Fundamental Indexation in Brazil: a competitive strategy?/Indexacao fundamental no Brasil: uma estrategia competitiva?/Indexacion fundamental en Brasil: ?una estrategia competitiva?
| Autor | Roquete, Raphael Moses |
1 Introduction
Portfolios weighted by the market value of stocks will not necessarily lead to an efficient relationship between risk and return. Haugen and Baker (1991), for example, showed that there are alternatives that lead to the same expected returns, but with lower volatility. Markowitz (2005) argues that even if the restrictions of long and short positions in the real world were taken into account, the market portfolio, often represented by a stock index weighted by market value, will not be efficient. Passive management usually seeks to imitate one of these indices.
Alternatives for building and weighting passive portfolios, therefore, have been offered. Arnott, Hsu, and Moore (2005) proposed portfolios weighted according to the fundamental indicators of companies and not on the basis of market prices. They claim that the traditional weighting by market value method is sub-optimal because prices may introduce noise and not fully reflect the fundamentals of the companies. The form of weighting proposed by these authors considers that the weight of each stock will be its proportion in the total value of the fundamental indicator selected. Assuming, for example, that the fundamental indicator selected is the company's revenue, whose amount is R$ 10 million, and the sum of the revenue of the companies being considered is R$ 100 million, the weight of this asset in the fundamental index (or portfolio) would be 10 percent in the period.
The objective of this article is to apply the fundamental indexation of Arnott et al. (2005) to the Brazilian market and check whether it leads to an excess return in relation to indexation by market value, between June 2003 and May 2015. The article is the first to address fundamental indexation exclusively in Brazil. It examines whether fundamental indexation generates abnormal risk-adjusted returns according to the Capital Asset Pricing Model (CAPM) and a five-factor model, which is an extension from Fama and French (1993). The article also contributes to extending the method proposed by Arnott et al. (2005) and employs an alternative weighting of the fundamental indices to mitigate the effects of extreme values on the weights. Finally, a contrast is made between the fundamental indices and actively managed stock mutual funds considering their behavior in periods of market highs and lows.
Fundamental indexation may be relevant to the investment fund industry in Brazil, which has been growing substantially. The Brazilian Association Financial and Capital Markets Entities (ANBIMA) (2016) reported that the assets under management (AUM) of investment funds for the 12 months before February 2016 exceeded R$ 3 trillion in contrast with R$ 400 billion in 1996. Despite this, indexed (passive) stock funds represent a much smaller portion of the total AUM of stock funds than those of actively managed stock funds. Data from February 2016 from ANBIMA showed that less than two percent of the AUM of stock funds was made up of indexed stock funds. On the other hand, according to the Investment Company Institute (ICI) (2016), indexed stock funds in the US continue to grow and accounted for about 22 percent of the mutual fund industry there in 2015. It was not possible to identify Brazilian indexed stock funds that adopt fundamental indexation, but stock funds that consider this type of weighting may be interesting products.
The theme of fundamental indexation in Brazil is also relevant because of the difficulty for managers to achieve a better yield than that of the yields from Interfinancial Certificates of Deposit (CDI) (Dana, 2015). Since the beginning of the Real Plan (July 1994) until May 2015, CDI performance was 3,626.15 percent in comparison with 2,269.94 percent for the main stock index in Brazil, the Ibovespa. Some international financial crises that occurred in the period also contributed to increasing the difficulty of stocks outperforming the fixed income market in Brazil.
The results reported here show that only the portfolio built according to the fundamental index based on free cash flow showed superior performance to the index weighted by market value (IBrX 100), and even then, only when the CAPM alpha is considered. When a five-factor model is considered, this result disappears, corroborating with international authors that claim that results favorable to fundamental indices are nothing more than evidence of a value stock premium. This article continues with a review of the literature, followed by details of the sample and methodology, a discussion of the results, and the conclusions.
2 Literature Review
2.1 Initial evidence in the US
Arnott et al. (2005), Treynor (2005), Hsu (2006), Hsu and Campollo (2006), and Arnott and Hsu (2008) reported that market value weighting might lead to sub-optimization of the risk-return ratio of a portfolio because stock prices are very volatile in relation to their fundamentals. Siegel (2015) argued that stock prices are susceptible to trades that are not related to the fundamental values of companies, but result from the activity of traders related to "liquidity" or "noise", who buy and sell shares influenced by tax reasons, to rebalance portfolios, or for other personal reasons and, therefore, are not necessarily the best estimate of the value of firms. The author called this condition the Noisy Market Hypothesis (NMH).
Arnott et al. (2005) proposed fundamental indexation whose selection, weighting, and rebalancing disregard the market value weighting. This strategy sets out that the weight of each asset in a portfolio is determined according to some fundamental index of the company, such as sales, dividends, cash flow, and earnings, among others. The objective of the authors was to deviate from weighting according to market prices, maintaining, however, the positive attributes of market value passive indexation, such as a low turnover when rebalancing the portfolios once a year, and liquidity, even though admittedly lower (Arnott, Hsu, & Moore, 2005; Arnott & West, 2006; Hsu & Campollo, 2006).
Arnott et al. (2005) argued that portfolios built according to fundamental indicators tend to have higher returns and less volatility than those weighted according to market value. These portfolios, represented by fundamental indices, would maintain the main benefits of a passive strategy, such as exposure to companies of higher market value, concentration in stocks with good liquidity (which reduces transaction costs), betas similar or lower than those of the indices weighted by market value, and a high correlation with the market. The authors built fundamental indices for the US stock market based on revenue, shareholders' equity, gross sales, dividends, cash flow, and number of employees, in addition to building an equally weighted aggregate index with all these measures except for gross revenues and the company's number of employees. The rebalancing of the portfolio occurred once a year. On average, the portfolios built outperformed the S&P 500 by 1.97% per year, with the same volatility for the period between 1964 and 2002.
Also in the US, Chen, Chen, and Basset (2007) estimated the fundamental weights of assets by means of a median of weights according to the historical market values of 1000 companies between 1962 and 2003, without employing their accounting information, such as in Arnott et al. (2005). The authors argue that this information changes slowly and that prices always converge to the fair price in accordance with the fundamentals. The results indicated the superiority of fundamental indexation. Amenc, Goltz, Lodh, and Martellini (2012) calculated the excess return from minimum variance portfolios, of the maximum Sharpe ratio, equally weighted portfolios, and ones based on fundamental indexation, in relation to the S&P 500 between 2003 and 2011. They showed that all these strategies were winners at some time, but it was not possible to argue that fundamental indexation is better.
2.2 Studies in other countries
Fundamental indexation has also been tested in other countries. Estrada (2008) mixed fundamental indexation with international diversification through a global fundamental index of dividends considering 16 markets that represented just over 90 percent of world market capitalization. The author compared it to an index based on the market value for the period between 1974 and 2005 and found an excess return at an average annual rate of 1.9% and a better return-to-risk ratio in favor of fundamental indexation. Hsu and Campollo (2006) created fundamental indices for 23 developed countries and a global one for the period from 1984 to 2004, finding superior results for the average of the fundamental indices of 2.8% p.a. and for the global one of 3.5% p.a. in relation to the corresponding indices computed by Morgan Stanley Capital International (MSCI), which are weighted by market value.
Walkshausl and Lobe (2010) created a global fundamental index considering 50 countries, of which 22 were emerging and 28 developed, and fundamental indices for each country in the period from 1982 to 2008. The aggregate index of each country was formed by the same allocations in four individual fundamental indices calculated with the value of shareholders' equity, cash flow, dividends, and sales. The authors found superior performance for the global fundamental index and for the fundamental indices of each country in relation to the index based on market value. However, this result was less prominent for the indices of each country after carrying out the Ledoit and Wolf (2008) bootstrapping procedure in the robustness tests. The authors found more positive results for fundamental indexation in developed countries than in emerging ones. This evidence was contrary to expected because the likelihood of prices not converging to the fundamentals would be greater in emerging markets, where volatility also tends to be higher. Brazil...
Para continuar a ler
PEÇA SUA AVALIAÇÃOCOPYRIGHT GALE, Cengage Learning. All rights reserved.
