« "Future Recession Risks" | Main
It's a schizophrenic world. On one side, there are lots of people worried about hyperinflation [0], despite forward looking indicators of inflation signalling quiescence [1] and actual price indicators going downward.
On the other are those who actually look at the data, and then conjoin their observations with information on the tremendous slack in the economy, and say that rapid inflation is unlikely. So while all eyes are on this Friday's CPI release, I assert that we don't really have to wait to find out the trajectory of inflation. In this post, I will highlight the fact that over certain horizons, we already have deflation; and for certain segments of the population, inflation has been at zero for a year already.
Core Inflation
First, consider standard indicators. The core CPI and core PCE deflator inflation rates are still positive.
For discussion of differences in coverage, weights, and construction between CPI and PCE, see this post. From these conventional indicators, it would seem we are still aways from deflation.
Measures of Inflation, by Income Group
It is important to keep in mind that the impact of price changes on the cost of living differs across income groups, as I discussed in this post. Figure 2 depicts the 3 month annualized growth rate in the CPI as it pertains to the 1st, 3rd, and 5th income quintiles.
A caveat: I call these guesstimates because I am not always certain that the categories I have selected match up with the series Kokoski uses in her calculations. In addition, I do not believe that my ex.-"energy" series matches up with the construction of the official CPI excluding energy and fuel series reported by the BLS. I welcome cross-checking. However, I think the deviations of the movements by quintile should be representative.
Clearly, (3 month annualized) inflation for the 5th quintile is very close to stall speed, at about half a percentage point in June 2010. In contrast, inflation being experienced by the 1st income quintile is nearly 1 percentage point. Another way to highlight this difference is to examine the (log) levels of the indices.
Notice that the 5th quintile CPI ex.-"energy" is essentially flat since 2009M07. Running a regression on first differences (remember log price indices are at the very minimum I(1), and arguably I(2) over some samples), one finds that the implied annualized inflation rate is 0.3 percentage points, and is not statistically significantly different from zero at the 10% msl. In contrast, the core CPI inflation rate over the corresponding period is 1 percentage point, and is significantly different from zero, with a p-value of 0.015 (both regressions using HAC standard errors, 2 lags).
Why should one care about the 5th quintile? If one is concerned about aggregate consumption, then it is of interest to know what happens to the top 20%. According to Moody's, about 60 percent of total consumption is accounted for by this quintile [2]. If the price level facing this group is falling, then they might either defer consumption in anticipation of yet lower prices, or either increase or decrease consumption in response to changes in wealth (depending on whether they are net creditors or debtors). (The overall CPI weights, according to Deaton, corresponds to about the 75th percentile in income.)
Additional Observations
The CPI-all inflation rates (3 month annualized) for the 3rd and 5th quintiles have been negative since May and April respectively. Even that for the 1st quintile has declined as of June. This matches up with the declining inflation for the standard CPI-all, since May.
We know that the CPI is upwardly biased, due to the use of fixed weights for the major categories. That means that if one could apply a Fisher ideal index to these data, for each quintile, one would probably obtain more negative inflation rates. [3]
Parting Thought
We have forward looking market-based indicators of expected inflation. One is the spread between the 5 year Treasury rate and the corresponding 5 year TIPS. Keeping in mind the problems with using the spread to infer inflation [4], here is the picture.
With these points in mind, and with tremendous slack in the economy, rapid inflation seems like the last thing one should worry about. Or crowding out, for that matter [5].
Update: 11:30am Pacific, 8/10/2010
Oscar Jorda has just sent me a note written by a UC Davis PhD student, Paul Gaggl, which reports the unconditional forecasts and associated fan charts for a simple four variable VAR (GDP growth, PCE inflation, Fed funds rate, ten year bond rate). Pay close attention to the top rightmost graph in the chart.
I'm not saying this particular VAR is the definitive word in how to model inflation, but I think it indicates the fact that based upon historical correlations, there is ample evidence to support the view that deflation is very possible. (The referenced Jorda/Marcellino paper is here)
In other news, WSJ RTE/Izzo reports:
A Wall Street Journal survey found that by a two-to-one margin Wall Street economists see deflation as a bigger threat to the U.S. economy over the next three years than inflation.
"Deflation is dangerously close," said David Resler of Nomura Securities, one of 53 economists surveyed by the Wall Street Journal. Among economists who answered the question, nearly two-thirds said that deflation poses the bigger risk to the economy over the next three years; the remainder said inflation is the bigger threat. That compares to an April survey, when the economists were split 50/50 over whether inflation or disinflation posed the bigger risk over the next year.
Posted by Menzie Chinn at August 9, 2010 04:43 PM
5-year inflation swaps paint a similar picture and are (presumably) a more direct market-based measure of inflation expectations.
Posted by: Nemo at August 9, 2010 05:35 PM
Disinflation or deflation alongside a zero bound makes things even worse. Ordinarily it is positive interest rates that are responsible for a downward sloping aggregate demand curve. But if interest rates are at a zero lower bound, then the slope of the aggregate demand curve is determined by the relative effects of lower prices giving us more buying power (a Pigou effect) versus the effect of deflation on debtors. With deflation debtors will cut back on consumption. Yes, owners of debt will have more income from debt payments of debtors, but more of that debt income will be saved. I think that's what the quintile charts show. Low income quintiles don't have a lot of options. The net effect is that the this debt effect overwhelms the real balance effect, and this implies that the aggregate demand curve will be (locally) upward sloping. And that means deflation not only leads to more deflation, but a rightward shift in the aggregate supply curve actually reduces output. A world turned upside down. We better hope this is nothing more than a theoretical musing.
Posted by: 2slugbaits at August 9, 2010 06:29 PM
Perhaps the current deflationary depression is a unique opportunity to monetize the debt without causing proportional inflation.
On the other hand, we're so far into uncharted territory, they'll probably blow something up when they try it.
Posted by: W.C. Varones at August 9, 2010 07:48 PM
Menzie: For the question at hand, another source to consider:
http://www.clevelandfed.org/research/data/inflation_expectations/index.cfm
Posted by: Phil Rothman at August 9, 2010 08:47 PM
Broad commodity indexes are up about 20% in the last 18 months. Equities up over 30%. TIPS and Treasuries are also up. In the wonderland of ZIRP, it's all good. Are all those markets predicting low inflation? Does the Treasury bond market reflect a sober calculation of the risks ahead or does it reflect a toxic cocktail of central bank manipulation (both foreign and US)along with leveraged speculators? Time will tell.
Posted by: tinbox at August 9, 2010 09:00 PM
Nemo: Thanks for the useful link.
Phil Rothman: Thanks! I have a plot from a month ago using this data here.
2slugbaits: Good observations. If anybody knows whether the top quintile is net debtor on paper assets (as opposed to real assets), please send along the info.
tinbox: I'll just note that survey based expectations (such as the one mentioned by Phil Rothman) are also trending down.
Posted by: Menzie Chinn at August 9, 2010 09:02 PM
"Broad commodity indexes are up about 20% in the last 18 months. Equities up over 30%"
Please, that is hardly impressive considering the losses that were obtained previously. If Big Capital doesn't start spending again, than "broad" base commodity indexes and the stock markets will all "correct" again and lose half of those 'gains'.
The ability to draw any inflation is if Big Capital starts spending again to give the American idiots.....err I mean people confidence to start spending again with the savings they have collected(this, they wouldn't have if a direct liquidation had taken place). Then the economy will pick up slack and job growth will resume.
That is the way the system is run. Big Capital always is the starter on the car.
Posted by: The Rage at August 9, 2010 10:41 PM
While the Treasuries minus TIPS estimate of inflation is theoretically ok, though I've never seen anyone actually use TIPS that way and the tax ramifications make the equation highly inexact, my point is that there are other markets that are hard to reconcile with the low inflation sentiment.
It is pretty clear that the housing element will keep overall inflation measures in check for many, many months. And the dismal employment situation will also have a significant downward impact on pricing. These aspects of the economy are well known and not really disputed.
But if the main function of the Fed over the next few years is to ensure that domestic housing prices (the weakest asset sector by far) do not fall significantly further, then what should reasonable, non-hyperinflationistas, expect prices for other goods and services to do?
Posted by: tinbox at August 10, 2010 03:16 AM
Menzie:
Read Full Article »