Purified Sentiment Indicators

This thesis constructs the corresponding price dynamics and PSI, basedupon the Market Sentiment Indicators and Sector Sentiment Indicators. Then wearrive at subtle results conditional on preconditions, by comparing the ProfitFactors of different trading rules. As are the followings:Investor sentiments affect stock prices, and conversely prices also affectsentiments. In theory, if we can eliminate the effects of price dynamics (velocity,acceleration and volatility) on Sentiment Indicator (SI), probably we canimprove the stability and predictability of SI. The sentiment indicator is namedPurified Sentiment Indicator (PSI), after price dynamics are removed from it.Relevant empirical works has been done by foreign researchers. So is itmeaningful to build up a PSI of China A share market? Can this indicator guidereal money investment? This paper tries to answer the above problems bycarrying out empirical research. Then three questions are raised:(1) Could SI bepredicted by price dynamics?(2) Does Contrary Opinion Theory hold in ChinaA share markets?(3) Which one could better predict the future price paths? Thisthesis constructs the corresponding price dynamics and PSI, based upon theMarket Sentiment Indicators and Sector Sentiment Indicators. Then we arrive atsubtle results conditional on different preconditions, by comparing the ProfitFactors of different trading rules. As are the followings:All SI is predictable using price dynamics (R squares for market SI rangefrom0.5to0.52, while those of sector SI are from0.53to0.72). Using crossvalidation, conditional on whether the Contrary Opinion theory holds or not, thisthesis gives both long and short trading rules for price dynamics, SI and PSI, andcompares the out of sample profit factors with those of stochastic trading. Insummary PSI dominates SI, while the latter do not obviously dominates stochastic trading. When the Contrary Opinion theory does not hold, pricedynamics indicators strongly outperforms stochastic trading, which demonstratethe strong price momentums in China A shares markets, which could beexplained by the herding psychology. When the Contrary Opinion theory holds,PSI has a better predictability for market index WIND A, which is composed ofmainly blue chips and not of small stocks that are easily affected by the investorsentiment. When the Contrary Opinion theory does not hold, deploying shortstrategies, PSI could assist in predicting the possible future trends of suchsectors as mining, household appliances, black metal, and commercial trade. Aswhether the Contrary Opinion theory holds, except sectors SI, most empiricalresults do not support the theory.In summary, it is meaningful to build up PSI for China A shares, and undermost conditions investor market sentiment has a positive feedback effect on thecorresponding index, which does not obviously support the Contrary Opiniontheory.