Currency Weekly. Our three core calls. Market focus pg 2. Quant indicators pg April Buy the fragile five

Transcription

1 April 1 Macro Currency Weekly Our three core calls Market focus pg 1. Buy the fragile five. EUR will come under pressure 3. Cable will fall The first quarter is now behind us. The question is: what does the rest of the year herald? 1 Buy the fragile five For us the trade to have on now is to be long EM, and in particular the fragile five. One feature of the market that is particularly striking is the fall in volatility which supports the long EM carry trade especially now that the current account adjustment is fully underway. EUR will come under pressure This is an important view given the size and the scope of the euro as a currency. However, from a consensus point of view it seems we are in the pack. Here we feel the differences between the Fed and the ECB will define the currency performance. The Eurozone is coming very close to deflation and some type of QE policy is just a matter of time, whilst the Fed is starting to talk about rate hikes. 3 Cable will fall This is perhaps our most controversial forecast and the price action is testing our resolve. Unlike our EM call we do NOT see a continuation pattern. Here we fear the UK balance of payments position will become problematic for the currency as the year progresses. We, as forecasters, have to decide what will become a major driver for the currency in the future and a current account deficit of 5.% of GDP is very high on our list. What is particularly disturbing is that a deficit of this size is usually associated with a late cycle blow out, not as a starting point for a recovery. Quant indicators pg 9 Regular updates of our quantitative indicators. This includes (i) our new positioning indicators; (ii) correlation analysis, both multi-asset (RORO) and G1 FX high-frequency; and (iii) indices that quantify the market s appetite for risk.

2 April 1 Our three core calls 1. Buy the fragile five The fall in volatility is a big plus for EM The main feature of the FX market so far this year has been the constant and relentless fall in volatility, with volatility in many currency pairs at year to date lows (charts 1 and ). The main exception to this theme is the RMB, which has seen volatility rise. This, however, is a policy choice rather than a marketdriven event. It is very hard to explain why we have had such a broad-based fall in volatility, especially given heightened geopolitical risk and talk of higher rates in the US. Low volatility always has the seeds of its own destruction built into it. Low volatility encourages carry trades and as positioning grows and grows, position sizes become untenable. This eventually gives way to a sell off/drawdown, often driven by a political or economic event, and then higher volatility ensues. However, it seems to us that we are now only at the beginnings of a carry cycle and a spike in volatility is therefore unlikely to be position driven. We have suggested that the best way to play this fall in volatility is to buy the high yielding fragile five currencies (TRY, ZAR, BRL, IND, INR). In our eyes we are only at the beginning of the carry cycle. 1.G3 vols are falling...and so is EM 3m Global Hazard Indicator, Jan 1 - Mar 1 % 5 Average 1M implied volatility of 'fragile 5' Average 1M historical volatility of 'fragile 5' % Jan-1 Jan-11 Jan-1 Jan-13 Jan-1 Jan-1 Jan-11 Jan-1 Jan-13 Jan-1 GHI is a composite implied volatility measure of USD, JPY and EUR Current account confusion Did not matter / Mattered / Does not matter It is worth discussing the current account argument and what it means for FX as it seems to operate like an on/off switch. In essence there is nothing wrong with a current account deficit either on a personal or national level. On the personal level, take the example of a young dynamic person with bright prospects who wants to start their own business. Now this young person has limited savings (S) but needs to invest (I) in order to realise future returns. By going into debt by using someone else s savings they can invest

3 April 1 more than they have saved (I must be greater than S). This is sustainable because there is a high level of confidence that future earnings will be more than enough to pay off the debt. The same thing happens on a country level the young and dynamic EM country offers superior rates of return but does not have enough local savings to invest to realise these future returns. So here too Investment must be greater than Savings and so the country imports capital and runs a sustainable current account deficit. When money is cheap and easy to come by this worked well perhaps too well. This describes the period from December when the Fed announced endless QE, until May 13 when the Fed changed tack and first discussed tapering. Fed the trigger for the drawdown The trigger for the drawdown and scramble out of EM and the birth of the fragile five was May 13. Suddenly and dramatically the market saw the prospect of the quantity of money being restricted and the price of money as measured by the 1 year bond rose dramatically. Capital went back to the US and the market became risk averse. Those vulnerable to a squeeze of global capital those with current account deficits suffered sharp falls in their currencies. The market then raised another awkward point: was the capital sent to these EM countries really being used for productive Investment? Perhaps it had been finding its way into Consumption. The combination of these factors saw the market start to force a current account (C/A) adjustment onto the fragile five. First and foremost the currency adjustment starts. The fall in the currency makes imports more expensive and exports more competitive the C/A mechanism is now underway. The currency is driven lower and lower to restore a healthy balance. Eventually reluctant central banks, fearing a currency problem and associated inflation, are forced to act. They are forced to raise rates. This is second stage of the process. Now the currency does not do all the heavy lifting since higher rates attract hot money and slows the currency fall. More importantly, these higher rates slow domestic demand thereby dampening imports. 3. Carry ends with a weaker currency a BoP adjustment starts with it BoP adjustment Carry Stage 1 Weaker currency Stage 1 Weaker economy Stag Higher rates Stage Lower rates Stage 3 Weaker economy Stage 3 Weaker currency Source: HSBC 3

4 IDR BRL ZAR TRY INR COP HUF PLN MXN MYR RON CZK KRW CLP BGN THB PHP SGD PEN ARS TWD RUB CNY Macro April 1 Hence the current account adjustment is fully underway, with first FX, then rates and lastly the real economy all playing their part. The point as shown by the schematic above is in a carry type world we go from stage 1 a weak economy to stage lower rates and finally this causes the currency to weaken. Importantly, in the BoP adjustment we start with a weaker currency then move to higher rates move higher and finally to a weaker economy. The weaker economy is an end point caused initially by the FX adjustment. Now that the economies are weakening some fear buying the currency, and that would be correct if we were in the carry world, but in the world of the BoP adjustment it is the end point of the adjustment and is not to be feared. In fact, it signifies that the adjustment is underway and the currency has done its job. So in our eyes the process is near completion in the fragile five. Couple this with low volatility and higher rates and these currencies are to be bought..em FX is making a comeback EM FX performace against the USD, 1 Feb - 31 Mar, % Flows are starting to recover Total foreign portfolio flows into 1 EM countries, USD bn USD bn USD bn Source: HSBC, IMF Differentiation will start soon So far we have seen a more or less collective rally in EM FX (chart ), although, some currencies still remained under pressure given individual factors. RUB was sold-off aggressively due to the Russia- Ukraine stand-off, and the RMB volatility has been driven by official policies rather than market driven events. Going forward, we expect greater differentiation. Asia, we believe North Asian currencies will generally outperform South Asian currencies, given their stronger balance of payments position and comparative cheapness on long-term valuation measures. In LatAm, we recommend being long BRL to earn carry and we have turned more bullish on COP. Finally in EMEA, we like the TRY for its carry appeal and the uncertainty surrounding the elections is gone for now. The ZAR should also remain supported given the current account deficit has begun to show some improvement. We also like the CEE currencies.. EUR to come under pressure This is an important view given the size and the scope of the euro as a currency. However, from a consensus point of view it seems we are in the pack. The Bloomberg consensus has the euro falling to 1.31 by year end vs the HSBC 1. spot forecast. Here we feel the differences between the Fed and the ECB will drive the currencies. The Eurozone is coming very close to deflation with the latest CPI print at a mere.5% YoY (chart ) whilst the Fed is on course to end QE and is even talking about higher rates at

5 April 1 some point in the future. The EUR falling relies on the ECB loosening monetary policy further and entering into some type of QE arrangement.. Eurozone inflation is at its lowest rate since 9 %. Headline inflation, % YoY Eurozone inflation Core inflation (ex. Food, energy, alcohol and tobaco), % YoY % , Eurostat In essence, we believe that the EUR continues to be a carry trade where currency movements are largely determined by movements in relative rate expectations. Our economists have long been of the view that the growth recovery in the Eurozone will not be strong enough to reverse the disinflation trend and that the ECB will have to resort to further stimulus in the coming months and eventually outright asset purchases to address the downside risks to price stability. At the same time, the Fed is in a process of unwinding their asset purchases. Moreover, at their latest meeting the FOMC revised up their rate projections for 15 and 1 and Yellen signalled that the first rate hike could be delivered as soon as six months after the end of QE tapering (according to our expectations the hike could be delivered in H 15). Chart 7 shows EUR-USD plotted against the expected gap between Eurozone and US 3M interest rates by the end of 15. This illustrates clearly the potential for EUR-USD to be a lot lower. 7. According to rate expectations EUR should be much lower EUR-USD Expected Dec'15 3M rate differential (EUR-USD, RHS) % Jun 3-Jul 3-Aug 3-Sep 3-Oct 3-Nov 3-Dec 3-Jan 3-Feb 3-Mar 3-Apr 5

6 April 1 3. Cable will fall This is perhaps our most controversial forecast and the price action is testing our resolve. There is no doubt that the UK economy has surprised to the upside, and this is pushing cable higher. For the moment the UK is winning the FX ugly contest. The debate on rates is now about whether the UK will raise rates late this year or after the elections which will be held in May 15. So with the economy doing so well why don t we just go with the flow? Unlike our EM call we do NOT see a continuation pattern. Here we fear the UK balance of payments position will become problematic for the currency as the year progresses. In the UK the overall current account deficit remains huge. However, for the moment the market is choosing to ignore this aspect. We as forecasters have to decide what will become a major driver for the currency in the future and a current account deficit of 5.% of GDP is high on our list. What is particularly disturbing is that a deficit of this size is usually associated with a late cycle blow out, not as a starting point of a recovery. On this basis we expect cable to fall to 1.5 by year end. The UK economic upswing has been driven largely by consumption supported, not by rising real incomes, but by a combination of dis-saving and wealth effects from a government-boosted housing market. We believe this imbalance in growth drivers will be echoed in rising concerns regarding the UK s sizeable current account deficit, which reached 5.% of GDP in Q 13 (chart ), the widest since 199. That deficit took around 1 years of expansion to accumulate, and it is an alarming place to be in the early stages of recovery.. UK current account deficit reached its 19s lows.% 1.%.% -1.% -.% -3.% -.% -5.% -.% UK Current Account, % of GDP % 1.%.% -1.% -.% -3.% -.% -5.% -.% Source: HSBC, ONS For the most of 13 and so far in 1, GBP performance has been driven by the carry considerations. Chart 9 shows GBP-USD plotted against expected differential between UK and US 3M interest rates by the end of 15. The market has consistently brought forward its expectation of interest rate rises and has driven cable up. However, with the Fed gradually unwinding its asset purchases and signalling first rate hike in H 15 while the BoE s rhetoric is rates to stay low carry considerations may begin to move in

7 April 1 9. Rate differential fully explain cables move %.7 Expected Dec15 3M rate differentials (UK-US) GBP-USD (RHS) Jan-1 Mar-1 May-1 Jul-1 Sep-1 Nov-1 Jan-13 Mar-13 May-13 Jul-13 Sep-13 Nov-13 Jan-1 Mar-1 1. the opposite direction. Moreover, strong consumption, which is at the heart of current UK recovery, will eventually falter as real earnings growth has actually been negative since 9. Once this starts to happen, the UK economy will start to look weak again, and the market s expectations of the first rate hike would be postponed driving the GBP-USD lower. However, the turn in rate expectations is only part of the story. In addition, GBP may find itself more in focus for its balance of payments fundamentals, a process that also points to considerable downside for the currency. The nature of the UK current account financing explains why the currency is becoming more vulnerable. Large portfolio inflows into the UK during the crisis have become large outflows in recent quarters. Portfolio outflows have been mostly larger than direct investment inflows, which means short term financing flows have been large enough to cover this capital flow deficit as well as the wider current account deficit. Chart 1 shows the short-term financing flows have indeed been building up. This is fine 1. Short term inflows into the UK have become substantial GBP bn 3 Implied short term funding flows (Q sum) GBP bn Source: HSBC, ONS 7

8 April 1 whilst the market believes the UK economic story but when this belief falters the fall in the currency will be fast and furious. It is also clear from the chart that since last year the implied short-term financial flows have been narrowing. As soon as doubts about the sustainability of the UK recovery emerge, the short-term inflows will be undermined even further and GBP would face the double whammy of questions over the trade deficit s sustainability alongside retreating interest rate hike expectations. Conclusion: the three core calls Buy the fragile five EUR will come under pressure Cable will fall In terms of belief and importance the HSBC FX core calls are to buy the fragile five, expect a big move lower in GBP and (despite its consensus nature) we think the EUR is overvalued. The fragile five call is playing out. Eventually this idea will get over crowded and carries the seeds of its own destruction but not yet. The cable call is our most non-consensus call and the one that there is the most resistance to. There is no doubt GBP is winning the ugly contest but the nature of the recovery and the huge external imbalance will mean a severe correction at some stage this year. Lastly and most in line with consensus we expect the relative policy differences between the ECB and Fed to undermine the euro.

9 April 1 Quant Indicators 1. HSBC Positioning Indicators (pg 1) The HSBC Positioning Indicators measure the degree to which the momentum community is either long or short of a currency pair. For exchange rates where position data is available from the IMM, we compare the two sources of data. Discrepancies between these two sources of data can be particularly informative about positioning and sentiment of fundamental FX traders.. Correlation Analysis: Multi-Asset (RORO) & G1 FX (pg 15) (a) RORO Index multi-asset correlations The RORO Index is at moderately low levels. However, correlations have risen recently. (b) Emerging Market RORO Indices Regional correlations within EM regions have increased recently. In addition, the regional EMcommon-factors remain strongly risk-on. (c) Equity RORO Index The Equity RORO index measures the strength of correlations within the main risky asset class of equities. The Equity RORO Index is significantly lower than the all-time highs seen in late 11, but is still above pre-crisis levels. (d) High-frequency G1 FX correlations We show the strength of the correlations between all G1 exchange rates, calculated using hourly FX price data. 3. Risk Appetite: OPRA Index and MRAI (pg 5) The OPRA index measures risk appetite based on the positions held in contracts with varying degrees of risk by speculative traders on US futures exchanges. The OPRA Index is in neutral territory; this indicates that speculative traders on the US futures exchanges have shifted their exposure in a way unrelated to the risk. The MRAI has generally moved sideways with high volatility since May 1. This is indicative of neutral risk appetite and is consistent with the RORO phenomenon. 9

10 April 1 HSBC Positioning Indices Current Value of HSBC Positioning Indices 1

11 April 1 HSBC EUR-USD Positioning Index Mar-1 Jun-1 Sep-1 Dec-1 Mar-13 Jun-13 Sep-13 Dec-13 Mar-1 HSBC EUR-USD Positioning Index Net EUR-USD IMM Speculative Position HSBC USD-JPY Positioning Index HSBC USD-JPY Positioning Index Net USD-JPY IMM Speculative Position -1 Mar-1 Jun-1 Sep-1 Dec-1 Mar-13 Jun-13 Sep-13 Dec-13 Mar HSBC GBP-USD Positioning Index HSBC GBP-USD Positioning Index Net GBP-USD IMM Speculative Position -1 Mar-1 Jun-1 Sep-1 Dec-1 Mar-13 Jun-13 Sep-13 Dec-13 Mar

12 April 1 HSBC AUD-USD Positioning Index HSBC AUD-USD Positioning Index Net AUD-USD IMM Speculative Position -1 Mar-1 Jun-1 Sep-1 Dec-1 Mar-13 Jun-13 Sep-13 Dec-13 Mar HSBC USD-CAD Positioning Index HSBC USD-CAD Positioning Index Net USD-CAD IMM Speculative Position -1 Mar-1 Jun-1 Sep-1 Dec-1 Mar-13 Jun-13 Sep-13 Dec-13 Mar HSBC NZD-USD Positioning Index HSBC NZD-USD Positioning Index Net NZD-USD IMM Speculative Position Mar-1 Jun-1 Sep-1 Dec-1 Mar-13 Jun-13 Sep-13 Dec-13 Mar

13 April 1 HSBC USD-CHF Positioning Index HSBC USD-CHF Positioning Index Net USD-CHF IMM Speculative Position -1 Mar-1 Jun-1 Sep-1 Dec-1 Mar-13 Jun-13 Sep-13 Dec-13 Mar HSBC USD-MXN Positioning Index HSBC USD-MXN Positioning Index Net USD-MXN IMM Speculative Position -1 Mar-1 Jun-1 Sep-1 Dec-1 Mar-13 Jun-13 Sep-13 Dec-13 Mar HSBC EUR-GBP Positioning Index HSBC EUR-GBP Positioning Index -1 Mar-1 Jun-1 Sep-1 Dec-1 Mar-13 Jun-13 Sep-13 Dec-13 Mar

14 April 1 HSBC EUR-JPY Positioning Index HSBC EUR-JPY Positioning Index -1 Mar-1 Jun-1 Sep-1 Dec-1 Mar-13 Jun-13 Sep-13 Dec-13 Mar HSBC USD-ZAR Positioning Index HSBC USD-ZAR Positioning Index -1 Mar-1 Jun-1 Sep-1 Dec-1 Mar-13 Jun-13 Sep-13 Dec-13 Mar ; 1

15 April 1 HSBC Risk On Risk Off Index Risk On Risk Off Index RORO Index RORO paradigm stronger The RORO index is at moderately low levels, but has risen from the lows seen towards the end of 13. RORO paradigm weaker This indicates that the risk on risk off phenomenon is less dominant than in the height of the crisis, despite having shown a recent resurgence. See Appendix A1 for more details of the methodology. Asset correlations with the risk on risk off factor RORO Correlations Strongly risk on Uncorrelated with RORO Strongly risk off The assets that were most highly correlated with the risk on risk off factor during the previous weeks were: Risk-on assets Dow Jones S&P 5 Risk-off assets VIX US Government Bonds Uncorrelated with RORO Wheat Natural gas EM regions are all positively correlated with RORO. 15

16 April 1 HSBC Emerging Market RORO Indices Emerging market risk on risk off indices EM RORO Indices Correlation between assets within each region strengthening Correlation between assets within each region weakening The regional indices have all increased recently. This indicates that regional correlations within EM regions have all strengthened recently. Countries included: Asia: Hong Kong, South Korea, Singapore, India, Taiwan, Malaysia, Thailand Latam: Brazil, Mexico, Chile EMEA: Czech Republic, Hungary, Poland, South Africa, Turkey Interpretation Risk on risk off is a truly global phenomenon that drives returns and causes high correlations across many different markets and geographic regions. However, there can still be variations in the strength of correlations between assets from different markets, as well as differences in the extent to which these correlations are driven by risk on risk off rather than region-specific factors. To quantify the strength of correlations in different emerging markets, we construct three EM RORO indices (shown in the chart above). A high index level indicates strong correlations between assets in that region. For example, when the Asia RORO index is high this implies that a single factor is driving returns across Asia, which leads to strong correlations between Asian assets. Similarly, high levels of the Latam and EMEA RORO indices imply that correlations are high in Latin America and EMEA, respectively. Strong correlations between assets in different regions can be caused by local phenomena as well as global RORO dynamics. To illustrate the importance of risk on risk off rather than local factors in driving correlations, in the bar chart on the previous page we show the extent to which the different regions are driven by the RORO factor. When a region is strongly driven by risk on risk off, it will have a high correlation with the RORO factor and will appear to the left of the bar chart. On the other hand, if regional correlations are not primarily driven by risk on risk off, but instead by other local factors, a region will be only weakly correlated with the RORO factor. The picture today Correlations within EM regions have increased recently. Methodology See Appendix A for more details of the construction methodology. 1

17 April 1 Correlation heat map Heat map showing correlations over the last days Reading the heat maps The heat map shows the correlations between different assets during the last days. Dark red regions indicate strong positive correlations. Dark blue regions indicate strong negative correlations. Yellow and green regions indicate weak correlations/uncorrelated assets. The picture today The heat map illustrates that market structure looks similar to pre-crisis times. Despite increased market focus on RORO, the typical RORO structure has not returned yet. However, see our recent piece RORO Returns: Risk On Risk Off rears its head again, 1 March 1 for more details on the recent RORO behaviour. 17

18 April 1 Correlations with the risk on risk off factor through time Rolling correlations of individual assets with the risk on risk off factor The charts show the strength of the correlations between individual assets and the risk on risk off factor through time. These correlations quantify the extent to which the different assets are driven by risk on risk off. A correlation close to 1 implies that the asset is strongly risk on; a correlation close to -1 implies that the asset is strongly risk off; and a correlation near zero suggests that the asset is not primarily driven by the risk on risk off phenomenon. 1

19 Correlation with RORO factor Correlation with RORO factor Correlation with RORO factor Macro April 1 G1-FX Correlations with the RORO factor Rolling correlations of G1-FX with the risk on risk off factor 1.5 AUDUSD USDCAD EURCHF EURUSD GBPUSD USDJPY NZDUSD EURNOK EURSEK

20 Correlation with RORO factor Correlation with RORO factor Correlation with RORO factor Correlation with RORO factor Correlation with RORO factor Macro April 1 Asia-FX Correlations with the RORO factor Rolling correlations of Asia-FX with the risk on risk off factor 1.5 USDCNY USDIDR USDINR USDKRW USDMYR USDSGD USDTHB USDTWD LatAm-FX Correlations with the RORO factor Rolling correlations of LatAm-FX with the risk on risk off factor 1 USDBRL USDCLP USDCOP USDMXN

21 Correlation with RORO factor Correlation with RORO factor Correlation with RORO factor Macro April 1 EMEA-FX Correlations with the RORO factor Rolling correlations of EMEA-FX with the risk on risk off factor 1.5 EURCZK EURHUF EURPLN USDILS USDRUB USDTRY USDZAR

22 April 1 HSBC Equity RORO Index Equity RORO Index EM RORO Indices Increasing correlation between individual equity returns The Equity RORO Index is above pre-crisis levels. However, it is still significantly below the all-time highs seen in late 11. Decreasing correlation between individual equity returns This indicates that equity moves are showing significantly more dispersion than during late 11, but are more similar than was typical in pre-crisis times. See Appendix A3 for more details of the methodology. Interpretation Whilst risk on risk off is inherently a cross-asset phenomenon, equities are the quintessential risk-on asset. When there is a perception in the market that correlations are high, it is important to determine whether it is simply a within-asset-class phenomenon or part of the wider global macro theme. The HSBC Equity RORO Index allows us to distinguish between high correlations which are specific to this main risky asset class and high cross-asset correlations, as measured in the original RORO Index, which indicate broader macro stress. The picture today At the moment the Equity RORO Index is above pre-crisis levels, but is still significantly lower than the all-time highs seen in late 11. This indicates that movements in individual equities are showing significantly greater dispersion than in late 11, but are more similar than was typical in pre-crisis times.

23 April 1 Correlation of sectors with Equity RORO factor Rolling correlations of individual sectors with Equity RORO factor These charts show the rolling correlations between the returns of individual equity sectors and the Equity RORO factor. Values close to +1 indicate that the sector is simply moving in response to changes in the Equity RORO factor. The closer the value is to, the more that sector is displaying sector-specific character. Interpretation Some sectors are currently showing moderately low correlations to the Equity RORO factor. This is consistent with the level of the Equity RORO index being much lower than the all-time highs seen in late 11. 3

24 April 1 G1 Exchange Rate Correlations In the linked document at the following url ( we show the strength of the correlations between all G1 exchange rates. If one has a view on how an exchange rate is going to move, this can be used to identify other trading opportunities by highlighting other currency pairs that move independently or in the same (or opposite) direction. The chart below is an example page from this document for AUD-JPY. The three bar charts show: The correlation of AUD-JPY with all other G1 crosses during the previous week; A comparison of AUD-JPY correlations during the previous week with a 1-week period 1-month ago; and A comparison of last week s AUD-JPY correlations with the average correlation during the previous month. To enable us to calculate correlations over periods as short as a week, we have used hourly price data. In the linked document, we provide similar charts for all other G1 crosses and more details of the methodology that we use to construct the charts. Example page from the linked correlation document: AUD-JPY correlations over the last week and versus the previous month Source: HSBC

25 Risk Appetite Decreasing Neutral Territory Risk Appetite Increasing Macro April 1 OPRA OPRA Index May-1 Nov-1 May-11 Nov-11 May-1 Nov-1 May-13 Nov Interpretation When the OPRA index is close to 1 it indicates that speculators have increased their exposure to risky assets, whereas a value close to -1 indicates that speculators have shifted their exposure to less risky assets. The picture today The OPRA Index is in neutral territory. This indicates that speculative traders on the US futures exchanges have shifted their exposure in a way unrelated to the risk. Methodology The OPRA index is based on the relationship between changes in the futures positions held by speculative traders in various contracts and the risk associated with holding the contracts. See Appendix B for more details of the methodology. 5

26 April 1 MRAI MRAI: Short-term picture Short-term picture The price-based risk appetite index has moved sideways with high volatility since May 1. Volatile and no clear trend This index is based on changes in prices and volatilities of assets that are known to be affected by risk appetite. See Appendix C for more details of the methodology. MRAI: Long-term picture Long-term picture The MRAI is in a long-term downward trend. Decreasing risk appetite Interpretation A positive trend in the MRAI implies increasing risk appetite whereas a negative trend implies decreasing risk appetite. The picture today The MRAI has been volatile and has shown no clear trend since May 1. This indicates that there is constantly changing appetite for risk, which is consistent with the risk on risk off phenomenon.

27 April 1 Appendix A1: RORO Methodology Market-wide correlation index HSBC Risk On Risk Off (RORO) Index The Risk On Risk Off (RORO) index takes the rolling correlations between the daily returns of the 3 assets listed in the table below and combines them into a single index. We construct the index by using principal component analysis (PCA) to decompose the 3 asset return time series into 3 principal components (PCs), which are mutually uncorrelated variables that explain the observed asset returns. The first PC represents the most important factor driving financial markets during a particular time period. In current market conditions, this factor can be considered to represent risk on risk off. That is, the paradigm in which the market either believes the future is bright risk on or that it is bad risk off. The proportion of the variance explained by the first PC then provides an indication of the strength with which this paradigm dominates markets. If the first PC dominates markets and explains a large proportion of the variance, this implies that market-wide correlations are strong, which is a key feature of the risk on risk off paradigm. In this scenario, this single factor is driving synchronized changes amongst many different markets; hence correlations are high. We define the RORO index as the variance in market returns explained by the first PC. An increase in the RORO index implies an increase in market correlations, whereas a decrease implies that market correlations have decreased. In constructing the index we focus on markets that have a large overlap in trading hours (Europe and North America and Asian currency markets). This enables us to track correlations on a daily basis without having to worry about the non-synchronicity of return time series. We also consider correlations between the different assets and the risk on risk off factor. These are the correlations between the different return time series and the first PC, and can also be considered to provide an indication of the extent to which risk on risk off is driving different assets. Assets included in the RORO Index Source: HSBC Equities Government bonds (1 year yields) Corporate bonds (yields) Currencies ( trade weights indices) Metals S&P US AAA USD Gold VIX Dow Jones Canada BAA EUR Silver Oil NASDAQ UK CHF Copper Natural Gas Russell Germany GBP Heating Oil FTSE 1 France JPY Wheat Euro Stoxx 5 AUD Soybean DAX CAD Cotton CAC NZD Other 7

28 April 1 Appendix A: EM RORO Regional emerging market correlations HSBC Emerging Market RORO Indices We produce Emerging Market RORO Indices for Asia, Latin America, and EMEA. We construct the indices using a similar methodology to that described in Appendix A1 for the cross-asset RORO index. For each region, we perform a principal component analysis (PCA) on the returns of a range of assets from that region. We then define each regional index as the proportion of the variance in the returns of assets in that region explained by the first principal component (PC). For the original multi-asset RORO Index the first PC represents the most important global macro factor driving returns across a wide range of different assets. When the RORO index is high, this factor is strong. The regional EM indices have an analogous interpretation. For example, when the Asia RORO index is high this implies that a single factor is driving returns across Asia, which leads to strong correlations between Asian assets. Similarly, high levels of the Latam and EMEA RORO indices imply that correlations are high in Latin America and EMEA, respectively. For each of the regions, we use both bond and equity data for the countries listed in the table below. To enable us to compare the regional indices, we use weekly price data to eliminate any effects due to the different time zones. This also allows us to compare these indices to the cross-asset RORO. We consider the correlation between the dominant market factor in the different regions and the main risk on risk off factor that we identify in our cross-asset analysis. This is the correlations between the first PC for each region and the first PC for the cross-asset returns. The strength of these correlations can be considered to provide an indication of the extent to which risk on risk off is driving returns in the different regions. Assets included in the EM RORO Indices Asia Latin America EMEA Hong Kong Brazil Czech Republic South Korea Mexico Hungary Singapore Chile Poland India South Africa Taiwan Turkey Malaysia Thailand Source: HSBC

29 April 1 Appendix A3: Equity RORO Equity market correlations HSBC Equity RORO Index The HSBC Equity RORO Index looks at all current members of the S&P 5 Index that have an appropriate data history back to 1 January 199. We use a similar construction methodology for this index to the one described in Appendix A1 for the RORO Index. To construct the Equity RORO Index we perform a principal component analysis (PCA) on the returns of all of the equities that we consider. We define the index as the proportion of the variance in the returns of these equities that can be explained by the first principal component (PC). This first PC is the most important factor driving the returns at any time. For the original multi-asset RORO Index the first PC represents the most important global macro factor driving returns across a wide range of different assets. When the RORO index is high, this factor is strong. For the Equity RORO, there is an analogous interpretation; however, in this case we are only looking at the risky asset class of equities. When the Equity RORO index is high it indicates there is a supercharged market beta dominating stocks correlations are high and individual identity is reduced. We use the two indices together to characterise the stress in the global macro environment. High correlations are generally an indication of market strain and have consequences for most asset classes. The two indices help understand the extent to which stress is confined to risky assets or is more comprehensive. 9

30 April 1 Appendix B: OPRA Methodology Position-based risk appetite index Open Positions Risk Appetite (OPRA) Index We use speculative positions from the CFTC Commitments of Traders report to measure risk appetite. We track changes in exposure of the speculative community to the various contracts listed in the table below and relate these changes to the risk associated with the contracts. We view it as a sign of high risk appetite when the speculative community has increased its exposure to the more risky assets more than for less risky assets. To measure this we calculate the rank correlation between changes in the speculative open interest and volatility. A rank correlation is used since this is less susceptible to outliers than a standard correlation. Since this is a correlation, the index will lie between -1 and +1. A value close to +1 indicates that speculators have been increasing their positions in risky assets across the board, with the largest percentage increase in exposure being in the riskiest assets. A value close to the minimum value of -1 indicates the opposite. If speculative positions have been changing in a way unrelated to risk, then the value of this index will be close to zero. Contracts included in OPRA Index Agricultural Drinks Metals Currencies Oil Other Corn Cocoa Platinum AUD LSCrude Lumber Oats Coffee Silver CAD Unleaded Rough Rice OJ Copper CHF Heating Oil Soybeans EUR Natural Gas Soybean Oil GBP Soybean Meal JPY Wheat Cotton Lean Hogs Live Cattle Source: HSBC 3

31 April 1 Appendix C: MRAI Methodology Price-based risk appetite index Market Risk Appetite Index (MRAI) The MRAI measures the aggregate level of risk appetite in the financial system using risk premia from various markets. The index is based on changes in price and volatility of several assets that are known to be strongly affected by risk appetite. A positive trend in the MRAI implies an increasing appetite for risk whereas a negative trend in the MRAI implies a decreasing appetite for risk. We construct the index using equally weighted z-scores of changes in the level of six inputs: the VIX and VDAX volatility indices; the Global Hazard Index, which aggregates the 3-month implied volatilities for EURUSD, USDJPY, and EURJPY; BAA and AAA corporate bonds spreads; and interest rate swap spreads. 31