This method is popular in the United States. In contrast to the hedonic regression and the mix-adjustment methods, which use large numbers of property characteristics, the repeat sales model observes the price development of a specific house over a period of time. By observing the price developments in a number of houses during overlapping time periods, general inflation trends can be inducted.

Example: One group of properties was sold in 2001, and then again in 2003, for prices 10 per cent higher on average. Another group was sold in 2002 and again in 2003 for 7 per cent higher prices. Thus, we can deduce that prices between 2001 and 2002 rose by about 3 per cent.

Problems

• Because the inflation rates of different property types (e.g. detached and terraced houses) might diverge, a shift in the sample data towards a certain property type can influence the estimated inflation rate. If, for example, the prices of detached houses rise by 5 per cent whereas the prices of terraced properties do not rise, a shift in the sample data towards detached houses will show a higher inflation because the characteristics are not taken into account.
• This method also ignores whether a house changes in condition or is altered, such changes will affect value and thus the index.