Would it make sense to do a 2SLS regression for Nominal GDP and Money Supply (using predetermined variables of government expenditure and investment)?

If it does make theoretical sense, is this because GDP and Money Supply are endogenous? I am unsure why GDP is endogenous... For example if there is an increase in money supply, then I know there is an increased in GDP (since AD increases); however, I fail to see how an increase in GDP has a feedback effect and causes an increase in money supply.

  • $\begingroup$ Could you edit your title to be more descriptive? This would provide a better incentive for users to offer you a good answer. $\endgroup$
    – Aarthi
    Commented Mar 7, 2012 at 19:01
  • $\begingroup$ Incidentally, economists don't use money supply measures (such as M1, M2, etc) anymore since they don't predict anything in recent decades. Instead, the Fed funds rate is what determines monetary policy. $\endgroup$
    – Ronaldo Carpio
    Commented Mar 18, 2012 at 5:30

1 Answer 1


Money supply is endogenous because policy makers adjust the money supply in response to GDP. While it is certainly possible to use 2sls for this problem (assuming you think you have some good instruments), the more common approach is to use systems of equations--vector autoregression (VAR) models, in particular.

See Sims (1980) for a discussion on the merits of using VAR for this type of problem.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.