Printer Friendly

The second-order structure of immigration law.

INTRODUCTION
I. OVERVIEW OF AMERICAN IMMIGRATION LAW
 A. First-Order Policy Preferences
 B. Second-Order Institutional Design
 C. Constitutional Restrictions
II. THE THEORY OF SECOND-ORDER IMMIGRATION DESIGN
 A. Information and Screening Devices
 B. Risk and Country-Specific Investment
 C. Other Factors
 D. A Comparison
 E. Positive Implications
III. THE SECOND-ORDER STRUCTURE OF AMERICAN IMMIGRATION LAW
 A. Ex Ante vs. Ex Post
 1. The shift from exclusion to deportation
 2. The gap between exclusion and deportation grounds
 3. Constitutional constraints on second-order design
 B. The Ex Post System: Illegal Immigration and Guest Workers
 1. Reconceptualizing illegal immigration
 2. The guest worker alternative
 C. The Ex Ante System: Family Migration and Alternatives
CONCLUSION


INTRODUCTION

Immigration law has long been underappreciated in the legal academy but, as so often happens, the pressure of current events is causing scholars to rethink old ideas. Two such events are of special importance. First, the U.S. government's reaction to the 9/11 attacks raised questions about the extent to which the norms of equal protection apply to noncitizens on American soil. In the immediate aftermath of the attacks, federal officials conducted sweeps in which they rounded up over a thousand noncitizens on alleged visa violations. Nearly all of these noncitizens were from predominantly Muslim countries. Many were detained for months without charges, and some were eventually released without ever being charged. (1) The Department of Justice subsequently pursued several other immigration initiatives that targeted noncitizens from predominantly Muslim countries. (2)

Second, the influx of undocumented aliens in the 1980s and 1990s, most of them Hispanic, has brought to the headlines the traditional litany of complaints about immigration: that noncitizens take the jobs of citizens, that they overwhelm social services, and that they cannot be assimilated if they arrive in excessive numbers. (3) In the early 1980s, undocumented immigrants made up a small fraction of the entering flow of immigrants, with approximately 130,000 crossing the border each year. (4) By the late 1990s, however, nearly one million undocumented immigrants were entering the United States each year--roughly the same number as were immigrating through legal channels. (5) Today, Congress is debating whether to overhaul the immigration system. The reform proposals include criminalizing unauthorized presence in the United States and creating a large-scale guest worker system. (6)

The current debates differ from many of the major historical debates about immigration policy inasmuch as they focus on what we will call second-order issues of institutional design rather than first-order issues of immigration policy. First-order issues address the ultimate ends of immigration policy: simplifying greatly, these ends entail admitting the state's desired quantity and types of immigrants. Quantity refers to the number of noncitizens granted lawful permanent residency every year. In the United States, the quantity of legal entries per year has varied between 400,000 and 1.8 million over the last three decades. (7) Type refers to the characteristics of the people who are granted lawful permanent residency--their status as family members of U.S. citizens, their work skills, their ability to assimilate, and related characteristics. Second-order design issues concern the legal institutions that are used to implement the first-order policy goals. Focusing on second-order design raises the question of how to screen applicants for admission so that the desired types are admitted and others are excluded.

The academic literature on immigration law and policy has largely neglected these second-order questions of institutional design. Its focus has been almost exclusively on two other questions: on the first-order question of how many, and whom, to admit; (8) and on the question of what role courts play in policing immigration policy. (9) But the second-order question is at least as important as the first-order question. Indeed, if a state makes poor second-order design choices, the importance of the first-order question fades: one cannot have an immigration policy if one cannot enforce it. And the role of courts cannot be understood apart from broader issues of institutional design. (10)

One of our contributions is to identify the importance of second-order questions and provide a framework for addressing them. A central design choice for any state, we argue, is the choice between ex ante and ex post screening. Under an ex ante approach, a state decides whether to accept a particular immigrant on the basis of pre-entry information, such as the immigrant's race or her educational achievement in her home country. In contrast, an ex post approach selects immigrants on the basis of post-entry information, such as her avoidance of criminal activity or unemployment in the host country. The ex ante approach generally, though not unavoidably, leads to a system of exclusion at the border; the ex post approach necessarily leads to a system of deportation. Much of the history of immigration policy can be characterized by one of these two main approaches, but little effort has been made to explain why a state would select one approach or the other.

We explore the advantages and disadvantages for a self-interested state of the ex ante and ex post approaches, as well as combinations of these approaches, and generate predictions about the circumstances under which one approach will dominate. We argue that immigration screening presents an information problem, and that the comparative effectiveness of ex ante and ex post screening turns in part on the solution to that problem. (11) The main screening advantage of the ex post system is that it uses more information (both about the immigrants and about the country's current needs) than the ex ante system does, which minimizes errors. The main advantage of the ex ante system is that it reduces the risk faced by potential immigrants that they will be deported, so that risk-averse noncitizens are more likely to enter and invest in the country than they are under the ex post system. There are also other important tradeoffs. The ex post system requires a probationary period that can be costly--if, for example, dangerous noncitizens commit crimes before being evaluated and deported, or if immigrants develop local ties that are disrupted by deportation. By contrast, the ex ante system can avoid these costs by excluding people at the border. Moreover, the two systems depend on different enforcement schemes, though it is unclear which is more costly: the ex ante system depends on the ability to control the border; the ex post system depends on the ability to detect noncitizens in the host country's territory. (12)

Our framework clarifies numerous positive and normative questions about immigration law. We argue that port-of-entry exclusion systems (which are predominantly ex ante) result in poorer screening than post-entry deportation systems (which are predominantly ex post), but also encourage risk-averse immigrants to make country-specific investments of value to the host country, and may be cheaper to enforce. The choice between the two systems turns in part on tradeoffs among these variables. We also argue that although the U.S. de jure system is highly (although not entirely) ex ante, the U.S. de facto system is predominantly ex post--this is the "illegal immigration system" that results from deliberate underenforcement of immigration law plus periodic amnesties. We explain why, for immigrants covered by the illegal system, the government might prefer such ex post screening to an ex ante regime. And we argue that the government might choose the illegal system over a legal version, which would be a large-scale guest worker program, because the illegal system skirts constitutional restrictions that would reduce the advantages of a legal program. We also use our theory to explore why the role of ex post screening in the American system has steadily increased over the last century, and discuss the potential second-order policy justifications for America's uniquely family-based immigration system.

Our plan is as follows. Part I contains a brief description of the three conceptual elements of immigration law and policy: the first-order policy goals; the second-order rules of institutional design; and constitutional restrictions. Part II sets out our theory of second-order institutional design: we show how information problems motivate the choice between ex ante and ex post screening and lay out several positive implications of our theoretical model. Part III uses the theory as a lens to analyze American law. It argues that many features of immigration law are consistent with the factors identified in our theory of second-order institutional design, and speculates that divergences between theory and practice may in part be attributable to constitutional restrictions.

I. OVERVIEW OF AMERICAN IMMIGRATION LAW

In this Part, we briefly survey immigration law in the United States to highlight the ways in which it reflects second-order design choices and constitutional constraints, not merely first-order policy goals.

A. First-Order Policy Preferences

A central goal of immigration policy for all states is, at a very high level of abstraction, to expand the polity by admitting desirable people. (13) Nonetheless, states have different attitudes about which potential immigrants are desirable under various circumstances. These differences lead states' first-order policy preferences to diverge along three main dimensions: with respect to the quantity of immigrants, the type of immigrants, and the terms of admission. (14)

Quantity. States can choose a range of numerical restrictions. At one extreme, a state permits no immigration; at the other extreme, a state permits unlimited immigration. (15) Nearly all states choose intermediate points, but there is still a great deal of variation. Prior to the twentieth century, the United States had no formal numerical limits; beginning in 1921, the United States imposed an immigration ceiling of 350,000 on the Eastern Hemisphere. (16) Over time, that ceiling ranged from 150,000 (in 1927) to 700,000 (in 1990) and since 1965 has covered the entire globe. (17) Because of exceptions to the ceilings, actual immigration has been somewhat higher; for example, in 2004 legal immigration exceeded 946,000. (18) By comparison, legal immigration in the same year was 202,300 in Germany, 175,200 in France, 88,300 in Japan, 235,800 in Canada, and 266,500 in the United Kingdom. (19)

Quantity restrictions can take various forms. As noted above, the U.S. federal government placed few formal restrictions on immigration prior to the 1870s. (20) During this period, the government often took steps to increase the size of the immigrant flow. (21) Even after the federal government began to place statutory limits on the types of immigrants who could enter, it did not place numerical restrictions on the overall size of the immigrant flow. The first statutory limits on the annual flow of immigrants were not adopted until shortly after World War I. (22) These initially temporary limits were codified in the national origins quota system a few years later, but even these quota laws applied only to the Eastern Hemisphere. (23) Only in the last forty years has the United States established relatively rigid global numerical restrictions on the annual number of immigrants the country will admit. (24)

Type. States also regulate the type of person who may immigrate. Some states use a point system that favors applicants with desired characteristics. (25) These typically include the ability to speak the native language, work skills, educational achievement, and propensity to obey the law. For example, Canada awards points to applicants who are highly educated; who speak English and French proficiently; who have work experience; who are between twenty-one and forty-nine years old; who have arranged for employment in Canada; and (under the category of "adaptability") who have an educated spouse or partner, have had prior work experience in Canada, or have a family relation in Canada. (26) The United States places more weight on family relationship, though it too favors immigrants who have desired work skills. (27) Before it imposed numerical restrictions, the United States did not have such elaborate and specific criteria for type, but it would be a mistake to think that the type of immigrant is a new concern. The Alien and Sedition Acts passed by the first Congress permitted the deportation of disloyal or subversive aliens, (28) and many states in the post-founding era had laws that permitted the expulsion (from state territory) of public charges and criminals. (29) Starting in 1875, Congress passed laws designed to exclude noncitizens on the basis of race (initially Chinese, then covering noncitizens from most of East and South Asia) and later on the basis of national origin (disfavoring, for example, southern Europeans). (30) Today, the United States treats a criminal record as an important indication that a person is of an undesired type. (31)

Terms of admission. States also differ in the status that they confer on those permitted to immigrate. At one extreme, a state may confer full citizenship on an immigrant; at the other extreme, a state may permanently deny an immigrant the legal incidents of citizenship. For example, while the United States places substantial constraints on the numbers and types of immigrants it admits, today it places relatively few conditions on their terms of admission. Most noncitizens admitted to lawful permanent residence in the United States have a relatively easy path to citizenship. (32) They must live in the country for five years before becoming eligible to naturalize, but this is nearly the only meaningful condition they must satisfy. (33) Many other states have been less welcoming. In Germany, for example, a guest worker system under which resident workers (and their children) were ineligible for citizenship was the norm for much of the twentieth century. (34) Moreover, the naturalization requirements have not always been so easy to satisfy in the United States. Until 1952, the United States restricted naturalization on the basis of race, which had the effect of permanently depriving some immigrants of access to full membership in the political community. (35)

We should be clear that we describe here only the immigrant admission system in the United States, not the system used to admit nonimmigrants. Nonimmigrants are those noncitizens admitted for a temporary period, such as tourists or employees who receive temporary authorization to work in the country. (36) Immigrants, by contrast, are admitted to permanent residence in the country--residence that is not contingent on retaining employment, learning English, and so forth--and are on a path to eventual citizenship. (For that reason admitted immigrants are typically referred to as "lawful permanent residents"). We focus on the structure of the immigrant system, because our interest here is in the system that the state uses to select those in the immigrant pool whom it considers desirable to add to the country's population and eventually to the citizenry. (37) As we explain below, however, a state might choose to use a temporary immigration system--such as a guest worker program--as a screening mechanism for potential permanent immigrants. (38)

This highlights one last important point about first-order preferences--a point that foreshadows the following discussion on second-order design. Though restrictions concerning immigrant numbers, types, and terms of admission often reflect a state's first-order preferences, they need not always do so. Because such restrictions are closely interrelated, a restriction along one dimension can be used as an instrument to advance a different first-order preference. Restrictions on terms of admission are, as we note above, one example of this. Numerical restrictions offer another example: they can reflect a first-order preference, but can also be used as a second-order mechanism to control the types of immigrants (and vice versa). The national origins quota system that Congress enacted in 1924 was designed to do just this: the quota law's formula was intended to constrict the flow of immigrants from southern and eastern Europe, whom Congress saw as racially inferior to their western and northern European counterparts. (39)

B. Second-Order Institutional Design

Second-order institutional design focuses on the sorting of applicants for immigration. How do states screen out undesired types so that only desired types will be admitted?

Substance. The American approach is formally embodied in three prominent features of U.S. immigration law: the requirements for admissibility, inadmissibility, and deportability. (40) An applicant is admissible if he satisfies the legal grounds of admission, of which the two most important are having family members living in the United States and qualifying for work that cannot be performed by an American worker. (41) But such an applicant can be excluded if he falls under any of the criminal, health, or national security grounds of inadmissibility specified in section 212 of the Immigration and Nationality Act (INA). (42) For example, under the INA, an otherwise admissible noncitizen is inadmissible if he has a prior conviction for a "crime involving moral turpitude," (43) if he has provided "material support" to a designated terrorist organization, (44) or if he is HIV positive. (45)

The INA also provides for post-admission screening, authorizing the government to deport certain noncitizens after they have been admitted to the country, largely on the basis of their post-entry activities but also if their original application was fraudulent. (46) The deportability provisions are similar to the inadmissibility grounds in section 212. For example, a number of the crime- and national security-related grounds of inadmissibility are also grounds of deportability. (47) However, the grounds of inadmissibility and deportability are not coextensive, as we will discuss in Part III. (48)

Parallel to this legal track, following which immigrants present themselves for admission and are evaluated under the admissions rules, there is an illegal track. Many immigrants sneak across the border with no documents, overstay tourist or temporary work visas, or use fraudulent documents to enter. (49) These unauthorized immigrants are often much easier to deport, both on substantive grounds and (as we will discuss shortly) as a matter of procedure, than are legal immigrants. (50) Though the presence of unauthorized immigrants is generally treated as an enforcement failure, we will argue that the illegal track reflects policy choices and has advantages for the government.

Procedures. The substantive criteria are applied in diverse procedural settings. For the most part, American immigration law reserves the most summary procedures for non-resident aliens who arrive at a point of entry (either along the border or at an airport) and seek admission to the country without documents or with fraudulent documents. These noncitizens are subject to a summary screening process known as expedited removal. (51) Under expedited removal, a single immigration official decides, on the basis of the noncitizen's documents and an interview, whether she is entitled to enter the country. If the official concludes that she is, then she is admitted; if the official concludes that she is not, then she is detained and removed from the country. (52) Noncitizens screened through the expedited removal process generally do not have access to a lawyer or other procedural protections, (53) and the government takes the position that they may not seek judicial review of the immigration official's decision. (54)

By contrast, lawful permanent residents whom the government attempts to deport are entitled to a formal removal proceeding. This proceeding begins with a hearing before an administrative immigration judge. (55) Informal rules of evidence apply, and the noncitizen has the right to be represented by a lawyer (though not at the government's expense) and the right to testify and cross examine witnesses. (56) At the conclusion of the hearing, the parties can appeal the immigration judge's decision to the Board of Immigration Appeals. (57) And in many cases, the noncitizen can also petition the federal courts of appeal to review final orders issued by the Board. (58)

While the above rules might suggest that the level of procedural formality turns on whether a noncitizen has entered the country, it is important to recognize that the INA sometimes authorizes the use of summary procedures for noncitizens who have already entered. Expedited removal, for example, can be used against recent undocumented immigrants apprehended within 100 miles of the border. (59) Conversely, immigration law sometimes provides more extensive procedures for noncitizens seeking entry--doing so, for example, in instances where the noncitizen seeking entry has previously been admitted to lawful permanent resident status in the United States. (60) Thus, the level of procedure provided to immigrants by modern American immigration law often turns more on whether the noncitizen has previously been lawfully admitted by the immigration authorities, rather than on whether the noncitizen has physically entered the cotmtry. (61)

C. Constitutional Restrictions

The state's second-order strategies are often shaped by constitutional law. For that reason, we briefly mention here constitutional restrictions on immigration law in the United States. Although the Supreme Court has not spoken clearly on this issue, conventional wisdom holds that constitutional restrictions on immigration law are less strict than they are with respect to other areas of the law. Simplifying greatly, we can summarize as follows the basic contours of when courts hold that the Constitution applies in immigration contexts with minimal, moderate, and ordinary force:

Minimal force. The government is least constrained by the Constitution when it takes an immigration-related action against a noncitizen who arrives at the border for the first time. Courts have at times suggested that the government may exclude such a noncitizen for any reason--even on the basis of race or some other ground that would ordinarily be constitutionally suspect. (62) Similarly, courts have sometimes intimated that the government is free to enforce its exclusion policy with whatever summary procedures it deems appropriate, on the ground that the Due Process Clause does not constrain the government in such situations. (63)

Moderate force. The government is more constrained by the Constitution when it takes an immigration-related action against a noncitizen who has achieved a greater constitutional status than a first-time arriving alien. Courts suggest that noncitizens may receive additional constitutional protections when they have entered the territory of the United States, when they have established significant ties to the United States, or when they have achieved a legal immigration status under statutory and regulatory immigration law. (64) So, for example, courts have held that the government is more constrained by the Constitution when it seeks to deport a lawful permanent resident (LPR) than when it seeks to exclude a first-time arriving alien. (65) This does not mean, however, that the Constitution applies with ordinary force in these contexts. The picture is mixed. The Due Process Clause prohibits the government from subjecting an LPR to the expedited removal procedures that it uses to screen first-time arriving aliens at airports. (66) But courts are divided over the extent to which substantive constitutional provisions such as the Equal Protection Clause and the Free Speech Clause apply to limit the power of the government to deport LPRs. (67)

Ordinary force. The government is most constrained by the Constitution when it regulates lawful resident aliens outside the immigration context, in purely domestic arenas. In the foundational equal protection case Yick Wo v. Hopkins, (68) the Supreme Court held that the Equal Protection Clause protected resident Chinese aliens from a racially discriminatory laundry ordinance in the same way that the clause would have protected citizens. (69) Courts have applied this holding to encompass a wide variety of domestic regulations and a broad spectrum of constitutional rights. The First Amendment has been held to prohibit the punishment of noncitizens for engaging in protected speech, (70) the Fifth Amendment to guarantee criminal process protections to noncitizens charged with crimes, (71) and so on.

As we will explain in Part III, these differential constraints are also crucial to understanding the structure of immigration regulatory policy in the United States.

II. THE THEORY OF SECOND-ORDER IMMIGRATION DESIGN

A central second-order design question is how to sort applicants for immigration, so that only the desired types are admitted, where the desired type is just the type of person who satisfies the criteria derived from a state's first-order immigration goals. (72) This design question is central because a person's type is information that is hidden from the state, and indeed often from the applicant herself. (73) For example, states want people who are assimilable. (74) Some applicants do not know whether they can assimilate easily; they just do not know whether they will fit in. Others, such as those plotting terrorist or criminal activity, do have this information. Neither kind of applicant will reveal her type to immigration authorities--the first because she cannot, the second because she cannot gain from doing so. The second-order design trick is to determine the immigrant's type, even when she is unwilling or unable to reveal it.

A. Information and Screening Devices

Ex ante. How does the state distinguish desired and undesired types? A simple but incomplete answer is that the government could demand that the applicant supply credentials that prove that she has the desired characteristics. The applicant could provide, for example, a diploma that shows that she has acquired desired high-technology skills abroad, or evidence of work experience that shows the same. If assimilability is a concern, the applicant could show that she has learned the native language or that she has family members in the country. In these ways the applicant would prove that she meets the state's criteria for the high-quality noncitizen.

To see the problem with this answer, one must distinguish between the type of the noncitizen and the evidence or proxy that might be correlated with type. Recall that type is private information; it will be revealed to the government only years later (perhaps never) when the noncitizen is productive, assimilated, and so on. Further imagine that type is a continuous variable, and any noncitizen can be placed somewhere along a line segment from zero (lowest type) to one (highest type). The government's first-order goals determine the place on this line segment that divides the type of noncitizen who should be admitted and the type of noncitizen who should not be admitted. For convenience, we will say that the dividing line is at 0.5. Thus, noncitizens who are of type 0.4 should not be admitted, while noncitizens of type 0.51 should be admitted. Because the noncitizen's type is private information, immigration authorities who examine her diploma and other pieces of evidence will be able only to infer the noncitizen's type within a confidence interval.

The breadth of the confidence interval depends on the degree to which the visible proxies are correlated with type. If some measure of education--years, grades, and so forth--were perfectly correlated with type, then the design problem would be easily solved. Immigration authorities would admit or reject on the basis of this measure of education. But immigration goals are more complex, and paper credentials are not necessarily accurate proxies for a noncitizen's type. Consider, for example, the factor of assimilability. Ability to speak English and the existence of family and friends in the United States may all be rough proxies for assimilability, but the best evidence may simply be the person's ability to live in the United States for an extended period of time. It may turn out that many non-English speakers with no family and friends in the United States can assimilate easily; and, further, that many people without much education or job training can be productive members of society. If so, even the simple proxies we have discussed--such as a diploma showing advanced studies or English-speaking ability--would exclude many high-quality noncitizens, yielding numerous false negatives.

The system we have described can be called a (pure) ex ante system. An ex ante system determines whether noncitizens will be allowed to enter and stay in the United States entirely on the basis of pre-entry credentials, credentials that are determined in advance and identified at the border. These credentials are a proxy for the person's type. Noncitizens who have these credentials are admitted; noncitizens who do not are excluded. The credentials typically include educational achievement, the possession of skills that are in local demand, wealth, health, and similar qualities. The American system, as we will discuss, is partially, but not fully, ex ante. It also has an important ex post component.

Ex post. To understand the ex post system of immigration, we begin with the pure version, which is rarely seen in the modern world, but which has some important precursors. The pure ex post system determines the noncitizen's type after she has entered and resided in the country for a period of time--what we will call the probationary period. Criteria for "admission"--that is, allowing the noncitizen to stay and eventually naturalize--are applied to the noncitizen's conduct after entry, rather than to her achievements or features prior to entry.

Suppose, for example, that the state admits anyone who applies for entry or, by lottery, more people than the state believes it will eventually want to stay. Noncitizens then remain for a probationary period of, say, five years. At the end of the period, the person is evaluated. If the noncitizen has steady, productive work, has learned English, has not committed any crimes, and otherwise has behaved herself, she is given the option to remain and naturalize; otherwise, she is deported.

The main advantage of this system over the ex ante system is that the state can use information that emerges from the noncitizen's residence in the United States (and, of course, pre-entry information as well). This information will often improve the correlation between the proxy and the type. A person with high-tech skills who is not able to find a job is most likely a low type; a person with no skills who is able to find a good job is likely a high type. Such people would be false positives and false negatives in the ex ante system, respectively, but they will be identified correctly and treated properly by the ex post system. Although the ex post system depends on a correlation between post-entry behavior and type (a person who commits a crime might actually be a high type), this correlation will often be stronger than the correlation between pre-entry behavior and type, especially in the case of unskilled workers, whose ex ante credentials are likely to be very similar to each other. Indeed, because the ex post system can use information about pre-entry characteristics as well, it cannot be less accurate than the pure ex ante system.

A closely related point is that the state will often have better information about its own needs at some time after, rather than before, a particular immigrant or group of immigrants has entered. So far we have assumed that the government knows at time 0 what its immigration needs will be at time 1. But a persistent problem for governments is that migrants needed for the economic boom at time 0 become unwanted competitors for jobs during the bust at time 1, or potential terrorists or subversives during a national security crisis at time 1. Ideally, a government would like to be able to screen immigrants on the basis of such information at time 1 as well as at time 0.

To see why this matters, imagine that the institutional design decision is made at year 0, and labor, security, or political conditions are revealed at year 1. Suppose further that the optimal type at year 0 is a function of conditions at year 1. In other words, the state does not know whether the floor for the optimal type will be 0.3 or 0.7 at year 1, and can only predict that each is equally likely. Under a system of only screening noncitizens when they first apply for admission, the state must predict conditions at year 1 and admit people on the basis of this prediction about the future optimal type. The state might, for example, set the floor at 0.5, which means that the stock of noncitizens at year 1 will be too large or too small. Under a delayed screening system, the state can wait until year 1 and then remove or retain noncitizens on the basis of an accurate assessment of the optimal type. (75) If the floor at year 1 is 0.3, then the state will deport everyone below 0.3; if the floor is 0.7, then the state will deport everyone below 0.7.

In sum, whether a "type" of immigrant is desirable depends both on the characteristics of a particular person and the needs of the government. Information about both will typically be more plentiful after entry than before. The advantage of the ex post system is that it permits the government to use this additional information prior to determining whether an immigrant will remain in the country and eventually become a citizen. (76)

B. Risk and Country-Specific Investment

While ex post screening provides more accurate information--both about an immigrant and about the state's needs--such screening does come at a cost. As we noted above, ex post screening entails delayed screening. Delayed screening creates a period of uncertainty for the immigrant. During the probationary period, the noncitizen knows that there is some chance that she will be deported on the basis of new information. This is a disadvantage of the ex post system: risk-averse noncitizens who do not know whether they will be retained may be reluctant to come to the country and make country-specific investments. (77)

Country-specific investments can be defined, by analogy to relationship-specific investments, (78) as investments whose return can be obtained only through continued residence in the country. For example, a noncitizen who learns American workplace norms makes an investment--the cost incurred in learning those norms--whose return she can obtain only though continued work in the United States. By contrast, a resident noncitizen who obtains transportable technical skills does not make a country-specific investment. Such investments could be labor-related, but they need not be: they could also be social, familial, and so on. The personal relationships that a noncitizen develops after entering the country, for example, are lost or seriously impaired if that noncitizen is deported. (79)

All else equal, it is generally better if the immigrant makes a country-specific investment than if she does not. If she does, she will generate more value; this value can be divided between her and the state. For example, if she learns workplace norms, she will obtain a better job, earn more money, and pay more taxes, all to the benefit both of the immigrant and the state. Thus, the state would like to encourage immigrants to make a country-specific investment; the problem is that the immigrant may fear that if she does, she will lose it, with the result that she will be worse off.

To understand the problem, consider the baseline case of the risk-neutral immigrant and a state with a perfect enforcement system. Both sides know that the state might decide, after entry, that the immigrant does not belong to the high type for reasons unrelated to the immigrant's post-entry conduct. Many states face volatile economic and security environments, which may suddenly and unexpectedly reduce the value of the continued presence of some or all noncitizens. An economic downturn may cause natives to pressure the government to remove foreigners who are perceived as competitors for scarce jobs; a security threat, such as a war, may lead to worry about the intentions of enemy aliens. What this means is that the optimal type can be a function of events that occur after entry. The optimal country-specific investment will thus be somewhat less than what it would be if the chance of the shock were zero. The risk-neutral immigrant will make this investment or something close to it, assuming that she obtains the return in the form of a salary or other benefit that is not excessively taxed.

In the real world, by contrast, immigrants are risk-averse, and immigration authorities make errors. In such a world, immigrants may fear that if they make costly country-specific investments, they will be deported in error. If they are risk-averse, they will be discouraged from optimal investment. For example, an immigrant may fear that she will be falsely convicted of a crime and deported, thus losing her country-specific investment.

A solution to this problem is for the government to commit not to rely on post-entry information when determining whether to deport the immigrant. This means that the noncitizen takes no risk of erroneous deportation once she is admitted on the basis of pre-entry information. With this security, she may make the optimal country-specific investment.

To sum up the argument so far, a central advantage of the ex post system is that it permits the government to use more information in determining whether a noncitizen is a desirable type; a central disadvantage is that it will excessively discourage risk-averse noncitizens from making country-specific investments and, thus, because the benefit of immigration must be lower, from immigrating in the first place.

A note on moral hazard. Our approach is very simple, and it could be complicated in many ways, but these complications do not change the basic story. One of our assumptions is that types the government deems desirable will make the optimal country-specific investment once admitted; but this assumption is not necessarily true. Consider instead the possibility that immigrants could obtain a positive return by refraining from making country-specific investments and "shirking" in additional ways--for example, by engaging in criminal behavior--even though they know they will be deported at the end of the probationary period. This could be true for desirable types as well as undesirable types. If this is the case, an ex ante system will be inadequate. All types will enter and then shirk after entry.

Desirable types, however, are often more deterrable than undesirable types. Suppose the government desires successful workers. Because its desired type of immigrant obtains a greater return from legal work, those immigrants' relative gains from, say, crime are lower. Thus, an ex post sanction on noncitizens who commit crimes or otherwise fail to make a country-specific investment will impose lower costs on the desirable type (who will invest instead of shirking) than on undesirable types (who will either incur the sanction or invest, but at higher cost to themselves). An appropriately calibrated ex post sanction will discourage entry from undesirable types but not from desirable types. (80) It is perfectly possible that the sanction would have to be greater than mere deportation; a criminal punishment might be necessary. The gains from such an approach would then have to be balanced against the costs--namely, that desirable types who are risk-averse might refrain from entering and investing lest they erroneously be classified as criminals, sanctioned, and deported. (81) If moral hazard by immigrants is a serious concern, the case for the ex post system improves, but it nevertheless remains possible that the ex ante system is superior because of the importance of risk-aversion and country-specific investment.

Moral hazard by the government is also a possibility. If immigrants undertake country-specific investments with the hope that doing so will improve their chances of making it through ex post screening, the government might benefit by ever-delaying that screening. Through delay, the government can obtain more benefits from the immigrants without bearing the costs of providing them with more secure status. If this sort of shirking by the government is a serious concern, the case for the ex ante system improves, assuming that the ex ante system can be constitutionalized or otherwise used as a constraint on the government.

C. Other Factors

There are other advantages and disadvantages of the different screening approaches. Two should be mentioned because of their prominence in immigration debates.

Enforcement costs. The ex ante system is more desirable if the government can effectively patrol the border. If the government cannot, then it can remove noncitizens only by patrolling the interior and deporting those who are discovered. In the United States, this typically happens when a noncitizen commits a crime and is captured by local police. (82) As long as removal occurs after entry, then the government might as well use ex post as well as ex ante information.

Controlling the border can be expensive, but whether it is expensive or not depends on numerous factors--including the length of the border, the ruggedness of the terrain, and the difference between quality of life on either side of the border. The last factor is worth emphasizing. The Mexican and Canadian borders are both very long, but Canadians are less interested in migrating to the United States than Mexicans are, because Canada is much wealthier than Mexico. Topography also matters: island nations like Britain and Japan can rely more heavily on the ex ante system than can the United States, because it is more costly for noncitizens to cross large bodies of water than to step across an invisible line.

While difficulty in enforcing the border makes ex post screening relatively more attractive, it also reduces the absolute effectiveness of screening altogether. After all, immigrants who are deported after ex post screening may simply sneak back across the border. For that reason, states without perfect border enforcement often attempt to deter deported immigrants from reentering by imposing criminal sanctions. (83) Relatedly, governments can and do supplement ex post sanctioning with border exclusion: people who have been deported generally are denied readmission. (84)

Residency-related costs. The ex post system also has its own distinctive disadvantage, which is that the noncitizen's residence in the state during the probationary period can bring with it costs that the ex ante system avoids. There are several potential costs. First, the noncitizen may commit crimes, and no realistic ex post sanctions, including criminal penalties, could reduce the rate of criminal activity to zero. (85) Applicants excluded at the border cannot commit crimes within the territory. So a nation with inadequate local policing or serious crime problems might prefer the ex ante system in order to keep potential criminals off its territory. Second, under the ex post system noncitizens may develop local ties during the probationary period. These ties can make deportation seem unfair, or harmful to citizens who are friends or relatives. The ex ante system avoids this cost as well.

D. A Comparison

Our theory suggests that a central advantage of ex post screening compared to ex ante screening is that it increases accuracy for the government by making available more information, and that a central disadvantage is that it reduces immigrants' incentive to make country-specific investments, and so possibly their incentive to apply for admission in the first place. More specifically, the ex post system becomes more desirable when:

(1) The correlation between pre-entry characteristics and type becomes weaker relative to the correlation between post-entry behavior and type (and type is private information or unknown to the immigrant himself). An ex post system is not necessary when type is racial, for example, because race can usually be determined on the basis of physical features, which can be inspected by the entry officer. The ex post system becomes more advantageous as assimilation becomes more important, because an immigrant's ability to adjust to living in a new society is much more difficult than race to determine on the basis of ex ante criteria.

(2) Economic and security conditions become more unpredictable. The government benefits from retaining the right to deport people (who might otherwise be satisfactory) on the basis of information about the government's needs that arises after entry. And given that the deportation judgment occurs after entry, post-entry activities can be used as an additional source of information about the quality of the noncitizen.

(3) Country-specific investment becomes less important. If country-specific investment adds a great deal to the value of the noncitizen for the host state, then the state, to encourage such investment, must promise the noncitizen not to deport him except under special circumstances. This reduces the value of deportation proceedings, which makes the ex ante determination relatively more valuable. (Relatedly, the ex post system is more attractive when immigrants are less risk-averse.) However, when country-specific investment is unimportant and post-entry moral hazard is a serious concern, then the case for the ex post system improves.

(4) The cost of detecting and deporting people on American territory declines relative to the cost of excluding people at the border. As border control becomes more costly, post-entry deportation becomes a more cost effective means of control, in which case post-entry conduct can be used to evaluate the noncitizen. If exclusion at the border is cheap, then this additional information might be forgone for the cost savings.

(5) Residency-related costs decline. When resident aliens are unlikely to commit crimes or cause other harms during the probationary period, the probationary period does not seriously create risks for citizens. This could be the case, for example, if areas in which noncitizens reside are effectively patrolled by local police, or if immigrants are more law-abiding than citizens. Similarly, when resident aliens are unlikely to form bonds with local citizens, the welfare loss and political costs of deportation after the probationary period will be relatively low. This might be more likely to be the case when noncitizens come from countries for which there are no existing immigrant communities into which they would otherwise be integrated.

As noted above, optimal immigration design will usually involve both ex ante and ex post controls. A state applies general criteria ex ante and excludes those who do not satisfy them; then the state evaluates the noncitizen with additional criteria ex post, criteria that are based on her performance after arrival. Indeed, the state might find it useful to have two or more "tracks" that different aliens are placed on. This possibility is suggested by the economics of contracts literature, which argues that employers that hire workers under conditions of asymmetric information should offer a menu of contracts. (86)

Consider, for example, an immigration system that offers three kinds of visas. First, people who can prove that they have skills that U.S. employers greatly demand are given the right to enter, plus the right not to be removed unless they commit a serious crime or become a public charge. Second, people who cannot prove that they have desired skills are given the right to enter (perhaps on a lottery basis), plus they can be removed for any reason or no reason. Third, people who cannot prove that they have desired skills are given the right to enter (perhaps on a lottery basis), plus they must make a payment or post a bond, which they forfeit if they are removed; they can be removed only if they commit serious crimes or become public charges.

Credentialed high types can enter with the first visa; uncredentialed but hard-working high types can enter with the second and third visas, in the latter case by borrowing against human capital. Low types cannot enter with the first visa because they do not have credentials and with the third visa because they cannot borrow against their human capital. They will also not want to enter under the second system as they will quickly shirk, be caught, and be removed, so that they gain less than the cost of entry. This is how the system would work in theory; in practice, of course, some uncredentialed high types would be screened out because they cannot borrow against their human capital, and some low types would obtain entry with the second visa and escape detection and removal. But this "menu" system--with an "easy-in-easy-out" track and a "difficult-in-difficult-out" track, and perhaps variations of each--should be superior to a system that treats everyone the same.

Whether or not a menu is used, the menu metaphor usefully highlights the way that a well-functioning immigration system should lead potential immigrants to self-select. With enough information about American immigration rules, potential migrants who satisfy the criteria will enter while those who do not will not enter, sparing the government costly enforcement resources. This is true regardless of whether there are multiple immigration tracks or just one.

E. Positive Implications

Several predictions emerge from our analysis. It is not our purpose here to test our predictions, which would be a large undertaking; instead, we provide each prediction and an illustration of the type of evidence that would confirm or undermine it. One should keep in mind that an adequate empirical test would need to control for all variables, so we do not mean to suggest that these predictions can be easily mapped to American or foreign history and practice. (87) All of our predictions assume that the government seeks to maximize the welfare of its own citizens, and not potential immigrants or other foreigners. (88) If this assumption is wrong because of political dynamics, constitutional restrictions, or some other reason, then the predictions will be false. We will return to these possibilities in Part III.

(1) As the value of a noncitizen for a country increases with the noncitizen's country-specific investment, the noncitizen will be granted greater protections against removal.

Consider, for example, that Japan and the United States compete for highly skilled migrants. (89) To function effectively in these countries, the migrant must learn the language. A migrant who learns English obtains a transportable skill that can be used in other countries--not just Canada, Britain, and Australia, but also many European countries where English is the de facto language of business. A migrant who learns Japanese obtains a much less transportable skill. Thus, learning Japanese is a particularly country-specific investment; learning English is not nearly as much of one. Thus, we predict that Japan and most other countries would provide greater ex post protections to migrants lured over to provide needed skills than would the United States or another English-speaking country.

(2) As the value of a noncitizen for a country is increasingly (negatively) correlated with exogenous factors such as security threats and economic downturns, the noncitizen will be granted fewer protections against removal.

Wealthy, populous countries are buffered against security and economic shocks to a much greater extent than poor and thinly populated countries are. Given an identical shock, a large country would gain less from removing noncitizens (as labor competitors, or threats) than a small country would. Thus, we predict that larger and wealthier countries provide greater ex post protections to migrants than smaller and poorer countries do, holding constant the proportion of migrants in the population. For example, a large country with few migrants like Japan would provide greater ex post protections than small countries with many migrants like the Persian Gulf States. (90)

(3) As ex ante screening technology improves relative to ex post evaluation technology, the immigration system is more likely to be ex ante than ex post, or to include more ex ante elements and fewer ex post elements.

Consider two countries, one of which wants to import only high-skilled labor and the other of which has a general labor shortage, including unskilled workers. It seems plausible that a country can more easily evaluate highly skilled labor on the basis of pre-entry criteria than low-skilled labor--where it is hard to distinguish people on the basis of their credentials. (91) If so, we would predict that nations that permit a small amount of highly skilled immigration relative to their labor force will rely more heavily on ex ante criteria, and that nations that permit a large amount of relatively unskilled immigration relative to their labor force will rely more heavily on ex post criteria. Japan is in the former category; the United States is in the latter category. A similar point can be made about the relative cost of guarding borders and patrolling the interior.

III. THE SECOND-ORDER STRUCTURE OF AMERICAN IMMIGRATION LAW

So far we have established three points: (1) that issues of second-order design, often overlooked, are critical to making sense of the structure of immigration law; (2) that one central design decision for any state is what mix of ex ante and ex post screening mechanisms it will use to shape its immigrant population; and (3) that, roughly, ex post systems provide more accurate and flexible screening than ex ante systems while discouraging country-specific investment by risk-averse noncitizens. In this Part, we use these ideas to evaluate prominent features of American immigration law and proposals for reform.

A. Ex Ante vs. Ex Post

1. The shift from exclusion to deportation

The government can pursue its first-order policy goals by excluding undesirable immigrants when they attempt to enter the country, or instead by deporting those immigrants at some point after their entry. An important question about the structure of American immigration law is why the United States chooses to use the mix of exclusion and deportation that it does. Part II suggests a partial answer to this question: the choice between exclusion and deportation depends in part on the comparative effectiveness of ex ante and ex post screening to accomplish different immigration policy objectives. (92)

To see this, consider the historical trajectory of America's reliance on deportation and exclusion. There has been a steady shift over time toward increased reliance on deportation and, consequently, on the ex post screening of immigrants. When the federal government first began to restrict immigration in the 1870s and 1880s, it relied almost exclusively on ex ante screening. This early immigration legislation was principally designed to exclude Chinese immigrants. The first statute, the 1875 Page Act, targeted Chinese women, requiring them to obtain certificates of immigration showing that they were not entering the United States "for lewd and immoral purposes." (93) The Page Act was followed by the Chinese Exclusion Act of 1882, which prohibited all Chinese laborers from entering the country. (94) Both statutes established purely ex ante screening mechanisms. A potential entrant was either a Chinese laborer or prostitute, or not. (95) If she was, she was excluded; if not, she was admitted.

In contrast, federal ex post mechanisms were extremely rare early in U.S. history. The sole exception was the short-lived Alien Act passed in 1789, which permitted the President to deport aliens suspected of subversive activity. (96) The Page Act contained no deportation provisions. (97) And although the Chinese Exclusion Act did permit the deportation of "any Chinese person found unlawfully within the United States[,]" (98) the deportation was still formally based on pre-entry information--the fact that the noncitizen was of Chinese national origin and did not already reside in the United States at the time the Chinese Exclusion Act was enacted--and thus on ex ante grounds. (99)

This nearly exclusive ex ante focus remained until 1907, when Congress for the first time clearly provided for the deportation of an immigrant solely on the basis of post-entry conduct. (100) The statute made deportable any noncitizen woman who engaged in prostitution or was found living in a "house of prostitution" within three years after entering the United States. (101) Congress expanded its ex post screening criteria over the following years, adding criminal convictions and advocacy of anarchy to the types of post-entry conduct that could get an immigrant deported. (102) In one important respect, however, the reliance on ex post information was restricted: nearly all of these provisions shared the feature of being time-limited. They made a noncitizen deportable only if she engaged in proscribed conduct within a certain number of years after entering the country. In the 1907 Act, for example, an immigrant who engaged in prostitution four years after entering the country did not become deportable. (103)

Over the last century, Congress steadily expanded the ex post screening system by augmenting the list of post-entry conduct that would make a noncitizen deportable. In 1922, Congress for the first time included certain drug convictions. (104) And the enactment of the Immigration and Nationality Act in 1952 broadened the definition of subversives subject to deportation and enlarged a number of other deportability grounds as well. (105) This expansion has accelerated in the last twenty years, as Congress has added additional grounds of deportability--particularly criminal grounds--in a series of immigration reform bills. (106) The growth in ex post screening has been augmented by two other important changes: first, Congress has extended the screening period, eliminating the statutes of limitation for most grounds of deportability; (107) second, Congress has made the screening system more categorical, eliminating many avenues of relief from deportation that in earlier periods were available to noncitizens who engaged in deportable conduct. (108) Today, more so than at any time in the past, immigrants must prove themselves by refraining from post-entry activities that run afoul of the statutory restrictions.

What might explain the ever-expanding reliance on ex post screening over time in American immigration law? One possibility is that first-order preferences have changed: American immigration law may have begun to rely more on ex post screening as it has become less racist. To the extent that past immigration policy was interested principally in racial sorting, there was little need for ex post selection mechanisms. Ex ante mechanisms were adequate because the type of immigrant that the government was trying to select could be identified on the basis of the immigration enforcement officer's determination of the immigrant's race, which was observable from the applicant's physical features (such as skin color). As immigration law became less openly racist over time, however, it would likely have become more difficult to identify "desirable" immigrants on the basis of information available at the time they entered the country. This slow shift in the first-order preferences of immigration regulators would have made ex ante screening less effective. And when ex ante screening mechanisms become less effective, ex post screening becomes comparatively more attractive.

Alternatively we might find the explanation in a shift in screening constraints rather than a shift of first-order preferences. Here there are a few possibilities. First, suppose that early policymakers were not so much racist as concerned about assimilation, and that they assumed that racial minorities could be assimilated only with difficulty. (109) The policymakers need not have believed that all minorities are unassimilable, just that, on average, minorities assimilate with more difficulty than immigrants of Anglo or perhaps German stock. This was a standard argument made by policymakers in that era. (110) But as America became more racially and ethnically diverse, racial and ethnic homogeneity no longer served as a reliable proxy for assimilability. Hence race (or close proxies like national origin) was dropped as an ex ante criterion. In its absence, ex post mechanisms became increasingly important. Though more costly than observing physical features, the government might have believed that evaluating an immigrant's post-entry conduct--criminal activity, public charge status, and the like--would provide better information about her assimilability. Thus, one argument for the shift from ex ante to ex post--consistent with our theory in Part II--is that as America became more diverse, the correlation between type (assimilability) and the ex ante proxy (racial/ethnic homogeneity) declined, so that the cost advantages of the ex ante system became less significant.

Second, changes in the difficulty of policing the border may have been an important variable. As Mae Ngai has noted, "Before the 1920s the Immigration Service paid little attention to the nation's land borders because the overwhelming majority of immigrants entering the United States landed at Ellis Island and other seaports." (111) Under these conditions it was comparatively easy to screen potential immigrants at the point of entry. By the 1920s, however, growing Mexican migration, the passage of the quota laws, and changing political conditions brought new pressure to police immigration across the Mexican and Canadian borders. (112) And as migration across land borders became a bigger and bigger part of the regulatory picture, the cost of exclusion increased. It was simply much more difficult to police the nation's long land borders than it was to police the seaports. (113) Because ex post screening becomes more attractive as the cost of deportation declines relative to the cost of exclusion, these structural changes in migration flows may also have made ex post regulation more attractive over time. (114)

2. The gap between exclusion and deportation grounds

Similar logic might help explain why modern immigration law sometimes treats pre-entry and post-entry behavior so differently. As we noted in Part I, many of the grounds of inadmissibility and deportability--such as the national security grounds--are essentially identical. (115) But not all are. Consider, for example, the way that current immigration law treats criminal behavior classified as an "aggravated felony." Congress in 1988 made deportable any noncitizen with a conviction for an "aggravated felony"--a term that the INA initially defined to cover serious drug trafficking offenses. (116) Since then the definition has been repeatedly expanded by Congress. (117) Today it sweeps in a broad swath of criminal conduct, including minor convictions--even some misdemeanors--that make the statutory label something of a misnomer. (118) Commentators often criticize the aggravated felony provision on the ground that it is too harsh. (119) Less often noticed is the fact that a conviction for an aggravated felony, which makes a lawful permanent resident deportable and ineligible for nearly any relief from deportation (regardless of how many years she has resided lawfully in the United States), (120) does not constitute a ground of inadmissibility. (121) In other words, a conviction that will not lead to the exclusion of a first-time arriving alien can lead to the deportation of a longterm permanent resident. (122)

Why would Congress make the ex post screening system stricter than the ex ante system in this fashion? First-order preferences are not the answer: no one thinks that a person who commits an aggravated felony in the United States is invariably worse than a person who commits an aggravated felony in a foreign country. However, second-order considerations might answer our question. A conviction in the United States might be better evidence of a person's type just because American authorities are familiar with and trust American criminal law and procedures, while being largely unfamiliar with and suspicious of the substance or procedures of foreign criminal law. (123)

Such a conviction also might be better evidence of the person's type because committing a crime in the United States might be better evidence of qualities the government deems undesirable--like unassimilability--than the commission of the identical crime in another country. For example, a person might commit minor crimes in a foreign country because the probability of detection and conviction is low and the punishment is trivial, or because this type of illegality is widespread and considered socially acceptable. The same crime, however, might have a different meaning in the United States, and, among immigrants, it may be that mainly those who fail to assimilate commit it. If so, commission of the crime might be a good ex post proxy of unassimilability but not a good ex ante proxy of unassimilability. (124)

In practice, the possibility that the aggravated felony category operates in this fashion may be implausible, given some of the minor conduct that it sweeps within its ambit. Our point here is not to defend this particular deportability provision but to explain why institutional concerns might cause a state to use different grounds for admission and deportation.

3. Constitutional constraints on second-order design

Our discussion to this point has emphasized how changes in screening constraints or first-order preferences might account for the mix of ex ante and ex post screening in American immigration law and its change over time. But constitutional limitations may prevent the government from choosing its preferred second-order design.

In Part I, we mentioned several such constraints, which collectively restrict the government's ability to regulate noncitizens who have lawfully entered the country while providing the government much greater leeway to restrict entry (and, to a certain extent, to deport those who enter unlawfully). The government may often be able to use summary procedures in the latter cases but not in the former, and equal protection norms apply to a much greater extent to territorial regulation than to regulation of admission and deportation. (125) Another constitutional constraint with similar effect is the guarantee of birthright citizenship in the Fourteenth Amendment. (126) By conferring citizenship on all children born within the territory of the United States, that amendment prevents the government from deporting children born to immigrants after they enter the country--regardless of the status of the immigrant parents. This deprives the government of some ex post screening options. If the government adopts a large-scale guest worker program, for example, it will not always be able to deport a guest worker's entire family if it decides to screen that worker out at the end of the probationary period. On the margin, the guarantee of birthright citizenship, like

the different requirements for due process and equal protection, makes the ex ante system more flexible and attractive for the government than the ex post system.

All of this implies, everything else being equal, that constitutional constraints often make it more costly for the government to rely on the ex post approach. As we noted above, the trend has in fact been in the direction of ex post despite the constitutional barriers. But constitutional restrictions may well cause the government to rely more on an ex ante system than it would like to, or to structure ex post screening in a way that attempts to evade constitutional restrictions. As we will explain below, the most important part of the modern ex post system--the "illegal immigration system"--might fit this pattern: it exists as a result of discretionary under-enforcement, which has the result of avoiding some constitutional constraints that would otherwise apply. (127) Of course, such strategic substitution by the government can undermine the benefits (in terms of constitutional values) of constitutional immigration law. Discrimination prohibited at the ex post phase might persist in ex ante form. If, for example, deportation procedures cannot be biased against Muslims, then the government may bias admission procedures against Muslims or reduce the national origin quotas from Muslim countries. It does not necessarily follow that the current constitutional restrictions should be abolished--perhaps instead they should be extended to the ex ante phase, though this could create additional difficulties--but it does make it more difficult to assess the benefits of existing constitutional restrictions.

B. The Ex Post System: Illegal Immigration and Guest Workers

1. Reconceptualizing illegal immigration

As we explained above, the legal immigration system is paralleled by a shadow illegal system. Through this illegal system, as many immigrants flow into the United States each year as through authorized immigration channels. As a result, it is estimated that more than eleven million unauthorized immigrants currently reside in the United States. (128) Both the public debate and existing scholarship typically view illegal immigration as an enforcement problem that needs to be solved. The high level of illegal immigration is seen as reflecting the government's failure to enforce the existing immigration rules. This premise has led to two lines of analysis. The first is positive: commentators attempt to determine what factors (institutional, political, etc.) lead to the enforcement failure. (129) The other is normative: commentators criticize the existing rules--often on the ground that the high levels of undocumented migration make clear that the economy needs the influx of low-skilled workers--and argue that the government should liberalize its admissions system to provide a legal pathway for these workers to enter. (130)

Our theoretical framework suggests a different way of understanding the illegal immigration system. That system can be seen as a de facto ex post screening system operated under the guise of an ex ante system.

This point has been obscured by the formal structure of law relating to undocumented immigration. As a formal matter, noncitizens who enter the country without authorization are deportable on that basis alone. (131) And when the government removes a noncitizen on the ground that she entered illegally, the government appears to be simply enforcing the ex ante screening rule after the fact of entry itself. The removal is based on information that existed at the time the person entered--specifically, it is based on the fact that the person entered without being admitted. (132)

Despite this initial impression, the present "illegal immigration system" operates substantially as an ex post screening mechanism. This becomes clear if one focuses on the enforcement decisions made by immigration officials rather than on the formal legal grounds on which undocumented immigrants are eventually deported. In theory, immigration authorities should pursue all illegal entrants. In practice, the Bureau of Immigration and Customs Enforcement (one of the successors to the INS) has for some time focused its enforcement efforts on those immigrants who are arrested for having committed non-immigration crimes. (133) In this way, the illegal immigration system operates substantially as a de facto ex post screening mechanism. On the front end, under-enforcement at the border permits large numbers of immigrants to enter the country without ex ante screening. Then these immigrants are screened at some later date on the basis of their post-entry conduct: those who avoid contact with the criminal justice system are generally permitted to remain in the country, while those who have a run-in with the police are often removed. (134) Contact with the criminal justice system, then, becomes the de facto proxy for type. (135) The implicit theory is that, at least for the pool of unskilled labor, authorities can better screen out undesired types by waiting for noncitizens to commit crimes and expelling them than by using some other proxy at the border ex ante.

Why might the government prefer an ex post system to an ex ante system in this context? To the extent that the illegal immigration system operates in significant part to supply low-skilled workers to domestic producers, as many have argued, ex post screening might be more effective than ex ante screening. It is difficult to select desirable low-skilled workers on the basis of pre-entry information. There are few objective criteria like education or prior work history that would be reliable indicators of the ability of a low-skilled immigrant worker to be a productive employee in the United States. By contrast, an applicant's post-entry employment record is highly relevant, often fine-grained information. This helps explain the Senate immigration reform bill's reliance on post-entry employment history as a central part of the eligibility screening process for amnesty. (136) And it suggests that a largely ex ante screening system--like the point system Canada uses to select high-skilled immigrants--would probably not work effectively for the immigrant pool currently affected by the illegal immigration system. (137)

Thus, the immigration agencies have structured their enforcement priorities in a way that transforms a central part of American immigration policy from a de jure ex ante screening system into a de facto ex post screening system. Moreover, if we draw back the lens and look at the illegal immigration system across time, it becomes clear that the system is ex post in a quasi de jure sense as well. The United States has periodically regularized the status of many of the undocumented noncitizens living in the country through amnesties or other mechanisms. Such status regularizations, which can legalize the noncitizens' permanent residence and provide a pathway to citizenship, may become a periodic feature of the contemporary American immigration landscape, as it already is in other countries. (138) In 1986, Congress passed legislation that legalized over two million undocumented immigrants. (139) And right now Congress is debating legislation that would legalize a substantial number of the undocumented immigrants who arrived after the last round of regularization. (140) These status regularizations make the existing illegal immigration system look like not just a de facto, but a de jure ex post system. Rather than just permitting undocumented residents to remain in the country so long as they do not engage in post-entry behavior that leads to their being screened out, the government eventually accords legal status and a path to full membership to those illegal aliens who satisfy the ex post screening criteria embodied in the amnesty program.

Consider, for example, the legalization program included in the immigration reform bill passed by the Senate in 2006. (141) The program does not permit all undocumented noncitizens to legalize their status and become lawful permanent residents. Instead, the program limits eligibility in a way that suggests that it is being used by the government as an ex post screen. Immigrants who have lived continuously in the United States for more than five years meet the baseline qualifications for the program if they have worked for at least three of those years and learned sufficient English to pass a language and civics test. A path to legal permanent residency is open to those who continue to work for six additional years after enactment of the bill, pay back taxes, and register with the Selective Service. (142) For immigrants who have lived continuously in the United States for more than two but fewer than five years, the requirements are more stringent: they must have been employed before enactment of the bill and have worked continuously during their time in the United States. (143) (Here, steady employment appears to serve as a partial substitute for longer residence in the country.) Individuals in both groups are ineligible for legalization if they have been convicted of a felony or three or more misdemeanors. (144)

The legalization program's requirements are consistent with the idea that those undocumented immigrants who qualify have demonstrated, through their conduct after entering the United States, that they should be admitted to permanent residency and eventually citizenship status in the United States. This explains why the government might choose to rely in large part on measures like post-entry success in the labor market and a clean criminal record to screen out some immigrants from the available pool. (145) The government might conclude that these ex post measures serve as superior substitutes for ex ante screening criteria.

Understanding the ex post structure of the illegal immigration system highlights the multi-"track" structure of American immigration law. The United States desires both high- and low-skilled labor, but a single immigration track might work poorly for both. So the United States offers two tracks. First is the legal labor immigration track reserved principally for highly credentialed immigrants. For this part of the market, the ex ante system does much of the work. A second track--the illegal immigration system we describe above--is available to less-credentialed, low-skilled workers. For this part of the market, the ex post system does much of the work. This description is an oversimplification: not all low-skilled workers are forced to immigrate illegally, for example, and demand for high-skilled workers has often led to their illegal immigration as well. Still, the rough two-track structure of the U.S. system fits well with the information theory we set out in Part II. (146)

2. The guest worker alternative

Once we see that the illegal immigration system and Congress's periodic amnesties create a large-scale ex post screening process for immigrants and need not simply reflect a failure to enforce existing immigration law, new questions emerge. We might ask why the government does not just formalize the system by replacing the illegal immigration system (including amnesties) with a legal guest worker program. The government could relax entry restrictions but then formalize the ex post screening elements that are today embodied in agency decisions about enforcement priorities and congressional decisions about periodic amnesties.

Congress is actually considering doing something like this right now. It has been debating the possibility of creating a new guest worker program, and the existing legislative proposals would formalize some elements of the present illegal system. The plan embodied in the Senate's immigration reform bill would dramatically increase the number of low-skilled workers admitted to the United States. (147) These immigrants would be authorized to enter the United States for an initial period of three years, and each immigrant could seek one three-year extension of this initial period. (148) During the period of admission, immigrants would be required to engage in approved work such as agricultural labor; if they become unemployed, they would be required to leave the country. (149) Significantly, under the Senate's proposal these immigrants would also have the opportunity to apply for lawful permanent residency at the end of their term. (150)

As with the illegal system, the main advantages for the government of a guest worker program are twofold: by allowing the government to admit people to fill temporary jobs without committing itself to keeping them, the program offers the informational and flexibility advantages of ex post screening. Such increased reliance on ex post screening is most appropriate for people whose labor value does not depend heavily on country-specific investments. There may, of course, be substantial countervailing costs associated with the system. (151) Our goal here is not to resolve the debates about the merits of introducing a new guest worker program in the United States--though our above discussion makes clear that the desirability of such a program must be measured against the current system of tolerated illegal immigration. Instead, we want to suggest one potential reason why Congress might prefer the existing illegal immigration system to a large-scale guest worker program: constitutional law.

As we discussed above, constitutional law will sometimes preclude an immigration regulator from adopting her preferred mix of ex ante and ex post screening mechanisms. (152) Current constitutional doctrine might discourage Congress from converting the existing illegal immigration system into a legal one. This conversion would have the effect of increasing the constitutional protections available to noncitizens in the system. Under the present regime these noncitizens often enter without authorization and are, consequently, never lawfully admitted to United States. Were the system formalized through expanded entry provisions and greater formal reliance on ex post screening, the same noncitizens would no longer lack legal status while they were in the United States. Most importantly, their greater legal status would likely entitle them to more constitutional protection than unauthorized immigrants. On the present Administration's view, for example, their legal status would mean the difference between having some constitutional rights with respect to removal instead of none. (153)

If immigrants who are presently part of the illegal system had greater constitutional rights, the government would have less flexibility to pick whatever ex post screening mechanisms it wished. For example, noncitizens acquire greater procedural protections as their constitutional status rises. (154) Accordingly, noncitizens who are in the country pursuant to lawful admission have procedural due process protections that preclude the government from deporting them on the basis of extremely summary procedures. (155)

From the government's perspective, therefore, there is a significant advantage to the present illegal immigration system: it skirts constitutional restrictions that might otherwise apply and, as a result, gives the government greater flexibility in deciding how to deal with the noncitizen population that currently immigrates to the United States through the illegal system. The government can use this flexibility to lower the cost of the screening system or to make it easier to achieve other policy goals--like expelling security risks as new threats surface or competitors for scarce jobs during economic downturns. As an example of this flexibility, consider INA section 235(b), the expedited removal provision that we described in Part I. (156) Under this provision, the immigration agencies are authorized to use summary procedures to remove noncitizens who entered the United States without being admitted, unless those noncitizens have been in the country for more than two years. (157) In practice this means that the Attorney General can authorize a single immigration enforcement official to order the removal of such a noncitizen, without a hearing or any other sort of process. (158) Were currently unauthorized immigrants instead admitted through an expanded legal system, it would almost certainly be unconstitutional to remove those immigrants by using the summary enforcement mechanism authorized in section 235(b). (159) The fact that the current illegal system does not confer any legal immigration status on those immigrants, however, helps preserve the possibility that the government can rely on summary mechanisms that would otherwise be constitutionally proscribed. (160)

C. The Ex Ante System: Family Migration and Alternatives

Family Migration. While the preceding Subparts have focused on ex post elements, American immigration law also contains substantial ex ante components. Like those of most liberal democracies, the formal ex ante screening system in the United States favors highly skilled workers who have skills that are in short supply in the native population. But to a much greater extent than other liberal democracies like Canada, Australia, and the United Kingdom, the United States also strongly favors immigration by persons who have family relationships with American citizens. (161) In this Subpart, we use our framework to briefly evaluate this approach.

Why might the United States place such a heavy emphasis on family relationships? One standard hypothesis is simply that the bias in favor of family reflects a first-order policy judgment: Americans care more about family intactness than people living in other countries do. The problem with this hypothesis is that it seems implausible that Americans care about family intactness more than other people do. To be sure, Congress has long emphasized the importance of reunifying families. When Congress established the family immigration system in the Immigration and Nationality Act of 1952, (162) for example, the congressional reports accompanying the legislation emphasized that Congress was consciously providing for "the preferential treatment of close relatives of United States citizens and alien residents consistent with the well-established policy of maintaining the family unit wherever possible." (163) Similar statements about the importance of family reunification accompanied the passage of the 1965 amendments to the INA. (164) These amendments strengthened the focus on family-based immigration at the same time that other countries were taking a different approach. Canada in the late 1960s established a point system for potential migrants designed to attract high-quality workers. A potential migrant could earn only a tiny fraction of the required points by having family already residing in Canada. (165) Still, it would be surprising if these divergent approaches were simply the product of different first-order preferences--especially in light of the fact that many other cultures have stronger family norms than Americans do.

Regardless of whether or not the first-order hypothesis is correct, our aim is to focus attention on second-order possibilities: the bias in favor of family might also be explained as an institutional design strategy in which family relationship serves as a proxy for a first-order immigration policy goal. Assimilability is one obvious potential goal; it may be that family members are more easily assimilated than other types of immigrants. Another possible goal might be the desire to achieve racial homogeneity. There is some evidence that the family focus was designed in part to achieve one or both of these goals. In support of the racial homogeneity hypothesis, the rules favoring family intactness in part replaced the national origin quota system. The national origin system was widely regarded as racist because it restricted immigration by Asians and southern and eastern Europeans (as well as by Africans and others who historically suffered from racial discrimination), (166) and historical evidence suggests that legislators who wanted to retain the national origin system settled for the family relation rules because they were expected to preserve the racial balance of the population (given that family members usually belong to the same race). (167)

However, as we have noted above, the national origin system can also be justified on the basis of concerns about assimilation rather than racism-though, to be sure, the two policies overlap. Supporters of the national origin system feared that noncitizens with different cultures and political views would obtain disproportionate political power and change the character of the republic. Whether or not this fear was justified, one can similarly understand the family relations system as based on the view that family members can be more easily assimilated than can others. (168) To be sure, this second-order hypothesis raises the question of why Americans would be more concerned with assimilation than people in other countries would be. The answer is far from clear, but it could be that the United States permits greater immigration because of persistent historical labor shortages, and so the failure to assimilate would be a larger social problem in the United States than in other countries. (169)

Alternative ex ante mechanisms. Economists have often proposed other, more decentralized ex ante screening mechanisms for immigration. Julian Simon and Gary Becker have proposed that visas for lawful permanent residence be auctioned off. (170) In terms of our framework, an auction would serve as a screen. The good types would pay the most, and would be allowed in; the bad types would not be able to pay as much, and so would be excluded. (171) Such a system would need to be accompanied by strong enforcement to prevent illegal entry.

A central problem with these approaches is that they overlook first-order goals that are of central importance to the government--such as assimilation. The auction approach is desirable only if willingness-to-pay is correlated with the types that the government deems desirable. This might be true at the very highest end, but otherwise we are doubtful. One problem is that people often have trouble borrowing against their human capital, so people with great skills may not be able to afford a visa. Although employers could finance visas in some cases, their incentives are not optimal, as they do not capture the full value of the immigrant's contribution when the immigrant's skills are not specific to one firm. (172) More important, the noncitizens' work skills are not the only relevant factor; many other factors shape a state's first-order immigration preferences. States often care considerably about assimilability, including an immigrant's commitment to liberal, democratic institutions, for which employers will have little concern. Yet from the state's perspective, highly assimilable people with low skills may be preferred to unassimilable people with high skills. An auction would screen out the former while allowing in the latter. (173)

Moreover, these alternative ex ante screening proposals overlook entirely the advantages of ex post evaluation. This is of considerable importance because the noncitizen applicant may not know whether she is assimilable prior to entry; she may be as ignorant about her character in this respect as the immigration authorities. If so, she will not know how much to pay for her visa.

CONCLUSION

Immigration scholarship has long overlooked important second-order issues about the institutional design of immigration systems. Our aim has been to highlight these second-order questions and to begin developing a theoretical framework for addressing them. We draw an analogy between the immigration system and the screening process by which employers choose employees, and argue that the economics of information, which has productively been used to analyze employment contracts, should be used to help understand immigration design as well. A central design question for states is whether screening should occur ex ante or ex post. Ex post systems are more accurate because they exploit more information--both about potential immigrants and about the government's preferences and needs. But they discourage country-specific investment, increase the risk that citizens will be harmed when deportation severs social and familial ties or when noncitizens commit crimes on home territory, and may be more costly to administer as well. We argue that many immigration controversies---concerning the use of exclusion versus deportation, criteria for admission, guest worker programs, and the severity of sanctions for illegal immigration--are wrongly thought to be exclusively controversies about first-order goals reflecting moral commitments. They are often controversies about institutional design and, as such, are amenable to institutional and empirical investigation.

(1.) See DAVID COLE, ENEMY ALIENS: DOUBLE STANDARDS AND CONSTITUTIONAL FREEDOMS IN THE WAR ON TERRORISM 22-35 (2003).

(2.) See Adam B. Cox, Citizenship, Standing, and Immigration Law, 92 CAL. L. REV. 373, 413-14 (2004). These initiatives included a special registration program that required noncitizens from such countries to report to the INS to be fingerprinted, photographed, and interviewed; the Absconder Initiative, which targeted noncitizens from "al Qaeda nations" for removal; and Operation Liberty Shield, which subjected asylum applicants from many predominantly Muslim countries to mandatory detention. Id.

(3.) See, e.g., PETER BRIMELOW, ALIEN NATION: COMMON SENSE ABOUT AMERICA'S IMMIGRATION DISASTER (1995); SAMUEL P. HUNTINGTON, WHO ARE WE?: THE CHALLENGES TO AMERICA'S NATIONAL IDENTITY (2004); Tom Brune, Laying Down the Law, NEWSDAY, July 18, 2006, at A19; Julia Preston, Texas Hospitals' Separate Paths Reflect the Debate on Immigration, N.Y. TIMES, July 18, 2006, at A1; Nicholas Riccardi, Immigration Hard-Liners on a High, L.A. TIMES, July 12, 2006, at A1; Allan Richter, Drawing Workers, and Some Critics, N.Y. TIMES, Apr. 30, 2006, [section] 14LI, at 1.

(4.) See JEFFREY S. PASSEL, PEW HISPANIC CTR., UNAUTHORIZED MIGRANTS: NUMBERS AND CHARACTERISTICS 5 (2005), available at http://pewhispanic.org/files/reports/46.pdf.

(5.) Compare Office of Policy & Planning, U.S. Immigration & Naturalization Serv., Estimates of the Unauthorized Immigrant Population Residing in the United States: 1990 to 2000 (2003), available at http://www.dhs.gov/xlibrary/assets/statistics/publications/Ill_ Report_1211.pdf, with Kelly Jefferys & Nancy Rytina, Office of Immigration Statistics, Dep't of Homeland Sec., U.S. Legal Permanent Residents: 2005 (2006), available at http://www.dhs.gov/xlibrary/assets/statistics/publications/ USLegalPermEst_5.pdf.

(6.) See John D. McKinnon, Bush Continues to Push His Plan on Immigration, WALL ST. J., July 8, 2006, at A4; Julia Preston, House and Senate Hold Immigration Hearings, N.Y. TIMES, July 6, 2006, at A12.

(7.) See Jefferys & Rytina, supra note 5, at 1 fig. 1.

(8.) See, e.g., JULIAN L. SIMON, THE ECONOMIC CONSEQUENCES OF IMMIGRATION (2d ed. 1999); MICHAEL WALZER, SPHERES OF JUSTICE: A DEFENSE OF PLURALISM AND EQUALITY (1983); Joseph H. Carens, Aliens and Citizens: The Case for Open Borders, 49 REV. POL. 251 (1987); Howard F. Chang, Liberalized Immigration as Free Trade: Economic Welfare and the Optimal Immigration Policy, 145 U. PA. L. REV. 1147 (1997); Kevin R. Johnson, Open Borders?, 51 UCLA L. REV. 193 (2003).

(9.) See, e.g., GERALD L. NEUMAN, STRANGERS TO THE CONSTITUTION: IMMIGRANTS, BORDERS, AND FUNDAMENTAL LAW (1996); Gabriel J. Chin, Segregation's Last Stronghold. Race Discrimination and the Constitutional Law of Immigration, 46 UCLA L. REV. 1 (1998); Stephen H. Legomsky, Immigration Law and the Principle of Plenary Congressional Power, 1984 SUP. CT. REV. 255; Hiroshi Motomura, Immigration Law After a Century of Plenary Power: Phantom Constitutional Norms and Statutory Interpretation, 100 YALE L.J. 545, 580-83 (1990). In addition to these two literatures, there is a large social science literature that focuses on migration's causes and consequences--economic, social, cultural, and otherwise. See, e.g., GEORGE J. BORJAS, HEAVEN'S DOOR: IMMIGRATION POLICY AND THE AMERICAN ECONOMY (1999); DETERMINANTS OF EMIGRATION FROM MEXICO, CENTRAL AMERICA, AND THE CARIBBEAN (Sergio Diaz-Briquets & Sidney Weintraub eds., 1991).

(10.) Although there is a small literature on second-order questions of institutional design, it lacks a theoretical framework. It focuses piecemeal on such questions as whether illegal immigration is best deterred by greater border or interior enforcement, by sanctioning employers who hire undocumented aliens or imposing sanctions on noncitizens themselves, or by giving authority to the federal government or to the states. See, e.g., Douglas S. Massey, Beyond the Border Buildup: Towards a New Approach to Mexico-U.S. Migration, IMMIGR. POL'Y IN FOCUS (Immigration Policy Inst., Washington, D.C.), Sept. 2005; Peter H. Schuck, Some Federal-State Developments in Immigration Law, 58 N.Y.U. ANN. SURV. AM. L. 387 (2002). These are all important issues, but they are only individual aspects of the overall institutional design of the immigration system and cannot be resolved without reference to a theory of that design.

(11.) As far as we have found, no one has recognized that immigration policy faces this asymmetric information problem. The most prominent economic work on immigration, such as that of George Borjas, addresses the first-order issues of type and quantity, with some attention to the problem of illegal immigration but mainly as an enforcement problem. See BORJAS, supra note 9.

(12.) Our argument is based on models of contracting and asymmetric information in the economics literature. See PATRICK BOLTON & MATHIAS DEWATRIPONT, CONTRACT THEORY (2005). Our argument that the academic literature should focus on second-order questions has analogies in many other policy areas--for example, the transition from the debate about whether markets are good or bad to the debate about how industries organize themselves. See JEAN TIROLE, THE THEORY OF INDUSTRIAL ORGANIZATION 1-3 (1988). The tradeoffs we discuss have also been identified in other contexts in the law and economics literature pertaining to risk regulation. See STEVEN SHAVELL, ECONOMIC ANALYSIS OF ACCIDENT LAW 277-85 (1987) (discussing ex ante and ex post approaches to regulation); William M. Landes & Richard A. Posner, The Economics of Anticipatory Adjudication, 23 J. LEGAL STUD. 683 (1994) (discussing the tradeoff between error costs and investment incentives).

(13.) We do not mean to suggest that immigration policy is driven solely by the goal of selecting the most preferred group of immigrants. In addition to using immigration policy as a selection mechanism, the United States has often used immigration law for many other purposes, including the promotion of other domestic or foreign policy objectives. For example, Congress has at times admitted, excluded, or deported immigrants specifically to send a signal to other states or to domestic political audiences. See, e.g., GEOFFREY R. STONE, PERILOUS TIMES: FREE SPEECH IN WARTIME: FROM THE SEDITION ACT OF 1789 TO THE WAR ON TERRORISM (2004). While these other aims also shape immigration policy in important ways, for the purposes of this Article, we are interested exclusively in the way in which states use immigration policy to shape the size and composition of the polity. Accordingly, we will focus on the selection-related aspects of immigration policy.

(14.) See BORJAS, supra note 9, at 176 (discussing quantity and type).

(15.) Actually, these extremes can be exceeded. On one side, states can expel people from the existing polity. See, e.g., MAE M. NGAI, IMPOSSIBLE SUBJECTS: ILLEGAL ALIENS AND THE MAKING OF MODERN AMERICA 71-75 (2004) (discussing the mass repatriation of Mexican Americans during the 1930s). On the other side, states can bribe (or coerce) people to immigrate (indeed, can conquer and assimilate populations). See, e.g., GAVAN DAWS, SHOAL OF TIME: A HISTORY OF THE HAWAIIAN ISLANDS (1968) (discussing the United States's colonization and assimilation of the Hawaiian kingdom). There is also the question whether a state should restrict exit--that is, emigration. Many states have historically restricted exit as well as entry. Nonetheless, for two reasons we focus exclusively on entry restrictions: first, we are interested in the way that immigration policy is used to select new members, not the way in which it might be used to compel continued membership; and second, modern liberal democracies uniformly permit unrestricted exit and the United States has never restricted exit in any significant way.

(16.) Quota Law of May 19, 1921, ch. 8, [section] 2(a), 42 Stat. 5 (introducing overall quota of 350,000 and limiting the admissions level of aliens of any nationality to three percent of the foreign-born persons of that nationality living in the United States in 1910). See generally U.S. Citizenship & Immigration Servs., Legislation from 1901-1940, at 3, available at http://www.uscis.gov/files/nativedocuments/Legislation%20from% 201901-1940.pdf.

(17.) Immigration Act of 1924, ch. 190, [section] 11(a)-(b), 43 Stat. 153, 159 (reducing ceiling to 165,000 or two percent of each nationality's population in the United States; in effect until June 30, 1927, when the level would be reduced to 150,000); Immigration and Nationality Act of 1965, Pub. L. No. 89-236, [subsection] 1, 21(a), 21(e), 79 Stat. 911,911,920-21 (limiting for the first time the number of immigrants from the Western Hemisphere and raising the combined ceiling to 290,000--170,000 from the Eastern Hemisphere and 120,000 from the Western Hemisphere); Act of Oct. 5, 1978, Pub. L. No. 95-412, [section] 1, 92 Stat. 907 (leaving ceiling at 290,000 while eliminating the Hemisphere quota split); Refugee Act of 1980, Pub. L. No. 96-212, [section] 203(a), 94 Stat. 102, 106-07 (lowering ceiling to 270,000); Immigration Reform and Control Act of 1986, Pub. L. No. 99-603, [section] 31 1(a), 100 Stat. 3359, 3434 (raising ceiling to 540,000); Immigration Act of 1990, Pub. L. No. 101-649, [subsection] 101(a), 112(a), 104 Stat. 4978, 4981-82, 4987 (raising ceiling to 700,000 for 1992-1994, setting the ceiling at 675,000 thereafter). Because refugees and close relatives are not counted toward the ceiling, actual immigration has been higher, as noted in the text.

(18.) ORG. FOR ECON. CO-OPERATION & DEV. [OECD], INTERNATIONAL MIGRATION OUTLOOK 30 tbl.I.1 (2006).

(19.) Id. These figures show that substantially greater numbers of immigrants came to the United States than these other large democracies. If the immigration flows are measured as a fraction of existing population, however, the picture is more mixed. In 2004, legal immigration to the United States constituted approximately 0.32% of the existing population. This is a larger fraction than in France (0.28%), Germany (0.25%), or Japan (0.07%), but a smaller fraction than in the United Kingdom (0.44%) or Canada (0.74%).

(20.) For a discussion of state laws that had the effect of regulating some immigration during this period, see Gerald L. Neuman, The Lost Century of Immigration Law (1776-1875), 93 COLUM. L. REV. 1833 (1993).

(21.) See ARISTIDE R. ZOLBERG, A NATION BY DESIGN: IMMIGRATION POLICY IN THE FASHIONING OF AMERICA (2006).

(22.) See Quota Law of May 19, 1921, ch. 8, 42 Stat. 5.

(23.) See Immigration Act of 1924, ch. 190, 43 Stat. 153.

(24.) See Immigration and Nationality Act of 1965, Pub. L. No. 89-236, [section] 1, 79 Stat. 911 (setting numerical ceilings on both Western and Eastern Hemispheres for the first time); Act of Oct. 5, 1978, Pub. L. No. 95-412, [section] 1, 92 Stat. 907 (combining for the first time the separate hemispheric quotas into a single global quota); Immigration Act of 1990, Pub. L. No. 101-649, [section] 101, 104 Stat. 4978, 4981 (setting current ceiling).

(25.) New Zealand, for example, requires that immigrants applying as skilled workers pass a test that matches their skills with the country's current needs. See Immigration New Zealand, The Skilled Migrant Category Points Indicator, http://www.immigration.govt.nz/pointsindicator.

(26.) See Citizenship & Immigration Canada, Six Selection Factors and Pass Mark, http://www.cic.gc.ca/english/skilled/qual-5.html.

(27.) See Immigration and Nationality Act (INA) [section] 201, 8 U.S.C. [section] 1151 (2007) (describing annual immigration limits); CHARLES GORDON ET AL., IMMIGRATION LAW AND PROCEDURE [section] 1.03(2)(e) (2006) [hereinafter IMMIGRATION LAW AND PROCEDURE].

(28.) Act of June 25, 1798 (Alien Enemies Act), ch. 58, 1 Stat. 570 (expired 1800).

(29.) See Neuman, supra note 20, at 1841-59.

(30.) See, e.g., JOHN HIGHAM, STRANGERS IN THE LAND: PATTERNS OF AMERICAN NATIVISM, 1860-1925 (2d ed. 1988); NGAI, supra note 15; LUCY E. SALYER, LAWS HARSH AS TIGERS: CHINESE IMMIGRANTS AND THE SHAPING OF MODERN IMMIGRATION LAW (1995).

(31.) See INA [section] 212(a)(2), 8 U.S.C. [section] 1182(a)(2) (listing criminal grounds of inadmissibility).

(32.) All states deny citizenship to some noncitizens who enter the country. That is because states admit noncitizens on a variety of temporary bases. As we explain below, however, our focus is principally on the process of permanent immigration. See infra text accompanying notes 36-37.

(33.) See T. Alexander Aleinikoff, Citizens, Aliens, Membership and the Constitution, 7 CONST. COMMENT. 9, 23 (1990) (noting that the United States does most of its selecting at the entry stage and not at the naturalization stage).

(34.) See Nicole Jacoby, Note, America's De Facto Guest Workers: Lessons from Germany 's Gastarbeiter for U.S. Immigration Reform, 27 FORDHAM INT'L L.J. 1569 (2004); see also Veysel Oezcan, Migration Policy Inst., Germany: Immigration in Transition (July 2004), http://www.migrationinformation.org/Profiles/display.cfm?ID=235.

(35.) The first naturalization law restricted naturalization to "free white person[s]." Act of Mar. 26, 1790, ch. 3, 1 Stat. 103. While the Fourteenth Amendment extended citizenship after the civil war to all persons born in the United States, race-based naturalization policies continued. All "races indigenous to the Western Hemisphere" were not added until the Nationality Act of 1940, ch. 876, [section] 303, 54 Stat. 1137, 1140, and Congress did not extend the right to naturalize to all Asians until 1952, see Immigration and Nationality Act, ch. 477, [section] 311, 66 Stat. 163, 239 (1952) (codified at 8 U.S.C. [section] 1422) ("The right of a person to become a naturalized citizen of the United States shall not be denied or abridged because of race or sex or because such person is married."). However, the effect of racially restrictive naturalization laws was lessened by the fact that, for much of this period, the United States prohibited immigration by those who were ineligible to naturalize. See Immigration Act of 1924, ch. 190, [section] 13(c), 43 Stat. 153, 162. For that reason, many potential immigrants of Asian ancestry who would have been ineligible to naturalize were also prohibited from immigrating to the United States.

(36.) See INA [section] 214, 8 U.S.C. [section] 1184 (setting forth the grounds for the admission of nonimmigrants to the United States).

(37.) A lawful permanent resident becomes a citizen by undergoing naturalization procedures. Because these procedures are not significant and at the same time many lawful permanent residents (LPRs) do not bother with them, we do not focus here on naturalization rules. However, we do not think that citizenship policy is unimportant. For a discussion of the connections between immigration policy and citizenship policy, see HIROSHI MOTOMURA, AMERICANS IN WAITING: THE LOST STORY OF IMMIGRATION AND CITIZENSHIP IN THE UNITED STATES (2006).

(38.) See infra Part III.B.2.

(39.) See ZOLBERG, supra note 21, at 258-64. Because the focus was on European immigration, the quota law did not apply to states in the Western Hemisphere. See Immigration Act of 1924, ch. 190, [section] 4, 43 Stat. 153, 155.

(40.) We focus on the immigration regulatory mechanisms that directly regulate potential immigrants, placing conditions upon their entrance and continuing residence in the receiving state. These are the policies at the core of what people usually consider to constitute immigration law. There are, of course, non-screening strategies a state might pursue to select its preferred mix of immigrants. For example, a state could attempt to alter directly the basic conditions that give rise to migration. If migration is driven in part by the domestic demand for labor, the state could try to reduce that demand by taxing or sanctioning employers who hired immigrants, by providing labor protections for some immigrants to make them more expensive employees, or so on. Similarly, if migration is driven in part by economic conditions in the migrants' home countries, the receiving state could give foreign aid to the home state in order to improve economic conditions and reduce these "push" factors. These approaches and many others--including state intervention in overseas conflicts--can be used to change the magnitude of migration, its character, or both.

(41.) See INA [section] 203(a), 8 U.S.C. [section] 1153(a) ("Preference allocation for family-sponsored immigrants"); INA [section] 203(b), 8 U.S.C. [section] 1153(b) ("Preference allocation for employment-based immigrants").

(42.) INA [section] 212(a), 8 U.S.C. [section] 1182(a).

(43.) INA [section] 212(a)(2)(A)(i)(I), 8 U.S.C. [section] 1182(a)(2)(A)(i)(I).

(44.) INA [section] 212(a)(3)(B)(iv)(VI), 8 U.S.C. [section] 1182(a)(3)(B)(iv)(VI).

(45.) INA [section] 212(a)(1)(A)(i), 8 U.S.C. [section] 1182(a)(1)(A)(i).

(46.) INA [section] 237, 8 U.S.C. [section] 1227.

(47.) See, e.g., INA [section] 237(a)(2)(A)(i), 8 U.S.C. [section] 1227(a)(2)(A)(i) (making deportable any noncitizen "convicted of a crime involving moral turpitude committed within five years ... after the date of admission"); INA [section] 237(a)(4)(B), 8 U.S.C. [section] 1227(a)(4)(B) (making deportable for terrorist activity "[a]ny alien who is described in subparagraph (B) or (F) of section 212(a)(3) [the inadmissibility section of the INA]").

(48.) See infra Part III.A.2.

(49.) See THOMAS ALEXANDER ALEINIKOFF ET AL., IMMIGRATION AND CITIZENSHIP: PROCESS AND POLICY 1096-1182 (5th ed. 2003) (discussing unauthorized migrants in the United States).

(50.) See infra text accompanying notes 62-67.

(51.) See INA [section] 235(b), 8 U.S.C. [section] 1225(b).

(52.) See INA [section] 235(b)(1)(A)(i), 8 U.S.C. [section] 1225(b)(1)(A)(i).

(53.) See INA [section] 235(b)(1)(C), 8 U.S.C. [section] 1225(b)(1)(C). There are some limited exceptions: the noncitizen is provided more procedural protections if she is found to have a credible fear of persecution if she returns to her home country, see INA [section] 235(b)(1)(A)(ii), (b)(1)(B), 8 U.S.C. [section] 1225(b)(1)(A)(ii), (b)(1)(B), or if she claims under oath to be a lawful permanent resident, see INA [section] 235(b)(1)(C), 8 U.S.C. [section] 1225(b)(1)(C).

(54.) See Li v. Eddy, 259 F.3d 1132 (9th Cir. 2001), vacated as moot, 324 F.3d 1109 (9th Cir. 2003).

(55.) See INA [section] 240, 8 U.S.C. [section] 1229a ("Removal proceedings").

(56.) See 8 C.F.R. [subsection] 1240.3, 1240.10(a)(4) (2007).

(57.) See id. [section] 1240.15.

(58.) See INA [section] 242, 8 U.S.C. [section] 1252 ("Judicial review of orders of removal").

(59.) See Press Release, Office of the Press Sec'y, Dep't of Homeland Sec., Department of Homeland Security Streamlines Removal Process Along Entire U.S. Border (Jan. 30, 2006), available at http://www.dhs.gov/xnews/releases/press_release_0845.shtm; see also Designating Aliens for Expedited Removal, 69 Fed. Reg. 48,877-01 (Aug. 11, 2004) (authorizing the use of expedited removal against noncitizens found within 100 miles of the Mexican or Canadian border who cannot demonstrate to an immigration enforcement official that they have been in the United States for more than fourteen days). Moreover, Immigration and Customs Enforcement (ICE) has authority under the INA to expand the use of expedited removal to cover the entire interior of the country and all non-admitted aliens who cannot demonstrate that they have been continuously present in the United States for at least two years. See INA [section] 235(b)(1)(A)(iii), 8 U.S.C. [section] 1225(b)(1)(A)(iii). In addition, expedited removal is not the only summary enforcement mechanism that applies to noncitizens who have already entered the country. Noncitizens who re-enter the country illegally after previously being deported are subject to a summary procedure known as "reinstatement of removal" regardless of how long they have been present in the United States. See INA [section] 241(a)(5), 8 U.S.C. [section] 1231(a)(5). And noncitizens who are not lawful permanent residents are subject to summary procedures when they are deported for certain criminal convictions. See 1NA [section] 238(a), 8 U.S.C. [section] 1228(a) (describing administrative removal for criminal aliens).

(60.) See INA [section] 235(b)(1)(C), 8 U.S.C. [section] 1225(b)(1)(C) (excepting those who claim to be lawful permanent residents from the expedited removal process).

(61.) The regulatory focus on admission rather than entry is in large part the product of Congress's comprehensive revision of the Immigration and Nationality Act in 1996. Prior to 1996, the INA differentiated principally between the concepts of exclusion and deportation. "Entry" marked the dividing line between exclusion and deportation, and "entry" typically meant physical entry into the United States. Thus, noncitizens who had entered--even surreptitiously--were subject to grounds of deportability rather than excludability; they were also entitled to the greater procedural protections of a deportation proceeding rather than an exclusion proceeding. See generally Matter of Lin, 18 I. & N. Dee. 219 (B.I.A. 1982). In 1996, Congress rewrote the INA, largely eliminating the exclusion-deportation line and replacing it with a focus on the concept of "admission." See Illegal Immigration Reform and Immigrant Responsibility Act of 1996, Pub. L. No. 104-208, div. C, 110 Stat. 3009-546. "Admission" rather than "entry" now marks the dividing line between the application of grounds of inadmissibility or deportability. And because admission is specifically defined as "the lawful entry of the alien into the United States after inspection and authorization by an immigration officer," this change puts immigrants who sneak across the border on the same footing as those who present themselves at the border. INA [section] 101(a)(13), 8 U.S.C. [section] 1101(a)(13). Both are subject to the grounds of inadmissibility, and both are more likely to be screened using summary procedures.

(62.) See, e.g., Chae Chan Ping v. United States, 130 U.S. 581 (1889); Shaughnessy v. United States ex rel. Mezei, 345 U.S. 206 (1953); Jean v. Nelson, 727 F.2d 957 (11th Cir. 1984) (en banc).

(63.) See, e.g., Mezei, 345 U.S. 206. Compare Borrero v. Aljets, 325 F.3d 1003 (8th Cir. 2003), with Rosales-Garcia v. Holland, 322 F.3d 386 (6th Cir. 2003) (en banc). This account almost surely overstates the extent to which the government operates without constitutional constraints when noncitizens present themselves for initial admission to the United States. Arriving aliens appear to have at least some minimal process rights--such as the right to file a habeas petition if in custody--and they may have other fights as well. See Kwong Hai Chew v. Colding, 344 U.S. 590 (1953); United States ex rel. Knauff v. Shaughnessy, 338 U.S. 537 (1950); Chae Chan Ping, 130 U.S. 581. Moreover, the government may be constrained in some situations by the rights of citizens when it acts to exclude noncitizens from the country. See Cox, supra note 2. Still, it is accurate to say that both substantive and procedural constitutional constraints are at their weakest when the government enforces immigration policy against arriving aliens.

(64.) See, e.g., Landon v. Plasencia, 459 U.S. 21 (1982); Yamataya v. Fisher, 189 U.S. 86 (1903); see also Linda Bosniak, A Basic Territorial Distinction, 16 GEO. IMMIGR. L.J. 407 (2002). While it is unclear under contemporary constitutional law what logical role each of these factors plays in determining a noncitizen's constitutional status, much turns on the answer. For example, if legal status were a necessary condition, noncitizens who have entered the United States without being admitted by the immigration authorities might lack any constitutional fights with respect to immigration actions taken against them and might, consequently, be treated by the federal government in the same fashion as arriving aliens.

(65.) Compare Zadvydas v. Davis, 533 U.S. 678 (2001), with Borrero, 325 F.3d 1003. See generally David A. Martin, Graduated Application of Constitutional Protections for Aliens: The Real Meaning of Zadvydas v. Davis, 2001 Sup. CT. REV. 47.

(66.) See Plasencia, 459 U.S. 21; cf. Yamataya, 189 U.S. 86.

(67.) See, e.g., Reno v. American-Arab Anti-Discrimination Comm., 525 U.S. 471 (1999); Dennis v. United States, 341 U.S. 494 (1951).

(68.) 118 U.S. 356 (1886).

(69.) See id. at 374.

(70.) See Bridges v. Wixon, 326 U.S. 135, 148 (1945) (citing Bridges v. California, 314 U.S. 252 (1941)).

(71.) See Wong Wing v. United States, 163 U.S. 228 (1896).

(72.) To be clear, we do not take a position as to what it is that makes a person fall into a particular "type"--that is, whether type is the product of innate characteristics or instead the product of social, economic, or other circumstances.

(73.) As will become clear, we analyze the government's second-order immigration design problem as a problem of hidden or private information, which is extensively discussed in the economics literature. See, e.g., BOLTON & DEWATRIPONT, supra note 12, at 47-99.

(74.) We use the term "assimilable" throughout in the broadest, most catholic sense--to refer to an immigrant's ability to adjust to living in a new society. Often, of course, the United States has also desired immigrants who are assimilable in a different, more controversial sense: immigrants who will abandon the cultural practices of their countries of origin and "Americanize."

(75.) Note that this later screening could be based on either ex ante or ex post information about the noncitizen, but most likely would include both.

(76.) To be clear, we should highlight the distinction between two important choices that a state must make when structuring its immigrant screening system. First is the choice between screening on the basis of pre-entry information or on the basis of information that develops post-entry. Second is the choice between screening at the point of entry or at some later time. Obviously, these two choices are connected. As noted above, a state cannot engage in ex post screening when an immigrant first applies for admission. But if a state chooses to screen immigrants at some time after entry, it can do so on the basis of either ex ante or ex post criteria. The following table provides the three combinations.
 Time of admission

Ex ante information Port of entry screening
Ex post information [None]

 Later time

Ex ante information Post-entry screening based
 on pre-entry information
 Post-entry screening based
Ex post information on post-entry information


Because our interest is principally in the information problem presented by immigration, we use the terms ex ante screening and ex post screening to capture the distinction in the information dimension, not the temporal dimension.

(77.) Cf. Landes & Posner, supra note 12 (describing the risk/information tradeoff in the context of risk regulation).

(78.) See generally BOLTON & DEWATRIPONT, supra note 12, at 489-91.

(79.) The legal structure of American immigration law accentuates the extent to which investments in personal relationships are country-specific. Noncitizens who are deported are inadmissible for at least ten years after being deported. See INA [section] 212(a)(9)(A)(ii), 8 U.S.C. [section] 1182(a)(9)(A)(ii).

(80.) The discussion in the text is based on an economic model that combines asymmetric information about type and ex post moral hazard. See BOLTON & DEWATRIPONT, supra note 12, at 228-32.

(81.) Specifically, risk-averse people who have already entered will invest more in order to avoid the punishment, but in order to avoid this cost, fewer risk-averse people will immigrate in the first place.

(82.) See infra text accompanying notes 133-34.

(83.) American immigration policy has long included criminal sanctions for immigrants who re-enter illegally after being deported. See INA [section] 276, 8 U.S.C. [section] 1326. These sanctions are greater than the criminal sanction for a first-time illegal entrant. See INA [section] 275, 8 U.S.C. [section] 1325.

(84.) U.S. immigration law does just this: noncitizens who are deported are inadmissible if they apply for readmission "within 10 years of the date of such alien's departure or removal (or within 20 years of such date in the case of a second or subsequent removal or at any time in the case of an alien convicted of an aggravated felony).... " INA [section] 212(a)(9)(A)(ii), 8 U.S.C. [section] 1182(a)(9)(A)(ii).

(85.) Cf. SHAVELL, supra note 12, at 279-81 (stating that an ex ante regulation is better when injurers cannot pay for harm done because of inadequate assets or detection problems).

(86.) See BOLTON & DEWATRIPONT, supra note 12, at 93.

(87.) For example, our predictions assume that states act as rationally self-interested actors. Political dynamics and other features may undermine this assumption.

(88.) We also ignore here the possibility that ex post screening increases the likelihood that the welfare of particular citizens is tied to the welfare of particular immigrants--where, say, a citizen develops social or familial ties with an immigrant. This possibility could also raise the relative cost of the ex post system, even for a government focused solely on the welfare of its own citizens.

(89.) See Ayelet Shachar, The Race for Talent: Highly Skilled Migrants and Competitive Immigration Regimes, 81 N.Y.U. L. REV. 148 (2006).

(90.) The Persian Gulf states have historically relied heavily on guest workers from Egypt and South Asia, but have maintained tight control on these temporary immigrants, conferring few benefits and resorting to strict cutbacks on work permits and mass deportations where perceived necessary. See ANDRZEJ KAPISZEWSKI, NATIONALS AND EXPATRIATES" POPULATION AND LABOUR DILEMMAS OF THE GULF COOPERATION COUNCIL STATES 5 (2001) ("[Expatriates] are temporary residents only and denied the possibility of obtaining citizenship; they are always dependent on a national who is responsible for all their legal and financial dealings; they have to leave the country once unemployed; and they are barred from any type of political involvement.").

(91.) See infra text accompanying note 136.

(92.) As we noted above, the classic distinction between exclusion and deportation in immigration law will often parallel the difference between ex ante and ex post screening. See supra note 76. But it need not always do so. Ex ante and ex post screening differ in the information that serves as the basis for the immigration decision. Ex ante screening relies on pre-entry information while ex post screening relies on new facts that develop after entry. While deportation by definition takes place after entry, it need not be based on new information that develops after entry. Because it often is (as a statutory matter) based on new information, however, the exclusion-deportation distinction is substantially correlated with the distinction between ex ante and ex post screening.

(93.) See Act of Mar. 3, 1875 (Page Act), ch. 141, [section] 1, 18 Stat. 477. The statute also prohibited felons and prostitutes from immigrating to the United States and criminalized the importation of prostitutes and "cooly" labor. Id. [subsection] 3, 4. In practice, the Act was enforced nearly exclusively against Chinese women. See Kerry Abrams, Polygamy, Prostitution, and the Federalization of Immigration Law, 105 COLUM. L. REV. 641, 698-702 (2005). 94. Act of May 6, 1882 (Chinese Exclusion Act), ch. 126, 22 Stat. 58.

(95.) Whether a potential immigrant fit into one of these two categories was not, oF course, always a simple factual matter. For a discussion about how immigration authorities enforced the provisions of these early statutes, see, for example, SALYER, supra note 30, and Abrams, supra note 93.

(96.) See Act of June 25, 1798, ch. 58, 1 Stat. 570, 570-71 ("[I]t shall be lawful for the President of the United States at any time during the continuance of this act, to order all such aliens as he shall judge dangerous to the peace and safety of the United States, or shall have reasonable grounds to suspect are concerned in any treasonable or secret machinations against the government thereof, to depart out of the territory of the United States...."). By its terms, the Act expired two years after its passage. See id. [section] 6.

One other potential exception is that certain elements of one of the Chinese exclusion laws of the 1890s may in practice have permitted a Chinese person to obtain readmission on the basis of a certificate that showed that he or she had developed significant contacts with the white community in the United States during her residence. Congress, in 1892, extended Chinese exclusion for ten years, broadened the coverage of exclusion to include most Chinese persons, and created a presumption that any Chinese person found in the United States was deportable "unless such person shall establish, by affirmative proof, ... his lawful right to remain in the United States." See Act of May 5, 1892 (Geary Act), ch. 60, [section] 3, 27 Stat. 25. The regulations implementing this provision required all Chinese immigrants to obtain a certificate of residence as proof of their lawful right to remain. But the regulations provided that the certificate could be obtained only on the basis of the testimony of two white witnesses. Had the regulations ever been widely enforced (they were not), they would have amounted to a de facto test of each Chinese immigrant's connections to the white community in the United States. See generally SALYER, supra note 30 (detailing the lack of enforcement of the certificate requirement).

(97.) It was not until 1907 that Congress passed legislation making immigrants deportable for engaging in prostitution after entering the country. See infra notes 100-01 and accompanying text.

(98.) Chinese Exclusion Act [section] 12, 22 Stat. at 61.

(99.) Id.

(100.) Prior to 1907, Congress had slowly begun to expand the statutory role of deportation. In 1891, for example, Congress made noncitizens deportable for one year following entry if they were found to have entered in violation of law. See Act of Mar. 3, 1891, ch. 551, [section] 11, 26 Stat. 1084, 1086. But this deportation provision simply reflected delayed ex ante screening, because deportation was based on pre-entry characteristics. The 1891 Act also made deportable "any alien who becomes a public charge within one year after his arrival in the United States from causes existing prior to his landing.... "Id. While this deportability ground turned in part on post-entry conduct--the immigrant's becoming a public charge--Congress's focus on ex ante characteristics is clear in its limitation of this provision to immigrants whose poverty arose from "causes existing prior to [entry]." Id.

(101.) Immigration Act of 1907, ch. 1134, [section] 3, 34 Stat. 898, 900.

(102.) See Immigration Act of 1917, ch. 29, [section] 19, 39 Stat. 874, 889 (making deportable "at any time within five years after entry ... any alien who at any time after entry shall be found advocating or teaching the unlawful destruction of property, or advocating or teaching anarchy, or the overthrow by force or violence of the Government of the United States"); id. (making deportable "any alien who is hereafter sentenced to imprisonment for a term of one year or more because of conviction in this country of a crime involving moral turpitude, committed within five years after the entry of the alien to the United States, or who is hereafter sentenced more than once to such a term of imprisonment because of conviction in this country of any crime involving moral turpitude, committed at any time after entry").

(103.) Immigration Act of 1907 [section] 3, 34 Stat. at 900.

(104.) See Act of May 26, 1922, ch. 202, 42 Stat. 596 (making deportable any noncitizen convicted of violating the statute's prohibition on the importation of or dealing in opium).

(105.) See Immigration and Nationality Act, ch. 477, 66 Stat. 163 (1952); see also E. P. HUTCHINSON, LEGISLATIVE HISTORY OF AMERICAN IMMIGRATION POLICY 1798-1965, at 307-13 (1981).

(106.) See Anti-Drug Abuse Act of 1988, Pub. L. No. 100-690, 102 Stat. 4181; Immigration Act of 1990, Pub. L. No. 101-649, 104 Stat. 4978; Immigration and Nationality Technical Corrections Act of 1994, Pub. L. No. 103-416, 108 Stat. 4305; Antiterrorism and Effective Death Penalty Act of 1996, Pub. L. No. 104-132, 110 Stat. 1214; Illegal Immigration Reform and Immigrant Responsibility Act of 1996, Pub. L. No. 104-208, div. C, 110 Stat. 3009-546.

(107.) See, e.g., Act of Mar. 26, 1910, ch. 128, [section] 3, 36 Stat. 263, 264-65 (eliminating the statute of limitations from the 1907 Act's ground of deportability for noncitizens who, after entry, practiced prostitution or were associated with a house of prostitution); Act of Oct. 16, 1918, ch. 186, 40 Stat. 1012 (eliminating the 1917 Immigration Act's statute of limitations on the deportability of anarchists); cf. Immigration Act of Feb. 5, 1917 [section] 19 (extending to five years the statute of limitations for deporting public charges). Today such statutes of limitation remain for only a few grounds of deportability. See, e.g., INA [section] 237(a)(2)(A)(i)(I), 8 U.S.C. [section] 1227(a)(2)(A)(i)(I) (making deportable noncitizens convicted of a single "crime involving moral turpitude committed within five years ... after the date of admission").

(108.) Prior to 1996, statutory relief from deportation was available under a variety of circumstances. All deportable noncitizens who could otherwise qualify for an immigrant visa--even those without lawful status--were eligible for suspension of deportation if they had lived for a sufficient period in the United States, were of good moral character, and could make a showing of extreme hardship. See 1NA [section] 242, 8 U.S.C. [section] 1254 (1994), repealed by Illegal Immigration Reform and Immigrant Responsibility Act of 1996, Pub. L. No. 104-208, div. C, [section] 308(b)(7), 110 Stat. 3009-546, 3009-615. For lawful permanent residents, somewhat more generous relief was also available under INA [section] 212(c). Congress significantly restricted the availability of relief from removal in 1996 when it consolidated the various relief provisions. See INA [section] 240A, 8 U.S.C. [section] 1229b. After 1996, for example, noncitizens convicted of "aggravated felonies" are categorically ineligible for relief from removal. See INA [section] 240A(a)(3), 8 U.S.C. [section] 1229b(a)(3). In addition to making the screening system more categorical, we should note that the modern restrictions on relief from removal have also altered the temporal dimension of the screening process. This is because relief from removal has historically offset partially the elimination of statutes of limitation from deportation provisions, as lengthy continuous residence in United States was often a central factor in the provision of relief from removal. See INA [section] 240A(a)-(b), 8 U.S.C. [section] 1229b(a)(b).

(109.) For a discussion of the relationship between purposeful and statistical discrimination, see David A. Strauss, The Myth of Colorblindness, 1986 SUP. CT. REV. 99, 113-16.

(110.) See ZOLBERG, supra note 21.

(111.) NGAI, supra note 15, at 64.

(112.) See id. at 52-53, 58-67.

(113.) While the cost of policing the border rose, it may also have been the case that improved recordkeeping and communications technology--allowing authorities to keep track of and share information about post-entry activities--reduced the cost of the ex post system.

(114.) Of course, there are other possible explanations as well. While we describe more fully below the ways in which constitutional law can shape the second-order structure of immigration, it is worth noting that the steady expansion of ex post screening through the proliferation of grounds of deportability might also have been connected to changes over time in constitutional immigration law. The expansion of deportation may track the story that Bill Stuntz has told about the expansion of substantive criminal law. As Stuntz has explained, the Warren Court's strengthening of constitutional criminal procedure made criminal prosecutions more difficult and expensive for the government. This created an incentive to try to circumvent these constitutional protections. One way to do that was to expand dramatically the scope of substantive criminal law. See William J. Stuntz, Substance, Process, and the Civil-Criminal Line, 7 J. CONTEMP. LEGAL ISSUES 1 (1996); William J. Stuntz, The Pathological Politics of Criminal Law, 100 MICH. L. REV. 505 (2001). Similar logic can help explain the expansion in recent years of deportability grounds. Rather than reflecting a change in first-order immigration preferences, the expansion might be a second-order strategy to preserve government discretion that has been whittled away by the expansion of due process protections available to noncitizens placed in deportation proceedings. These protections have slowly expanded over the course of the last century (though perhaps not quite as dramatically as criminal procedural connections expanded under the Warren Court). See, e.g., Motomura, supra note 9; Peter H. Schuck, The Transformation of Immigration Law, 84 COLUM. L. REV. 1 (1984). In practice, such due process protections make it more difficult and costly for the government to deport a noncitizen. But expanding the grounds of deportability and eliminating forms of relief from deportation can help reduce these costs. Thus, the expansion of deportability grounds can help augment the discretion of immigration enforcement agencies by relocating that discretion to the charging stage of the enforcement process.

(115.) See supra note 47 and accompanying text.

(116.) See Anti-Drug Abuse Act of 1988, Pub. L. No. 100-690, [section] 7344(a), 102 Stat. 4181, 4470-71.

(117.) See Immigration Act of 1990, Pub. L. No. 101-649, [section] 501, 104 Stat. 4978, 5048; Immigration and Nationality Technical Corrections Act of 1994, Pub. L. No. 103-416, [section] 222, 108 Stat. 4305, 4320; Antiterrorism and Effective Death Penalty Act of 1996, Pub. L. No. 104-132, [section] 440, 110 Stat. 1214, 1276; Illegal Immigration Reform and Immigrant Responsibility Act of 1996, Pub. L. No. 104-208, div. C, [section] 321, 110 Stat. 3009-546, 3009-627.

(118.) See INA [section] 101(a)(43), 8 U.S.C. [section] 1101(a)(43); IMMIGRATION LAW AND PROCEDURE, supra note 27, [section] 71.05(2)(d) (examining case law interpreting the breadth of "aggravated felony"); Dawn Marie Johnson, Note, The AEDPA and the IIRIRA: Treating Misdemeanors as Felonies for Immigration Purposes, 27 J. LEGIS. 477 (2001).

(119.) See, e.g., Teresa A. Miller, Citizenship & Severity: Recent Immigration Reforms and the New Penology, 17 GEO. IMMIGR. L.J. 611, 633-35 (2003).

(120.) See INA [section] 240A(a), 8 U.S.C. [section] 1229b(a) (making noncitizens "convicted of any aggravated felony" ineligible for cancellation of removal); INA [section] 208(b)(2)(B)(i), 8 U.S.C. [section] 1158(b)(2)(B)(1) (making such noncitizens ineligible for asylum).

(121.) Compare INA [section] 237(a)(2), 8 U.S.C. [section] 1227(a)(2), with INA [section] 212(a)(2), 8 U.S.C. [section] 1182(a)(2).

(122.) This does not mean that there are no criminal convictions that make noncitizens inadmissible. As we explained in Part I, the INA uses criminal convictions as grounds of both inadmissibility and deportability. And a conviction that constitutes an aggravated felony can also fall under one of the grounds of inadmissibility--if, for example, the conviction also constitutes a crime involving moral turpitude. See INA [section] 212(a)(2)(A)(i)(I), 8 U.S.C. [section] 1182(a)(2)(A)(i)(I) (defining "crime involving moral turpitude" as a ground of inadmissibility). Even where there is overlap, however, convictions that constitute aggravated felonies result in different treatment. As noted above, for example, a noncitizen with an aggravated felony conviction is ineligible for almost all forms of relief from removal. See IMMIGRATION LAW AND PROCEDURE, supra note 27, [section] 71.05(2)(c). A noncitizen with a conviction for a crime involving moral turpitude is not subject to the same bars on relief. See id. [section] 71.05(1)(e).

(123.) This cannot be the entire story. For while the INA's deportability grounds cover only aggravated felony convictions "after admission," INA [section] 237(a)(2)(A)(iii), 8 U.S.C. [section] 1227 (a)(2)(A)(iii), those convictions can technically be for an "offense in violation of the law of a foreign country" (though in practice they almost never are), INA [section] 101(a)(43)(U), 8 U.S.C. [section] l101(a)(43)(U). Still, the sort of skepticism described above has been common from the time of the earliest federal immigration law. The first federal statute to make a person excludable on criminal grounds, the Page Act, made excludable immigrants "who are undergoing a sentence for conviction in their own country of felonious crimes...." Act of March 3, 1875 (Page Act), ch. 141, [section] 5, 18 Stat. 477 (repealed 1974). But the statute exempted crimes that were "political" or that were "growing out of or [were] the result of such political offenses"--an acknowledgment that felony convictions in other countries sometimes would not provide reliable information about whether or not an immigrant was desirable. Id. This exception was carried forward in subsequent immigration statutes and remains in the modern INA. See INA [section] 212(a)(2)(A)(i)(I), 8 U.S.C. [section] 1182(a)(2)(A)(i)(I). More generally, skepticism about foreign law and legal judgments is an entrenched feature of American law, though there are some exceptions. See Eric A. Posner & Cass Sunstein, Chevronizing Foreign Relations Law, 116 YALE L.J. (forthcoming 2007).

(124.) There are alternative theories as well. One possibility is that the harsher treatment of post-entry convictions reveals that the power to deport is being used as an additional criminal sanction--rather than being used as an immigration screening device. While much has recently been written on the intersection of immigration law and criminal law, this possibility is too often overlooked. See generally Miller, supra note 119; Daniel Kanstroom, Deportation, Social Control, and Punishment: Some Thoughts About Why Hard Laws Make Bad Cases, 113 HARV. L. REV. 1890 (2000).

(125.) See supra text accompanying notes 62-71.

(126.) See U.S. CONST. amend. XIV, [section] 1 ("All persons born or naturalized in the United States, and subject to the jurisdiction thereof, are citizens of the United States and of the State wherein they reside."); see also United States v. Wong Kim Ark, 169 U.S. 649 (1898).

(127.) See infra Part III.B (discussing the illegal immigration system).

(128.) See PASSEL, supra note 4.

(129.) See, e.g., NGAI, supra note 15, at 129-53; Frank D. Bean & David A. Spener, Controlling International Migration Through Enforcement: The Case of the United States, in INTERNATIONAL MIGRATION: PROSPECTS AND POLICIES IN A GLOBAL MARKET 352 (Douglas S. Massey & J. Edward Taylor eds., 2004); Wayne A. Cornelius & Takeyuki Tsuda, Controlling Immigration: The Limits of Government Intervention, in CONTROLLING IMMIGRATION: A GLOBAL PERSPECTIVE 3, 7-15 (Wayne A. Cornelius et al. eds., 2d ed. 2004); Monica Heppel & Demetrios Papademetriou, Government Intervention and the Farm Labor Market: How Past Policies Shape Future Options, in FORUM FOR TRANSNATIONAL EMPLOYMENT 38 (Amber Lopez, Cal. Inst. for Rural Studies ed., 2001); Philip Martin, Does the U.S. Need a New Bracero Program?, 9 U.C. DAVIS J. INT'L L. & POL'Y 127 (2003); Pia M. Orrenius & Madeline Zavodny, Do Amnesty Programs Reduce Undocumented Immigration? Evidence from IRCA, 40 DEMOGRAPHY 437 (2003).

(130.) See, e.g., Howard F. Chang, The Immigration Paradox: Poverty, Distributive Justice, and Liberal Egalitarianism, 52 DEPAUL L. REV. 759 (2003); Johnson, supra note 8; cf. NGAI, supra note 15.

(131.) INA [section] 212(a)(6)(A)(i), 8 U.S.C. [section] 1182(a)(6)(A)(i) ("An alien present in the United States without being admitted or paroled, or who arrives in the United States at any time or place other than as designated by the Attorney General, is inadmissible.").

(132.) This highlights the way in which the distinction between ex ante and ex post screening does not always track the distinction between screening at the border and screening after entry. Post-entry screening is still ex ante to the extent that it is based on information that existed at the time of entry rather than on new facts that developed after entry. See supra note 76 and accompanying text.

(133.) Immigration and Customs Enforcement (ICE) conducts investigations covering five primary classes of activity: crime, illegal employment, fraud, immigrant smuggling, and status violations. In recent years, criminal investigations have been the bread and butter of ICE's enforcement actions, making up 70% of ICE's apprehensions in fiscal year 2003. See OFFICE OF IMMIGRATION STATISTICS, DEP'T OF HOMELAND SEC., 2003 YEARBOOK OF IMMIGRATION STATISTICS 147 (2004) [hereinafter 2003 YEARBOOK], available at http://www.dhs.gov/xlibrary/assets/statistics/yearbook/2003/2003Yearbook.pdf; see also IMMIGRATION & NATURALIZATION SERV., 2001 STATISTICAL YEARBOOK OF IMMIGRATION AND NATURALIZATION SERVICE tbl.61 (2002), available at http://www.dhs.gov/xlibrary/assets/statistics/yearbook/2001/ ENF2001tables.pdf. Moreover, initiatives like the Institutional Removal Program (in which ICE works with state and local law enforcement to remove incarcerated noncitizens), Operation Community Shield (in which ICE uses immigration law to target gangs), and the Fugitive Absconder Program (in which ICE focuses on the capture and removal of fugitive aliens, particularly those who have committed crimes) further enhance the agency's focus on criminal aliens. In light of these enforcement priorities, "[f]ederal immigration officials[] ... maintain that the vast majority of illegal immigrants detained and deported are people convicted or charged with serious crimes." Paul Vitello, Path to Deportation Can Start with a Traffic Stop, N.Y. TIMES, Apr. 14, 2006, at A1; see also Your World with Neil Cavuto (Fox News television broadcast June 15, 2006) (transcript available at 2006 WLNR 10374973) (interview with Julie Myers, Assistant Secretary for Immigration and Customs Enforcement). Over the last decade, ICE has invested substantially fewer resources in other enforcement activities, such as workplace raids, employer investigations, and enforcement against noncriminal visa overstayers. See William Branigin, INS Shifts 'Interior' Strategy to Target Criminal Aliens; Critics Say Plan to Curtail Work-Site Raids Will Hurt Immigration Compliance, WASH. POST, Mar. 15, 1999, at A03; see also 2003 YEARBOOK, supra, at 157 tbl.39 (showing a plunge from 17,554 employment-related arrests in 1997 to 445 such arrests in 2003 and from 865 issuances of Notices of Intent to Fine employers to a low of 53 such issuances in 2002); Eric Lipton, Report Finds U.S. Failing on Overstays of Visas, N.Y. TIMES, Oct. 22, 2005, at A13 ("[T]he inspector general [of the Department of Homeland Security], Richard L. Skinner, predicted that a 'minuscule' number of [visitors who overstay their visas] were ever likely to face deportation, an action generally taken only if a person has a criminal history and is detained."). And while ICE has in recent months undertaken a few high-profile workplace enforcement actions, see, e.g., Nicholas Riccardi & Nicole Gaouette, Employers' Immigration Pains, L.A. TIMES, Dec. 17, 2006, at A37, these activities do little to alter the agency's overall enforcement priorities. Instead, they appear more likely to be related to the ongoing immigration reform debates in Congress.

(134.) In fact, because ICE focuses on criminal immigrants and faces significant resource constraints, the agency often declines to pick up undocumented immigrants who are detained by local law enforcement for having committed a minor infraction or appearing to be in the country illegally. See Julia Preston, New Scrutiny of Illegal Immigrants in Minor Crimes, N.Y. TIMES, June 20, 2006, at A10 ("Immigration agents, overwhelmed by a decade-old surge in illegal immigration to Colorado, said they had neither the time nor the resources to pick up the illegal immigrants [arrested by local authorities] whose violations were not grave."); Vitello, supra note 133 ("There are simply not enough immigration agents to respond every time a suspected illegal immigrant is arrested for driving with an invalid license.") (quoting a spokesman for ICE).

(135.) The government identifies deportable immigrants in the criminal justice system in several ways. One way is through prison screening: ICE regularly interviews prisoners in a number of prison systems in order to locate criminal aliens, and several state and local law-enforcement agencies have recently entered into agreements with the federal government under which the law enforcement agencies are delegated authority to do the screening themselves. See, e.g., Lance Pugmire, Immigration Check at Inland Jail Is OKd, L.A. TIMES, Sept. 21, 2005, at B3 (discussing agreements between federal immigration officials and Los Angeles and San Bernardino counties allowing local law enforcement officials to screen for illegal immigrants); see also 2003 YEARBOOK, supra note 133, at 150 (discussing the Institutional Removal Program, whereby DHS cooperates with local law enforcement to apprehend incarcerated immigrants).

(136.) See Comprehensive Immigration Reform Act of 2006, S. 2611, 109th Cong. [section] 601 (passed by the Senate on May 25, 2006).

(137.) Even if this is true, the heavy reliance on ex post screening might be objectionable on other grounds. For example, ex post screening will often impose greater costs than ex ante screening on immigrants and their families. Cf. supra text accompanying note 85 (discussing the costs ex post screening can impose on residents with whom immigrants develop relationships). One might argue that, as a moral matter, those personal costs should weigh heavily in the policy calculation. Our explanation is meant only to suggest that there are reasons why a self-interested government might prefer to rely principally on ex post screening to select certain kinds of immigrants.

(138.) Nancy H. Chau, Strategic Amnesty and Credible Immigration Reform, 19 J. LAB. ECON. 604, 605 (2001).

(139.) See Immigration Reform and Control Act of 1986, Pub. L. No. 99-603, 100 Stat. 3359.

(140.) Comprehensive Immigration Reform Act of 2006, S. 2611, 109th Cong. [section] 601 (passed by the Senate on May 25, 2006); Maura Reynolds, GOP Sends Mixed Messages on Immigration, L.A. TIMES, Aug. 27, 2006, at A17.

(141.) Comprehensive Immigration Reform Act of 2006.

(142.) Id.

(143.) Id. There are other complications for this group of immigrants. Although they too are provided a path to eligibility for legal permanent residency, the path is less straightforward than it is for those who have been in the United States for more than five years. Under the bill, immigrants in the country for more than two but less than five years are initially granted only a three-year visa, after which they must briefly leave the United States before returning. Id.

(144.) Id. This criminal history requirement is considerably stricter than the existing crime-based grounds of inadmissibility and deportability. See INA [section] 212(a)(2), 8 U.S.C. [section] 1182(a)(2); INA [section] 237(a)(3), 8 U.S.C. [section] 1227(a)(3).

(145.) The ex post screening criteria contained in the Senate bill are different from--and in several ways more stringent than--those embodied in the INA's deportability grounds, which apply to all admitted noncitizens. See supra note 144. There are, of course, a number of reasons why Congress might adopt more stringent ex post screening criteria as part of its regularization legislation. The differences might be driven by the fact that immigrants covered by the regularization legislation were never subjected to any ex ante screening; by the fact that the regularization program is attempting to select for different sorts of immigrants than are other parts of our immigration policy; or simply by the fact that different political dynamics constrain Congress's efforts to adopt a legalization program.

(146.) See supra text accompanying note 86.

(147.) See Comprehensive Immigration Reform Act of 2006 tit. VI.

(148.) See id. [section] 403.

(149.) See id. (stating that, subject to some exceptions, "the period of authorized admission ... shall terminate if the alien is unemployed for 60 or more consecutive days").

(150.) See id. [section] 408. While these immigrants do not, as a result of their employment history and residence in the United States, automatically qualify for LPR status, they appear to face considerably lower screening requirements than potential employment-based immigrants who have not previously lived and worked in the United States. See id. (easing the adjustment of status requirements for a noncitizen who has been employed in H-2C status for some time).

(151.) Critics of guest worker programs have raised a number of concerns about such programs. These concerns, which focus on harms to the guest workers, to domestic workers, or to the receiving country more generally include: (1) that guest worker programs increase the exploitation of immigrant workers or perhaps even create a de facto caste system in which American citizens have the full privileges of citizenship while nominally temporary guest workers become second-class citizens; (2) that they harm those domestic workers who are the least economically secure; (3) that they exacerbate rather than ameliorate the problem of unauthorized immigration; and (4) that they undermine existing mechanisms of assimilation, even for immigrants who are not part of the system. See, e.g., BORJAS, supra note 9; NGAI, supra note 15; Martin, supra note 129; Cristina M. Rodriguez, Guest Worker Programs and the Threat to Immigrant Assimilation, 2007 U. CHI. LEGAL F. (forthcoming). These concerns are made more salient by the troubled history of guest worker programs in the United States, Germany, and other countries. In the United States, the Bracero guest worker program that ran from WWII until the late 1950s is thought by some to have led to most of the above problems. See, e.g., NGAI, supra note 15. Germany's high-profile guest worker program has been similarly criticized. See Jacoby, supra note 34. To be sure, it would be a mistake to conclude from these examples that guest worker programs were never worthwhile. The details of any potential program would be crucially important. A central problem with the German program was that children born to guest workers (and the children born to those children) were not granted German citizenship. See id. While the presence of he guest workers, children, and grandchildren raised serious concerns about assimilation, deportation came to be regarded as both unfair and politically infeasible. A U.S. program would obviously not have this structural issue because of the birthright citizenship clause in our Constitution. And more generally, of course, the United States has traditionally had more success assimilating noncitizens than have European countries.

(152.) See supra Parts I.C, III.A.3.

(153.) Cf. Respondents-Appellees' Opposition to Request for Release from Detention Pending Appeal at 9, Nadarajah v. Gonzales, 443 F.3d 1069 (9th Cir. 2006) (No. 05-56759) (arguing that "[a]liens arriving at our borders who seek admission have no constitutional right to be admitted or paroled into the United States" and that, consequently, the petitioner had no due process right to contest his detention, even where that detention was indefinite).

(154.) See, e.g., Landon v. Plasencia, 459 U.S. 21 (1982); supra text accompanying notes 64-67.

(155.) See Yamataya v. Fisher, 189 U.S. 86 (1903) (holding that the Constitution confers procedural due process rights on noncitizens in removal proceedings); cf. Zadvydas v. Davis, 533 U.S. 678 (2001) (holding that due process required by the Constitution precludes the unreasonably lengthy detention of a noncitizen subject to a final order of removal).

(156.) 8 U.S.C. [section] 1225(b); see supra text accompanying notes 51-54.

(157.) INA [section] 235(b)(1)(A)(iii), 8 U.S.C. [section] 1225(b)(1)(A)(iii) ("The Attorney General may apply [expedited removal] to any ... alien[] ... who has not been admitted or paroled into the United States, and who has not affirmatively shown, to the satisfaction of an immigration officer, that the alien has been physically present in the United States continuously for the 2-year period immediately prior to the date of the determination of inadmissibility under this subparagraph."). While the Attorney General is currently not required to use expedited removal for all noncitizens within this statutory class, recently proposed legislation would make expedited removal mandatory for all noncitizens in the country for less than two years. See ALISON SISKIN & RUTH ELLEN WASEM, CONG. RESEARCH SERV., IMMIGRATION POLICY ON EXPEDITED REMOVAL OF ALIENS 18 (2006), available at http://opencrs.cdt.org/rpts/RL33109_20060118.pdf.

(158.) INA [section] 235(b)(1)(A)(i), 8 U.S.C. [section] 1225(b)(1)(A)(i). Narrow exceptions are available for a noncitizen who indicates a fear of persecution in his home country or who claims under oath to already have been lawfully admitted as a permanent resident or a refugee. See INA [section] 235(b)(1)(A)(ii), (b)(1)(C), 8 U.S.C. [section] 1225(b)(1)(A)(ii), (b)(1)(C).

(159.) Cf. Plasencia, 459 U.S. 21.

(160.) Again, of course, we do not mean to suggest that this is the only reason why Congress might prefer the illegal system. An alternative theory is that the tolerance of illegal immigration allows the government or public to maintain an "enticing myth" about American norms--"to avoid seeing ourselves as the sort of people who exploit the vulnerability of outsiders by holding a formal competition within our borders." Saul Levmore, Unconditional Relationships, 76 B.U. L. REV. 807, 818 (1996).

(161.) See Kara Murphy, Attracting the Best and the Brightest: The Promise and Pitfalls of a Skill-Based Immigration Policy, IMMIGR. POL'Y IN FOCUS (Immigration Policy Inst., Washington, D.C.), Dec. 2006, available at http://www.ailf.org/ipc/infocus/infocus_ 120506.pdf. See generally Nora V. Demleitner, How Much Do Western Democracies Value Family and Marriage?: Immigration Law's Conflicted Answers, 32 HOFSTRA L. REV. 273 (2003); Susan Martin & B. Lindsay Lowell, Competing for Skills: U.S. Immigration Policy Since 1990, 11 LAW & BUS. REV. AM. 387 (2005). Between 2003 and 2005, the United States admitted between 489,082 and 649,201 family-based immigrants per year. This amounted to between fifty-eight and seventy percent of the immigrant visas awarded in these years. See OFFICE OF IMMIGRATION STATISTICS, DEP'T OF HOMELAND SEC., 2005 YEARBOOK OF IMMIGRATION STATISTICS 18-19 tbl.6 (2006), available at http://www.dhs.gov/xlibrary/assets/ statistics/yearbook/2005/OIS_2005_Yearbook.pdf (authors' calculations based on the table). In contrast, Canada during this period admitted between 63,352 and 65,125 family-based immigrants, which constituted twenty-four to twenty-nine percent of the immigrant visas awarded. See CITIZENSHIP & IMMIGRATION CANADA, ANNUAL REPORT TO PARLIAMENT ON IMMIGRATION 2006, at 17 tbl.3 (2006), available at http://www.cic.gc.ca/english/pdf/ pub/immigration2006_e.pdf; Citizenship & Immigration Canada, Facts and Figures 2004: Immigration Overview: Permanent and Temporary Residents (July 31, 2005), http://www.cic.gc.ca/english/pub/facts2004/overview/1.html.

(162.) Immigration and Nationality Act, ch. 477, 66 Stat. 163 (1952) (codified at 8 U.S.C. [section] 1101 et seq.).

(163.) H.R. REP. No. 82-1365 (1952), as reprinted in 1952 U.S.C.C.A.N. 1653, 1691.

(164.) S. REP. No. 89-748 (1965), as reprinted in 1965 U.S.C.C.A.N. 3328, 3332.

(165.) See Shachar, supra note 89, at 171-74. It is notable, however, that the Canadian scheme allowed for the contemporaneous migration of the approved worker's family. See id.

(166.) Gabriel J. Chin, The Civil Rights Revolution Comes to Immigration Law: A New Look at the Immigration and Nationality Act of 1965, 75 N.C.L. REV. 273 (1996).

(167.) See ZOLBERG, supra note 21. Of course, the family migration rules did not end up having this effect. Instead, there is evidence that they have contributed to the influx of Hispanics into the United States by providing a legal pathway for Hispanic immigrants to bring relatives to the United States--in some cases without being subject to the immigration quotas. See INA [section] 201(b)(2)(A)(i), 8 U.S.C. [section] 1151(b)(2)(A)(i) (exempting immediate relatives of U.S. citizens from the numerical restrictions on immigration).

(168.) For discussion of the role that family structures can play in the assimilation of individuals, see generally THE NEW MIGRATION: AN INTERDISCIPLINARY READER (Marcelo M. Suarez-Orozco et al. eds., 2005).

(169.) See, e.g., HUNTINGTON, supra note 3.

(170.) See SIMON, supra note 8, at 357-63; Gary S. Becker, An Open Door for Immigrants--the Auction, WALL ST. J., Oct. 14, 1992, at A14.

(171.) A related idea is Michael Trebilcock's mandatory insurance scheme, under which anyone could enter the United States as long as she purchases insurance against becoming a public charge. See Michael J. Trebilcock, The Law and Economics of Immigration Policy, 5 AM. L. & ECON. REV. 271 (2003).

(172.) Alternatively, the noncitizen could borrow on pain of deportation if she defaults on the loan. For a discussion, see Levmore, supra note 160, at 811.

(173.) Trebilcock's scheme suffers a similar defect. Like the auction, it ensures that the noncitizens create welfare gains for the host society. But, also like the visa auction, mandatory insurance does not take seriously first-order concerns about assimilability--and, unlike the visa auction, does not even take seriously a state's interest in controlling the size of its population.

Adam B. Cox * and Eric A. Posner **

* Assistant Professor of Law, University of Chicago Law School.

** Kirkland and Ellis Professor of Law, University of Chicago Law School. Many thanks to Ahilan Arulanantham, Douglas Baird, Larry Frankel, Jacob Gersen, Bill Landes, Stephen Legomsky, Tracey Meares, Thomas Miles, Hiroshi Motomura, Richard Posner, Matthew Price, Cristina Rodriguez, Matthew Stephenson, Lior Strahilevitz, Adrian Vermeule, David Weisbach, and participants in workshops at the University of Chicago, University of Notre Dame, and the University of Texas-Austin, for helpful comments. Thanks also to Sarah Waxman for excellent research assistance.
COPYRIGHT 2007 Stanford Law School
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2007 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Cox, Adam B.; Posner, Eric A.
Publication:Stanford Law Review
Date:Feb 1, 2007
Words:25331
Previous Article:Intuition or proof: the social science justification for the diversity rationale in Grutter v. Bollinger and Gratz v. Bollinger.
Next Article:Choosing immigrants, making citizens.
Topics:


Related Articles
Setback to state-level protection for immigrants.
Choosing immigrants, making citizens.
Obama administration refuses to E-Verify.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters