Quant Concepts: Does it really help to hold?

Comparing one strategy with a stop-loss, and another without, CPMS's Emily Halverson-Duncan tests whether it helps to stick it out in downturns

Emily Halverson-Duncan 27 March, 2020 | 1:18AM
Facebook Twitter LinkedIn



See more episodes of Quant Concepts here

Emily Halverson-Duncan: Welcome to Quant Concepts, our first ever virtual office edition. With the COVID-19 coronavirus on the rise worldwide, we've seen many impacts including social distancing, the rise of work from home, and of course, volatility in the markets. The latter has caused significant investor pain and uncertainty and continues to beg the question, how much further can the declines go? While I can't answer that question, one question I can address is whether or not exiting your holdings during periods of market volatility has any benefit over the long term.

In today's video, I'm testing whether to stick it out when your holdings are declining, or whether to cut your losses and run. Given the popularity of income investing, I've chosen to create a strategy that searches for companies with growing dividends and compare that with the same model with a price stop-loss added as a sell criteria. Recall that a stop-loss refers to selling a stock once it declines past a specific threshold. For the purpose of this video the stop-loss will be set at a 10% decline. So, let's take a look at how to build that.

First off, as always, we're going to go through and rank our universe of stocks. Here, our universe is the Canadian CPMS universe which has 696 stocks as of today. The factors we're going to look at to rank first off are five-year dividend growth. So, in this case, we want a higher five-year dividend growth for a company to be ranked higher on the list. We're looking for expected dividend growth that looks at what a company is expected to pay out in dividends over the next 12 months. And again, we want that value to be higher. And lastly, we're looking at quarterly earnings surprise, which looks at whether or not a company is beating or missing the expected earnings. And again, we want that to be a higher value.

From there, we're going to go ahead and screen out the stocks we don't want to own so we end up with a list of stocks that we're willing to buy. Some of the screens we're applying here – for the five-year dividend growth, we want that to be in the top third of peers which today has a value of 9.47% or higher. Expected dividend growth, we want to have a positive value for that. So, we want companies that are actually increasing their dividends or expected to increase across the next 12 months. We look at dividend payout ratio using earnings per share, and we want that to be less than or equal to 75%, the reason being, there as we don't want them paying out too much of their earnings in dividends, but also retaining some that they can use for growing the company. And five-year sales growth, which looks at whether a company is growing their sales across the last five years, and we want that to be positive. The stop-loss that we're applying here on the buy side is, we want to buy stocks that have across the last 30 days a price increase greater than minus 10%. So, meaning, if the stock has already dropped more than 10%, we wouldn't buy it. On the sell side, if we are already holding a stock and then that price went down more than 10% across the last 10 days or the last 30 days, we would actually sell it out of the model. The other sell criteria that we have are dividend payout ratio if it rises above 80%, that's too high and we'll sell it out of the model and five-year dividend growth, if it falls into the bottom half of peers, which is 3.21% today or below, we would sell it from the model.

So, let's go ahead and back test this and see how it did. Our back test here is actually going to be a little bit unique. Instead of running it against the broad-based index like we usually do, we're actually going to run it against the same model, but with no stop-loss applied. So, all those same rules we just talked about will still be in there but no 10% stop-loss.

The time period for this back test is going to be from January 2004 until February 2020. And there are 15 stocks that are going to be selected throughout this time period. What you're looking at right now is the top 15 stocks that would have been purchased in January 2004. And then those buy and sell rules would be applied going forward. A few key things to note about this back test is, there's no survivorship bias. So, if any stocks were to go bankrupt in this timeframe, the historical data does still stay in our system. So, you'd still be seeing that reflected should you be holding any of those stocks. And there's no forward-looking bias as well. All this data was collected point in time. So, you'll be looking at a very good representation of what would have happened had you followed either of these models.

So, the maroon line that we're seeing there is the stop-loss version. And then, the orange line is the same model but with no stop-loss applied. I still have a few more years to run here. One thing I'm noticing right off the bat is that in 2008, it looks like the dividend grower strategy without the stop-loss actually seemed to have what looks like a smaller decline than the one with a stop-loss. But we'll confirm that after the back test is run.

Just about finished here. So, across this timeframe, the model with the stop-loss performed 11.6% annualized across 2004 to 2020. That's actually an underperformance if we had (indiscernible) by just under a percent. So, what that says is, adding that stop-loss in actually didn't improve the performance; it actually took away from it slightly across that timeframe. Turnover is actually a little on the higher side. The original model has a turnover of about 18%. So, at 81, you're more or less four times more turnover, meaning that you're trading a lot more frequently within the model.

More specifically, though, we want to look at volatility. So, I can see here, downside deviation, which looks at volatility of negative returns, for the stop-loss strategy is 7.5, for the non-stop-loss of 7.2. So, it is actually a little bit higher despite adding in that stop-loss criteria.

One other thing that we'll look at is specifically 2008, which is another year, of course, the credit crisis where the market did quite poorly. So, what we're looking at here for the 10% stop-loss version, the returns were minus 33.62 in that year. For the no-stop-loss version, it was minus 25.89. So, actually, not having the stop-loss would have benefited you about 7.73% in 2008. So, adding that stop-loss would not have helped you.

So, all that to say, despite all the market volatility that's going on and the pain that it can cause investors, changing your strategy right now may not be the best idea. It's better to trust your initial judgment, stick the course and see how things will play out over the long term, because over the long term, we've seen that trusting your original portfolio plan tends to be the best bet.

For Morningstar, I'm Emily Halverson-Duncan.

How will COVID-19 affect your portfolio?

Find out all you need to know here

Facebook Twitter LinkedIn

About Author

Emily Halverson-Duncan  Emily is Director, CPMS Sales at Morningstar

© Copyright 2022 Morningstar, Inc. All rights reserved.

Terms of Use        Privacy Policy