Random Thoughts On Inflation

Includes: RINF
by: Scott Sumner

Over the past 7 years, I've frequently criticized the Fed's predictions for inflation and RGDP growth. This is not based on my skill as a forecaster, I don't have any. Rather I try to infer market forecasts, by looking at things like TIPS spreads (for inflation) and TIPS yields (for RGDP growth).

I do recognize that these market predictions can be flawed, for instance TIPS spreads fell to implausibly low levels during the banking crisis of late 2008, perhaps because T-bond prices were distorted by illiquidity in the asset markets. Nonetheless, during normal times, I think TIPS spreads can be useful. Certainly, the TIPS market correctly called the Fed's earlier over-optimism about achieving 2% inflation.

But now I'm a bit torn, as the current TIPS spreads, about 1.3% on 5-year bonds, seems implausibly low to me. TIPS spreads are based on the CPI, which tends to run higher than the PCE inflation rate targeted by the Fed. So they are implicitly calling for barely over 1% PCE inflation. Here's Steve Williamson discussing a recent post by Larry Summers:

We can use inflation swaps data, as Summers does; we can look at the break-even rates implied by the yields on nominal Treasury securities and TIPS; and we can look at survey measures. What do people who do forecasting for a living, and who have access to all of that data, say? The Philly Fed's most recent Survey of Professional Forecasters [SPF] has predictions of PCE headline inflation for 2016 and 2017, respectively, of 1.8% and 1.9%, which is pretty close to the 2% PCE inflation target. A Wall Street Journal survey shows a CPI inflation forecast that seems roughly consistent with the December FOMC projections for PCE inflation. So, it seems that "most available data," filtered through the minds and models of professional forecasters, suggests no less optimism than the FOMC is expressing in its projections, about achieving 2% inflation in the future.

This is an interesting way to frame the issue. Williamson talks about the fact that forecasters can incorporate market forecasts, whereas I'd tend to look at things in exactly the opposite direction. In an efficient market, asset prices reflect all available information, including the consensus of professional forecasters. I'd prefer the market forecast, if we could be sure that the asset prices/spreads actually reflected market expectations. But can we?

You might wonder why I am skeptical of the TIPS spreads, which seem to imply just over 1% PCE inflation over the next 5 years, given that PCE inflation has averaged just over 1% during the past 4 years. And indeed the trend has been downward. But I see this as reflecting the effects of lower commodity prices, especially oil. Since commodity prices tend to follow something close to a random walk, there's no reason to extrapolate those declines into the future. I have no idea where oil will be in 5 years, but the best guess is probably not too far from where it is now. In that case, we might prefer to extrapolate the CPE core inflation rate, which has averaged about 1.5% over the past 5 years. But even that may be a bit too low, as it includes a period where the dollar strengthened significantly in forex markets. On the other hand, maybe some of the impact of the strong dollar has not yet worked its way into core inflation, recall that PPP is a long-run concept.

After reading Williamson's post, I looked up the track record of the SPF, and it looks to me like it made roughly the same sorts of mistakes as the Fed. Here are the headline PCE inflation forecasts, with actual numbers in parentheses. The forecasts were for Q4 over Q4 inflation, made in Q4 right before the year being forecast:

2008:Q4 1.8% (1.2%)

2009:Q4 1.3% (1.3%)

2010:Q4 1.4% (2.7%)

2011:Q4 1.7% (1.8%)

2012:Q4 2.0% (1.2%)

2013:Q4 1.9% (1.1%)

2014:Q4 1.8% (0.4%) estimated based on November over November

Too small a sample to draw any conclusions, but this does support one of Williamson's complaints about Summers' post. Williamson criticized Summers' claim that the Fed treats the 2% inflation figure as a ceiling, not a target. Lots of my commenters agree on that point, partly based on the past few years. But I've never been fully convinced. Yes, the Fed has generally undershot inflation since 2008, but perhaps this was mostly an honest mistake. The fact that the SPF tends to have made very similar errors to the Fed in recent years, strongly suggests that these were in fact honest mistakes.

For what it's worth, here are some predictions:

1. Both the Fed and the SPF will continue to forecast roughly 2% inflation, in the years ahead.

2. At some point, the SPF will stop being biased (if they are currently biased) and they'll figure things out. At that point, PCE inflation will begin averaging 2%.

3. But not yet. I think both the Fed and the SPF still put too much weight on Phillips Curve models of inflation, and thus I put some weight on the TIPS spreads.

4. But the 5-year TIPS spread seems too low to me, so I will split the difference and forecast about 1.5% PCE inflation for the next couple years, and then closer to 2% after that.

5. However, when the next recession hits (and I have no idea when that will be) I'd expect inflation to again decline. And unless the Fed moves toward NGDP targeting (which makes inflation countercyclical), this undershoot may revive complaints that the Fed treats the 2% figure as a ceiling.

PS. If the Fed and the SPF are right about inflation expectations, then TIPS would seem like a much better investment than T-bonds. The expected 5-year return would be perhaps 80 basis points higher on TIPS, and they are if anything lower risk than T-bonds. The only downside is slightly less liquidity, but in an absolute sense, TIPS are still a very large, deep and liquid market. For commenters who work in finance, why would something like an insurance company hold T-bonds instead of TIPS?