Noah Smith’s Conclusion on DSGE Models

In my last blog post I summarized and reflected on Noah Smith’s article The Most Damning Critique of DSGE. The blog post expanded on Smith’s opinion of utilizing DSGE models in the financial sector but I neglected to note one of the most important points made. Noah claims that there are numerous different variations of DSGE models that can be used, so it is difficult for firms to choose the “correct” one.  The potential benefits of using DSGE models, the main one being that they satisfy the Lucas-critique, have been overshadowed by the costs of using them or determining the most appropriate one to use.

Smith determines that the lack of use of  DSGE models, specifically Lucas-robust  DSGE models, in finance markets are an indicator that “DSGE fail  the market test” (Smith 2014). This means that financial modelers who would benefit directly from superior market returns uniformly do not use DSGE models, thus strongly suggesting that DSGE models are not useful for macroeconomic predictions. I found this extrapolation to be interesting, but not necessarily supported firmly in the article.

Noah Smith’s blog post can be accessed through the following link