US Bond Yields and the Rediscovery of Normality

What chance of a bond market collapse as the Fed starts to lift rates?

The final stage in the US Fed's long edging-back to 'normality' will be the first interest rate rise - still generally expected in June. This provokes two  major worries. First, there is the generally unacknowledged problem that for the foreseeable future Fed interest rate decisions are much more exposed to policy error than before the Great Recession.  The reason for this is that the Fed (and its market watchers) can no longer deploy the Taylor Rule (of thumb) to judge what should be done, and what is likely to be done, because there is no robust technical consensus over what US potential output might be. For the foreseeable future, we are more likely to discover the limit to potential US output by tripping over it than by calculation.  

The second worry is this: we may have to rediscover how how bond yields are determined if they are not simply wrangled by the Fed. The whole point of zero interest rate policies and, in extremis, central bank quantitative easing, is precisely to over-ride market mechanisms in order to command the longer-end of the yield curve.  Under such a regime, what matters for bond markets is disproportionately what the central bank decides to do.  

That's what's coming to an end, and already in the expected absence of such security, there are plenty of scary scenarios about what happens next.  This piece is an attempt to offer some prompts. 

The starting point is 'fair value yield' models, which essentially attempt to locate where bond yields 'should be' by accounting for three sets of information: 

• expected movements in the central banks' core short-term interest rate; 
• expected movements in inflation;
• expected movements in GDP growth 

In practice, these models worked tolerably well when central banks were not intervening so aggressively in the long end of the market. The chart below shows the history, and casts forward using Bloomberg consensus forecasts for interest rates, inflation and growth. 

At present, the prevailing 1Q 10yr treasury yield of just under 2% compares with a fair value yield of around 3.1%.   Further, it suggests that if GDP growth steadies to around 2.8% in 1H16, inflation rises to around 2.1% and Fed Funds rates rise to 1.45% by 2Q16, fair value yields would rise to around 4.2%.

That suggests plenty of room for yields to rise in the short to medium term. But historically speaking, does this suggest that current yields are still carrying an unusual discount inherited from the Fed's years of bond-buying?  If we look at the history of deviations from fair value, it turns out that the current 1.2% discount to 'fair value' is not historically unprecedented even in the absence of central bank activism.  Valuations are also not stretched so dramatically from 'fair value' as they were prior to the 'Taper Tantrum' of mid-2013.

In the short term, then, whilst one should anticipate yields rising, the case for a major bond crash is not compelling. In the medium term, of course, the fate of yields will be largely determined by how the economy (and the Fed) performs - which is, of course, the point about 'normality'. 

We need not leave it there: it would be nice to be able to offer some explanation of why bond yields deviate from 'fair value' in the way they have over time - acknowledging that 'fair value models' do not provide complete explanations.  The most obvious explanation is the uncertainty and error which must accompany any set of forecasts or expectations. But in addition, one tool I suggest is movements in the private sector savings surplus (measured as the current account balance minus the quarterly change in federal debt in public hands).  There is a very good reason why one would expect movements in the PSSS to have some impact on government bond yields. What the PSSS measures is, after all, the cashflow passing between the private sector and the financial system. In the simplest of institutional set-ups, if the private sector has a savings surplus (ie, it spends and invests less than it earns during a period), it deposits that surplus in the bank. The bank, by definition, can invest that cash only in either foreign assets or government bonds. 

But this is unlikely to be a simple one-way relationship, since, conversely, one might expect significant movements in bond yields to make an impact on private sector saving/spending decisions. 

Such a feedback relationship means one can't be dogmatic about causation. Nevertheless, it strengthens the expectation that major movements in the private sector savings surplus to be reflected in bond yields (and vice versa). 

Now compare the deviation of bond yields from 'fair value' with movements in the PSSS during the last 25 years. 

What it suggests is that the current gentle movement downwards in the PSSS offers some support for the current structure of bond yields.

And this is where this week's 4Q current account balance information matters: 4Q's current account deficit was estimated at $113.5bn, which suggests the deficit was running around 2.3% of GDP during 2014, essentially unchanged from 2013, and relatively stable.  In addition, it suggested that the PSSS for 2014 came in around 1.5% of GDP, and, like the current account balance, appears to have been almost stable during the last two, after drifting down gently since 2009.  In the context of the threat of bond-yield volatility during the re-discovery of normality, that stability in the core cashflow measure of the US economy is genuinely comforting.