Quant Concepts: ESG metrics to mitigate risk

Ethics and personal beliefs aside, CPMS's Ian Tam shows us that screening for responsible corporate policies can pay off

Ian Tam 1 October, 2019 | 8:05AM

 

 

See more episodes of Quant Concepts here

Ian Tam: Welcome to Quant Concepts. As the idea of using environmental, social and governance factors when investing continues to proliferate globally, investors often ask whether these factors actually add any value to an investment strategy. Our partners at Sustainalytics have a firm belief that analyzing a company's corporate policies around topics like carbon footprint, customer data privacy, and board gender composition can often be an infected indicator of monitoring risk. Ethics and personal beliefs aside, these types of policies are likely intuitive to investors who require that a company's policy structure is sound. Today, I'll use Morningstar CPMS in conjunction with data from our partners at Sustainalytics, Inc. to exemplify how leveraging ESG metrics might change a quantitative investment strategy.

So, to start off with, what I'll do is, build a fairly aggressive model using Morningstar CPMS. And of course, I'm going to pick certain factors that speak to that style of investment. So, taking a look at the pie chart on my screen, you'll see I have six factors. Each of them are going to be fairly short term and aggressive in nature. So, for example, we're looking at quarterly earnings momentum and quarterly sales momentum. Each of these measures, the last four quarters for earnings or sales compared against the same number one quarter ago. So, again, very short term in nature. I'm also going to look at a couple of technical indicators. So, for example, a three-month price change and a nine-month price change, again, just measuring market sentiment and the direction that the stock is moving. And of course, we prefer the stock to be moving in the upward direction. We're also going to look at analyst sentiments. So, these are opinions from sell-side institutional research analysts. And what we're doing is, looking at today's consensus estimate on earnings compared against what it was three months ago, so roughly one fiscal quarter ago. We're also going to look at quarterly earnings surprise. So again, this is what the company has reported versus what the analyst expectations were just prior to the company reporting. So, each of these factors is fairly short term in nature. They move very quickly relative to other longer-term factors. And what's this is designed to be is a fairly aggressive momentum-oriented model.

So, based on those six factors, we're going to take all the companies in our U.S. CPMS database, which has about 2,150 companies today, and we're going to rank them based on those six factors. You can see they're equally weighted on my screen. For me to consider buying a stock, I'm going to look at the top quartile, the top 25% of stocks. I'm also going to put a market cap minimum on my list of stocks just to make sure that they're reasonably liquid. Today, I'm using a market cap minimum of US$22 billion. That represents the median U.S. market cap today. So, those are my buy rules. For my sell rules, I'm simply going to say that if the stock dropped below the top 35% of my universe, I'm going to sell that stock.

So, to start with, let's just have a look at how this momentum-oriented model would work over a fairly long period of time. So, of course, I'll use Morningstar CPMS to back test this for you.

So, what's happening in the back test, of course, is we're taking some amount of cash, here we're using $1 million. We're using the data from August of 2009 as a starting point, we would have picked 15 stocks that met my requirements at that point in time, with no more than four per sector. If any of the stocks fall below the top 35%, I sell the stock, and I replace it with the next highest-ranking stock. So, based on that very simple momentum-oriented model, you can see that the annualized return on that is fairly market leg. So, we're looking at about a 13% annualized rate of return, that's without any fees or transaction cost. That's a little bit worse off than the market. And the important thing to note here, it's got a fairly high turnover, meaning that we're trading in and out of stocks fairly rapidly throughout the year. So, here, we're looking at 121% turnover. So, on a 15-stock model, you're trading more than the entire portfolio every year, on average.

More importantly, we also want to look at the Sharpe Ratio. So, Sharpe Ratio here is 0.7%. And as a quick reminder, Sharpe ratio is your risk adjusted return. So, how much are you getting in return for the amount of risk that you're taking on. So, unfortunately, a Sharpe Ratio of 0.7 is actually a little lower than buying the market. So, this is just a very simple example of a momentum-oriented model.

So, going back to the theme today, perhaps using some ESG factors within the model might actually help modify the returns a little bit or modify the results. So, what I've done here is built a second model, this time adding a very substantial waiting on what Sustainalytics calls the overall ESG score. Now, ESG scores are fairly new development in the market. Many people think of them as a negative screen. So, for example, if you're in the oil and gas industry, you might get a poor ESG score. That's actually not how it works. ESG scores are based on the corporate policy that are in place at each company. And they're going to be measured against other stocks in a similar sector. So, for example, Suncor, just as a plain example, might have better corporate policies around carbon footprint than another company in that sector.

So, there are three pillars for ESG scores, environmental, social, and governance. And depending on which sector the company belongs to, those three pillars are going to be weighted a little bit differently. And the weighted average of those scores basically results in the overall ESG score, as defined by Sustainalytics. So, again, each ESG score is going to be relative to stocks in its own sector or against global peers. So, what I've done here is, I've put up a very substantial weighting in the Sustainalytics overall ESG score. So, on the screen, you can see it's worth now 50% of my model. So, only half of my model is based on the momentum factors that I showed you earlier, and the other half is based on the Sustainalytics' overall ESG score.

I've kept the model pretty much the same other than the fact that in order for me to buy a stock, the company has to have a fairly good Sustainalytics ESG score. So, other than that, I kept the model exactly the same. So, just to illustrate or exemplify the differences between using an ESG overlay and not using an ESG overlay on the same model, I'm going to run the back test again with these modifications.

So, again, we're starting in August 2009. Many viewers may ask why I'm not including 2008. Sustainalytics data actually starts in August of 2009. So, I wanted to make sure we have the complete data set available across the stuff that I'm looking at for the full length of the back test.

So, having a look here, the return on a total return basis is about 12.3%. So, on the surface, it looks like this model did a little worse off than without ESG scores. However, if we look a little bit closer, you'll see a couple of large differences. First of all, the turnover is substantially lower. Now, this kind of makes sense because corporate policies don't often change very quickly. In fact, they change probably less than once a year. So, having a fairly stable factor in my model immediately will reduce the turnover in my model and hopefully reduce the volatility of the model as well. So, the turnover here is about 23% on the same 15-stock portfolio.

The second thing, of course, I'm going to look at is the Sharpe Ratio. So, again, per unit of risk, am I getting a higher return? So, you can see here that the Sharpe Ratio is actually 1, so very close to the market, and certainly better than had I not used ESG overlay in my model. Finally, we're going to look at the defensive characteristics of this model. In all the down quarters between 2009 and today, this strategy beat the market 71% of the time. In comparison, if I look at the original model, that model beat the market in down quarters only about 30% of the time. So, again, in this particular case, using ESG scores, environmental, social and governance scores, actually helps the risk-adjusted returns even though not on an absolute basis, but when you factor in the amount of risk you're taking, it looks a little bit better. So, the stocks that meet the requirements to be purchased today in the model are listed in the table accompanying the transcript to this video.

For Morningstar, I'm Ian Tam.

About Author

Ian Tam  Contributor