Analyzing the quality factor for Brazil.
| Autor | Bortoluzzo, Adriana Bruscato |
-
Introduction
An extensive literature on risk factors explains returns on assets. One of the most famous asset pricing models, created by Fama and French (1993), has undergone several updates over the years. Asness et al. (2019) suggest a new risk factor that remunerates the investor, known as "quality minus junk" (QMJ). The intuition behind creating this factor is that higher-quality assets, all else being constant, should be more expensive than low-quality assets, eliminating higher risk-adjusted returns for a portfolio long on high-quality assets and short on assets with low-quality characteristics.
Studies show that the most profitable companies generate greater return for the investor (Novy-Marx, 2013); companies with low leverage have a higher alpha (George and Hwang, 2010; Penman et al., 2007); companies with a lower risk of credit tend to perform better; and growing companies have better returns than those that grow little (Mohanram, 2005). Asness et al. (2019) connect these disparate ideas, presenting a common characteristic among them: the quality of the company. This factor can explain a positive relationship between prices and quality. But what would a quality company be? For the authors, a quality company is one that grows, is highly profitable, bears low bankruptcy risks, and, hence, has a low cost of capital.
At the time of writing this paper, Brazil was going through a period with the risk-free rate at historically low levels, and negative real interest rates. Institutional investors and individuals were seeking to diversify their investments, accepting more risk. According to data from the Brazilian stock exchange (B3), the number of active accounts for individuals increased 107% in 2019, and in 2020, 92%. (1) We believe that studying the impact of stock picking on the return of stock portfolios becomes more relevant in such a scenario. In this study, following the suggestion of Chagas et al. (2020) to advance the relevant literature on Brazil, we apply the model proposed by Asness et al. (2019) for Brazilian stocks, to assess whether Brazilian investors are remunerated for being exposed to higher quality assets. The construction of the QMJ factor can also be used in a practical way to study the performance of funds or investors. Frazzini et al. (2018) find that the notorious investor Warren Buffet, adept at value investing, tends to buy high-quality, safe, and cheap companies. These factors may explain the significant alpha generated in his company over time.
With data collected from Brazil for the period from 2000 to 2018, our results suggest that the quality of companies is a characteristic that persists over time. That is, a company that is said to be of quality today should keep this status in later periods. In addition, we note, as do Asness et al. (2019), that the quality of the company has a positive relation to its price. Finally, we see that the QMJ factor, despite presenting high returns in the period analyzed, does not have statistical significance as a new risk factor, even when the explanation of traditional models is improved.
This paper is organized in five sections. After this introduction, we review the relevant literature, and present our methodology. We divide the results section between analysis of the persistence of quality and its impact on the price of assets and analysis of the QMJ factor. In the last section, we conclude and suggest directions for future study.
-
Literature review
Markowitz (1959) was one of the first scholars to consider not only returns, but also the risk involved for an investor. He divides the risk to which the investor is exposed into two components: systematic risk, which is not diversifiable; and specific risk, which is. From this theory, Sharpe (1964) develop a model, based on the study by Tobin (1958), which is known as the capital asset pricing model (CAPM). One of the implications of the model is that a diversified investor can only achieve greater returns when exposed to a greater market risk, the sensitivity known as beta.
CAPM is still widely used in practice and in academic studies, despite evidence that market risk is not sufficient to fully explain returns on assets and portfolios. Based on the CAPM, Jensen (1968) develops a measure of the skill of a portfolio manager, which consists of the constant (alpha) of the regression equation commonly used for empirical tests of the CAPM. Several researchers find evidence of returns consistently above those postulated by the CAPM (statistically positive alpha). These pieces of evidence are commonly called anomalies. One of the oldest anomalies studied is the value factor. Graham and Dodd (1934) document that companies with value perform better than companies with lower value, known as growth companies (growth). This anomaly has been explored in depth by other authors such as Basu (1975, 1983) and Stattman (1980), who show, by different metrics, that the value factor can explain part of the returns on assets.
Banz (1981) find evidence of a negative relationship between the size of the company and its return, so that small companies have higher returns than large companies. This anomaly, called the size factor, has persisted for at least forty years. It is incorporated by Fama and French (1993), together with the value factor, and the market risk premium, into the model today known as the Fama and French three-factor model.
Jegadeesh and Titman (1993) find another anomaly: the momentum factor. According to them, it is possibly linked to the behavior of investors. The authors realize that buying companies that performed well and selling those that did not perform so well in the past generates abnormal returns for investors. One of the authors' main arguments is that the market reacts slowly to new information specific to firms, a phenomenon known as underreaction. The strategy of maintaining a portfolio long on winning shares (with good past performance) and short on losing shares (with low past performance) is the momentum factor. Carhart (1997) also finds evidence of the influence of the momentum factor on asset returns. He suggests a four-factor model, adding this factor to the earlier three factors of Fama and French.
Through a financing constraint model, Frazzini and Pedersen (2014) prove that a strategy of maintaining a portfolio long on low-beta assets and short on high-beta assets, known as the "betting against beta" (BAB) factor, produces a positive excess return, even with beta zero. According to the model, it presents positive returns when there is liquidity in financing, and negative returns in liquidity crises. The authors obtain positive alphas, not only for CAPM, but also for several other pricing models with more factors, in the U.S. and in nineteen other developed markets.
Some investigators discuss the idea that investors obtain abnormal returns if they maintain a strategy of being long on quality companies and short on companies considered to have low quality. But how to classify and create a quality measure? Graham (1973) uses metrics of indebtedness, stability of profits, and growth of dividends, together with metrics of value (price- tobook and price-earnings) to show that choosing cheap and quality companies compensates investors. Greenblatt (2010) defines quality as a high return on invested capital, combined with a high net profit margin. Novy-Marx (2013) finds that companies with high gross profitability generate excess return not explained by Carhart's four-factor model (Carhart, 1997). Based on the discounted dividend model, Fama and French (2015) add two new factors: profitability and investment, to the Fama and French three-factor model. They note that including these factors reduces the relevance of the value factor to explain excess returns. However, Martins and Eid Jr (2015) show that, in Brazil, these two new factors have low explanatory power.
Asness et al. (2019) combine several variables in a z-score methodology based on Gordon's (1962) growth model to create the factor known as "quality minus junk" (QMJ). The authors realize that the quality characteristic is not fully priced in the actions. They discuss whether this positive adjusted return can come from human behavior, regulatory constraints, or the wrong specification of risk models. The excess return persists in the North American data, and is also observed in twenty-three more countries. In Brazil, as far as we are aware, this factor has not been tested. It was suggested as a topic for future research by Chagas et al. (2020), on the possibility of gains from a well-founded stock portfolio.
-
Methodology and Database
We initially defined the universe of all companies listed on the Sao Paulo stock exchange. For each company, we chose the main stock, based on Bloomberg's classification of primary stocks. In all, we collected a sample of 355 shares listed between 2000 and 2018. Also, in each month, for a share to be eligible for our analysis, it must have had an average volume greater than 500 thousand Brazilian Reais (BRL) per day, on more than 80% of trading days during the previous year. These filters follow those used by the Center for Research in Financial Economics (NEFIN) at the University of Sao Paulo (USP) in its studies related to risk factors. Thus, we guarantee that we are collecting a universe of shares available over the period analyzed, including shares that are no longer listed, and a universe that can be traded without major problems.
We collected balance sheet data, as well as price and market value data, from Bloomberg. We calculated returns in BRL, and adjusted them for dividends and other corporate events. We calculated excess returns on the 30day deposito interbancario (DI) swap, which is a future-interest-rate contract traded on B3. It is more liquid than the government bonds secondary market, and can be considered to...
Para continuar a ler
PEÇA SUA AVALIAÇÃOCOPYRIGHT GALE, Cengage Learning. All rights reserved.
