Most empirical analysis of property development treats the sub-components of the construction industry as independent of each other. For example, models of housing construction typically do not consider any possible relationship with industrial or commercial construction. New building and repair and maintenance are rarely modelled jointly. But, in fact, there are a number of ways in which changes over time may be interdependent. For example, since similar labour skills are required across the sub-sectors and labour supply is not perfectly elastic, expansion in one sector might impose constraints on others. Furthermore, economic theory suggests that, under some conditions, housing investment crowds out industrial and commercial investment in a general equilibrium framework. In general, therefore, if there are any interdependencies, the presumption is that the relationship is negative. In this paper, we test the interdependencies between the sectors, concentrating particularly on the relationship between new housing and industrial construction. We find that, although in the short run the expected negative correlation occurs, this is not true in the long run, based on Johansen tests for British data since the sixties. Indeed, in the long run, the relationship is positive ñ movements in the two are complementary. At first sight, this result is counter-intuitive, at least in an aspatial setting. However the result can be explained once a spatial dimension is added to the analysis, taking account of firm location decisions. Weak exogeneity tests indicate that, for the industrial sector, îjobs move to workersî rather than ìworkers moving to jobsî as standard residential location theory might suggest. Therefore a positive relationship occurs between housing and industrial construction, particularly in southern England, as newly-forming and relocating firms seek out highly skilled workers who, in turn, seek out high quality housing locations.