limbeya v. Holder, 764 F.3d 894 (8 th Cir. 2014). Lin v. Mukasey, 521 F.3d 22 (1 st Cir. 2008a). Lin v. Mukasey, 534 F.3d 162 (2 nd Cir 2008b). Matter of J-Y-C-, 24 I&N Dec. 260 (BIA 2007). Shrestha v. Holder, 590 F.3d 1034 (9 th Cir. 2010). Wang v. Holder, 569 F.3d 531 (5 th Cir. 2009). STATUTES CITED Immigration and Nationality Act of 1965, 8 U.S.C.A. §1231 and 1252. REAL ID Act: Emergency Supplemental Appropriations Act for Defense, the Global War on Terror, and Tsunami Relief, 2005, PL 109-13, 11 May 2005, 119, Stat 231.
MIXeD MeSSa GeS & BOUNDeD Ra TIONalITY 15 AUTHOR BIOGRAPHIES How to cite this article: Stobb, Maureen, Banks Miller, and Joshua Kennedy. 2022. Mixed messages & bounded rationality: The perverse consequences of real ID for immigration policy. Children & Society, 00, 1–18. https://doi.org/10.1111/psj.12486
Maureen Stobb is an Associate Professor of Political Science at Georgia Southern University. Her research focuses on the expansion of judicial power relative to the legislature and the executive, particularly in the policy areas of immigration and citizenship. Her research has been published in various outlets including The Journal of Law & Courts, and Justice System Journal. Banks Miller is an Associate Professor of Political Science at the University of Texas at Dallas. His research focuses on judicial decision making, intellectual property policy, and immigration policy. Recent work has been published in the Journal of Law & Courts and American Politics Research. Joshua B. Kennedy is an Associate Professor of Political Science at Georgia Southern University. His research of late focuses on political control of the administrative state, and he has also published in the area of unilateral presidential power. His research has appeared in American Politics Research, Research & Politics, and Presidential Studies Quarterly, among other outlets. Mean Std. Dev. Min. Max. Asylum Grant 0.35 0.47 0 1 Circuit Cred. 0.74 0.44 0 1 Congressional Cred. 0.48 0.50 0 1 Executive Cred. 0.45 0.50 0 1 Circuit Perc. Democratic 0.49 0.17 0.07 0.83 Congress GOP 0.54 0.50 0 1 Executive GOP 0.53 0.50 0 1 IJ Liberalism 0.08 0.97 −1.49 4.24 Trade (Logged) 15.34 3.04 10.96 20.27 Detention 0.32 0.69 0 2 Attorney 0.68 0.47 0 1 PTS 3.35 0.92 1 5 Post 9/11 0.64 0.48 0 1
STOBB eT al .16 We include a host of control variables in Model 2 in the paper and we describe them here. First, we include a bilateral measure of trade (logged to account for skew) between the U.S. and the country from which an asylee is fleeing. The reason for including this variable is to help control for the economic relationship between the U.S. and the sending state. We include a measure of the level of repression in a state using the Political Terror Scale ( PTS), with higher scores indicating greater levels of repression. Detention status is a variable provided by EOIR that accounts for whether an asylee has never been in detention (=0), was previously detained but is not at the time of the hearing (=1), or is currently detained (=2). Attorney is a variable that is coded dichotomously to indicate whether an asylee has retained counsel, based on data also provided by EOIR. Finally, post-9/11 is dummy variable to indicate whether the hearing takes places after the year 2001.
Table A2 displays model 2, presented in the paper, but includes the coefficients for the control variables. Note, throughout the appendix coefficient indicated with an * are statistically significant at p < 0.05
(two-tailed). Each of these variables behaves as we would expect. For instance, we would expect asylees fleeing highly repressive states to be significantly more likely to receive asylum and the large, positive, and statistically significant coefficient indicates that this is the case. Similarly, we expect that asylees who are represented by counsel should be significantly more likely to receive asylum, and indeed this is confirmed in the model. Model 2 Coeff. S.E. Credibility Standards Circuit Cred. −1.51* 0.29 Congressional Cred. 0.81* 0.20 Executive Cred. 0.91* 0.24 Political Control Circuit Perc. Democratic 1.06* 0.29 Congress GOP 0.15* 0.04 Executive GOP 0.46* 0.06 Interactions Percent Dem * Circ. Cred. 2.93* 0.73 Cong. GOP * Cong. Cred. −0.50* 0.10 Exec. GOP * Exec Cred. −1.27* 0.19 Control Variables IJ Liberalism 0.07 0.04 Trade −0.07* 0.01 PTS 0.24* 0.05 Detention −0.58* 0.08 Attorney 0.65* 0.14 Post-9/11 −0.21* 0.05 Constant −1.80 0.24 N 902,601 Years 26 Immigration Judges 687 Controls Yes Wald χ 2 1291.10 (p = 0.000) PRE 0.11
MIXeD MeSSa GeS & BOUNDeD Ra TIONalITY 17
Table A3 below shows the modeling results used to generate Figure X in the text, exploring the effect
of IJ ideology on IJ decision making around the implementation of REAL ID.
Table A4 illustrates the model we use as the basis for the claim in the paper that excluding the Trump
presidency does not alter our conclusion with respect to the responsiveness of IJs to the White House after the implementation of REAL ID. The coefficient for the Exec. GOP * Exec Cred. Interaction is similarly sized and signed as that in the model presented in the paper and it remains statistically significant. Although interpretation of the interaction coefficient is not itself enough to confirm uniformity across the models we supplemented this model with a graphical exploration of the results that confirms their similarity to those presented in the paper. [psj12486-tableappendix-0006] The data provided by the EOIR can be found on their website at https://www.justice.gov/eoir/foia-library-0. The original dataset included approximately 9 million immigration proceedings. Each immigration case (“idncase”) can have multiple proceedings (“indnproceeding”) and hearings (“idnschedule”). We narrowed the proceedings to removal proceedings only using EOIR data on the case type (“case_type”). Next, we used EOIR data identifying whether the noncitizen filed an application for asylum and/or withholding of removal (“Tbl_Court_Appln), and reduced the dataset to only asylum cases. We eliminated cases cleared via prosecutorial discretion and riders to lead cases from our dataset. We then narrowed the data to include only the first substantive hearing on the merits. The result was one hearing per asylum case. IJ liberalism interaction Coeff. S.E. Credibility Standards Circuit Cred. −0.10 0.15 Congressional Cred. 0.64* 0.21 Executive Cred. −0.43* 0.2 Political Control Circuit Perc. Democratic 1.59* 0.36 Congress GOP −0.32* 0.05 Executive GOP −0.34* 0.13 Interactions IJ Liberalism * Circ. Cred. 0.09 0.08 IJ Liberalism 0.09* 0.04 Constant −1.19 0.17’ N 941,375 Years 27 Immigration Judges 716 Controls No Wald χ 2 353.69 (p = 0.000) PRE 0.03 *p < 0.05.
STOBB eT al .18 Trump excluded Coeff. S.E. Credibility Standards Circuit Cred. −2.02* 0.28 Congressional Cred. 0.71* 0.19 Executive Cred. −0.03 0.20 Political Control Circuit Perc. Democratic 0.83* 0.29 Congress GOP 0.06 0.06 Executive GOP 0.24* 0.09 Interactions Percent Dem * Circ. Cred. 4.73* 0.65 Cong. GOP * Cong. Cred. −0.45* 0.07 Exec. GOP * Exec. Cred. −0.44* 0.11 IJ Liberalism 0.09* 0.04 Constant −1.33 0.19 N 739,168 Years 22 Immigration Judges 555 Controls No Wald χ 2 393.21 (p = 0.000) PRE 0.09 *p < 0.05.