The US yield curve and the next recession.

The current US  economic expansion started in July 2009 and is already  much longer than the post-war average, although still  below the 10-year record duration set in the 1990’s, while closing in on the no.2 spot, set at 106 months in the 1960’s. A near term downturn is not inevitable but history suggests is likely at some point over the next few years. Forecasters are  poor at predicting recessions, and so there is interest in other potential signals. Equity markets generally turn down ahead of the real economy but they can and do fall without that precipitating a decline in GDP, so there is a risk of a false signal. The relationship between short term interest rates and long rates  (the yield curve) is another indicator of note, and in the US has proven  remarkably accurate ahead of the last seven downturns. Specifically, a yield curve inversion ( 10 year yields below 2 year yields or as some prefer, 3- month rates) has proven to be an excellent signal of a US recession a year or so ahead.

Why the success as a signal? Longer term  bond yields  carry a risk premium and are therefore generally higher than short rates, and may also be influenced by specific demand/supply factors at different maturities. For example, banks generally buy shorter term bonds, while pension funds seek much longer maturities. Expectations about the path of short term rates over the period are the most important factor, however, which in the US amounts to expectations about the Fed’s monetary policy and inflation. If policy is tightened in response to a booming economy or above target inflation longer term rates tend to rise, albeit by less than the move in short rates (the curve flattens) and may eventually invert if the market believes that  short rates have peaked and will eventually start falling . The inverted yield curve may also help precipitate a downturn because it dampens margins for the banking sector (banks borrow short and lend long)

The US yield curve is not currently  inverted but it has flattened appreciably; the 2-10 year spread has fallen from a peak of 260 basis points in late 2013 to just under 60 now, having started the year at 135.The recent pace of flattening has prompted much market debate  particularly as short rates are still very low by historical standards.

The Fed is  widely forecast to raise short rates again as early as December , and has signalled that it expects to tighten further in 2018, yet  10-year yields have fallen in absolute terms over the past month and are well below the highs in yield recorded earlier in the year.Maturing Treasuries are no longer being fully reinvested and, all things equal, the Fed’s decision to steadily reduce its holdings of bonds might be expected to push yields up. Some argue that issuance is shifting towards  the shorter end of the yield curve, so supporting longer dated paper, which in any case is still in strong demand as a ‘safe’ asset and  such assets are relatively scarce as central banks elsewhere are still buying.

A bigger factor may simply be that the market is convinced, at least for now, that  US inflation will continue to disappoint the Fed and remain below the 2% target, so implying that short rates will not rise to the extent the FOMC expect. The current  core inflation rate is only 1.3% ( the consumption deflator ex food and energy)  and was last (briefly) above 2% in early 2012. Most Fed governors believe that inflation will eventually start to accelerate as wages belatedly respond to the extremely low unemployment rate, but that view is not universally shared. Indeed, the minutes from the most recent FOMC meeting point to growing doubts as to whether sub-target inflation is indeed ‘transitory’.

Yield curve models are currently giving a low probability of recession in 2018 it has to be said ( the New York Fed’s model indicates around 10%) but the yield curve certainly bears watching given the recent trend.

 

Eligible Irish bonds a constraint on extending QE

The ECB has flagged the possibility of adjusting its asset purchase programme (QE) to  boost economic activity and move inflation closer to target. At the moment the Bank is buying around €60bn of assets a month, with over €50bn in the shape of Government bonds, with the intention of continuing until at least September 2016. One practical concern regarding extending that timeframe  is the supply of eligible bonds and that may well become more of an issue in the Irish market as we move through 2016.

The nominal value of Irish Government bonds  currently outstanding is €125bn, with some €95bn falling within the maturity range eligible for QE (over 2 years and under 30 years). Bonds yielding below -0.2% are also excluded under the current criteria, which affects a number of countries, notably Germany, but not Ireland. The ECB can buy up to 33% of bonds issued, so that implies an Irish figure of  €31bn.

QE started in March and purchases in the Irish market have averaged €0.8bn a month, bringing the total to €6bn at end-October. Plenty of scope left, therefore, except that the ECB and the Irish Central Bank  already hold Irish bonds, and that figure is included in the QE calculation. The Central Bank  acquired €25bn of bonds as part of the Promissory note deal, with €7bn  maturing within 30 years. The ECB’s holdings arose from the Securities Market Program , which operated for a time in 2010, and amounted to €9.7bn at the end of last year. We do not know the current position or how many will mature over the next two years but from the average maturity (4.5 years) it would seem reasonable to assume that  €7bn would redeem after 2017 and hence fall within QE eligibility.

On that basis the ECB and the Irish Central Bank may own an additional €14bn of Irish eligible bonds, bringing the total to €20bn  when adding the QE purchase to date. Absent any other changes that would mean that the ECB could buy an additional €11bn, which at the current pace of purchase implies a maximum of 14 months, taking us to the end of December 2016 or just three months beyond the current timeframe.

On the face of it, then, the scope for extending Irish QE is very limited, although two factors are likely to give the ECB some leeway, albeit not a great amount. The first is additional supply from the NTMA, with perhaps  €10bn issued in 2016, which would boost the eligible bond total by that amount if over 2 years in maturity. That would allow €3.3bn in additional QE , an extension of four months.  In addition , the Central Bank is required to sell at least €0.5bn of its bond holding next year and any sales would leave greater room for ECB buying of bonds held by the market. On that basis Dame Street may well sell far more than the minimum.

What’s driving bond yields up?

The ECB has been delighted with the response to its asset purchase programme, and indeed the initial reaction from  all asset classes, from bonds through to equities and FX, was both significant and supportive of the Bank’s attempts to stimulate economic activity. The ECB first announced its intention to buy  private sector debt last September, with the euro trading at $1.29, and the single currency subsequently declined to under $1.05  following the January decision to extend QE to government bonds and the  commencement of purchases  in early March. European stock markets rose sharply in the months after the January decision and bond yields continued the trend decline begun last autumn; Irish 10-year yields fell to a low of 0.65% and the German equivalent traded at 7bp, with negative yields prevalent in that market up to an including the 5-year maturity.

The picture looks rather different today.  Government bond yields have risen sharply amid very volatile trading, with 10-year yields in most markets back up to levels seen last October. German  4 and 5-year yields  are now positive again  and the major European equity markets have fallen by around 10% from the highs, with the euro also gaining ground, trading above $1.12.  QE is still  proceeding according to plan and the ECB’s balance sheet is expanding as intended ( €2.42 trillion at end- May from €2.15 trillion at end-2014)   so the fall in asset prices has prompted some  puzzlement, with  a number of  explanations vying for supremacy.

One approach emphasizes  bond  fundamentals, starting with real interest rates and the outlook for economic growth. The  macro data in the euro zone has tended to surprise to the upside in recent months and there was some modest upward revisions to near-term growth forecasts  but the consensus projections for the next few years have not really changed, with most still expecting a sub 2% expansion in the EA.  Similarly the outlook for the global economy has not materially changed (if anything,  the growth forecast for this year have moved lower) so it does not seem likely that real interest rates have suddenly moved higher.

Nominal bond yields are also determined by  inflation expectations (plus a risk premium) and again  forecasts  for EA inflation have not materially changed of late, including those from the ECB,  which foresees a gradual return to annual inflations rates approaching 2%.  Actual inflation has turned positive, it has to be said, so perhaps the deflation scare has abated, although it was always difficult to know if that was really a major concern for investors. Expectations on one of the ECB’s most closely watched measures (the 5 year five year forward inflation swap) are  around 1.75%, which is well up from the sub 1.50% lows but not signaling any inflation scare.

Some peripheral bond markets have fared worse than others during the sell-off  (Portugal for example) but a generalized contagion from Greece is not evident, at least not yet, given that 10-yr bund yields  have also risen sharply, by over 80bp in the past 6 weeks.

Other explanations emphasis market conditions. Issuance in some markets has been higher than expected, for example, including corporate debt. Lack of liquidity may also be  a factor, as a consequence of banks having to hold more  regulatory capital. This , alongside the Volckler rule, has persuaded many market-makers to hold less inventory, with the result that a given degree of selling will have a much greater impact on the market price than it would have done a few years ago. Certainly the scale of intra-day volatility (up to 16bp in 10-year bund yields) is far higher than normal, supporting the idea of thinner markets.

Another  explanation highlights the different types of bond buyers, each with varying risk  tolerances and trading objectives. Banks are required to hold more liquid assets under new Basel regulations, and  so  have bought shorter-dated bonds even at negative yields , particularly as for some the alternative is a deposit with the ECB at an interest rate of -0.2% (overnight ECB deposits are still high, at  €100bn). Credit conditions are improving in the EA, however, with a modest pick up in lending to the private sector, so  some banks are finally using the ample liquidity available to support credit creation to firms and households.

Hedge funds and other traders are looking for a short term return and here the predominant  trading style may be a factor- momentum trading is the order of the day for many, which explains why a trend already well established can persist long after some feel it has lost touch with fundamentals.  The problem arises when the trend  changes and many are then heading for the door, which is suddenly crowded. The lack of liquidity  is exacerbating the downdraft.

‘Real money’ investors, such as pension and insurance funds, are also important, but usually ‘buy-to-hold’ and generally players at longer maturities. They are therefore  less likely  to get caught up in a specific trading style and may well step in following a sharp  sell-off, so putting a floor in the market.

All these explanations, fundamental and market related, are not mutually exclusive, of course, and I suspect the sell-off owes most to the  recent inflation data and the acceleration in monetary growth, with the exit from a crowded trade also playing a big role. One should also keep the correction in perspective- bond yields (government and corporate ) are still extremely low by normal standards and hence  nominal financial conditions  remain unusually  loose, even if a little tighter of late.  In the shorter term it may well be  the actions of  the Fed, rather than the ECB, that helps determine the next big move in EA yields.

Why buy Bonds with negative yields ?

Negative bond yields are no longer a rarity across the Euro area,  accounting  for over half the government debt at issue in some countries (Germany, the Netherlands and Finland) and well over a third in others (Austria, France and Belgium).  Moreover, what was generally confined to shorter term debt is now extending along the yield curve, and  many now expect German 10-year yields  (currently 0.15%) to  follow Switzerland into negative territory.

Low bond yields are one thing but negative yields are a rare if not unique phenomenon. The former may be generated by a flight to quality but if widespread imply that investors expect short term interest rates to stay low for a long time. That in turn signals an expectation of limp growth and little or no inflation for a prolonged period.  Nonetheless, very low yields still mean a positive return, albeit a limited one: if I buy the German 10-year benchmark, which pays a coupon of 0.5% per annum, I will receive €5 per €100 invested in interest , offset by the capital loss on the bond ( it is trading at  €103.35). This will reduce my total return over 10 years to just €1.65.

That level of nominal yield is obviously  very problematical for savers or for the pension funds that are investing the savings of companies or households. That meagre return is also nominal, of course, and would mean a substantial loss in real terms even with very low inflation over the period.

Nonetheless,  any holder to maturity will not face a nominal loss, in contrast to that  awaiting  an investor with the same time horizon  buying a bond at a negative yield . Take the  2%  Jan 2022 Bund, which is priced at €113.80. Over  the 6.7 years  to maturity the interest will amount to €13.40 but this will be offset by the capital loss of €13.80, ensuring a negative nominal return.

Why  would anyone buy a bond which gives a loss if held to maturity? Some argue that investors are now  simply buying on the expectation that someone else will buy it at a higher price, a classic bubble, but there are other explanations. In the Swiss case investors may believe that the currency will appreciate  significantly, so ensuring a positive return for a non-Swiss  buyer. One doubts if many expect the euro to  outperform most of the other major currencies, however, so other factors are at work. One is QE, in that the ECB is a buyer in the market at any yield above -0.2%. The ECB will not buy all the bonds at issue, however,( the limit is 33% , at least for now)  so  investors  will still be left holding two-thirds of the market.

A second rationale relates to banks, the main buyers of shorter-dated bonds.  For them, any excess liquidity deposited with the ECB costs them 0.2% so any yield above that, even if negative, is viewed as a plus. The implication is that banks would also prefer to park liquidity in bonds, however low the yield, than lend it to firms or households – such lending requires higher capital backing and in general carries  a higher perceived risk. One should also remember that banks are also now required to hold a specified amount of assets in liquid form, as part of the Basel 111 regulatory changes, which  in effect means a greater demand for government bonds at the same time as the ECB has entered the market as a buyer.

For investors as a whole the low or even negative return on bonds is supposed to act as an incentive to switch to other assets, including equities and corporate debt, although again, regulatory constraints for pension funds and insurance companies may make a significant switch into riskier assets problematical.

In the short term, then, a combination of QE and pessimism on growth and inflation has led to a collapse in the risk free rate of return, with the possibility of 10-year yields and beyond turning negative. That has serious practical implications for savers and those relying on annuities in retirement. At another level, it throws up difficulties for asset valuation models, as the risk free alternative is now a negative number. How all this ends is anyone’s guess but there is a paradox at its heart; if QE stimulates growth and leads to a rise in inflation over the medium term, perhaps due to a much weaker currency, it makes negative bond  returns in real terms all the more likely for anyone buying to hold at these levels.

 

 

The Fed, Policy Rules and Market rates

It is a long time since the US Federal Reserve last raised rates ( in June 2006, from 5% to 5.25%)  and one has to go back over a decade to find the start of a tightening cycle (in June 2004, from 1% to 1.25%).Consequently there are many in the financial markets who have never  experienced a period of rising interest rates although that may be about to change, judging by rate expectations; the Fed Funds rate is currently trading at 0.12% ( the Fed’s target is 0-0.25%)  and the futures market is pricing in a strong probability of a rate rise over the late-summer and autumn this year.

Of course the Fed itself has also changed tack in terms of monetary policy, in response to the improvement in the real economy, including strong employment growth and a decline in the unemployment rate. Consequently the FOMC terminated its asset purchase programme last year and has now dropped its forward guidance on rates, indicating that the decision will now be data dependent. That does not mean that a rate rise is inevitable ( some FOMC members are worried that inflation is low and may not pick up as anticipated)  and the exact timing is clearly debateable, but what is interesting is the divergence between the market’s rate expecations over the next few years and that indicated by the published views of the FOMC.

One standard method of predicting the future path of Fed policy is using the Taylor rule, first put forward in 1993. The original  formulation  envisaged an equilibrium Fed funds rate of 4%, with the Fed adjusting policy if growth moved away from trend or if inflation deviated from target (assumed to be 2%). A popular variant of the model uses an empirical relationship between unemployment and economic growth (Okun’s law) to derive a rule which includes the unemployment rate  as a proxy for full employment instead of GDP or

r = 1 + 1.5PCE – (UR -URF)

where r= the Fed funds rate, PCE is the annual inflation rate ( the Fed prefers the consumption deflator to the CPI) , UR is the unemployment rate and URF is the rate associated with full employment

We can use such a model to project the Feds fund rate implied by the latest forecast made by the FOMC members(based on the  central tendency). That envisages the unemployment rate falling to 5.25% by the fourth quarter of 2015, and as such around the Fed’s idea of full employment ( an unemployment rate between 5.2% and 5.5%). The FOMC expects inflation ( at 1.3%) to remain below target this year but to  move up to just under 2% by 2017, and based on those forecasts the Taylor rule implies a Fed funds rate of 2.9% at end-15, 3.9% at end-16 and 4% in 2017.

Indeed, the rule indicates that Fed policy should have been much tighter at the end of last year ( around 2%) . The severity of the Great Recession and the sluggish nature of the economy have persuaded central bankers that this time is different, however, and Janet Yellen, in a speech in June 2012 *  put forward reasons why policy should remain looser for longer and discussed two alternatives to the Taylor rule as outlined above.

The first , a Balanced rule, doubled the weight on the unemployment variable. This did imply a different rate trajectory given the level of unemployment at that time but such a model now shows little difference from the basic Taylor rule, because  the unemployment rate  is so close to full employment- the  path of  the Fed funds rate is virtually identical in the two models.

Yellen went on the put forward a third approach – optimal control. This involves using a range of Fed funds rate in a model of the US economy  in order to minimize a loss function, with the latter set in terms of the deviation of unemployment and inflation from target levels. That approach did imply a longer period of zero rates and a slower tightening cycle, albeit  also rising to around 4% in the medium term. We do not know what that optimal control approach states now but based on the charts published in the Yellen  paper  it may imply a Fed funds rate of 1.25% by end-15 and 3% by end-16.

The Fed will not slavishly follow any model in setting policy and the data may disappoint but what is striking is the divergence between market rates and that implied by any of the three models if the fed ‘s economic forecasts are broadly right. For example, 3- year bond yields are currently trading at 1% with the 2-year at 0.63%, which implies a 1-year rate in two years time of 1.7%. If the market is wrong in its implied pessimism on inflation and growth  there is a very big bond market accident waiting to happen.

* Yellen, ‘Perspectives on Monetary Policy’ , Federal Reserve, June 2012

 

 

 

How Many Irish bonds can QE buy?

The ECB’s expanded  QE is due to start in March and a figure of €12bn is often reported in terms of the amount of Irish sovereign debt that the Central bank can buy. The case is not that clear-cut , however,  and the limit may be only €9bn from June.

Bonds purchased in any EA  country  under QE are subject to a number of restrictions. The first limits the share  of the total each country can buy. In Ireland that means about 1.7% of  the €60bn per month  QE target, which includes private and  EU supranationals as well as government debt. The second limits central bank buying to 25% of any issue. A third puts a  33% ceiling  on the amount of   any issuer’s debt that can be held . The latter only makes sense relative to the 25% limit if existing central bank holdings are included .

Indeed,the ECB does already own some sovereign debt, purchased under the Securities Market Program (SMP) , which stands at €144bn having fallen from over €200bn (via bonds maturing). We do not know the  current breakdown of that holding by country but the ECB did publish that data as at end-2012. At  that time €14bn of Irish government debt  had been purchased, 6.5% of the then total, which implies about €9bn today, assuming redemptions were broadly proportional over the past two years.

The Irish Central Bank also owns government debt stemming from the Anglo Promissory note. Part of that was repaid  in 2012 via the issuance of €3.5bn  of the 5.4% 2025 bond. The Bank announced it had sold a portion of that in 2013 but presumably still owns around €3bn of that issue. In addition, the Central bank received €25bn in long term bonds as part of the Promissory note deal in 2013   and again has sold a small amount, leaving €24.5bn.

Another QE stipulation is that  only bonds with a maturity  between 2 and 30 years are eligible, which in Ireland’s  case gives a current figure of €86bn. That implies  €8.5bn of central bank holdings are within that range ( including €5.5bn of the Prom note bonds) which alongside the €9bn SMP figure gives €17.5bn or 20% of the total at issue. The issue limit therefore leaves only 13% open to further purchase, which is just under €11bn

Finance Minister Noonan stated that the Central bank had ‘ample room ‘ to purchase Irish debt. However, the situation changes in mid-year as  €3bn of the Prom note bonds redeem in June 2045 (i.e. would then fall within the 30 year limit), so  from then on the CB’s eligible  holdings rise to €20.5bn or 23% of the total, implying less than €9bn could be added.

Some SMP holdings will mature over the next 18 months and the Central bank will sell some of its Prom note bonds, so giving some room for additional QE, The NTMA will also issue new debt (perhaps another €11bn this year) which  will qualify as long as it is over 2 years so raising the remaining limit for the Central bank rule. Perhaps not quite as ‘ample’ then as some think, in the short term and with more moving parts.

 

Are current euro sovereign yields irrational?

The Irish Government issued a 15-year bond recently at a yield of 2.49% and the 10-year benchmark is trading at under 1.6%, a far cry from the double digit levels seen in the latter just a few years ago. Sovereign yields across the euro area have plunged of course ; 10-year yields in Italy are down at 2.3%, Spain is trading at 2.1% while Portugal is just over 3%.The perceived risk of default  over the next five years has clearly changed dramatically and judging by Credit Default Spreads is now around 10% for Italy, 17% for Portugal and less than 5% for Ireland.

German yields are lower still and bund yields can be considered the nearest thing to a risk-free rate in the euro area. Consequently, a 10-year bund rate of just 0.8% implies that the market expects short term  euro interest rates to stay low for a long time and then to rise only slowly; the implied 5-year forward yield in five years time is under 1.5%.Interest rates are also extremely low in other currencies but higher than bunds, with the US 10-year yield at 2.34% and the UK  trading at 2.1%. The 5-year forward rates also tell a  very different story- the US implied yield is just over 3% , with the UK at 2.8%.So the bund curve indicates  that investors expect inflation to stay low for a long time and not to threaten the ECB’s 2% target. This , in turn, indicates an expectation that growth will remain weak in Germany, with the prospect of  secular stagnation clearly seen by many investors  as more than just an academic debating point.

That scenario  may or may not materialize but it would appear to be a plausible rationale for the current level of bund yields. The problem is , though, that such a scenario would be very negative for other euro zone economies, particularly those with extremely high debt burdens and needing an internal devaluation to become more competitive within the zone. A prolonged period of very low nominal GDP growth, possibly with price deflation, would severely damage fiscal capacity and via the denominator put further pressure on already stretched debt ratios; Italy’s ratio in q2 of this year  was 134% , for example, with Portugal at 129%.Indeed, Italy’s GDP is already falling steadily and in real terms is back to the level recorded in early 2000.

Buyers of peripheral euro sovereign bonds can point to Draghi’s ‘whatever it takes’ mantra and  it would seem that the market views full-blown QE (i.e. buying sovereign debt)  by the ECB as inevitable and that somehow this will save the day. The Governing Council is clearly split on the issue, however, and some opposed the current plan to buy Asset Backed securities which explains why Draghi is at pains to add the qualification ‘within our mandate’ to his statement that there is unanimous support for additional non-standard policy measures.- clearly some members feel that the ECB could be acting ultra vires. So QE is not inevitable but if it did emerge (through a majority vote perhaps) it is still problematical relative to experience elsewhere. The banking system in Europe  is the main source of credit, unlike the US where the disintermediation of banks is the norm, so QE may not have much impact. Moreover Central banks in the US, the UK and Japan have bought their own sovereign bonds but the ECB has no sovereign bonds to buy- it would have to choose among the sovereign bonds of the 19 member states. How would it proceed if it chose a €1,000bn target- would it buy in proportion to the country weights within the euro area, or target higher yields? The former would imply the purchase of large amounts of  bunds and say only €12bn of Irish bonds. In addition the ECB would be buying at extremely low yields ( elevated prices) so adding a high degree of market risk to the credit risk inherent in QE.

More fundamentally, transferring ownership of some sovereign debt from one group of investors to another (in this case the ECB) does nothing to change the debt burden unless of course one believes that the ECB will effectively tear up the bonds or hold to maturity and then pass the proceeds back to the respective governments. One doubts if that would be acceptable to the Germanic school within the Governing Council, at least on any scale that would make much difference to high debt countries, but such concerns do not seem to weigh much on investors at the moment.