The Anatomy of Speculation; a Regional Analysis

In an earlier paper we explored the dynamics of speculation in the housing market using national data for England. We argued, firstly, that the increase, in real terms, in house prices over the past thirty years or so has been primarily driven by the shortage of land permitted to be used for development, by the planning system. This meant that the price of developable land had risen more, over all, than house prices. Thus we would expect that the price of dwellings which use land extensively, i.e. bungalows and detached houses, would have risen more than the price of dwellings which use land intensively, i.e. apartments and terraced houses. We also argued that when prices were seen to be changing during a cycle householders not already in owner occupied housing were the only ones who could ëspeculateí by choosing to buy when they saw that prices were rising and choosing not to buy when they saw that prices had stopped rising. This hypothesis was suggested by the fact that apartments, according to the national data, rose in price faster during a boom than other dwellings and fell more in the ensuing down turn. Econometric analysis tended to support both these hypotheses. From data which covered all housing transactions in England and Wales, we obtained price indices for five dwelling types ñ bungalows, detached houses, semi-detached houses, terraces, and apartments. We found that these five indices did not cointegrate, which indicated that they did not tend to converge, a finding which surprised a number of housing market economists, but which tended to confirm, or, rather, did not refute, our first hypothesis. We also found that the price index for apartments ëGranger causedí the other price indices, and that no similar Granger causative relationship existed between the other four price indices, a finding which tended to confirm our second hypothesis. In the current paper we subject the hypotheses to more stringent tests by using regional data for England and Wales. Looking at regional data we are able to explore the possible convergence or divergence of the various regional price indices, and the extent to which particular regional sectors lead others. Previous regional analysis of the UK housing market suggests that the market is led by London and the South East. Our exploration sets out to determine whether the apartments market in this region ëGranger causesí changes in all other regions, or only in the South East, and also to determine whether the apartment prices in other regions ëGranger causeí changes in other sectors of the housing market in their own regions.