Determining inflation forecasts is a complicated procedure that not all economists agree upon. The Consumer Price Index (CPI) is often used as the measure of inflation. Forecasts are typically given in the short versus the long term. Short-term forecasts tend to rely on historical inflation rates, while long-term forecasts consider other economic factors. Inflation forecasts are inherently risky, as many events that affect the inflation rate are essentially unpredictable.
Before forecasting inflation, a method of measuring inflation must be established. In the United States, the CPI is the most common measure used. The CPI works by keeping track of a variety of basic goods and services, such as food and medicine. Each item in the calculation is weighted based on its impact on cost of living. This way, the CPI can keep track of how these costs change over time—a common definition of inflation.
Short term is often considered to be from several months to one year, while long term is ten years or more. In the short term, forecasts might focus heavily on the techniques of technical analysis. Technical analysis uses past data—in this case, past CPI inflation rate data—without much speculation on future events. More recent data, such as data from the previous few months, is generally weighted more heavily than distant past data.
Proponents of technical analysis as a means of determining inflation forecasts cite the unreliability of other methods. They claim that historical inflation forecasts that tried to incorporate economic trends such as employment have erred in their predictions. They argue that momentum in an economy can have an important effect in the short term, and that this momentum is best detected in the trend of inflation itself.
Inflation forecasts longer than a few months can benefit from adjusting to seasonal variations. Historically, inflation rates in the US have been lower from May to July and from November to December than during other months of the year. This has been a very regular cycle, so economists are highly confident it will repeat itself in the future. Forecasts that take this into account are said to be seasonally-adjusted.
Long-term inflation forecasts tend to use methods of fundamental analysis. This involves broadly considering factors in the global economy and geo-political landscape. For example, the power of large central banks can be significant for inflation rates. Moreover, inflation is highly dependent on the political stability of an economy.
There will always be uncertainty with inflation forecasts because certain events are beyond the predictive scope of economists. Natural disasters, for example, can cause prices for basic building commodities to rise. Likewise, stock market crashes are practically unpredictable. Finally, a large holder of currency reserves, such as China, can suddenly release money into the economy, which has the effect of expanding the money supply and increasing inflation. With these possibilities in mind, inflation forecasts are best considered as practical guidelines in the case that world events go as planned.