Working Paper: NBER ID: w0904
Authors: John F. Helliwell; Paul M. Boothe
Abstract: In this paper we estimate and compare several alternative exchange rate models that have received wide attention, but little comparison, during the 1970s. In order to compare purchasing power parity (PPP), nominal interest rate parity, real interest rate parity, and portfolio balance models, we first strip each down to its essential core and undertake comparable single-equation tests of both 'hard' and 'easy' (more and less constrained) versions of each model. We then embed each of the 'hard' versions in a new macroeconomic model of Canada, and assess their implications for the impacts of monetary and fiscal shocks. Using annual Canadian data from the 1950s and 1970s, all of the models have single-equation errors of about 3%, except for the 'hard' versions of PPP and real interest parity, which are heavily rejected by the data. In a macroeconomic context, the models have modestly different implications for the effects of fiscal shocks, and diverge more widely under monetary shocks.
Keywords: exchange rate models; purchasing power parity; nominal interest parity; real interest parity; portfolio balance
JEL Codes: F31; E52
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Purchasing Power Parity (PPP) (F31) | depreciation of the domestic currency (F31) |
| Real Interest Parity (RIP) (E43) | depreciation of the domestic currency (F31) |
| Nominal Interest Parity (NIP) (E43) | different response to monetary shocks (E19) |
| Portfolio Balance Model (G11) | different response to monetary shocks (E19) |
| Fiscal Expansion (E62) | initial multiplier responses are similar across models (C30) |
| Fiscal Expansion (E62) | diverge significantly over time (J19) |
| Choice of Exchange Rate Model (F31) | dynamic responses of the macroeconomic model (E19) |