Join our community of smart investors

When betas mislead

The idea of a share's beta is misleading: a share's response to a rise in the aggregate market varies depending on why the market has risen.
February 20, 2020

I wrote recently that a handful of lead indicators has successfully predicted annual changes in the All-share index. Which poses the question: do these indicators also help predict returns on specific sectors? The answer is yes, to some extent.

My table shows the sectors most and least sensitive to the four most important lead indicators for the aggregate market.

The first column shows the sectors that do best after the market has been cheap, in the sense that the dividend yield on the All-share index has been high. These include engineers, life insurers and IT. The weakest responders to a cheap market tend to be defensive stocks – with the notable exception of construction stocks. These rise least after the market has been cheap, and fall least after it has been dear.

Sectors' response to lead indicators  
ValuationsSentimentPortfolio rebalancingMomentum
WinnersWinnersWinnersWinners
EngineersITEngineersElectronics
ITGeneral financialsTransportTelecoms
Life insurersElectronicsITUtilities
TransportMediaGeneral retailersNon-life insurers
General financialsTravelTravelChemicals
LosersLosersLosersLosers
TobaccoTobaccoNon-life insurersGeneral retailers
UtilitiesFood retailersPharmaceuticalsMiners
Oil & gasUtilitiesFood retailersEngineers
ConstructionFood producersOil & gasIT
Food retailersChemicalsUtilitiesBanks
Based on annual changes since 1996  

The sectors most sensitive to overseas investors’ sentiment (as measured by foreigners’ net buying of US equities) include IT again, but also general financials and electronics. These are the sectors that do best after sentiment has been depressed and worst after it has been euphoric. Again, the sectors least sensitive to sentiment have been defensive ones.

A third predictor of aggregate returns has been the ratio of the money stock in OECD countries to the MSCI world index. When investors have lots of cash and few equities they are likely to rebalance their portfolios towards equities, causing the market to rise. The main beneficiaries of this process have been engineers and IT again, but also general retailers. Stocks that under-perform again tend to be defensive ones.

Finally, we have momentum, as measured by the ratio of the All-Share index to its ten-month average. Here, the picture is different. Sectors that do best when the market has momentum include telecoms and utilities, whilst miners, engineers and IT stocks have tended to lag behind.

The fact that different sectors are most responsive to these different factors tells us something significant – that our four main lead indicators capture different aspects of market dynamics. A rise in the market following a period of under-valuation has different winners and losers from a rise driven by momentum, for example.

Innocuous as it seems, this fact undermines one of the key concepts in finance – that of beta, in the sense of a stock’s sensitivity to moves in the overall market.

True, high-beta stocks such as IT and engineers have tended to do best during market rallies caused by the index’s prior cheapness or by global investors rebalancing their portfolios towards equities. But they have tended to lag behind during momentum-driven rallies. The high-beta mining sector is actually not sensitive to the market’s cheapness or to foreign investors’ sentiment. And the high-beta construction sector is actually less sensitive than most others to the aggregate market being misvalued.

Whether a stock rises a lot or a little when the aggregate market rises depends on why the market goes up. To understand equities’ response to a rise in the aggregate market requires us to know why the aggregate market has risen. A rally caused by a correction to past misvaluations, for example, has different effects to one caused by momentum. A share’s beta is a measure of its average move when the market moves. But there is always variation around an average.

All this should remind us of the wisdom of a point often made by Scott Sumner at George Mason University: never reason from a price change.

Which brings us to a problem. Collectively, our lead indicators are pointing to only a moderate rise in the All-Share index this year, because they are telling different stories. While the high dividend yield and recent foreign selling of US stocks are bullish indicators, the fact that the global money-price ratio is low is a worrying sign.

What we would like ideally, therefore, are stocks that are responsive to low aggregate valuations and to weak global investor sentiment, but not responsive to the money-price ratio. Only one sector fits this bill: non-life insurers.

Which tells us something. All sectors are to some extent correlated with the overall market: if equities generally do well most rise, and if the aggregate market slumps most equities fall, albeit to different extents. True, beta can be misleading if we rely on a precise estimate of it. But the fact that betas are usually greater than zero is a useful fact. Which means that there are severe limits to how much you can spread risk by using equities alone.