While the market seems to be ignoring bad economic news—inflation a case in point—it is vital to be aware of the potential impact of such shifts. Stock sensitivities to major economic variables, and the correlations between them, can help reveal a portfolio’s vulnerabilities.
In this article we examine stress-test results to determine the ex-ante performance impact of the variables in the Axioma Worldwide Macroeconomic Projection Equity Factor Risk Model (Macro Model). We stressed several of the economic variables in the model to see what would happen to the STOXX® USA 500 index, as well as the active performance of single- and multi-factor style indices built from that parent index (in this case the US-based industry-neutral variations).
We found that several of the single-factor indices could be expected to outperform or underperform the underlying index by more than 1%, with the Multi-Factor version experiencing the smallest average expected active impact, adding to the list of its diversification benefits.
This is the fourth in a series of blog posts highlighting the new Axioma Macroeconomic Projection Equity Factor Risk model. The first focused on economic exposures and their impact on return over the full model history for a broad market benchmark. The next post took a shorter-term view, analyzing the macroeconomic contribution to the disappointing active performance of the STOXX Global 1800 AX Low Risk Index for the first couple months of 2021. And the third post compared the economic exposures of various style factor portfolios, how they evolved over time, and their impact on performance.
We start with macroeconomic exposures…
Exhibit 1 shows the active exposures of the six style indices, along with those of the STOXX USA 500, to all the variables in the Macro Model. These exposures can also be thought of as betas or sensitivities, so that, for example, a 1% increase in the US Term Spread corresponds to a 3.49% increase in the USA 500 index, all other things being equal. We have bolded the variables we stress tested but have included exposures for all 14 of the variables in the model. We have also highlighted the style index with the highest exposure (in green) and the lowest (in pink). We see that for all economic variables some indices have positive exposures and others negative, implying very different impacts, depending on the nature of the style factor portfolio. For example, the Size portfolio (which favors smaller names) is likely to be hurt by an increase in the US BBB corporate spread, whereas the exposure suggests this would be a positive for the Low Risk portfolio. The intuition makes sense, as higher spreads, meaning higher financing costs, are likely to hurt smaller names, but they may also drive investors to lower risk stocks.
Exhibit 1. Macroeconomic Factor Exposures on May 31, 2021
Now we test the potential impact on market return…
One of the basic tenets of stress tests is that the magnitude of the stress should be big enough to be meaningful, but the move should be plausible. For our economic factor tests, we started with values that were roughly three standard deviations above or below the average monthly move, and then chose the closest round number for ease of exposition. We tested eight of the 14 factors using Axioma Portfolio AnalyticsTM. The stress tests use the exposures and the correlation structure that underlies the risk model as of the test date, in this case May 31, 2021. The output (expected total returns for the market and expected active returns for the style portfolios) does not take into account specific risk in the indices, which in some cases is significant.
Exhibit 2 shows the test results for the STOXX USA 500 index. The widening of credit spreads in the US, and even more so in Europe, is associated with falling stock prices. Interestingly, some of our other stressors—which could be viewed as potentially negative for equity markets—would actually be positive for equities at this point in time. The index has a positive exposure to inflation, as it has since the end of 2007. Similarly, an increase in non-oil commodity prices would also boost share prices, while higher oil prices would have little impact.
Exhibit 2. Stress Test Results, STOXX USA 500
And lastly we look at the active return of style indices…
We ran the same tests on the STOXX USA 500 series of industry-neutral factor portfolios, which are based on the STOXX USA 500 index. In Exhibit 3 we show the expected impact of each of our stressors in active return space.
The charts have the same range for the X axis, so we quickly see that the Low Risk index, which tends to have the highest magnitude exposures to the macro variables of any of the six indices, is also the most vulnerable, with potentially large moves related to several of our test variables. Among the style indices, Low Risk has the highest negative sensitivity to US inflation and is therefore likely to see the biggest shortfall. In addition, although the sensitivity to EU inflation is slightly positive, the correlation of EU inflation to other model factors (most likely US inflation) filters down to the negative return for the index. In contrast, the Low Risk portfolio’s positive exposure to credit spreads suggest it is likely to outperform should they widen, reflecting the index’s defensive nature.
Whereas Low Risk has had a negative sensitivity to US inflation for most of the history of the Macro Model, Momentum’s exposure has usually been fairly close to zero, albeit on the positive side of zero. The current reading, however, is just below the historical high level of sensitivity, and is currently much higher in magnitude than any of the other indices except Low Risk. The index should benefit handsomely, should inflation increase by 1%. Momentum also has higher magnitude of exposures to a number of other factors. As for US inflation, many of them are at unusually high or low levels, meaning the reaction to stress would be bigger than normal.
Quality, also a somewhat defensive strategy, sees expected reactions in the same direction as Low Risk, but of somewhat lower magnitude.
Size was relatively less sensitive to most of the economic factors, with the stark exception of the US BBB corporate spread. Still, given its other low exposures and the correlation structure, even a big move in the spread would be expected to have a relatively muted impact. Most of the stresses would hurt Small relative to the parent index, but rising commodity prices or inflation in general would help it.
Value would be helped most from higher inflation (as compared with the impact of other factors, not in contrast to other style indices), but as we see for Size, the magnitudes of expected active returns are relatively small. The sensitivity of Value to the term spread variables in particular is surprisingly low, and indeed it is currently much closer to zero than usual.
Finally, the magnitude of the expected impact on active risk for our stress tests is lower for the Multi Factor index than it is for any of the other indices we tested. This result is not surprising, as one of the intents of the Multi Factor index is to diversify away some of the risks of the individual factors. This diversification of style factors also diversifies the economic exposures, and also pushes more risk away from factors and into specific risk. As we noted earlier, the stress tests only use the factor exposures to determine the potential impact.
Exhibit 3. Expected Active Returns Under Selected Stress Scenarios
Stress tests are another arrow in the risk manager’s quiver, designed to inform a portfolio or risk manager where the portfolio may be vulnerable beyond the traditional factors in the risk model. Having a macroeconomic view that takes into account not only exposures but also correlations to paint a picture of what might happen to the portfolio “if…” can prompt action, or at least help explain performance. Currently, with news of higher interest rates, inflation, oil prices, etc. dominating the headlines, it might be just the time to run some tests on your active portfolio.
For more insights and research from the Applied Research team, please click here.
 Three of the economic variables—Commodity, Gold and Oil—are much more volatile than the other economic factors and, because of that, their exposures tend to be much lower. For example, a 10% move in the oil price is relatively routine, whereas even a 50-basis point change in the corporate bond spread is relatively rare.
 Those that have been the biggest focus of investor interest in the last few months.
 The Axioma Risk portfolio enables the tester to construct more sophisticated tests, in particular ones in which the user can choose alternate periods to construct the correlation matrix. As we wanted to use current correlations, the Axioma Portfolio Analytics platform met our relatively simpler needs for these tests.
 This has generally been a period of low inflation, which is typically positive for equities. Based on observations prior to the start of the model in 2004, we believe that high inflation is negative for stocks, as companies cannot raise prices to keep pace with their increasing costs, and as central banks have to tighten.