MCNK program

The MCNK program is a MatLab program with an Excel interface that is set up to replicate the results in Dees, Pesaran, Smith and Smith (DPSS, 2013). However, it is reasonably flexible to allow for alternative setups. It can be downloaded from the link below including all relevant documentation:

MCNK_August2014 (zip, 9.2 MB)

The associated Excel interface file is set-up to estimate a 33 country MCNK model, over the period 1980Q2-2011Q2. The estimation is conducted by inequality constrained IV, using lagged and contemporaneous foreign variables as instruments, subject to NK theoretical restrictions. The model is then solved to provide estimates of identified supply, demand and monetary policy shocks. Within a country, supply, demand and monetary policy shocks are orthogonal, though shocks of the same type (e.g. supply shocks in different countries) can be correlated. Impulse response functions and variance decompositions are computed, allowing for both direct channels of international transmission through regression coefficients and indirect channels through error spillover effects. Bootstrapped error bands are also provided for the cross country responses of a shock to the US monetary policy, as well as global supply and demand shocks.

The country-specific models include a Phillips curve determining inflation, an IS curve determining output, a Taylor Rule determining interest rates, a real effective exchange rage equation, with the addition of the price of oil (the US providing the numeraire). The MCNK thus explains output, inflation, short interest rates and exchange rates, but not long interest rates and equity prices which are included in the GVAR of Dees, di Mauro Pesaran and Smith (2007). This is because there are no widely accepted standard models for the term structure or equity premium. The MCNK has 131 endogenous variables all expressed as deviations from their steady states, which are estimated as long-horizon forecasts from a reduced-form cointegrating global VAR.

Note that the US real exchange rate variable does not enter the GVAR model as an endogenous variable (given that the US is the numeraire country). Consequently, the US real exchange rate deviation is not obtained directly from the multivariate decomposition of the GVAR, as discussed in DPSS (2013). See the MCNK documentation for more details.

As an alternative measure of steady states DPSS (2013) use the familiar Hodrick–Prescott (HP) filter for real output and following the literature assume that the other variables, namely inflation, the interest rate and the real exchange rate are stationary, and thus their steady states can be viewed as constants. Gauss code that estimates the trend path of the real output series using the HP filter can be downloaded from the link below, from which output deviations can then be computed:

HP_August2014 (zip, 135.1 KB)

*Minor discrepencies noted between these results and those reported in the supplement of DPSS (2010) (relating to the reduced-form cointegrating global VAR used to obtain a measure of the steady states), are due to a slight change in the calculation of the trade weights for the construction of the foreign US exchange rate in the MatLab code. Specifically, in the gauss code that generated the DPSS (2010) results, the trade weights used to construct the foreign US exchange rate were not re-weighted in light of the US domestic exchnage rate variable not entering the US model. In the MatLab code they are.