Abstract: To systematically analyze the relationship between public expectations and economic fluctuations, we construct a system to analyze economic climate based on public opinion by text mining. Through the creation of a financial sentiment dictionary and natural language processing, we quantify the public's judgment on the economic situation reflected in multi-channel news, that is, the financial sentiment score, which could be used to measure the chaos of economic activity. Because the financial sentiment score can be obtained and updated in real time, it is forward-looking and can be used to portray public expectation. In the empirical analysis, there is indeed a linkage between economic fluctuation and public expectation. The empirical analysis finds that official information tends to reversely realize, and free public expectation and professional expectation could self-realize, that is, optimism brings economic growth, and pessimism brings economic decline. A further Granger test finds that there is a two-way causal relationship between economic fluctuations and public expectations, verifying their self-realization relationship. For the mechanism and path of self-expectation realization, we believe that news information shows the consistent and stable preferences of rational individuals, i.e., the financial sentiment score. Such emotion is finally transmitted to the economic activities of entrepreneurs, causing self-realization. This finding could help the public and regulators further understand the interaction between expectation management and macroeconomic regulation.

Key words: adaptive expectation, economic fluctuation, text analysis, public sentiment