market risk insights - Hermes Investment Management

MARKET RISK INSIGHTS

How to evolve in the new investment climate Eoin Murray Head of Investment

Hermes Investment Office Q4, 2017

For professional investors only

www.hermes-investment.com

OUTCOME #10 Contributed to the development of Japan’s first corporate governance code, encouraging companies to apply best-practice governance and work in the interests of long-term investors

MARKET RISK INSIGHTS Q4 2017

It is not the strongest of the species that survives, not the most intelligent that survives. It is the one that is the most adaptable to change. Charles Darwin Naturalist, 1809-1882

In markets, change is inevitable; for investors, adaptation is a choice. And this quarter we have chosen to introduce a new slice of intellectual DNA into our ever-evolving risk analysis gene pool. While our five-factor matrix – covering volatility, correlation, stretch, liquidity and even risks – has served us well in preparing clients for changing market conditions, we’re confident the addition of environmental, social and governance (ESG) analysis is a step up the evolutionary curve. Historically, we have shied away from tackling ESG risk for a number of reasons, including: A wish to avoid potentially being considered as alarmist, given that

our ESG convictions are typically stronger than consensus views;

A view the sheer complexity of the topic could not be captured

within the scope of this quarterly publication, given that important updates are typically released annually; and,

A fear of muddying a highly important issue with ethical or political

overtones: we consider ESG for investment purposes, not political or ethical agendas.

But ESG has become such a central component of how we (and many of our clients) think about investing, that including it now seems part of a natural selection process. As Niccolo Machiavelli noted, however, taking the lead in “the introduction of a new order of things”1 is no easy task. The great Italian Renaissance strategist, though, also wrote: “The one who adapts his policy to the times prospers, and likewise that the one whose policy clashes with the demands of the times does not.2” Our times are undoubtedly moving into an ESG era with an increasing number of asset owners and investors demanding a more thoughtful and holistic approach to the business of investing money. We are heartened by the shift in investor conversations from governance issues alone to broader environmental and social concerns.

Given the many years of post-global financial crisis (GFC) deflationary psychology perhaps it was inevitable that the first signs of co-ordinated global growth would be greeted with wild enthusiasm. The globally-synchronised uptrend in bond yields has been matched pace-for-pace by new highs in US equity markets, whether in large caps (S&P500), small caps (Russell 2000) or sectorally-focused indices (Nasdaq). But we scratch our collective heads at the euphoria of asset markets. Growth is hardly spectacular. Central bank liquidity (and yes, the taps are still largely on) has, to an extent, floated equity prices higher. And as we have documented many times before, corporate buybacks have also flowed into higher stock prices, boosting EPS growth by lowering the share count rather than driving organic growth. US stock boosterism has been pumped up further by a low dollar with multinational companies getting the double-whammy from currencytranslation gains and stronger export markets. Of course, the prospect (however vague and/or unlikely) of tax cuts, too, provides another fillip to animal spirits. Whether those pressures are creating a sustainable market ‘reflation’ or merely supplying the laughing gas to a late-cycle bubble party remains to be seen. On the downside, capital markets have seen little long-term investment despite the favourable liquidity/credit environment featuring cashed-up corporates. As short- and intermediate-term interest rates rise, how will firms fare as they attempt to roll over a staggering sum of maturing debt at higher interest rates? We have talked plenty before about defaults under the current ‘cov-lite’ paradigm, with updated evidence presented here: Figure 1: Share of leveraged loans that are ‘cov-lite’ or missing traditional investor protections 0.8

Furthermore, it’s clear that issues such as climate change present investors – and the broader global community – with a relatively short, and shrinking, time period to adapt. Action today is even more necessary. We have adapted. However, the ESG analysis – which we will tackle in bite-size chunks over the coming quarters – simply provides another angle on market risk that we also continue to view through our five existing lenses. Since our last quarterly publication global capital markets have become increasingly schizophrenic: torn between a global reflationary growth story and fears about the ‘lower-for-longer’ scenario known, somewhat disparagingly, as Japanification. 1,2

0.7 0.6 0.5

% share

2

0.4 0.3 0.2 0.1

0.0 2007 Europe

2008

2009

2010

2011

2012

2013

2014

US

Source: Wall Street Journal, Hermes as at October 2017

Machiavelli, Niccolo. The Prince, translated by N.H. Thomson. Vol. XXXVI, Part 1. The Harvard Classics. New York: P.F. Collier & Son, 1909–14; Bartleby.com, 2001.

2015

2016

2017

HERMES INVESTMENT OFFICE

Our concerns about recovery in the next downturns remain starkly illustrated in the sharp, upward curve of the debt data graph above. Elsewhere, the Federal Reserve’s September 2017 ‘Survey of Consumer Finances’ found that the wealth share of the bottom 90% fell from 33.2% in 1989 to 22.8% in 2016, while the richest 1% increased their proportionate share of the pie from just below 30% to 38.6%.

The unnatural calm in the face of real-world events could, perhaps, be a classic pre-storm state. We measure this tendency to over-relax via our bespoke Complacency Indicator, which compares current volatility measures to historical trends. Under our modeling we can see that markets often track through long periods of calm broken by violent outbursts of volatility during a crisis.

The Fed data clearly shows benefits from post-crisis unconventional monetary policy have largely accrued to a very narrow segment of the population – a wealth disparity that increases the risk of rising populism in the US and beyond.

The two important characteristics of volatility are ‘long memory’ and ‘jumps’: the latter relates to the propensity of volatility to spike, while the former suggests that deflation of the spikes is far slower than the initial jump.

With such a range of asymptomatic risks to consider we need more than simple diagnostic tools to discern the underlying condition of the broader financial system.

One way to track this is by comparing volatility high points to the sum of the volatilities for the days involved in the jump. Using this technique, the higher the read-out on the Complacency Indicator, the more ‘ready-for-action’ markets appear to be.

While the data is US-centric, the St Louis Fed metric provides a decent proxy for global financial ‘wellness’, considering the still-critical influence of the world’s biggest economy on the rest of the world. Theoretically, the underlying index variables should move together as levels of financial pressure in the economy change. As per the chart below, index values above zero indicate above-average financial stress, and vice-versa: Figure 2: Financial stress

120 100 80 60 40 20 0

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

As explained in our previous issue, the St Louis Fed stress benchmark tots up 18 different financial variables (seven related to interest rates, six to yield spreads, and five others) into a single, weeklyupdated index.

Figure 3: Complacency

VIX & Complacency indicator

Last quarter, for example, we added the St Louis Fed Financial Stress Index to our bag of indicators, which we once more refer to for potential early warning signs of declining economic health.

LT average

VIX

LT average

Complacency indicator

Source: Hermes, Bloomberg, CBOE as at October 2017

6

The chart above covers three major jittery market periods: the 1998 Long Term Capital Management/Russian default crisis; the so-called ‘dotcom’ collapse in 2001; and, the three or four years from just before the GFC until 2011. While we refer to equity markets (via the VIX) in this analysis, the trend holds equally for other asset classes.

5 4 3 2 1 0 -1 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

-2

Our Complacency Indicator remains at worrying lows, clearly suggesting the market is poorly prepared for a change in mood. As we have noted before, complacency does not guarantee negative outcomes, but we believe that investors should continue to be cautious and prepared for any eventuality.

Source: Federal Reserve Bank of St. Louis, Hermes as at October 2017

The year may be almost at an end but 2017 still has the capacity to shock. The shift from one monetary regime to another is unlikely to progress without accompanying convulsions in the market.

If the St Louis Fed index tells the whole story then investors appear to have entered an almost Zen Buddhist realm of bliss. The stress measure has sustained the close-to historically low readings we saw last quarter.

In our consistently-held view, the current calm – reflected in low volatility and generally quiescent conditions – may hide deeper market vulnerabilities that could catch complacent investors and fund managers unawares.

And yet the global backdrop has featured intensifying political unrest (most recently manifested in the independence stand-off between Catalonia and the Madrid-based central Spanish government) along with further signals from central bankers that the easy-money days are coming to an end.

To remain ahead of the risk curve we measure forward-looking ex-ante risk from as many angles as possible. Our models, built on the everincreasing piles of patterned data produced by the apparent chaos of daily market moves, help us discern likely hazards ahead.

St Louis Fed Financial NBER recession indicator

But we are risk-aware, not risk-averse.

3

4


Our forward-looking, evidence-based approach offers a unique multifaceted guide to generating returns through informed risk-taking – albeit under the proviso that any map of the future inevitably includes some projection into unknown terrain. We have long recommended that investors consider the full gamut of risks beyond pure financial market metrics. Our new ESG measure is a worthy addition to the adaptive techniques necessary for investors to survive in a risky world, which we will track in this issue as below: 1. Volatility 2. Correlation risk 3. Stretch risk 4. Liquidity risk 5. Event risk 6. ESG risk

Figure 4: Moving averages of selected volatility measures 4 3 2 1 0 -1 -2 -3 2008 VIX

Summary Key risks highlighted in this report: Geopolitical uncertainty, in its various manifestations, continues

to weigh most heavily in 2017

How will the markets react to unwinding Unconventional

Monetary Policy (UMP)

Correlations point to choppier waters ahead Liquidity risk still remains the most likely source for the

propagation of contagion

VOLATILITY: WHY MARKETS MAY GET THE UMP Typically the go-to measure of financial risk, volatility has largely been missing from markets for years. In our view, the unnaturally-subdued range of asset price swings is due to widen as economic, political and leverage concerns eventually weigh-in on markets. Regardless of the current calm, we must scan for the early signs of volatility across multiple asset classes and geographies using the best tools available. As per previous issues, we present updated data across the complete set of volatility measures at our disposal with at least two – the VIX term structure and the dispersion indicator – bumping up the dial a little. Firstly, we turn to Figure 4 that shows the 52-week moving average of the VIX, the Merrill Option Volatility Expectations (MOVE) Index, the Deutsche Bank FX Volatility (Currency VIX) Index and the expected volatility of the Bloomberg Commodity Index (Commodity VIX). They measure the implied volatility of equity markets, government bond markets, currency markets and commodity markets respectively, and have been standardised to make them directly comparable. Furthermore, each of them represent the market’s expectation of future volatility, and are often viewed as a benchmark of risk appetite.

2009

2010 MOVE

2011

2012

2013

Currency VIX

2014

2015

2016

2017

Commodity VIX

Source: Hermes, Bloomberg, CBOE, Deutsche Bank, Bank of America Merrill Lynch as at October 2017

Across the board, long-term implied volatility measures continued to drop for all asset classes over the quarter, with currency and commodity volatility dropping most rapidly. Equity and bond volatility eased off at a less precipitous rate. This pattern is roughly consistent with the daily raw (un-averaged) data. Despite the apparent easing of risk concerns, we believe such low volatility conditions are unlikely to be sustained for the rest of the year – and probably obscure other emerging dangers.

FRASER LUNDIE, CO‑HEAD OF CREDIT

CREDIT CARE: WHY VOLATILE MEMORIES WILL PAY OFF The global hunt for yield, particularly of the short-dated variety, has seen dramatic demand for bank loans of late. Indeed, bond and loan markets are converging as the former experiences ‘protection erosion’ while the latter adopts ‘covenant light’ as standard operating procedure. This trend is a boon for companies and sponsors alike, as optionality and pricing power are with them. However, the rapidly-evolving debt markets leave investors requiring significantly more flexible, global approaches in order to achieve the risk-adjusted returns of the past. That said, and despite this year’s rally, we see plenty of opportunities in 2018 where default rates remain very low, aided by balance sheet health, earnings momentum, and a robust technical backdrop – all of which continues to be underpinned by central bank liquidity. But dynamic, nimble management and the sizing of positions in the riskier parts of the market will be necessary. Long memory will be a bonus for experienced investors who can recall a time before the current unprecedented run of low volatility – and position portfolios with due care.

The Hermes Investment Office Independent of the investment teams, the Hermes Investment Office continuously monitors risk across client portfolio and ensures that teams are performing in the best interest of investors. It provides rigorous analyses and attributions of performance and risk, demonstrating our commitment to being a transparent and responsible asset manager


The VVIX, a risk-neutral forecast of large-cap US equity index volatility, has been designed to flash early warning signals of potential market ruction ahead.

16 14

10

Index

Jun 18

May 18

Apr 18

Mar 18

Feb 18

Term structure of VIX futures contracts at 12 July 2017 Term structure of VIX futures contracts at 11 October 2017 Source: Hermes. Reuters, CBOE

It is worth acknowledging that the liquidity in the more distant contracts will not yet be great. Nonetheless, we can safely conclude that the market appears to be anticipating some uptick in volatility (note the steepening of the curve for early 2018), and is consistent in expecting it to rise as we move through time.

12

8 6 4 2

VVIX

10.0

This quarter we again offer the full term structure based on all available contracts out to midway through next year, and note how the curve has changed.

Figure 5: The volatility of volatility

0 2009

12.5

Jan 18

As wider macro issues come into play and central bank liquidity eases we expect volatility to shift. We urge investors to remain cautious of leveraging positions too far as the year comes to a close.

15.0

Dec 17

Not only do we expect volatility to increase during the remainder of 2017, we also anticipate that markets will retain the effects of volatility shocks more deeply in their collective conscience (long memory will return).

17.5

Nov 17

Given that we anticipate further shocks with sharp surges in volatility, those same investors will be forced to cut their positions, leading to self-reinforcing position-shedding.

Figure 6: Term structure of volatility

Futures price

Most importantly, sustained low implied volatility can lull investors into gearing up portfolios – or via other techniques to ramp up exposures – in order to supercharge returns.

2010

2011

2012

2013

2014

2015

2016

2017

Moving average

Source: Hermes, Bloomberg, CBOE as at October 2017

We also like to look at cross-sectional dispersion as a volatility measure. It can be thought of as a measure of the various opportunities available for stock-pickers in equity markets, reflecting the best-to-worst range at particular time periods – we track the measure through time. Other things being equal, elevated levels of cross-sectional volatility speak to higher return dispersion across assets. Figure 7: Cross-sectional dispersion of stock returns 20

As the chart above shows, forward-looking volatility expectations spiked during August on the back of a minor disturbance in markets, but fell back towards longer-term lows by the end of the quarter.

We also test market states, and develop investment strategies, by investigating the term structure of volatility – via VIX futures contract data. Comparing the price levels of different futures expirations provides a good view of VIX term structure. Since volatility is a measure of systematic risk, in that sense the VIX term structure suggests the trend of development of future market risk. If the VIX is upward-sloping, it implies that investors expect to see the volatility (risk) of the market going up in the future. If the VIX is downward-sloping, it indicates that investors expect to see the volatility of the market falling in the future.

16 14 12

%

However, with ongoing global political unrest on the rise and the likelydisruptive effects on financial markets of UMP to consider, we expect forward-looking gauges such as the VVIX will flicker upwards over the rest of 2017.

18

10 8 6 4 2 0 2007

2008

2009

2010

Cross-sectional volatility

2011

2012

2013

2014

2015

2016

2017

Cross-sectional volatility (moving average)

Source: Hermes, Bloomberg, FTSE as at October 2017

Dispersion picked up a little this past quarter in Europe, but still remains low, far lower than would be helpful to active managers. In the US it is close to its 25th percentile, while in the UK and China it nudged above median.

5


We expect the measure to edge higher as markets wake up to the consequences of unwinding UMP. Stock-pickers everywhere will be hoping for just such an improvement to the cross-sectional volatility picture, and we feel that there should be sufficient uncertainty in the remainder of this year to deliver on that.

Taking the end of June 2017 and the end of September heat maps of cross-asset correlations, not a great deal has changed over the quarter.

To summarise, the majority of our volatility measures suggest that at present we reside in a low-risk environment, with dispersion on the upswing. However, the tranquil surface could hide emerging disruptive forces – including economic, political and financial leverage risks – that may cause a splash before year-end.

The Morgan Stanley Global Correlation Index below measures correlations across asset classes, geographies, sectors and factors, and intra-markets – it plainly fell late last year and has remained at levels more commonly last seen pre-GFC.

CORRELATION RISK: PREPARE TO BE SURPRISED Correlation assumptions underpin most asset allocation and investment strategies. Indeed, investors depend on the assumed relationship between asset classes to manage overall portfolio volatility risk.

If volatilities rise in the final quarter as we expect, the prevailing correlation paradigm could well be challenged.

Cross-asset global correlation has remained stubbornly in a band from 2007, with assets, asset classes and geographies largely moving together. There is still a strong hint that this background has given way to one of lower correlations all round despite a modest rise in the latest quarter. The recent rise in equity/bond correlation has increased focus on the potential deleveraging pressure from risk parity funds and other volatility-targeting strategies.

Correlations: March 2017 Unfortunately, correlation is not a fixed quality, behooving us to closely Global HY US NonFin HY Constrained monitor the subtly-shifting dynamics of asset class relationships over MSCI EM EU N-FinaFixed&Float HYC time and to clearly define what we mean by MSCI theEUROPE concept. MSCI NORTH AMERICA

TOMMASO MANCUSO, HEAD OF MULTI-ASSET

BBG Agriculture BBG Industrial Metals BBG Energy MSCI JAPAN Australia Govt Bonds Generic Germany Generic Govt 10Y Euro Generic Govt Bond 10Y BALTIC DRY INDEX Japan 10 YEAR JGB FLOAT BBG Livestock Global Broad Market BBG Precious Metals

For instance, two variables with the same long-term trend could have a negative, short-term correlation coefficient, over-emphasising the level of diversification available between them.

UPSIDE SURPRISE: WHY GOOD TIMES MIGHT BE BAD FOR BONDS

Information regarding the long-term trend should be taken into consideration when assessing diversification. Given that correlation is typically measured with respect to mean values, we must also account for sample trend in our analysis below: BBG.Precious.Metals Global.Broad.Market BBG.Livestock Japan.10.YEAR.JGB.FLOATING.RA BALTIC.DRY.INDEX Euro.Generic.Govt.Bond.10Y.Yield Germany.Generic.Govt.10Y.Yield Australia.Govt.Bonds.Generic.Y MSCI.JAPAN BBG.Energy BBG.Industrial.Metals BBG.Agriculture MSCI.NORTH.AMERICA MSCI.EUROPE Eu.N.FinaFixedFloat.HYC MSCI.EM US.NonFin.HY.Constrained Global.HY

While investors are focusing, quite understandably, on the implications of a market correction – or worse – the alternative scenario also holds significant risks. In a scenario wherein growth and inflation surprise on the upside, we would most likely see a market regime shift.

Figure 8: Correlation heat maps

Correlations: Correlations:June June2017 2017

Since 1997, the multi-decade stability of the stock-bond relationship has led to the gradual adoption of the flight-toquality narrative (i.e. stock down, bonds up), and over time, such investment narratives can be mistaken for absolute truths.

Euro.Generic..Govt.Bond.10Y.Yield Euro.Generic..Govt.Bond.10Y.Yield Germany.Generic.Govt.10Y.Yield Germany.Generic.Govt.10Y.Yield BALTIC.DRY.INDEX BALTIC.DRY.INDEX Japan.10.YEAR.JGB.FLOATING.RA Japan.10.YEAR.JGB.FLOATING.RA BBG.Livestock BBG.Livestock Australia.Govt.Bonds.Generic.Y Australia.Govt.Bonds.Generic.Y MSCI.JAPAN MSCI.JAPAN Global.Broad.Market Global.Broad.Market BBG.Precious.Metals BBG.Precious.Metals BBG.Energy BBG.Energy BBG.Industrial.Metals BBG.Industrial.Metals BBG.Agriculture BBG.Agriculture MSCI.EUROPE MSCI.EUROPE MSCI.NORTH.AMERICA MSCI.NORTH.AMERICA Eu.N.FinaFixedFloat.HYC Eu.N.FinaFixedFloat.HYC MSCI.EM MSCI.EM US.NonFin.HY.Constrained US.NonFin.HY.Constrained Global.HY Global.HY

Global HYHY Global USUS NonFin HYHY Constrained NonFin Constrained MSCI EMEM MSCI EUEU N-FinaFixed&Float HYC N-FinaFixed&Float HYC MSCI NORTH AMERICA MSCI NORTH AMERICA MSCI EUROPE MSCI EUROPE BBG Agriculture BBG Agriculture BBG Industrial Metals BBG Industrial Metals BBG Energy BBG Energy BBG Precious Metals BBG Precious Metals Global Broad Market Global Broad Market MSCI JAPAN MSCI JAPAN Australia Govt Bonds Generic Australia Govt Bonds Generic BBG Livestock BBG Livestock Japan 10 10 YEAR JGB FLOAT Japan YEAR JGB FLOAT BALTIC DRY INDEX BALTIC DRY INDEX Germany Generic Govt 10Y Germany Generic Govt 10Y Euro Generic Govt Bond 10Y Euro Generic Govt Bond 10Y

However, if stocks look expensive then bonds look even more so (see earning yields vs bond yields below). In this environment, even under a moderate reflationary scenario bonds would not only disappoint from a return perspective but also lose their much-vaunted hedging quality as the stockbond correlation would most likely turn positive.

Figure 9: Global Correlation Index 60

Source: Hermes, Bloomberg as at October 2017

50

Index

40 30 20

Global Correlation Index Source: Morgan Stanley, Bloomberg, Hermes as at October 2017

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2005

2006

0

2004

10

2003

MSCI.NORTH.AMERICA

MSCI NORTH AMERICA MSCI EUROPE BBG Agriculture BBG Energy BBG Industrial Metals BBG Livestock Japan 10 YEAR JGB FLOAT MSCI JAPAN Australia Govt Bonds Generic Germany Generic Govt 10Y Euro Generic Govt Bond 10Y US NonFin HY Constrained EU N-FinaFixed&Float HYC MSCI EM Global HY BBG Precious Metals Global Broad Market BALTIC DRY INDEX

BBG.Energy BBG.Agriculture MSCI.EUROPE

BBG.Industrial.Metals

Japan.10.YEAR.JGB.FLOATING.RA BBG.Livestock

Eu.N.FinaFixedFloat.HYC US.NonFin.HY.Constrained Euro.Generic.Govt.Bond.10Y.Yield Germany.Generic.Govt.10Y.Yield Australia.Govt.Bonds.Generic.Y MSCI.JAPAN

Correlations: September 2017

BALTIC.DRY.INDEX Global.Broad.Market BBG.Precious.Metals Global.HY MSCI.EM

6


We further delve into the correlation analysis by examining how the current statistical measures stack up relative to history. This ‘correlation surprise’ metric – as with any statistical-based tool – is not an absolute gauge.

Figure 12: Correlation signal 1.0

Correlation

0.00

0.2

80

0.0 -0.2

60

-0.4

40

-0.6

-0.05

20

-0.8

-0.10

-1.0 2004

-0.15

Correlation

-0.20

Correlation signal

100

0.4

Figure 10: Correlation surprise and returns Subsequent one-month annualised return

120

0.6

But nonetheless, we can see that spikes in correlation surprise are frequently followed by periods of disappointing returns.

2006

2008

2010

2012

2014

2016

2017

0

Correlation signal


-0.25 -0.30

Russell 3000

MSCI Emerging Asia

MSCI Emerging Markets

MSCI Europe

MSCI China

Subsequent one-month annualised return

700000 600000

Over the last few decades the range of investment options has multiplied across asset type and geography. While the expanding investment universe has brought the benefits of diversification, portfolio construction and risk management, techniques have become increasingly homogenous.

500000 400000 300000 200000 100000 0 2003 2005 2007 Correlation Surprise Index

During the last quarter our correlation stability measure stayed flat in contrast to the previous period. Notably, our signal has been generally more unstable post-GFC, indicating greater caution should be used when looking at correlations. We expect instability to pick up again in the remainder of 2017. In our view, traditional methods of portfolio diversification that rely principally upon historical measures of correlation have become less effective. Furthermore, the potential for regime change in cross-asset correlations remains at the elevated levels typical of recent times.

Figure 11: Correlation surprise in the global equity universe

Index

140

0.8

2009

2011

2013

2015

2017


Our correlation surprise indicator remained high throughout the quarter with a flurry of activity mid-August, suggesting highly unusual behaviour. We will keep a close eye on this critical signal. From time-to-time the correlation surprise metric will give false-positive signals, but we think it is wise to remain cautious about any portfolio assumptions with respect to cross-asset relationships. Indeed, the variation in correlation levels of assets or asset classes is one of the biggest conundrums for investment managers. Assets that at one time appear to be uncorrelated often become highly correlated during periods of market stress: conversely, those that are highly correlated may de-couple at a later time. This instability of correlation is further aggravated by time-dependency in the volatility of the correlation coefficient. At times, correlations appear to fluctuate within a tight range, at others, we see fluctuations in the sign of correlation over very short time periods.

As their strategies conform investors tend to react in a similar fashion when faced with the same new information, increasing the gap risk in correlations. We anticipate sharp movements in correlation, which investors will need to handle by combining different portfolio construction and risk management methods.

STRETCH RISK: WHY INVESTORS CAN’T BE SURE OF THE FLOOR As discussed above, volatility provides the most obvious, and initial, glimpse of rising market risks. But the tight focus on price movements alone can fool investors into missing other market risk insights. Our ‘stretch risk’ analysis – which we apply across various asset classes – is a good way to diversify away from the volatility-heavy crowd. Stretch risk allows us to identify assets that trend in one direction for a considerable period of time. A ‘stretched’ asset typically features suppressed headline volatility that obscures true underlying risks. For instance, very expensive assets that often have very low volatility, and despite sizable downside tail risk, are deemed perfectly safe by traditional risk models. In the past we have illustrated stretch risk with reference to different parts of the credit market – this quarter we focus on the European high yield market. The hunt for yield in Europe continues apace with the index at all-time lows (see chart below), suggesting a degree of complacency in credit markets.

7


Overall, we can find excellent examples of stretch risk in both equity and bond markets, either from a momentum or extreme valuation perspective.

Figure 13: Stretch risk – high yield in Europe 30 25

The injection of liquidity from UMP has led to an unstable floor for downside risk, which we see continuing to develop in unpredictable ways throughout the remainder of this year.

20 15

LIQUIDITY RISK: CONTAINED BUT UNDER PRESSURE

10

2017

2015

2016

2013

2014

2011

2012

2010

2009

2007

2008

2005

2006

2003

2004

2001

2002

1999

2000

1997

0

1998

5

BofA Merrill Lynch Euro High Yield Index Semi-Annual Yield to Worst (% Daily, Not Seasonally Adjusted)

We have consistently argued that investors must closely scrutinise the link between market between market risk, funding and monetary liquidity.

Source: Hermes, Reuters as at October 2017

Under those definitions, funding refers to the ease of borrowing, whereas monetary liquidity reflects the ease of monetary conditions. They influence market liquidity, through market-making activity and bank funding respectively.

Meanwhile across the Atlantic, the mature stage of the US business cycle may not bode well for corporate bonds if fiscal policy (a great deal of hope is pinned on far-from-certain tax cuts) fails to re-energise the economy.

And, importantly, liquidity is often the first casualty of any market dislocation.

US corporate debt shows a far higher beta to equities on the downside than on the upside. With the Fed beginning to unwind its balance sheet (albeit at a slow pace) in the US, and the ECB leaning in the same direction, credit returns may begin to look less-distorted relative to the same stage in previous cycles. Junk bond spreads in the US also remain close to cyclical lows, in spite of known issues in the energy and retail sectors. Even under a very slow and measured version of quantitative tightening (QT), though, we would anticipate spreads moving wider.

Today, the ‘TED spread’ and the ‘Credit spread’ are the two most closely-followed metrics for funding and liquidity risk. The TED spread focuses on the difference between interest rates available in the interbank market and on short-term US government debt (T-Bills), typically at a one- or three-month view. Whereas the Credit spread generally targets the spread between corporate bonds and government bonds, again at a comparably short maturity – essentially an indicator of perceived credit risk, linked closely to the potential for default in the corporate bond market. Figure 15: Funding and credit risk

Figure 14: Stretch risk – US corporate debt

10

50%

3000

48%

2500

6

2000

4

44% 1500 42%

2 0

1000

US non-financial corporate debt as % GDP Source: Reuters, Hermes as at October 2017

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2002

2003

2001

2000

1999

40% 38%

8

%

46%

1998

8

500

S&P500

We can combine total US corporate bonds outstanding with collective loans outstanding to give a picture of all US non-financial debt ranked against GDP. The current debt-to-GDP reading of more than 45% in this measure has historically pre-figured significant equity market corrections. Valuations can also become very stretched without the appearance of increased volatility. In this scenario, assets or markets become extremely cheap or expensive through continual small price movements – one of the reasons that equity volatility is so low at the moment. However, such valuations rarely persist, with the relevant asset or market likely to snap back to fair(er) value.

-2 2007 2008 2009 2010 TED spread

2011

2012

2013 2014

2015

2016

2017

Credit spread


As the graph above reveals, Credit and TED spreads have remained at subdued levels for a considerable period of time with the latter shaking off a brief rise in the autumn of last year. The rise in the LIBOR, which we noted earlier this year continued beyond the impact expected from structural reforms, has now stabilised. We continue to monitor the LIBOR closely, suspecting it could provide an indication of a deeper problem, symptomatic of broader issues in the offshore dollar market in particular.


By identifying ‘crowded’ trades, we are better able to identify potential trigger sources of liquidity risk. To that end, the monthly survey conducted by Bank of America Merrill Lynch of global fund managers provides some clues.

Figure 17: The Hui and Heubel ratio for Bund futures 0.10 0.09 0.08 0.07 HHL ratio

Dislocations can occur even in highly-liquid markets. The liquidity of an asset often depends on whether the direction in which you wish to trade matches the broader market sentiment. We also must consider the quantity of the asset to be traded as another variable that influences liquidity.

0.06 0.05 0.04 0.03 0.02

Figure 16: Fund managers answer the question: where do you think the most crowded trades currently are?

0.01 Jul 16

Jan 17

Jul 15

20-period moving average for the euro Bund


The ratio shows liquidity conditions remained relatively stable during the quarter, with no more than the usual number of illiquidity spikes.

Figure 18: Kyle’s lambda 30% 25% 20% 15% 10%

Effect of a trade equalling 2% of average daily volume Source: Hermes, Bloomberg as at October 2017

“FX Swaps and forwards: missing global debt,” (by Borio, Claudio et al)published by Quarterly Review as at September 2017.

29 Sep 17

11 Jun 15

0%

23 Oct 16

5% 27 Jan 14

As in our previous Market Risk Insights we include once again the Hui and Heubel ratio for Bund futures: the ratio measures intra-day price movement relative to the ratio of traded volume to either market capitalisation or open interest.

By estimating the basis point impact of a trade of 2% of the average daily volume traded, we can compare the cost of accessing liquidity at any point in time in our data history, irrespective of changes in underlying market structure.

14 Sep 12

We believe that concerns over liquidity risk in the corporate debt market remain highly relevant. Alongside our credit portfolio managers we are keeping a close eye on these dials.

Kyle’s lambda is a measure of liquidity resilience in the sense that it captures how much equity prices move with order flow, or effectively, price impact. In that sense it is very similar to the Amihud illiquidity metric.

26 Aug 10

Positions in some US equity sectors remain hot, principally banks and Nasdaq technology names. European equities, meanwhile, remain in favour with institutional portfolios as corporate bond positions stay broadly stable.

This quarter we turn again to, Kyle’s lambda, which seeks to throw further light on liquidity conditions in equity markets.

13 Apr 09

We have seen some further rotation in fund managers’ most-crowded trades during the latest quarter. Reflecting the crypto-currency craze ‘long bitcoin’ has made its first appearance in these charts while most popular US dollar trades have switched from long to short positions.

Liquidity, as we have always maintained, will be the most likely transmission mechanism for contagion should any significant shocks upset the current market détente.

30 Nov 07

40%

18 Jul 06

30%

5 Mar 05

20%

22 Oct 03

10%

9 Jun 02

0%

Jul 17 Aug 17 Sep 17 Source: Hermes, Bank of America Merrill Lynch as at October 2017

25 Jan 01

Other

13 Sep 99

Long Bitcoin

However, it remains moot whether years of UMP (low rates together with quantitative easing) have eroded the ability of central banks to contend with downturns. If so, should liquidity dry up confidence could likewise evaporate abruptly, with bond markets in particular set to suffer.

1 May 98

Short RMB

5 Aug 95

Long Banks

17 Dec 96

Long EM equities

2% ADV Monthly Average Impact

Long Nasdaq

3

Jan 16

Jul 14

Jan 15

Jan 14

Jul 13

Jan 13

Jul 12

Jan 12

Jul 11

Euro Bund

Short Govt Bonds Long US/EU Corp Bonds Long Eurozone equities Long EM debt & FX

-10%

Jul 10

Short USD Long US Small Cap

Jan 11

Jan 10

0.00

9

Our model – as per the graph above – picks up some of the major market dislocations in recent decades. The good news is that the market impact for global equities tightened only modestly for the third quarter of 2017.

Figure 19: Turbulence index – future returns

On that basis, our liquidity concerns are limited to the bond and offshore dollars markets – for the time being. But liquidity issues will likely prove influential in amplifying any local shocks into global concerns.

Turbulence index


We attack the problem by using non-standard models, which give investors a better understanding of possible outcomes stems by stresstesting portfolios and detailed scenario analysis. These represent the minimum standard for risk management in today’s investment world. If we can successfully identify periods in advance in which asset prices behave uncharacteristically – as seen during the GFC turbulence – then we may be able to minimise portfolio drawdowns by adjusting portfolios appropriately in advance.

MSCI Emerging Asia

MSCI Europe

MSCI China

Full sample annualised return Annualised return following most turbulent period Annualised return following most non-turbulent period Source: Hermes, Bloomberg

We analyse market turbulence by identifying the statistical unusualness of the current risk environment, in terms of both volatility and correlation. Our analysis demonstrates that most turbulent periods typically precede significant drawdowns across a number of asset classes and markets. Times of financial turbulence are typically persistent and provide lower rewards for risk-bearing than normal times. As such, this measure could be used to construct portfolios that would be relatively resilient to turbulence via a conditioning process. Figure 20: Turbulence index – global equities 30 25 20 15 10 5

2017

2016

2015

2014

2012

2013

2011

2010

0 2009

Naturally, event risk proves largely immune to discovery by financial statistics alone.

MSCI Emerging Markets

2008

But real-world risks inevitably influence how investors view market opportunities, which requires us to create some framework to measure those effects.

Russell 3000

2007

Like the North Korean missile tests, investors have watched most potentially explosive global events – including Brexit and the election of Donald Trump – splash apparently without harm in the ocean of market resilience.

-1.0

2006

Markets have almost casually shrugged off worries following the constant bouts of political upheaval the world has witnessed over the last couple of years.

-0.6 -0.8

2005

EVENT RISK: MOOD MODERATES DESPITE THE RADICALS

-0.2 -0.4

2004

Our analysis suggests the decline in liquidity is structural and permanent: getting paid for illiquidity may seem attractive but it must be balanced against the need to absorb a tidal wave of selling should that recur.

0.2 0.0

2002

Rather alarmingly, the BIS found non-banks outside the US owe large sums of dollars off-balance sheet through these types of FX derivatives. Although much of the US dollar debt will be purposed for FX hedging, the potential for maturity mismatch is very real: a particularly daunting scenario considering the BIS report estimates the off-balance sheet figure exceeds the $10.7tn of on-balance sheet dollar debt.

0.6 0.4

2003

For example, the Bank for International Settlements (BIS) recently posed the question: “What would balance sheets look like if the borrowing through FX swaps and forwards were recorded on-balance sheet, as the functionally equivalent repo debt is?”3

0.8

Turbulence Index

10

Turbulence Source: Hermes, Bloomberg, MSCI as at October 2017

Our Turbulence Index picked up the mid-August move that was only briefly sustained. The uptick in the Turbulence Index implies that markets were behaving less normally relative to their own recent past. The measure ended the quarter at low levels again. The final tool that allows us to estimate market fragility is the Absorption Ratio, which captures the market’s ability to cushion shocks: perhaps best thought of as a measure of systemic risk. We use principal components analysis to determine the extent to which the largest risk factors dominate the entire risk factor set. When markets are particularly vulnerable to shocks, a handful of factors will explain the vast majority of risk, whereas when markets are less fragile, we see the Absorption Ratio fall.


Our global policy uncertainty indicator reached a 16-year peak around Brexit, only to be surpassed by a new high around the time of the US election. The metric has declined substantially since those levels, as the prevailing view appears to be that markets can remain immune to political uncertainty but not so to its economic cousin.

Figure 21: Absorption Ratio 0.85 0.80

Index

0.75

Despite the sanguine market take on geo-politics, there are flashpoints aplenty – notably North Korea and Brexit – with populism also back high on the agenda in Europe in Austria, Germany and Spain (with the latter spiraling towards an internal crisis as the central government moves to impose direct rule on Catalonia). We expect policy uncertainty to remain at modest levels throughout the remainder of 2017.

0.70 0.65 0.60 0.55 2015

2016

2013

2014

2011

2012

2010

2009

2007

2008

2005

2006

2003

2004

2001

2002

1999

2000

0.50

Absorption Ratio Source: Hermes, Bloomberg, MSCI as at October 2017

We continue with our expanded multi-asset Absorption Ratio, capturing information for 17 different asset classes. The metric was at relatively subdued levels throughout 2016, but rose noticeably during the first quarter of the year. It is clearly picking up some cross-asset vulnerability, which is unsurprising given the movements between bonds and equities late last year. We would expect to see further upward movement in this measure as the year progresses. Political risk and its impact on markets should never be ignored, even if it is difficult to measure quantitatively. One way to tackle this problem is with a metric based upon the frequency of economic policy uncertainty coverage. Using an unstructured data approach, policy uncertainty appears to be a leading indicator of increases in equity market volatility. At a macro level, increased policy uncertainty tends to foreshadow declines in economic growth and in employment. More explicitly, the policy uncertainty indicator chimes with the average correlation of stocks, in the sense that equities tend to be driven by a single macro factor when policy uncertainty is high. Figure 22: Economic Policy Uncertainty 350 300

Index

250 200 150 100 50

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

0

Source: Economic Policy Uncertainty, Hermes IM as at October 2017

4

Regardless of the multi-jurisdictional pop-up political crises, investors remain buoyed by slowly improving macro statistics suggesting a global economy on the mend. However, event risk, incorporating political and policy uncertainty, is a constant feature of financial markets. Our principal metrics for capturing it, the Turbulence Index and the Absorption Ratio, are at moderate levels and broadly in agreement.

ESG RISK: TIPPING POINT FOR INCLUSION Over the last few years, ESG factors for investing have evolved from a rudimentary fringe concern to a sophisticated component of mainstream risk analysis. As aforementioned, while ESG has long been an essential part of our research process we have – for several reasons – held off including it formally in this publication. Until now. We believe most investors today view ESG as an essential item for consideration, not an optional feel-good add-on. But ESG is a large – and growing – field of risk management with everimproving data flows providing new insights for investors. It is beyond our scope to cover the whole ESG universe in this quarterly publication. Instead, we will highlight ESG issues that crop up during the quarter in question with an emphasis on how they may affect future investments – beginning in this issue with the perennial headlinegrabber of climate change. Climate change itself is, of course, too big to contain here in full but we note below just a handful of relevant research pieces that have recently crossed our desks. Firstly, a new study in the Proceedings of the National Academy of Sciences4 shows that each degree of global warming will reduce global harvests by about 3% to 7%, depending on the type of crop. The study says US agriculture is particularly vulnerable, with corn the most sensitive, showing a potential 7.4% production decline for each degree celsius of warming. Recently, we were also fortunate enough to hear Professor Tim Lenton from Exeter University speaking at a Thinking Ahead Institute symposium. Professor Lenton presented his work on tipping points, which struck a real chord.

“Temperature increase reduces global yields of major crops in four independent estimates,” published by PNAS as at July 2017

11

12


Figure 24: Potential tipping elements in the climate system

Source: PNAS, Hermes as at October 2017

A ‘tipping point’ refers to a critical threshold at which a tiny change can significantly alter the state or development path of a system. The study’s authors used expert survey information to look at the sensitivity of likely climate change impacts. That data in turn can suggest some early alert techniques to detect proximity to tipping points – providing a level of ESG risk information that investors can incorporate into their decision-making processes. A further study that grabbed our attention examines efforts by firms to meet ambitious carbon emissions targets5. With a novel dataset compiled by the Carbon Disclosure Project (CDP), the authors found that while using stretch goals alone was effective, the addition of monetary incentives/bonuses seemed on the whole to undermine the achievement of those targets. Studies like these give us strong clues about how to analyse the relative riskiness of company strategies to address climate change through corporate behaviour modification. We would encourage all investors to give the research full consideration. ESG analysis offers investors another, potentially lucrative, method of managing portfolio risk. By introducing ESG into our risk mix we aim to help investors adapt quality research insights to practical portfolio solutions.

5

CONCLUSION: TIME FOR A CORRECTION? The whack-a-mole nature of risk keeps investors guessing as to where, and when, it might appear next. Historical pattern analysis undoubtedly provides us with some hints about the current, or imminent, manifestation of market risk – but statistical sifting does not yield up the whole picture. We believe investors are best-served by considering risk from multiple angles using both qualitative and quantitative tools. At a macro level, more hawkish noises from central banks have failed as yet to derail markets. It is still not clear whether central bank actions are driven by a concern that another risky asset bubble has been inflated, or fears they could lose control of the economy should inflation rise on the back of modest economic growth. History suggests that economic data surprises have far greater capacity to spook markets than geopolitical risks – central banks will need to tread carefully as they attempt to ‘normalise’. In this environment, we would encourage investors to shy away from any complacency implied by benign risk measures. Consumer survey data comes with a slight health warning, but it is interesting to test the water in terms of obtaining the views of the retail investor in the US.

“The Effect of Target Difficulty and Incentives on Target Completion: The Case of Reducing Carbon Emissions,” published by the Accounting Review as at May 2016.


Figure 25: Consumer surveys 70

240

65 60

190

In this quarter, many of our risk indicators remain subdued but with enough movement on some dials to reinforce the need for caution.

140

Nonetheless, the market still offers opportunities for suitably skilled and risk-aware active managers.

55 50 45 40

90

35 30

40

25 20

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

-10

Probability of an increase in stock market price in the next year Current value of stock market investments Source: University of Michigan, Hermes IM as at October 2017

The survey finds US households feeling very positive about the stock market in the year ahead. In fact, the data seems to indicate US investors believe that there has never been a better time in the last decade or more to buy shares than now. Somewhat counter-intuitively, a majority of respondents also think equity markets are over-valued. Central banks inform policy decisions by tapping into data like this in the hope of discerning a more accurate probability of certain events occurring – like market crashes. Unsurprisingly, these sorts of secondguessing exercises – based, after all, on subjective consumer assessments – carry certain caveats. For all that, the graph below from the Minneapolis Federal Reserve, based on options prices, appears to support the current benign market consensus. Figure 26: Implied probability of a crash 0.40 0.35

Probability in (%)

0.30 0.25 0.20 0.15 0.10 0.05 0.00 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Percent chance of a -20% or greater fall in the market

Source: Minneapolis Federal Reserve, Hermes as at October 2017

According to the Minneapolis Federal Reserve, the probability of -20% or greater decline in US equity prices over the next year implied by the options market is comfortably less than 10% and trending down. Now even the Minneapolis Fed will probably admit the figure is a rough approximation rather than representing the true likelihood of a substantial correction.

6

However, when we stack that result against the third most expansive equity market run in living memory, a less than 10% probability of a major market fall within the next 12 months looks extraordinarily optimistic.

“The Great Liquidity Crisis’ investor note,” published by JP Morgan as at September 2017.

If we are indeed headed for a correction of some sort, then understanding your investment timeframe is paramount: depending on how your portfolio is adapted to your investment horizon may well represent the difference between calamity and opportunity if risk runs wild. Overall, our six-factor risk outlook shows: Volatility – must pick up from current low levels soon; Correlation risk – We concur with the JP Morgan investor note published this September that warned: “Over the past two decades, most risk models were (correctly) counting on bonds to offset equity risk. At the turning point of accommodation, this assumption will most likely fail”6; Stretch risk – Different market variables get tighter and tighter – in the face of greater leverage (as a result of lower volatility), we can only hope that the inevitable unwind will be orderly; Liquidity risk – Liquidity concerns refocused back to the bond markets, with other asset classes representing a lower chance of liquiditysourced contagion; Event risk – The focus seems to have shifted from political and policy uncertainty back to economic conditions, which for the time being are flashing modest growth – look for those conditions to wobble as the unwinding of UMP takes a firmer hold; ESG risk – ESG risks are for real and for now – investors have the opportunity to incorporate these in their decision-making.

13

14



15

HERMES INVESTMENT MANAGEMENT We are an asset manager with a difference. We believe that, while our primary purpose is to help savers and beneficiaries by providing world class active investment management and stewardship services, our role goes further. We believe we have a duty to deliver holistic returns – outcomes for our clients that go far beyond the financial – and consider the impact our decisions have on society, the environment and the wider world. Our goal is to help people invest better, retire better and create a better society for all.

Our investment solutions include: Private markets Infrastructure, private debt, private equity, commercial and residential real estate High active share equities Asia, global emerging markets, Europe, US, global, and small and mid cap Credit Absolute return, global high yield, multi strategy, global investment grade, real estate debt and direct lending Multi asset Multi asset inflation Stewardship Active engagement, advocacy, intelligent voting and sustainable development

Offices

London | New York | Singapore

For more information, visit www.hermes-investment.com or connect with us on social media:

This document is for Professional Investors only. The views and opinions contained herein are those of Eoin Murray, Head of Investment, and may not necessarily represent views expressed or reflected in other Hermes communications, strategies or products. The information herein is believed to be reliable but Hermes does not warrant its completeness or accuracy. No responsibility can be accepted for errors of fact or opinion. This material is not intended to provide and should not be relied on for accounting, legal or tax advice, or investment recommendations. This document has no regard to the specific investment objectives, financial situation or particular needs of any specific recipient. This document is published solely for informational purposes and is not to be construed as a solicitation or an offer to buy or sell any securities or related financial instruments. Figures, unless otherwise indicated, are sourced from Hermes. The distribution of the information contained in this document in certain jurisdictions may be restricted and, accordingly, persons into whose possession this document comes are required to make themselves aware of and to observe such restrictions. Issued and approved by Hermes Investment Management Limited (“HIML”) which is authorised and regulated by the Financial Conduct Authority. Registered address: Lloyds Chambers, 1 Portsoken Street, London E1 8HZ. HIML is a registered investment adviser with the United States Securities and Exchange Commission (“SEC”). BD00977 0002032 10/17

Certified ISO 14001

Environmental Management

www.hermes-investment.com