In standard economic theory wages are simply prices on the labor market determined in equilibrium by the supply and demand of labor. Looking across space within countries it is however standard to find a relative large variation in local wage levels. At the same time labor is often very mobile again within countries.

For the United States of America the U.S. Census reveals that in 1990 four percent of white males age 25 to 34 moved across state lines, with the cumulative effect that almost a third of this group no longer resided in the state in which they were born.

When individuals move the supply of labor changes and one could expect wages to equalize based on the assumption that the migration itself was motivated by the one state offering higher wages than another. Seen in the light of this it can seem a little puzzling that spatial wage differentials nonetheless continue to exist. What are the main explanations offered by economic theory?

  • $\begingroup$ Many people do not want to move due to family connections, etc. And for people in the lower economic categories moving can be too expensive, especially when you consider the cost of accommodations in the destination. $\endgroup$
    – Hot Licks
    Commented Dec 10, 2020 at 16:26
  • $\begingroup$ @Jesper Hybel Could you be confusing different things? Wages are simply prices on the labor market. How is that comparable to your "space within countries"? If you think it's standard to find a variation in local wage levels, can you Post three or four examples? Why care if the US Census said that in 1990, 4% of 25to-34 white males crossed state lines, cumulating in a third no longer residing anywhere? If you expect wages to equalize due to one state offering higher wages, what other factors do you dismiss? In reality, how could "spatial" wage differentials seem puzzling? $\endgroup$ Commented Dec 10, 2020 at 22:06

3 Answers 3


There are several explanations for this in the literature (the order does not necessarily reflect importance of each explanation).

  1. Different Price Levels: Non-trivial portion of the wage differential is due to different price levels/cost of living. Once you adjust for different price levels and compare real wages gap narrows. This is obvious but since you do not mention in your question whether you are talking about real or nominal wages I thought it is useful to mention that.
  2. Regional Institutional Differences: For example Blackaby and Murphy (1991) argue that inter regional real wage differentials are at least partially due to labour market institution heterogeneity (unionisation and the bargaining system). This applies especially to unions/federations like US where there can be large heterogeneity in regional laws & regulations.
  3. Regional Heterogeneity/compensating differentials: Even within one country the geography is usually heterogenous which creates opportunity for compensating differentials (Farber and Newman, 1989). If you are offered two jobs that are same in all aspects but one job requires you to relocate somewhere to Alaskan wilderness and the other job requires you to relocate to some pleasant state with nice beaches you will likely be willing to accept lower salary from the job in nicer place because you are compensated by the fact that you are living in a nice place. Some regional differences could be also explained by agglomeration economics (e.g. urban vs rural areas have different economic implications), see for example Combes & Gobillon (2015).
  4. Worker Heterogeneity: The wages would only equalize among workers that are homogenous across all work relevant characteristics. However, people are different from each other and often people who are similar tend to cluster together. For example, would it be surprising to find that average wage in silicon valley is higher than average wage in some farming community? This would be just comparison of apples to oranges. Hence looking at average wages can be deceptive and research that takes into account worker heterogeneity find that the wage differentials shrink substantially (see Gerking and Weirick, 1983; Dickie and Gerking, 1987; Farber and Newman, 1989 and Combes, Duranton and Gobillon, 2008).
  5. Regional Differences in Human Capital: Workers are not just compensated for their labor but also get return for their human capital. Human capital endowments might differ across regions (perhaps due to regional differences in quality of education) leading to wage differentials (see Duranton and Monastiriotis, 2002; Motellón, López‐Bazo & El‐Attar, 2011).

I think the above summarizes all the major factors but there might be some further ones as well for that you can see the sources above and ones cited therein.

  • $\begingroup$ '+1' nice answer. Does economies of agglomeration matter? $\endgroup$ Commented Dec 7, 2020 at 15:20
  • $\begingroup$ @JesperHybel yes I think these would fall under the regional heterogeneity (some areas can be considered urban others rural). Also network effects relate to labor heterogeneity. Similar people can locate close together due to network effects - and this intra-regional homogeneity would result in inter-regional heterogeneity. I am sure that other concepts from agglomeration economics might play role as well. $\endgroup$
    – 1muflon1
    Commented Dec 7, 2020 at 15:24
  • 1
    $\begingroup$ The cost of living argument seems very circular. Why is San Francisco so expensive? Because a lot of the people who want to live there can afford to pay a lot. Why do they have so much money? Because they get paid a lot. Why do they get paid a lot? Because it's so expensive to live in San Francisco. Why don't tech companies just move to San Jose? $\endgroup$ Commented Dec 8, 2020 at 17:48
  • 1
    $\begingroup$ @JesperHybel I can’t believe that published literature would suggest that there is really a circular argument there (since that is logical fallacy why would that even be then used as an argument in research). Again circular argument is not the same as saying there is reverse causality or endogeneity, or general equilibrium effects multipliers and what not $\endgroup$
    – 1muflon1
    Commented Dec 9, 2020 at 15:44
  • 1
    $\begingroup$ Those are, I agree, also a valid points. $\endgroup$ Commented Dec 9, 2020 at 15:45

I would add two more, and rather primitive, factors in @1muflon1 long list:

  1. Labor input is not perfectly mobile with respect to the wage. We should never forget that the concept of a utility function allows for arbitrary "goods" to be "utility enhancing". It follows that migrating in chase of a better wage does not always prevails when this may mean that the worker will leave behind some intangible utility-enhancing goods, like for example a sense of belonging in a specific place, or a network of friends, kin and extended family.

  2. Information is not complete nor perfect. Even in today's era of information overdose, "demand for labor" is not what the employers demand, but what the workers know that the employers demand. And not all job opportunities are posted on-line.

  • $\begingroup$ '+1' Good points. Point 1 could perhaps also include more tangible but localized goods such as amenities in a classic Roback (1982) Wages, Rents, and the Quality of Life, kind of thinking? $\endgroup$ Commented Dec 7, 2020 at 17:26
  • 2
    $\begingroup$ @JesperHybel I would include these in the "Regional Heterogeneity / Institutional differences" category. $\endgroup$ Commented Dec 7, 2020 at 17:30
  • 1
    $\begingroup$ Point 1. People who have benefitted from higher education tend to be mobile when job-seeking. People with little academic ambition or attainment tend to stay near family and friends - provided their situation is bearable. If extreme poverty reigns then working-class people will move, given the opportunity. This is most likely to happen where there is no government-provided social security. The phenomenon of complacency due to govt payouts in Britain gave rise to the much discussed "Get on your bike" slogan express.co.uk/news/uk/230675/… $\endgroup$ Commented Dec 7, 2020 at 23:13
  • 4
    $\begingroup$ If that network provides free childcare, cheap accommodation, etc. the benefits may be financially measurable, unlike a sense of belonging - though separating the financial from the other benefits of having family around would be tricky. Information, and the ability to act on it, isn't timely either - notice periods, the need to find a house/school/job for a partner all impose delays. In specialised fields trying to relocate despite the 2-body problem can mean years of double job-hunting. $\endgroup$
    – Chris H
    Commented Dec 8, 2020 at 9:24

tl;dr Even simple homogenous systems don't necessarily equalize across space.

Unbiased systems aren't necessarily homogenous.

Naively, one might assume that if we have a large unbiased system, its members will tend to form one homogenous community. But that's not the case.

For example, consider a pure substance in a glass jar at its triple point. Then there can be three stable phases: one gas, one liquid, and one solid. There doesn't need to be an external bias or forced separation to maintain this; it's often just favored over a single homogenous phase.

Likewise, have you seen oil and water separate after being mixed? Actually, both phases tend to have both water and oil, just in different ratios. Just one phase tends to have a high ratio of water-to-oil while the other has a high ratio of oil-to-water. So it's not a full separation nor single ratio, but rather two different ratios coexisting.

Other answers have made good points about how biases can prevent everything from equalizing out across space. The point here is just that we wouldn't tend to expect everything to equalize out across space even if there weren't biases.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.