January 16, 2008
Will inflation fears restrain the Fed?
I think not, and here's why.
The Bureau of Labor Statistics reported today that the seasonally adjusted consumer price index increased by 0.3 percent between November and December, implying an annual inflation rate of 3.4%. If you take the percent change in the seasonally adjusted December 2007 CPI over the seasonally adjusted December 2006 CPI, you get an inflation rate for 2007 of 4.1%. That's the highest December-over-December growth in the seasonally adjusted CPI since 1990, which seems to be the basis for claim by many media outlets that this was the highest inflation rate in 17 years. But that metric is a bit odd, since if you're looking at December-to-December comparisons, I'd prefer to start with the seasonally unadjusted numbers, and for that matter it's not clear why you'd focus just on December-over-December changes as opposed to, say, November-over-November changes. A monthly graph of the year-to-year change in the seasonally unadjusted CPI reveals six observations since 1992 that were bigger than December 2007 (one of which is November 2007). Notwithstanding, there's no question that CPI inflation has been coming in at higher values recently than it had for much of the last two decades.
The Fed bases its actions not on what inflation has been, but rather on what it anticipates for the future. Price changes in food and energy have historically exhibited less serial correlation than most other categories. For example, even though energy costs, according to the BLS, rose at a 37% annual rate in 2007:Q4, it's not a particularly good forecast to anticipate they'll also increase an additional 37% during 2008. Forward-looking Fed policy decisions are more apt to respond to the core CPI, which excludes food and energy items. This grew at a 2.9% annual rate in December and 2.4% over the last year.
|Measure||Past month||Past 6 months||Past year|
|Trimmed mean PCE||3.2||2.5||2.4|
For that matter, the Fed has been paying more attention to the measure of inflation derived from the implicit price deflator on personal consumption expenditures. Based on the November values for this, we'd calculate inflation at a 3.6% rate over the last year, or 2.2% if we excluded food and energy. An alternative strategy to deleting food and energy is to delete the extreme outliers in both the up and down directions-- the Dallas Fed's trimmed mean PCE measure calculates 2.4% inflation for 2007.
In testimony in June 2006, Fed Chair Ben Bernanke referred to the then-prevailing annualized 6-month changes in core CPI of 2.8% and core PCE of 2.3% as
at or above the upper end of the range that many economists, including myself, would consider consistent with price stability and the promotion of maximum long-run growth.
All of which suggests to me that, if the Fed did not have important concerns about what is happening to real economic activity, it would be raising rather than lowering interest rates at the moment.
But of course, the Fed does have some very big concerns about real activity, and the question is to what extent the recent inflation numbers will dissuade them from lowering rates aggressively. Here is what the Fed Chair said in his remarks last week:
Even as the outlook for real activity has weakened, there have been some important developments on the inflation front. Most notably, the same increase in oil prices that may be a negative influence on growth is also lifting overall consumer prices and probably putting some upward pressure on core inflation measures as well. Last year, food prices also increased exceptionally rapidly by recent standards, further boosting overall consumer price inflation. Thus far, inflation expectations appear to have remained reasonably well anchored, and pressures on resource utilization have diminished a bit. However, any tendency of inflation expectations to become unmoored or for the Fed's inflation-fighting credibility to be eroded could greatly complicate the task of sustaining price stability and reduce the central bank's policy flexibility to counter shortfalls in growth in the future. Accordingly, in the months ahead we will be closely monitoring the inflation situation, particularly as regards inflation expectations.
How exactly is the Fed monitoring that anchor for inflation expectations? I suspect that one of the key series they look at is the nominal interest rate on long-term Treasuries. If investors did have significant worries about inflation, you'd expect this to be creeping up. But in fact the 10-year Treasury yield has been making new lows. Hard to imagine how someone who believed inflation was going to continue to romp along at 4% would have any interest in a 10-year security yielding only 3.72%.
Regardless of how you interpret that indicator, a second thing the Fed is looking at is the spread between inflation-indexed Treasuries and nominal yields, which offers a nice hedge against inflation (as measured by the "headline" CPI, not the core measure) for anybody who wants it. That spread has also been very steady at around 2.3%. Again, if you're expecting inflation over 3%, you're a fool to buy the nominal securities rather than the TIPS. A gap of 2.3% is pretty convincing evidence that buyers of U.S. Treasuries do not anticipate 3% CPI inflation to continue.
As long as those two series stay in their recent territory, the Fed thinks it has the maneuvering room to be aggressive about addressing the dangers of an economic downturn and financial collapse. And that's why we'll see at least a 50-basis-point cut in the fed funds target at the next meeting, despite the "highest inflation rate of the last 17 years".
Posted by James Hamilton at January 16, 2008 02:31 PMdigg this | reddit
Listed below are links to weblogs that reference Will inflation fears restrain the Fed?:
» January 16, 2008 from PrefBlog
Rule #1 states that the world always looks more interesting than it really is, an idea mentioned in a previous post, The Bond Market is Excitable. James Hamilton of Econbrowser took a look at the retail sales numbers that had everybody so excited yeste... [Read More]
Tracked on January 16, 2008 06:24 PM
I suspect that one of the key series they look at is the nominal interest rate on long-term Treasuries. If investors did have significant worries about inflation, you'd expect this to be creeping up. But in fact the 10-year Treasury yield has been making new lows.
Hmm, where have we heard that before?
Ha, ha, ha, ha, ha. JDH agrees with ME!
JDH, one reply to my T-Bill theory is that there is now a bubble in T-bills, and that there were 6 years in the 1970s where BEFORE TAX, inflation adjusted yields on T-Bills were negative. So buyers of T-Bills may have a longer term horizon that does not reflect shorter term inflation expectations.
Would you happen to have an opinion on those counter-arguments?
Posted by: Buzzcut at January 16, 2008 07:15 PM
Who buys Treasuries, and what rate of inflation are they experiencing relative to the headline CPI? We know that those near the poverty line, who must put larger portions of their income into food, energy, and health care, currently experience an inflation rate significantly higher than the headline figure; but they buy darned few Treasuries.
Posted by: Michael Cain at January 16, 2008 07:31 PM
Although I understand the reason for having "Core inflation", I think the Fed (and all central banks) should take non-core CPI as being the guide for making monetary policy.
What is disturbing about America's 4.1% inflation rate is that the natural inflationary effects of a declining currency have yet to be felt.
At some point, Bernanke and the Fed will be faced with a choice of fighting inflation or boosting growth. As the 1970s taught us, a monetary policy that focuses upon growth and employment produces neither growth nor employment and does not even solve inflation. Instead, a sole focus upon inflation will solve inflation and, eventually, produce growth and employment.
Let's hope Bernanke sees reason - and that means raising rates and killing inflation. That may seem disastrous during a recession but, hey, Volcker did it. It's really the only alternative.
Posted by: One Salient Oversight at January 16, 2008 08:17 PM
The Cleveland Fed uses an adjusted Treasury-Tips comparison to estimate inflation expectations. Their latest (dailY) calculation suggests more inflation ahead. Here is the url
Posted by: wogie at January 16, 2008 09:32 PM
Buzzcut:What would you mean by a bubble in Treasuries?
Wogie:I'm not sold on the Cleveland adjustments as the way to use these numbers. Seems odd that the "adjusted" series is so volatile and the simple-minded spread is so steady.
Posted by: JDH at January 16, 2008 09:46 PM
JDH, I think you may have lost the forest in looking at the trees.
The US has maintained low inflation with moderate growth only because money has been pouring in from developing nations. The money has been essential to fund large governmental deficits generated in part by the occupation of Iraq and Afghanistan. It has also financed consumption by keeping interest rates low.
Events can change: developing nations could invest internally (raising interest rates), we could end the occupations (lowering deficits and interest rates). Americans could be forced to cut back on consumption for reasons other than price (tighter regulation of mortgages, for example), lowering interest rates. Declining growth could raise deficits and interest rates.
Yet our national experience is that at the end of every sustained conflict, inflation has raged out of control as the federal government attempts to lower deficits while maintaining growth. In the present case, there is a lot of bad economic news coming down the pike: rising retirement, rising medical costs, unfunded pension liabilities, and so on.
And in previous conflicts, we won or we lost nothing in geostrategic advantage. It looks increasingly probable to me that the outcome in Iraq will be a clear loss, but in any event, it will not be a clear win, i.e., the US will not suddenly transform the Middle East into a peaceful paradise. Most likely, problems in Pakistan and Afghanistan will force a shifting of troops out of Iraq and a return of violence. Victors of wars get a clear economic advantage that helps to rein in inflation. Losers, like post-WW I Germany, get disadvantages that fuel it.
I agree that the CPI data do not show a trend. But the big picture (as well as Barry Ritholtz's Big Picture) makes a good case that inflation will become a problem in the near future.
Posted by: Charles at January 16, 2008 11:28 PM
Okay, inflationary expectations remain anchored because TIPS and long bond rates are low. Foreign savers (private and public) are willing to accept the low rates, so no domestic savings are needed at the present time. The Fed can deal with the current crisis by lowering rates, and worry about the national savings rate later (or not at all since this is not one of its mandates).
Posted by: Dropping Rates at January 17, 2008 04:45 AM
The picture changes substantially if you look at the PPI "Seasonally adjust annual rate for the 3-months ended December 2007".
Finished Goods: +13.3%
Finished consumer foods +9.4%
Finished energy goods: +51.9%
Intermediate materials supplies & components: +15.0%
Intermediate food & feeds: +19.2%
Intermediate energy goods: +55.3%
Materials for non-durables: +20.3%
Think those are scary? Look at the head of the supply chain!
Crude materials for further processing: +59.6%
Foodstuffs and feedstuffs +19.2%
Crude energy materials +129.5%
Posted by: Disagreeing at January 17, 2008 06:09 AM
I like your post but it has a common mistake normally made within the academia that I think it is relevant for the discussion. Please, correct me if I am wrong.
You can't just compare 10 yrs nominal bonds versus 10 years Tips to extract inflation expectations (besides the usual liquidity and inflation risk premiums that complicates that comparison). There is a substantial difference in duration between the two bonds due to the way TIPS are structured. Once you duration adjust the respective yields and take into consideration differences in liquidity and inflation risk premiums, the graph for inflation expectations look quite different from the one you showed here.
Anyway, I think it is a relevant issue right now and we should be careful with any quick comparison.
By the way, I like your blog a lot.
Posted by: Alejandro at January 17, 2008 07:17 AM
Perhaps you could clarify your point, Alejandro. Are you stating that the way that the Fed constructs a "constant maturity yield" from coupons and times to maturity is flawed?
Posted by: JDH at January 17, 2008 07:41 AM
Professor, there's nothing wrong with constant maturity yields from the Fed.
The first point I want to make can be seen in the following link: http://www.clevelandfed.org/research/inflation/TIpS/index.cfm
You can see there that once you adjust for liquidity and inflation premium, the chart looks very different and it is no longer that steady lately.
The second point is that people tend to compare the current 10 year nominal on the run treasury bond with the 10 year Tips to extract inflation expectations and that is wrong due to the larger duration of TIPS. Even if you use constant maturity yields for both from the Fed (and please correct me here if I am mistaken), there is an important discrepancy in durations which obscures the comparison.
Posted by: Alejandro at January 17, 2008 08:25 AM
Using the effective T-bill rate as some kind of 'proof' that 'everybody' knows that inflation really isn't what is, seems awfully sloppy.
The obvious cause of the low rate of return on T-Bills is people jumping out of stocks (which are losing what 2% a week right now!?) and into T-Bills. No one is looking at whether that is a 'good' rate of return, they are looking at whether it is better than what they would get if they left in the stock market.
That in no way proves that the 'all seeing all knowing market' has 'told us' that inflation is really contained.
All it tells us is that lots of people think 3% positive is a pretty good return right now. And they are right.
The rate of return on T-Bills is not causatively coupled to inflation. Your argument confuses causation and correlation, and I can assure you that even a .90 R^2 has periods where the two vars do different things.
Your usually better than this.
Posted by: Mike at January 17, 2008 08:26 AM
Mike, we're not talking about T-bills (maturity less than 1 year) but 10-year bonds. Are you saying that people are so concerned that they want to lock in a negative real return for 10 years? Surely nominal T-bill yields will go up substantially some time over the next 10 years if inflation continues at a 4% rate, in which case you're better buying T-bills rather than bonds today. It's the bond yield I'm calling attention to. And you don't mention the nominal-TIPS spread.
Posted by: JDH at January 17, 2008 08:41 AM
Alejandro, my understanding is that the Cleveland adjustments are precisely for the premia that you say you are abstracting from. And my graph is constructed from the constant-maturity 10-year nominal and TIPS.
Posted by: JDH at January 17, 2008 08:44 AM
I previously charted the same TIPS/Treasury spread you show but shortened the vertical scale to better see what has happened since the financial crisis began in Aug. Inflation expectations have increased since the beginning and end of August. The increase may or may not be material, but I think the spread is likely to continue to drift upward the more the Fed contemplates additional rate cuts.
Posted by: GWG at January 17, 2008 10:30 AM
in your answer to Mike you asked why capital fleeing the stock market would buy 10 year bonds rather than 1 year T-bills.
As an investor I can assure you that I don't normally plan on holding bonds to maturity, but I like the idea that I could if the price development isn't favorable. As my broker puts it: You are being paid to wait.
So, if I was fleeing the stockmarket today, I would only buy 1 year T-bills if I was 100% sure that the stockmarket would be booming in less than one year (I'm not!).
If I thought that we are heading into a multi-year stockmarket slump (note: were talking stock prices, not recession here), then I would buy bonds that had at least 5 years maturity, but preferrably less than 30 years, and that had maximum liquidity and then third I'd look for yeild. That tends to favor the 10 year bond.
Posted by: bellanson at January 17, 2008 10:32 AM
A couple of points about implied inflation expectations from TIPS:
(1) The Fed itself is a big (in fact the biggest) holder of treasuries (as I discuss in more detail at http://reservedplace.blogspot.com), so treasury yields can be expected to anticipate Fed transactions. In his well-known speech in November 2001 (which I believe was a big reason Bernanke was appointed Fed Chairman, and in which incidentally he referred to the Michigan survey of long term inflation expectations - currently 3.5%), Bernanke offered to increase Fed purchases of long bonds in a slump. The market may well be discounting a growing probablility of this (the Fed do buy TIPS, but less so, and would probably start Bernanke's plan with shorter treasuries).
(2) The implied measure of inflation is the cpi, and this has a history of being lowered by methodological change. If inflation rises awkwardly when growth is slow and entitlement spending high, there will be pressure for further cpi-lowering fiddles. Treasury prices may also discount this.
Posted by: RebelEconomist at January 17, 2008 01:03 PM
Charles there was no post Korea inflation spike. Post conflict inflation occurs when governments try to print money to pay off war debt. They don't magically appear just because a war happened. This is exactly what happened in post WW 1 Germany
Posted by: oops at January 18, 2008 12:04 PM
You said "serial correlation" where you meant "serial autocorrelation". Correlation has to be with something and in this case it is with itself, in its past data.
Posted by: W P Gardner at January 18, 2008 12:14 PM
I do appreciate Menzie and your efforts in maintaining Econobrowser. Its my absolute favorite for eclectic economic discussion.
I do have a very sore spot relative to CPI however. Manipulation of these statistics via hedonics, re-definition, and re-weighting has, in my opinion, made CPI a convenient fiction enabling inflation adjusted entitlements to be eroded, and inflationary monetary policy to be obscured. For instance, if CPI-U was currently calculated with the 1980 CPI methodology, it currently would be approaching 12%. (See John Williams at http://www.shadowstats.com/alternate_data ).
Or, you might look at The Economist commodity price index, which lists the last year's US Dollar denominated commodity inflation as:
- All Items - 28.4%
Euro denominated commodity inflation for the same period:
- All Items - 13.5%
At its core, (no pun intended), CPI manipulation allows the reporting of real GDP to continue to be reported as positive, while in reality, its been almost continually negative since the second quarter of 2000.
Last, Treasury and Agency bonds sell at very low interest because of supply and demand. IE: There are too many dollars chasing too few secure investment vehicles, driving yield down. As long as the FED continues to increase the money supply (M3 dollar growth is currently approaching 17%/yr), US government and agency securities will continue to have low yields. America pays for the low yield on its government securities in inflation and real asset value erosion. Eventually the general public will wake up and demand action. Voelker, where are you? We need you now!
Posted by: MarkS at January 18, 2008 12:18 PM
W.P. Gardner, here's a link to the world's leading graduate text on serial correlation, if you're curious to learn more about the topic.
Posted by: JDH at January 18, 2008 12:40 PM
I don't understand how increasing money supply causes real estate value erosion.
Please add color.
Posted by: KevinM at January 20, 2008 09:57 AM
I wrote real asset value erosion not real estate erosion. However, for discussion purposes, real estate is convenient.
Increased money supply results in lower interest rates and higher risk tolerance. In the U.S. and some countries in Western Europe, this has resulted in unprecedented real estate development and speculation. I have recently read reports that the inventory of private housing in the U.S. increased about 45% from 2000-2007. This increase in supply has swamped demand, and has resulted in the recent precipitous fall in real estate market values, (currently about -7% nationwide on a nominal basis, and about -20% if adjusted for the drop in the dollar index).
In a nutshell, over-expansion of the money supply dilutes real asset values everywhere in the economy. Think of the dollar as a share of stock (a security) in the American economy.
Posted by: MarkS at January 21, 2008 11:58 AM
Oops, you're correct that there wasn't an inflation spike post-Korean war. That inflation spike occurred during the war (see Big Picture). The reasons why the spike occurred earlier than one expects it to occur are debatable. But wars do tend to distort the relationship between price and scarcity.
Posted by: Charles at January 21, 2008 12:04 PM