News or Noise? Signal Extraction Can Generate Volatility Clusters From IID Shocks

We develop a framework in which information about firm value is noisily observed. Investors are then faced with a signal extraction problem. Solving this would enable them to probabilistically infer the fundamental value of the firm and, hence, price its stocks. If the innovations driving the fundamental value of the firm and the noise that obscures this fundamental value in observed data come from non-Gaussian thick-tailed probability distributions, then the implied stock returns could exhibit volatility clustering. We demonstrate the validity of this effect with a simulation study.