In this issue, I would like to recommend a paper involving behavioral finance "Paying Attention: Overnight Return and The Hidden Cost of Buying at The Open", written by Henk Berkman et al., this paper used data from more than 3,000 listed companies in the United States to conduct an empirical analysis on whether there are abnormal overnight returns during the opening period and the transaction costs incurred by investors. The conclusion of this paper is useful for quantitative researchers. It has some reference value for ordinary investors, because the article uses data to prove a problem - retail investors chasing high prices and buying high-opening hot stocks are "giving money" to institutions. If you have such a bad trading habit, you may want to take a look. this paper.
The authors of this paper conducted an in-depth study of a type of stock price behavior in which prices open high and then fall back. This concentration of price behavior occurs in stocks with high retail investor concentration recently, especially when retail investor sentiment is high, and is more obvious for stocks that are difficult to value and have high arbitrage costs. Research points out that retail investors will have to pay additional transaction fees when buying stocks of these companies with higher attention at the opening of the market, which means that the slippage of the transaction will be relatively large.
The authors of the paper are based on a theory discovered by Barber and Odean in 2008 - that retail investors will flock to attractive stocks, so-called hot stocks. Barber and Odean point out that for retail investors, buyers and sellers face different decision-making horizons. Just imagine, when retail investors want to buy stocks, they must choose from thousands of stocks, and they will naturally limit their stock selection to popular stocks. The difference is that if retail investors want to sell stocks, they face a much smaller range because they can only choose from the stocks they hold, and retail investors generally do not engage in short selling. Barber and Odean proved that when stocks have higher returns for several consecutive days, retail investors will simply go long on the next trading day.
Based on the theory of behavioral finance, investors will be affected by emotions when trading. Driven by emotions, retail investors tend to act in groups. They trade the same stocks at the same time and move in the same direction. Therefore, trading against these retail investors may bear high costs and risks, because under the influence of this collective price behavior, emotionally based trading may cause the stock price to deviate from its fundamental value in the long term. The authors of the
paper deepened and expanded their research based on the conclusions of Barber and Odean. The authors used order flow and price information to predict intraday patterns. Based on this, the author puts forward two important hypotheses. The first hypothesis is that after an attractive event occurs on day t, retail investors will be guided to trade at the opening price on day t+1, especially when retail investor sentiment is high. This is because when the market opens, it is the first opportunity for retail investors to buy stocks. The second hypothesis is that such event-triggered buy-side trading will lead to higher opening prices than other trading sessions during the day, especially when the stock is difficult to price effectively and there are high arbitrage costs.
The author of the paper used the 3,000 largest U.S. listed companies from 1996 to 2008 when verifying the above hypothesis. In a preliminary analysis, the authors examined the opening and closing midpoints and found a significant daily average overnight return of 0.1%, as well as a -0.07% return reversal. The above phenomenon is more significant in some stock samples with a low proportion of institutional holdings. The author analyzed a sample group of stocks with a high proportion of retail holdings and high attention (hot), and found that in this sample set, the average overnight return was 0.43%, while the day's return reversed to -0.45%. This conclusion means that for these popular stocks with low institutional holdings, if the trading time is postponed from the opening to the closing time, the annualized return can be increased by approximately 112.5%. The
authors' results are consistent with evidence from two other papers cited in the article, Branch and Ma (2008) who found a negative relationship between overnight returns and subsequent trading day returns.They explain that this trend may be related to the microstructure of professional traders and market makers at the opening of the market. In the same year, Cliff, Cooper, and Gulen found that their observation of negative daily returns was a severe challenge to traditional asset pricing models, because these models were unable to predict negative average returns. The researchers of these two papers emphasized that this behavior represents a new abnormal behavior and hope that there will be a theory to explain it. The theory of the author of this paper is based on the behavior of retail investors buying at the opening triggered by paying attention to hot topics, which can uniformly explain the above abnormal behaviors.
The author of this paper used 2 indicators to proxy stock popularity. The first one is yesterday's squared return as an indicator that can attract retail investors. The second one is based on the previous assumption of Barber and Odea - retail investors are pure buyers of popular stocks, so the proportion of active buying volume in the previous day to the overall trading volume is used as another indicator. The author of
found through data analysis that the purchasing intensity of retail investors in popular stocks during the opening period was significantly higher than that of unpopular stocks. Secondly, within one hour of opening, the purchasing power of retail investors of popular stocks was also significantly higher than that of unpopular stocks. In a sample of 3,000 of the largest U.S. companies, it was found that popular stocks had a significant overnight return of +13 basis points. In contrast, the out-of-favor stocks only had +3 basis points overnight gains.
From the above table, it can be seen that the two indicators that measure the attention of retail investors are significantly positively correlated with CTO (close to open overnight return), and are significantly negatively correlated with the next day OTC (open to close intraday return), confirming the author's assumptions. However, it is still uncertain whether the reversal returns of popular stocks on the next day come from the overestimation of the opening or the underestimation of the closing. The author further researched this issue.
The above picture shows the time series distribution of popular stock prices calculated by the author based on empirical data. It can be seen that popular stocks exceeded the 95% percentile at the opening, which is obviously overvalued, and then fell back to the 95% range very quickly. . It can be seen that intraday reversal returns are caused by higher opening prices. The above conclusion warns US stock investors that they should not directly buy popular stocks at the opening. They will get better costs by opening positions one hour after the opening. At the same time, the authors of the paper found that retail investors also face higher hidden transaction costs. The slippage they need to bear when buying within the opening hour may be twice as much as at other times.
In addition, the actual application of the degree of hot stocks opening higher and falling back to A-shares needs to be explored and verified. The reason is that the trading system of U.S. stocks is different from that of A-shares, and the actual magnitude of the fall may also be different. Our thinking is that since A-shares cannot be directly traded at t+0, the magnitude of the decline may lag behind that of U.S. stocks. At the same time, A-shares still have the problem of daily limit, and the high opening effect of popular stocks may not be Obviously, since there is no daily limit system for U.S. stocks, the opening price can be completely overestimated. (Part of it comes from Nasdaq’s special data. In order not to cause misunderstanding among investors, we will not describe the use of special data. If you are interested, please contact us.
Disclaimer
Reprinted from Shiling Research and issued by The research report is provided for information purposes only and does not constitute any advertising or investment advice under any circumstances. Investors must note that any investment decisions they make based on it have nothing to do with this company, its employees or affiliated institutions.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-