The Welfare State versus the Common Labor Market: Which to Dismantle?
* Department of Economics, Michigan State University, East Lansing, MI 48824, USA, e-mail: wilsonjd{at}msu.edu
 |
Abstract |
|---|
 Top Abstract 1 Introduction2 The model3 Equilibrium redistribution...4 Equilibrium immigration...5 Extensions and discussion6 Concluding remarksAppendix 1References |
|---|
Migration by low-income workers limits the ability of a country to redistribute income, since more generous income supplements attract additional workers into the country, reducing wages and raising the cost of the program. This article studies the role of immigration controls, which allow the government to raise the real incomes of existing immigrants without causing additional immigration. Paradoxically, immigration controls may lead to higher equilibrium levels of immigration in a common labor market, and those low-income individuals left behind in the source countries may be better off. Simply stated, a host country benefits more from immigrants when they are not impoverished, and immigration controls enable the country to eliminate impoverishment. Thus, the country is willing to increase the number of immigrants that it allows within its borders. After obtaining this insight from the basic model, the article discusses some extensions and qualifications.
|
1 Introduction |
|---|
|
TopAbstract1 Introduction2 The model3 Equilibrium redistribution...4 Equilibrium immigration...5 Extensions and discussion6 Concluding remarksAppendix 1References |
|---|
The problems associated with income redistribution in a common labor market are well-known. If a single jurisdiction attempts to redistribute income from high-income residents to mobile low-income residents, then it will attract additional low-income individuals, until the utilities of low-income residents fall to the levels available outside the jurisdiction. Consequently, there will be a "race to the bottom" in welfare payments, with a system of many jurisdictions engaging in little or no income redistribution. To the extent that high-income individuals are also mobile, they will be induced to leave a jurisdiction attempting to redistribute income, further reducing its redistribution capabilities.
As an important example, consider the concerns over the possibility that free migration in the European Union could lead to the "dismantlement" of the EU's "welfare state". These concerns have been heightened by the expansion of the EU in 2004 to include ten new countries (Cyprus, Czech Republic, Estonia, Hungary, Latvia, Lithuanaia, Malta, Poland, Slovakia and Slovenia). At the time of expansion, average wages in the new member states were substantially below the average in the old member states, and the old member states tended to offer generous welfare programs. As a result of the fears of mass immigration and its negative effects, transitional rules were included in the Accession Treaty, allowing old member states to restrict immigration for up to 7 years (the "2-3-2" system). With the addition of Romania and Bulgaria to the EU in 2007, the elimination of all labor movement restrictions has been further delayed until 2014. As a result of these restrictions and other impediments to labor mobility within the EU, including cultural and language differences, the feared mass migration has not occurred.
While these restrictions to free mobility are intended to be transitional, an important question is whether the ultimate goal should necessarily be free migration. The purpose of the present article is to provide an argument for restricting immigration as a means of facilitating income redistribution.
The literature on income redistribution in a common labor market has analyzed other methods for encouraging jurisdictions or countries to redistribute income. As modeled by Wildasin (1991), the "rich" possess altruistic preferences that induce them to redistribute income to the "poor", but the public good aspect of income redistribution results in its undersupply. In particular, income redistribution is a public good in the sense that more of it in one jurisdiction benefits all jurisdictions by raising the well-being of the mobile poor in all jurisdictions. Indeed, the amount of income redistribution goes to zero as the number of jurisdictions goes to infinity, since a single jurisdiction cannot affect the incomes of the mobile poor in this limiting case. Wildasin's solution to this problem is a system of matching grants on the subsidies that jurisdictions give to the poor. A properly designed system of grants does indeed support an efficient equilibrium, though Raff and Wilson (1997) emphasize the informational problems that a central government would encounter if it tried to implement such a grant system. Second-best grant policies of the type considered by Raff and Wilson seem more practical, but even these policies require the existence of a strong central authority. In yet another approach, Dreze, Figuieres and Hindriks (2006) propose a system of voluntary matching grants, based on an adjustment process under which grants are set at each point in time according to information obtained from jurisdictions about their aggregate willingness to pay for each grant. The cost of the grant is also allocated according to willingness-to-pay considerations. This process leaves all jurisdictions better off at each point in time and converges to a stationary solution that is Pareto efficient. This is an interesting idea, though it removes the central government's informational problems by assuming that the jurisdictions are well-informed about each other's production technologies and tastes for altruism.
The introduction of immigration controls in the present article expands the set of income-redistribution policies beyond the subsidies paid to the mobile poor. These controls split the common labor market into multiple labor markets. They are set noncooperatively by individual governments and therefore provide a method for income redistribution when central-government intervention is either not available or not effective.
At first glance, restrictions on immigration might seem to be an anti-redistribution device. However, they remove a major impediment to income redistribution in a common labor market: By restricting inflows of migrants, a jurisdiction can raise the net incomes of its resident poor above the net incomes earned by similar individuals elsewhere. For this reason, the jurisdiction is now able to redistribute income, even in a system of many jurisdictions.
While it is reasonably clear that immigration controls will improve the well-being of those individuals who are allowed to immigrate, it is less clear what happens to those individuals who are left behind in the source countries. Indeed, a major argument against immigration controls concerns the well-being of the latter. To investigate this issue, I examine a system of countries in which income redistribution is motivated by a more elaborate specification of altruism than the one employed by Wildasin. Because migration equalizes the incomes of the mobile poor in Wildasin's model, he is able to model altruistic preferences by including a single income variable for the poor in the utility functions for the rich. In the current article, the degree of altruism is allowed to differ between immigrants and those left behind.
Paradoxically, I find that the use of immigration controls may lead to more immigration. The basic intuition centers on the insight that these controls enable a host country to optimally redistribute income to low-income residents without this redistribution being thwarted by the entry of new residents. But raising the incomes of immigrants to socially preferable levels, the host country then makes immigration more attractive from its viewpoint, giving it stronger incentives to let more of them enter the country. Applying the law of diminishing marginal productivity, this increase in immigration improves the well-being of similar workers who do not immigrate.1
Returning to the EU case, several responses to the concerns raised by EU enlargement have been proposed, including the differential treatment of immigrants, based on their length of stay in a host country. For example, German welfare benefits might not be fully extended to immigrants until they had resided in Germany for a specified number of years. Sinn (2004) proposes a policy of "delayed integration" along these lines.2 To the extent that new immigrants respond to economic incentives, the denial of benefits should reduce the amount of immigration. But if the immigrants are able to accurately calculate expected lifetime welfare, free migration would still leave immigrants no better off than similar individuals left behind in source countries, in terms of their expected lifetime welfare. In contrast, the current article demonstrates that immigration controls have the potential to increase immigration while making immigrants better off.
Richter (2004) proposes a delayed integration policy under which immigrants remain assigned to their country of origin for fiscal purposes for a given period of time after emigrating. In his model, permanent assignment to the country of origin (the "origin principle") is optimal only in an economy with benevolent governments. Delayed integration responds to the tendency of governments to engage in wasteful expenditures unless they are forced to compete for the tax payments of mobile workers. The current article does not model waste in government, but it does include possible incentives to tax mobile workers, in the form of higher "congestion costs" associated with an increased residential population. Immigration controls then enable governments to limit these costs without the need to tax low-income immigrants as a means of discouraging immigration. To the extent that governments are not benevolent, these controls may discourage efficient behavior in the countries of origin by limiting migration opportunities. However, nonbenevolent host countries may also choose to forgo such controls if they are then able to increase revenue by taxing immigrants.
De Giorgi and Pellizzari (2006) propose a more harmonized welfare system across the European Union, undertaken through the introduction of a uniform minimum income program. Their estimates show that, for a minimum income set at the average of similar programs in the EU, this system would cost about three quarters of what is currently spent on housing and social assistance benefits. As they acknowledge, however, such a system would have distributional effects that would render it politically infeasible if there were no offsetting income transfers across countries. But implementing such transfers requires a level of coordination that is difficult to achieve. Again, the analysis in the current article is motivated by such difficulties. Moreover, uniformity in welfare systems entails inefficiencies, to the extent that different countries possess different preferences towards redistributing income.
The next section of the article describes the model without redistribution activities. Section 3 then introduces decentralized redistribution without controls on immigration. These controls are then investigated in Section 4, Section 5 discusses extensions, and Section 6 concludes. Appendix 1 contains the proof of a key optimality condition.
|
2 The model |
|---|
|
TopAbstract1 Introduction2 The model3 Equilibrium redistribution...4 Equilibrium immigration...5 Extensions and discussion6 Concluding remarksAppendix 1References |
|---|
Consider a system of countries that are initially linked by a common labor market. Following Wildasin (1991), I aggregate all private goods and services into a single aggregate commodity, "consumption", which is produced by competitive firms in each country, using two inputs, internationally mobile labor and an immobile factor.3 For our purposes, it is useful to assume only two types of countries, those that receive immigrants and those that supply immigrants. The former are called the "host countries", and the latter are called the "source countries". Although labor is mobile across countries, it is fixed in supply for the system of countries as a whole, called the "common labor market". In particular, there are a fixed number of mobile "poor workers", each of whom supplies a unit of labor.
Production exhibits constant returns to scale, and factors are paid their marginal products, leaving no other sources of income. But increasing the supply of mobile poor workers in a country lowers their wages, because there is then a rise in the amount of labor per unit of the immobile factor, yielding a diminishing marginal product of labor. In other words, the wage in a host country, w, is a decreasing function of the number of poor residents there: wh = w(nh) for host country h, where nh is the number of poor residents and w(nh) declines with nh. Using uppercase letters for source countries, we similarly have Ws = W(Ns) for source country s. All source countries are treated as identical, so they contain the same numbers of poor workers and pay the same wages in equilibrium. The host countries are similarly identical.
To model income redistribution, let us interpret the owners of the fixed factor in host countries as the "rich", who possess altruistic preferences that lead them to support income transfers to the poor, in the form of a subsidy per worker. Let zh denote this subsidy. For now, this subsidy is uniformly provided to all poor residents, but Section 5 considers the case where natives and immigrants are treated differently. Poor consumption in host country h is ch = w(nh) + zh. I also allow the host countries to be characterized by "external congestion costs", represented by the function, E(nh), which is increasing in nh, reflecting marginal congestion costs. One interpretation is that there is crowding in the usage of public goods and services, the supplies of which are treated as exogenously fixed. These costs will play a role in the determination of zh and subsequent immigration controls. To focus on the income redistribution activities of the host countries, I assume for simplicity that the source countries do not also redistribute income, and I assume away taxes and congestion costs, allowing poor consumption in source country s to be given by the wage: cs = W(Ns).4
To model migration from source to host countries, assume that the number of mobile poor initially residing in source countries is sufficiently numerous to yield an equilibrium where some of these workers immigrate to the host countries (given the chosen income subsidies). This immigration causes wages to rise in source countries and fall in host countries. An equilibrium is obtained when consumption levels cs and ch equal a common value, c:5
|
| (1) |
The specification of altruism takes into account the possibility that the rich view residents and nonresidents differently. In particular, suppose for now that the rich in a given country h care about the well-being of a fixed number of poor, P, but let us allow the intensity of their caring to depend on whether the poor reside within h or outside of h. These assumptions are captured by the following weighted sum of utilities:6
|
| (2) |
, β and 
are welfare weights that take into account differences in preferences towards consumption by poor residents in h, in other host countries, and in the source countries, respectively. By assumption, PS, P–h and nh sum to the fixed quantity, P, which may be interpreted as the total number of poor residents in h and "potential immigrants" to h, interpreted as individuals with ties to h, such as family connections or cultural similarities.
It seems reasonable to assume that P is less than the entire population of poor workers in the common labor market; otherwise, the welfare for a small country would be primarily based on the utilities of the nonresident poor, unless their welfare weights were infinitesimally small. But the results that follow assume only that there remain some potential immigrants in the source countries after the equilibrium amounts of immigration to host countries occur.7 We shall also see that the weight on resident utilities need not exceed the weights on nonresident utilities, although this is the natural assumption. We do assume throughout that is positive, and we later assume that β is at least as great as , reflecting as great or greater concern for the well-being of other host-country poor compared with source-country poor. This last assumption may be viewed as indicating greater cultural ties and proximity of host countries with one another, compared to host and source countries.
|
3 Equilibrium redistribution without immigration controls |
|---|
|
TopAbstract1 Introduction2 The model3 Equilibrium redistribution...4 Equilibrium immigration...5 Extensions and discussion6 Concluding remarksAppendix 1References |
|---|
Let us start with the case of no immigration controls. The host governments then play a Nash game in subsidies, with the government of country h choosing subsidy zh to maximize welfare Vh, given the subsidies chosen by the other governments of host countries. For zh to solve this optimization problem, a marginal change in zh must leave welfare Vh unchanged, taking into account its impact on immigration. A unit rise in zh attracts
nh additional more poor workers into country h, lowering the wage until consumption levels are equalized across jurisdictions. If country h is sufficiently large, the resulting drop in source-country populations will have a noticeable positive impact on the marginal productivities of poor residents remaining there, causing their wages to rise. As a result, the common consumption level obtained by the mobile poor will rise throughout the common labor market. Let c denote this change in consumption, and let PS and P–h denote the resulting (negative) population changes that occur in the source and other host countries as workers are induced to move into h. At the optimum, the resulting change in country h's welfare must equal zero, giving us the following optimality condition (derived in the Appendix 1):
|
| (3) |
Thus, increasing zh has two effects. The first term gives the welfare effects of the resulting rise in poor consumption, c. While the poor inside and outside the country benefit, there is also the direct cost, nh, of raising each poor resident's consumption one unit. Second, the rise in zh leads to a higher nh, which directly affects the weighted sum of utilities by shifting weight towards resident utilities. If the welfare weights are all the same ( = β = ), then this effect is zero. But the rise in nh also affects welfare through any difference between the tax (minus subsidy) imposed on mobile workers and the marginal congestion cost, –zh –ME. Setting the sum of these welfare effects equal to zero gives the optimal nh for country h.
Let us focus on the "small country" case, where the number of countries in the common labor market is sufficiently numerous for no single country to be able to have an important influence on the common consumption level received by the poor. Then the first of the two terms in (3) becomes negligible, leaving the second term to be equated to zero. For future use, we may rewrite this condition more simply as follows:
|
| (4) |
nh = 1), and f–h and fs represent the fractions of any marginal inflow of immigrants that originate from other host countries and the source countries, respectively (so f–h + fs = 1). If there were no altruism, then each worker would pay a tax equal to the marginal congestion cost. In symbols (4) would become, –zh = ME. This marginal-cost-pricing rule is standard in the local public economics literature. But if altruism is added to the model, with more weight placed on poor residents than nonresidents (i.e. exceeds β and ), then zh can turn positive, reflecting the desirability of subsidizing the resident poor, even in the absence of market power in the common labor market. Indeed, we could conceivably obtain a case where there is overprovision of income redistribution in a first-best sense.8 In what follows, I rule out this case by assuming that the Nash equilibrium among governments is characterized by the underprovision of income redistribution in the sense that giving another dollar to a poor resident raises welfare more than the cost of this dollar: MU > 1. We can ensure that this inequality holds in equilibrium by making the marginal congestion cost, ME, sufficiently high, since high congestion costs lead to the taxation of poor residents in a common labor market, thereby effectively redistributing income from the poor to the rich. |
4 Equilibrium immigration controls |
|---|
|
TopAbstract1 Introduction2 The model3 Equilibrium redistribution...4 Equilibrium immigration...5 Extensions and discussion6 Concluding remarksAppendix 1References |
|---|
Assume now that in addition to providing subsidies to poor residents (zh for country h), host governments may also use immigration controls as a strategy variable, defined as an upper bound on nh. My first observation is that these controls will indeed be used. Suppose instead that the host countries continued to compete using the zh' s alone. Then any single country would benefit by unilaterally fixing its nh at its current level and using its subsidy zh to raise the poor consumption ch above the level offered elsewhere. With nh fixed, there would be no change in the behavior of any other country, and so this redistribution would be purely lump sum. Since the level of redistribution is assumed to be below the first-best level in the absence of immigration controls (
MU > 1), the ability to redistribute income to the poor without inducing an inflow of new immigrants must raise country h's welfare.
Thus, the equilibrium is characterized by immigration controls. With these controls, each host country redistributes income to poor residents until the social marginal benefit equals the marginal cost. In symbols,
|
| (5) |
First-order condition (3) remains the rule for how many immigrants to admit, except that now the optimality of the consumption of poor residents, ch, implies that there is no welfare impact from a marginal change in this consumption [i.e. nhMU–nh = 0 in (3)]. Moreover, when nh rises, there is no change in consumption in other host countries, since they too are restricting entry of immigrants, thereby insulating them from changes in h's immigration policies Thus, all that is left in the first of the two terms in (3) is the change in consumption within source countries, which is positive because a rise in host country h's immigrant population attracts residents away from these countries, increasing the marginal product of labor there. But if, we consider a host country without significant market power in the common labor market, then this change in cs can be ignored. In this case (3) reduces to
|
| (6) |
By comparing the two optimality rules, it is possible to show that immigration controls may actually lead to more immigration in equilibrium, leaving everyone better off, provided a weak condition on the welfare weights is satisfied. The exact proposition is stated as follows:
Proposition 1. Assuming countries redistribute income to maximize welfare, as defined by (2) with β 
, and the number of countries is sufficiently large,10 then introducing immigration controls leads to a rise in the equilibrium level of immigration, leaving all poor workers better off, regardless of where they reside.
To see why this proposition holds, let us restate (4) and (6) together, for ease of comparison:
|
| (4) |
|
| (6) |
, as assumed. But then the left side of (6) is nonnegative, evaluated at the equilibrium values of variables in the absence of immigration controls [where (4) holds].
With immigration restricted, each host country now has an incentive to increase the subsidy provided to poor residents (or reduce the tax) until the optimality condition, MU = 1, holds. The resulting rise in consumption ch equals the rise in the subsidy, zh. Since MU > 1 until optimality is achieved, u(ch) rises more than zh, increasing the left side of (6). All host countries then have an incentive to raise their immigrant populations until (6) is re-established, while adjusting their subsidies to maintain the optimality condition, MU = 1. Thus, giving host governments access to immigration controls actually causes immigration to rise above the equilibrium level in the absence of these controls.
The basic idea behind this argument may be simply explained. Immigration controls enable a country to optimally raise the consumption of its poor residents, rather than merely cause an influx of new immigrants that prevents poor consumption from being significantly increased. But with immigrants being able to consume at the level that is optimal from country h's viewpoint, an additional immigrant is more highly valued by the host country, that is, the left side of (6) rises. Simply stated, the country benefits more from new immigrants when they are not impoverished, and immigration controls enable the country to eliminate impoverishment. Thus, the country is willing to increase the number of immigrants that it allows within its borders. When more poor workers travel to host countries, the law of diminishing marginal productivity implies that those left behind in the source countries are better off.
We have considered the case where host countries do not possess significant amounts of market power on the common labor market. If the host countries do possess market power, then a single host country is able to raise the incomes of poor workers in the absence of immigration controls. But doing so is accomplished only by drawing more immigrants into the country, thereby improving the well-being of those poor workers left in source countries until they no longer have an incentive to immigrate. Immigration controls enable the country to make its poor residents better off without attracting additional immigrants. In other words, the consumption level of poor residents is no longer positively related to the amount of immigration, eliminating one source of the marginal benefit of immigration. For this reason, incentives to attract immigrants may decline when a large country is able to impose immigration controls. As a result, it appears possible for Proposition 1 to fail in the case of large countries with substantial market power. On the other hand, the large-country case is one where countries have greater ability to redistribute income without immigration controls. The beneficial effects of such controls are stronger in the small-country case, where countries are least able to redistribute income in their absence.
|
5 Extensions and discussion |
|---|
|
TopAbstract1 Introduction2 The model3 Equilibrium redistribution...4 Equilibrium immigration...5 Extensions and discussion6 Concluding remarksAppendix 1References |
|---|
While the formulation of welfare used in this model allows for a wide variety of welfare weights, other formulations give different results. This section discusses one extreme alternative, and then turns to intermediate cases. Finally, I amend the model to allow the host countries to discriminate between natives and immigrants in the provision of income transfers. Doing so provides further support for immigration controls.
To begin, suppose that only a fixed number of host country natives, 
, receive weight in the welfare function, independently of how much immigration occurs, and that poor workers residing outside the host country receive no weight:
|
| (7) |
In this case, condition (4) reduces to
|
| (8) |
Now suppose that immigration controls are implemented, and a positive subsidy is chosen to redistribute income to poor residents (zh > 0), raising poor utilities above those available elsewhere. But under a positive subsidy, the entrance of a new resident lowers welfare, by the amount by which ME exceeds –zh. Thus, the government should reduce the number of resident poor. Doing so raises the wage, w(nh), allowing the subsidy to be reduced without altering the well-being of the poor. To the extent that it can lower nh by reducing the amount of immigration, it will do so. The process continues until either (8) is satisfied, implying a positive tax, or (8) is not yet satisfied but there are no more immigrants to exclude. In the latter case, if there existed host country natives not among the set of citizens that are included in the welfare function, then the host government would want to move them out of the country too, but I assume this is not possible.
The previous welfare function, given by (2), yields such different results because, at the margin, immigrants receive weight in this function. To better understand the difference, let us rearrange the original optimality condition under immigration controls, given by (6), in the same form as (8):
|
| (9) |
= in (9)]. Instead, the critical difference between (8) and (9) is that MME falls below –zh as income is redistributed to poor residents, reflecting their increased value at the margin to the host country. This decline produces the incentives to expand immigration once immigration controls are allowed. In contrast, ME in (8) is not affected by income redistribution, because there are two classes of poor residents: those that receive weight and those that do not, and the latter includes immigrants. Welfare function (2) seems more defensible than (7) from a normative perspective. But an important question is whether the optimization of either welfare function would be the outcome of a reasonable political process. Grossman and Helpman (1994) and others have formulated models in which special-interest politics implies the maximization of a weighted social welfare function. Welfare function (7) seems most applicable to cases where lobbies represent fixed sets of domestic poor residents, who naturally do not want to bring more immigrants into the economy because their presence raises the cost of redistribution, lowering the equilibrium amount. On the other hand, Proposition 1 might apply to the case where some groups of potential immigrants are represented by lobbies, perhaps through their connection to current immigrants.
Between these two extremes is the intermediate case, where immigrants receive some weight but not as much as natives. Suppose, in particular, that natives receive the higher welfare weight of 
> . Then the rule for the optimal number of immigrants remains (6) [or (9)] in the case of immigration controls, but the rule for the optimal level of income redistribution, is altered from (5) to
|
| (10) |
|
| (11) |
u(ch)–zh in (6); the subsidy is too high, implying that the marginal immigrant need no longer be more valuable as a result of income redistribution. If is sufficiently low relative to , then immigration controls will lead to less immigration, harming those left behind in source countries.
These results suggest that it would be useful to model the endogenous determination of the relative weights that natives and immigrants receive in the measure of welfare. One approach would be to model a political process in which immigrants and natives engage in lobbying to influence the government's income redistribution activities. In any case, we may expect the effective weight that an immigrant receives in the measure of social welfare to grow with the time spent in the host country. The fixed values of and may be viewed as capturing within a static framework the outcomes of this dynamic process of assimilation.
On the other hand, changing welfare weights over time suggest the possibility of a time-consistency problem. If a host government could commit to a particular policy, then it could optimally choose its present and future immigration restrictions, along with income transfers (which might vary across time). With a time-consistent policy, it would not matter whether the government committed to future policy choices now or made its decisions at a later date. In the absence of unforeseen events, policy choices made later would be consistent with choices for the same future time period made now. But government decision-makers and their constituencies change over time, and future decision-makers may possess different objectives functions than current decision-makers.
For example, a government might allow a relatively small number of immigrants to enter the country initially, based partially on their low productivities and low weight in the government's welfare function. But as these immigrants start participating in the political process (perhaps indirectly through the lobbying efforts of supporters), then their weights in the welfare function might increase (a rise in ), inducing a relaxation of future immigration restrictions. Recognizing its inability to maintain future controls on immigration, a current government might then over-restrict immigration, or even eliminate it, in contrast to Proposition 1.
Related concerns arise when there is return migration, whereby immigrants return to their home countries after working in host countries for various periods of time. In the static model developed here, immigration controls restrict only the total supply of immigrant labor, but the model can be extended to account for restrictions on both entry into the host country and length of stay. In practice, there could be special provisions for return immigration, specifying the number of temporary immigrants with various attributes, recognizing that allowing temporary immigration benefits both source and host countries. If immigration were thought to be truly temporary, then a country might be willing to admit a relatively large numbers of immigrants, with the goal of obtaining the "gains from trade" with them, consisting of the exchange of labor for goods, without the perceived burdens associated with the process of cultural assimilation and its effects on the country's own cultural attributes.11 But once present in the economy, large numbers of immigrants, and their children, would potentially affect the political process, perhaps leading to permanent status. If a country recognized its inability to fully commit to a temporary-immigration program, then it might again respond with overly restrictive immigration standards.
Returning to the current model, note that once the welfare function discriminates between natives and immigrants, the government also has an incentive to discriminate in its provision of income transfers, providing a lower subsidy to poor immigrants than to poor natives. But by uncoupling the transfers received by natives and immigrants, we now have separate optimality conditions for these transfers, and the transfers received by immigrants again satisfy (5). Moreover, the previous rules for the optimal level of poor residents, given by (3), remain valid, but with nh now representing the number of immigrants, and u(ch) denoting their utility. Note, in particular, that there is no change in the natives’ consumption level in these rules, because the subsidy provided to natives can be raised to offset any fall in the wage resulting from the influx of new immigrants.
Once discrimination between natives and immigrants is allowed in the provision of income transfers (or in-kind benefits), governments have weaker incentives to reduce the size of the "welfare state" in response to unrestricted immigration. In other words, a country can maintain a large welfare state without inducing large numbers of workers residing in other countries to immigrate there with the goal of obtaining sizable income transfers. This discrimination between natives and immigrants naturally benefits natives at the expense of immigrants, providing potential immigrants with weaker incentives to immigrate.12 The results reported here suggest, however, that even in the face of this weaker immigration incentive, immigration controls may still have value. By restricting immigration, the government can raise the well-being of those immigrants who are allowed to immigrate to levels above those available outside the source countries (adjusted for migration costs). Following the previous argument, we then see that these immigrants become more valuable to the host country, and so the country is willing to expand their numbers.
In a dynamic context, the different treatment of natives and immigrants described above corresponds to the use of "delayed integration" policies suggested for dealing with migration within Europe, whereby immigrants are not provided with the full benefits of the host country's welfare system until they have resided there for a specified period of time. The results reported here suggest that immigration controls still have value, even when coupled with delayed integration policies. But their value is reduced, since free migration no longer has potentially catastrophic consequences for the maintenance of the welfare state.
To the extent that immigration is rationed, there remains the issue of how this rationing is accomplished. Although the theoretical model developed here abstracts from differences in immigrants, such differences would likely form the basis for immigration controls in practice, including skills, relatives already in the host country, and nationality. The mix of attributes used in practice would again depend on the government's objective function. In addition, the current analysis does not address the costs of rationing, including enforcement costs, and the implications of illegal immigration.
|
6 Concluding remarks |
|---|
|
TopAbstract1 Introduction2 The model3 Equilibrium redistribution...4 Equilibrium immigration...5 Extensions and discussion6 Concluding remarksAppendix 1References |
|---|
The preceding analysis has demonstrated that allowing countries to restrict immigration may not lead to substantial reductions in migration relative to the levels occurring in a common labor market. In fact, Proposition 1 presents reasonable conditions under which immigration controls actually lead to an equilibrium in which more immigration occurs, and in my model, more immigration means that those left behind are better off because their productivity rises as their numbers decline. The basic idea behind Proposition 1 is that countries are able to use these restrictions to raise the consumption levels of immigrants, and under the appropriate altruistic preferences, immigrants with higher consumption levels are more attractive to natives and are therefore allowed to be more numerous. If immigration were unrestricted, the government's inability to raise the incomes of migrants would lead it to offer them little or no income transfers, or even to tax them, and fewer of them would choose to immigrate.
Although Proposition 1 provides conditions for the uncoordinated use of immigrations restrictions to improve welfare, both in host countries and in source countries, the use of these restrictions still does not produce a fully efficient equilibrium. In the case of unrestricted migration, Wildasin (1991) demonstrates that income distribution is "underprovided" from an efficiency viewpoint, when modeled as a "public good" for all countries within the common labor market. Proposition 1 describes how restricting immigration reduces this underprovision problem, but it does not eliminate it, due to the public-good nature of altruism. In particular, raising the net incomes of a given jurisdiction's immigrants benefits not just that jurisdiction, but also other jurisdictions. Thus, we continue to have underprovision of income redistribution, though not the zero provision that would characterize a system of many jurisdictions. Moreover, there are now separate inefficiencies associated with the choices of income transfers and immigration levels, both of which now serve as strategy variables in the policy game played among host countries. It is interesting to note here that allowing "foreign aid" might lead to more restrictive immigration policies, since it provides an avenue by which host countries can aid source-country natives, in addition to allowing them to immigrate. On the other hand, remittances might conceivably lead to less restrictive immigration controls, to the extent that they substitute for foreign aid. These issues deserve further study.
I have discussed a number of alternatives to Proposition 1, emphasizing the sensitivity of the results to the treatment of immigrants in the welfare function, along with problems associated with committing to a particular immigration policy that may no longer be optimal in the future. These concerns suggest that further research would be needed before embracing immigration restrictions. But the results suggest that the advantages of unrestricted migration may not be as clear-cut as currently thought.
|
Appendix 1 |
|---|
|
TopAbstract1 Introduction2 The model3 Equilibrium redistribution...4 Equilibrium immigration...5 Extensions and discussion6 Concluding remarksAppendix 1References |
|---|
Derivation of Equation (3)
Country h's optimization consists of choosing zh to maximize Vh, given by (2). With free migration, equilibrium is obtained only when all poor workers receive the same consumption, regardless of residence:
|
| (A.1) |
|
| (A.2) |
|
| (A.3) |
|
| (A.4) |
|
| (A.5) |
|
Acknowledgments |
|---|
I am grateful to the editors of this issue, David Wildasin and Thiess Buettner, two reviewers, and Peter Birch Sorensen for helpful comments and suggestions.
|
Footnotes |
|---|
1 Our analysis assumes that those who do not immigrate are paid their marginal products, net of any social costs they impose on the economy (e.g. costs associated with public good usage). If these poor workers also received income supplements from their home governments, then additional migration would reduce the number of workers receiving these supplements, perhaps leading to more generous supplements.
2 Specifically, he proposes that the new EU constitution be amended to read, "Every person migrating legally from one EU country to another for the purpose of working there is, in principle, entitled to social security benefits and social advantages of the host country. However, during an initial waiting period the host country may limit tax-financed benefits. Non-employed persons must direct claims for social assistance at their home countries regardless of ... their country of residence" (p. 703).
3 The assumption of perfect labor mobility is made to highlight the difficulties that individual countries encounter in redistributing income. Adding mobility costs to the model would increase the ability of individual countries to redistribute income.
4 More generally, congestion could be added without altering the analysis, if W(Ns) were redefined as the wage net of efficient congestion taxes. Recall footnote 1 for a discussion of further extensions.
5 By equating private consumption levels across locations, I am abstracting from other factors that might influence migration decisions.
6 For simplicity (2) has been specified as quasi-linear, thereby eliminating income effects. In my working paper, Wilson (2006), a more general form of the welfare function is considered, allowing welfare to depend nonlinearly on the numbers of poor workers in the given host country, other host countries, and source countries.
7 Which particular mobile workers emigrate from a particular country is not determined by the model and may therefore be specified by the given host country. For example, a given host country h may include in its welfare function only nonresident poor residing in a couple of source countries and a couple of host countries. Only these individuals would be allowed to enter country h. But for a symmetric equilibrium, the set of migration restrictions imposed by all host countries must imply that each source country's population declines the same amount as a result of immigration. This assumption could be relaxed at the cost of a more complex analysis.
8 If ME were zero, and = 1 but β = = 0, then (4) would give zh = u(w + zh). If zero income implies zero utility [i.e. 0 = u(0)], then the only way for this equality to hold under the assumption of diminishing marginal utility is for the marginal utility at w + zh to fall short of one: MU < 1. This means that transferring a dollar from rich to poor residents, while holding fixed the number of the latter, actually reduces welfare, as measured by Vh in (1). But the assumption that β = = 0 implies that the rich exhibit altruism only towards the resident poor.
9 This claim concerns net changes in immigrants. In some cases, country h might be able to attract additional immigrants from other host countries, but they would then be replaced by new immigrants from other source countries, leaving the total numbers of immigrants in the other countries fixed at the levels determined by immigration controls. I assume that only net changes in immigrant populations matter for the computation of welfare.
10 This condition can be satisfied by increasing either the number of host countries or the number of source countries.
11 Verbon and Meijdam (2008) construct a model that includes the disutility that immigrants impose on natives as a result of cultural differences between natives and immigrants. They also model the gains from trade resulting from immigration, in the form of less rationing of the "services" produced by immigrants when their prices are fixed, or lower prices of services when these prices are flexible. In the case of fixed prices, the natives choose to restrict immigration to the point where services are rationed.
12 In our basic model, immigration controls may be implemented as a form of discrimination in income transfers, whereby a fixed number of natives and immigrants receive subsidies, but potential immigrants outside the country are informed that they will not receive these subsidies, or might even be taxed, thereby inducing them not to migrate.
|
References |
|---|
|
TopAbstract1 Introduction2 The model3 Equilibrium redistribution...4 Equilibrium immigration...5 Extensions and discussion6 Concluding remarksAppendix 1References |
|---|
-
De Giorgi G, Pellizzari M. "Welfare Migration in Europe and the Cost of a Harmonised Social Assistance". In: IZA Discussion Paper No. 2094. (2006) Bonn, Germany.
Dreze J, Figuieres C, Hindriks J. "Voluntary Matching Grants can Forestall Social Dumping". (2006) Working paper, CORE.
Grossman GM, Helpman E. "Protection for Sale". American Economic Review (1994) 84:833–50.[ISI]
Raff H, Wilson JD. "Income Redistribution with Well-informed Local Governments". International Tax and Public Finance (1997) 4:407–27.[CrossRef]
Richter WF. "Delaying Integration of Immigrant Labor for the Purpose of Taxation". Journal of Urban Economics (2004) 55:597–613.[CrossRef][ISI]
Sinn H-W. "EU Enlargement, Migration and the New Constitution". CESifo Economic Studies (2004) 50:685–707.
Wildasin D. "Income Redistribution in a Common Labor Market". American Economic Review (1991) 81:757–74.[ISI]
Wilson JD. "Protecting the Welfare State from International Migration". (2006) Michigan State University. Working paper.
Verbon H, Meijdam L. "Too Many Migrants, Too Few Services: A Model of Decision-making on Immigration and Integration with Cultural Distance". Journal of Population Economics (2008) forthcoming.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||