In Praise of Saville
Steve Saville wrote a very nice piece on inflation entitled The Non-Stop Inflation. Although we disagree on the "end game" it sure is nice to see someone properly explain what inflation is, what it is not, and why measuring the CPI is fraught with so many problems as to make it useless. Steve wrote:
The correct definition of inflation is an increase in the supply of money that CAUSES a decrease in the purchasing power of money, but we usually define it as simply an increase in the supply of money. This is done, in part, for the practical reason that it's impossible to measure changes in the purchasing power of money on an economy-wide basis.
It is not possible to measure changes in the overall purchasing power of money because it is not possible to come up with a meaningful number that represents the average price level within an economy. There are, of course, price indices such as the CPI that purportedly represent the average price level, but these indices are generally worse than useless because they are calculated in such a way that they are guaranteed to paint a misleading picture. And in any case, even an honest attempt to determine the average price level would fail. The reason is that even if it were somehow possible to determine a meaningful number that represented the average of things as different as eggs and new cars, a calculation that was valid today would, in a dynamic economy, be obsolete tomorrow.
To illustrate just one of the insurmountable challenges of coming up with a price index that accurately represents the EFFECTS of inflation let's consider the hypothetical example of the $2 widget made in the US. In our example we'll assume that monetary inflation within the US over many years pushes up manufacturing costs such that these widgets can no longer be made in the US and sold at a profit for anything less than $3; in other words, if nothing else changes then the inflation will cause the dollar's purchasing power to fall from 0.5 widgets (2 dollars = 1 widget) to 0.33 widgets (3 dollars = 1 widget). But something else does change. Instead of raising the selling price of the widget from $2 to $3 in response to the increase in costs, what actually happens is that our hypothetical widget manufacturer keeps the selling price the same and opens a factory in a part of the world where manufacturing costs are much lower. Therefore, in our example inflation causes the dollar's purchasing power to fall and this, in turn, leads to the closure of a factory in the US and an increase in the quantity of imported goods. Inflation has obviously had an important effect, but it's unlikely that any price index will capture this effect.
There can be no real meaningful discussion of inflation or deflation unless one agrees on a definition. I am pleased to see that in spite of one of us being an inflationist and the other a deflationist, we have independently surmised what the real problem is: expansion of money supply and credit.
In Inflation: What the heck is it? and Inflation Monster Captured, I wrote many reasons why attempting to measure inflation by looking at the CPI was a fool's game. It is good to see someone who has a radically different view of the end game than I do at least agree as to what the problem is.
Given that the problem is a rampant expansion of money and credit, the solution can not possibly be a "price stability" policy that attempts to measure consumer prices. With all the hedonic adjustments (many justified but probably all overstated), with all the nonsense about core inflation vs. non-core inflation, with all the imputed economics, with all the understating of medical costs, and with enormous discrepancies between rental costs vs. housing ownership costs, there is not a person on this earth that could possibly know 2% price increases if it hit them smack in the face.
Right now I am in a huge debate with someone on Silicon Investor. It seems that even otherwise reasonable people (which I think both of us are), simply can not come to agreement about whether or not quality improvements should be factored into the CPI.
My position is that quality improvements should be factored in otherwise one is not comparing apples to apples. Take autos for example: antilock brakes, seat belts, air bags, improved gas mileage, air conditioning, CD players, electronic ignition, fold down seats, windows that safely shatter, better gas mileage, intermittent wipers, reduced emissions, etc, are all quality improvements that have appeared in cars over time.
As compared to 1924 even the amount of raw materials going into cars has gone up dramatically (copper, electronics, rubber, aluminum, steel, glass, etc). On a component basis alone (forgetting about quality improvements like intermittent wipers and safer glass) I just do not see how one can reasonably claim that the difference in price of a car today vs. 1924 can all be chalked up to "price inflation". Yet somehow there is a disagreement as to whether or not either component changes or quality improvements should be factored in. No wonder people have wildly differing views of price increases. In practice, it seems that people see the every price increase as "inflation" but fail to see or account for obvious improvements in quality. Getting agreement as to how to measure quality improvements is of course impossible if one can not even get agreement that it should be done at all.
To be sure, the government overstates those quality improvements and does so on purpose so as to lower the CPI. This makes Social Security and other cost of living adjustments (COLAs) as small as possible. Such manipulation is of course is another reason why CPI price targeting is wrong.
While I agree with Saville about the necessity of tracking money M3 is probably not the best way to do it. Shostak has pointed out problems with the "M-series" of money supply measures in Making Sense of Money Supply Data as well as other articles that he has written.
The distinction between money and credit it is an important one, and each needs to be tracked separately given the extent of our fractional reserve lending. In Japan, the government continually injected money for 18 years in an attempt to fight deflation but credit contracted at a greater rate nearly the entire time. I believe the same will happen here.
For those that want to compare my end game scenario with Steve's here they are.
Saville: The End Game