(Note: In this "Inflation and Yields" series, I am outlining the fundamental relationships between and within yield and price complexes. In Part I of this article, we noted that deflated commodity and producer prices have been highly correlated with equity yields since 1871, in much the same way that producer prices were correlated with bond yields under the gold standard, a phenomenon Keynes called Gibson's Paradox. What is the connection? Part II looked at Britain. Part III looked at commodities in America.
Part IV examines the revolution in inflation, and V will look at how this revolution has impacted services, technology, and stocks.)
The Revolution in Prices, 1914-1959
Experience has finally caught up with Fisher's Theory!
[P]erhaps the Gibson phenomenon is no longer with us now that the markets have learned their Fisher.
--Milton Friedman and Anna Schwartz (1976)
Under the gold standard, nominal CPI prices were effectively aligned with nominal wholesale and commodity prices, which is why the level of CPI was often correlated with yields. The link was weaker but unmistakable, which is probably why nearly every economist who has addressed Gibson's Paradox has treated it as a general price phenomenon, rather than a wholesale or commodity price one.
It is clear, however, that since 1914 (or 1971, or whichever date you prefer), that although deflated commodity prices are highly correlated with equity yields, deflated consumer prices are not.
Whether or not one accepts my belief that nominal prices under the gold standard were real prices, if he or she can accept my argument that Gibson's Paradox shifted from nominal prices to deflated prices, then he or she must accept that consumer prices are no longer correlated with yields. It would be mathematically impossible.
If that is the case, then consumer prices, previously loosely tied to commodity prices, are now an independent phenomenon. And, of course, that is what we have seen in recent decades. Commodity prices, which are frequently cited as either evidence of inflation or the source of inflation, have both fallen during inflationary decades (1980-2000) and risen dramatically during periods of low inflation (2000s). Consumer prices have become consistently inflationary, while commodities rise and fall in decades-long sprints.
The only alternative to this analysis of the changed behavior of consumer prices that I can imagine would be to push forward a second deflator to measure 'real' CPI and commodity prices. I would not be opposed in principle to that, but one would have to determine how that measure had behaved over the last few centuries, as well. Let's say, you were to use your own preferred measure of inflation, the GDP deflator, some measure of "core inflation," gold, silver, seashells, whatever. You would still find that there had been a revolution in relative prices.
And, that means that the chronic decline of the dollar since we removed ourselves from the gold standard has impacted different price sectors quite differently.
That might be a bit abstract, but we can demonstrate this a couple of ways.
First, we can try to gauge when deflated commodity prices really started to track with equity yields and, second, when CPI stopped tracking equity yields.
As we have seen, it's difficult to objectively pinpoint when deflated commodity prices really began to correlate with equity yields, because one could argue they have been correlated for as long as we have reliable records of equity yields (again, as far back as 1871). That is precisely the observation that ignited this reappraisal of Gibson's Paradox.
CPI is a different matter, however. Even taking into account that CPI was more weakly correlated with yields over the long run, it appears that that correlation was already weakening quite significantly by the 1920s and '30s.
In the chart below, I've spliced the bond yield onto the earnings yield and calculated a centered moving average to give an impression of what equity yields have theoretically looked like long-term. I have also spliced the wholesale price index (WPI) onto the Grilli-Yang Commodity Price Index (GYCPI) to create an impression of what deflated commodity prices looked like. And, then I've put alongside CPI.
By the mid-1950s, it's already pretty clear that with respect to CPI, Gibson's Paradox had long gone, and it never looked back.
There is a more conventional way to look at this process, however.
In an economic system where inflation is taken for granted, we tend to look at, and think in terms of, year-on-year movements of prices. Below is year-on-year changes in CPI, cotton, and the GYCPI from 1800-2010. Because the GYCPI only goes back to 1900, I am using cotton as a proxy for commodities generally.
The important thing to keep your eye on is the blue line (CPI). It has always been less erratic than commodities (and producer prices generally), but it's behavior, both absolutely and relatively, has clearly changed by the middle of the last century.
If anything can be said for continuity in the behavior of CPI, it is that it's year-on-year movements begin to behave the way absolute price levels used to behave under the standard formulation of Gibson's Paradox. As Friedman and Schwartz wrote in "From Gibson to Fisher," by the late 1960s, the markets had finally "learned their Fisher." But, there was more to it than that.
Setting aside everything that has been said here so far, there are at least two ways to account for this change in CPI. First, there may have been a change in the composition of the consumer price index. Second, there may have been a change in the behavior of consumer prices more generally.
In previous articles, I believe I over-emphasized the first reason. Modern inflation indexes include a number of goods (especially new-fangled technological goods) and services (not just individual goods, but whole classes of goods and services) that were not included in previous versions of indexes. But, I believe we can show that there has been a clear transformation in the absolute and relative behavior of prices of non-commodity goods and services, and that these changes have occurred almost simultaneously. Moreover, the change in the behavior of prices coincides with a revolution in the composition of consumer prices, not only because we have better measures for things like services and technological goods, but because there was a shift in the structure of developed economies.
Think of it this way:
As I have shown in my account of Gibson's Paradox, commodity prices made a transition from nominal to deflated prices probably with the establishment of the Fed, and without a doubt by mid-century. The fundamental pattern of their behavior did not change; we simply had to take into account the change in the currency.
Generally, this has resulted in reduced deflated commodity prices (and increasingly lower equity yields, or higher P/E ratios). That is to say, real commodity prices have retained their cyclicality but generally been suppressed.
The real prices of a number of other goods have also been depressed but no longer display the cyclicality we associate with Gibson's Paradox. I have in mind technological goods, chiefly.
The only long-range proxies for technological goods I can find are measures for plant and machinery and for motor vehicles. If we deflate them by CPI, I think one can notice that, like commodities, they are depressed, but they are also increasingly immune to the Gibson effect.
In other words, deflated prices for goods, "real prices," in general have been in decline, but commodities, although similarly in decline, have remained true to their Gibsonian nature.
That is why I contended in Part III that the correlations for some producer price sectors are partly misleading. In Part V, I will go back over this argument using year-on-year changes, and it should become more evident.
I want to note first that, among the oldest producer price indexes, textiles and furniture most strongly reflect this breakdown in any Gibson effect, although this change does not seem to have occurred simultaneously. This will become significant in Part V.
(The chart below is a bit difficult to read, but keep your eye on the extreme right hand side, noting that some categories did not turn up with commodities, others only weakly so).
It appears that producer indexes that would be least influenced by primary commodity prices deviate from the Gibson Paradox the most.
Therefore, for now, I would like to group all goods into two broad categories: a) those whose real prices are in a state of absolute decline and b) those who appear to be in a state of decline but still retain varying degrees of the Gibsonian cyclicality (especially commodities; chemical, agricultural, and metal producer goods; and fuel). It is difficult to categorize food in this bifurcation, but we can set it aside for now.
So, all the classical categories of goods, as well as the new classes (technological goods) and their deflated prices are in a clear state of decline, although some possess an undiminished cyclicality (commodities).
But, if deflated goods prices are in a state of decline across the board while nominal prices are constantly rising, all of this inflation is emanating from some source outside of goods.
I mean services.