Market sophistication, an analysis before COVID-19: Valuation considering the most relevant companies of the NASDAQ-100 applying PLS-SEM algorithms

Abstract

Investors can develop behaviors that may be considered irrational from the classical economics' perspective. This research aims to generate reflective behavioral constructs, which measure the economic effects of investor decisions in the stock market. The applied algorithms' fundament is Partial Least Squares (PLS) - Structural Equation Models (SEM).

Through this approach, the analysis of dependent and independent variables that are not identically independent o randomly distributed is feasible; the advantage of this methodology is that it can apply to small samples by leveraging PLS/SEM. Complex relationships, including categorical variables, improve the models' reliability and validity by reducing the random term with an appropriate collinearity measurement.

Leverage a time series categorization for investor behavior, and path modeling is a novel procedure. Three investor categories are defined: winners, indifferent, and losers, which interact within the NASDAQ-100 stock market and the top ten enterprises in capitalization, deploying seven different phases or emotional stages ranging from financial panic to market euphoria, which reflects the investor’s behavior.

Therefore, in winner enterprises, emotional and contagious effects show market bubbles, specifically on these stocks. The results also helped test the Prospect Theory's central idea: investors tend to be less excited about gains and suffers more from losses in the decision-making process. Also, to assess investor confidence and sense the dynamic market stages to find if it is undervalued or overvalued due to the investors' feedback, interactions, and sense of the market interactions.