ACTIVE ASSET ALLOCATION IS IT WORTH IT?

Transcription

1 INFORMATION FOR INVESTMENT PROFESSIONALS ACTIVE ASSET ALLOCATION IS IT WORTH IT? MULTI ASSET SEPTEMBER 2018 Craig Nowrie Client Portfolio Manager In a world where volatility is heightened, managing overall risk is key. The need for a dynamic asset allocation portfolio is more prevalent than ever. Over the five years since we launched the Threadneedle Dynamic Real Return Fund we have exceeded our performance objective with volatility akin to global bonds. Background Five years ago we launched the Dynamic Real Return Fund, which aims to deliver an inflation-plus return (equivalent to the Consumer Price Index plus 4%) over the medium term, with up to two-thirds the volatility of global equities. The fund aims to navigate markets by protecting investors capital when threats appear on the horizon and participating when opportunities develop by dynamically managing the asset allocation. Simply put: if equity markets are up 20%, a good result for the fund would be up 12%; by contrast, when equity markets are down 20% then an acceptable result for the fund would be down 5%. This asymmetry appeals to investors and should enable us to deliver equity-like returns over the long term, with less volatility. Over this five-year period we have managed to beat our CPI +4% objective by, with volatility akin to global bonds. When talking to clients we have reflected on when we have performed well and when we have missed opportunities. While we are pleased to have delivered these strong risk-adjusted returns, as an inquisitive team we have been asking ourselves could we have done better. We therefore tasked our independent investment risk team to assess how well we have actually done over time. But before going into more detail on the resulting million portfolios simulation, it is Issued September 2018 Valid to end April 2019 Page 1 of 15

2 important to take a step back and outline the key decisions that we make in constructing the Dynamic Real Return Fund. Firstly, we start with a blank sheet of paper with the following allocation constraints: Equities 0%-75% Cash & Fixed Income 0%-100% Property 0%-20% Commodities 0%-20% Alternatives 0%-10% The first step in building the portfolio is identifying the risk environment we believe we are currently in; effectively, will taking risk be rewarded or not? We express this in terms of five possible risk quintiles, as shown in Figure 1. Figure 1: the five risk buckets Risk for this strategy is capital loss, and we can therefore position the portfolio to protect our investors capital when we perceive there to be a number of threats on the horizon. Conversely, we can participate when we believe opportunities are aplenty and we will be rewarded by taking more risk (ie, good risk premia are available). This sets the parameters for the maximum and minimum risk allocations of the portfolio the minimum being 100% in short-dated government bonds and the maximum 75% in equities. Issued September 2018 Valid to end April 2019 Page 2 of 15

3 The million portfolios simulation Our independent risk team then ran a million portfolios simulation consisting of all the possible portfolios within the asset allocation limits above. They then removed those portfolios that were deemed not to have enough risk (ie, the overall risk of the portfolio was below the risk of fiveyear government bonds) and those that were too high risk (ie, those portfolios that had risk above 75% equity volatility). So, for example, given the asset allocation restrictions it is theoretically possible for a portfolio to consist of 75% equities and 25% in emerging market local bonds, but the risk of this portfolio would be above that of 75% in equity volatility and as such would be excluded. The remaining portfolios were then attributed to the five risk buckets in Figure 1, from Strongly Dislike risk to Strongly Favour risk. We then could see how each of these hypothetical portfolios, held statically over a one-year period, would perform (using index returns) and compare this to the actual return of the Dynamic Real Return Fund and to the potential returns within each of the five risk buckets. This allowed us to observe whether we selected the correct risk bucket and if we selected a more or less successful combination of asset classes within this bucket. The charts that follow show the return of each of these portfolios, with the height of the distribution being the proportion of instances a random portfolio returned a certain amount (shown in the horizontal axis). Our view is that this analysis can be applied to any long-only, active asset allocation fund, allowing investors to review the returns achieved within each of these risk buckets and compare them to the return of their individual manager. Issued September 2018 Valid to end April 2019 Page 3 of 15

4 June 2013 one-year return In the first year the Dynamic Real Return Fund was launched it returned 6.2%, and throughout this period we were in the Neutral risk category. Figure 2: 2013 risk quintile return frequencies 1.6% Strongly Dislike Dislike Neutral Favour Strongly Favour Dynamic Real Return Fund Dynamic Real Return Fund, 6.2% Over the first annual period the selection of risk category didn t have a meaningful impact on return, though interestingly the distribution of the Dislike risk would, theoretically, have resulted in the worst performance. What was more important in this period was the selection of assets within each risk bucket, as can be seen by the width in returns. In other words, it was possible to deliver strong positive or indeed negative returns for any given risk exposure. With the Dynamic Real Return Fund in the Neutral risk category over this period (see Figure 3), within this risk category our asset selection was relatively favourable in other words, we are on the right side of the distribution curve and within the 25% percentile of potential portfolios. Nonetheless, on reflection there were opportunities that we missed within this risk bucket. Issued September 2018 Valid to end April 2019 Page 4 of 15

5 Figure 3: 2013 Neutral return frequency Neutral Dynamic Real Return Fund Dynamic Real Return Fund, 6.2% Issued September 2018 Valid to end April 2019 Page 5 of 15

6 June 2014 one-year return The fund returned 7.7% over the one-year period from June Throughout this time we were more cautious on risk so were in the Dislike risk category, having moved from Neutral. Figure 4: 2014 risk quintile return frequencies 1.8% Strongly Dislike Dislike Neutral Favour Strongly Favour Dynamic Real Return Fund 1.6% Dynamic Real Return Fund, 7.7% On reflection it would have been preferable to have allocated more risk, though the distribution within both these risk categories were relatively wide with a strong skew to the positive. Avoiding the Strongly Dislike risk category was also preferable. The Neutral risk category distribution was actually slightly worse than the Dislike risk category. So overall we did a reasonable job in selecting a preferential risk category, but again, selecting the right asset classes within the risk budget was the most important aspect. So how did we do within this category? Issued September 2018 Valid to end April 2019 Page 6 of 15

7 Figure 5: 2014 Dislike return frequency Dislike Dynamic Real Return Fund Dynamic Real Return Fund, 7.7% Within the Dislike risk category (Figure 5) we were at the sixth percentile so did a good job in selecting the correct mix of asset classes, avoiding the left-hand side where there were a number of negative returns. Issued September 2018 Valid to end April 2019 Page 7 of 15

8 June 2015 one-year return Over the one-year period from June 2015 the fund remained in the Dislike risk category and returned 5%. Unlike the previous year, in this period the greatest impact on returns was selecting the right risk category, in that it paid to take on more risk as during this period risk premia were generally well rewarded. This is clearly shown in the distribution of returns in Figure 6. Figure 6: 2015 risk quintile return frequencies 2.0% Strongly Dislike Dislike Neutral Favour Strongly Favour Dynamic Real Return Fund 1.8% 1.6% Dynamic Real Return Fund, 5.0% Over this one-year period we failed to take enough risk. Furthermore, our selection of assets within the risk category was poor, leaving the fund in the 97 th percentile. To put the fund s return in context, over this one-year period UK equities (FTSE All-Share) delivered a return of 2.2%, global equities returned 13.9%, global equities hedged to sterling returned -3.4%, UK gilts returned 13.5%, and global bonds unhedged returned 3. The impact of the result of the referendum to leave the European Union on the potential return outcomes was clearly substantial and came through a sharp currency movement. As a sterlingbased volatility-controlled fund, currency is a portfolio lever to be carefully managed. While outsized gains were in retrospect on offer to portfolio managers willing to concentrate risks in Issued September 2018 Valid to end April 2019 Page 8 of 15

9 anti-sterling bets, these were not positions that would be in keeping with the risk-management deployed in the portfolio to date. If we look at the previous month s annual return (May 2015), before the referendum and the subsequent devaluation of sterling, the return to risk appears fairly even. Figure 7: 2015 Dislike return frequency Dislike Dynamic Real Return Fund 1.6% Dynamic Real Return Fund, 5.0% As with previous years the spread within the risk category was relatively wide, once again highlighting the importance of active asset allocation. The portfolio allocation was disappointing, suffering from too little interest rate risk and too little overseas currency exposure. Exposure to equity risk, credit risk and commodity risk were all unhelpful. Issued September 2018 Valid to end April 2019 Page 9 of 15

10 June 2016 one-year return Over the one-year period from June 2016 the fund was again in the Dislike risk category and returned 8.3%. The selection of the risk category was relatively important in that you were paid to take on more risk; though interestingly around a fifth of the portfolios in each of the risk categories resulted in an absolute negative return over the year. Figure 8: 2016 risk quintile return frequencies 2.0% Strongly Dislike Dislike Neutral Favour Strongly Favour Dynamic Real Return Fund 1.8% 1.6% Dynamic Real Return Fund, 8.3% Issued September 2018 Valid to end April 2019 Page 10 of 15

11 Over this period the fund was in the Dislike risk category (Figure 9), and within this we were within the third percentile of portfolios. In other words, the assets we selected and our active allocation over this period was well rewarded. Figure 9: 2016 Dislike return frequency Dislike Dynamic Real Return Fund 1.6% Dynamic Real Return Fund, 8.3% Issued September 2018 Valid to end April 2019 Page 11 of 15

12 June 2017 one-year return Over the one-year period from June 2017 the fund was in the Neutral risk category and returned 4.7%. The selection of the risk category was relatively unimportant in that you were paid to neither take on more risk or be more defensive. Approximately 40% of portfolios in the Favour and Neutral risk allocation buckets resulted in absolute negative return. Figure 10: 2017 risk quintile return frequencies Over this period the fund was in the Neutral risk category (Figure 11), and within this we were within the fifth percentile. In other words, our selection of assets was strong in this period. Issued September 2018 Valid to end April 2019 Page 12 of 15

13 Figure 11: 2017 Neutral return frequency Neutral Dynamic Real Return Fund 2.5% 2.0% 1.5% 0.5% Dynamic Real Return Fund, 4.7% The initial results of this analysis have proven to be very interesting to us as a team. We will continue to evolve this million portfolios work to improve and refine the data. That said, it is good to reflect on the results and consider if there are any areas of our process we can improve. The results do indicate, as one would expect, that the allocation of risk is more rewarded in certain periods than others. The one common element within each risk category is that the divergence between the best- and worst-performing asset class in any risk bucket was around 20%, highlighting the importance of asset allocation. Issued September 2018 Valid to end April 2019 Page 13 of 15

14 Summary Since the inception of the Dynamic Real Return Fund five years ago we have produced strong risk-adjusted returns for our clients delivering a return of per annum above our CPI+4% target with risk akin to government bonds. The million portfolios work has indicated that at times we managed to identify the most beneficial risk category and then, more often than not, a better than average mix of asset classes to produce these strong returns. However, this was not the case in every time period. There have been periods where, with the benefit of hindsight, we should have taken more risk was a good example where up until the end of June we had delivered strong risk-adjusted returns, but after the referendum on leaving the European Union we should have taken the risk allocation to the maximum. Looking at risk-adjusted returns over the full period, the average Sharpe ratio for the fund is above the median Sharpe ratio for all median quintiles, and in absolute return terms we are above the median return for the three lower risk buckets but behind over the Favour and Strongly Favour median. This work has not changed our thought process or how we would have managed our asset allocation over the years, but it will make us reflect more at times of stress to see if we are too conservative in managing the strategy. We firmly believe that the dynamic allocation approach used by the fund is key to achieving a good real return with less volatility than equities. Not being beholden to any asset class is critical, as ultimately we are accountable for every basis point of portfolio return. To date, we have successfully managed the active asset allocation with great success and we look forward to continuing to navigate these challenges for our investors going forward. Issued September 2018 Valid to end April 2019 Page 14 of 15

15 Important information: For investment professionals only, not to be relied upon by private investors. Important Information: Past performance is not a guide to future performance. Past performance is shown gross of fees. Your capital is at Risk. The value of investments and any income is not guaranteed and can go down as well as up and may be affected by exchange rate fluctuations. This means that an investor may not get back the amount invested. The mention of any specific shares or bonds should not be taken as a recommendation to deal. Threadneedle Opportunity Investment Funds ICVC ( TOIF ) is an open-ended investment company structured as an umbrella company, incorporated in England and Wales, authorised and regulated in the UK by the Financial Conduct Authority (FCA) as a Non-UCITS scheme. Subscriptions to a Fund may only be made on the basis of the current Prospectus and the Key Investor Information Document, as well as the latest annual or interim reports and the applicable terms & conditions. Please refer to the Risk Factors section of the Prospectus for all risks applicable to investing in any fund and specifically this Fund. The above documents are available in English only and may be obtained free of charge on request from Columbia Threadneedle Investments at PO Box 10033, Chelmsford, Essex CM99 2AL.This material is for information only and does not constitute an offer or solicitation of an order to buy or sell any securities or other financial instruments, or to provide investment advice or services.this document is not investment, legal, tax, or accounting advice. Investors should consult with their own professional advisors for advice on any investment, legal, tax, or accounting issues relating to an investment with Columbia Threadneedle Investments. The analysis included in this document has been produced by Columbia Threadneedle Investments for its own investment management activities, may have been acted upon prior to publication and is made available here incidentally. Any opinions expressed are made as at the date of publication but are subject to change without notice and should not be seen as investment advice. This document includes forward looking statements, including projections of future economic and financial conditions. None of Columbia Threadneedle Investments, its directors, officers or employees make any representation, warranty, guaranty, or other assurance that any of these forward looking statements will prove to be accurate. Information obtained from external sources is believed to be reliable but its accuracy or completeness cannot be guaranteed. Issued by Threadneedle Asset Management Limited. Registered in England and Wales, Registered No , Cannon Place, 78 Cannon Street, London EC4N 6AG, United Kingdom. Authorised and regulated in the UK by the Financial Conduct Authority Columbia Threadneedle Investments is the global brand name of the Columbia and Threadneedle group of companies. columbiathreadneedle.com. Issued September 2018 Valid to end April 2019 Page 15 of 15