This is quite a lengthy but interesting piece which forms part of a PhD submission. I have snipped out a portion of the introduction that makes some interesting statements about the role of new in trader behaviour and the impact of news upon sophisticated and non sophisticated investors. The author uses an interesting technique for defining unsophisticated based upon levels of reading comprehension and therefore the ability to respond to complex pieces of information.
This paper makes four contributions to the literature. First, I show that sophisticated investors are less likely to trade in response to new information. I flag transactions that occur within 5 days of a news article being published and estimate the effect of news on the sell-or-hold decision of each individual investor in the sample. Using the same measure of trading propensity I find that sophisticated investors tend to trade more often than unsophisticated investors unconditionally. This suggests that the trading motives of sophisticated investors are different from unsophisticated investors in that they tend to trade more for reasons such as portfolio rebalancing and less in response to new information.
Second, I show that unsophisticated investors trade using relatively larger individual transactions. For each transaction in the sample I determine the fraction of stocks sold relative to the investor’s total holding in the specific stock and estimate the effect of sophistication on relative transaction sizes. Hence, even though sophisticated investors trade more frequently each individual transaction is smaller. This is consistent with previous literature that finds that turnover, the compound effect of trading frequency and transaction sizes, is larger for unsophisticated investors (Anderson, 2013).
Third, I provide evidence suggesting that sophisticated investors are more likely to use complex information that requires higher levels of reading comprehension, and that this leads to better stock-picking performance. Using LIX (Bj¨ornsson, 1968), the Swedish equivalent of standard readability measures such as Gunning-fog and Flesch-Kincaid (Flesch, 1948; Gunning, 1952; Kincaid et al., 1975), I divide my sample of articles into two groups, complex and non-complex, based on how difficult they are to comprehend. I find that while sophisticated investors, compared to unsophisticated investors, are less likely to trade following non-complex articles, they are more likely to trade following complex articles. Furthermore, comparing stock prices at the date of the transaction to the price of the same stock 12 months later I find that the probability to sell a stock that subsequently decreases in price is higher if the transaction occurred in response to a complex article, and particularly so for sophisticated investors.
Fourth, I document a gender gap in individual investor reactions to news. I find that males trade more often and use larger individual transactions both in response to news as well as unconditionally. While the gender gap is present among both the unsophisticated and the sophisticated investors, sophistication greatly reduces the magnitude of this gap. While the gender gap is present and 3 sizeable in the propensity to trade as well as in relative transaction sizes, I find no systematic evidence for a gender gap with regards to stock-picking performance. In interpreting my findings I relate to the theoretical model of overconfi- dence in Odean (1998). Modelling overconfidence as a tendency to overvalue the relevance and precision of new information Odean (1998) predicts that overcon- fidence increases trading volumes and decreases returns.
Furthermore, Odean (1998) predicts that markets underreact to “abstract, statistical, and highly relevant information” and overreact to “salient, anecdotal, and less relevant information.” I also relate to the empirical findings that investor sophistication (1) reduces behavioral biases and investor mistakes (Feng and Seasholes, 2005; Calvet et al., 2007, 2009a), and (2) is strongly linked to reading comprehension (Flesch, 1948; Gunning, 1952; Kincaid et al., 1975). I argue that with investor sophistication reducing behavioral biases such as overconfidence we would expect household trading behavior to be affected in two ways: (1) sophisticated investors, more likely to discount the relevance of new information, will be less likely to act on new information; (2) conditional on acting on new information, sophisticated investors are more cautious about the precision of the new information, and hence react less strongly. Since unsophisticated investors exhibit lower levels of reading comprehension we would expect unsophisticated investors to be less likely to act on new information that is complex in nature. Finally, the better stock-picking performance of sophisticated investors can be understood as a result of them being more selective in the news they choose to act on, as well as them being able to comprehend more complex information
The central issue for us as system traders is that information needs to be carefully managed and filtered – to me news has always represented little more than noise that clouds the investing/trading process.