Pages

Monday, September 1, 2014

Another Bond Market Conundrum?

Is the U.S. economy in the midst of another bond  market "conundrum"? The last time we had one was in 2005 when former Fed chairman Alan Greenspan became perplexed over long-term interest rates failing to rise with the tightening of monetary policy. Some observers see something similar happening today. They note that the Fed has been tightening monetary policy with its tapering of QE3 and yet the benchmark 10-year treasury interest rate has been falling since the beginning of 2014. This conundrum gets even more interesting when one looks at the five-year treasury interest rate. It has hardly budged since the beginning of the year even as the 10-year interest rate has steadily declined. Jim Hamilton calls these developments the 'bond market condrum redux.

So what could be driving these developments? Robin Harding and Michael Mackenzie suggest it is the worsening economic conditions in the Eurozone and its implications for ECB monetary policy. Here is Harding-Mackenzie:
The link between US monetary policy and US bond yields has fallen apart this year, showing how fears of deflation in Europe are driving global financial markets. According to analysis by the Financial Times, the correlation between five- and ten-year Treasury yields has fallen to its lowest level on record, with US bonds appearing to track European monetary policy instead.

[...]

Five-year bond yields closely reflect the path of interest rates that markets expect from central banks. Ten-year yields normally move in tandem. But so far in 2014, the US ten-year has fallen from 3 per cent to 2.4 per cent – even as news on the economy has got stronger – while the US five-year yield has barely changed.

That is unprecedented: no other global shock going back to the 1960s has ever caused US five and 10-year yields to diverge like this. In recent months, the US 10-year yield has been more correlated with falling five-year yields in Europe.
I think Harding-Mackenzie may be on to something with the Euro zone-U.S. treasury yield connection. It is worth taking a closer look. To do so, we first need to recall that long-term interest rates can be broken down into the following components:

(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 holding long-term treasuries over short-term ones. For example, if a financial panic causes a "rush to safety" by investors and they are more willing to hold long-term treasuries, then they will demand less compensation and the term premium will decline. The other component, the average short-term interest rate, is a nominal interest rate and via the Fisher relationship can be further decomposed into real interest rate and expected inflation terms:

(2) long-term interest rate = (average expected real short-term interest rate + average expected 
     inflation) +  term premium 

The average expected real short-term interest rate is typically tied to the expected growth of the economy.1 If the U.S. economic outlook is improving, it should increase and vice versa. Over the short-run it should track the business cycle while over the longer run it should track trend economic growth. It is also called the real risk-free interest rate and approximates what many macroeconomists call the natural interest rate.

Okay, with all that said we can now take a look at these components and assess the Harding-Mackenzie explanation for recent interest rate movements. The figure below shows these series since 2013 (It is constructed using the Adrian, Crump, and Moench (2013) term premium estimates and implied inflation forecasts from TIPs.)


This figure shows two interesting developments since the beginning of 2014. First, the term premium has been steadily falling. This decline could be explained by increasing worries over the Eurozone driving investors on the margin to treasury securities. This is not exactly the Harding-Mackenzie story, but it does provide a potential link between what is happening in Europe and long-term U.S. yields. 

Second, the real risk-free rate has been steadily climbing since the beginning of the year. This is consistent with the gradually improving U.S. economic outlook. As I have shown before, short-term swings in this real risk-free interest rate track the business cycle fairly closely. Consequently, this second development should not be surprising.

So on one hand the term premium is arguably falling because of Eurozone woes, while on the other hand the real risk-free rate is rising because of the improved U.S. economic outlook. The term premium decline, however, is larger and on net is dragging down the 10-year treasury yield (given stable inflation expectations).

But what about the 5-year treasury yield? The figure below provides the same decomposition as above and shows similar developments:


The big difference here is that changes in the term premium and real risk-free rate are much closer in size and therefore closer to offsetting each other. The graph below summarizes these differences. It shows similar increases in the real risk-free yield for both treasury securities, but much larger declines in risk premiums for the 10 year security.


In my view, then, the real bond market conundrum is why has the term premium falling more for the 10-year treasury? Has the flight to safety associated with the Euro crisis been weighted more toward long-term treasury securities? Or is it something else?

Update: See this Bruegel review of blogs for more the bond market conundrum.

 1It is technically tied to the expected growth of productivity, the expected growth of the labor force, and household's discount rates. See here for more on these points.

11 comments:

  1. Would be interesting to see the spread -- real risk-free rate on 5s vs 10s -- over the long haul, and also zoomed in on recent years. Could also include the term-premium spread. I can't imagine the 5- versus 10-year inflation-expectation spread would be all that interesting, but...

    ReplyDelete
  2. David: Exactly how did you compute this decomposition? Can you describe it simply here, or do I need to do two days of reading of some obscure literature? Help me out here!

    ReplyDelete
  3. David, all I am doing is filling in data for equation two above. The only hard part is getting term premium estimates, but fortunately the smart folks in macro-finance have figured out how to estimate it and some are even posting it online with regular updates. The one I used above is from the Adrian, Crump, and Moench (2013) paper that can be found here. The other tricky part to get is expected inflation. I used breakeven inflation from TIPS from FRED since it is available at a daily frequency. I used the Philadelphia Fed Survey of Forecasters for quarterly inflation forecasts when I did my Washington Post piece on secular stagnation where I used this decomposition.

    So with an observed nominal yield, a term premium, and an expected inflation series you can solve for the real risk-free interest rate in equation two.

    ReplyDelete
  4. My answer to the conundrum is the Fisher Effect. As we see ECB nominal rates heading increasingly lower due to economic stagnation, lower inflation follows with a natural real rate independent of monetary policy. There is an accompanying downward spiral of ever weaker demand.
    As we see the US economy pick up and expectations of the Fed raising nominal rates, the real rate will rise as inflation expectations have your 2% upper bound. I agree with you on that 2% upper bound of inflation. The Fed will react to head off upward inflation momentum. They have to keep a reliable lid on inflation.
    So the Fisher Effect explains the conundrum. I agree with the likes of Steve Williamson that the Fed should have already started raising nominal rates. Japan fell into that trap of stabilized low rates and deflation developed. Europe should have already raised their nominal rates too in order to avoid the Japanese style moderate deflation which is coming their way.

    ReplyDelete
  5. Edward,

    Didn't the ECB already try raising rates in 2011? Right before they had a second recession?

    ReplyDelete
    Replies
    1. Garrett,
      Euro area ticked up their benchmark rate from 1% to 1.5% for part of 2011. Their core inflation rose during the tick up in the benchmark rate. Core inflation has been falling ever since 2012 when it became clear that the ECB would commit to low rates for a long time. This is the Fisher Effect.
      Great Britain started to contract in 2010. Loans to private sector were actually rising during the tick up in the benchmark rate. Germany did not go into recession. The periphery countries lowered wage costs because they could not devalue their currency.
      The Euro area has become a net exporter since 2011, which implies that there was a decline in domestic consumption which raised the area's "national" savings. That decline in domestic consumption took place 2010 and 2011. Do you remember the wage disinflation policy from Germany in Agenda 2010? This policy spread through Europe and there was a contraction in demand, not only at the household level but at the public level too. This was a primary cause of their recession. After the recession, the Euro zone became a net exporter due to the declining wage shares which triggered the domestic recession.
      Wage shares fell most dramatically in Spain, Greece, Portugal and Ireland. The recession was centered most in these periphery countries. Credit markets were tight. Debt levels were still high.
      in 2011, Inflation began to rise in the face of wage cutting. Firms were trying to be more profitable. Raise prices and cut costs. This is not a good mix for society though. So the ECB acted to restrain inflation.
      There is much more to the story than just... The ECB caused a recession by raising rates. There was an adjustment process to become a net exporter area. The process was most damaging to the periphery countries.

      Delete
  6. This theory has been floating around as to why bonds (and perhaps the real term premium) is "decreasing"

    http://blog.alliancebernstein.com/index.php/2014/05/30/how-low-can-the-30-year-treasury-yield-go/

    ReplyDelete
    Replies
    1. Did you decompose the 2005 time period to see term premium was behind that divergence?
      And has the ending of or operation twist itself impaired the yield curve in some way making comparison difficult?

      Delete
  7. Japanification...the real threat to modern economies, not inflation. Please tell the central bankers...

    ReplyDelete
  8. Everyone makes the assumption that bond-holders and buyers judicious assessments (modeling), are always at least halfway correct. That is not even necessarily so. There's absolutely nothing perplexing about current levels and no need to fine-tune their calculations.

    ReplyDelete
  9. The actual conundrum is why all the talk is about what rates did, instead of what they will do. Anyone that understands money and central banking already knows what's "in the cards" (seemingly like a magician).

    ReplyDelete