Why is the US risk contribution suddenly so big?
In our recent Q4 2018 Risk Review webinar and research paper, we noted a sharp increase in the US’s contribution to overall risk in global developed-markets equity benchmarks. Between the third quarter of 2016 and the final 3 months of 2018, the percentage contribution from US stocks to FTSE Developed short-horizon risk rose from 47% to 66%, while the market-value weight remained fairly stable around 58%.
Part of this can be explained by a recent surge in predicted volatility for US stock indices, which ranked among the riskiest of the major markets in Q4 2018. This is in stark contrast to most of the previous quarters, in which US large-cap stocks were among the least volatile. Between October and December alone, short-horizon risk for the Russell 1000 shot up 13 percentage points from 8% to 21%—a more than 2.5 times increase and one of the biggest jumps since at least 1982.
However, the higher standalone volatility only explains part of the increased share of overall risk, as other markets also saw their predicted volatility rise. In fact, our analysis from a multi-asset class perspective showed that this was to a large part a currency effect. In the second half of 2016—the run-up and immediate aftermath of the US presidential election—the dollar had exhibited a very high correlation with the US stock market. This is a common phenomenon during periods of political uncertainty. We observed similar co-movements of share prices and the euro around the French presidential campaign in spring 2017 or the Italian coalition talks in mid-2018.
When a region’s currency is very highly correlated with its stock market, gains in foreign investments will be, at least in part, offset by FX losses. While this is detrimental to the total portfolio return, the inverse relationship between equity and exchange rate returns also results in a significant overall risk reduction.
We noted this peculiarity in our FX Trumps Correlation whitepaper from January 2017. At the time, Axioma’s global multi-asset class portfolio showed an unusually low percentage risk contribution of only 12% from US equity holdings, despite a corresponding market-value weight of 30%. Our analysis revealed that this was due to a strong positive relationship between the stock market, government bond yields and the value of the US dollar, all of which were boosted by the expectation of lower taxes and increased infrastructure spending under the newly inaugurated Trump administration.
The graphs below show the volatility contributions from the various asset classes in the model portfolio and the correlations between the major risk factor types driving their returns as of Nov. 25, 2016. The correlation matrix on the right-hand side highlights a strong inverse relationship between equities on the one hand and FX and interest rate returns on the other. The calculations are based on security returns. Therefore, a negative number for equity versus interest rates means that bond prices go down as share prices rise and vice versa. The negative value for the FX/equity relationship implies that foreign currencies depreciated against the USD as stock markets went up. Thus, the portfolio derived a lot of diversification from being invested in fixed-income and non-USD securities. Or conversely, investing more money into the US stock market would have improved both the risk and return characteristics of the portfolio.
Two years later, the picture is completely different. While government bonds still provide diversification benefits in the current flight-to-quality environment, exchange rates are now almost uncorrelated to the share prices. This is primarily due to the constantly changing focus of FX markets from geopolitical issues, such as the China-US trade war, which tend to adversely affect the USD, to the idiosyncrasies of Brexit and other government crises in Continental European countries. While this at least results in a lower risk contribution from non-US equities, it also means the majority of the volatility in the portfolio comes from US stocks. The latter now account for 63% of total risk of the multi-asset class portfolio—more than twice their monetary share of 30%, as can be seen on the left-hand side graph below, which shows risk contributions as of Jan. 18, 2019.
In the current climate of political uncertainty, it is hard to predict which way cross-asset correlations are going to turn next. If the focus remains on Europe, the euro is more likely to move in line with the stock market. For investors outside the single currency, this means increased equity volatility, while their European colleagues could reduce their overall risk by diversifying into other regions.
US investors, on the other hand, could experience similar diversification benefits, if market participants were to turn their attention back to geopolitical risks, for instance, in the case of a renewed deterioration of the trade war between the US and China. In such an environment, safe-haven assets, such as the government bonds, gold, or the Japanese yen and Swiss franc, could help reduce both portfolio risk and losses. The inverse relationship between share prices and exchange rates against the USD would also likely lower the risk contribution from US equities within global benchmark indices.
For regular updates on the current risk environment and cross-asset correlations, subscribe here to Axioma’s weekly Risk Monitor newsletters.