Spot Variance Path Estimation and Its Application to High-Frequency Jump Testing

Bos, Charles S.; Janus, P.; Koopman, Siem Jan

doi:10.1093/jjfinec/nbr013

Cited by 18 publications

(13 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…in which κ(∆t) could increase even a 100-trial average is performed. Note that the loss of the aggregational Gaussianity in short time scales is not only a phenomenon appearing in this model, but also an observation in the real market [56]. In fact, Johnson et al [39] had pointed out that a time scale of ∆t ≥ 60 min is generally required in order to approximate the probability distribution of return as Gaussian distribution.…”

Section: Aggregational Gaussianitymentioning

confidence: 86%

Development of an agent-based speculation game for higher reproducibility of financial stylized facts

Katahira

Chen

Hashimoto

et al. 2019

Physica A: Statistical Mechanics and its Applications

View full text Add to dashboard Cite

Simultaneous reproduction of all financial stylized facts is so difficult that most existing stochastic process-based and agent-based models are unable to achieve the goal. In this study, by extending the decision-making structure of Minority Game, we propose a novel agent-based model called "Speculation Game," for a better reproducibility of the stylized facts. The new model has three distinct characteristics comparing with preceding agent-based adaptive models for the financial market: the enabling of nonuniform holding and idling periods, the inclusion of magnitude information of price change in history, and the implementation of a cognitive world for the evaluation of investment strategies with capital gains and losses. With these features, Speculation Game succeeds in reproducing 10 out of the currently well studied 11 stylized facts under a single parameter setting. nancial time series [1]. The 11 currently reported stylized facts can be listed as follows: volatility clustering, intermittency, heavy tails, the absence of autocorrelation in returns, slow decay of autocorrelation in volatilities, volume/volatility correlation, aggregational Gaussianity, conditional heavy tails, asymmetry in time scales, leverage effect and gain/loss asymmetry. They are quite nontrivial features which have been well studied for many years in different markets and for various instruments.The reproducibility of the stylized facts is a prerequisite for the financial market model. However, to reproduce them with a conventional market model, whether a simple or sophisticated stochastic process, is not very easy due to the high level of multiplexity of the market. With the contribution of econometrists, several well known stochastic models have been constructed to yield typical stylized features, for example, the autoregressive conditional heteroscedasticity (ARCH) and generalized autoregressive conditional heteroscedasticity (GARCH) processes [2,3]. Nonetheless, to the best of our knowledge, most existing models fail to reproduce the whole set of stylized properties simultaneously.On the other hand, an alternative approach mainly advocated by econophysicists, which adopts agent simulation based on a set of simple rules, has been used to study the stylized facts as well. The simulation enables the reproduction and analysis of complex price dynamics in an artificial market with autonomous agents acting as the traders. Compared with previous stochastic process models, which are designed to describe the traits of financial time series at the macroscopic level, this method is a bottom-up approach which explains these features at the microscopic level, that is, the level at which traders' decision-making and trading behaviors are taking into account.There are mainly two reasons for adopting agent-based models. First, the agent-based simulation is almost the only way to analyze the effects of traders' behavior to the whole market. All time-series data obtained from the real market, such as price and volume, are aggregational ones, which i...

show abstract

Section: Aggregational Gaussianitymentioning

confidence: 86%

Development of an agent-based speculation game for higher reproducibility of financial stylized facts

Katahira

Chen

Hashimoto

et al. 2019

Physica A: Statistical Mechanics and its Applications

View full text Add to dashboard Cite

show abstract

“…We also compare the accuracy of σt iK 2 with the TTS estimator and the modulated bipower-type estimator, respectively, employed by Sun (2019) and Bos et al (2012) in estimating the SV. To approximate 𝜎 t iK 2 , the TTS estimator used by Sun (2019) is defined as follows:…”

Section: F I G U R Ementioning

confidence: 99%

“…Assuming the drift is driven by a linear process of the efficient price, Laurent and Shi (2020) propose some correction to the SV estimator used in Lee and Mykland (2012) by taking the drift effects into account. However, both Bos et al (2012) and Laurent and Shi (2020) do not theoretically justify the consistency of their tests in the presence of the noise. Sun (2019) assumes the existence of arbitrarily high moments for the noise terms, which implies that the noises have thin tail distributions.…”

mentioning

confidence: 96%

Testing for jumps with robust spot volatility estimators

Sun

2023

Statistica Neerlandica

View full text Add to dashboard Cite

Jumps in the paths of efficient asset prices have important economic implications. Motivated by the issue of testing for jumps based on noisy high‐frequency data, we develop a novel spot volatility estimator, which is obtained by minimizing the sum of some Huber loss functions, and use it as an ingredient for jump detection. This type of estimators is uniformly consistent in estimating the spot volatilities of the efficient price at numerous time points. We further demonstrate the consistency of the proposed jump test based on the property of the novel spot volatility estimator. We show that in finite samples, the proposed volatility estimator and the test perform favorably compared to some competitors through Monte Carlo simulations. We also illustrate our methodology with the stock prices of Apple and Microsoft.

show abstract

“…We will set K = ⌈N 3/4 ⌉ below. Such a kind of spot variance estimator was studied in Bos et al [11]. Then we choose…”

Section: Choice Of the Threshold Processesmentioning

confidence: 99%

Estimation of Integrated Covariances in the Simultaneous Presence of Nonsynchronicity, Microstructure Noise and Jumps

Koike

2013

SSRN Journal

View full text Add to dashboard Cite

We propose a new estimator for the integrated covariance of two Itô semimartingales observed at a highfrequency. This new estimator, which we call the pre-averaged truncated Hayashi-Yoshida estimator, enables us to separate the sum of the co-jumps from the total quadratic covariation even in the case that the sampling schemes of two processes are nonsynchronous and the observation data is polluted by some noise. It is the first estimator which can simultaneously handle these three issues, which are fundamental to empirical studies of high-frequency financial data. We also show the asymptotic mixed normality of this estimator under some mild conditions allowing infinite activity jump processes with finite variations, some dependency between the sampling times and the observed processes as well as a kind of endogenous observation errors. We examine the finite sample performance of this estimator using a Monte Carlo study.

show abstract

Spot Variance Path Estimation and Its Application to High-Frequency Jump Testing

Cited by 18 publications

References 43 publications

Development of an agent-based speculation game for higher reproducibility of financial stylized facts

Development of an agent-based speculation game for higher reproducibility of financial stylized facts

Testing for jumps with robust spot volatility estimators

Estimation of Integrated Covariances in the Simultaneous Presence of Nonsynchronicity, Microstructure Noise and Jumps

Contact Info

Product

Resources

About