Relatively little is known about the Office of Financial Research. The OFR is an Office within Treasury established by Congress “to serve the Financial Stability Oversight Council, its member agencies, and the public by improving the quality, transparency, and accessibility of financial data and information; by conducting and sponsoring research related to financial stability; and by promoting best practices in risk management.” In short, this is where some of the smarter people in the US Treasury department can (or should) be found, and certainly those who are paid to anticipate the risks to the financial system. The reason the OFR is relevant is because earlier this week it released its annual Report to Congress that “identifies threats to financial stability and tools to monitor them.” Or, stated simply, in its 152-page report, the US Treasury revealed what it believes are the biggest threats to the US financial system.
The four main risks (in our opinion) as listed by the Treasury are Credit risk (the bond bubble identified by Jeremy Stein in early 2013), Duration risk (the impact on balance sheets from a 100 bps blow out in rates), Impaired Trading liquidity, as noted most recently by the TBAC, and finally Emerging Market capital outflow spillover risk.
From the US Treasury’s Office of Financial Research:
Since the last OFR annual report in July 2012, uncertainty about U.S. interest rates led to a broad sell-off across global markets, unveiling important fragilities. Figure 3 shows the performance of global asset markets relative to a three-year average and the relative shift in risk appetite. Green represents a lower appetite for risk, for example, as represented by lower equity prices,
wider credit spreads, or lower Treasury yields, while red represents a higher appetite for risk. The dotted line — an average of five asset classes — suggests that despite the decline in prices of certain risky assets, overall risk appetite remains above levels prevailing at the time of the last annual report. In particular, risks in the euro area have abated as its banking and sovereign debt crisis has morphed into a more manageable economic recession risk. Liquidity has improved, equity values have recovered, and credit spreads have tightened on sovereign debt issued by Greece, Ireland, Italy, Portugal, and Spain.
There was a sizeable correction in asset markets in May 2013 and June 2013 on expectations that improvements in the U.S. economy could prompt the Federal Reserve to taper the asset purchase program sooner than expected. Underperformance was most pronounced in emerging markets, with sovereign external and local currency debt spreads widening, implied default risk rising, and currencies coming under pressure.
Poor performance was also pronounced in high-risk sectors and sectors that had previously benefited most from excess liquidity. Those concerns have since partly abated and equity and credit markets have recovered, but the episode was an important mini-stress test for markets that could presage the potential reaction to monetary policy tightening when it does occur. In early October 2013, increased sovereign risk concerns related to the U.S. debt ceiling impasse and government shutdown led to a sharp rise in interest rate volatility, a widening in near-term sovereign credit default swap spreads, and a rise in measures of risk in short-term secured and unsecured funding markets.
The episode was short-lived, and most sovereign risk measures returned to earlier prevailing benign levels after the debt ceiling was temporarily extended. Although market conditions have since calmed, challenges remain. Threats to stability can be generally categorized in two ways: (1) cyclical or structural, and (2) inside or outside the financial system (Figure 4).
Grouping threats in this way helps focus on the causes behind each threat, rather than just symptoms, although some threats contain both cyclical and structural causes. Many of the threats previously flagged by the Office and the Council in their respective annual reports remain relevant.
This section highlights the following potential threats:
• the risk of runs and asset fire sales in repurchase (repo) markets;
• excessive credit risk-taking and weaker underwriting standards;
• exposure to duration risk in the event of a sudden, unanticipated rise in interest rates;
• exposure to shocks from greater risk-taking when volatility is low;
• the risk of impaired trading liquidity;
• spillovers to and from emerging markets;
• operational risk from automated trading systems, including high-frequency trading; and
• unresolved risks associated with uncertainty about the U.S. fiscal outlook.
These risks in isolation do not necessarily lead to systemic weakness. But, in combination, they may leave the financial system more susceptible to adverse shocks.
* * *
And the main threats (in our opinion) as summarized by the OFR:
Among nonfinancial corporations, leverage has been increasing since 2012, and there are reasons to be concerned about a potential deterioration in corporate balance sheets once interest rates begin rising.
After the crisis, nonfinancial corporations managed their balance sheets conservatively to reduce debt and build liquidity, while profits grew at an accelerating rate. Since 2010, however, leverage on investment-grade and high-yield corporate balance sheets has been rising (see Figure 11). Early in the cycle, most of that increase was at corporations with strong credit ratings and low debt. More recently, weaker companies have followed suit. Corporate cash buffers have been steadily diminishing, reversing the hoarding that took place earlier.
Underwriting standards continue to weaken by some measures. Companies with low credit ratings have been among the biggest issuers of new debt, with recent transactions turning more aggressive. There has been a spate of payment-in-kind bonds, which pay interest or dividends to investors with additional debt. Also, less-strict terms are being used in legal covenants attached to leveraged loans. Another sign of weaker underwriting standards are dividend-refinancing loans, which increase leverage through the financing
of shareholder dividends by reducing the capital stock that buffers a firm from insolvency.
Relatively easy financial conditions often are accompanied by, or lead to, a compression in risk premiums and higher asset prices. Loans with weaker covenants (covlite loans) carry less stringent borrower obligations and represent one example of mispriced credit risk. With fewer investor protections for cov-lite loans, expectations of recovery on default are lower. Consequently, a cov-lite risk premium should exist to account for lesser creditor protection (fewer covenants) and lower expected recoveries. Based on historical recovery and default rates, these loans should command a risk premium of 30 to 35 basis points, but are currently priced below or only on par with other comparable loans requiring stronger protections for lenders. By the same token, despite the deterioration in fundamentals, corporate borrowing costs and the spread investors are willing to pay per unit of balance sheet risk are at historically low levels, implying a lower price of credit risk and greater risk of a sharper adjustment in reaction to an adverse shock.
Duration and Interest Rate Risk
Investment portfolios now face growing duration risk — the risk that investors will incur outsized losses in the event of an unexpected rise in interest rates as a result of exposure to long-dated, fixed-rate bonds. Courtesy of a long period of low yields, low volatility, and investors’ search for yield, duration risk is at recent historical highs. Portfolio allocations to fixed income instruments also remain above the recent historical trend, despite the rise in yields in May 2013 and June 2013 (see Figure 12). Thus, losses from a given change in interest rates would be larger than in the past.
These positions increase the vulnerability for some market participants to outsized losses that could be difficult to absorb in the event of an unanticipated increase in long-term rates. To assess the degree of vulnerability, we simulated an adverse interest rate shock to estimate losses by bond funds from an instantaneous parallel shift in the yield curve of 100 basis points from current levels. We then compared the impact of such losses in today’s context to loss rates from a similar hypothetical scenario during the three previous periods of U.S. monetary policy tightening. Losses during each tightening cycle are calculated by averaging monthly estimated losses, where the Barclays Capital U.S. Aggregate Bond Index is used as a proxy for duration and mutual fund bond holdings are based on data from the Investment Company Institute. Figure 15 shows that losses could rise to nearly $200 billion (or 5.5 percent of GDP), underscoring that current bond portfolios are vulnerable to a sudden, unanticipated rise in long-term rates.
Interest rate risk extends beyond nonfinancial bond portfolios. On the asset side, banks have increased their holdings of longer-term assets, leaving them more exposed to interest rate risk. On the liability side, U.S. banks have seen dramatic growth in their non-interest bearing deposits relative to total banking system liabilities. The ratio now stands at a 30-year high. It is unclear how much of the growth is attributable to structural factors or cyclical factors. Challenges exist for banks and regulators in modeling the behavior of these deposits as interest rates rise. There is a nonnegligiblerisk that deposits would shift to alternative, higher-yielding investments as rates rise.
In the event of an adverse interest rate shock, policymakers would likely adopt actions aimed at tempering the rise, for instance through communication and fine-tuning policies. However, determining the underlying drivers of the rise could be challenging. For instance, the roughly 100 basis point rise in long-term rates that took place during the May–June period mostly reflected an increase in term premiums (the extra yield needed for investors to hold a long-term bond instead of a series of short-term bonds) rather than short-rate expectations (see Adrian and Fleming, 2013).
To understand this rise in the term premium, we evaluate the statistical relationship between the term premium and its drivers. Decomposing the term premium is a challenging task, in part because the term premium itself is unobservable. Following Gagnon and others (2010), we constructed a model in which the term premium (the difference between long-term and short-term bond yields) on 10-year U.S. Treasury securities is a function of macroeconomic fundamentals and uncertainty, volatility in financial markets, and supply factors. We estimated the model over the past 22 years, and assessed drivers of increases and decreases in the term premium during the pre- and post-crisis periods.
Figure 16 summarizes our main findings. During most of the 1990s, the term premium steadily declined, driven predominantly by an improvement in macroeconomic factors, as unemployment and inflation decreased steadily. By contrast, interest rate volatility, reflecting interest rate uncertainty, was a key driver of the rise in the term premium from late 1998 to 2000. Beginning in 2008, the Federal Reserve’s asset purchase program became an important driver of the decline in the term premium, while macroeconomic factors became less important. During the most recent period, our model suggests that increased interest rate volatility has more than accounted for the rapid rise in long-term rates, reflecting increased difficulty evaluating the future direction of interest rates. Although our model is imperfect, the preliminary findings suggest that changes in the term premium will be strongly tied to investor perceptions of the future path of nontraditional monetary policy as the Federal Reserve pares back its asset purchases.
Impaired Trading Liquidity
Impaired trading liquidity — the inability to execute large trades without having a significant impact on market prices — could aggravate some of the threats already discussed. Market liquidity measures show a mixed picture. The current high levels of central bank liquidity may be masking some weakness in trading liquidity. Within the corporate bond market, some evidence indicates that liquidity is more bifurcated than before the crisis. Liquidity has become increasingly concentrated, with large, investment-grade bonds showing the strongest liquidity, while some smaller, high-yield issues have become less liquid. The gap has widened as broker-dealers’ securities holdings have shifted toward larger, more frequently traded corporate bonds. The growth in exchange-traded funds within the corporate bond market increases the potential to weaken market liquidity during periods of market stress (see Figure 20 and OFR, 2013).
The sources of diminished trading liquidity are not fully understood. A commonly cited source is reduced broker-dealer capacity and a higher premium for the risk of holding inventory. Broker-dealer inventories of fixed-income instruments have declined since 2007, particularly for corporate bonds. The shift in inventories has occurred against the backdrop of an expanding corporate bond market, reducing the ability of broker-dealers to act as shock absorbers during market stress (see Figure 19). Other changes since the crisis may have also affected structural market liquidity, including shifts in the investor base, risk appetite, and trading behavior.
Foreign Risks: Spillovers to and from Emerging Markets
Accommodative monetary policies in advanced economies, strong domestic fundamentals in select emerging markets, and a structural increase in investor allocations have led to strong cross-border portfolio flows to emerging markets over the last few years. Foreign flows have predominantly targeted emerging market bonds, with cross-over and nondedicated emerging market investors increasing their footprint. In some emerging markets, domestic policies have encouraged local companies to expand debt to high levels and boost leverage (see IMF, 2013e).
Increased sensitivity between the U.S. risk-free rate and emerging market capital inflows has increased the vulnerability of capital flows to a sudden increase in U.S. rates. A reversal in capital flows could highlight vulnerabilities that have built up, particularly where sovereigns and corporates have become dependent on capital inflows to meet near-term borrowing and refinancing needs. An abrupt reversal in inflows would be damaging for countries with external imbalances or near-term refinancing needs.
Yield-seeking capital flows across borders, driven by both external and domestic factors, have driven a decline in local bond yields. Markets for emerging-market bonds have grown increasingly more sensitive to changes in U.S. interest rates (Figure 21). Rises in yields for 10-year Treasury bonds have been accompanied by a depreciation in emerging-market currencies, higher bond yields, and weakness in equity valuations.
The sell-off in emerging markets that began in late May illustrates what could happen once U.S. monetary conditions tighten. Higher U.S. interest rates coincidedwith a pullback in capital flows to emerging markets and increased instability in emerging market assets. The first phase of the sell-off was concentrated in highly liquid proxy trades (trades that use one asset class to take positions in another asset class — for example, positions in commodity producers’ assets to proxy for China). The second phase saw a more pronounced sell-off in assets that had been the primary beneficiaries from excess liquidity since the start of the Federal Reserve’s quantitative easing program (see Figure 22). The third phase reflected further differentiation, with heightened volatility in emerging market assets with the weakest domestic fundamentals (for example, weaker relative growth prospects, low or eroding foreign exchange reserves, large external financing needs, high levels of leverage, or limited policy buffers). On these measures, markets in several countries look vulnerable, including Turkey, South Africa, India, Indonesia, and Brazil (see Figure 23).
Sustained volatility can feed on itself, spilling over toother risk assets as losses trigger fund redemptions and asset sales. Emerging markets are generally more resilient as an asset class than in the past, thanks to liberalized exchange rates, more prudent macroeconomic policies, and issuance of debt in local currencies rather than in dollars. But vulnerabilities remain, including the buildup of corporate debt and leverage (see IMF, 2013e), rapid nonbank credit growth (see insert on Financial Intermediation in China), and diminished policy buffers. Emerging markets are now larger and more connected to developed markets, which means stress can be more readily transmitted directly or indirectly to the U.S. through various conduits, including funding, foreign exchange, credit, and growth channels.
An increase in U.S. policy rates could create challenges for overseas central banks seeking to maintain a looser monetary policy stance. An already challenging policy environment for certain emerging markets with less capacity to absorb external shocks increases the risk of a policy error. Figure 24 shows significant differentiation in monetary, external, and fiscal buffers. Since our last annual report, buffers have eroded in Indonesia, Malaysia, Peru, India, and South Africa. The thinner cushion means less room for stimulus, more difficulty in managing external shocks, and a greater risk of a policy error.