There is No "Market": Risk Factors and Scenarios
We look at scenarios using risk factors. Why factors? Because there are few “pure” stocks.
What does inflation mean for the market? What does recession mean? We can’t say, because there is no “market”. It is like asking, “What will higher oil prices do to the world”. Do you mean Germany or Saudi Arabia? United Airlines or Exxon?
Instead, we look at scenarios using risk factors. Why factors? Because there are few “pure” stocks. Stocks are affected by many factors. Those that are within, say, the consumer durables sector might also be affected by energy, or by China. Or take Amazon. It is in retail, but has a high factor loading to technology because of its cloud services.
There is an array of risk factors — multidimensional across sectors, countries, and investment styles. Every scenario affects risk factors differently. Every asset is affected differently by each of these factors, and every portfolio — including any index — is built up lego-like from these.
So a better question to ask is: What risk factors will be most affected? And then ask, how exposed is my portfolio to these factors?
It is one of the great advantages of posing risks in factor space and then seeing how those factors feed through to assets, rather than trying to look at risk directly with assets. I’ve made this point elsewhere; it is central to how we develop scenarios at Fabric. It is why we look at our relationship with MSCI and our ability to use their Multi-asset factor model as being so valuable for risk assessment.
Using inflation as an example, we’d expect basic materials to be less affected by inflation than downstream product like retail. We’d also expect real estate and mortgages to be adversely affected because inflation leads to higher nominal rates.
We have the benefit of hindsight to see if such intuition is correct. The bar chart gives a sense of this using MSCI’s Multi-asset factor model that is at the core of our risk factor approach. Since the start of the year, we see that oil and gas, chemicals, and metals factors all do well. Real estate and mortgages, and retail and consumer durables factors, which are at the end of the supply chain, do poorly.
The result is different behavior for different stocks in the throes of the scenario. Look at these two ETFs, one for energy and one for consumer discretionary. Just as the factors vary, so do the stocks and the portfolios with high loadings in the various factors.
The chart is focused on industry factors; there also are country and style factors, and the story will be the same there. For example:
▪️Companies with high leverage will be in the negative territory. Leverage looks at long-term debt, but we can assume a need to roll it, and it will be done at a higher cost, both due to rates and credit spreads.
▪️Size is a positive. Higher size means more market power, and so costs can be kept down, and in some cases price can be raised. Low size means cost takers. So hurt more
▪️Value is a negative. This is a little iffy. My argument is that value often means brick and mortar, and a cost structure wedded to real factors of production. By contrast, high growth tends to be biased toward companies driven by intangibles.
▪️Some argue that growth is hurt because high rates discount earnings further out.
▪️Emerging markets is a positive. Another bit of a reach. But EM tend to be commodity producers, and the ultimate end of the supply chain is insulated from inflation.
The point is that scenarios don’t affect “the market”. Each scenario will affect different parts of the market, based on how these factors respond to the scenario and based on the stock’s loading on different risk factors.
Access a better way to understand and work with risk, powered by MSCI’s factor model.
Rick Bookstaber
CO-FOUNDER AND HEAD OF RISK
Rick Bookstaber has held chief risk officer roles at major institutions, most recently the pension and endowment of the University of California. He holds a Ph.D. from MIT.
Access a better way to understand and work with risk, powered by MSCI’s factor model.
© 2022 Fabric Risk