Measuring potential GDP is inherently difficult. As we have demonstrated in class, in the short run, the outcomes we observe in the economy are based on short-run supply and aggregate demand, and these variables may move in ways that are counter to the long run ability of the economy to generate output (i.e. potential GDP). But all is not lost. There are a number of variables that are helpful in working out what is happening to potential output:
- the labour market. If the gap between the unemployment rate and the natural rate is falling, then the actual level of output will be moving closer to potential output. This is not entirely informative, since the natural rate of unemployment is unknown, and varies through time (see previous post).
- investment. If firms are increasing their level of investment, then potential GDP should be increasing, as workers have more capital to work with.
- productivity. If the level of output per worker is increasing, the rate of increase in productivity times the rate of growth in the labour force should give us a good estimate of changes in potential output.
- capacity utilisation. There are various ways of trying to measure the extent to which firms are operating at their maximum capacity.
In general, we have a better idea of potential output after the fact. Partly this is because some economic data (for example investment and productivity) is only released with a lag, so it is impossible to know how these are varying in real time. But there is an additional reason: with time we simply have more data that might be informative for determining potential output today, for the following reason.
We normally think of potential output as changing relatively smoothly over time (at least more smoothly than actual output), and moving towards potential output, as firms and workers adjust wages/prices/expectations in response to shocks. If workers and firms are rational (in that they set prices and wages based on all available information), then the difference between actual output and potential output is due to shocks that could not be anticipated by workers and firms. Then the level of actual output should be equal to the level of potential output on average. So if we put a smooth line through actual output, we can argue that we have an estimate of potential output. There's no exact science to doing this, but the most common method involves the HP filter (http://economics.about.com/library/glossary/bldef-hodrick-prescott-filter.htm).
If we want to construct a smooth line over some data to work out the level of potential output today, we only have the data up to today to use. As time passes, we will have more data points, and so the level of that smooth line today will be estimated more precisely.
No matter how we construct potential output, it will always be an imperfect measure of the ability of the economy to produce output.