In a recent research note we showed that under low interest dynamics valuation could drastically increase and we used Microsoft as an example. We hypothesized that aggressive valuations did not only require low interest rates but also predictable earnings so that the stock mimics a bond income stream. In today’s research note we show that higher earnings predictability increases the forward valuation premium in the equity market, and we show how the S&P 500 has a higher frequency of companies with high degree of earnings predictability.
Highest earnings predictability comes with 30% valuation premium
If we combine all the companies in the S&P 500 with those in the STOXX 600 index and group them by their earnings predictability (R-squared of EPS – definition is explained further down) then an interesting pattern emerges. Higher earnings predictability is rewarded by higher forward valuation. The 10% of companies with the highest earnings predictability have a 12-month forward EV/EBITDA metric that is 30% above the overall US equity market. This means that earnings predictability is a quality factor that explains some of the variance in valuations, but other factors such as business models, geographic footprint, earnings growth and return on invested capital may also help explaining the variance in valuations.
Our future analysis of earnings predictability will look into whether earnings predictability drives future returns across equities.
S&P 500 companies have higher earnings predictability
If we plot the relative frequency of R-squared brackets of 2%-points, then we clearly see that the S&P 500 has a higher degree of companies with high earnings predictability compared to the STOXX 600 index. This is driven by the higher concentration of information technology companies in the S&P 500 which are dominated by software companies. Higher earnings predictability is a defensive characteristic as the certainty over futures earnings goes up and this can explain why investors have preferred US companies over European companies. If we combine the higher earnings predictability of S&P 500 with the fact of higher earnings growth over the past 10 years, then we have the recipe for the enormous outperformance of US equities over European equities.
Accenture vs Adobe
Earnings predictability can explain some of the variance in valuations but within each decile of R-squared we observe large differences. Accenture and Adobe have R-squared metrics of 0.993 and 0.983 respectively but Accenture is only valued at 17.6x on 12-month forward EV/EBITDA whereas Adobe is valued at 31.4x on the same valuation metric. While both valuation metrics are far higher than the average of the equity market their large difference indicates that other crucial factors can explain valuation. If we look at earnings growth, we observe that Adobe’s earnings have grown by 17.5% annualized since 2011 whereas Accenture has grown earnings by around 8% annualized. In other words, a high earnings predictability is not worth as much if the growth rate is 3% as for utilities compared to a highly scalable software company such as Adobe.
The table above shows the ten largest companies in the S&P 500 and STOXX 600, and their respective earnings predictability. The market cap weighted average shows that the largest companies in the S&P 500 have a much higher earnings predictability than the largest companies in the STOXX 600 index. This difference is largely explained by the US technology companies such as Apple, Facebook and Google.
How is earnings predictability defined?
The earnings predictability is defined by the R-squared metric which describes the fraction of variance in the dependent variable that can be explained by the independent variable. In this case the dependent variable is the earnings per share from continued operations over the past 20 quarters, 10 semi-annuals or 5 yearly earnings per share numbers depending on the reporting frequency. The independent variable is time. The metric goes from zero to one where one indicates that EPS fits perfectly linear over time and thus if the relationship holds future EPS is easily predictable from the past. In this analysis we removed financials as the valuation metric EV/EBITDA is not measurable for this sector.