Join our community of smart investors

Examining the quality factor

Former hedge fund manager Stephen Clapham explores ways of identifying quality stocks
Examining the quality factor

I try to read widely, as I think it’s essential for investors. Some fund managers produce excellent content and I am therefore sharing some thoughts from my review of two studies looking at quality stocks. The first is Amundi Asset Management’s Revisiting Quality Investing, a 100-page paper from last year. The second is a study by Invesco which its chief investment officer (CIO) recently discussed at a conference.

The Amundi paper set out the best way to define quality as a factor. Stocks were grouped into quintiles according to various quality criteria and their performance following the global financial crisis analysed, from 2007-20. There was a striking difference between the lowest and highest quality stocks:

The authors divided the quality baskets into different factors – based on profitability, safety, earnings quality and investment – see diagram.

By using two metrics to define each factor, there is less chance of misclassification (whether these are the best bases to look at is a subject for another article). The results in performance terms for each are shown in the chart, again breaking each factor into quintiles.

The profitability metric, which is really a returns-based metric, shows the most significant performance variation between best and worst quintiles, and it’s here we should focus. The authors suggest that the risk-adjusted return will improve by using all the factors together as shown in the multi-dimensional scores in the table, but while multi-dimensional might be more effective for quants funds, it’s obviously of less use to private investors.

The table highlights a small improvement overall in returns and a lower volatility, but it’s not consistent across geographies and I just don’t think many serious investors care as much about volatility as the academics (and perhaps some wealth managers).

Table 1 Factor-based performance statistics, June 2007-May 2020
 Multi-ProfitabilityEarningsSafety Investment
 dimensional quality  
Ann. Return8.20%7.70%4.20%4.40%2.90%
Ann. Volatility6.10%7.10%6.20%7.30%8.00%
Risk-Adj. Return1.341.080.680.60.36
North America     
Ann. Return9.40%8.70%5.70%5.10%1.70%
Ann. Volatility8.30%9.20%7.20%9.50%11.30%
Risk-Adj. Return1.130.950.790.540.15
Ann. Return1.90%5.50%-2.10%0.60%2.00%
Ann. Volatility8.90%10.60%12.80%12.60%8.10%
Risk-Adj. Return0.210.52-
Europe ex-EMU     
Ann. Return7.30%5.20%2.60%2.20%5.80%
Ann. Volatility9.70%10.80%11.40%11.70%11.30%
Risk-Adj. Return0.750.480.230.190.51
Ann. Return8.80%4.60%2.30%7.00%3.40%
Ann. Volatility8.40%10.20%9.50%10.80%8.90%
Risk-Adj. Return1.050.450.240.650.38
Pacific ex-Japan     
Ann. Return1.00%5.10%0.40%-2.80%1.60%
Ann. Volatility13.20%11.50%14.10%11.10%14.40%
Risk-Adj. Return0.070.440.03-0.250.11

Globally, profitability is the best factor by far, but combining it with the others does improve results slightly (by 0.5 per cent as highlighted, which is significant) and lower volatility (1 per cent as shown in the line below the highlights). Let’s just focus on the advantage of the multi-dimensional vs returns alone.

Table 2 Comparing profitability with multi-dimensional factors
Ann. Return8.20%7.70%0.50%
Ann. Volatility6.10%7.10%1.00%
Risk-Adj. Return1.341.081.24
North America   
Ann. Return9.40%8.70%0.70%
Ann. Volatility8.30%9.20%0.90%
Risk-Adj. Return1.130.951.19
Ann. Return1.90%5.50%-3.60%
Ann. Volatility8.90%10.60%1.70%
Risk-Adj. Return0.210.520.4
Europe ex-EMU   
Ann. Return7.30%5.20%2.10%
Ann. Volatility9.70%10.80%1.10%
Risk-Adj. Return0.750.481.56
Ann. Return8.80%4.60%4.20%
Ann. Volatility8.40%10.20%1.80%
Risk-Adj. Return1.050.452.33
Pacific ex-Japan   
Ann. Return1.00%5.10%-4.10%
Ann. Volatility13.20%11.50%-1.70%
Risk-Adj. Return0.070.440.16

For simplicity, I show the performance improvement and reduction in volatility by looking at the arithmetical differences and the risk returns using a geometric calculation. It’s interesting that in two regions out of five, it’s better to use the returns or profitability metric alone. I don’t know why this should be, but it strongly suggests that returns are a safer filter. And that’s consistent with what we would have expected.

Interestingly, when the profitability metric is analysed between its two components, it’s gross profitability that is much superior.

Understanding the drivers is only one part of the story, however, so I wanted to drill down further into the returns and look at the persistence of returns.


Persistence of returns

Invesco has produced some interesting research about persistence of returns and how this affects stock market relative performance. In one particular study they divided the MSCI EMU Index (currently 233 stocks in the universe, weighted to France, Germany and the Netherlands) into 16 buckets. The starting return on invested capital (ROIC) is divided into quartiles and they then looked at where each stock lies in the ROIC quartile five years later to get the 16 buckets. 

So if you start in the top quartile and end in the top quartile, you are in the Q1 starting bucket and Q1 ending bucket, which occurred for 56 per cent of the Q1 stocks. They then showed Q1 stocks that end in Q2, Q3 and Q4, and also the relative performance over a five-year period. This was done starting in July 2010 and ending in July 2015, then repeated monthly until September 2021. The sample size is in my view too small to be conclusive, but it’s interesting that it corresponds closely to my hypothesis.

The most important point to note is that the most frequent bucket pairs are Q1/Q1 and Q4/Q4. In other words, good companies tend to stay good and bad companies tend to stay bad. And you get well-rewarded for buying the first and avoiding the second, with a relative five-year performance in each case of 5 per cent. This is a sensible strategy to follow.

I was quite surprised that buying a bad company that improved significantly didn’t get rewarded more highly – buying a Q4 company that improved to Q1 only beat a company that started in Q1 and ended in Q1 by 3 per cent. Buying a Q3 company that ended in Q1 was slightly better. This profit improvement strategy is the one I followed when I was at the hedge funds and that delivered hugely better relative performance.

Another way of looking at the data is to segment it by the ending quality bucket, again with Q1 being a good company and Q4 being a bad company.

If you end making high returns you will make good money, and if you end making poor returns you will lose money versus the benchmark. Stocks that stay in the middle perform like the average and if they slip performance is significantly weaker.

As I say, the sample size is small and I don’t place too much emphasis on it, but I would like to look for some wider studies, and I am sure there will be a number. But buying quality has worked and good companies tend to stay good companies which sustains the relative performance.