Global financial markets: between uneasy calm and turbulence

Abstract

Financial markets experienced alternating phases of calm and turbulence in the past year, as prices in core asset markets remained keenly sensitive to monetary policy developments. Investors also closely followed growing signs of economic weakness in the main EMEs, especially China. Bond yields in advanced economies continued to fall, in many cases to historical lows, while the share of outstanding government bonds trading at negative yields reached new records. Low yields reflected low term premia as well as a downward shift in expected future short-term interest rates. Investors turned to riskier market segments in a search for yield, thereby supporting asset prices despite already high valuations. Unease about such valuations, coupled with concerns about the global outlook and about the effectiveness of monetary policy in supporting growth, resulted in recurring selloffs and bouts of volatility. Markets appeared vulnerable to a sharp reversal of high valuations. Some outsize bond price movements point to changes in market liquidity, but lower leverage should support more robust market liquidity under stress. Financial markets also exhibited persistent market anomalies that spread further, such as a widening cross-currency basis and negative US dollar interest rate swap spreads. These anomalies partly reflected market-specific supply-demand imbalances, sometimes reinforced by the impact of central bank actions on hedging demand. They also reflected shifts in the behaviour of large dealer institutions, which are now less active in arbitraging the anomalies away.

Full text

In 2015 and 2016, financial markets experienced alternating phases of calm and turbulence. As in prior years, prices in core asset markets were keenly sensitive to monetary policy developments. Weaknesses in the main emerging market economies (EMEs), especially China, were again watched closely. Relative to a year earlier, by end-May 2016 equity prices were lower; credit spreads higher; the dollar had depreciated against most currencies; and bond yields were reaching new lows.

As bond yields fell to historical troughs in a number of countries, the share of outstanding government bonds trading at negative yields reached new records. Low yields reflected low term premia as well as a downward shift in expected future rates. In response, investors sought returns in riskier market segments, which supported asset prices. Standard metrics, such as nominal GDP growth in the case of bonds or historical price/earnings (P/E) ratios for equities, pointed to signs of overvaluation. Unease about such valuations, coupled with concerns about the global outlook for growth, resulted in recurring sell-offs and bouts of volatility. Markets appeared vulnerable to a sharp reversal of high valuations. Some outsize bond price movements point to changes in market liquidity, but lower leverage should support more robust liquidity under stress. At the same time, the persistence of low yields could worsen financial system weaknesses over the longer term, through a number of channels.

Persistent market anomalies spread further. Examples included a widening of the cross-currency basis and negative US dollar interest rate swap spreads. These anomalies partly reflected market-specific supply-demand imbalances, sometimes reinforced by central bank actions. They also resulted from shifts in the behaviour of large dealing institutions, which are now less active in arbitraging the anomalies away.

The first section reviews market developments during the past year. The second looks more closely at market valuations, with an emphasis on the role of very low nominal and real interest rates and market liquidity conditions in fixed income markets. The third explores the factors behind the emergence and persistence of certain market anomalies in recent years.

A year of alternating calm and turbulence

Markets experienced periodic bouts of turbulence in 2015 and the first half of 2016, which alternated with phases of uneasy calm. A common theme was the progressive downward revision to the global growth outlook, particularly for EMEs (Chapter III). As in previous years, markets were closely attuned to central bank decisions. But market participants' confidence in monetary policy's ability to steer the economy appeared to falter (Chapter IV). By May 2016, equity indices, commodity prices and advanced economy benchmark bond yields were below their levels of a year earlier, while both corporate and EME sovereign credit spreads were higher. Measures of volatility were stable or higher (Graph II.1). EME currencies, while recouping some of their losses, were in most cases weaker against the dollar compared with a year earlier.

The first episode of market turbulence started in the third quarter of 2015, when the growth prospects of a number of large advanced and emerging economies were downgraded. The spotlight shone especially brightly on China, which for several years had been seen as the global growth engine. A rapid rise in equity prices in the first half of the year, fuelled in part by heavy retail margin purchases, was reversed sharply over the summer (Graph II.2, left-hand panel). The Shanghai Shenzhen CSI 300 Index collapsed from a high of 5,354 on 8 June 2015 to 3,026 on 26 August, while the high-tech-oriented ChiNext board fell even more dramatically. In August, Chinese authorities altered their exchange rate mechanism, allowing the renminbi to depreciate sharply against the US dollar (centre panel). Such events shook confidence in China's ability to achieve a "soft landing" scenario after years of rapid credit-fuelled growth.

Concerns about China quickly spread to other economies and asset prices in August and early September. EME currencies weakened against the dollar (Graph II.2, right-hand panel) while the euro and yen strengthened. Equity prices plunged, particularly in EMEs (Graph II.1, left-hand panel). Credit spreads started to widen again (Graph II.1, centre panel). Volatility rose, especially for equities and commodities (Graph II.1, right-hand panel). The plunge in commodity prices weakened the economic prospects of commodity-exporting countries and of commodity-producing firms, some of them heavily leveraged and with a large weight in key equity and credit indices (Graph II.3).

Markets stabilised in October 2015, but in most cases did not recoup their summer losses. Continued strong data for the United States reinforced the expectation that the Federal Open Market Committee would at last tighten policy at its December meeting. The likelihood of divergent monetary policies between the United States, on the one hand, and the euro area and Japan, on the other, contributed to renewed dollar strength (Graph II.4). However, when the hike did take place on 16 December, ending eight years of near zero nominal policy rates, markets took it in their stride, as the move had already been fully incorporated into expectations.

The new year nevertheless brought a second bout of turbulence. The triggering event was again economic news about China, as weak data signalled a sharper than anticipated growth slowdown. Moreover, the arrival of fresh data soon led to a downgrade in growth estimates for a number of other countries. From the start of the year to mid-February, commodity markets weakened, with oil prices falling below $30 per barrel, and major global equity indices dropping 10-20%. US high-yield spreads widened by almost 200 basis points over this period. The high-yield market was particularly vulnerable to turbulence since US energy producers had issued a large volume of lower-quality debt in recent years.

In January, banks came under particular pressure. Their credit spreads widened sharply and equity prices fell before partially recovering in the following months (Graph II.5, left-hand panel). Signs of stress were particularly evident in the pricing of European bank debt, with the prices of contingent convertible obligations (CoCos) diving precipitously (Graph II.5, centre panel). Clearly, investors were still learning about the risks associated with such relatively recent financial instruments. Bank profits, especially in Europe, were seen as coming under growing pressure as policy rates were cut further into negative territory and as an increasing proportion of European government bonds traded at negative yields (Chapter VI). Consistent with this development, ever deeper negative euro interbank rates initially went hand in hand with wider bank credit spreads (Graph II.5, right-hand panel). Bank credit spreads then partially recovered, following the ECB's announcement on 10 March of various policy measures, including the possibility for banks of borrowing from the ECB at its negative deposit rate under certain conditions.

Weaker growth led to expectations of a longer-lasting, and in some cases stronger, monetary policy stimulus. Government bond yields fell across the major economies. The expected path of Federal Reserve tightening shifted downwards. Central banks explored new stimulus measures, including expanded asset purchases by the ECB and a shift to negative policy rates by the Bank of Japan (Box II.A). But whereas in the past these measures might have buoyed markets, in early 2016 they met with an indifferent or even negative response. While the end-January announcement by the Bank of Japan of negative rates led at first to a rebound in the Japanese stock market and a depreciation of the yen, Japanese banks' stock prices soon fell sharply and the yen strengthened as investors seemingly began questioning whether monetary policy would be able to successfully stimulate the economy.

The market episodes of July-September 2015 and January-February 2016 shared a number of common elements. Weaker global stock and credit markets, and weakness in EME bonds and currencies, pointed to "risk-off" behaviour on the part of investors and expectations of slower growth. Commodity prices weakened, especially oil prices, highlighting the important role of oil producers in global equity and credit indices.

Low-rate environment drives asset valuations

Government bond yields of advanced economies continued to fall during the period under review (Graph II.4, left-hand panel). By May 2016, medium- and long-term bond yields stood well below the already very low levels observed in June 2015, irrespective of whether monetary policy rates had been reduced (Germany, Japan, Sweden), left unchanged (Switzerland, UK) or raised (US) since then (Graph II.6, left-hand and centre panels). Moreover, having been given an upward jolt following the Bank of Japan's decision to move to negative rates, the stock of sovereign bonds trading at negative yields across the globe soared to new highs, reaching close to $8 trillion by end-May (Graph II.6, right-hand panel).

Such low levels of interest rates and yields are in most respects historically unprecedented. The near zero short-term interest rates seen in the United Kingdom and the United States today represent the lowest levels observed since the Great Depression, while current negative short-term rates in Germany and Japan are unparalleled (Graph II.7, left-hand panel). Nominal 10-year bond yields, at between -0.1 and 1.8% for these four countries, are also at or near record lows (Graph II.7, centre panel). And while current 10-year real yields are not unprecedented when compared with ex post real yields since 1900, they are at levels not seen since the inflationary 1970s (Graph II.7, right-hand panel).

The historically low bond yields coincided with low estimated term premia. In fact, estimates indicate that a significant part of the decline in nominal and real bond yields in recent decades reflects a secular decline in term premia, which are also at historical troughs (Graph II.8, left-hand panel).

Sudden jumps in term premia led to sharp yield increases in the second quarter of 2015, in particular in the euro area (Graph II.8, centre panel). Euro area and US term premia estimates then stabilised before edging down again in the second half of the year. By contrast with historical experience, where US yield movements have tended to drive those in other currencies, US bond yields came under pressure when euro area bond yields fell, as investors chasing higher yields moved into US Treasuries. And during the turbulence of early 2016, a global flight to quality pushed down premia on government bonds further still.

Alongside term premia, expectations of future interest rates also played a role (Graph II.8, right-hand panel). Between May and December 2015, the expected average short-term interest rate in the United States over a 10-year period rose some 40 basis points. This rise came to a halt and was partly reversed in early 2016 as investors reassessed US monetary policy prospects. In the euro area, the expectations component played a smaller role but still contributed to lower yields.

Large-scale central bank purchases, possibly reinforced by financial institutions' behaviour, weighed heavily on yields. For example, by end-2015 the Eurosystem had increased its holdings of euro area government debt to almost 17%, while the Bank of Japan held around 32% of outstanding Japanese government bonds. Hedging by institutions such as pension funds and insurance companies may have further boosted demand for government securities. And banks increasingly favoured sovereign bond holdings, in part owing to financial regulatory reforms, but also due to increased demand for collateral in financial transactions.

In this environment, some observers have asked whether government bonds in the main currencies might be overvalued. Although it is difficult to define overvaluation in the context of government bonds, various views have informed the debate. Some have argued that both the natural real rate of interest and, to a lesser extent, expected inflation have fallen substantially for the foreseeable future in a number of advanced economies (Chapter IV). Others have noted that economic growth, productivity, inflation and other macroeconomic fundamentals could well revert to their pre-crisis levels. Even disregarding term premia, this second line of argumentation would view current bond market valuations as overly rich.

Unusually low prevailing term premia are another piece of the puzzle. Bond prices would be overvalued if zero or negative term premia turned out to be unsustainable. True, term premia may not necessarily return to their pre-crisis levels: for instance, inflation may be viewed as less of a long-term threat than in previous years, or investors' risk aversion may have changed. That said, it seems unlikely that deeply negative term premia such as the ones estimated for the euro area can persist indefinitely. The question then becomes when, and how fast, premia will normalise.

Comparing long-term bond yields with the evolution of nominal GDP suggests that yields are indeed currently on the low side. Over the past 65 years or so, the broad trends in nominal GDP growth and 10-year bond yields have lined up reasonably well across the United States, Japan, Germany and the United Kingdom (Graph II.9). Most likely, the real bond yields and expected inflation components of nominal yields have fluctuated in tandem with real GDP growth and inflation, respectively. Currently, bond yields are somewhat below nominal GDP growth in all four countries.

If, for whatever reason, bond yields are "too low", they could snap back at some point as market expectations adjust. The size and potential disruption of a reversal have less to do with the characteristics of day-to-day liquidity provision than with the incidence of forced sales and financial institutions' capacity to absorb the hit. Despite the higher frequency of large price moves in recent years, there is no convincing evidence of a structural decline in fixed income market liquidity (Box II.B). That said, there is no doubt that such liquidity will evaporate under stress, as it has always done in the past. Signs of liquidity illusion in the growing asset management industry suggest caution. At the same time, stronger bank balance sheets, lower broker-dealer leverage and better liquidity risk management should all support more robust liquidity and help contain any damage from periods of illiquidity (Chapter VI).

Alternatively, persistently low yields could end up having pernicious effects on the economy and become to some extent self-validating. By sapping banks' profitability and resilience, low yields may reduce banks' capacity to support the economy (Chapter VI). They may also distort financial and real economic decisions more generally, for instance by encouraging unproductive firms to maintain capacity or by inflating asset prices, thereby weakening productivity (Chapter III). And they may encourage further debt build-up, which could make it harder for the economy to withstand higher rates (Chapter IV).

Equity valuations have also come under scrutiny. As bond yields fell globally post-crisis, equity prices rose, pushing up valuation metrics such as P/E ratios. Stocks naturally became increasingly attractive relative to bonds while lower real interest rates boosted the discounted value of future corporate earnings. Partly as a result, cyclically adjusted P/E ratios in the United States and the United Kingdom have stood well above their historical averages in the past few years (Graph II.10, first two panels) - a possible sign of stretched valuations. The corresponding P/E ratio for Japan has been below average, but that average has arguably been inflated by the financial bubbles of the 1980s (Graph II.10, third panel). Moreover, the run-up of US equities in recent years has coincided with sharply higher leveraged positions (Graph II.10, last panel). And equity prices have also been supported by very strong share buybacks, particularly in the United States. The nervousness observed in global equity markets over the past year, with repeated sharp price corrections, hints at investors' unease with current valuations.

Corporate credit markets were subdued over the past year. Both investment grade and high-yield credit spreads were higher in May 2016 than a year before in the United States and the euro area as well as across EMEs (Graph II.1, centre panel), although euro area spreads benefited from expected ECB purchases of corporate bonds. In part, the general rise in spreads reflected the weakening economic outlook, and in particular the rapidly deteriorating creditworthiness of the energy sector (Graph II.3).

Signs of a turn in the default cycle helped widen corporate credit spreads. In the United States, the rise in the speculative grade default rate that had begun in early 2015 continued uninterrupted throughout the review period, and credit spreads followed upwards (Graph II.11, left-hand panel). For instance, according to Moody's estimates, having risen from 1.8% at the end of 2014 to 4.4% in April 2016, the default rate was projected to rise above 6% by the beginning of 2017. In Europe, the speculative grade default rate also edged up in the past year, albeit less than in the United States (Graph II.11, centre panel). The sharper rise in the US default rate was due partly to the relatively stronger importance of the energy sector there: a greater number of corporate defaults in the United States occurred among energy firms than in Europe, where only one energy sector default has been recorded in the last three years (Graph II.11, right-hand panel).

Exchange rate developments also appear to have played a role in the pricing of credit risk. In periods during which the dollar strengthened strongly relative to the euro, US investment grade and high-yield spreads tended to rise sharply, and vice versa (Graph II.12, left-hand and centre panels). This pattern was evident among corporates in both the energy and non-energy sectors. Moreover, from 2015 onwards, with the euro relatively weak vis-à-vis the dollar, euro high-yield credit spreads tended to trade significantly below those in the United States (Graph II.12, right-hand panel). In the preceding years, before the dollar had strengthened appreciably, euro credit spreads had been instead close to or higher than their US counterparts. While the strong influence of oil producers on US credit spreads has played a role, the relationship is also present among non-energy firms.

One possible explanation behind this pattern could be changing perceptions of the growth outlook - and hence of corporate credit quality - resulting from exchange rate swings. Similarly, a stronger dollar goes hand in hand with tighter US financial conditions (Chapter III). Another possibility is that easier monetary policy in the euro area tends to weaken the euro while at the same time inducing European investors to seek higher-yielding investments such as euro area corporate bonds.

Market anomalies spread

The years since the Great Financial Crisis have been marked by a number of anomalies in financial markets. Pricing relationships that in previous times would have been rapidly eliminated through arbitrage have instead proved surprisingly persistent.

Cross-currency basis swap spreads offer one example (Box II.C). Such spreads normally adjust to eliminate arbitrage opportunities between a forward position in a given currency and the same position as constructed by a foreign currency swap and a forward in a different currency. Indeed, this covered interest parity relationship is taken for granted in economics textbooks. The relationship broke down during the crisis, reflecting counterparty risk and funding liquidity shortages in certain currencies. Since the crisis, concerns about banks' credit quality and liquidity shortages have diminished - but the anomaly has persisted.

A second example is US dollar interest rate swaps (Graph II.13). Normally, the fixed rate offered on swaps against floating rate payments is a small, positive spread over US Treasury yields. After all, the counterparty credit risk of the banks involved in the transaction is somewhat higher than that of the US government. During the crisis, investors' flight to safety and growing counterparty risk concerns drove these spreads to high levels. But in 2010, and again in late 2015, the spreads were actually negative for US dollar swaps, while remaining positive for euro swaps (Graph II.13, left-hand panel).

These anomalies reflect a combination of factors.

One set of factors relates to supply and demand conditions in underlying asset markets. For example, central banks' large-scale purchases of government securities represented a new and largely price-insensitive source of demand. The end of Fed purchases, reinforced by sales of Treasury bonds by some EME official reserve holders, may have created temporary excess supply that drove US bond yields above dollar swap rates. An indication of these conditions is that dealers' inventories of US Treasuries soared as they stepped in to meet the supply (Graph II.13, centre panel).

A second, related set of factors involves hedging demand. Shifts in the perceived likelihood and timing of higher US rates will shift US dollar borrowers' demand for receiving fixed rates via swaps. The same is true for those attempting to hedge currency risk via cross-currency basis swaps. In currency swap markets, given exceptionally low yields, there have been signs that large institutional investors boosted their investments abroad on a currency-hedged basis, putting strains on the corresponding instruments' prices. The footprints of hedging demand have also been visible in the implied volatility of swaptions (options to enter into swaps), which jumped (particularly in euros) in early 2015 (Graph II.13, right-hand panel).

A third factor relates to a reduced willingness or ability to arbitrage, an activity that requires both capital and funding, and is balance sheet-intensive. Large dealer banks play a central role in maintaining these arbitrage relationships, either through their own position-taking or by providing funding to others such as hedge funds. In response to the large losses incurred on their trading activities during the Great Financial Crisis, banks in many jurisdictions have reappraised the risk/return trade-off of their business lines, including proprietary trading and market-making. Moreover, structural reforms, such as the US Volcker rule, as well as bank capital and liquidity requirements, have been tightened to enhance bank resilience and induce a more accurate pricing of risks, reducing the scope for banks to tightly price away arbitrage opportunities (Chapter VI).¹

¹ See L Andersen, D Duffie and Y Song, "Funding value adjustments", mimeo, 10 March 2016, who argue that well documented pricing "anomalies" in derivatives markets can be traced back to so-called funding value adjustments that incorporate the cost of funding the cash or collateral needed to enter or maintain unsecured derivatives positions - a reflection of more comprehensive collateralisation and wider bank funding spreads post-crisis.