Rogers-Castro Migration Schedule

Because moving is tied to life-course events that themselves cluster by age. Infants move because their parents do, school-age children are relatively immobile, and the late teens and twenties concentrate leaving home, education, entry into the labour market, and family formation, producing the sharp labour-force peak. Retirement and later-life care needs create the secondary humps. Rogers and Castro showed that this life-course logic yields a curve of stable shape across societies, which is precisely why a single parametric family fits so widely.

Let the data decide. Use the seven-parameter schedule when only the childhood decline and labour-force peak are visible. Move to nine parameters when there is a clear retirement hump around the sixties, and to eleven parameters when there is also an upturn at the oldest ages reflecting moves into dependency or care. Adding components you cannot see in the rates risks overfitting and unstable estimates, so favour the most parsimonious version that reproduces the genuine features of the profile.

They serve two roles. First, the smoothed schedule itself provides clean age-specific rates for projection or further analysis, free of sampling noise. Second, the parameters translate into interpretable indices — the mean age of the labour-force peak, how sharply it rises and falls, and the spacing between childhood and labour curves — that compress a whole age profile into a few comparable numbers. These let you say, for instance, whether young-adult moves peak earlier or are more concentrated in one population than another.