Entrepreneur: Start & Grow Your Business

Migration, fixed costs, and location-specific amenities: a hazard analysis for a panel of males.


by Huffman, Wallace E.^Feridhanusetyawan, Tubagus

Since Schultz' important paper on human capital (Schultz 1961), human migration has been an important topic of economic research. Much of the human capital literature on internal migration has emphasized expected net earnings benefits as the major factor driving human migration decisions, e.g., see Schwartz (1976), Schlottmann and Herzog (1981), Herzog and Schlottmann (1984), Sandefur (1985), Pissarides and Wadsworth (1989), and Detang-Dessendre and Molho (1999). A number of recent studies have also used aggregate data on migration, but aggregate data contain considerable migration against the economic gradient and are hard to interpret, for example, see articles by Deller et al. (2001), Huang, Orazem, and Wohlgemuth (2002), and Hunter, White, and Little (2003). Models underlying these studies assume that an individual moves in the first period in which he is made better off at a new destination than at the origin, but this approach ignores the timing decision, i.e., an individual should move when the payoff from migration is at a maximum rather than when it is first positive. Also, these studies ignore individual heterogeneity that can bias parameter estimates. Moreover, fixed costs and location-specific amenities are important to migration decisions, and they have received less attention (Greenwood 1997). Exceptions, however, are Mueser and Graves (1995), Deller et al. (2001), and Hunter, White, and Little (2003).

This article presents econometric estimates of the adult working-age male hazard function of interstate migration. The hazard function is the conceptually correct statement of the migration decision, i.e., it gives the probability that an individual moves at a given point in time, given that he has not yet moved. As such, it easily accommodates the migration timing decision. Furthermore, we include individual heterogeneity in the model to reduce omitted variables bias. The empirical hazard rate model is derived for an individual who has a finite-life and consumes leisure, purchased goods and local amenities, and incurs significant fixed costs of moving. The econometric hazard model of migration uses data on the length of an individual's resident spells, and the spells for this study are obtained from following a sample of adult males for a twenty-year period. These males were first interviewed in 1968, when they were nineteen to forty-five years of age, and hence, after twenty years of migration experience, the oldest males are sixty-five years of age. This is a relatively large amount of migration experience, given that we are primarily interested in the migration behavior of economically active males rather than males contemplating immediate retirement. The econometric results show a strong negative effect of men's real wage difference between origin and destination and of fixed costs associated with a move, and a positive effect of the local crime rate, a disamenity, on the hazard of interstate migration. Farmers and other self-employed men who own above-average location-specific assets have an unusually low hazard rate of interstate migration compared to wage earners.

The story unfolds in the following sections. First, a very brief summary of the economic problem is presented. Second, we present the econometric model and data, and third, we present the empirical results. The final section contains conclusions.

The Conceptual Model of Internal Migration

Males are born into a particular region, move with their parents until they are eighteen years of age, and then are assumed to make independent decisions about their residence for the remainder of their life. We assume an adult male receives utility from consuming his leisure time, purchased goods, and local amenities. Local amenities, representing location-specific culture, climate, topography (e.g., parks, access to the sea, mountains, plains), and environmental conditions (crime, pollution, congestion) are a type of local public good to an individual. An adult male chooses between staying at his current residence, the origin (o), or migrating to a new area or destination (d), and he is uncertain about future real wage and amenity outcomes at these locations. Let his expected indirect utility function for each year be [V.sub.j]([w.sub.jt], [x.sub.jt]), where [w.sub.j] is the expected real wage and [x.sub.j] is an indicator of the expected local amenities in location j,j = o, d (Greenwood 1997, pp. 668, 677). Let all expected relocation costs associated with moving from o to d in t, except for the foregone earnings, be represented by [c.sub.dt], and to simplify, assume that [c.sub.dt] is fixed and invariant with the distance moved. Also for simplicity, assume that local amenities and relocation costs can be measured in real wage units.

An adult male is assumed to choose a residence that gives him maximum utility. To further simplify, assume that he is risk neutral, that migration does not affect his length of remaining life, and ignore discounting. He then migrates when the summation of net real benefits is a maximum, provided that the summation is positive. (1) Let this maximum be:

(1) -[n.summation over (t=1)]([w.sub.ot] + [x.sub.ot] - [c.sub.d1] + [n.summation over (t=1])([w.sub.t] + [x.sub.dt]) [much greater than] 0

where n is his number of remaining years of life.

In equation (1), clearly [[summation].sup.n.sub.t=1]([w.sub.ot] + [x.sub.ot]) < [[summation].sup.n.sub.t=1]([w.sub.dt] + [x.sub.dt]) and [[summation].sup.n.sub.t=1]([x.sub.ot] - [x.sub.dt]) < [[summation].sup.n.sub.t=1]([w.sub.dt] - [w.sub.ot]). Hence, over n the summation of the difference in expected value of amenities between the origin and destination is less than the sum, over n, of the difference in expected real wage at the origin and destination. However, only if [[summation].sup.n.sub.t=1]([x.sub.ot] - (x.sub.dt]) = 0 can we say for certain that the expected real wage difference between the origin and destination will be positive. For example, if a destination gives an individual higher expected amenity value relative to the origin, it may be lifetime-utility-maximizing for him to migrate from o to d even, when the accumulated wage difference after migration is negative. (2)

Define D as an indicator variable taking a value of 1 if [NG.sub.t] = -[[summation].sup.n.sub.t'=t+1]([w.sub.ot] + [x.sub.ot]) - [c.sub.d1] + [[summation].sup.n.sub.t'=t+1]([w.sub.dt] + [x.sub.dt]) is a maximum (and positive) and 0 otherwise. Then the probability of an adult male migrating from o to d can be represented by the following probability statement:

(2) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

The following comparative static results hold:

(3) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

(4) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

(5) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

Hence, the probability of an adult male migrating from o to d is increasing in his destination wage [w.sub.d] and amenity [x.sub.d], and decreasing in his wage and amenity at the origin. It is also decreasing in the fixed cost of migration.

In a static environment, a strong economic incentive exists for a finite-life individual to migrate from o to d early, perhaps in the first period, or to stay at his current location. However, given that a move from o to d is completed, it is possible for additional moves to be optimal.

The Econometric Model and Data

At any point in time, each adult male is located at some residence location (o). Furthermore, as time in a particular place increases, he accumulates information on local conditions (knowledge of social networks, local culture, local businesses, local personal friends, and local amenities) that can be expected to strengthen his ties to a place and reduce mobility. However, if a change occurs in family composition such as a child leaves home, the local real wage declines, or if distant economic conditions improve, he may choose to move to a new location (d). Therefore, we expect the migration hazard, i.e., the probability that the resident spell ends at this time, to be a function of covariates (Greenwood 1997) and to be time dependent. Also, we can incorporate individual heterogeneity into the model. Because the migration hazard can accommodate these features, the hazard analysis of migration reveals a rich picture of interstate migration relative to a discrete choice model. (3)

The Hazard Model of Migration

Unlike discrete choice models of migration, the hazard rate model of migration treats the length of the resident spell as the dependent variable. For a given individual, define T as the duration or length of time that he has resided at a particular location, and t as a particular realization of T. T has an associated cumulative distribution function F(t) and probability density function f(t). An adult male's hazard of migration is represented as the limiting probability that a resident spell is completed at t + [DELTA], given that it has lasted until time t, or:

(6) H(t) = [lim.sub.h[right arrow]o][P.sub.r](t < T [less than or equal to] t + hIT > t)/h

= f(t)/[1 - F(t)] = f(t)/S(t)

where S(t) = [P.sub.r](T > t) is the individual's survival function for his current location. S(t) expresses the probability that a resident spell is of a length of at least t (Kiefer 1988; Lancaster 1990; Greene 2003, pp. 790-94).

We expect an adult male's hazard of migration to be time dependent and depends on a set of covariates, X. This is a so-called accelerated failure-time model (Greene 2003, p. 769). In addition, individual heterogeneity may exist because of (i) individual-specific unmeasured effects, e.g., intensity of psychological costs of moving, (ii) measurement error in X, or (iii) measurement error in the duration of a resident spell. Following others, for example, Heckman and Singer (1985), we impose the Weibull distribution on the density function for residency at a particular location (t), which permits constant, decreasing, or increasing time dependence of the hazard function for migration determined by the sign of [sigma]. If [sigma] is one, then the hazard of migration is not time dependent. We define [upsilon] to be individual-specific unmeasured heterogeneity, and assume [upsilon] is distributed gamma with unit mean and variance [theta]. It is incorporated as in Heckman and Singer (1985) or Greene (2003, pp. 797-798). Hence, a male's mixed resident-survivor function is

(7) S(t, X, [beta], [sigma], [theta]) = [{1 + [theta][[t exp(X[beta])].sup.1/[sigma]}.sup.-1/[theta].

His associated hazard function of migration is

(8) H(t, X, [beta], [sigma], [theta])

= [[S(t, X, [beta], [sigma], [theta])].sup.[theta]] (1/[sigma])[t.sup.(1/[sigma])-l][[exp(X[beta])].sup.1/[sigma]]. (4)

If [sigma] is not significantly different from zero, the hazard of migration is monotone in duration. An important feature of this specification is that the effect of heterogeneity is increasing in [theta], but as [theta] goes to zero, heterogeneity vanishes. (5)

Some variables in X for the ith individual, say [X.sub.ij], change over time and are jointly determined with duration. For example, an individual who has children and chooses a place to reside may make a joint decision. When this is the case, [X.sub.itj] is typically assigned its value at the beginning of the resident spell, say [X.sub.ij0], (Lancaster 1990; Greene 2003, pp. 790-800). Other variables are time varying but not endogenous to duration, e.g., marital status, and actual value during the spell can be included as a regressor.

The Data

Individuals in this study are working-age adult males of the Panel Study of Income Dynamics (PSID). We use the Survey Research Center (SRC) Sample, which was a probability sample of all U.S. households in the contiguous forty-eight states in 1968 and not the Survey of Economic Opportunity sample, which drew mainly from low-income households. Critical to this study, the PSID identifies the state of residence. These males were first surveyed for the PSID in 1968 when they were nineteen to forty-five years of age. Males were surveyed annually starting in the year after they completed school and were reinterviewed annually until they retired, died, or disappeared. We have data on twenty years of migration experience for 915 men where 193 of them had at least one move in the twenty-year period; 10.6% moved once, and the remaining 10.5% moved more than once (table 1).

From the full number of males, we derived a total of 865 resident spells having known starting dates, and 1,268 residence spells that have adjusted starting dates. The 1,268 open and "closed" spells are the total number of observations in our econometric hazard rate of interstate migration analysis. The adjustment closes the migration interval when it is open on the left, i.e., when we do not have data on the starting date. This provides a relatively large amount of information on resident-location decisions of working-age adult males and variation in length of resident spells. However, the oldest-aged males are sixty-five years, and some of them have retired or are contemplating retirement. (6) The PSID has major advantages over cross-sectional micro-data sets on migration, because we have twenty years of information on migration decisions, and for the most part we know the individual's attributes at the start of each resident spell. Data for these adult males are supplemented with data from other sources for their resident area.

For working-age males, internal migration over a long distance is generally associated with a change in employment, whereas short-distance migration is frequently associated with a change only in residence or housing. Since the latter is not of interest to us, we choose to define migration decisions for adult working-age males as interstate moves, which is the most frequently used area designation for migration studies in the U.S. States have fixed geographical boundaries over time and are exhaustive in their geographical coverage. Occupational licensing practices and union membership policies are set at the state level, and the state government is a major fiscal and jurisdictional authority. (7) Some important amenity benefits of states include the quality of local public goods, such as public schools and park areas, air and water quality, crime rate and "mildness" of the climate. Much of the important variation in topography is also captured at the state level. Other migration studies that have analyzed interstate migration include Sandefur (1985), Brinig and Buckley (1996), Clark, Knapp, and White (1996), and Pashigan (1979). (8)

For our study, an adult male is defined as a long-distance migrant if he moves across a state boundary. Although states differ in area, population, employment, and number of cities, we ignore these differences. Adult males and families may be affected marginally by the distance of a move to a new destination, but a major part of the cost is fixed. Costs of moving for an adult male with a family include: time spent weighing the decision to move, finding new places to shop, finding schools for kids, finding a church to attend, and finding and making new friends. They also include the cost of selling a house or ending a lease at the origin, finding a new house or apartment at a destination, and the cost of loading one's possessions at the origin, and unloading and putting them in order at the new destination.

Since destinations and their location-specific amenities are a choice, all potential working-age males face a similar common fixed amenity-effect of potential destinations. Hence, when they live in different places, the local amenity attributes that differ among them, and that is relevant to their decision to migrate is the amenity attributes of their current location (Huang, Orazem, and Wohlgemuth 2002).

The PSID identifies the state of residence of each adult male in the survey We focus on a twenty-year period, starting in 1968.

The Empirical Hazard Function of Migration

The systematic part of the hazard function of interstate migration is specified as:

(9) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where i = 1, 2, ..., n, denotes sample males, s = 1, 2, ..., [C.sub.i], denotes the resident spells, W([ot.sub.is]) denotes the real wage for ith male at his current location or the origin o in time t during spell s, W([dt.sub.is]) denotes the real wage for the ith male at a new destination in spell s, AGE is the age of the ith male at the beginning of spell s, EDU([t.sub.is]) denotes the years of schooling completed by the ith male at the beginning of spell s, DSLFEMP([t.sub.is]) denotes an indicator for the ith male being self-employed or being a farmer at the beginning of spell s, UNION([t.sub.is]) denotes an indicator for the ith male being a labor union member at the beginning of spell s, UNEM([t.sub.is]) denotes the ith male's unemployed status in resident spell s, MARR([t.sub.is]) denotes the ith male's marital status in spell s, CHILD([t.sub.is]) denotes the ith male's number of children at the start of spell s, CRIME([t.sub.is]) denotes the crime rate at the origin for the ith male relative to the U.S. average in spell s, PARKS([t.sub.is]) denotes for the ith male the share of origin or of the resident state that is in state and federal parks relative to the U.S. average at the start of spell s, JAN([t.sub.is]) denotes for the ith male the thirty-year average January temperature at the origin or resident state relative to the U.S. average at the start of spell s, JULY([t.sub.is]) denotes for the ith male the thirty-year average July temperature at the origin or resident state relative to the U.S. average at the start of spell s, and DWHITE([t.sub.is]) denotes the ith male's race.

The set of covariates is largely defined as the beginning of a resident spell, especially for those variables that seem most likely to be jointly determined with the probability that a residence spell ends--EDU, DSLFEMP, UNION, and CHILD. (9) For other variables that are time varying, but that are not jointly determined with the probability of a spell ending, we use a summary of the variable over the resident spell, for example, MARR. (10) Also, we choose to incorporate the effect of unemployment on the probability of a resident spell ending as the average amount of time in the spell that the individual was unemployed. (11)

Expectations about the signs of the [beta]'s are as follows. If a working-age male's real wage increases at his current location [W([ot.sub.is])], this will reduce the attractiveness of an interstate move, other things equal, and reduce his hazard of interstate migration. An increase in his real wage at a new destination [W([dt.sub.is])], other things equal, will increase his hazard of interstate migration. Although there is a stronger incentive for young adult males to migrate than older preretirement males, the hazard of interstate migration may not peak at a young age for working males (Greenwood 1997, pp. 655-656). Because an adult male's life is finite and the life cycle plays an important role in the timing of his human capital investment and family decisions, the marginal effect of his age [AGE([t.sub.is])] on the hazard of interstate migration may be nonlinear. Thus, [AGE.sup.2]([t.sub.is]) is also included as a regressor. We expect [[beta].sub.3] > 0 and [[beta].sub.4] < 0.

The next seven variables are associated with a working-age adult male's cost of migration. A male who has more education is expected to have a higher hazard of interstate migration, other things equal (i.e., [[beta].sub.5] > 0). Other studies have shown that an individual's education is associated with the ability to acquire and process information, and to make efficient decisions (Schultz 1975; Schwartz 1976; Huffman 1977), and this seems likely to extend to interstate migration. Moving to a new location carries some uncertainty, and additional education, which aids the acquisition and interpretation of information, can greatly reduce this uncertainty. Moreover, Detang-Dessendre and Molho (1999, 2000) have shown that an individual's schooling increases his/her hazard of migration.

If a working-age male is a farmer, he has control over farmland now and has knowledge of the local land market, but it may be difficult for him to acquire farmland in a new location. If he is a professional or trade association member, his income is linked to a local clientele base that also ties him to his current location (Pashigan 1979; Goss and Paul 1990). If he is a union member, he frequently has nontransferable seniority and pension rights. These attributes of an individual's local occupation are expected to increase his utility at his current residence relative to a new location and to reduce his hazard of interstate migration (i.e., [[beta].sub.6], [[beta].sub.7] < 0). When a working-age male experiences temporary unemployment, this reduces his opportunity cost of searching for a new location (Herzog and Schlottmann 1984; Pissarides and Wadsworth 1989; Goss and Paul 1990), and we expect his hazard of interstate migration to increase (i.e., [[beta].sub.8] > 0).

Spouses can either reduce or increase interstate mobility, depending on their current job and amenity matches. Having children is an irreversible decision, and children only know what they have experienced. Hence, for an adult male having school-aged children is expected to increase the psychological and monetary costs of an interstate move (Mincer 1978; Greenwood 1997, pp. 701-705) and to reduce the likelihood of a resident spell ending, e.g., [[beta].sub.10] < 0.

Local amenities at the origin, relative to those at a new destination, are expected to affect a working-age male's hazard of interstate migration (Greenwood 1997, pp. 676-677). We focus on the crime rate, land in state and national parks, and long-term normal July and January temperatures. A higher local crime rate against persons and against real property is a negative local public good for non-criminals, imposing psychic and self-protection costs and lowering the residents' utility, other things equal. Hence, we expect a larger value of CRIME to increase a working-age male's hazard of interstate migration, i.e., [[beta].sub.11] > 0, or to reduce resident duration. In contrast, local area parks provide a positive local public good, and can be expected to reduce a male's hazard for interstate migration (i.e., [[beta].sub.12] < 0). Long-term average January and July temperatures play an important role in home utility bills and affect the types of winter- and summer-season outdoor recreational opportunities that are available in an area. Some effects of weather, which have both cost of living and amenity dimensions, should be incorporated into local wage rates and create compensating differentials across states, but other effects may not (Roback 1988). If there are other effects, then [[beta].sub.13] and [[beta].sub.14] will be statistically different from zero.

White working-age males are expected to have a higher hazard of interstate migration, other things equal, than nonwhites (i.e., [[beta].sub.15] > 0), because discrimination against nonwhites limits the number of destinations where they can expect to be made better off by a move relative to their current location. In particular, Filler (1992) has shown that whites in the United States have greater opportunities for superior location moves than nonwhites. His finding suggests white males, on average, will experience a shorter duration at any location than nonwhite males, other things equal.

The Wage at Potential Destinations

An adult working male's expected real wage at a new location is an important variable for the hazard of interstate migration. It reflects a location choice that is endogenous, however, and is not generally available. Hence, we include a proxy variable for this opportunity wage (Wooldridge 2002, pp. 63-67, 83-86; Greene 2003, pp. 86-88). This proxy variable should have the property that it is correlated with the "true" but unobserved wage, but uncorrelated with the error term in the decision to migrate. Otherwise it should not be part of the structural model. If we assume that the U.S. interstate labor markets for males are approximately in equilibrium, then a hedonic wage equation fitted to data on male wages pooled across states provides a method for valuing individual and location-specific attributes (Hoch and Drake 1974; Kenny and Denslow 1980; Roback 1982; Rosen 1986; Topel 1986; Tokle and Huffman 1991). In particular, state labor market units provide valuable information about the parameters of the male-wage offer equation.

Following Kenny and Denslow (1980), Adams (1985), Topel (1986), and Tokle and Huffman (1991), we adopt a real hourly wage equation where differences in the cost of living are being adjusted for over time using the implicit price deflator for personal consumption expenditures (U.S. Dept. of Commerce) and across geography using state average land prices, normal January and July temperatures, extent of urbanization, and the crime rate:

(10)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where [EDU.sub.iy] denotes years of schooling completed by ith male in yth year, [EXP.sub.iy] denotes labor market experience computed as the ith male's AGE - EDU - 6 in years, [RACE.sub.iy] denotes an indicator for the ith male being white, [PLAND.sub.ky]/[P.sub.y] denotes the real price of land in state k where the ith male resides in year y, [URBAN.sub.ky] denotes the proportion of the population in the ith male's resident state k that is urban in year y, [CRIME.sub.ky] denotes the crime rate in the ith male's resident state k in year y, [JAN.sub.k] denotes the thirty-year average January temperature in the ith male's resident state k, [JULY.sub.k] denotes the thirty-year average July temperature in the ith male's resident state k, [PJOBGR.sub.ky] denotes predicted job growth in the ith male's resident state s in year y, [PURATE.sub.ky] denotes the predicted unemployment rate in the ith male's resident state s in year y, [RSHOCK.sub.ky] denotes the relative employment shock in the ith male's resident state s in year y, [RURATE.sub.ky] denotes the residual unemployment rate in the ith male's resident state s in year y, and [DS.sub.ky], [DW.sub.ky], and [DNC.sub.ky] denote that the ith male resides in census region South, West, or North Central, respectively. TIME and TIME-squared are included to allow for a possible long-term negative trend in male real wage rates (Blundell and MaCurdy 1999). (12) The [[zeta].sub.j]s are unknown parameters, and [[epsilon].sub.iky]' is a zero mean random disturbance term.

In equation (10), the real price of land ([PLAND.sub.ky]/[P.sub.y]) is a key proxy for local housing prices, and [URBAN.sub.ky], [CRIME.sub.ky], [JAN.sub.k], and [JULY.sub.k] are proxies for various local amenity attributes. By including the regional fixed effects and a time trend, we control for other omitted variables, and by doing so, our parameter estimates in equation (10) should have better statistical properties.

Sample mean values of the variables entering the hazard of interstate migration and the real wage are reported in tables 2 and 3, respectively.

The Empirical Results

This section reports the empirical results for the wage equation and for the hazard function of interstate migration.

A Male's Hedonic Wage Equation

The wage equation (equation (10)) is fitted to all of the observations for the 915 adult males in the PSID, pooled over twenty years, starting in 1968. Results are reported in table 3. The performance of the fitted male wage equation is generally in agreement with results reported in other studies. A one-year increase in a male's schooling increases his real wage by about 7.5%. An increase in his experience has a positive effect on his real wage, but at a diminishing marginal rate. The maximum effect of EXP on a male's wage profile occurs at twenty-six years of experience (approximately forty-five years of age). All other measured variables held equal, white males earn 11% more than nonwhites. These results are consistent with those reported by Neal and Johnson (1996), but lower than those by Topel (1986).

Male real wage rates are shown to differ because of local cost-of-living and amenity differences. Both the real price of land (PLAND/[P.sub.y]) and congestion, as reflected in URBAN, have positive effects on the real male wage rate. The land-price effect on the male wage compares favorably to the findings of Kenny and Denslow (1980) and Tokle and Huffman (1991), but the effects of URBAN are greater in this study. Roback (1982) does not include a variable that is a good proxy for the cost of housing in her earnings equations, and this may color the interpretation and comparison of those results. However, Roback (1988) includes the local cost-of-living index in some of her wage equations.

Local amenities have significant effects on the male real wage. A one percentage point increase in a state's crime rate increases the male real wage by about 1.5% (significant at the 1% level), consistent with findings reported by Roback (1982). Roback (1988), however, did not find statistically significant effects of the crime rate in her wage equations, with one exception. The impacts of climatic amenities on the male wage are conditioned by the inclusion of regional dummy variables. However, a higher average January (or July) temperature for an area, holding the region constant, reduces the male real wage by 4 (17) percent per 10 degree increase, suggesting positive amenity value for higher values, other things equal. Roback (1982) includes weather variables in her earnings equations and finds that heating-degree days, snowfall, clear days, and cloudy days have statistically significant effects. She, however, largely excludes weather variables in her 1988 article (Roback 1988). (13)

State labor-market characteristics have a significant effect on male wage rates. A one percentage point increase in the anticipated job growth rate for the resident state (PJOBGR) increases the male real wage by 6.0%, and a one percentage point increase in the anticipated unemployment rate for the resident state (PURATE) increases the male real wage by 4.3%. This reflects the fact that over the long run, he must be compensated for taking the higher risk of being unemployed. The latter result is greater by a factor of three than those obtained by Topel (1986) and Tokle and Huffman (1991). The signs of the coefficients for unanticipated job growth (RSHOCK) and unemployment (RURATE), which cannot be factored into long-run decision making, are consistent with the results in Topel (1986) and Tokle and Huffman (1991).

Historically, the U.S. has had broad regional differences in real wage rates, and our results show that they persist over the twenty-year period of our data, even after controlling for land prices, urbanization, crime rates, and climate. Compared to the Northeastern region, the male real wage in the South is 8.9% lower and 13.7% lower in the West. However, the male real wage rate in the North Central region is not significantly different from that in the Northeast region. Consistent with other evidence, the results show a statistically significant negative trend in the male real wage rate of slightly less than 1% per year.

The Hazard of Interstate Migration

The empirical hazard function of adult male interstate migration is fitted to the data on resident spells over a twenty-year period starting in 1968, which is a large amount of information on migrat