Home > Good One, Investing, Theory, Trading > How To Exceed Expectations (Or At Least Keep Up)

How To Exceed Expectations (Or At Least Keep Up)

As we wait for something interesting to happen in the markets – or at least for some volume! – I thought I would write about a way that investors can, should they choose to, invest in a security that is directly linked to inflation expectations. Tomorrow there is useful economic data in the form of Retail Sales (Consensus: 0.3%/0.3% ex-auto) and Initial Claims (Consensus: 375k vs 372k last), although this last is subject to huge error bars because the weeks just before and just after the new year are very hard to seasonally adjust. Still, it’s data. Otherwise, the market continues to chop around with little volume and seemingly little conviction. It’s a good time to consider other approaches, so I am going to do that today.

With monetary policymakers in virtually every corner of the globe furious pumping liquidity into the world’s economies, it is no surprise that many asset markets are not cheap. Equities are expensive, although less so than they were last year; nominal bonds are terribly expensive. Inflation-linked bonds generally sport real yields below zero (out to 10 years) and insubstantial real yields beyond that. Commodities look cheap as a whole, but even though commodity indices are as diversified as equity indices many investors have a hard time putting a huge weight in commodities because of the sense that they are “risky.” Corporate inflation-linked bonds are doubly expensive, with real rates quite low and credit spreads tighter than they should be given the economic outlook. How then can an investor protect him/herself from the possibility of inflation moving higher?

I have been an advocate for commodity indices, of course, which tend to do well when real yields are low and in the early stages of inflationary surprises. But there is another way now that retail investors can fairly easily be long inflation expectations.

First, let me explain some basics. When we talk about nominal yields, such as normal Treasuries have, we recognize that they are made up of several parts. If I borrow money from you, you will first assess the real cost of money – how much more stuff do you want to have at the end of the loan, in order to convince you to defer consumption and lend me the money? That is the real interest rate. Second, you will evaluate how much less the dollars you receive from me at the end of the deal are likely to be worth, compared to the dollars you pay me at the beginning of the deal. This is an adjustment for expected inflation. There is a third adjustment, probably, that relates to the uncertainty of the real return when the nominal yield is fixed; this extra premium is called the inflation risk premium.[1] The Fisher equation is approximately:

y=r + i,

where y is the nominal yield, r is real yield, and i is expected inflation plus the risk premium. Because this latter quantity is what you need to realize if you buy an inflation-indexed bond with a real yield of r, in order to break even against a nominal investment that pays a yield of y, this is called breakeven inflation, or BEI.

This background is necessary to understand the following statement: TIPS are not “inflation-protected” in the sense that they do better when inflation rises. TIPS are real rate instruments, whose real price depends only on the real yield to maturity. The nominal value of a TIPS bond depends on the actual inflation realized over time, but this just means that a TIPS bond is immune to inflation. The real return of a TIPS bond depends only on the yield of the bond at purchase, if it is held to maturity.[2]

And so, a TIPS bond is just like a nominal bond in the sense that if its yield rises, its price falls. The Barclays Capital 1-10y TIPS Index returned 9.00% for 2011. That wasn’t because inflation was 9%, but rather it was mostly because real yields fell over the course of the year. The flip side is also true: if real yields rise, then TIPS will decline in value (although as I said, if held to maturity you will receive the real yield the bond sported when you bought it). And guess what will happen when inflation really picks up? You got it: real yields, along with nominal yields, will likely rise. While real yields should rise less than nominal yields in such a circumstance, it will not feel like “inflation protection” if your TIPS lose 9% when inflation is positive 4%.[3]

Now, I’m concerned about inflation, and while I think TIPS will beat nominal bonds handily over the next five years they may both have negative returns. I could simply short nominal bonds (and I have, via the TBF ETF), but if I am wrong – or if the Fed simply holds down nominal yields forever – that won’t produce the outcome I want. I would like to be long “i”, or inflation expectations. An institutional investor can do that by buying inflation swaps, or buying TIPS and shorting nominal bonds. A retail investor can buy a TIPS ETF (such as TIP) and short nominal bonds with a different ETF (such as TBF), but this is clunky, and since the durations may not match up well it requires a fair amount of work to get the right hedge. But there’s another way, which I will discuss in one moment.

Before I do, let me show one or two charts as context for what I have said recently. Quoting again from my week-ago comment,

It is incredible to me that with monetary conditions as accommodative, globally, as they have been in decades, inflation swaps are still nearly as cheap as they have ever been. That’s truly striking. What is the 10-year downside to a long inflation position from these levels? One-half percent per annum? Seventy-five basis points per annum? How low can inflation be over the next 10 years, especially with central banks apparently willing (and even anxious) to produce as much liquidity as is needed?

Let me provide a graphical answer to that question.

The chart below (source: Shiller and BLS) shows compounded 10-year inflation rates since the late 1800s.

Compounded inflation below 2% has been very rare since 1914, and the Fed is clearly leaning one way here.

I have drawn two lines on this chart. The vertical line indicates the date of the formation of the Federal Reserve; the horizontal line shows the current level of 10-year inflation breakevens (Treasury yields minus TIPS yields). Since the formation of the Federal Reserve, you can see that a 10-year period of inflation below 2% has been exquisitely rare, with the exception of the Depression when the Federal Reserve erred and tightened policy. So, if you are buying inflation below 2%, your realistic downside (especially with a Chairman who is acutely aware of the Fed’s failing in the Great Depression) over ten years is probably on the order of 50bps.[4]

The institution of the Federal Reserve created an institution whose purpose is to prevent deflationary depressions, and who has historically pushed prices higher – sometimes gently, and sometimes not so gently – over a long period of time. What may be surprising to see is how oftenwe have experienced periods of high inflation compared to episodes of tame inflation. While this histogram isn’t necessarily the purest way to address that question, since the periods overlap, it gives some sense for how frequent the “tails” of inflation are:

A frequency distribution breakdown of the previous chart (since 1914).

The maximum 10-year inflation rate was 8.8%, with about 30% of all observations above 5%. Note that these are not 1-year inflation numbers, but 10-year compounded inflation. The compounding matters. 2% compounded for 10 years is 21.9%. 8% compounded for 10 years is 115.9%.

This is what you’ll get if you own TIPS for 10 years. You’ll get the real yield of TIPS (currently negative out to 10 years) plus compounded inflation. If you think we can get 6% or 8% inflation, then the fact that the real yield is 0% instead of 1% isn’t that big a deal. But I would like to put on a trade that responds when inflation expectations themselves start to rise.

Deutsche Bank recently issued a pair of PowerShares exchange-traded notes (ETNs) that trade with the symbols INFL and DEFL. Information, and the prospectus, can be found at links here, here, and here. In full disclosure, I have bought some INFL for my own portfolio.

These instruments are an interesting way for retail investors to play for a potential increase in inflation expectations. The notes, which are issued by Deutsche and so are exposed to Deutsche Bank credit,[5] are designed so that they expect to rise $0.10 if inflation expectations rise 0.01% (1 basis point – I am here, and henceforward, speaking of INFL although the inverse holds for DEFL). So a 1bp rise in inflation expectations equals (if the security is at $50, as it is now) a 0.20% rise in the security’s price. Put another way, the buyer of INFL has roughly a 20 modified duration, which is pretty long for a bond-like instrument. The underlying index consists of 5-year, 10-year, and 30-year TIPS and short positions in 5-year Note, 10-year Note, and Ultra Treasury futures, in roughly the proportions TIPS are represented in the bond universe. The ETN also earns a T-Bill return and carries fees of 0.75% per annum.

The securities are supposed to maintain that 1bp=$0.10 relationship even as price rises and falls within some bounds, so that if INFL declines to $30, the modified duration rises to 33.

(Because this would eventually cause modified duration to head towards infinity as the price declines, the securities have a feature that causes the security to split if the price goes above $100 or reverse-split if the price goes below $25; however, the 1bp:$0.10 relationship would remain the same, so that if the price declines below $25 for a few days your modified duration will suddenly decline from 40 to 20; if the price rises above $100 your duration will go from 10 to 20. This is not necessarily a bad feature, since it means you will eventually get longer in a rally, but it isn’t analogous to normal convexity since with normal positive convexity you would get long in a rally, then less-long as price declined again. In this case, if price went to $100 and then reversed, you’d essentially have double the duration on the way down. So I suggest keeping a close eye on the ETN and being sure to adjust your exposure manually from time to time to remain within your risk tolerance. Having your exposure change via a split is also quintessentially unlike a normal equity’s behavior in a split, in which your exposure remains constant even though the number of shares doubles.)

There is a market-maker who presents orderly two-sided markets some $0.12 wide, or roughly 1.2bp on breakevens. That’s not bad at all – professional inflation traders don’t face markets much tighter than that. I don’t know the size commitment of the market-maker and have no direct knowledge of his dedication to maintaining these markets.

Are there warts to the structure? Sure. The weird ‘convexity’ is off-putting, although it can be managed. The float is currently smaller than I’d like (only $4mm each side at issue), although that’s true of any new ETF or ETN and I would expect it to increase over time. The use of futures instead of cash for the nominal position complicates analysis somewhat, although there is probably no easy way around it. There is an additional drag on return that comes from the financing of the long-TIPS position at LIBOR. Ordinarily, you would finance TIPS at the repo rate (about 20-40bps lower over time), and then you would earn a somewhat lower repo rate on the short Treasuries position. By selling futures, the repo rate earned on the bond short is embedded in the futures convergence, which makes it quite difficult to analyze the true total cost of the structure. The long-INFL investor earns T-Bills, plus implied repo on the futures position, minus Libor, minus fees. When T-bills are at zero and LIBOR at 0.20%, that plus the fees make the cost around 0.95% per annum. When T-Bills are at 1% and LIBOR at 1.30%, the net cost is essentially zero (1% + 1% – 1.30% – 0.75%). When T-Bills are at 2%, INFL should appreciate over time although that should be considered against your opportunity cost.[6]

Does Deutsche make money on the structure? Of course. But they make less they would with many structured notes, and these ETNs are, in my opinion, actually a useful way that retail investors can achieve a particular exposure. I don’t expect to hold this position until the ETNs mature, but with breakevens near 2% it is in my opinion a good way to bet on expectations rising over the next few months or years.


I appreciate that many of you have suggested to friends and colleagues that they follow this author – thank you very much! I also suspect that people you know may have concerns about inflation or be interested in learning more about inflation and hedging inflation, and that you can help them by introducing them to people who have something to say on the topic. I would appreciate your generous referral of these contacts of yours to our website, and would like to express my appreciation by sending a copy of my book, Maestro, My Ass!, to those of you who point people our way. (Please let your friends know that they should mention you when they fill out the contact form, so that I can know where to send my gift of thanks.) And thanks again.

[1] I say “probably” because although Fisher included the inflation risk premium in his work, it has never been clear to me why the provider of money would demand protection for the uncertainty of his real return while the user of money would not demand protection for the uncertainty of his real cost. To me, it isn’t clear which effect will dominate, and so I suspect it is entirely possible that the “inflation risk premium” can even be negative. We certainly see this phenomenon in some commodities futures curves. Anyway, since we can’t directly observe and separately trade the risk premium, it’s usually folded in with the breakeven.

[2] …unless there is sufficient deflation that the floor on the principal kicks in. This has never happened, but it adds a complexity to TIPS and requires that many statements about TIPS include the phrase “except if…”

[3] This fact leads some managers to make the false claim that “TIPS don’t hedge against inflation.” Of course they do. They hedge almost perfectly against inflation over the horizon from purchase to maturity. However, they do not hedge month-to-month or year-to-year inflation very well if they have a long time to maturity, because the short-term price change of the bond swamps the inflation accretion.

[4] Of course, past results are no guarantee of future returns. Anything can happen. 10-year inflation could go to -10%, or perhaps worse expectations could go to -10% even while inflation was rising, leading to a loss on the trade I’m about to mention. Consult your financial advisor.

[5] However, note that since the ETNs can be delivered in blocks to Deutsche on short notice, it is best thought of as short-term credit even though the notes themselves have a long maturity.

[6] I’m cuffing all of these relationships for the purposes of this column. The point is that right now the structure is about as expensive as it is going to be, and as rates rise INFL should have some positive net carry over time.

  1. bixbubba
    January 11, 2012 at 5:30 pm

    Hi Michael.

    Thanks for this! I’ve been waiting for it as you know. I’m hoping you can elaborate on fn 5. All of this is very tempting, but I want none of exposure to DB default risk. Can you say a bit about what kind of risk an investor is facing should the worst case scenario happen to DB?

    • January 11, 2012 at 6:00 pm

      So Deutsche is obligated to repay these notes just like any other notes. However, in the case of a normal Deutsche note, you have to wait for maturity before demanding payment. In this case, if you gather a whole bunch of the securities and present them, Deutsche must pay you right then. So if a big investor thought DB was about to go under, and the securities were trading at a discount, it could scoop up a bunch of them, present them to the firm, and get money 1 day before the company fails. You can’t do that with your size, nor I with mine, but because the risk to Deutsche is basically one day – if I present today, the money is in my account today – it’s much less risk. Once they actually default, of course, you’re pari passu with other investors but until that point it shouldn’t trade with much if any of a credit spread.

  2. onebir
    January 12, 2012 at 4:03 am

    Good to know about INFL. That said, part of the problem facing the US definition of ‘inflation’ has changed quite a bit. I think the same is true in Europe (notably the HICP measure uses a geometric mean, underweighting large changes relative to the old national indices which were mostly additive means).

    I suspect the 10 year inflation rate histogram would show a much more pronounced rightward bias using a more consistent inflation measure. 🙁

    • January 12, 2012 at 8:09 am

      You mean left-ward bias, right? If we were able to go back and measure inflation in the old days using modern methods?

      Or are you one of those who thinks that they still haven’t built a car to beat a Studebaker? 🙂

      • onebir
        January 31, 2012 at 3:52 am

        Well, if the car ‘costs’ $3.99 after hedonic adjustment, how could I go wrong… Tho perhaps the higher metal content of the Studebaker would make for a better investment 😉

        But seriously, as I understand it, if you’re going back to 1914, or even 30 years, there’s consistency issues with the published data. (Maybe Shiller’s data adjusts for that?)

        So either old methods carried forward (a la shadow stats) or new methods rolled back (probably impossible for lack of data)…

    • January 31, 2012 at 7:46 am

      Shadow Stats doesn’t even carry the methods forward, because the data was gathered in different ways and the underlying data he needs are no longer available. It’s yet another way we can tell he’s just making it up to try and sell his product.

      Hedonic adjustment is actually a very small effect in CPI, despite what hacks like John Williams will tell you. The aggregate effect of upward and downward quality adjustments is less than 0.1% per year. And that’s including all QUALITY adjustments, of which hedonic adjustments are just a small subset.

      You have a good point about consistency issues, but I’d apply the same reasoning I do to prove Williams is wrong to this issue: taken over 100 years, even a tiny error in measuring prices would create a huge error in the net outcome. 2% inflation for 100 years is 624% inflation. 3% inflation is 1,821% inflation, so the price level would be wrong by THREE TIMES and if we were measuring 2% and the actual number is 4% (4,950% inflation) our price level would suggest 1/8 of the actual change in prices. I suspect the errors over time tend to mean-revert, so that even though the methods in 1935 weren’t very good and may have been off by 1% or 2% in any given year, they didn’t accumulate.

      The simple fact is that it is easier to tell what inflation was over a longer term than what it was over a shorter term. Over 50 years, just look at what has happened to the price of wages, and houses, and you’ll get a pretty decent idea of what aggregate inflation was. But that relationship may not hold in the shorter-term.

  3. BOB
    January 12, 2012 at 10:31 am

    What specific investments would you recommend in commodities? I agree with you that they should have some upside, but I am unfamiliar with investing in them.

  1. No trackbacks yet.

Leave a Reply

%d bloggers like this: