X

Stay ahead of the curve with CGD Development Updates

(if not affiliated type "individuals")
Working Papers

An Identity Crisis? Testing IMF Financial Programming - Working Paper 9

8/1/02
Download PDF
Share

The IMF uses its well-known "financial programming" model to derive monetary and fiscal programs to achieve desired macroeconomic targets in countries undergoing crises or receiving debt relief. Financial programming is based on monetary, balance of payments, and fiscal accounting identities. This paper subjects the identity-based framework to a variety of tests. All of the identities contain large statistical discrepancies, which weakens the case for them as a "consistency check." Financial programming assumes a one for one relationship from the identity between the policy variable (e.g. domestic credit) and the outcome variable (e.g. money supply) posited by financial programming, because the other variables in the identity are assumed to be exogenous with respect to the policy variable. This assumption fails in the data, as all the coefficients of outcome variables on policy variables depart from a unitary coefficient. The elasticity of inflation with respect to excess money growth (money growth — real output growth) is significantly less than one, and shows a high variance in the data. Changes in velocity account on average for 57% of the change in the price level. Velocity is non-stationary. Imports are not significantly related to long-term disbursements in most countries. The median income elasticity of imports is 1.36 and the dispersion of import elasticities in the data has a majority of the distribution outside the usual range used in country projections. Using import availability to predict growth leads to a forecast error more than twice that of the naïve model that growth is a random walk. Government deficits do not have a one for one link with domestic credit creation, as predicted by the identity approach. In sum, the financial programming approach is flawed because it does not take into account the endogeneity of virtually all the variables in each macroeconomic identity, the instability of its simple behavioral assumptions, and the large statistical discrepancies in all the identities. Accounting identities do not a macro model make.