The land-value surface in suburban Washington, D.C., changed dramatically o
ver the decade of the 1980s. This article explains these changes in terms o
f the decentralization of jobs versus socioeconomic trends. Contemporaneous
correlation among selected variables needs to be controlled with reduced f
orms and SES techniques. But all explanatory variables except distance from
some unchanged point are determined simultaneously. Predetermined variable
s control for this double-endogeneity issue.
Land values in 1990 have a U-shape with respect to distance from the U.S. C
apitol Building after controlling for other variables. The data indicate th
at this is the result of demographic changes rather than the development of
suburban employment nodes: polycentric SUE theory is rejected. Land values
are an increasing function of lagged land values, a decreasing function of
work at home. Moreover, work at home is attracted by low structural densit
y and high socioeconomic status as well as low land values. This supports t
he argument that demographics and technological innovations have shaped the
land-value surface; baby boomers are seeking low-density housing for work
and family life.