Presentation on theme: "Migration Within England and Wales and the Housing Market Gavin Cameron, John Muellbauer and Anthony Murphy Oxford University and Nuffield College, Oxford."— Presentation transcript:
Migration Within England and Wales and the Housing Market Gavin Cameron, John Muellbauer and Anthony Murphy Oxford University and Nuffield College, Oxford RES Annual Conference, Nottingham, 20 April 2006
Modelling Regional Migration Important to understand determinants of regional migration For example, will an increase in housing supply in South East (if achieved) lead to greater migration and thus have little impact on housing affordability? Important feedbacks between regional migration, household formation, tenure choice, earnings, employment and unemployment and house prices.
Gross and Net Migration Focus on gross and net migration between regions (GORs or combinations of GORs) Data from NHS patient register from 1975 to 2003 Data scaled by gross migration between all GB regions and by regional working age population
Theory Utility maximising households choosing location. High relative earnings and employment opportunities encourage (discourage) in- migration (out-migration) and net migration. High relative house prices discourage it. Why?
Different House Price Effects Cost of living effect. However user cost of housing implies expected relative appreciation offsets relative costs. Credit constraints e.g. loan to earnings ceilings. Higher housing stock per head of population is likely to be attractive for in-migration. Captures effects not mediated thru prices of owner occupied houses e.g. other tenures. Downside risk: recent negative returns in a region, relative to other regions, are likely to discourage in-migration.
Contiguity and Commuting Contiguity matters. Near migration more likely than distant migration. Hence near house prices (hp) matter more than far relative hp. However for labour market, commuting – migration trade-off can weaken or reverse role of near versus far alternatives. See Gordon (1975), Mohlo (1982), Jackman & Savouri (1992) and Cameron & Muellbauer (1998).
Construction of Labour and Housing Market Variables Consider relative log earnings (ln earn) and log house prices (ln hp). F 1 (ln earn r ) = (1 - λ 1 conn r )(ln earn r - ln earn GB ) + λ 1 conn r (ln earn r - contig ln earn r ) F 2 (ln hp r ) = (1 – λ 2 conn r )(ln hp r - ln hp GB ) + λ 2 conn r (ln hp r – contig ln hp r ) conn r is a measure of connectivity (ratio of connected boundary to total boundary). We expect λ 2 to be positive and exceed λ 1, which may be negative.
Labour and Housing Market Variables (Contd) The sign and size of the 1 and 2 coefficients reflect the commuting-migration trade-off. Commuting, instead of migrating, is an option when relative labour market conditions in region r are good, but not when relative housing market conditions are good etc. In practise, we use λ 1 = -0.6 and λ 2 = 0.4. Same λs used for all labour and housing market variables.
Migration Equations Migration equations simplify to partial adjustment model with variable speed of adjustment. Speed varies with rate of housing transactions ptran. We include region-specific double fixed effects and time trends. We also allow some parameter heterogeneity in the Greater London and South equations.
Typical Equation Dependent variable is change in migration rate. Include forecast earnings growth and house price appreciation.
Net Migration Results Results in line with theory although relative unemployment works better than employment. The inactive may be less likely to migrate or respond to economic signals? Housing variables – hp level, expected hp appreciation, downside risk and housing stock / working age pop – all significant, correctly signed and of a plausible magnitude. Downside risk is lagged rate of return on housing if negative or zero otherwise.
Net Migration Results (Contd) If levels hp term omitted then earnings incorrectly signed and unemployment borderline significant! (See Table 2 in revised results.) Expected capital gains important factor explaining high migration into South. The estimated contributions of relative earnings and house prices, and the housing stock / working age population vary by region. The former is very important in the North whereas the latter drives net migration in Greater London.
Further Results Equations fit reasonably well. Adjusted R 2 range from 0.58 to 0.91. Some mild autocorrelation. Cross equation and other common restrictions accepted. Estimated parameters on labour and housing market explanatory variables in gross inflow (outflow) equations have the same sign as in net migration equation. (See Table 3 in Revised Results).
Further Results (Contd) Modelled net migration by broad age group - those aged 15 to 24, 25 to 44 and 45 to 64. House prices matter somewhat more than earnings as age increases.
Summary Extend previous research: More data; Hopefully better model (better specification of equilibrium correction); Model Greater London region; Model gross as well as net migration; Model net migration by age group.
Summary (Contd) Economic conditions, in both the labour and housing markets, exert strong influence on regional migration. On the one hand, strong labour market conditions, as exemplified by low unemployment rates and high earnings, draw migrants into regions. On the other hand, strong housing market conditions can prevent movement since commuting may often be an alternative to migration.
Summary (Contd) This can be thought of as giving rise to a migration equilibrium where high house prices choke off migration caused by strong labour market conditions. Expected capital gains in housing, however, can offset high levels of house prices, an effect ignored in previous literature. Migration can also be influenced more directly by the availability of housing relative to population without this being mediated through prices.