Two approaches in modelling the asset price process are regression model and stochastic model. What would be the major differences and pros and cons on risk management between the two approaches?Can regression model be used as a risk management tool (such as headging and derivative pricing)?

Is this a homework question? I don't mind helping with them, but if it is it would be a lot clearer with the course title and text.On the surface the question doesn't make much sense. The only asset price model I can think of that could be labeled "regression" is a covariance approach used in the Capital Asset Pricing Model and Riskmetrics, for examples. All interesting asset price process models are stochastic. "Reduced-form" is used in so many different ways that it doesn't clarify your question much. In finance, it is used almost entirely to refer to Beaver-Altman type predictions of bankruptcy based on accounting ratios.While most finance professionals understand its use in older econometrics literature (Lucas and Sargent), almost all useful financial models are reduced-form in this sense. In more recent econometrics, calling someone's model reduced-form seems to mean, "okay, you're right and I can't get an answer at all, but that's because I'm a deeper thinker than you." Structural/Reduced-form is the kind of hair-splitting that drove a lot of us out of economics into a field where hard data matters more than politics.