Below is a summary (and extension) of my post-CPI tweets today. You can follow me @inflation_guy.
- CPI +0.3%/+0.1% with y/y core figure dropping to 1.9%. That will be only by a couple hundredths on rounding, but it’s still a decline.
- Looks like core was 0.129% rounded to 3 decimal places. y/y went from 1.956% to 1.933% so a marginal decline.
- RT of Bloomberg Markets @themoneygame: Consumer Price Changes By Item http://read.bi/1nx0sUf
- Core goods still -0.2%, core services still +2.7%, unchanged from last month. [ed note: I reversed these initially; corrected here]
- All of this a mild miss for the Street, which was looking for +0.19% or so, but I though the Street was more likely low.
- Major groups accel: Apparel, Recreation, Educ/Comm, Other (19.7%). Decel: Food/Bev, Transp, Med Care (38.9%), Housing flat.
- There’s your real story. Recent drivers: medical care, which is a base effect and oddly reversed. That’s temporary. Also>>
- >>big fall in non-rental/OER parts of housing: insurance, lodging away from home, appliances. Those are not as persistent as rents.
- Primary rents went to 3.153% from 3.058%; OER unch at 2.640% from 2.638%. The rest will mean-revert.
- College tuition and fes at 4.142% from 4.001%.
- 60.5% of all low-level categories accelerating (down from 70.5%). Still broad but not as broad.
- Actually looks like Median CPI could downtick today.
This is why I try very hard to resist the urge to forecast the monthly CPI, and admonish investors (and even traders) to resist trading on the data. Chairman Yellen is right about this: the data are noisy, so one month can be almost anywhere. This month, there was a reversal in the recent rise in y/y medical care inflation. But that rise was due to base effects, which aren’t going away, so forecasting medical care inflation to continue to accelerate is more a statement of mathematical likelihood than it is an economic forecast. And it’s all the more surprising then when it reverses.
This month’s figure makes it a fair bit harder for my forecast of near-3% for 2014 on core or median inflation to come to pass, although it bears noting that median inflation (even though it may downtick later today) is still within striking distance. Since median is currently the better measure, and will be for much of this year, I won’t back off my forecast yet. Another weak month, though, would cause me to ratchet down the target simply because it becomes harder to hit as time becomes shorter.
However, I expect several months this year will exceed +0.3% on core inflation. And it is worth remembering that core inflation faces easy year-ago comparisons for the rest of the year. In July of last year, the seasonally-adjusted m/m core inflation figure was +0.167%; in August it was +0.138%; in September it was 0.132%; in October +0.124%; in November +0.175%; and in December +0.101%. So, even if core inflation only averages +0.2% for the rest of the year, core will still be at 2.3% by year-end. If core inflation averages what it has been for the last four months, we’ll be at 2.4%. What that means is that (a) my forecast of something near 3% doesn’t represent a massive acceleration, although we only have half a year to get there, and (b) anyone forecasting less than 2.3% by year-end is actually forecasting a deceleration in inflation from recent trends.
The breadth indicators also took a mild breather this month, with the proportion of the CPI that is accelerating (looking at low-level categories) dropping to around 60% from around 70% in May. As with the other analysis, however, we should be careful not to read too much into one month since this figure also jumps around a lot. Interestingly, the proportion of categories where the year-on-year change is at least 2 standard deviations above zero – so that we can reject the ‘deflation’ meme for these categories – is basically unchanged from last month at 24%. As the chart below shows, we last saw a level this high in 2006, which is also the last time that core CPI ran at 3%.
Housing inflation is now back below my model’s projections, inflation breadth is still high, and the persistent parts of CPI are maintaining their levels or advancing while a few of the skittish parts are retreating (or at least not yet converging to the mean). There is nothing here to indicate that the three months of accelerating core CPI were the aberration; in fact to me it appears that the June figure was the aberration. That question will be answered over the balance of the year. In the meantime, inflation markets remain priced at levels so low that even if you’re wrong in betting on higher inflation, you don’t lose much but if you’re right, you do very well. In my view (although admittedly I may be biased), most investors remain significantly underweight protection against this particular risk.
Heading into the CPI print tomorrow, the market is firmly in “we don’t believe it” mode. Since the CPI report last month – which showed a third straight month of a surprising and surprisingly-broad uptick in prices – commodity prices are actually down 4% (basis the Bloomberg Commodity Index, formerly known as the DJ-UBS Commodity Index). Ten-year breakeven inflation is up 1-2bps since then, but there is still scant sign of alarm in global markets about the chances that the inflation upswing has arrived.
Essentially, no one believes that inflation is about to take root. Few people believe that inflation can take root. Indeed, our measure of inflation angst is near all-time lows (see chart, source Enduring Investments).
…which, of course, is exactly the reason you ought to be worried: because no one else is, and that’s precisely the time that often offers the most risk to being with the crowd, and the most reward from bucking it. And, with 10-year breakevens around 2.22%, the cost of protecting against that risk is quite low.
I suspect that one reason some investors are less concerned about this month’s CPI is that some short-term indicators are indicating that a correction in prices may be due. For example, the Billion Prices Project (which is now Price Stats, but still makes a daily series available at http://bpp.mit.edu/usa/) monthly inflation chart (shown below) suggests that inflation should retreat this month.
However, hold your horses: the BPP is forecasting non-seasonally-adjusted headline CPI. The June seasonals do have the tendency to subtract a bit less than 0.1% from the seasonally-adjusted number, which means that it’s not a bad bet that the non-seasonally-adjusted figure will show a smaller rise from May to June than we saw April to May, or March to April. Moreover, the BPP and other short-term ‘nowcasts’ of headline inflation are partly ebbing due to the recent sogginess in gasoline prices, which are 10 cents lower (and unseasonally so) than they were a month ago.
But that does not inform on core inflation. The last three months’ prints of seasonally-adjusted core CPI have been 0.204%, 0.236%, and 0.258%, which is a 2.8% annualized pace for the last quarter…and accelerating. Moreover, as I have previously documented the breadth of the inflation uptick is something that is different from the last few times we have seen mild acceleration of inflation.
None of that means that monthly core CPI will continue to accelerate this month. The consensus forecast of 0.19% implies year/year core CPI will accelerate, but will still round to 2.0%. But remember that the Cleveland Fed’s Median CPI, to which core CPI should be converging as the sequester/Medical Care effect fades, is at 2.3% and rising. We should not be at all surprised with a second 0.3% increase tomorrow.
But, judging from markets, we would be.
This is not to say I am forecasting it, because forecasting one month’s CPI is like forecasting a random number generator, but I think the odds of 0.3% are considerably higher than 0.1%. I am on record as saying that core or median inflation will get to nearly 3% by year-end, and I remain in that camp.
So, the Fed’s tightening is almost done.
Chairman Yellen informed Congress that a “high degree” of easing is needed given the slack in the labor market. This is in keeping with the Fed’s ongoing thematic presentation of “tapering is not tightening,” but of course tapering is indeed tightening. Call it “easing less” if you like, but going from “providing lots of liquidity” to “providing less liquidity” to “providing no added liquidity” is tightening.
I would argue that providing no added liquidity – which is where the Fed is headed, with the taper due to be completed in the autumn – is neutral policy, not an easy policy. But the Fed, like many observers, confuses the level of interest rates with the degree of accommodation. That is confusing a price (the interest rate) with a flow, but it seems not to bother them very much. (I explain the distinction, which is crucial to monetary policymaking, in this article.)
Now, whatever the Chairman thinks she’s saying, what she means is that the Fed isn’t going to be raising interest rates soon. This is partly because the main tool they had been planning to use, the reverse repo facility, isn’t as simple a solution as they believed at first. This isn’t terribly surprising; as I (and others) have been pointing out in presentations and articles for a while it isn’t trivially easy to drain $2 trillion in reverse repo transactions, even if you can do $2 billion with ease. The pattern is familiar, and should be mildly discomfiting:
- At first, the Fed thought to unwind the massive purchases of Treasuries by simply selling them. The original argument was that the Fed pushed rates lower by buying Treasuries, but selling them wouldn’t raise interest rates. This sort of perpetual motion machine never made much sense, and at some point it became clear that if the Treasury started to unwind the SOMA portfolio securities and rates rose, it would likely not be sufficient to drain all of the excess reserves, since the average selling price would most likely be lower than the average purchase price.
- The Fed then thought to just let the securities in the SOMA roll off. Then someone noticed that because of the TWIST program, the Fed doesn’t own many short-dated Treasuries, so that letting QE gradually drain itself would take more than a decade.
- No problem; we’ll just conduct massive reverse repo operations to drain a couple trillion dollars from the system. The link above shows that the Fed’s newly discovered skepticism on that matter; the website Sober Look recently had a good article on the topic as well.
None of this is surprising to people who actually have market experience; unfortunately, over the last decade or so the level of actual market expertise at the Federal Reserve has dropped significantly so they are re-discovering these things the hard way. Now, the focus is on interest on excess reserves (IOER) as the main tool for raising rates eventually.
All of this confusion is one reason that the Fed will move only slowly to ‘normalize’ interest rates. They’re simply not sure how they’ll do it. The problem with IOER is that we have no idea how sensitive the level of reserves it to the amount of interest paid on reserves…since we have never done this before. But to the Fed, that’s no problem because they don’t seem to care about reserves – they only care about the level of interest rates, which at the end of the day don’t matter nearly as much as the growth rate of the money supply.
And so US and UK money supply growth rates are both in the 6-7% range, and interestingly median inflation in the US recently accelerated to 2.3% while core inflation in the UK surprised everyone today by rising to 1.9% (as of April). Commercial bank credit growth in the US over the last 13 weeks has risen at a 10.4% pace, the highest rate since early 2008 (see chart, source Federal Reserve).
Slowing QE has not, evidently, slowed money supply growth, and this is one reason the Fed insists that tapering is not tightening. Unfortunately, this doesn’t mean that the Fed is right, but that they are wrong twice: first, tapering is tightening. Second, changing the pace of addition to reserves does not matter for growth in the money supply (and, hence, inflation) when there are enormous piles of inert reserves already. Picture a huge urn filled with coffee. The spigot at the bottom controls the pace at which coffee leaves the urn, and adding more coffee to the top of the urn has essentially no effect.
So money supply growth, and corporate loan growth, is currently not under control of the Fed in any way. Interest rates are under their control, but interest rates don’t cause changes in the money supply but rather the other way around. Here is another analogy: a robust harvest of corn pushes corn prices lower, but if the government officially sets the price of corn very low it does not cause a robust harvest of corn. This is exactly what the Fed is trying to do if they attempt to control the money supply by changing interest rates.
It actually is worse than this. Raising interest rates will tend to increase money velocity, a relationship which has held very well for the last two decades. I have written about this quite a bit in the past (see for one example this article from last September), but I – like many monetary economists – have often struggled with the fact that there was a regime shift in the early 1990s which messes up the beauty of this fit (see chart, source Enduring Investments).
We have recently resolved much of this problem in our own modeling. The following chart uses three (unstated here, but included in our quarterly inflation outlook to clients) inputs to model M2 velocity, and the regime shift is largely absent. Suffice to say that with a model that makes sense and fits a much wider range of history, we are even more confident now that any Fed move to hike interest rates, rather than to drain reserves, would be a mistake.
The bottom line is that it is good news that Yellen is not planning to hike interest rates soon. It is bad news that she is not planning to drain reserves any time soon. But the Fed is perilously close to making its big policy error of this cycle. Stay tuned.
A very common refrain among stock market bulls these days – and an objection some made to my remarks yesterday that markets are still not making sense – is that the low level of interest rates warrants a high multiple, since future earnings are being discounted at a lower interest rate.
My usual response, and the response from far more educated people than me, like Cliff Asness who published “Fight the Fed Model” back in 2003, is that low interest rates explain high multiples, but they do not justify high multiples. High multiples have always historically been followed – whether explained by low interest rates or not – by poor returns, so it does no good to say “multiples are high because rates are low.” Either way, when multiples are high you are supposed to disinvest.
But I thought it would also be useful, for people who are not as familiar with the argument and only familiar with the sound bite, to see the actual data behind the proposition. So, below, I have a chart of year-end Shiller P/E ratios, since 1900, plotted against year-end 10-year nominal interest rates.
Note that it is generally true that lower nominal interest rates are associated with higher multiples, although it is far more clear that higher nominal interest rates are associated with lower multiples, whether we are talking about the long tail to the right (obviously from the early 1980s) or the smaller tail in the middle that dates from around 1920 (when 5% was thought to be a pretty high interest rate). But, either way, the current multiples represent high valuations whether you compare them to high-rate periods or low-rate periods. The exception is clearly from the late 1990s, when the long downtrend in interest rates helped spark a bubble, and incidentally spurred the first widespread discussion/excuse of the so-called “Fed model.” If you take out that bubble, and you take out the 1980s high-rates tail, then there is left just a cloud of points although there does seem to be some mild slope to it from lower-right to upper-left.
But in short, the data is hardly crystal clear in suggesting that low interest rates can explain these multiples, never mind justify them.
More interesting is what you get if you compare P/E ratios to real rates. Because equities are real assets, you should technically use a real discount rate. Since real economic growth in earnings should be reflected in higher real interest rates generally, only the incremental real growth in earnings should be discounted into higher values today. This eliminates, in other words, some of the ‘money illusion’ aspect of the behavior of equity multiples.
I haven’t seen a chart like this before, probably because the history of real interest rates in the U.S. only dates to 1997. However, using a model developed by Enduring Investments (and used as part of one of our investment strategies), we can translate those historical nominal rates into the real rates we would have expected to see, and that allows us to produce this chart of year-end Shiller P/E ratios, since 1900, plotted against year-end 10-year real interest rates – using Enduring’s model until 1997, and actual 10-year real interest rates thereafter.
I find this picture much more interesting, because there seems to be almost no directionality to it at all. The ‘tail’ at upper right comes from the late 1990s, when again we had the equity bubble but we also had real rates that were higher than at equilibrium since the Treasury’s TIPS program was still new and TIPS were very cheap. But other than that tail, there is simply no trend. The r-squared is 0.02 and the slope of the regression line is not statistically different from zero.
And, in that context, we can again see more clearly that the current point is simply at the high end of the cloud of historical points. The low level of real interest rates – actually quite a bit higher than they were last year – is of no help whatsoever.
None of that should be particularly surprising, except for the buy-and-hope crowd. But I thought it constructive to show the charts for your amusement and/or edification.
The Employment number these days is sometimes less interesting than the response of the markets to the number over the ensuing few days. That may or may not be the case here. Thursday’s Employment report was stronger than expected, although right in line with the sorts of numbers we have had, and should expect to have, in the middle of an expansion.
As the chart illustrates, we have been running at about the rate of 200k per month for the last several years, averaged over a full year. I first pointed out last year that this is about the maximum pace our economy is likely to be able to sustain, although in the bubble-fueled expansion of the late 1990s the average got up to around 280k. So Thursday’s 288k is likely to be either revised lower, or followed by some weaker figures going forward, but is fairly unlikely to be followed by stronger numbers.
This is why the lament about the weak job growth is so interesting. It isn’t really very weak at all, historically. It’s merely that people (that is, economists and politicians) were anticipating that the horrible recession would be followed by an awe-inspiring expansion.
The fact that it has not been is itself informative, although you are unlikely to see economists drawing the interesting conclusion here. That’s because they don’t really understand the question, which is “is U.S. growth unit root?” To remember why this really matters, look back at my article from 2010: “The Root of the Problem.” Quoting from that article:
“what is important to understand is this: if economic output is not unit root but is rather trend-stationary, then over time the economy will tend to return to the trend level of output. If economic output is unit root, then a shock to the economy such as we have experienced will not naturally be followed by a return to the prior level of output.”
In other words, if growth is unit root, then we should expect that expansions should be roughly as robust when they follow economic collapses as when they follow mild downturns. And that is exactly what we are seeing in the steady but uninspiring job growth, and the steady if not-unusual return to normalcy in the Unemployment Rate (once we adjust for the participation rate). So, the data seem to suggest that growth is approximately unit root, which matters because among other things it makes any Keynesian prescriptions problematic – if there is no such thing as “trend growth” then the whole notion of an output gap gets weird. A gap? A gap to what?
Now, it is still interesting to look at how markets reacted. Bonds initially sold off, as would be expected if the Fed cared about the Unemployment Rate or the output gap being closed, but then rallied as (presumably) investors discounted the idea that the Federal Reserve is going to move pre-emptively to restrain inflation in this cycle. Equities, on the other hand, had a knee-jerk selloff on that idea (less Fed accommodation) but then rallied the rest of the day on Thursday before retracing a good part of that gain today. It is unclear to me just what news can actually be better than what is already impounded in stock prices. If the answer is “not very darn much,” then the natural reaction should be for the market to tend to react negatively to news even if it continues to drift higher in the absence of news. But that is counterfactual to what happened on Thursday/Monday. I don’t like to read too much into any day’s trading, but that is interesting.
Commodities were roughly unchanged on Thursday, but fell back strongly today. Well, a 1.2% decline in the Bloomberg Commodity Index (formerly the DJ-UBS Commodity Index) isn’t exactly a rout, but since commodities have been slowly rallying for a while this represents the worst selloff since March. The 5-day selloff in commodities, a lusty 2.4%, is the worst since January. Yes, commodities have been rallying, and yet the year-to-date change in the Bloomberg Commodity Index is only 2% more than the rise in M2 over the same period (5.5% versus 3.5%), which means the terribly oversold condition of commodities – especially when compared to other real assets – has only barely begun to be corrected.
I do not really understand why the mild concern over inflation that developed recently after three alarming CPI reports in a row has vanished so suddenly. We can see it in the commodity decline, and the recent rise in implied core inflation that I have documented recently (see “Awareness of Inflation, But No Fear Yet”) has largely reversed: currently, implied 1 year core inflation is only 2.15%, which is lower than current median inflation – implying that the central tendency of inflation will actually decline from current levels.
I don’t see any reason for such sanguinity. Money supply growth remains around 7%, and y/y credit growth is back around 5%. I am not a Keynesian, and I believe that growth doesn’t matter (much) for inflation, but the recent tightening of labor markets should make a Keynesian believe that inflation is closer, not further away! If one is inclined to give credit in advance to the Federal Reserve, and assume that the Committee will move pre-emptively to restrain inflation – and if you are assuming that core inflation will be lower in a year from where it (or median inflation, which is currently a better measure of “core” inflation) is now, you must be assuming preemption – then I suppose you might think that 2.15% core is roughly the right level.
But even there, one would have to assume that policy could affect inflation instantly. Inflation has momentum, and it takes time for policy – even once implemented, of which there is no sign yet – to have an effect on the trajectory of inflation. Maybe there can be an argument that 2-year forward or 3-year forward core inflation might be restrained by a pre-emptive Fed. But I can’t see that argument for year-ahead inflation.
Of course, markets don’t always have to make sense. We have certainly learned this in spades over the last decade! I suppose that saying markets aren’t making a lot of sense right now is merely a headline of the “dog bites man” variety. The real shocker, the “man bites dog” headline, would be if they started making sense again.
In keeping with the topic of the month, I present this chart.
I really wanted to make the x-axis the compounded inflation rate since the World Cup began, but the data is just too difficult to find for many of these countries. Nevertheless, we see the broad outlines of the thesis in this chart. If you want to be excellent at soccer, inflate your economy.
The correlation between soccer wins and inflation (I arbitrarily decided to only include countries which have appeared in eight or more World Cups, so that there is some chance that they have some wins) is only 0.31, but notice the two blue dots at the upper left. I would argue that at least Germany has an inflation-driven history, although since the 1980s they have had fairly low inflation. One might argue the same with Italy, albeit to a lesser extent. If we exclude those two aberrations, the correlation rises to a whopping 0.67!
Ok, sure, this is somewhat spurious – it is largely driven by the fact that two of the winningest teams are Brazil and Argentina, which have quite a history of inflation as well as of soccer. But if the ECB discovers this, it should make sure all of the retail shops in Europe know…and they’ll have widespread support for inflation.
Suddenly, there is a bunch of talk about inflation. From analysts like Grant Williams to media outlets like MarketWatch and the Wall Street Journal (to be sure, the financial media still tell us not to worry about inflation and keep on buying ‘dem stocks, such as Barron’s argues here), and even Wall Street economists like those from Soc Gen and Deutsche Bank…just two name two of many Johnny-come-latelys.
It is a little surprising how rapidly the articles about possibly higher inflation started showing up in the media after we had a bottoming in the core measures. Sure, it was easy to project the bottoming in those core measures if you were paying attention to the base effects and noticing that the measures of central tendency that are more immune to those base effects never decelerated much (see median CPI), but still somehow a lot of people were taken by surprise if the uptick in media stories is any indication.
I actually have an offbeat read of that phenomenon, though. I think that many of these analysts, media outlets, and economists just want to have some record of being on the inflation story at a time they consider early. Interestingly enough, while there is no doubt that the volume of inflation coverage is up in the days since the CPI report, there is still no general alarm. The chart below from Google Trends shows the relative trend in the search term “rising inflation.” It has shown absolutely nothing since the early days of extraordinary central bank intervention.
Now, I don’t really care very much when the fear of inflation broadens. It is the phenomenon of inflation, not the fear of it, which causes the most damage to society. However, there is no doubt that the fear of inflation definitely could cause damage to markets much sooner than inflation itself can. The concern has been rising in narrow pockets of the markets where inflation itself is actually traded, but because we trade headline inflation the information has been obscured. The chart below (source: Enduring Investments) shows the 1-year headline inflation swap, in black, which has risen from about 1.4% to 2.2% since November. But the green line shows the implied core inflation extracted from those swap quotes, and that line has risen from 1.2% in December to 2.6% or so now. That is far more significant – 2.6% core inflation over the next year would mean core PCE would exceed 2% by next spring. This is a very reasonable expectation, but as I said it is still only a narrow part of the market that is willing to bet that way.
If I was long equities – which I am not, as our four-asset-class model currently has only a 7.4% weight in stocks – then I would keep an eye on the search terms and for other anecdotal evidence that inflation fears are starting to actually rise among investors, rather than just being the probably-cynical musings of people who don’t want to be seen as having missed the signs (even if they don’t really believe it).