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.
Both posts were excellent and I do think that central banks need to tell a better story than just changes in CPI (or whatever proxy they're using).
Anyway I commented on the original post at Catallaxy (it is safe to ignore the rest of the thread).
http://catallaxyfiles.com/?p=3446#comment-83412
Posted by: Sinclair Davidson | February 27, 2008 at 04:18 AM
Good posts on a most important topic -especially for Austrian-school economists. Sometimes I feel that there are not enough Austrians specialising in monetary policy, which is truly a shame.
I have two comments on Frederic's points:
1. Hedonic price adjustments are questionable as a feature of a CPI index, particularly if that index is used as a monetary policy target. If there is an increase in quality due to productivity increases or higher differentiation, and there is some repercussion in the form of higher prices, first of all it is extremely hard to measure it properly (if at all). And what if there is a shift in relative qualities? (as there are shifts in relative prices) Even more difficult to measure. If Von Mises objected to simple CPI indexes as trying to measure something that is impossible to measure, the problem is compounded with nontransparent statistic alchemy such as hedonics or substitution coefficients.
Worst of all is the result that when there are productivity increases the CPI goes down, and that allows the central bank to be more inflationary without affecting the CPI in the short term. The consequence is that the Central Banks redistributes wealth from consumers towards the financial sector, the State and itself.
2. A fall in the CPI is not deflation. It can perfectly come from falling prices due to productivity increases or more competition. Deflation should be associated with the simultaneous contraction of money, credit and prices.
Another issue is that the argument about relative prices should be dealt rigorously. Now Central Banks are taking the increase in commodities prices or past increases in housing prices as "relative increases", "exogenous", "coming from supply constraints" -all of them excuses for applying expansionary monetary policies in the meantime. However, when those prices eventually fall, you bet that they want to include them or take them into consideration... to apply expansionary monetary policies.
Posted by: Mitch | February 27, 2008 at 03:16 PM
I agree with all your comments. Because of the relative success of central banking in the last 20 years, we have come to forget that central banks are (de jure) monopolies and will act accordingly whenever possible. And the RBNZ is no exception.
Posted by: Frederic Sautet | February 28, 2008 at 03:02 AM
Frederic,
The concept of hedonic adjustment is one of those things that sounds reasonable superficially, especially if its employment gives desired results, but which is actually economic nonsense.
If I buy a new computer for $500 every year, and the speed increases by, say, 100% annually and the hard disk storage space doubles each year, how should this impact the exchange value of the dollar?
The answer is, it shouldn't, not in any practical sense. Only if the new models unleash a significant increase in actual economic output, might it matter.
Why would the exchange value of the dollar care if the computer is able to spend twice as many microseconds waiting for your next keystroke, or if the hard disk has twice as much unused space for papers that you will never have time to write?
What matters is what you spend, the $500. You may well feel more subjectively satisfied with the new models, but this is not a proper basis for computing the exchange value of the dollar.
In general, you can't buy last year's $500 new computer for $250 because it will no longer be available. It will now be heading towards antique status as falling costs and competition force it out of production as new models take over.
Regards, Don
Posted by: Don Lloyd | February 28, 2008 at 10:09 AM
What does it mean that inflation is overstated? I would argue that that inflation is in fact understated, as it is principally a monetary process, and that technological and productivity improvements will hide the effects of said policy.
Posted by: kurt | March 04, 2008 at 09:54 PM