Inflation Expectations: A Primer 31 comments
an article to
-
Font Size:
-
Print
- TweetThis
By James Kwak
Only a few years ago, the accepted remedy for a recession was for the Federal Reserve to lower interest rates - namely, the Federal funds rate. Now, however, the economy has been stuck in recession for over fifteen months and the Federal funds rate has spent the last several months at zero. (The Fed funds rate cannot ordinarily be negative, because one bank won’t lend $100 to another bank and accept less than $100 in return; it always has the option of just holding onto its $100.) As a result, the Fed has resorted to other policy tools, most notably large-scale purchases of agency and Treasury securities, funded by creating money. (Here’s James Hamilton’s analysis.)
As the Fed’s monetary policy plays a more prominent role in the response to the economic crisis, there will be more talk of inflation or, more accurately, inflation expectations. While inflation is what affects the purchasing power of the money in your wallet, inflation expectations are what affect people’s behavior in ways that have a long-term economic impact. Take the case of wage negotiations, for example: a union that believes inflation will average 5% over the life of a contract will demand higher wage increases than a union that believes inflation will average only 1%. Once those higher wages are built into the contract, the employer is forced to raise prices in order to cover those wage increases, and inflation begins to ripple through the economy.
One of the major objectives of modern monetary policy is to control inflation expectations, because controlling inflation expectations is the first step to controlling inflation. If there is a short-term burst of inflation - as we had a year ago, if you look at headline inflation numbers that include the prices of food and energy - the macroeconomic consequences can be limited if people believe that the Fed can and will bring inflation under control.
Unfortunately, it is impossible to know exactly what people’s inflation expectations are; in fact, it may not even be a sensible question, since different people have different understandings of what inflation is. However, there are three main approaches to estimating inflation expectations.
1. Inflation-indexed government bonds. (If you need a refresher on how a bond works, read the first part of this article.) A traditional bond is a stream of payments that is fixed in nominal terms: for example, $100 in 10 years, and 6% interest, paid semi-annually ($3 every 6 months). Such a bond is not inflation-indexed; if inflation goes up, the purchasing power of that $100 goes down, and it’s too bad for the bondholder.
An inflation-indexed bond, by contrast, pays an amount that is indexed to some measure of inflation. In the U.S., where these bonds are called Treasury Inflation Protected Securities (TIPS), we use the Consumer Price Index. A TIPS bond may have a $100 face value and pay a 2% interest rate. However, every 6 months, that $100 face value is adjusted to reflect the change in the CPI, and the interest payment is calculated as a percentage of the adjusted value of the bond. Then, after 10 years, the bondholder gets back not $100, but $100 times the ratio between the CPI at the end of the period and the CPI at the beginning of the period. This way the bondholder is guaranteed a 2% real return (assuming he paid $100 for the bond), no matter what the rate of inflation is in the interim.
The implied inflation expectation, then, is the difference between the yield on an ordinary bond and the yield on an inflation-indexed bond with the same maturity. If the 5-year Treasury has a yield of 4% and the 5-year TIPS has a yield of 2%, then inflation expectations for the next five years are (about) 2% per year. The reasoning is that in order to buy the regular bond as opposed to the inflation-indexed bond, an investor has to be paid a higher yield to compensate him for the level of inflation that he expects.
Actually, in addition to expected inflation, the Treasury investor also has to be paid an inflation risk premium because, all things being equal, it is better not to have inflation risk than to have it. So the implied inflation expectation is actually slightly less than the spread between the regular and the inflation-indexed bonds. If you didn’t follow that, don’t worry, just remember that, roughly speaking, Treasury yield = TIPS yield + expected inflation.
2. Inflation swaps. These are a type of derivative contract, where the payments under the contract depend on the value of an inflation index, such as the CPI. The swap has a nominal value of, say, $100, but $100 never changes hands. Instead, at the end of some period of time, party A pays party B a fixed rate of interest on $100 - say 2.5% per year. At the end of the period, B pays A the cumulative percentage change in the inflation index over the period. Assuming A has $100 in his pocket, he has now hedged the inflation risk on that $100, because no matter what happens, at the end of the period he will get an amount that compensates him for the impact of inflation on his $100. The price of this hedge is $2.50 per year. (Because these are over-the-counter contracts, there are many variations on this, including swaps with periodic coupon payments.)
For the same reasons described above, the implied inflation expectation is roughly 2.5% per year: party B thinks inflation will be less than 2.5% per year, and therefore is willing to take 2.5% and pay the amount of inflation; party A thinks inflation will be more than 2.5% per year, and therefore is willing to pay 2.5% per year to get the amount of inflation back. So the market clears at 2.5%. (Actually, for the exact same reasons as with bonds - party B has to be paid an inflation risk premium for absorbing the risk in this trade - the inflation expectation is slightly less than 2.5% per year. There are also some complications having to do with the lag in the publication of inflation indices, but let’s ignore that for now.)
One curiosity is that the inflation-indexed bond method and the inflation swap method can produce different estimates. Theoretically this should not happen, because if two products that will have the same price in the long term (since they are based on the same index) have different prices today, there should be an arbitrage opportunity. Why this happens in practice is discussed on pp. 5-6 of this Bank of England paper. (Thanks to Bond Girl for pointing out the paper.)
3. Surveys. You can also just ask people what they think inflation will be. Economists ordinarily prefer markets, under the principle that when people are paying money they are signaling what they really believe. But if you think there are sufficient problems with the markets you may want to go with surveys. Tim Duy has a post with a number of charts, including one of an inflation expectations survey.
So what do things look like today?
This is the historical graph for implied U.S. inflation over the next 5 years, based on TIPS. Remember, you are looking at 5-year inflation expectations as they changed over the last year.
In the dark days of October-December, inflation expectations were clearly negative: that is, the market was expecting deflation over a 5-year period. Things have picked up, but inflation expectations are still around 0.6% - far less than the 1.7-2.0% targeted by the Fed. And that 0.6% is before adjusting for the inflation risk premium, so inflation expectations are actually lower than the chart shows.
And these are current inflation expectations over various time horizons, again derived from inflation-indexed bonds. Note that they are sorted by value, not by time.
TIPS are not very liquid compared to regular Treasury bonds, and the implied inflation expectation numbers are sensitive to aberrations in both the Treasury and the TIPS markets (which have both been pretty aberrant recently). For example, if there is a shortage of TIPS of a given maturity, then the TIPS yields will be artificially low and the implied inflation expectation will be artificially high. Still, it seems like inflation expectations are on the low side, even when it comes to the 10- and 20-year time horizons. (I don’t know what’s going on with the 2-year number: when I look at the underlying bonds, it seems like it should be about -0.1%.)
For more on measuring inflation expectations, there is a short primer from the San Francisco Fed, as well as the Bank of England paper mentioned above.
Related Articles
|
-
The Seeking Alpha 100
See allThe top 100 stock
market authors
selected for publication in the last week
Seeking Alpha 100 » -
Fastest Climbers
Those with the largestSee all
increase in followers
in the last week
Fastest Climbers » -
Top Commenters
Top commenters,
as ranked by their peers-
1
User 353732
-
2
Moon Kil Woong
-
3
doubleguns
-
4
optionsgirl
-
5
Mayascribe
Top Commenters » -
1
-
Top Instabloggers
The 100 most activeSee all
Instabloggers.
Top Instabloggers » -
Top StockTalkers
TopSee all
StockTalkers.
Top StockTalkers »
5 year inflation expectation is only 0.6%. So all talk of inflation/hyper infation etc must end.
I belong to the deflation camp, at least TIPS support me.
"TIPS are unique among asset classes in that they are guaranteed to lose purchasing power if held to maturity. As you know, the accretion (the portion of the TIPS value that increases with the CPI) is taxed, and it is taxed yearly, not at maturity or sale. It is also taxed as ordinary income just like the interest portion (if there is any). Even if you assume that the CPI accurately tracks your cost of living (probably laughable), you are guaranteed to lose purchasing power due to taxes. These instruments do not have enough absolute yield to compensate for the taxation of CPI-derived accretion. You might argue that you can hold TIPS in a sheltered account and turn a profit, but it's probably not true unless you really believe that the CPI overstates your cost of living. I'm sure there's someone whose lifestyle and purchasing patterns make this true, but for the other 99.99% of us, it's a sure thing: TIPS are a loser. And why should this surprise us? If you were the Treasury, selling TIPS is almost an ideal scam: you get to tax both the accretion and the interest, and you get to decide what the accretion factor is going to be! It's as if I could issue you a bond and then decide each year how much interest I want to pay you, also knowing that I'm going to get 1/3 of it back from you no matter what. Madoff had nothing on these guys for the sheer brass of it all."
TIPS are never going to be my thing but I clipped it anyway. A few others tried to shoot it down but nothing really convinced me that Bearfund was too far wrong. I reposted this clip once before and will not do so again.
As for inflation expectations over the next few years I expect that the cut backs by businesses will reduce further capacity at the same time as the money supply dramatically increases. More dollars chasing fewer goods = more, not less, inflation. at the same time the correction in housing prices will also go on. As always, depending on how you index your inflation rates will show different results but the bottom line will be inflation that the average consumer will notice.
You were correct to mention that the TIPS market has been aberrant lately (last 6-8 months). At one point the real yield on 10yr TIPS was HIGHER than 10yr Treasuries, reflecting deflation expectations. This is silly, because TIPS have a par value as a floor, and do not deflate below their par value.
Because I am interested in inflation protection, can you tell me more about the US Govt swaps?
How would I buy such a security? - is there a symbol or a site to go to?
If China has 100 apples, and we have 100 dollars, so we can buy one apple with one dollar.
If we inflate our currency to 1000 dollars, you think they will sell one apple for one dollar?
We print a lot money during WWII, remenber the war bonds? We paid with post war inflation, 30% in a few years.
Forget China printed 10 trillion a day in 1948? the inflation rate was 1% every hour.
TIPS rely on CPI, and so do the inflation swaps you used as an example. So you are asking us to measure "inflation expectations" based on the government's CPI number. Is that what you want us to do?
More strong-dollar jawboning.
if the govt is the house, setting the odds and manipulating the games, they must consider all the bettors.
Another factor is the future payment on the national debt. If inflation were to increase, the interest cost of the debt would skyrocket and choke off the demand feeding inflation, until inflation and interest rates declined. Future monetization would be necessary to support future inflation. The current monetization will have a muted inflationary effect.
I apologize for not thanking you for this very useful article (in my previous comment).
Everything is taxed - have a suitable tax hedging strategy for specific investments.
CPI being rigged - it is the way it is - everything is rigged - entire Wall Street is rigged - analysts/estimates - it all is a giant Ponzi scheme. But that is the system -we are all a part of it and have to make the best out of it. CPI has obvious limitations and flaws - price/quality adjustments (hedonics), home rents are used as factor not home prices, etc etc.
TIPS offer a very good inflation hedge, the current inflation expectations are low - this is not a Govt. created number. These are the expectations of investors - if you think inflation will go higher - bet on it - buy TIPS and make a killing.
Great article.
We shall see. In the meantime, simple math tells me that it makes no sense at all to be paid a small amount of "dollars" for a day's work, when vast quantities of these same "dollars" can be clicked into existence at the whim of one man.
Thanks!
1. A US minimum wage worker can earn a Big Mac in about 20 minutes. A Chinese worker doing similar work nees to work for 2 hours to do the same;
2. Items that are not easily substituted- such as top universities, Manhattan real estate, designer jeans- maintain price power;
3. Minimum wages act as price controls, and in a deflationary environment just mean underutilization (i.e., unemployment);
4. Unemployment around 10 percent would engender a major political upheaval;
5. Anytime the Fed inverts the yield curve with its rate hikes, banks, credit, and the economy suffers;
6. There are three billion people on the planet living on less than $2.50 a day;
7. Substitutionary labor and technology are available to mitigate inflation, but not as quickly accessible as in the past- scale and complexity mean it takes longer time and less uncertainty for this new capacity to come on line;
8. There are energy and water challenges that can be met with very sizable investment- and if not met, their cost will be so overwhelming that it will be obvious that this investment is warranted.
If I put all this in a grinder and turn the wheel, what comes out is this: The Fed will keep rates low (probably a little too long), inflation will heat up, and if they're smart they won't invert the yield curve so that credit will continue to expand. There will be momentary periods of price madness as companies realize new competitors are coming in and make hay while the sun is shining, but if played correctly, a global economic boom could be just over the horizon...
I'm sorry, but I don't buy this widely held opinion. I know it's popular among pro-stimulus economists, but I think they're missing an important point.
It seems pretty obvious to me that the market value of equities is NOT equivalent to "money". It can't circulate. For the owner of a security to utilize that wealth in any way other than an investment, she or he has to sell it to another holder. The original holder now has liquid money to spend as she or he wants, but the new holder has reduced his or her ready funds by the same amount. Thus the amount of circulating money has been unchanged.
Wealth does not equal money. Money is but one special form of wealth.
So, increasing the circulating money supply to "replace" the evanescent wealth evaporated by the meltdown should in fact lead to inflation. Now we may be "lucky" and see that inflation only in the price of the assets propped up. In other words, we may reflate the bubble in assets but not experience rising prices in other areas of the economy. But I'm pretty doubtful that the Fed can pull off that particular legerdemain again.
On Apr 10 06:31 PM Tao wrote:
> Doesn't new money replace all the money that disappeared out of equities
> last year?
>
> Great article.
We have a fiat currency and once everyone in the world learns quadrillion (1,000,000,000,000,000) comes after trillion the hurting will begin. Nobody knows the outcome or the unintended consequences of throwing trillions into the fray.
It is my pleasure to announce the much anticipated New World Currency! As deficits skyrocket and new paper money is printed its nice to know that there is a safety net waiting for the US Dollar. Welcome back to the gold standard friends!
Greenspan advocated a return to the gold standard. I must agree. The Gold Eagle has been minted in the following denominations.
1 ounce: $50 face value
1/2 ounce: $25 face value
1/4 ounce: $10 face value
1/10 ounce: $5 face value
So, your "new" dollar is worth about $17.81 in current paper dollars, or 1/50th of an ounce of gold, at the time of this writing. Conveniently the Silver Eagle has a one dollar face value and assumes the value of silver as 1/50th that of gold. No need to mint new coins in denominations less than a dollar as the current ones will suffice.
This is not a matter of "if" its a matter of "when" this becomes the new accepted form of currency.
Surveys measure something different from breakeven spreads and swaps. In fact, it's hard to know exactly what surveys do measure, but they can capture two things the other metrics cannot: the expectations of those who do not or cannot participate in the mentioned markets, and (potentially) the extent to which those surveyed believe the CPI matches their definition of "inflation".
An additional metric not mentioned here is the gold market, which unfortunately also reflects a number of variables that are neither past inflation nor inflation expectations, but is in the long run probably the only truly accurate measure of *past* inflation. But it too is inadequate. I feel very comfortable using gold to say that today's dollar purchases around 2.5% of what it did in 1920. We can argue whether it's really 2% or 3.3%, but that number is not far wrong. But by the same logic, the dollar buys as much today as it did in 1980 - clearly an absurd conclusion. We can play games with moving averages but the reality is that the gold market harbors a lot of highly leveraged participants, dark pools, and suppliers (central banks) that act perversely with respect to their own profits. It is also subject to nebulous factors like "fear" that can cause price levels to diverge from reality; it is more correct to state that if the gold market is capable of measuring anything at all, it measures momentary confidence in the fiat money and fractional-reserve system. Inflation and inflation expectations are but one component of that confidence, so it will not do for our purposes either.
The trouble is that the CPI-U tells you that today's dollar buys about 47% (CPI-U=212.193 as of 2/09, 1982-4=100) as much as it did around that same time. I will contend that no person who was actually responsible for managing the finances of a household in 1982 and still is so today would afford that figure any credibility whatsoever. That is, while gold contains a huge amount of noise, the CPI-U is much less noisy but contains large systemic errors. Therefore, it's very important to state explicitly that TIPS and swaps measure CPI-U expectations (or expectations of whatever they're indexed to if not the CPI-U), not any useful definition of inflation. Worse, most of these instruments are held by one of two types of entity: large insurers and pension funds with CPI-indexed liabilities, and large banks. In the former case, the investor does not care whether the price accurately reflects either expectations or reality; they are holding to maturity and the proceeds, whatever they are, will suffice for their intended purpose. The banks, of course, are enormously leveraged and actually profit from a decline in the dollar's purchasing power, so a small real loss on a TIPS or a swap becomes a gain against a much larger real loss by the bank's depositors and other creditors.
In short, then, the two markets described are dominated by two factors rendering them of little value for the stated purpose: a dubious index and participants with little or no incentive to ensure that the prices they pay and receive are meaningful indicators of their own inflation expectations. Throw in a hefty helping of illiquidity and we can simply disregard these markets altogether.
So we're left with surveys. These at least attempt to measure expectations of the right people: those for whom the "cost of living" has some meaning and who are likely to hold most or all of their assets in cash over the period of interest (even if only on payday). Unfortunately most of those surveyed do not understand what inflation really is and in most cases do not even have sufficient data to determine whether their past estimates of it were correct, making future estimates nothing more than guesses. It would be more reasonable to simply ask them how rapidly they expect the cost of living to increase in coming years relative to the rate at which it increased in the last N years, and use non-numeric answers (much more slowly or not at all, more slowly, about the same, more rapidly, much more rapidly). Any significant balance toward the latter two should be a strong warning signal to central bankers, regardless of the numbers those surveyed would have assigned if given the chance. But today's surveys are not conducted in this fashion, so we continue to read that 800 randomly-selected adults expect 5-year inflation of 3.2%, a stunning example of precision in the complete absence of accuracy or even context.
A new approach to macro price prediction is needed. Some of the data described here may be valid inputs, but a rigorous many-factor approach with backtesting is required, just as it is for any other technical exercise. I do not know whether such an approach will yield better results than chance would indicate, but it could not be worse than any of the indicators mentioned here.