Google Trends could provide a neo-narrative approach to monetary policy analysis
If one go through the huge empirical literature on the real and nominal effects of monetary policy one will realize that most of this research is based on extremely simple and often very wrong measures of monetary policy. Hence, it is very common to use the central bank's key policy rate as a measure of the monetary policy stance. Anyway Market Monetarist would of course tell you interest rates is a very bad measure of the monetary policy stance and any conclusion drawn from empirical analysis of monetary policy based on interest rates is likely to be flawed. Unfortunately an analysis of monetary policy based on the money base is not much better as that is only telling us something about the supply of money, but nothing about the demand for money and therefore the development in the money base at base gives a incomplete picture of the monetary policy stance. Furthermore, using the interest rates or the money base as measures of the monetary policy stance tells us nothing about central bank guidance and therefore completely misses the fundamentally very important expectation channel in monetary policy. A good example of this is the Federal Reserve announcement of QE2 in August 2010. Fed chairman Bernanke announce a new round of quantitative easing (QE2) in at the end of August 2010, but the policy was not actually implemented before November 2010. New Keynesian, New Classical and Market Monetarist economists would all agree that fed started easing in August 2010 rather than in November 2010, but if one had looked the development in the US money base then one would have concluded that monetary easing did not start before November 2010. This means that any empirical analysis of monetary policy based on changes in the money base (or interest rates) will seriously underestimate the actual impact of monetary policy as the important expectational channel is totally ignored.