What happens when you over-account for dividends in investment decisions

By Jeffrey Cohen, President and Founder
​US Advanced Computing Infrastructure, Inc.

We have traditionally given stocks 100% (this is a variable setting) credit for dividends paid. This means we assume that the day a stock no longer pays a dividend, also known as the ex-dividend date, its stock price will fall by the value of the dividend (ex-dividend date). Since our model is looking to find expected returns for stocks, we 'add back' the value of the dividend so the expected return is included.

During a time of stable pricing, low interest rates, and when dividends are scarce, this is a necessary adjustment. The difference between a stock paying a 2% dividend and one paying zero is immaterial, but imagine the master limited partnership (MLP) or the Real Estate Investment Trust (REIT) that could be paying 7% or 10% during a time could be very material.

So, we gave back 100% of the value of the dividend to the 'future expected return' of each stock, and the model ran well, giving us a mix of dividend paying and non-dividend paying stock in the optimal portfolio.

A few weeks ago, we removed a set of stocks that had 14% or higher dividend yields (actual dividends paid / average closing price). 

This week we added them back to the list.

What we found is that the stock list has been dominated by common stocks that paid very high dividends in the past year, and where their stock price has fallen. This seems like a great idea, buy low and sell high, except that many of those dividends were likely one-time events. These were companies often enjoying a Covid-related windfall. In other cases, the dividend was 'normal' but the stock fell by a very large percentage, so that the previous dividend now looks enormous.

This week, the stocks selected by the model in the optimal portfolio were dominated by companies that had paid a high dividend, and we are uncertain whether those dividends would repeat. 

We have turned off our dividend adjustment, and the stocks picked in our optimized portfolio still have risky stocks, but at least we don't have a dividend repeat risk.

This week's CQNS Long portfolio has 12 stocks held evenly, and we will be watching that portfolio carefully to see if the market is acting like it is expecting a rise over the next year.

We set our model to assume a 7.00% capital appreciation and a 1.63% dividend yield for all stocks (or a good proxy like 1/3 of each of the following: S&P 500, Nasdaq 100 and Russell 2000).

​Good luck to all.