New methods for construction of historical price indices, with an illustration from Norway, 1777–1920*

This paper reviews some methodological and practical problems encountered in the construction of historical price indices. The underlying data sets in such studies are often characterized by heterogenous and incomplete price series. It is shown that by using the repeat sales method for constructing the subindices for individual commodity groups, some of the main problems can be overcome. The procedures are illustrated by material from the construction of monthly price indices for Norway from the year 1777 to 1920. The price indices shed new light on two great wartime inflationary episodes in Norway: 1807–17 and 1913–20. In spite of a 61-fold increase in the price level in the first period and a four-fold increase in the second, it is found that, after inflation had been brought under control, prices reverted to a level consistent with the purchasing power parity principle.