COVID-19 RESOURCES AND INFORMATION: See the Atlanta Fed's list of publications, information, and resources; listen to our Pandemic Response webinar series.

About


The Atlanta Fed's macroblog provides commentary and analysis on economic topics including monetary policy, macroeconomic developments, inflation, labor economics, and financial issues.

Authors for macroblog are Dave Altig, John Robertson, and other Atlanta Fed economists and researchers.

Comment Standards:
Comments are moderated and will not appear until the moderator has approved them.

Please submit appropriate comments. Inappropriate comments include content that is abusive, harassing, or threatening; obscene, vulgar, or profane; an attack of a personal nature; or overtly political.

In addition, no off-topic remarks or spam is permitted.

June 13, 2011

Core cuts both ways

With the six-month average of annualized headline inflation running just over 5 percent, this Wednesday's consumer price index (CPI) report looms a little larger than usual. While it is dangerous to predict such things, there is every reason to believe that the measured increase in CPI inflation for May could be quite low. And there is every reason to believe that this softness will persist into June.

The reason is quite simple. Movements in gasoline and fuel prices really do push around the headline inflation number, and at long last it looks like that movement is in the downward direction. Here's the relevant picture:

110613


Let's assume that, annualized, CPI excluding food and energy rises 2.1 percent, as it has so far this year, and food and nongasoline energy prices each rise at 5 percent. Then when you plug in gasoline prices already realized for May and EIA predictions for June, you get a 0.4 percent rise in the headline CPI in May and a 0.7 percent decline in June (both rates annualized).

Despite some probable relief on the headline inflation number, I remain aware of what that relief means and what it does not. Earlier in the year, Atlanta Fed president Dennis Lockhart had this to say:

"I want to contrast inflation to the cost of living. In casual language, we often interpret a rise in the cost of living as inflation. They are not the same thing. Cost-of-living increases are a result of increases in individual prices relative to other prices and especially relative to income. These relative price movements reflect supply and demand conditions and idiosyncratic influences in the various markets for goods and services…

"… The Fed, like every other central bank, is powerless to prevent fluctuations in the cost of living and increases of individual prices. We do not produce oil. Nor do we grow food or provide health care. We cannot prevent the next oil shock, or drought, or a strike somewhere—events that cause prices of certain goods to rise and change your cost of living."

Two points, then. First, even if things evolve as the chart above suggests, the level of gasoline prices will remain relatively high by recent standards. Low inflation readings for the next couple of months would therefore leave the cost-of-living high by recent standards, a fact that is not lost on us here at the Atlanta Fed.

Second, as President Lockhart's comments reveal, we were reluctant to attribute the run-up in fuel prices to monetary policy. And I imagine we will be equally reluctant to credit monetary policy with any relief from that trend.

In fact, I will fearlessly predict that, should our guess here about headline inflation in the next couple of months be proved accurate, we will point to core inflation measures as reason to look through some very low inflation readings. See these comments from President Lockhart's most recent speech for some additional perspective:

"Are the recent outsized increases in headline inflation the best signals of the inflation trend going forward? Or are other statistics—like core inflation or measures of inflation expectations—yielding a truer picture of what lies ahead?

"The answer matters a lot. And it certainly weighs heavily on my thinking. I would not hesitate to support an exit from our current policy stance if I believed that the headline inflation number of the past six months is really indicative of the underlying trend inflation rate. I don't believe this to be the case. And I am wary of tightening monetary policy in the face of quite ambiguous economic circumstances unless doing so is absolutely necessary to meet the FOMC's price stability mandate."

And here's why it matters, quoting President Lockhart from an interview with Reuters last week:

"In the interview, Lockhart said maintaining an easy Fed policy stance should ensure the moderate U.S. recovery does not fall off the rails."

But…

"How high would the bar be for further Fed easing?

"It would take 'a significant deterioration as reflected in the overall economy, a set of deflationary signals and also unemployment numbers that rise dramatically. Last fall (when the Fed launched QE2), by some measures, we were seeing declining inflation expectations that were headed in a pretty negative direction, and that was happening pretty rapidly. We were in a disinflationary environment. So the risk of deflation was plausible. We acted and the situation has turned around. That was, at least in my way of thinking, very central to supporting the policy. We don't have anything remotely like a deflationary risk at the moment, short of a shock of some kind.' "

If what I suggest above—falling headline inflation, stable core inflation—comes to pass in the near term, don't expect me to start ringing the disinflationary bell.

This isn't the last word on the usefulness of core inflation statistics, and the debate is certain to rage on (see here and here, for recent installments). But I do hope that this post is remembered the next time the Federal Reserve is accused of hiding inflationary pressures behind the rhetoric of core.


Update: The May CPI report is in, and once again the facts trump arithmetic. We got the flip in the headline/core measures right; the rest, not so much.


Update: Here's one way to construct a synthetic CPI:

110613b


CPIHeadline(T) is the calculated value of the seasonally adjusted CPI in month T > December 2010. CPIHeadline(Dec2010) is the reported seasonally adjusted headline CPI for December 2010. CPIFood, CPICore, CPIGas, CPIOtherFuel, and CPIHHenergy are the seasonally adjusted component level CPI indexes for the above mentioned components. Note that this formula will not exactly replicate the January through April 2011 headline CPI since the components are not granular enough. When using the above formula to compute CPIHeadline(Apr11) and CPIHeadline(May11), we compute one-month headline inflation in May 2011 as the percent difference between the values of CPIHeadline(Apr11) and CPIHeadline(May11), both calculated with the above formula (i.e., do not use the reported value of CPIHeadline(Apr11)). In our calculations we used the same inflation assumptions for "other motor fuels" and "household energy" (both increase at a 5 percent annualized rate in May and June). By Patrick Higgins, an economist at the Atlanta Fed


David AltigBy Dave Altig, senior vice president and research director at the Atlanta Fed

March 7, 2011

One chasm that really isn't one

Floyd Norris, writing last Friday in the The New York Times, fretted about "The Chasm Between Consumers and the Fed." We here at the Atlanta Fed share some of that concern, and indeed the Times article quotes from a speech by our president Dennis Lockhart on just that subject from last month. But then Norris takes a turn I didn't expect. Norris's Times article includes the following chart…


…and the article proceeds:

"The Fed's goal is to keep the inflation rate at or near 2 percent, and it does not expect a significant increase for at least a few years. … The stock market is generally thought to do better when inflation is falling, but Martin Fridson, the global credit strategist for BNP Paribas Investment Partners, points out that is not always the case. There is, he says, a time when inflation is too low.

"The accompanying chart, based on a report by Mr. Fridson, shows that from the 1940s through the 1990s, there generally was a relationship. The more inflation declined in a decade from inflation in the previous decade, the better the stock market did.

"But in two decades, the 1930s and the first decade of this century, inflation fell from already low levels and the stock market suffered. 'Below a certain level of inflation,' Mr. Fridson said, 'a further decline reflects economic weakness more than it reflects a salutary reining in of excessive monetary creation.'

"If that is correct, then it could be that both investors and those simply concerned with promoting economic growth should, as Mr. Fridson wrote, hope that Mr. Bernanke 'fails in his stated goal of holding inflation to 2 percent or less.' "

It was all good, up to that last paragraph. As President Lockhart reiterated in a speech today (emphasis added):

"I'll explain the technical rationale of my Reserve Bank in supporting the scope of LSAP2 [the second round of large-scale asset purchases] last November.

"Through the summer and into the fall of last year, our internal forecasts at the Atlanta Fed were calling into question whether the policy stance at the time assured progress toward the committee's growth and price stability objectives. In more normal times, these circumstances would have prompted a cut in the FOMC's [Federal Open Market Committee] target for the federal funds rate. This approach would be (would have been) the prescription of the so-called Taylor rule which relates policy rate moves to forecast 'misses' on the Fed's sustainable growth and stable inflation objectives."

As we've argued in macroblog before, keeping inflation from falling below that "certain level of inflation" reflecting "economic weakness more than it reflects a salutary reining in of excessive monetary creation" was exactly what President Lockhart has offered in defense of implementing LSAP2, and in support of claims to its success.

There remain plenty of policy questions on which intelligent well-intentioned folk can disagree, but on the assertion that it is wise to guard against too much disinflation, we are in agreement. No need to find disagreements that aren't really there.


Photo of Dave Altig By Dave Altig
Senior vice president and research director at the Atlanta Fed


December 22, 2010

An inflation (or lack thereof) chart show

Over at TheMoneyIllusion, Scott Sumner takes a shot at what he refers to as "Disinflation Denial." His point is that prior to the recent run-up, "commodity price indices fell by more than 50%." Thus, if the run-up in commodity prices suggests loose policy now, they must have been signaling tight policy earlier.

I am hesitant to endorse the view that any subset of prices gives us a clear view of inflation trends. What I do endorse in the Sumner piece is the advice that "the Fed look at a wide range of indicators." I can tell you that is exactly what we do at the Atlanta Reserve Bank and, as just one example within the Fed System, in this post I'll review the battery of indicators that we are currently looking at here. Most of these will be no surprise, but I find it useful to occasionally see them in one place. So here we go. (Note that throughout this blog post I will focus most of my comments on the consumer price index [CPI], but most of what I say also applies to the personal consumption expenditure [PCE] price index as well.)

First up, of course, are the so-called (and often maligned) core measures of inflation. I am completely sympathetic to the view that the traditional core index, which subtracts out food and energy components, is a somewhat arbitrary cut of the price statistics. For that reason, Ipersonally tend to lean more heavily on median and trimmed-mean measures.


In Atlanta, we have been monitoring a newer core inflation measure, called the "sticky-price CPI," jointly developed by Mike Bryan and Brent Meyer (of the Atlanta and Cleveland Feds, respectively). As described by Bryan and Meyer:

"Some of the items that make up the Consumer Price Index change prices frequently, while others are slow to change… sticky prices [those that are slow to change] appear to incorporate expectations about future inflation to a greater degree than prices that change on a frequent basis… our sticky-price measure seems to contain a component of inflation expectations, and that component may be useful when trying to gauge where inflation is heading."

Like the other core measure, the sticky-price CPI shows a pronounced downward movement over the past several years, with some sign of (an ever-so-slight) recovery as of late.


Though I disagree with the assertion that core measures are a convenient way to ignore unpleasant movements in the overall CPI—there is evidence that core measures are useful in predicting where total CPI inflation is heading—it is almost surely a bad idea to ignore what is happening to headline statistics. (After all, in the end it is the average of all prices with which we are concerned.)

Here too, the evidence suggests, at the very least, there is scant evidence that disinflation has left the scene:


I find it useful to take at least two more cuts at the overall price data. One, which has a decidedly short-term focus, involves examining the distribution of price changes in the broad categories that make up the headline CPI. Though a popular criticism of Fed policy—discussed and critiqued at Econbrowser—tries to deflate deflation concerns by reciting a number of prices that are rising, it is obvious that one could just as easily tick off a reasonably large list of prices that are falling:


(The individual colors in the chart represent different components of the CPI. The underlying data can be found from this link to the explanation of the median CPI.)

The graph of the November price change distribution is actually somewhat encouraging. What it tells us is that almost half of the price changes in the CPI market basket, weighted by their shares of total consumer expenditures, fell in the (annualized) range of 0 percent to 2 percent. Furthermore, about as many price changes were below this range as they were above it.

A closer look at the prices that fall in the 0 percent to 2 percent category, however, reveals that individual price changes are skewed to the downside of the range:


On a month-to-month basis, the distribution of individual prices does shift around, so these statistics are nothing more than suggestive short-run snapshots (but I believe they are informative nonetheless).

At the other end of the temporal scale is a look at how inflation has behaved over time. If the central bank had a long history of missing its stated inflation objectives, we might feel very different about an inflation rate that is below what Chairman Bernanke has referred to as "the mandate-consistent inflation rate" of "about 2 percent or a bit below" than we would if average prices were hewing pretty close to the target path. As I have previously noted, over the past 15 years or so, the Federal Open Market Committee (FOMC) has delivered an average inflation rate, measured as growth in the PCE price index, that is wholly consistent with this mandate. Here's the case in a graph, adjusting the mandate-consistent inflation rate to account for an assumed upward bias in the CPI relative to the PCE index:


Actually, those short-run complications are mostly associated with falling expectations of inflation. In my last macroblog post, I argued that the stabilization of market-based CPI inflation expectations and the associated decline in the perceived probability of deflation should arguably be counted as a success of the Fed's current policy stance. The latest on market-based expectations was included in our previous macroblog item. For completeness, survey-based expected long-term inflation remains somewhat below the levels prior to the onset of the recession:


I believe this is basically the bottom line: whether we look at headline inflation (straight-up, component-by-component, or in terms of the long-run trend), core inflation measures (of virtually any sensible variety), or inflation expectations (survey or market based), there is little a hint of building inflationary pressure.

While I don't dismiss the usefulness of looking at other indicators (stock prices, bond prices, foreign exchange rates, commodity prices, and real estate prices are on Scott Sumner's list; I would add various measures of labor costs to mine), you have to be pretty selective in your attentions to build the contrary case.

But feel free. We'll keep watching.

Photo of Dave Altig By Dave Altig
Senior vice president and research director at the Atlanta Fed

December 17, 2010

What's behind the recent rise in Treasury yields?

David Beckworth, who blogs at Macro and Other Market Musings, posted a comment regarding macroblog's post "What might monetary policy success look like?" from December 2. Beckworth's comment specifically mentioned this chart…


… as part of this question:

"How did you create the latter figure [shown above]? Using the Fed's own constant maturities series (for both the nominal and real yield), the figure I come up with is less impressive. It shows a turnaround in inflation expectations about the time QE2 is promoted by Fed officials, but then inflation expectations stall and remain far from the 'mandate-consistent inflation rate.'
"Here is a post where I placed one such graph."

And here's the graph of expected inflation from Beckworth's post:


The series shown in the Beckworth chart has a different economic meaning than the chart shown in the original macroblog post (as was suggested by another commentator to our earlier post).

The chart Beckworth shows in his referenced blog post is the five-year Treasury Inflation-Protected Securities (TIPS) spread (the difference in nominal and real Treasury yields at five-year maturities). And so when he states, "This figure shows average annual expected inflation over the next five years has been flatlining around 1.55% over most of November" it means just that: it's examining the next five-year period (2010–15). I've reposted below an updated version of this chart, along with the 10-year TIPS spread. Since Beckworth's comment on macroblog, the five-year TIPS spread has widened about 13 basis points, depending on the measure you're using.


The chart used in the December 2 macroblog post is a different measure altogether. It's the five-year/five-year forward break-even inflation rate—not the TIPS spread. This chart shows a measure of expected inflation in the five-year period beginning five years from now. So this chart shows what investors expect to be the cumulative change in the consumer price index beginning in November 2015 through November 2020. Put another way, it's the realized inflation that would provide an equivalent return to both the nominal Treasury securities and the real TIPS securities. An updated picture is provided below.


Thus we're talking about apples and oranges in two respects: (1) these two charts cover different periods (2010–15 versus 2015–20); and (2) the two calculations themselves are different (taking a simple nominal-real spread versus the 5-year/5-year forward calculation).

Now what's the point of all of this, besides highlighting the minutiae of measuring inflation expectations? Resurrecting Beckworth's question and answering it help illuminate the recent concern about increases in Treasury yields. Indeed, since the November Federal Open Market Committee (FOMC) meeting, longer-dated yields have risen considerably, with the 10-year bond's yield up 86 basis points, for example. But the recent movements in nominal and real yields can be placed in two categories: (A) from when the Federal Reserve began signaling consideration of further asset purchases (late August) to the November FOMC meeting, and (B) the post-November FOMC meeting period. In period A, nominal yields were relatively flat while real yields declined somewhat, indicating a healthy rise in inflation expectations from the lows seen this summer (this change is shown by the increase in the TIPS spreads and breakeven inflation rates during the period).  In period B, the rise in nominal yields has been primarily driven by a rise in real yields (not unanchored inflation expectations).


As Martin Wolf wrote in Tuesday's Financial Times on this issue, "To understand what is going on, we need to distinguish the role of shifts in real interest rates from that of shifts in inflation expectations." As is evident in the charts, and in one of Beckworth's most recent posts, real rates have risen alongside nominal rates—a sign that inflation expectations are now relatively stable.

By Andrew Flowers, senior economic research analyst in the Atlanta Fed's research department