Wednesday, April 29, 2009

GDP surprise: inflation

The economy contracted at a 6.1% annual pace in the first quarter, more than expected. The contraction was driven mainly by a huge inventory drawdown, which in turn was a reaction by business to the big slump in spending in the prior quarter. That will almost certainly turn around for the better in the current quarter, because spending is now growing again—real consumer spending rose at a 2.2% rate in the first quarter, and all the green shoots out there confirm that activity is bouncing in many areas of the economy.

The big news for me, however, was the 2.8% annualized rise in the GDP deflator, which is the broadest gauge of inflation. The chart above shows the year over year change in the deflator, and as you can see, there's not a lot going on, with inflation today about where it has been on average since the early 1990s. This is one of those "dogs that didn't bark," because conventional wisdom says that the huge (6.2% annualized) contraction in GDP over the past six months should have produced at least a little bit of deflation. That inflation is instead still alive and well is big news, because it says the conventional view of inflation is wrong. Inflation is not, as most people think, a function of the strength of the economy, but rather, as Milton Friedman says, a monetary phenomenon. Money has not been tight enough in recent years to bring inflation down. Since monetary policy is now quite accommodative, we should expect to see inflation drifting higher over the course of the year. This runs counter to what the market and the Fed are expecting, and that could easily overshadow the economy's likely recovery.

TIPS still offer good protection against higher-than-expected inflation, since they are priced to the assumption that inflation will be much lower than average over the next several years. Treasury bond yields, on the other hand, are going to be under tremendous pressure to rise.

Full disclosure: I am long TIP and TIPS, and long TBT at the time of this writing.

13 comments:

Mark Gerber said...

Hi Scott,
Thanks as always for your great blog. I greatly appreciate your economic analysis and charts, and I sure enjoy your photos and Argentina travel logs.

I have a request for you if you would be so kind: Would you please provide a tutorial about the various measures of inflation? For example, the GDP/CPE deflator and CPI seem to be track each other fairly well, but these produce a vastly different result from the pre-Clinton era CPI (ie. pre-Boskin commission CPI), as demonstrated by the chart at http://www.shadowstats.com/. The current CPI metrics (CPI & CPE deflator) seem to produce much lower inflation results than older CPI metrics (pre-Boskin and pre-Regan). I know one big difference in the current CPI formula is the hedonics and geometric mean, but it seems to me that most of the reduction in inflation in the newer CPI calculations come from changes to the housing component, imputed rents in particular. The bottom line is I wonder why you believe the CPE deflator is more correct than the pre-Boskin CPI formulation?

dave said...

Scott,
I too share your concern about inflation.A $900 gold price signals that inflationary pressures are already in the system, once the economy and demand picks up the measures of inflation are going to move up at an alarming pace.

The problem with quantitative easing is deciding when to stop.What signal does the FED have that will tell them when they have gone too far in one direction or the other?

Brian said...

Hi Scott -

The Dick Morris column that Drudge links to today predicts inflation rise - do you think this is inevitable? Perhaps just moderate inflation? How do we now compare to the economic models of Argentina and others?

Scott Grannis said...

dave, you seem to be forgetting that we have brilliant people running Treasury and the Fed, and surely they will be smart enough to figure out how to withdraw the trillions they have injected at just the right time to avoid any inflation problems. (just kidding of course)

I fully expect that the Fed will make a mistake (yet another in a series) going forward, since what they are likely going to watch for as a signal (positive GDP growth) is going to come out with a significant delay. Ideally they should be watching real time indicators such as gold, commodity prices, the yield curve, spreads, etc. Those indicators are saying they ought to begin withdrawing liquidity soon, if not now.

dr. j said...

Is it possible for you to post your holdings as you did today? I found it helpful. I have just started buying the DBC and DBA as a way to benefit from inflation. I also use a fund that hedges with TIPS and is short the 30 year treasury. Given your experience in your previous life with WAM, what fixed income categories, if any, look to hold value at this time? What would you recommend for income investors?

ps. My comments from a couple days ago should have read the "fall of 2007", not 2008. It must be a post 50 thing.

Mark A. Sadowski said...

Scott,
I noticed the odd GDP deflator result myself. All the odder still when one considers the following. There are 14 subcomponents to the GDP implicit price deflator and 10 out of the 14 were negative in the first quarter. The other four, services (+1.1%). equipment and software (+0.5%), national defense (+2.4%) and federal nondefense (+2.7%) were all less than the aggregate implicit price deflator.
How is this possible?

The answer is of course that the GDP underwent some drastic changes in terms of it's relative components. In other words the weighting of the price index changed in such a way that those components that had inflated more in the past acquired a greater weight and vice versa. The reality is when you look at the individual components we had significant price deflation in the first quarter.

From the Federal Reserve's perspective the target is the core PCE. The core PCE result for the first quarter is not in yet but it appears it will be about 1.6%. This is slightly below the implicit target of 2.0%.

So I have to take issue with the following statement:
"This is one of those "dogs that didn't bark," because conventional wisdom says that the huge (6.2% annualized) contraction in GDP over the past six months should have produced at least a little bit of deflation."

The current standard (New Keynesian) economic model of inflation looks at two things: output gaps, or changes in output gaps, and the anchoring of inflation expectations. How output gaps affect inflation depends on how well inflation expectations are anchored. In the United States inflation expectations (core PCE) seem to have been fairly well anchored at about 1.9% since 1994. In the period that preceded that, from the late 1960s through the early 1990s, inflation expectations were not well anchored at all.

During the period when inflation was not well anchored output gaps generated disinflation and so inflation fell (as for example during the early to mid 1980s). On the other hand, when the economy was above potential inflation accelerated (as for example during the late 1960s and during much of the 1970s).

A key reference to consider in this context is the non inflation accelerating unemployment rate or NAIRU. NAIRU has been variable over time. It was over 6% during most of the 1970s. Today it stands at 4.8%. A good rule of thumb for estimating the output gap is Okun's Law. The Okun multiplier for the United States is currently about 2.1. Thus with unemployment at 8.5% in March, the output gap in March was about 2.1 x (8.5 - 4.8) = 7.8%.

Once inflation expectations became well anchored in the mid 1990s, the effect of output gaps on inflation noticably changed. When unemployment decreased core inflation tended to be above core inflation expectations. When unemployment increased core inflation tended to be below core inflation expectations. Thus during this latter period there has been no evidence of disinflation or inflation acceleration, and consequently core PCE inflation has bounced around between 1.3% and 2.3%.

So as long as inflation expectations remain well anchored expect core inflation to be near 2%. With unemployment rising for the forseeable future it should be somewhat below 2%. When unemployment starts to fall it should be somewhat above 2%.

In short, the as yet unreported core PCE result in the first quarter will be as expected.

Scott Grannis said...

Brian: rising inflation is not inevitable yet. But the odds are definitely in favor I think. Gold at $900 is a good indicator of higher inflation. The dollar is not strong, even though it's up from its lows. The TIPS/Treasury spread is rising, but does not yet predict a new inflation high. The Fed has a huge job ahead, draining liquidity as the market's demand for liquidity declines. They will be reluctant to move too soon or too fast, for fear of damaging the economy; that is the key consideration. Already there are economists warning that they risk another Japan scenario (prolonged deflation/stagnation) if they tighten too soon.

Scott Grannis said...

dr j: I try to post my holdings when I make any comment on the attractiveness of an asset class or a particular stock. Assuming we get a moderate increase in inflation, this will prove very bad for Treasury bonds. TIPS should hold their own, though real yields might rise somewhat. The payoff for TIPS will come with time as the CPI rises; I wouldn't count on price gains at this point. High quality bonds should also hold their own, unless the rise in inflation is significant. High yields (junk) bonds should do very well, since a rising inflation environment is nirvana for those who have a lot of debt. Emerging market debt should also do very well for the same reason. Equities traditionally do poorly in times of rising inflation, but I don't think that will be the case this time, given how depressed valuations still are.

Scott Grannis said...

Mark: the PCE Core measure of inflation was 2.3% annualized in the first quarter by my calculation (I'm using the monthly data). I think that is remarkable given how weak the economy has been for the past six months. I think it is consistent with the fact that monetary policy has been very accommodative. It does not lend any support to the theory that inflation is a function of the economy's output gap.

As for inflation being a function of how well expectations are "anchored" I think that is just silly. It's one of those squishy variables that the Fed can use to explain away a mistake.

Mark A. Sadowski said...

Scott,
The March core PCE results were embargoed until this morning (8:30 AM EDT). PCE excluding energy and food averaged 118.195 in the first quarter 2009 versus 117.749 in the fourth quarter 2008 (Table 9):

http://www.bea.gov/newsreleases
/national/pi/2009/xls/pi0309.xls

The percentage increase was thus 0.379% and this works out to an increase of 1.524% at an annual rate (geometric of course), slightly lower than I was expecting (1.6%) but nevertheless quite close. (The overall PCE, which was released yesterday, was -1.0% at an annual rate of course.)

As for your assertion that this does not lend support to the idea that core PCE is a currently a function of changes in the output gap and inflation expectations I beg to differ. The empirical behavior of core PCE has been quite consistent since about 1994. I suspect most if not all professional macroeconomic forecasters form their core PCE forecasts based on these two inputs.

dave said...

Scott,
I beg your pardon,I completely forgot that the Fed is now completely staffed with genius caliber people.

I would however love to see some transparency from the Fed, maybe they could tell us what the dollar will be worth when they are through with their Q-easing.

PS I like your sense of humor!

Scott Grannis said...

Mark: For the three months ended March, the core PCE deflator rose at an annualized 2.3% rate. That's what I was referring to, not the change in the quarterly average.

Most economists probably do what you say, but I don't, and I think they will be proven wrong.

Former Deacon said...
This comment has been removed by the author.