Market-Based Price Autocorrelation

TL;DR

This paper models how the randomness in trade values and volumes influences market prices and their autocorrelation, providing a framework to predict price behavior based on trade statistics.

Contribution

It derives the dependence of price moments and autocorrelation on trade value and volume statistics, highlighting the impact of trade randomness on price dynamics.

Findings

Price autocorrelation depends on trade value and volume correlations.

Market trade randomness significantly affects price volatility.

Forecasting autocorrelation requires predicting trade statistical moments.

Abstract

This paper assumes that the randomness of market trade values and volumes determines the properties of stochastic market prices. We derive the direct dependence of the first two price statistical moments and price volatility on statistical moments, volatilities, and correlations of market trade values and volumes. That helps describe the dependence of market-based price autocorrelation between times t and t-{\tau} on statistical moments and correlations between trade values and volumes. That highlights the impact of the randomness of the size of market deals on price statistical moments and autocorrelation. Statistical moments and correlations of market trade values and volumes are assessed by conventional frequency-based probabilities. The distinctions between market-based price autocorrelation and autocorrelation that are assessed by the frequency-based probability analysis of price time series reveal the different approaches to the definitions of price probabilities. To forecast market-based price autocorrelation, one should predict the statistical moments and correlations of trade values and volumes.