State Space Models

All state space models are written and estimated in the R programming language. The models are available here with instructions and R procedures for manipulating the models here here.

Tuesday, March 8, 2011

About

The title for this blog is taken from an influential book by Nelson Goodman (here) Fact, Fiction and Forecast. In the book, Goodman explores deep issues raised by simple statements: "Global temperature is increasing" (fact, but how do we know this?), "Global temperature would not be increasing without human interventions such as fossil fuel burning" (a fictional counterfactual statement implying a causal link) and "Global temperature will increase above 2 degrees C in 2100" (a forecast, but how can we know the future?)

We don't just have to focus on the Intergovernmental Panel on Climate Change (the IPCC). Consider the current Global Financial Crisis that started in 2007: The financial crisis was caused by a meltdown in the subprime mortgage market (fact?). Had there been stronger financial regulation, the Global Financial Crisis would never have happened (counterfactual?). Had the US government not bailed out the big banks, the financial crisis would have been even deeper (counterfactual). There will continue to be financial crises in the future since financial crises are an inevitable part of the capitalist system (forecast).

Of course, I could go on (and will in this blog)! Nelson Goodman's work has been extended by Judea Pearl (here) in Causality: Models, Reasoning and Inference, by Kevin D. Hoover Causality in Macroeconomics and by others. I hope to discuss their work in future posts.

I also hope to use systems models to establish historical facts, run historical counterfactuals and make forecasts. My results have to be carefully evaluated against the philosophical work on causality: How do I know the historical data I am using is factual (does anyone believe statistics being published by authoritarian regimes)? Are my models good enough to be used for serious historical counterfactuals (just how would that be done)? Finally, how can we really know anything about the future?

No comments:

Post a Comment