2017: Factor Performance in Review

Transcription

1 2017: Factor Performance in Review By Monty Joshi, CFA Portfolio Manager With the onset of the new year, readers are likely being inundated with reports on 2017 stock market performance. Among these year-end reports, factor-based updates have proliferated, in line with their increased popularity. However, it is important to remember that unlike a simple cap-weighted index, the performance of a factor is NOT a universal number, and can be significantly influenced by portfolio construction technique. To illustrate this, consider the spread between returns of the two largest ETFs (by assets) for each of four factors shown below (US Large Cap only): Table 1: Leading US Large Cap Factor ETFs ETF pairs with highest level of assets within each factor 1

2 The differences are remarkable particularly for the Quality and Momentum ETFs, and with yearover-year differences of over 100 bps for the others. In our view, there are two key determinants of a given factor portfolio s performance: 1) how one chooses to score the stocks for selection, and 2) how one chooses to weight the selected stocks. The decision on how to weight the chosen stocks often gets lost in the more intuitive, or glamorous, world of stock selection. To truly understand your investment, it is important to view both these dials in turn. At Optimal we generally focus on the four factors highlighted above: Momentum, Quality, Low Volatility, and Value. A fifth factor, Size, is introduced implicitly through our portfolio construction, specifically what we refer to as risk-mitigated weightings of holdings (aka smart beta), which generally tends to underweight the largest stocks (relative to their weights in a cap-weighted index). Size exposure is a secondary outcome; our primary objective of weighting this way is to spread the risk and maximize the diversification potential across the portfolio, resulting in lower volatility (specifics of our portfolio construction process can be found in our Factor Focus Volume I) in Review The S&P 500 index is, of course, a cap-weighted index. How did the size effect, via the market-cap weighting of the stocks within the index, drive its performance? To answer this, let s look at two funds: SPY (S&P 500 ETF) and RSP (S&P 500 Equal-Weight ETF). These are ETFs based on the same stocks, but in RSP s case the stocks are equal-weighted. Therefore, the only difference is weighting method: RSP underweights the larger stocks and overweights the smaller stocks of the universe, relative to SPY. So, this becomes a handy way to quantify the capweight effect within the large cap universe. 2

3 Table 2: RSP vs SPY Table 2 shows that SPY had unusual outperformance against RSP, to an extent, in fact, not seen since the financial crisis (in 2008, SPY outperformed RSP by 3.46%). This is a strong reversal of the multiyear trend coming into the year. To try and disentangle the effect of size from our assessment of factor performance, then, let s first look at last year s factor performance against the index when the stocks are cap-weighted (like the index), but retaining our standard stock selection techniques. The results are in Table 3, where the CW suffix signifies that the portfolios are cap-weighted. Table 3: Cap-Weighted Factor Performance* * See Disclaimer on last page regarding the calculations of factor returns shown Momentum and Quality cap-weighted factors performed very well. Low Volatility, the most defensive factor was, unsurprisingly, a laggard in the strong bull market. On the other hand, cap-weighted Value was (unusually) a laggard in a strong market cycle. How unusual? In every one of the prior six years in which the S&P 500 enjoyed a double-digit return, the cap-weighted Value factor outperformed it, and by a healthy average of 4.7% (see Table 5 below for performance by year). 3

4 Now, to better understand the impact of weightings on factor performance, let s look at the factors using our risk-mitigated (smart beta) weighting schemes. The OAM signifies that the factor portfolios here are weighted with our standard risk-mitigation smart beta techniques. Table 4: Smart-beta weighted Factor Performance* * See Disclaimer on last page regarding the calculations of factor returns shown Momentum still held up well, but the other smart-beta weighted factors could not outperform the extremely strong cap-weighted market index. In part, this is due to the implicit Size (i.e. small) exposure that comes from the weighting process, and which negatively impacted performance, and in part this is due to the very strong performance of the Market factor compared to LowVol and Value. Longer-term ( ) Over a longer-term period, for each factor except Momentum, the smart-beta weighted factor portfolio outperformed its cap-weighted counterpart. Furthermore, each of the smart-beta weighted factors outperformed the index. The table below summarizes calendar year results, as well as full-period, annualized results, for both the cap-weighted factors and the OAM-weighted factors: 4

5 Table 5: Long-Term Factor Performance (Cap-weighted and OAM-weighted)* * See Disclaimer on last page regarding the calculations of factor returns shown 5

6 Summary So, to highlight our key insights: Factor portfolio returns dispersion is very significant across vendors, even for the same factor it is essential to understand both the selection and the weighting process to really understand what you are investing in Even allowing for some dispersion across styles, Value was an unusual laggard in a bull market year While the market remained extremely strong for most of the year it is critically important to stay disciplined and remain diversified by harvesting all available factor premia, including recent laggards like Value and Low Vol. The long-term has borne out the value of smart beta in providing systematic and scientific risk mitigation as well as incidental exposure to Size factor premia 6

7 DISCLAIMER Factor performance statistics shown are back-tested only. No representation is being made that OAM will or is likely to achieve comparable performance results in the future to those shown above. In fact, there are frequently sharp differences between a hypothetical performance record and the actual performance record subsequently achieved by live trading. The investment process described in this document is subject to change at OAM s discretion, based on OAM s work with clients to create a suitable portfolio. Hypothetical, back-tested results were achieved by means of the retroactive application of a simulation model and, as such, the corresponding results have inherent limitations, including: 1. the results do not reflect the results of actual trading using client assets, 2. the results do not reflect the investment of dividends in the same way that an investor might, nor the deduction of transaction, custodial or advisory fees, the deduction of which would have the effect of decreasing model performance results, and 3. back-tested performance may not reflect the impact that any material market or economic factors might have had on the adviser s management of actual client assets. 7