Paul Walker reacted to my post on inflation and the RBNZ (see here). It is not clear, in his view, that central banks can distinguish relative price changes from price inflation (defined as an ongoing increase in the prices of all goods in terms of money). Relative price changes may impact the CPI in the short run. Thus, by having a price stability objective defined in terms of CPI, a central bank such as the RBNZ may react to relative price changes when it should remain neutral. I agree with Walker's view and I don't think my post implied otherwise. Here is my view of the issue (and those who are specialists of the discipline are welcome to comment).
The CPI is an imperfect measure of inflation, as it may capture, in the short run, relative price changes (which are not to be confused with inflation). However, a pure relative price change will only create a temporary change in the CPI, as it will be followed by a counteracting change once households adjust their spending pattern. However, the impact on the CPI may last for a while (depending on how it is defined). It is also complicated by the fact that households may borrow, which may dampen relative price effects to some extent (as their budget constraints vary). But borrowing should not have any impact if it is based on pure savings transfers. However, if household debt is based on money creation, it will have an inflationary effect until it is serviced (i.e. the reflux). Also, productivity increases may make it difficult to capture changes in relative prices from inflationary price increases, as they improve households’ budget constraints and mask the two effects to some extent.
Another issue with the CPI is the statistical measurement of true inflation. When statisticians tell us that there is one percent inflation, it may be overstated. For instance, the Boskin Commission came up with the view that the CPI in the US was overstating inflation by 1.1%. This is due to the difficulty of capturing quality improvements, new products, etc in the CPI (e.g. new cars priced at the same nominal price but that are of improved quality). From what I recall Stats NZ uses hedonic pricing measurements to capture the measurement problem (they may use other techniques as well), and they still have a bias in the realm of 0.5 to 1 percent.
This being said, in the medium to long run, one cannot hide inflation, as it always shows up in the CPI. This is why the inflation target central banks use should cover a period long-enough to enable central banks to adjust for the problems associated with measuring true inflation. One could have a -1% to +1% range (or -1.5% to +1.5%) over a two to three year period for instance, which is what the NZ Reserve Bank Act tried to achieve without complete success. The problem is that this could entail some deflation at times and deflation is politically difficult to sell. Deflation also reduces the seignorage of the central bank and thus is not in its interest.
In any case, it is crucial to distinguish the short run from the medium to long run because of the inability of central banks to assess correctly short run changes in the demand for money. I understand that the stance of RBNZ is that it shouldn't respond to “first round” effects (see its Monetary Policy Statement), meaning that it should let relative price changes follow their course without altering its policy.
The fundamental issue here is the extent to which central banks can assess changes in the demand for real balances in the economy in the short run. They cannot do it well. To some extent, central banks grope in the dark, as Mises explained (and the graph depicting inflation in my previous post is an illustration of this). This is why we should go back to a system of commodity money where the market, not central banks’ officials, is in charge. In the mean time, I believe a well-defined objective of price stability can go a long way to improve outcomes under a monopoly central bank system (which has happened in the last 15 years in NZ), even if price stability as such is difficult to define. In any case, remember that there is no money (including commodity money such as gold and silver) that is a perfect measuring rod of value: the value of money (in terms of all the other goods) will always vary to some extent (cf. Mises’s Theory of Money and Credit). Because of the nature of money, a perfectly unchanging price level (i.e. perfect price stability) is not achievable. However, good money is. As often, the question is one of alternative institutional systems.