Have you ever wondered how well stock prices and trading volume can be analyzed together? Scientists have come up with a tweaked version of an existing model, called the Barndorff-Nielsen and Shephard stochastic volatility model. Instead of trying to figure out instantaneous variance, which is hard to observe, this model Looks at something more concrete: trading intensity. This could be the number of trades or trading volume.

The cool thing is, they’ve developed a special estimator that uses a method called martingale estimating functions. This estimator works in a special bivariate model that doesn’t follow the usual rules of diffusion but can handle sudden jumps. It’s like trying to predict the stock market, but with sudden leaps factored in.

The model assumes that both stock prices and trading volume are recorded at fixed intervals, and these observations stretch over a long period. The estimator was proven to be very reliable (consistent) and follows a predictable pattern (asymptotically normal). Plus, we get a clear idea of how tightly these values are connected through an explicit expression of the asymptotic covariance matrix.

To see how well this estimator works, scientists ran a finite sample experiment along with a statistical analysis on two big companies: IBM from the New York Stock Exchange and Microsoft from Nasdaq. Both were observed over five years.

Exploring Stock Prices and Trading Activity with a New Model Twist

Actions