Inflation forecasts and methodology


Inflation forecasts and methodology

Central banks have embraced a policy framework known as inflation targeting for forecasting of the future inflation. This approach is usually characterized with the announcing of inflation targets and explicit acknowledging stable and low inflation as the ultimate long-term objective for the monetary policy. Important features for the inflation targeting is transparency of policy that is in existence and is increased by communication and being clear on the plans as well as the objectives of policymakers. Also it requires increased accountability in the central bank objectives.

Inflation targeting argues that an approach may not be operational due to time, that monetary policy affects inflation and also difficulties in forecasting inflation. These problems are the reasons for policy making and therefore should target money and exchange rates as well as other variables that are more directly controlled (Kohn and Robert, 309).

Intermediate targets for money and also exchange rates are stable or not stable on long-run objectives. When they do so, they are used with other information in targeting long run inflation.

The rate of inflation is fundamental determinant for the discount rate that is used to calculate the investment present value. Changes in the rate of inflation affect market valuation of stock. There are various ways to construct and forecast this inflation rate and also the rationale of the methodology.

Consumer price index (CPI) relies on simple analysis but more technical in forecasting the inflation rate. The technical analysis mean that prediction using Consumer Price Index that has passed and is related to an inflation rate data than prediction using economic data e.g. trends of commodity prices and employment wages( Bernanke and Mishkin, 122).

The inflation rate of a given month for the next 12 months is more related to the past inflation rate in behavior for that month than the inflation rates for other months past.

Inflation rate momentum for the forecast is more important than the reversion. This means that inflation rate does not undergo reversion on its trend quickly on long term and short time changes in Consumer Price Index which are indicative of future changes and not changes in Consumer Price Index on a distant past.

Political cycle and fiscal policy are significant on rate of inflation behavior. This means inflation rate should use at least 4 years rate in history and also consider their increment.

Recent research shows that the accuracy of different methods of inflation forecasting using future consumer price index varies with time and technique (Bernanke and Mishkin, 134).

Forecasting using CPI requires construction of various different forecasts of a year before the annual CPI inflation by use of variables and method. Variables are CPI inflation, measures of inflation and economic activity and also the inflation expectations. These may be obtained through survey. Our study involved constructing many different forecasts of one-year-ahead annual CPI inflation using a number of variables and methods and then comparing the accuracy of those forecasts. Variables included CPI inflation, core measures of inflation, measures of economic activity, and inflation expectations obtained from surveys (Bernanke and Mishkin, 143).

Practitioners mostly rely on empirical models of inflation forecasts. These include the time series and cross section which are interesting in assessing the content of money on inflation.

Time series model

This uses autoregressive model (AR) that relates the current inflation, constant and its lags. The lags used are lag polynomial autoregressive and moving average coefficients lag. For comparison purposes it is useful ton consider bi-variate vector autoregressive (VAR) method of inflation (Kohn and Robert, 310).

Dynamic factor model

An alternative empirical method is the general dynamic factor model (GDFM) which is employed in forecasting using large section of data. The content of information on a large set of variables is represented by the small set of indicators called common factors which is in a dynamic setting. Consequently composite indicator is constructed and used to forecast the inflation (Kohn and Robert, 330).


Empirical findings on forecast combinations are on the lecture notes and suggest that simple combinations are not easy to beat this due to the importance of estimation error in parameter on combination weights. These methods are used to reduce these errors. This is in shrinkage or combination methods in which they ignore the correlations that exist between the forecasts. Forecasts that are not based on the model leads to poor performance .Trimming and clustering of models using similar forecasting performance can yield considerable improvements in performance mostly where large numbers of forecasts are used. Also shrinkage method on simple combination of forecasts weights improves the performance of the method.

Works cited

Bernanke and Mishkin, Inflation Targeting: A New Framework for Monetary Policy. Journal of   Economic             Perspectives, 11, (1997), Print

Kohn and Robert. When is an aggregate of a time series efficiently forecast by its past. Journal    of Econometrics, Elsevier, vol. 18(3), 1982. Print.