The Determinants of Relative House Prices: a reply

the sample data used in the regression analysis reported in his paper (1971) on which I commented in a review article (1973) on house price studies. As data for different years was apparently not used, obviously my statement about possible biases was invalid. He also raised a point about Table 11, which summarised the principal results of the studies surveyed. Wilkinson felt that the sample size given for his analysis should have read 2,708, the sum of the 1960 and 1964 data whereas, as he notes in his comment, the results shown are only based on 1960 data and a sample of 557. To put 2,708 would once again create the impression that data from the two years had been combined. Turning to the more general theoretical comments, the relevance of theories postulating a negative relationship between accessibility and house prices for empirical work in the U.K. was questioned. It was argued by Wilkinson that certain conditions of household behaviour and city structure, which form necessary assumptions for the theory, were unrealistic in the context of the housing markets of British cities, with the possible exceptions of London and Bristol. In order to examine the relationship between prices and travel costs, resi-dential location theory has followed the usual procedure of simplifying the real world situation and most of the literature initially deals with a model in which the only factors entering housing decisions are price, housing space and accessiblity costs. In reality, many other attributes of housing will exist which will influence relative prices. The situation posed by the theory is not that these other attributes are unimportant but that accessibility to workplaces and other activities will affect willingness-to-pay and the residential locatio