139 the model while treating the past and future values as fixed. The loop is repeated until
changes in the values of the endogenous variables between successive iterations become less than a specified tolerance. This method is often referred to as the Fair-Taylor
method.
7
5.4.4 Baseline Simulation Results
In this section, we will present our first set of simulation results for the Indonesian SSMM in the dynamic-deterministic DD mode by comparing the baseline simulation
results of the model with actual historical outcomes. In DD baseline simulations, our aim is to examine how the model performs when used to forecast many periods into the
future. Specifically, the results shown in Figures 5.2–5.8 illustrate how our model would have performed if we had used it back in 1983 to make a forecast for the Indonesian
economy over the next twenty years, assuming that we had used the correct paths for the exogenous variables.
8
All the structural equations perform well during the period prior to the Asian financial crisis. However, the simulated total output, non-oil output, money supply, and
exchange rate show deviations from actual outcomes after 19971998—a period when Indonesia was experiencing economic turmoil—although they do seem to follow the
general trends in the data. In the case of the exchange rate, the discrepancy between the simulated and actual values most likely reflects a time-varying risk premium on the
Indonesian Rupiah which our model did not allow for. The oil output and the inflation
7
Note, however, that the usual Fair-Taylor algorithm includes a particular handling of terminal conditions the extended path method that is slightly different from the option that we use.
8
In reality, we would not have known these values at the time the forecasts were generated.
140 equations show a good fit for the entire period under study. As for the Taylor rule, the
simulated series broadly captures the cycles in the domestic interest rate but it is not able to reproduce the latter’s volatility. This may not matter much since we are going to
experiment with different coefficients in the rule when performing stochastic simulations later.
Figure 5.2 Baseline Simulation Results Endogenous Variable: Total Output
40000 50000
60000 70000
80000 90000
100000 110000
120000
1984 1986
1988 1990
1992 1994
1996 1998
2000 2002
2004 TOTAL OUTPUT Baseline
TOTAL OUTPUT Actual
141
Figure 5.3 Baseline Simulation Results Endogenous Variable: Non Oil Output
30000 40000
50000 60000
70000 80000
90000 100000
110000
1984 1986
1988 1990
1992 1994
1996 1998
2000 2002
2004 NON OIL OUTPUT Baseline
NON OIL OUTPUT Actual
Figure 5.4 Baseline Simulation Results Endogenous Variable: Oil Output
5000 6000
7000 8000
9000 10000
1984 1986
1988 1990
1992 1994
1996 1998
2000 2002
2004 OIL OUTPUT Baseline
OIL OUTPUT Actual
142
Figure 5.5 Baseline Simulation Results Endogenous Variable: Inflation
-.04 .00
.04 .08
.12 .16
.20
1984 1986
1988 1990
1992 1994
1996 1998
2000 2002
2004 INFLATION Baseline
INFLATION Actual
Figure 5.6 Baseline Simulation Results Endogenous Variable: Money Supply
100000 200000
300000 400000
500000 600000
700000
1984 1986
1988 1990
1992 1994
1996 1998
2000 2002
2004 MONEY SUPPLY Baseline
MONEY SUPPLY Actual
143
Figure 5.7 Baseline Simulation Results Endogenous Variable: Exchange Rate
100 200
300 400
500 600
1984 1986
1988 1990
1992 1994
1996 1998
2000 2002
2004 EXCHANGE RATEBaseline
EXCHANGE RATE Actual
Figure 5.8 Baseline Simulation Results Endogenous Variable: Domestic Interest Rate
.0 .1
.2 .3
.4 .5
.6 .7
1984 1986
1988 1990
1992 1994
1996 1998
2000 2002
2004 DOMESTIC INTEREST RATE Baseline
DOMESTIC INTEREST RATE Actual
144
5.5 Scenario Analysis
