Open Economy Model
Tomas Motl, OGResearch, Course for ESF, Spring 2021
Open Economy Model
Introduction
Out of Model Calculations
Example: Biden stimulus package, analysis by the Global Projection Model Network
Tunes / Expert Judgment
Soft Tunes
Hard Tunes
External Assumptions
Introduction
We have model ( ).
We have initial condition ( ).
What do we get when we run a simulation into future? More or less smooth convergence
to steady-state. That's not a forecast, because it doesn't incorporate all available
information.
We can (need to) add shocks that will add dynamics. The shocks represent all future
expected events that will meaningfully impact macroeconomic variables - tax increases,
commodity price changes, fiscal policy, ...
Because QPM is reduced-form model, often one real life event is represented by a
combination of several shocks.
Out of Model Calculations
Model doesn't tell us the size of the shocks. Often we have to do calculations outside of
the model. Takes usually (far) more time than the work with model itself.
Example: Biden stimulus package, analysis by the Global
Projection Model Network
Tunes / Expert Judgment
Tunes are shocks we impose on the forecast. Also called expert judgment.
Consider model:
where , .
Soft Tunes
Simply set value of the shock directly.
,
We know the size of the shock, but we don't know what the resulting value of the variable
will be.
Hard Tunes
We set the value of a variable and ask Matlab/IRIS to calculate shock consistent with that
value. We need to specify which shock will explain the tune. Let's say we choose .
Because => .
Consider e.g. expected increase in policy rate. We know that policy rate will go up by 50
bp. We could calculate the shock needed to give us precisely the value, but it's pain. Hard
tuning the policy rate is much easier.
Note that choice of the shock matters. Policy rate can go up because of:
policy error (policy shock)
expected inflation increase (inflation shock)
expected tightening in foreign economy (foreign rate shock)
...
External Assumptions
There are variables in the model related to the foreign economy: foreign output gap,
foreign interest rate, oil price, food price, ... We will impose values for these variables over
the forecast horizon as hard tunes. Our model is not supposed to forecast the foreign
variables.
We need to get forecast for these variables from a reliable source. The GPMN, Consensus
Forecast, Bloomberg Consensus, ...