Estimating future returns on equities has been somewhat vexing in recent times given the impact of share buybacks. Share buybacks by companies in the S&P 500 have exceed the payout in dividends for 12 of the past 16 years (2000-2016). The challenge is separating what cash return is received now and what the growth of those cash returns will be in the future. If we were to only to use historical returns it wouldn’t matter the mix, but when forecasting based on projected payouts, growth in payouts, and current valuation levels, it becomes very valuable to estimate the different drivers of returns.
In the latest Financial Analyst Journal (Q3 2017) I was excited to see authors Straehl and Ibbotson 1 tackle this challenge by attempting to formalize forecasting models that are independent of the payout policy versus traditional models where buybacks are implicitly reflected in the growth estimate. They extended their work by comparing the results of stock returns with the real economy to verify the model’s macro-consistency and therefore usefulness to practitioners.
For instance the long-term real return on equities from 1970 to 2014 was ~6.1%. However to incorporate a view on current yields (and thus valuations) and keeping growth rates macro-consistent, i.e. growing no faster than the economy – a projected return will likely be lower than the historical rate. Our planning software’s current projected return expectations are currently 4.5% for US large cap equity.
The authors created two models independent of payout reflecting two types of hypothetical investors: a pro rata buyback investor and a cap-weighted investor. The pro rata buyback investor, who receives cash for participating in the buyback, fits closest with our own cash flow valuation methodology for the S&P 500 in terms of estimating payouts and fundamental growth based on earnings retained and not paid out. Using the pro rata buyback investor model, the long-run expected real return (i.e. net of inflation) based on prices as of December 2014 was 4.87%, which included a 3.2% total payout yield and growth of 1.67%. In 2013 our planning software’s default projected rate was 5.0% for blended US large cap compared to the 4.5% rate estimated today.
Who knows if others will join in working on formal solutions to isolate buybacks from growth but at least we aren’t alone in attempting to separate them when determining the drivers of a forecast. Furthermore, it was nice to see projected returns based on the pro rata buyback investor model line up closely with both our planning software’s projected return estimates as well as our own work on valuing the S&P 500, which uses both total payout (dividends and share buybacks) yield and growth in payouts.
- Straehl, Phillip U. and Ibbotson, Roger G., The Long-Run Drivers of Stock Returns: Total Payouts and the Real Economy, Financial Analyst Journal, Third Quarter 2017, 32-52 [return]