The low level of long bond yields continues to puzzle even the Federal Reserve Chairman Alan Greenspan. As he recently pointed out, 10-year US Treasury yields are around 80bp lower than they were last year despite a 200bp increase in the Federal Funds Rate. Could the increase in “structural” demand for bonds have more to do with these low yields than he thinks? By Phyllis Reed.
Yet again, bond bears have been disappointed this year. Despite rising short rates and the threat of higher inflation caused by the increase in commodity prices, long-term yields have remained underpinned by increasing investor demand. Is this bid for duration and yield here to stay or will money start to flow out of bonds?
Long bonds, arguably, may now look historically expensive given the already flat yield curves and valuations versus equities. But is it right to use history as a guide in the face of potentially long-lasting changes in the market’s appetite for bonds? Can money really start to flow back into stock markets given pension fund and insurance company constraints?
The macro-economic outlook
Although demand has become an important factor for fixed-income, the debate over the direction of the US economy remains very important given implications for global growth and market pricing. Will US economic activity rebound in Q2/Q3 or will the recent “soft patch” in growth amount to something more?
On balance, both economic forecasters and the Federal Reserve continue to believe the former. According to forecasts published by Consensus Economics, US GDP is expected to grow by a respectable 0.9% on the quarter in Q2 and Q3.
The Federal Reserve also believes the outlook remains favourable and that the recent slowdown in growth is transitory. In its May FOMC minutes, the Fed pointed out that policymakers should not overreact to a comparatively small number of disappointing indicators, especially when economic fundamentals appear to support continued solid expansion. In the Fed's view, low interest rates, robust underlying productivity growth and strengthened business balance sheets were expected to support economic growth at a pace sufficient to gradually eliminate remaining slack in resource utilisation.
The short-end of the bond market, on the other hand, continues to take a more pessimistic view of the US economy. If the Federal Reserve continued to "tighten at a pace which is measured" and raise rates by 25bp at each of the FOMC meetings remaining this year, Fed funds would end the year at 4.25% instead of the 3.75% currently priced into Fed fund futures. This deviation in pricing can only really be put down to the fact that the market already discounts disappointing growth and a “pause” in Fed tightening.
Outside of the US, rate expectations have also come down recently.
In the UK, further weak retail sales, CIPS and housing market data have prompted quite a large shift downwards in rate expectations, with short-sterling futures now discounting around a 50bp drop in rates by year-end.
In the Eurozone, the market now also appears to be moving towards pricing in some probability of easing, after a raft of
disappointing economic data and growing pressure on the ECB to cut rates in the wake of the rejection of the EU constitution by French and Dutch voters. Although the ECB has been reluctant to respond to pressure in the past, markets are beginning to think things could be different this time around.
As Arnaud Achour, head of DCM origination at SG CIB, points out, "Trichet [ECB governor] has for the time being rejected calls for a rate cut, and still believes that the best remedy for low growth and high unemployment is to deliver low inflation." However, SG CIB anticipates that slower economic activity may prompt the ECB to cut rates in Q3 to 1.75% and to 1.50% in Q4.
With a good deal of bad economic news already priced into the short-end in the US and Europe, where does this leave the long bonds, especially given the overall flatness of yield curves?
Yield curves: historical perspectives
The recent outperformance of long maturity bonds in the face of higher short rates in itself is not surprising given the relationship between official interest rates and the shape of the yield curve. As Chart 1 illustrates, the yield curve in the US tends to flatten as rates go up and steepen when rates go down – as simple as that.
However, it is also clear from the chart that the US yield curve is now actually much flatter than it would be given the actual level of short-term interest rates. Indeed, the current curve slope of around 37bp between 10 and two-year yields has normally been consistent with a Fed funds rate of around 5%.
Another stark contrast this time around is the weight of the large government borrowing requirement on the yield curve. Last time the US yield curve was this flat – in 2001– the Federal government had a surplus of $128bn compared to an expected deficit of around $350bn this year. Normally, a deficit this size would put a steepening bias on the yield curve.
So what has caused long-term yields to remain so low this time around? Out of all of the potential reasons pointed out recently by Greenspan, increased fixed-income demand is cited as very important. In addition to the $200bn in US Treasury debt amassed by Asian central banks in 2004 alone, demand from other investors has picked up recently as well. Indeed, March US TIC (Treasury International Capital) data show a significant pick-up in net purchases of $42bn by other overseas investors, more than making up for the $15bn in net sales seen by foreign official institutions (the first net sales figure since 2003).
In addition to this overseas demand, Federal Reserve funds flow data also highlight an increase in pension fund demand, with defined benefit pension funds making net purchases of fixed-income assets of $25bn in 2004, the highest level in over 10 years.
But what are the prospects for pension funds’ appetite for bonds, particularly given the potential for foreign central bank demand to wane further in the event of a change in the Chinese exchange rate regime? Is pension fund demand significant enough, in itself, to prompt a further deviation in yield curve shape versus the trends in interest rates and government borrowing? Could the yield curve, in fact, invert?
Remember, last time the US curve was inverted – in 2000 – interest rates were much higher (6.5%) and the Federal government had a surplus. But as seen in the UK back in 1999–2000, when the strict Minimum Fund Requirement was in place, demand for long-dated fixed-income from pension funds can have a sizeable flattening effect on the yield curve. Could upcoming and existing regulatory changes prompt the same situation to occur in the US and Europe?
Bond valuations versus equities
Historic comparisons also leave long bonds looking potentially expensive versus equities. In the US, for example, the earnings yield pick-up on the S&P 500 relative to the US Treasury 10-year yield is now well over 100bp – the most attractive levels seen since 1980.
But as is the case with the yield curve shape, are these historic yardsticks as relevant as they have been in the past – in the face of a potential further increase in investor demand for fixed-income? Closer examination of the outlook for pension funds in an environment of regulatory changes strongly suggests that the answer is no. The yield curve may be flat and equities may be cheap, but that is unlikely to mean much to a variety of investors who still have to buy fixed-income.
Pension fund demand: still the big story?
The demand for fixed-income being generated by pension funds has undoubtedly become a major focus for markets this year, prompted largely by the changing regulatory landscape in a number of countries. Although the shift towards fixed-income is, by itself, hardly news, upcoming changes in pension fund regulation – particularly in Holland – are expected to further increase demand for long-dated fixed-income.
From 2006, the new Dutch supervisory framework for pension funds, known as the FTK (Financieel Toetsingskader), is set to come into effect and will encompass a number of major changes to the current system. Most importantly, liabilities as well as assets will be marked to market. Instead of the fixed rate of 4% as used under the current system, the present value of pension fund liabilities will be calculated by discounting net cashflow using the actual risk free interest rate term structure. The Dutch Central Bank (DNB) has decided that the swap curve will be used to calculate this term structure.
Against this background, the present value of liabilities will become more volatile and sensitive to moves in interest rates across the yield curve. Consequently, assets used to cover these liabilities must more closely reflect interest rate risk if coverage ratios requirements are expected to be maintained. In addition, the maturity of fixed-income assets should also more accurately reflect the maturity of the liabilities.
At present, a large “duration gap” exists between Dutch pension funds’ fixed-income assets and their liabilities, with the average maturity of bond assets estimated to be five to seven-years and the average liabilities around 15-years. The new regulations are expected to prompt a sizeable closing in this gap as pension funds lengthen the duration of their assets to more accurately match their liabilities.
Although Holland is a small country, the numbers are still likely to be influential given the large private pension liabilities already outstanding. According to numbers published by UBS Global Asset Management, Dutch pension fund assets amounted to US$708bn in 2004, equating to 124% of GDP.
As Jitzes Noorman, fixed-income strategist at Rabobank, says, “Given a duration gap of roughly 10 years, the fact that only 40% of investment portfolios are invested in bonds – and assuming an average cover ratio of around 110% – one could argue that due to the FTK, Dutch pension funds need to double their fixed-income portfolios via the purchase of ultra-long bonds. This would result in an additional demand of about €170bn for ultra-long bonds. As the current outstanding volume of Euroland government bonds with a maturity in excess of 20-years is around €295bn, such a development would result in a marked premium for ultra-long bonds.”
The introduction of FRS17 in the UK and recent changes in pension fund regulations proposed by the Bush administration, encapsulate similar changes, with liabilities and assets more transparently being marked to market and corporate bond yields being used to discount the present value of liabilities.
The bottom line is that pension funds are continuing to come under pressure to reduce balance sheet volatility brought on by holding assets which do not appropriately match their liabilities. Undoubtedly, increasing fixed-income investment will help to reduce this volatility.
This asset reallocation towards bonds still looks a long way from complete, as evidenced by UBS numbers for 2004 which show US, UK and Dutch pension fund fixed-income weightings of 34%, 24% and 39%, respectively.
Although this represents a small increase over the years in the US and UK, these weightings are still seen to be far too low for pension funds with maturing liabilities, i.e. increasing members who are retired or getting closer to retirement.
This is particularly the case in the UK. As John Ralfe, independent pension consultant and former head of corporate finance at Boots, comments, “The closure of most UK defined benefit schemes to new members in recent years has produced a step change increase in the maturity of their liabilities. But changes in asset allocation are only just starting to reflect this, with UK pension schemes continuing to hold most of their assets in equities rather than bonds."
Consequently, it is very likely that we will continue to see significant demand for bonds from this sector as weightings increase in the future to levels which are deemed to be more appropriate.
The numbers could potentially be quite large. For example, as a very rough estimate, each 1% increase in US private pension funds’ fixed-income weighting would unleash approximately $75bn in new demand.
To put this in context, it would only take a 3% increase in US weightings to completely replace the entire stock of US Treasuries held in China, and a 10% increase (albeit unlikely in the short-term) to replace substantial Japanese outstanding holdings of around $700bn. In other words, pension fund buying could easily fill any potential demand vacuum caused by Asian central banks if and when China moves to change/abandon the yuan’s fixed peg to the US dollar.
From the issuer's perspective, the outlook looks set to remain favourable in this environment. As SG's Achour points out, "Low yields and flat yield curves have driven down cost of capital and, unlike equity, bonds are non-dilutive."
In terms of maturity preference for issuers, pension fund demand is also playing a role. "Both the UK Debt Management Office and French Tresor cited the need for pension funds and insurance companies to match lengthening liabilities as the main reason for issuing over a 2055 maturity," said Achour. In addition to 50-year issuance, there has been increased speculation the US government will re-initiate 30-year issuance in the face of this demand.
A number of historic measures may suggest that long maturity bonds are expensive at current levels. Yield curves, particularly in the US, are flatter than they have been in the past given the level of short rates and the weight of government supply.
Bond valuations also looked stretched when compared with equities, with earnings yields on US stocks at their most attractive levels versus bonds since the 1980s. However, history does not always give us the best guide to the future, particularly when market conditions are changing.
The asset allocation shift towards bonds from pension fund managers is a long way from complete, with these investors needing to more accurately match their liabilities with fixed-income assets. By itself, this increase in structural demand could challenge valuation yardsticks of the past, leaving fixed-income markets in a position to continue to outperform. Bond bears beware.