Testing for neglected monetary effects on inflation
The ICM equation for aggregate consumer price inflation in Table 8.12 contains three key sources of inflation impulses to a small open economy: imported inflation including currency depreciation (a pass-through effect), domestic cost pressure (unit labour costs), and excess demand in the product market. Monetary shocks or financial market shocks may of course generate inflation
impulses in situations where they affect one or more of the variables associated with these inflation channels. In this section, we will investigate another possibility, namely that shocks in monetary or financial variables have direct effects on inflation which have been neglected in the ICM. Results for Denmark in Juselius (1992) indicate that ‘monetary variables’ are important explanatory variables in an empirical model for Danish inflation and that they have clearly significant direct effects. In the following, we test the robustness of the ICM inflation equation in Table 8.12 with respect to neglected monetary effects on inflation, simply by subjecting this equation to a sequence of omitted variables tests. In turn we test the significance of current and lagged money growth (Amt, ■ ■ ■ , Amt-4), real interest rate (RT—A4p)t-i, interest rate spread (RB — RT)t-i, the lagged equilibrium correction term from the broad money demand equation ecmmdt_i, credit gap (cr — cr*)t_i, and all gap variables from the P*-models above (ygapt_i, A4mgapt_i, A4pgapt_i, RBRTgapt_i, rmgapt_i).
The results in Table 8.16 show that neither of these variables are significant when they are added to the ICM price equation. The same results hold for these variables irrespective of whether we test their significance simultaneously or include the variables one at a time. The lagged equilibrium correction term for broad money, ecmmdt_i, is clearly insignificant when it is added to the price equation. This is an important result, since it provides corroborative evidence that prices are weakly exogenous for the parameters in the long-run money demand relationship. This is a plausible finding from a theoretical point of view, and it is also in line with empirical evidence found in a series of previous studies, including Hoover (1991), Bardsen (1992),
Omitted variable tests (OVT) for neglected monetary effects
on inflation in the ‘reduced form’ ICM price equation
Interest rates, excess money and credit
Fovt (5,109) = 0.2284[0.9494] Fovt (1,113) = 0.0328[0.8565] Fovt (1,113) = 0.3075[0.5803] Fovt (1,113) = 0.1302[0.7189] Fovt (1,113) = 0.5173[0.4735]
‘Gap’ variables from the P-star model gdpgapt_i Fovt (1,113) =
A4mgapt_i Fovt (1,113) =
A4pgapt_i Fovt (1,113) =
RBRMgapt_i Fovt (1,113) =
rmgapt_i Fovt (1,113) =
Joint all five above Fovt(5,109) =
Hendry and Ericsson (1991), Engle and Hendry (1993), and Hendry and Mizon (1993).