The Database

For the most important articles of consumption, long series of institutional prices were collected mainly from two sources: Posthumus' Nederlandsche Prjsgeschiendems (1943/64), and a recently published database of prices paid by the cloister of Leeuwenhorst between 1410 and 1570 (de Moor 2000), which fills some of the gaps left by Posthumus' work.

In the first-generation CPIs, the price with the largest weight was wheat or rye (or malt, oatmeal, etc.). In Holland, especially in the cities, the making of bread was already concentrated in specialized bakeries during the fifteenth and sixteenth centuries, which meant that consumers did not buy grains but bread, of which the price was set by the city. During the seventeenth century, indirect taxation on milling increased sharply, which meant that the margin between the price of rye and the price of rye bread became much larger than it had been in the early sixteenth century.⁷ During the seventeenth and eighteenth centuries, bread prices increased much more than the price of rye (see Table 7.2), and measuring real wages in terms of rye clearly gives a biased impression of their development.

Another major change is the inclusion of a series on rents, an item of expenditure that is missing in all first-generation CPIs. On the basis of a vast database of houses owned by institutions and by the city, Lesger (1986) has constructed chain indices for the development of rents of houses in Amsterdam between 1550 and 1850. His method of constructing the indices has been criticized by Eichholtz and Theebe (1998, 1999). They state that a chain index does not reflect the housing market correctly, because rental contracts were often for periods longer than one year, and only in the years that a contract was renewed did a new market valuation of the property occur (whereas a chain index ‘assumes’ that in the intervening years the rent and therefore the market value remains the same).

Lesger's database of 48,620 yearly rent figures, they used only the information which related to the renewal of contract periods, that is, to actual changes in the observed rents. Using regression analysis they estimated on an annual basis the average change of these rents, and in this way created a repeated-rent index for the 1550—1850 period, which is also by its nature corrected for changes in the quality of the houses. It is possible to compare the long-term changes in these series with the average rents of all houses in Holland in different periods (Table 7.1). The comparison shows that in the very long run the different series developed in a similar way: a very strong increase in rents between the 1560s and 1630s, followed by a much more moderate increase in the next century, then stagnation during the eighteenth century, and even some decline between 1806—8 and 1816—26. The average rent per house rose, however, more than the repeated-rent index, which may be an indication of changes in the quality of the housing stock. The differences between Lesger's chain index and Eichholtz and Theebe's repeated-rent index appear to be rather small, but are nevertheless significant, especially for the early period.

The very rapid increase in rent levels in the period after 1560 is confirmed by other data. In Alkmaar, for which we have a complete and consistent set of data for 1561 and 1632, average rents went up by 428% (van den Berg and van Zanden 1993). The decline of rents during the second half of the eighteenth century is corroborated by Pot's (1994) research into house rents in Leiden. He found a fall of about 5—10% between 1750 and 1808, which is consistent with Lesger's data. These data and the comparison made in Table 7.1 also justify, in my view, the use of the Amsterdam repeated-rent series to estimate the development of rents in Holland.

The other series were directly derived from the two price histories mentioned, although some problems did occur.

Table 7.1 A comparison between the development of rent levels according to Lesger's data for Amsterdam (chain index and repeated rent index) and the average rents per house according to the registers of taxes on real estate, 1560/ 61-1806/08 (1806/08=100)

unweighted average of the series A, B, and C.

Sources. Lesger (1986), Eichholtz and Theebe (1998), data on average rents per house: Soltow and van Zanden (1998: 28, 136; van Zanden 1987: 576).

prices and beer prices seem to have been almost identical). After 1740, the institutions did not buy beer any more; it had been replaced by coffee and tea and to some extent by jenever (Dutch gin). Therefore, after 1740 beer prices were linked to an unweighted index of the prices of coffee, tea, and jenever.

The reconstructed series of the long-term development of prices are in themselves interesting. It appears, for example, that rents increased more rapidly than all other prices, which is what one would expect in a region of rapid demographic growth and urbanization. But beer prices increased almost as rapidly (and even more rapidly than rents between 1550-74 and 1575-99); the heavy indirect taxation of beer helps to explain the strong increase in its relative price. Rye prices increased almost tenfold during the three and a half centuries covered by these data, but the prices of other agricultural products—butter, meat, and peas—went up even more, especially during the century after 1575 when the Dutch economy went through its Golden Age. The relative increase in the prices of butter and meat testifies to a strong growth in the demand for these luxury goods. The growing disparity after 1550-74 between the prices of rye (which is representative of all grains) on the one hand, and beer and bread on the other hand, is probably caused by increased taxation and perhaps by the stagnation of productivity growth in these industries. It is striking that the prices of other industrial commodities and of fish increased less than the

Table 7.2 The development of the most important series (1550/74 — 100)

Sources.

prices of agricultural commodities; especially textiles (and paper, not shown in the table) became much less expensive in the long run.

4.