Continue active refreshing of this index's data?

Continue active refreshing of this index's data?

Qontigo Insight Quarterly — April 12, 2023

Qontigo Insight Q1 2023 Risk Review

by Applied Research Team

Top Takeaways:

  • Although it seemed as if the potential banking crisis sent volatility soaring, that was not the case, as predicted volatility fell in most country and regional benchmarks. Industry risk increased, but other determinants of risk fell
  • In the US, index and active risk according to the statistical model exceeded that of the fundamental model, suggesting an unidentified risk “bubbling under the surface”
  • Trading volume remained low through most of the quarter, driven higher only by increased trading in financials
  • From the standpoint of the macroeconomic projection model, Term Spread factors had highly negative returns, but the Global index’s negative exposure to the EU Term Spread meant a significantly positive contribution. Overall, macro factors contributed substantially to Q1 returns until the last few weeks, when specific returns took over
  • Asset dispersion declined in Q1, but we show that is not necessarily bad
  • Factor returns were high in magnitude around the March crisis, but settled down and were generally in line with expectations in Q1. Momentum and Dividend Yield had very tough quarters in the US, not quite as bad elsewhere
  • In the Global model, Value fared better among smaller names and on the short side
  • Dividend Yield and Volatility exposures became negatively correlated with Momentum, while Momentum’s risk model correlation with Earnings Yield increased

Short-horizon predicted volatility

Source: FTSE Russell, Qontigo
  • The STOXX USA 900 remains among the highest risk of all the major indices we track closely, while Canada and Australia have the lowest risk
  • The UK and Japan were the least volatile at the end of last year; they both saw risk increase in Q1 whereas it fell in all other major regions
  • The Global 1800, Global 1800 ex-USA, Global Total Market and USA 900 all saw risk fall more than 15% proportionally

STOXX USA 900 – Short Horizon vs. Trading Horizon forecasts

  • The trading horizon model risk forecast was below the short horizon for much of the last 12 months
  • However, it reacted much more quickly when SVB collapsed, and investors feared contagion
  • When markets calmed down, the trading horizon model reverted back quickly as well

Statistical-fundamental spread: Is something bubbling under the surface?

  • The spread between the statistical and fundamental risk forecasts for the STOXX USA 900 was about 2% by quarter end, a 95th percentile score relative to the past 10 years and the highest level in a year and a half
  • No other major region has a positive spread
  • The statistical model may be detecting a risk not seen by the fundamental model

STOXX USA 900 Ax Value Index: Risk spreads apparent in active space as well

Trading volume has ticked up, but only because of Financials

Source: Qontigo
Source: Qontigo
  • The global market’s move up has been on relatively low volume, although it has increased recently
  • The increase in trading volume recently, however, has been entirely the result of understandably higher volume in Financials

Macro factor returns

Source: Qontigo
*Normalized: (Actual Return – Long-Term Average)/Realized Standard Deviation
**Rank: Percentile relative to long-term history. Bold means top or bottom decile
  • Term Spread returns across regions were among the most negative as compared with the model’s historical quarterly returns, with EU and GB Term spread returns at least two standard deviations below average
  • GB Inflation also produced quite a negative return

Macro factor influence was substantial in Q1

  • For most of Q1, macro factors were substantial drivers of index performance
  • Interestingly, stock specific contribution took over after the SVB failure and its spillover

Q1 2023 performance attribution, STOXX Global 1800

  • Most notable was the positive return generated by the negative exposure to US BBB Spreads; contribution reversed sharply at the outset of the bank crisis but did start to recover at the end of the month
  • The negative exposure to the EU BBB Spread started the quarter with a positive contribution, but the contribution turned negative in the second half of the quarter
  • Negative exposure to EU Term Spreads and a positive exposure to Gold had a positive impact on index return
  • Positive exposure to oil started to impact returns negatively in March, but bounced back
  • Contact your Qontigo representative for similar analysis on other indices or different time periods

Average monthly asset-return dispersion*

Source: Qontigo
  • Dispersion is a function of correlation (the higher the correlation the lower the dispersion) and volatility (the lower the volatility the lower the dispersion)
  • Average dispersion backed off the levels of Q4, and were generally below average in Q1 – the US is a notable exception
  • High dispersion is a potential advantage for managers who could successfully distinguish between the winners and losers, as the reward to “being right” is higher than average (but so is the penalty to “being wrong”)

*Cross-sectional standard deviation of monthly returns for stocks included in each benchmark

Another note on dispersion

Source: Qontigo

Dividing history into times when dispersion in the US was higher than 10% vs when it was lower shows that only Liquidity and Size returns had significantly different returns (in the statistical sense) by environment

Most factors’ returns remained within a 2-standard-deviation range, but Momentum and Dividend Yield struggled

Source: Qontigo
Note: an “*” indicates that the return was two standard deviations or more away from the long-term average, with the standard deviation defined as the risk forecast at the beginning of the period. The highest and lowest regional return for each model is highlighted.

A closer look at US Style returns US4 and US5 Trading Model

  • Dividend Yield had a very tough quarter, as did Momentum
  • Low Volatility but High Market Sensitivity were better bets in Q1
  • Exchange Rate Sensitivity had a big return, and may have impacted portfolio returns more than expected
  • Returns from the trading horizon model do not stand out as particularly high in magnitude

*Normalized: (Actual Return – Long-Term Average)/Realized Standard Deviation
**Rank: Percentile relative to long-term history. Bold means top or bottom decile

Red denotes typically compensated factors with big returns opposite to the long-term average
Source: Qontigo

US Factor exposure correlations

  • Dividend Yield and Momentum, and Volatility and Momentum are now negatively correlated – Dividend Yield has lost its momentum, and low Volatility gained momentum
  • Momentum and Value have become more negatively correlated over the past two quarters, so Value seems to be losing momentum as well
  • Most other factor pairs’ correlations (shown and not shown) have not changed substantially

Global factor alternative portfolios

Source: Qontigo

Notes: Returns are scaled to a factor exposure of 1 for comparability. Differences over 1% for the quarter and 2% YTD are highlighted. All Cap portfolios use the All Cap Global Developed Axioma Market Portfolio as the investment universe, and Large Cap uses the corresponding Large Cap portfolio.

We created variations of factor-mimicking portfolios with exposure to the factor in questions and no exposure to any other factor:

  • Monthly rebalancing rather than daily
  • Large-cap universe rather than all-cap
  • Shorting only allowed up to the benchmark weight (“Long Only”)
  • Long-only and large-cap universe constraints had substantial impacts in Q1 for several factors, most notably Value, which seems to have worked better on the short side and in smaller names
  • In contrast, Earnings Yield did better in Large Cap and on the long side
  • Low Volatility also worked better on the short side, but had better returns in large cap