Changes to the methodology for surveying inflation expectations
Inflation expectations play a vital role in an inflation targeting system. A central bank operating such a system attempts to steer the inflation expectations of consumers, producers and other economic agents towards its inflation target. If expectations are close to the target, it becomes much easier to hit that target. Inflation expectations also play a pivotal role in the CNB's inflation forecast, and the central bank pays close attention to the results of statistical surveys of inflation expectations.
The CNB started regular measurement of inflation expectations among financial market analysts in May 1999. In June 1999 the survey was extended to businesses and households. The monthly statistical surveys of 15 selected domestic and foreign analysts gives the CNB an idea of inflation expectations on the financial market. The quarterly surveys of households are based on a large and randomly selected sample of 600 households and give the CNB an idea of consumers' inflation expectations. The survey of businesses is also performed quarterly, covering 120 managers of major businesses from all sectors of the economy. This gives the CNB a rough picture of producers' inflation expectations. Over time, the CNB's information requirements have changed and several methodological changes have accordingly been made to the surveys in recent years. The most important are as follows:
- financial market participants are only asked about the annual consumer price inflation index they expect at the 12-month and 3-year horizons; a question on future net inflation was dropped in June 2001;
- the number of financial market analysts surveyed has been gradually increased from 10 to 15;
- the business and household sector respondents are asked about their inflation expectations at the 3-year horizon as well as at the 12-month horizon; a question on year-end inflation was dropped in January 2003;
- businesses and households are no longer informed about the latest inflation figure during the surveys (dropped for businesses in June 2002 and for households in March 2003);
- up until March 2003, outliers in the responses of businesses and households were not excluded; expectations were computed as the arithmetic mean of all the answers. Since June 2003, 5% of the maximum responses and 5% of the minimum responses of businesses and households have been removed and average expectations have been computed as the arithmetic mean of the remaining 90%.