International Economic Geography Migration dr hab. Bart Rokicki Chair of Macroeconomics and Foreign Trade Theory Faculty of Economic Sciences, University of Warsaw
What are the motives for migration? Responding to neoclassical signals Enhance welfare Seek better jobs Individual versus family decision-making (intergenerational factors) Amenity based
Classical Model Assumptions Perfect competition in all markets Classical production function (CRTS) Factor migration costless and no barrier to movement Factor prices are perfectly flexible Factors of production are homogeneous Owners of labor and capital have perfect information (in all regions)
Equilibrating Effects of Interregional Labor Migration Real wage W 2 South s Labour Market S S 1 S S Real wage North s Labour Market S N W 1 W 1 Real wage L 2 L 1 D S D N Employment Employment Real S S 1 wage S N 1 S S 2 L 1 S N W 2 W* W* W 1 L 2 D S L* Employment L* L 1 D N Employment
Equilibrating Effects of Interregional Labor Migration Assume disturbance to labor supply in SouthRegion declining birth rates for example. Wage rate increase to W 2 in Southand a real wage differential emerges. As a resultlabor will move from Northto South. Labor supply curvewill move to the right in South(and to the left in North). Hence, labormigration removes wagedifferential and the economyreturnsto an equilibrium.
Interregional migration the UK case Net migration flows 1991-98 South West South East London Eastern West Midlands East Midlands Yorkshire and the Humber North West and Merseyside North East Northern Ireland Scotland Wales -500-400 -300-200 -100 0 100 200 300 Net migration in thousands
Interregional migration the US case
Classical model - conclusions Classical model oversimplifies role of income differentials (especially as many households now have >1 wage earners) Need to differentiates between autonomous migration (e.g. request of a company) as opposed to anticipatory moves that may be of two types: o Speculative ocontracted Former likely to require higher income differentials than the latter
Classical model conclusions(2) Researchers tend to make a distinction between odecision to move o Choice of destination Not clear role of income differences vis a vis other factors (amenities, better quality of life, climate, and stage in life cycle or family issues)that include both pecuniary and non-pecuniary benefits For example, differences in regional amenities may lead to persistent equilibrium differences Employment opportunities may also play a significant role Even the adjustment process described by neo-classical model may not work because of differences in skill requirements We may observe persistent unemployment differentials even with differences in real wages (thoughevidence in the UK that unemployed are 1.8 times more likely to migrate than those employed)
Costs of migration Costs of Migration include pecuniary costs (e.g. selling house in say a difficult market) and non-pecuniary (leaving friends, community environment etc.) In developing economies role of assets and credit constraints may play a vital role Lower flow levels of migration in developing countries than developed countries Aroca, Hewingsand Sonis(2002) explored a second stage of migration, the movements of the labor force among regional labor markets since today the urban population in Latin America averages close to 90 percent. Theyreveal that the migration mechanism is efficient when workers have access to borrowing from financial institutions; without this access, migration is shown to be inefficient from the point of view of the market equilibrium.
Migration Decision-Making Process Given the shortcomings of the neoclassical approach Aroca, Hewingsand Sonis(2002) migration model takesinto account decision-making process. For any time period, we can separate the labor force between: othosewho want to stay where they work othosewho want to migrate. For the latter, there are three potential outcomes oworkerswho migrate as soon as they make the choice because they can afford the moving cost (MC) oworkerswho have to save money until they can afford the MC oworkerswho never can migrate because they can never afford the MC.
Migration Decision-Making Process(2) Need to clarify the role of assets and credit constraints in the decision to migrate from an efficiency point of view Focus on the realization time of the workers who made the decision to migrate (movers). View the lack of assets and the credit constraint as intervening obstacles (Lee, 1960) in conditioning the market efficiency of migration in developing countries. In addition, different types of migration will be considered (e.g., speculative and contracted, following Molho1986) Also consider the behavior of workers facing different employment status, and also different levels of wealth in an imperfect capital market context.
Migration Decision-Making Process(3) The modelwill offer insights to explain why unemployment rate differentials among regions are not the equilibrium ones. It will also show that an inefficient migration process could make poor people worse off One of the major differences between developed and developing economy migration: o financial ability of migrants to move oquality and quantity of information that circulates about the relative advantages of two or more locations in the context of the choice set. Notwithstanding these differences, the volume of migration in developing economies has continued to rise with a dominating rural-urban flow (see, for example, Becker and Mills, 1986).
Decision-Making model - assumptions Workers make decisions under certainty: oifdecide to move from region i to region j it is because they have found a job in region j(contracted migration). oin addition, they know the salary they will earn. This information allows them to make an evaluation about the profitability of moving. Only employed worker has access to credit market Decision to migrate, assumes ability to afford the moving costs. These can be covered by: oborrowing, odrawing down assets, osaving an amount each period until they accumulate enough money to pay the moving costs
Decision-Making model - implications Assumptions imply that migration mechanism could be inefficient if the worker cannot afford the moving costs. Hence, selective migration may be based on wealth or assets, with a differential regional effect osending Region: The more endowed leave, further exacerbating the status of the economy of these regions (see Vanderkamp1970, for example); oreceiving Region: In the destination regions, the impacts may be positive since along with the additional wealth, the in-migrants may bring embodied occupational capital that further enhances the competitive position of these regions
Unemployed workers When an unemployed worker makes the decision to move from region ito region j, we assume that she or he has found a job. The worker has to assumethe migrationcosts(mc): o transportation costs and o housing price differentials between region iand region j. Traditional neoclassical theory assumes that moving costs can be covered by obtaining a loan from the perfect capital market, an obligation that would be retired with proceeds from future income. Difficult for unemployed worker to access credit markets, so only option is to draw on any accumulated assets.
Unemployed workers(2) Set the migration decision problem of an unemployed worker as a maximization of the intertemporal utility function subject to a constraint where the assets have to be greater or equal to zero Consider income to be wages (may own land) The current consumption is equal to the total consumption minus the minimal consumption that is necessary to survive. This minimal consumption and the income necessary to finance it are not in the model because they cancel each other when we set the dynamic constraint. The implication of this assumption is that we can maximize the utility function subject to the condition that current consumption has to be greater or equal to zero
Time Line PERIODS t 0 I n i t i a l t M M o v i n g T R e t i r e d
Migrant: Unemployed Worker PRE-MIGRATION Intertemporaldiscount rate Interest rate Rate of return on assets+incomeconsumption This first set of constraints describes the saving process in order to reach the necessary amount of money to pay the moving costs.
Migrant: Unemployed Worker (2) POST-MIGRATION WhereI(t):
Migrant Unemployed Worker(3) Consumption path: Assetpath:
Model constraints The first set of constraints describes the saving process in order to reach the necessary amount of money to pay the moving costs. The second set of constraints describes the worker's situation in the new job in the region j. Given the assumption that workers have perfect information, they can evaluate a priorithe profitability of migration -find the optimal path to do so, given the constraint set that they face
Maximum Principle The maximization problem can be solved using the following Hamiltonian: The Maximum Principle implies that c(t) and a(t) will be optimal if there exist non-negative functions λ(t), γ(t) such that:
Role of assets Assume: The unemployed worker has an intertemporal discount rate greater than the interest rate (high values of time preference); Mainmotivation for saving is the sole objective of reaching the necessary amount to pay the moving costs; Therefore, after moving, we can assume that a(t M )= 0. This isdueto the factthatthe onlymotivationfor savingisthe sole objective of reaching the necessary amount to pay moving costs.
No migration case Also consider the case where an unemployed worker does not migrate, even though he or she has found a job in another region. This worker faces the same constraint as the one we described earlier, but his or her previous saving, initial assets, or current income are too low to make the migration possible at the time when it is still profitable.
Non Migrant Unemployed Worker Consumption path: Assetpath:
Employed Worker Assume that the worker will migrate only if he or she has a job in another region, which pays a wage high enough to make it profitable to migrate. Then, the employed migrant worker's problem can be set as:
Employed Worker(2) Difference from the unemployed migrant worker's case: oemployed worker can obtain a loan in order to finance the moving cost ohas larger current consumption, that is reflected by the fact that a(t) can be greater than, equal to, or lower than zero. The unique constraint is that when he or she will retire (at time T), all loans have to have been repaid. If moving is profitable, then the worker will do so as soon as possible. Then we can expect that the t M will be equal to zero. In this case the market is highly efficient in reassigning workers. It could explain the findings that migrants have moved according to market forces, but these movements have not been strong enough to equalize regional differentials.
Employed Worker(3) Employed worker has a continuous path for c * (t)and a discontinuous one for a * (t). Given the fact that a(t)can be negative, then the employed worker uses the capital market to smooth his or her consumption path. The access to the credit market allows workers the possibility of separating the decision of how much to consume from the decision of how much to save. Therefore, the workers will consume more in the present and save more in the future given that ρ> r was assumed.
Migrant Employed Worker Consumption path: Assetpath:
Uncertainty Introduce post-migration uncertainty; the worker does not know whether he or she will obtain a job in the new region. Assume that this probability depends positively on the number of vacancies and negatively on the unemployment rate in the destination region. Furthermore, it is assumed that the worker knows the probability of obtaining a job in the new region. Therefore, in the spirit of the Harris and Todaromodel (1970), the worker will maximize the expected value of his or her utility from consumption.
Uncertainty(2)
Uncertainty and Risk Aversion Also explore the effect on the worker's consumption path and migration decision of the post-migration uncertainty and risk aversion. Find that under post-migration uncertainty, the worker will be less likely to migrate compared with the certainty situation. othis is true because the expected value of the income in the new region always will be lower than the income of an employment worker at the destination region. osecondly, we can see that the larger the probability of finding a job at a destination region, the more likely it is that the worker will migrate
Decision-Making model - conclusions Transport costs play a negative role in the migration decision and their effects are larger the poorer are the people who are contemplating migration; Given the fact that the people most affected by moving costs are the unemployed and poor, the migration mechanism is less effective in depressed regions; The migration mechanism is an efficient mechanism for reallocating people when workers have access to borrowing from financial institutions; Lack of information about the situation in other regions plays a negative role in the migration decision;
Decision-Making model conclusions(2) The higher the probability of obtaining a job at the destination region, the greater the effect on the migration decision; The larger the differential between the employment income and the unemployment income at the destination region, the larger the probability of migration This type of sorting may contribute to explanation of Partridge et al. (2012) findings that migration has declined in US since 2000 Migration now much more selective with lower income/skills workers moving less frequently
Interregional migration in the US Housing Bubble
Why there is a decline in interregional migration? Decline in role of natural amenities as a force of population redistribution post 2000 Continuing role of urban centers Migration was primary labor supply response to asymmetric labor demand shocks before 2000 Post 2000, primary response is change in the local employment rate response of labor force participation rate Increased variability of labor demand shocks result in households being more risk averse about moving National labor market slack offering more local supply Increased labor mobility across industries reduced regulation, reduced unionization and impacts of globalization
Why there is a decline in interregional migration? (2) Decline in military transfers and ageing of US population If response more local then adjustment to shocks may underline need for more place-based economic development policies But do people follow jobs or jobs follow people? Partridge and Rickman (2003) found that labor-demand shocks are more important than migration labor-supply shocks (although the latter account for majority of state employment fluctuations) What if migration not motivated by regional wage differentials? What if wage differentials confined within a limited range? Lim (2011) explored whether wage-differential migration continues to exist. He claimedthatthe main flows of interregional migration caused by wage differentials are more likely to befoundamong the regions where regional economic structures are dissimilar.
Migration and trade Horiba (2000): ointerregional migration and commodity trade are complementary in US owith imperfect labor mobility and imperfect commodity mobility Patterns will be contrasted between the one within the inertia where wage differentials are relatively smaller and the other beyond the inertia where wage differentials are relatively larger. Inertia can represent by the non-linearity between wage differentials and migration as shown in Basile and Lim (2006). Inertia between wage differentials and migration: ointerregional migration does not increase proportionally with real wage differentials between origin and destination regions
Inertia and migration Non-linear Relationship Inertia(persistence) betweenwage differentials & interregional migration Migrants do not move in response to wage differentials within inertia o Low curve inertia log of population migration flows -0.2 0.0 0.2 0.4 0.6 0.8 1.0-0.6-0.4-0.2 0.0 0.2 0.4 0.6 log of wage differential
Issues Confining role of interregional trade on wage differentials between origin & destination Marginal changes of interregional migration in response to wage differentials Comparative studies on the migration pattern considering the regional economic structures Patterns are contrasted between one within inertia vs. the other beyond the inertia
Trade and migration Regional price adjustment with interregional trade 1 2 2 1
Trade and migration(2) A change in the supply is shown with Ain region 2 and with B in region 1. The price gap between regions 1 and 2 has been narrowed from P 12 to P 12' As a result, the gap in nominal wages between regions 1 and 2 will be reduced, reducing interregional real wage differentials inertia Interregional trade plays a role to confine the real wage differentials among regions within a certain threshold level. Key role of transportation costs (NEG) help make this happen
Empirical findings With larger volume of trade from ito j, wage differentials aresmaller Migrationmotivated by wage differentials within inertia is only around 25% of the wage differential driven migrations beyond inertia Within the inertia range, the impact of similarity in regional economic structure measured by economic distance plays the more important role than the wage differentials. Migrants moving from an origin to a destination with the smaller wage differentials where two regions have similar economic structures tend to be more receptive to the marginal change in the wage differentials. Moving to the regions with similar economic structure (74.3%) Moving to the regions with higher wage rates (45.2%) o 42.2% beyond inertia o 3.0% within inertia
Concluding Remarks Similar flow patterns between migration and trade o More flows among the regions with the similar economic structure o This pattern strengthens over time Why more flows among the regions with similar structure? otrade mainly due to the increasing role of IIT o Migration possibly due to the prevalence of contracted migration Within inertia range owith the expectation of jobs moving to a region with relatively lower wage rate especially when the regional economic structures are similar, people rarely migrate Regional economic structures are becoming more similar over time. Similar regional economic structure will result in the decreases in wage differential driven migration