Abstract
Based on the dictates of market efficiency theory, stock prices immediately adapt to new information and there is no way to "beat the market". However, Behavioral Finance studies have shown that the market, based on the nature of the new information, reacts excessively (over-reaction) or too little (under-reaction).

The over-reaction represents the excessive, disproportionate and optimistic reaction of investors to the availability of new information regarding certain securities on the market. This disproportionate reaction suggests that investors are overly influenced by random events and this reaction causes a change in the prices of the securities in the period following the publication of the new information, prices that no longer reflect the fundamental value of the securities.
For De Bondt and Thaler (1985), the irrational investor has an overly optimistic reaction in the initial period with subsequent correction (a trend reversal) in the long run that tends to bring the price of the security back to its correct value. It may happen, therefore, that good news on a stock at a given time conditions its price for long periods, making it overvalued compared to the current value of expected future cash flows (fundamental value, Price Earning). This behavior of investors, which is clearly in contrast with the hypothesis of the efficiency of financial markets, can be generated by the representativeness mechanism that intervenes in the decision-making process of the individual or from the subject's over-confidence and is favored by imitative behavior .
Representativeness is based on a limited number of simple and fast rules, ie investors' choices are formulated on the basis of stereotypes rather than on a logical and rational mechanism. Therefore, the estimate of the probability of an event seems to depend on how much it appears to be similar to a given category of events, regardless of the observable frequency and the sample size. The tendency to ignore the size of the sample often leads to errors, such as the so-called gambler fallacy analyzed by Kahneman and Tversky (1974), ie the propensity of investors to believe that the frequency of an event of a large sample can be observed also with reference to a small sample. For example, in gambling, individuals believe that a random event has a greater chance of occurring because it has not occurred for a certain period of time; in the lotto game the cd. late numbers.
By imitative behavior, on the other hand, we mean the phenomenon through which individuals act in the same way, without there being any coordination between individuals (flock effect). The individual often finds it difficult to find the right answers, especially in times of uncertainty. For this reason it tends to pay attention to the actions and choices of other subjects. Furthermore, he believes that, following the behavior of other individuals (herd), the responsibility for any losses is not his, unlike when he adopts a strategy contrary to the trend. This effect, but also the general over-reaction of investors to new information, may underlie anomalies and cases of market failure such as financial bubbles. In a further study, Kahneman and Riepe (1998) noted that the human mind is always looking for patterns (patterns), causes, coherent stories and is strongly focused on adopting the hypothesis that a random factor may be present behind any sequence remarkable of events. Consequently, investors tend to over-interpret random patterns and are unlikely to persist over time. They react to recent events and their own experiences without paying enough attention to events that have not been directly experienced or preserved over time.
Unlike over-reaction, the under-reaction effect occurs when the arrival of new information on a particular security is only partially transferred into purchases / sales and therefore investors under-react to the spread of this information . The prices of the securities, after the emission of new information, tend to move slowly with respect to what would be correct to expect. In general, the under-reaction phenomenon has a shorter time horizon than that of the over-reaction, even if in the long term the prices of securities return to be equal to their fundamental value.
This sub-reaction of financial agents can be explained by the phenomenon of conservatism analyzed by Edwards (1968). It represents the slow updating of expectations in the face of the arrival of new information. Edwards finds that investors who behave conservatively pay little attention or even ignore information from recent earnings announcements, but cling to their previous beliefs based on past income. Scholars like Fama have tried to demonstrate how the effects of over-reaction and under-reaction within the market are able to compensate themselves and therefore do not constitute a reason to believe the hypothesis of efficiency of the financial markets to be inaccurate. Other important economists such as Barberis, Shleifer and Vishny (1997) have tried to develop models capable of explaining the behavior of financial agents. They believe that these traders' attitudes are dictated by the statistical weight (statistical weight) and strength (strength) of the announcements with which new information is provided. Under high-weight announcements (that is, information that has an important statistical relevance), but with reduced strength, under-reaction phenomena usually occur, while in the opposite situation it is usual to be faced with over-reaction of investors. Furthermore, Barberis, Shleifer and Vishny generally believe that in the case of under-reaction the expected value of future returns obtained following positive news is greater than that obtained for bad news, while in the case of over-reaction, the expected value of the returns deriving from a series of good news is lower than that following a series of negative news.

References
BARBERIS N., Shleifer A. and Vishny R. (1997) A model of investor sentiment. Journal of Financial Economics 49 (1998) 307-343
DE BONDT W. and THALER R. (1985). Does the stock market overreact? The Journal of Finance
Vol. 40, No. 3
EDWARDS W. (1968). Conservatism in human information processing. In Formal representation of human judgment, Wiley.
KAHNEMAN D. and TVERSKY A.(1974). Judgement under uncertainty: Heuristics and Biases. Science, 185(4157), 1124-1131.
KAHNEMAN D. and RIEPE M.(1998). The Psychology of non-Professional Investor. The Journal of Portfolio. Management, 24, 52-65.