Example of precisely inaccurate information

January 4th, 2008

Share |

Carrying on from yesterday’s article, we will give an example of highly precise numbers that may not necessarily be highly accurate as well. As we said yesterday in Confusion between precision & accuracy,

In another example, are the precise CPI figures that are being released monthly very useful? Given that we distrust price indices, we do not see the point of attaching too much significance into the minute details of only one month?s reading. It is the job of the professional economists and analysts to be highly precise in their measurement of numbers. Whether these numbers are accurate or not is completely another matter.

If you notice, the growth in the general price levels (price index, e.g. CPI) is usually expressed as a percentage figure that is precise up to one decimal point. We can argue that this is an example of a highly precise number. But are these figures accurate? As we explained in How much can we trust the price indices (e.g. CPI)?,

But this idea of price index is flawed and is easily subject to manipulations.

Now, if the mainstream economics? definition of inflation is flawed, then the whole idea of the price index is flawed too. Therefore, no matter how much ?enhancement? that statisticians and economists can think of to improve the ?accuracy? of the price index, it is an exercise of futility.

Therefore, beware of price indices! It is not as ?scientific? as you think!

The price index is a very important number. It is used to derive real GDP growth from the nominal GDP growth. From the growth (or contraction) of real GDP, we can then define when an economy is technically out of (or in) recession.

Now, if the price index is a logically invalid number (let alone accurate), then how accurate will real GDP growth figures be for capturing the growth of output of an economy? If this figure is inaccurate, then how accurate will it be for defining when an economy is technically in recession? In that case, how useful will it be to be so precise in defining the exact point for which the economy is in technical recession?

Today, there is a lot of talk in the financial markets about the probability of the US entering a recession. But really, what is the point of drawing a line in the sand to mark out whether the US is in recession or not? Here, we are more concerned about being vaguely right (accurate) on where we are in the business cycle then to be precise on which side of the recession line we are in. As investors, being accurate is far more important then being precise.