Friday, January 9, 2015

Don't Worry, Be Happy: Falling Treasury Yields Edition

Long-term treasury interest rates are falling again with the 10-year treasury briefly dipping below 2% this week. Some observers see this decline in yields as an omen for the U.S. economy:
The United States economy is accelerating... Yet a huge bond market with a strong track record for predicting economic problems is flashing a warning sign right now.The prices of Treasury bonds are rallying fiercely. The slide in oil prices has elevated concerns about growth in the global economy, and investors, as they do in times of stress and uncertainty, are seeking out the safety of government bonds. 
The rally in global government debt is pushing their yields, which move in the opposite direction from their price, to astonishing lows... “Make no mistake, these low levels of rates are challenging the notion that we are going to see robust and constant growth,” said George Goncalves, a bond market analyst with Nomura. In other words, the bond market is raising the specter that a period of economic growth that may have already felt lackluster to many Americans could be on the verge of losing steam. 
So is this dire view of the falling interest warranted? I am not convinced that it is, but before I explain why it is worth noting that even monetary authorities are bewildered by it. According to the Wall Street Journal's Jon Hilsenrath (AKA the "Fed Wire"), Fed officials are not sure how to interpret what is going on with yields: 
Falling long-term interest rates pose a quandary for Federal Reserve officials... If falling yields are a reflection of diminishing inflation prospects, as is typically the case, it ought to prompt the Fed to hold off on raising short-term interest rates in the months ahead. If, on the other hand, lower long-term rates are a reflection of investors pouring money into U.S. dollar assets, flows that could spark a U.S. asset price boom, it might prompt the Fed to push rates higher sooner or more aggressively than planned. 
[...] 
The Fed’s next policy meeting is three weeks away. It is clear officials will spend a considerable time debating the correct response to a perplexing lurch down in long-term rates.
The worried market observers and the perplexed Fed officials should take a deep breath. The Adrian, Crump, and Moench (2013) method of decomposing treasury yields paints a far more benign story, one that signals the U.S. recovery is on a solid footing.

To see why, we first need to recall that long-term interest rates can be broken down as follows:
(1) long-term interest rate = average short-term interest rate expected over same horizon + term premium
The term premium is the added compensation investors require for the risk of holding long-term treasuries over short-term ones. For example, if investors are worried that the Eurozone crisis is about to flare up again and desire to hold more U.S. treasuries, they will demand less compensation to hold the long-term securities. This will drive down the term premium. The term premium is also the component of the long-term interest rate the Fed was trying to manipulate with its large-scale asset purchases. 

The other component, the average short-term interest rate, is a nominal interest rate and via the Fisher relationship can be further decomposed into a real interest rate and an expected inflation:
(2) long-term interest rate = (average expected real short-term interest rate over same horizon + average expected inflation over same horizon) +  term premium
This average expected real short-term interest rate is often called the real risk-free interest rate since it is free of investor's risk considerations, the Fed's tinkering with risk premiums, and the expected path of inflation. This interest rate measure, consequently, tracks the fundamentals of the economy and is equivalent to the average expected path of the 'natural interest rate'. 

By looking at these components we can make sense of what is driving the fall in yields. We can also look to the real risk-free interest to see what it implies about the health of the U.S. economy. The Adrian, Crump, and Moench (2013) decomposition of the 10-year treasury yield into these components is below:




What we see is that changes in inflation expectations and the term premium are both behind the decline in the 10-year treasury interest rate. This suggest that there may be concerns about future inflation--though this might also be reflect the temporary drop in inflation from declining oil prices--and that there has been a rush into treasuries because of the worries about the Eurozone and China.

But there is more. After being negative for several years, the real risk-free interest rate has been steadily climbing and is now positive. This only happens when the economic outlook improves as seen in the figure below. It shows a close relationship between the real risk-free interest rate and the business cycle:




So the upward trend of the real risk-free rate implies we are in the midst of a solid recovery in the United States. This interpretation is supported by the spate of positive economic news shows. Yes, the economic problems in Europe and China could eventually harm the U. S. economy.  But for now the U.S. economy seems to be in the clear. 

So be careful when interpreting long-term treasury yields. They might be signalling a robust recovery even if they are falling.

13 comments:

  1. what did you use for 10-yr inflation expectations?

    ReplyDelete
    Replies
    1. I used 10-year breakevens. It is not perfect, but the story does change much if you use higher frequency measures from survey data like the Survey of Professional Forecasters.

      Delete
  2. David,
    The negative term premium is an interesting output of the Adrian model.

    T-bond buyers should be willing to instead buy s.t. TIPS and write a "swaption" on TIPS yields. This option would give the counterparty the option to receive (real) floating and pay (real) fixed during the tenor of the T-bond, with a strike at the current s.t. TIPS (real) yield.

    The twist, given the Adrian model, is that this swaption would currently be written at a cost to the seller. That is, it would have negative premium (corresponding to the negative term premium).

    I don't imagine you get such a price from JPMorgan for an equivalent synthetic. I must be missing something.

    ReplyDelete
    Replies
    1. Diego,
      Do you have an idea of how much the volume of these swaptions has changed in the past month?

      Delete
    2. Diego,

      I am not sure I completely understand the transaction, but aren't you describing one based on TIPs where the ACM is simply a decomposition of the observed nominal yield based on a macro finance model?

      For what it is worth, the Kim-Wright model, which is also based off of a macrofinance model, also shows a recent decline in the term premium that has turned negative. It can be found here: http://www.federalreserve.gov/pubs/feds/2005/200533/200533abs.html

      Delete
    3. David,
      The term premium is the compensation investors demand for accepting the risk of future s.t. rate volatility. A willingness to pay someone to give you risk seems to violate the assumption of risk aversion.

      I wonder what Adrian would say about the theoretical basis for a negative term premium? No doubt he would have an answer. However, I will note he himself has written:

      "It’s also instructive to compare the term premium with investors’ uncertainty about future Treasury yields (see chart below). We can measure this uncertainty by the Merrill Lynch three-month MOVE index, which summarizes options-implied expected volatility of Treasury yields of different maturities. Over the period when data are available on the index, the term premium is highly correlated with it. Hence, the level of uncertainty about future Treasury yields appears to be associated with the level of term premia."

      http://libertystreeteconomics.newyorkfed.org/2013/04/do-treasury-term-premia-rise-around-monetary-tightenings-.html


      Delete
  3. The United States economy is accelerating...”

    How so? Rates-of-change in monetary flows (our means-of-payment money times its transactions rate-of-turnover), have been decelerating during QE3’s operations. The proxy for real-output peaked in July. This December is now 45 percent of July 2014 levels.

    For the economy to accelerate, there must be an increase in aggregate monetary demand - which can come about only as a consequence of an increase in the volume and/or transactions velocity of money.

    Current inflation and inflation expectations drive long-term yields. The proxy for inflation peaked in January 2013. It is now 1/3 of January 2013 levels.

    Quantitative easing (when remunerating excess reserve balances), has been decidedly contractionary (other variables have been more important).

    Our economy is being run (and our students are being taught), by people who don’t know the differences between money and liquid assets (Chicago school, et al).

    ReplyDelete
  4. Seems to me that the declining prices of commodities over the last few months (Google this) and the yield curve being flatter than last year (see: http://www.treasury.gov/resource-center/data-chart-center/interest-rates/Pages/Historic-Yield-Data-Visualization.aspx - and use the 'Enter Comparison Date' feature to compare to points from last year) means that the US economy is expected to slow in 2015. But the yield curve is not yet inverted, so no recession is being forecast by the bond market.

    ReplyDelete
  5. Long-term rates are a function of the moving average of long-term money flows. While roc's in MVt peaked in Jan 2013, the moving average didn't peak until July (with rates finally peaking in Sept). Currently the moving average for long-term rates points to a trough in rates (peak in bond prices), during May 2015.

    ReplyDelete
  6. Rates-of-change in monetary flows, our means-of-payment money times its transactions rate-of-turnover, or the proxy for inflation, has fallen by 2/3 since January 2013.

    Rates-of-change in the proxy for real-output held up until July 2014, but has since fallen by 1/2.

    So aggregate monetary purchasing power, or as Keynes called it, nominal-gDp (which is a proxy for all transactions in Irving Fisher's "equation of exchange"), has taken a big hit.

    The Fed has unintentionally set up a liquidity squeeze, collapsing commodities, artificially raising the dollar's exchange value, and suppressing longer-dated bonds. Where are the nominal-gDp targeting advocates when they're desperately needed?

    ReplyDelete
  7. I'm wondering if we might be seeing an inversion of the pseudo-inflation driven by oil price shocks in the early 1980s. If I'm recalling the MMT primers correctly, the Fed tightened in appropriately because it mistook the supply shock-driven price increase in oil as indicative of of inflationary pressure.

    Is it now the case that observers may mistake the current oil price reduction as creating deflationary pressure? In a classical sense, it is deflationary (as the total quantity of oil is now denominated with far fewer dollars). But should the Fed consider this evidence of a need for more or extended QE?

    I suspect that we will see a PPI reflection of the oil price decline soon, and this may yet be stimulative to broader spending as it frees up discretionary spending.

    It's hard to value it, but there is certainly something economically consequential about how people view $2 gasoline.

    ReplyDelete
  8. US economy is in difficult financial situation today. People have huge financial difficulties and they try to solve them with the help of easy and quick loans payday . As for me it’s the best solution in such situations because these loans can really help in emergencies. You can take them 24/7 via internet. Of course these loans can’t solve the problems of the country but can help people who need cash.

    ReplyDelete
  9. Real interest rates are cyclical - peaking in 1989, 1999 & 2007. I'm not sure that's news.
    But a real rate below 1% on bonds does not seem like solid evidence of a "solid recovery."
    Better than zero, but not much.

    ReplyDelete