Memory

  • 详情 Long and Short Memory in the Risk-Neutral Pricing Process
    This article proposes a semi-martingale approximation to a fractional Lévy process that is capable of capturing long and short memory in the stochastic process together with fat tails. The authors use the semi-martingale process in option pricing and empirically compare its performance to other option pricing models, including a stochastic volatility Lévy process. They contribute to the empirical literature by being the first to report the implied Hurst index computed from observed option prices using the Lévy process model. Calibrating the implied Hurst index of S&P 500 option prices in a period that covers the 2008 financial crisis, they find that the risk-neutral measure is characterized by a short memory in turbulent markets and a long memory in calm markets.
  • 详情 Forecasting Stock Market Volatility with Realized Volatility, Volatility Components and Jump Dynamics
    This paper proposes the two-component realized EGARCH model with dynamic jump intensity (hereafter REGARCH-C-DJI model) to model and forecast stock market volatility. The key feature of our REGARCH-C-DJI model is its ability to exploit the high-frequency information as well as to capture the long memory volatility and jump dynamics. An empirical application to Shanghai Stock Exchange Composite (SSEC) index data shows the presence of high persistence of volatility and dynamic jumps in China’s stock market. More importantly, the REGARCH-C-DJI model dominates the GARCH, EGARCH, REGARCH and REGARCH-C models in terms of out-of-sample forecast performance. Our findings highlight the importance of accommodating the realized volatility, volatility components and jump dynamics in forecasting stock market volatility.
  • 详情 Memory and Beliefs in Financial Markets: A Machine Learning Approach
    We develop a machine learning (ML) approach to establish new insights into how memory affects ffnancial market participants’ belief formation processes in the field. Using analyst forecasts as proxies for market beliefs, we extract analysts’ mental contexts and recalls that shape forecasts by training an ML memory model. First, we find that long-term memories are salient in analysts’ recalls. However, compared to an ML benchmark trained to fit realized earnings, analysts pay more attention to distant episodes in regular times but less during crisis times, leading to recall distortions and therefore forecast errors. Second, we decompose analysts’ mental contexts and show that they are mainly shaped by past earnings and forecasting decisions instead of current firm fundamentals as indicated by the ML benchmark. This difference in contexts further explains the recall distortion. Third, our comprehensive memory model reveals the significance of specific memory features and channels in analysts’ belief formation, including the temporal contiguity effect and selective forgetting.
  • 详情 Investor Memory and Biased Beliefs: Evidence from the Field
    We survey a large representative sample of retail investors to elicit their memories of stock market investment and return expectations. We then merge the survey data with administrative data of transactions to test a model in which investors form expectations by selectively recalling past experiences similar to the present cue. Our analysis not only uncovers newstylized facts about investor memory, but also provides support for similarity-based recall as a key mechanism of belief formation in ffnancial markets. Market ffuctuations affect investors’ recall: positive market returns cue investors to retrieve episodes of rising markets and recall own performances more positively. Recalled experiences explain a sizable fraction of cross-investor variation in beliefs and dominate actual experiences in explanatory power. We also show that recalled experiences can drive out the explanatory power of recent returns for expected future returns, ruling in a memory-based foundation for return extrapolation.
  • 详情 Volatility Long Memory on Option Valuation
    Volatility long memory is a stylized fact that has been documented for a long time. Existing literature have two ways to model volatility long memory: component volatility models and fractionally integrated volatility models. This paper develops a new fractionally integrated GARCH model, and investigates its performance by using the Standard and Poor’s 500 index returns and cross-sectional European option data. The fractionally integrated GARCH model signi?cantly outperforms the simple GARCH(1, 1) model by generating 37% less option pricing errors. With stronger volatility persistence, it also dominates a component volatility model, who has enjoyed a reputation for its outstanding option pricing performance, by generating 15% less option pricing errors. We also con?rm the fractionally integrated GARCH model’s robustness with the latest option prices. This paper indicates that capturing volatility persistence represents a very promising direction for future study.
  • 详情 China’s Stock Market Integration with a Leading Power and a Close Neighbor
    Current integration and co-movement among international stock markets has been boosted by increased globalization of the world economy, and profit-chasing capital surfing across borders. With a reputation as the fastest growing economy in the world, China’s stock market has continued gaining momentum during recent years and incurred growing attention from academicians, as well as practitioners. Taking into account economic and geographical considerations, the US and Hong Kong are considerably the most comparable stock markets to China. As the usual vector error correction model (VECM) could overlook the long memory feature of cointegration residual series, which can in turn exert bias on the resulting inferences, we chose to employ a fractionally integrated VECM (FIVECM) in this paper to investigate the long-term cointegration relations binding China’s stock market to the aforementioned stock markets. In addition, by augmenting the FIVECM with multivariate GARCH model, the return transmission and volatility spillover between market return series were revealed simultaneously. Our empirical results show that China’s stock market is fractionally cointegrated with the two markets, and it appears that China’s stock market has stronger ties with its neighboring Hong Kong market than with the world superpower, the US market.
  • 详情 Dynamic Behaviors of Mix-game Model and Its Applications
    This paper proposes a modification to Minority Game (MG) by adding some agents who play majority game into MG. So it is referred to as Mix-game. Through simulations, this paper finds out that the fluctuations of local volatilities change a lot by adding some agents who play majority game into MG, but the stylized features of MG don’t change obviously except agents with memory length 1 and 2. This paper also uses mix-game to model Shanghai stock market and to do prediction about Shanghai index.
  • 详情 Long Memory in Stock Trading Volume : Evidence from Indian Stock Market
    In this paper, we have examined the long memory property of Indian stock market by analyzing the trading volume series. Given the absence of trading volume index data, we have constructed trading volume series for the Indian stock market. We used maximum likelihood method to analyze the constructed trading volume index. The estimation of ARFIMA model, obtained a signi cant parameter for the order of fractional integration, and this could be consistent with the long autocorrelations observed in the trading volume series. The ndings that stock trading volume is a long memory process is robust, given di erent estimating methods, different subsamples, temporal aggregation and tests on individual stocks. Because of the conditional heteroscedasticity in the series, we have also carried out ARFIMAGARCH procedures to check whether long persistence were robust in the presence of conditional heteroscedasticity.