Seven Centuries of Real Income per Wage Earner Reconsidered

Abstract

In 1955, Phelps Brown and Hopkins published their research into the course of real wages in England between 1250 and 1950 (1955, 1956, 1957, 1959). Since other available wage data for the earlier periods generally are of a discursive or a piecemeal nature, their series seeks to fill an important gap. And that gap is likely to increase in importance, if anything. For we are now beginning to see the long span of English demographic and economic history utilized to test various population theories (e.g., Lee, 1973, 1978). When we are confronted with these increasingly exacting uses of the series, its accuracy must come into question. This paper discusses one source of variation in the Phelps Brown and Hopkins index, and presents alternative computations which differ substantially from the original series. Phelps Brown and Hopkins created their index from data on money wages and prices. The price data had to be put into an index form before money wages could be deflated into real wages. They elected to compute a Laspeyres index. To determine the weights that were attached to different kinds of expenditures, Phelps Brown and Hopkins divided products into two general classes, food and non-food. They point to evidence showing that in 1450-1460, in 1790 and in 1904-1911 Englishmen spent between 77.5 and 81 per cent of their income on food (1956, p. 297). It would seem that, since the percentage spent on food did not vary, the composition of output remained unchanged and, hence, the weights for both Paasche and Laspeyres indices are similar. But Phelps Brown and Hopkins also discuss (1957, p. 306) changing relative prices between industrial and food products. As they show, industrial prices fell relative to food prices. This being the case, it is clear that the real quantity of industrial goods consumed rose relative to the real quantity of food products consumed. And since these real quantities changed, the weights assigned to a Laspeyres index differ from those in a Paasche index. Furthermore, not only do the weights differ; the conclusions derived from the two indices also differ substantially. We compute two alternatives to the Phelps Brown-Hopkins index as follows. We begin with changing food and industrial (non-food) prices and with fixed percentages of money income spent for food and non-food purchases. Phelps Brown and Hopkins (1957, p. 306) give us food and non-food price indices for 1401-1700. We extend these indices to 1900. We determine the change in relative prices between 1400 and 1900, and assume that the same pattern of change existed between 1250 and 1400. We then extrapolate that rate of change between food and non-food prices backward from 1400 to 1250. Using this procedure, we find that the food price index divided by the non-food price index in 1900 is 0 31; in 1400 it is 0.93, and we estimate it to have been 1 8 in 1250. Then, taking Brown and Hopkins' price index, and using the relationship we established between food and non-food prices in 1250, we derive the non-food and food price indices for 1250. Using the non-food and food price indices (still assuming that