by Jaime Caruana
Today my topic is the global bond market, a subject of keen interest to asset managers. The bond market in major currencies plays a much larger role than funding the gap between government spending and revenues. Nowadays the yields on such bonds ramify widely in the real economy and financial markets. Corporate treasurers use these yields to assess the value to shareholders of real investment projects. These yields determine how much households pay on their fixed rate mortgages. And they have become the cornerstone of valuation of other long-term assets, ranging from equities and real estate, to gold and foreign exchange.
Futures and options based on benchmark bonds are used to hedge other long-term bonds. Long-term sovereign bonds help pension funds and insurance companies to measure the value of their liabilities and are their core investment. Today, however, major bond markets are sending signals that are very hard to decipher. A previously unthinkable portion of major government bond markets is trading at negative yields. More than $7.5 trillion of sovereign bonds in major countries is trading at negative rates as of 11 November 2016. Investors are paying governments for the privilege of lending to them! You know this, but my theme today is that there is a risk that we lose our sense of wonder in facing that fact.
I shall consider in order the three stories that we tell ourselves about why long-term yields are so low: low growth, monetary policy and market dynamics. I hope to leave you today with two suggestions. First, that we risk putting too much weight on slow growth, or even secular stagnation, in our understanding of these very low yields. And second, that there is a corresponding risk that we put too little weight on unconventional monetary policy and market dynamics. One implication is that asset managers should reflect upon snapback risk – and risk managers should make sure that they do.
Low growth, perhaps secular stagnation
The most widely credited story for why major government bonds trade at such low yields is the combination of low growth and sub-par inflation. Low real growth is often ascribed to an excess of desired saving over investment. This excess is itself variously attributed, first, to Asian savings (the “savings glut”), second, to a technology-driven decline in the price of investment goods, third, to corporate caution, or fourth, to a managerial agency problem that leads to firms not investing their cash flow.1 The second, and sometimes the first, has been dubbed secular stagnation. This evokes the 1937 pessimism that presumed the American frontier was closing and technology had nothing more to offer. These are all stories that could result in growth being relatively low for the indefinite future. In emerging market economies, there are a variety of reasons to foresee growth headwinds in the future. Credit cycles are maturing and China is rebalancing from export-oriented manufacturing to services. Its approach to the international frontier in manufacturing and the peaking of its labour force all suggest slower growth. Elsewhere, productivity growth has fallen and in several countries demographics suggest slower growth. Inflation is generally below targets in Japan, Europe and the United States. Inflation outcomes are much more diverse in emerging market economies, with important economies struggling to bring down inflation in excess of declared norms, but, on balance, it is comparatively low there too. Putting together low real growth and sub-par inflation in major economies, nominal GDP growth and nominal bond yields have fallen broadly together since 1980. By this rough metric, the trend towards lower bond yields in recent years makes sense. However, in these major bond markets, current yields are a little on the low side compared to nominal GDP . In the US Treasury market, the German bund market, the Japanese government bond market and the British gilt market, nominal GDP has been growing at a rate noticeably higher than the 10-year bond yield. Both the coincidence of this observation and the observed size of the current deviation are worth noting.
As with inflation performance, matters are more diverse in emerging market economies, but there is still evidence of low bond yields. Asian yields are broadly below nominal GDP growth, although not by unusual margins. In Poland and Turkey as well, bond yields are below nominal GDP growth. In the commodity-exporting economies of Latin America, Russia and South Africa, where central banks are contending with the inflationary consequences of exchange rate depreciation, bond yields are above nominal GDP growth. To sum up, government bond yields have fallen below nominal GDP growth in the G4 currencies and many emerging market economies. While the margin is not large relative to historical experience, the breadth of the observation across many bond markets should give us pause. But we should not neglect the other two stories, monetary policy and market dynamics.
Monetary (balance sheet) policy of central banks
Central banks have influenced global bond yields in a variety of ways. They have set their short-term policy rates at low, even negative, levels; they have guided market participants to expect such rates to be low in the future; and they have published low forecasts of their future equilibrium rates. The Federal Reserve’s Federal Open Market Committee, for instance, published a median long-term federal funds rate projection of 3.8% in July 2015, but only 2.9% in September 2016.2 In addition, central banks have bought bonds and other assets, so-called quantitative easing (QE). Most commentary by market participants concerns the flow of central bank bond purchases, but I would like to draw your attention to the accumulating stock of bonds held by officials. The flow numbers are impressive: with the Bank of England joining the Bank of Japan and the ECB, purchases are heading to $200 billion per month.
But the real measure of the official thumb on the scale in major government bond markets is not the flows, but the stocks, and these are now very substantial indeed. What is often overlooked is that QE comes on top of long-standing ownership of major government bonds by official reserve managers. Here the policy is the sale of domestic currency – official foreign exchange intervention to hold down the domestic currency – and the investment of the proceeds in major government bond markets is a byproduct. Thus, by official holdings, I mean not only those of central banks implementing QE, but also those of foreign exchange reserve managers. The upshot of both intentional and incidental official bond buying is that officials hold a large and growing fraction of G4 government bonds. The fraction of US, euro area, Japanese and UK government bonds in official hands is high : a third of government bonds in those economies are held by central banks and other reserve managers. While this fraction declined a bit in 2015 as a result of official reserve decumulation, it is stabilising in 2016 as reserve drawdowns slow and the ECB and Bank of Japan accelerate purchases and the Bank of England resumes them. Over half of the US Treasury market is held by the official sector.
What is clear is that QE in one major bond market tends to push down yields in other major bond markets. This is evident from a range of studies using yields of different frequency and various techniques. As central banks remove duration from the market and thereby depress yields, investors scan the globe searching for yield with which to fulfil promises to policyholders and future pensioners. One symptom of this search for yield is the cross-currency basis, which makes hedging dollar investments more expensive for Japanese and euro area investors. In sum, the official thumb on the scale in global bond markets should not be underestimated. And what the central bank gives, it can take away – or market participants can fear that it will take away, especially if they are paying more attention to flows than to stocks. Just in July-August 2016, Japanese government bond yields rose sharply after the Bank of Japan said that it would perform a comprehensive assessment of its policy. Reactions can be swift, as demonstrated in the first of the series of market moves that came to be known as the taper tantrum in the United States.
Market dynamics
Just as official holdings should not be underestimated, neither should the effect of international market dynamics. These have arguably amplified the decline of rates from fundamental reasons and from official purchases. Before I lay out five such market dynamics – I am sure that this audience could add to my list – let me comment on Robert Shiller’s call that the US bond market is in a bubble. Shiller defines a bubble as a sort of thought virus, a highly contagious feedback loop “where price increases generate enthusiasm among investors, who then bid up prices more, … until prices get too high. During that period, people are motivated by envy of others who have made money doing it, regret in not have participated and the gambler’s excitement. Stories develop that justify the bubble …”. While terming the bond market a “new normal” bubble, he recognises that it does not have all the characteristics of a bubble, in that it features less greed and more fear. He suggests that people are buying bonds to protect themselves, because they feel anxious about their future. My own view is that, while bonds might have, through various social processes, reached prices that cannot be sustained, it is not all that useful to call this situation a bubble. Many investors, including perhaps those in this room, are far from enthusiastic about the prospects of the negative-yielding bonds in the portfolios that they manage as professionals. As I shall argue, institutional constraints may be forcing some of you to do at work what you are not doing at home in managing your family’s finances. The bond market strikes me as less a case of irrational exuberance and more a case of irrational, or even rational, despondency. Let me now turn to the five market dynamics. First, insurance companies that are short-duration have chased yields down. Data published by the European Insurance and Occupational Pensions Authority (EIOPA) show that German and Austrian insurers have a particularly large mismatch between the shorter duration assets and their longer duration liabilities. As bond yields fell, such duration-mismatched investors had to redouble their purchases of longer-dated bonds, contributing to a downward spiral of yields. More generally, insurers’ reluctance to part with long-term government bonds means that the price impact of central bank purchases tends to be magnified. Second, the decline of bond yields has come at a time when many pension funds are shifting from equities to bond investments. In some cases, firms are closing defined benefit plans and funding them with bonds or annuities. Some have called this de-equitisation. In the United States, the Federal Reserve’s flow of funds data show that pension funds’ weight on equities and mutual fund shares hit 58% in 1999 and 55% in 2005, but was only 45% in the second quarter of 2016. Third, the particular constraints put on QE can create market dynamics. In particular, the ECB constraint against buying bonds yielding less than the ECB deposit facility rate (currently –40 basis points) leads market participants to anticipate longer-duration purchases when bond yields threaten to become sufficiently negative in the euro area. The lower yields go, the farther out the yield curve the ECB must purchase. Fourth, extrapolative expectations of further rate cuts or continued low volatility can drive down bond yields and implied volatility. For instance, skewed demand in hedging markets, ie strong demand for protection against a further fall in yields, was seen in the run-up to the “bund tantrum”. Investment strategies designed to profit from low volatility may have played a role in the flash rally in US bonds in October 2014. Fifth, negative yields in Japan and Europe and narrow corporate spreads in Europe have put pressure on US bond market yields and spreads. This pressure recalls the lack of responsiveness of US bond yields to the Federal Reserve tightening in 2004–06, the so-called “conundrum”. US investors, despite the Federal Reserve’s mooted tightening, look at negative government bond yields abroad and see the purchases from abroad, and, for a time at least, accept lower yields on US Treasury bonds. If too much weight is placed on the low growth story and not enough on the official holdings and market dynamics explanations, market participants may leave themselves vulnerable to what has become known as snapback risk. It is useful to distinguish two different types of snapback risk. These correspond to a decomposition of bond yields into two components: expected short-term rates over the life of the bond (the expectation component) and the rest, also known as the term premium. The latter is usually substantially positive, reflecting the demand by investors for compensation above and beyond expected short-term rates for holding an instrument with considerable price risk. The estimated term premium is at an unusually low level on both sides of the Atlantic. So a snapback could take the form of a sudden rise in market participants’ expectations of short rates over the medium term or a sudden rise in the term premium.
Both could go together, but the term premium may prove less controllable. Central bank communication can guide expectations of short-term rates. And central banks may use short-term rate setting, forward guidance and bond purchases to prevent the sudden decompression of the term premium. But market dynamics could play a particular role in a sudden rise in the term premium. The taper tantrum in the United States in the second quarter of 2013 featured a rise in the US term premium that seemed to spill over into Europe. For its part, the bund tantrum of the second quarter of 2015 featured a rapid rise in the euro area bond market. One can see both types of snapback on either side of the Atlantic in 1994. That year, the Federal Reserve started a tightening cycle of short-term rates, and in response investors revised their expected short-term rates upwards. Meanwhile, the Bundesbank was actually lowering short-term rates and the Bank of France and others were following suit. Nevertheless, bond yields in Europe followed those in the United States very closely, thanks to a widening of the term premium. This is a helpful example although today central banks communicate their actions much better.
Markets are moving as we speak, and we would do well to recall that bond yields have been rising unevenly since July, reflecting both higher expected short-term rates and the term premium. Last week’s events are striking, and a first reading of them points to the transatlantic transmission of higher bond term premia. The US 10-year Treasury bond yield rose 30 basis points last week and, according the analysts at the Federal Reserve Bank of New York, this reflected a similar increase in the term premium. European 10year bond yields went up too, in a range from about half to as much as the rise of the US Treasury yield. Our preliminary assessment is that the European bond yield increases reflected a rise in the term premium as well.
Conclusions
To conclude, in understanding low global bond yields, it is easy to overstate the influence of slow-growth fundamentals and to understate the role of central bank actions and internal market dynamics. The BIS view is that very low bond yields that are sustained for a very long time cannot be an equilibrium. This is because they tend to lead to financial imbalances that can misallocate resources and lead to lower productivity. If a boom ends in financial instability, it can have huge economic costs and further sap productivity. If it turns out that the pessimists are right about the slow growth fundamentals, I would hope that we will not resign ourselves to low productivity. Nor should we forget that persistently low rates do not just reflect exogenous forces but also shape events. Instead, we should make the structural reforms necessary to restore productivity growth, which is so important to the standard of living of our children and grandchildren. For asset managers, the appropriate interpretation of the signals coming from the bond market is a key call. If central bank actions or market dynamics matter, then portfolio managers and risk managers might well give some thought to the consequences of a change in them.
The author, Jaime Caruana is General Manager, Bank for International Settlements.
