Portfolio Construction: The last free lunch
Robust portfolio construction has become the key to success in factor investing
- Changing economic and financial regimes can affect individual factor performance
- Practitioners are adopting multi-factor portfolio construction with risk-based weightings to harvest premia across market and economic cycles
- Diversification is the main way for investors to avoid risky factor-timing decisions
Take it to the bank: strategies pursuing the low-volatility factor premium tend to underperform the market in periods when interest rates are rising.
So with interest rates on the rise, it is fair to wonder when – and how deeply – that historical relationship might bite into a portfolio’s returns.
The answer: it may not matter much.
That is, not to institutional investors that are implementing factor investing the right way.
According to practitioners who discussed the state of the art in factor investing with IPE, the right way to capture risk premia is through diversification – the portfolio construction approach that random-walk theorist Burton Malkiel called the closest thing to a free lunch on Wall Street.
Factors work – but each factor’s performance varies in reaction to the underlying financial and economic regime that prevails at any given time. To avoid mono-factor volatility, practitioners advise institutions to use a multi-factor strategy with fixed factor weightings – one that is constructed after analysing an investor’s current portfolio to determine which factor exposures are embedded in existing holdings.
In short, robust portfolio construction has become the key to success in factor investing. That means building a well-diversified portfolio, stress-tested against drawdown scenarios to ensure risk exposures remain balanced during periods of market volatility. Practitioners also warn against over and underweighting factors in response to macro events – the potential return from such changes is not statistically significant, and is difficult to harvest in the systematic fashion required to capture factor premia.
Systematic implementation is the key feature of factor strategy, says Stephan Kessler, lead portfolio manager, alternative investment strategies, at Goldman Sachs Asset Management. “Factors are generally embedded in the return patterns of financial markets,” says Kessler. “That means by and large we would not override factor behaviour depending on macroeconomic views or the performance of individual markets, equities or sectors.
“The current rate hikes are not likely to have much impact on equities, in which most factor investing occurs, as long as the increase occurs at a moderate pace,” Kessler says. “There’s a lot of research based on historical data that shows that rate hikes during a strong economy shouldn’t cause stock market crashes. The main impact of the rate hike will play out in fixed-income factor strategies, and to some extent in FX factor strategies.”
Kessler says that GSAM looks at the factor world mainly in terms of carry, value, momentum and structural themes. “Each of these themes carries a return premium and has a certain premise,” he says. An FX carry strategy, for example, seeks to “harvest the interest rate differential between countries” by buying currencies of high-yielding countries and selling currencies of low-yielding countries. “The larger the difference between countries, the more opportunity you have,” he adds. With the US raising rates ahead of other major central banks “we believe carry is likely to benefit from the larger dispersion in rates that we are going to see.
“Rising rates are also likely to present opportunities in the momentum premium in fixed-income and forex markets, which tend to benefit from trends in rates,” Kessler says. “Central banks move in cycles, and once they’ve started hiking the tendency is to continue to move in that direction for a prolonged period of months or sometimes years. Factor strategies in forex and fixed income can harvest those directional opportunities, as well as the term structure opportunity that arises as the yield curve is expected to flatten when central banks start to hike rates. That’s an attractive result of this changing regime.”
While the regime change will benefit some factors more than others, this is not an argument to trade in and out of those factors. “We don’t tend to time strategies,” Kessler says. “When we put our portfolio together we ensure it is very balanced and well-diversified across a broad set of factor strategies. In our portfolio we’ve got an FX carry strategy, an FX value strategy and an FX trend strategy.
“Each will tend to do a bit better or a bit worse depending on the macro environment, but we do not over or underweight any of the factors. While we do find certain tendencies on how strategies react to the macroeconomic environment,” Kessler says, “it’s hard to harvest these systematically and significantly in terms of returns. What’s more, each factor’s performance plays out over a long period of time and, as far as we see, the links between the macroeconomic environment and weightings have not been statistically significant. So we continue to have a very balanced portfolio across strategies.”
In short, reacting to fluctuations in economic conditions distracts from the task at hand – harvesting the premium available from each factor. Individual investment styles and factor premia can be out of favour for long periods, says Jason Williams, a fund manager at Lazard Asset Management. “But making a bet on when one country or sector will reverse course and outperform any other is very difficult to do. Investors are much better off neutralising such exposures and focusing on investment anomalies that can systematically be targeted on a consistent basis.”
Lazard builds factor portfolios “in which active returns are not driven by macro exposure,” Williams says, instead targeting risk premia from sentiment, which includes momentum, value, growth and quality. “We are seeking to own stocks that represent in some way every one of those factors,” he says. As a result, “we’re well diversified and it’s a question of getting the averages right”.
The resulting strategy targets tracking error of 2% to 4%, with an objective of exceeding market returns by about 200 basis points on an annualised basis, towards the top rung of active equity managers on a five-year basis. The low tracking error – “almost benchmark-like,” according to Williams – reflects the portfolio’s consistent exposure to “investment anomalies where we think there’s good underlying economic rationale and we have techniques to extract alpha”.
Williams contends that asking whether or not to underweight low-volatility strategies now that US rates are rising is the wrong question. “That seems to come from a perspective of looking at individual strategies, and you shouldn’t be looking at factors that way in the first place.” Low-volatility strategies are today’s systematic way of creating what was previously called defensive equity investing, Williams says. Such stocks became bond proxies under extreme low interest rate regimes, and many “developed quite a substantial premium to the market as they came into favour”. But with synchronised global growth, the low-vol factor has lagged, he notes, with the MSCI minimum volatility index trailing the market by about 20% since the Brexit referendum and consumer staples, “the poster boy of defensive equity,” trailing by 27%.
That has brought the relative valuation of low-vol defensive equities back in line. While the prospects for low-vol strategies in a rising rate environment have recently attracted a great deal of attention, “what’s snuck up on investors is the enormous outperformance of growth stocks,” he says. “They’ve rarely been more expensive versus the market – attention now should be focused on the valuation of growth stocks.”
Capturing shifts in the relative performance of risk premia without resorting to in-or-out timing marks an innovation in the use of factors. “Multi-factor configurations are the next generation of factor and smart-beta investing,” says Koen van de Male, global head of investment solutions at Candriam Investors Group. While individual factors work over the very long run, trying to anticipate their performance cycles is “very dangerous because the factors individually will have huge drawdowns,” he adds. A better approach is risk-based weighting, but “within very narrow boundaries because done in an aggressive way the danger is that you end up with a mono-factor exposure and lose the diversification benefit when you absolutely need it”.
The new interest-rate regime is a case in point. The stocks comprising the low-volatility and quality factors, which are vulnerable to lagging the market when rates rise, have already adjusted. “Yields have already risen, quite a lot in the US, and you cannot say those factors are overly expensive today,” says van de Male. “It’s a bit too late to underweight these factors if one were using dynamic weighting.”
Van de Male recommends investors maintain a balanced multi-factor portfolio. “I wonder if it’s even worth the effort to try to time a little bit, because the implementation is quite complicated and the room to improve performance is very limited,” he says. “We stick with risk-based fixed weights for factors.It’s important in a multi-factor portfolio that each factor has more or less equal impact.”
There’s still a role for single-factor strategies for institutional investors – to complete or rebalance the risk allocation of a well-diversified portfolio of factor premia, says Yaz Romahi, chief investment officer of the JP Morgan Asset Management quantitative beta strategies (QBS) team. “It’s possible that in the last two years your portfolio became more exposed to growth,” he says. “if you never thought of your portfolio as a factor portfolio, this can help balance that exposure.”
Romahi also expects to see more long-short factor offerings. The original research on factors contemplated the use of long-short strategies to capture factor premia over long periods, he says, for example to be systematically long value stocks and short expensive stocks.
Ultimately, the challenge facing investors remains the same, says van de Male. “We need to identify which regime we are in today and which regime we’re likely to face tomorrow,” he says. It may be tempting to take a view on what the future holds, “but then you lose diversification and if you miss the regime in which we will operate tomorrow, you might end up with some problems.”
For institutional investors, then, the message is clear: stay diversified, and lunch is free.
Sensitivity training: low-vol strategies and rising rates
Investors are asking whether rising interest rates mean it is time to abandon low-vol strategies.
Identified by academics in the 1970s, the low-volatility anomaly contradicts a core principle of finance, that higher risk must be rewarded by higher returns.
Because low-vol stocks are typically dividend-paying companies that often carry high debt loads in sectors like utilities, investors turned to low-vol equities to offset the loss of bond income during the past decade’s zero-rate monetary regime.
But the search for yield pushed the prices of low-vol stocks to a substantial premium to the market, mitigating the benefit of investing in the factor, and now rising rates are fuelling concerns that low-vol strategies may not continue to deliver consistent performance.
Doubts centre on the interest-rate sensitivity of the stocks that comprise portfolios built to capture the low-vol anomaly. Amundi Asset Management says it has found a way to mitigate interest-rate exposure while maintaining most of the benefit of a low-vol factor strategy.
“Low-vol strategies are still appealing when interest rates are rising if we integrate interest-rate sensitivity into the portfolio,” says Bruno Taillardat, global head of smart beta and factor investing at Amundi. “The low-vol factor tends to react negatively when interest rates are moving up because the stocks are drawn from defensive sectors such as telecom and utilities that tend to be more leveraged.”
The characteristics of low-vol stocks that lead to underperformance as rates rise are what allow the low-vol anomaly to outperform over a full economic cycle. In a recent research paper, Amundi found a way to preserve those beneficial effects by making small changes in stock weights that create a low-vol portfolio with an interest-rate sensitivity that matches the benchmark, in this case the S&P 500. Weighting changes are determined by a systematic methodology that accounts for the effects of changes in the level of interest rates and the slope of the yield curve on stocks in the low-vol universe.
But there is a trade-off. As the figure shows, when the interest-rate sensitivity of the low-vol portfolio declines, investors forego some of the low-vol effect and accept higher volatility. But the change is small, Amundi writes, and “does not seem to disrupt the characteristics of the low-volatility strategy in a significant way”.
Perhaps the best solution is to sidestep the problems that arise when investing in a single factor. That is what Amundi has done with the recent launch of three new equity multi-factor funds that use a dynamic allocation approach to maintain an equal risk contribution from each factor.
“In a multi-factor framework, you don’t need to integrate sensitivity of interest rates to low-vol factors,” says Taillardat. “By definition, when you combine the low-vol factor with a value factor and a momentum factor, which can react very positively to interest rates, you diversify, and interest-rate sensitivity is embedded.”