117 results from the Indonesian SSMM under the baseline scenario are discussed next in
Section 5.4. Then we perform simulations under several economic scenarios in Section 5.5. Stochastic simulations and alternative policy rule experiments will be presented and
analyzed thoroughly in Section 5.6. In doing so, we assess the effectiveness of alternative monetary policies for stabilizing Indonesian macroeconomic fluctuations based on a
policy frontier.
5.1 Datasets
In this section, we briefly describe the datasets used—their sources can be found in the Data Appendix. We also plot the graphs of the time series in that appendix. The domestic
output variable is represented by the quarterly real GDP of Indonesia from 1983:Q1 to 2004:Q1 seasonally adjusted, decomposed into non-oil output and oil output. We
employ 1993 as the base year for our study.
1
For oil output, we have to resort to interpolated data from 1983 to 1992. We employ the Chow-Lin 1971 method for best
linear unbiased distribution and interpolation of time series by related series, and use the oil price in Rupiah as the related series to convert annual oil output into quarterly oil
output.
2
The short-term nominal interest rate used in our analysis is the 1-month Sertifikat Bank Indonesia SBI rate, ranging from 1983:Q1 to 2004:Q1. Government expenditures
data is from 1983:Q1 up to 2004:Q1 and interpolated from the annual frequency using the
1
The reason is partly because the Indonesian statistical office started to compile quarterly data from 1993 see footnote 2.
2
The annual data for oil output from Indonesia’s statistical yearbook has two base years, namely 1983 and 1993. We rebased the real oil output from the 1983 base year to the 1993 base year using growth rates.
Subsequently, we obtained the quarterly oil price series as described in the text.
118 cubic spline method. The inflation rate is calculated by taking the difference in the
natural logarithm of CPI from 1983:Q1 to 2004:Q1, which approximate quarterly growth rates. We specify the inflation target as a constant path in three different sub-periods
1983:Q1–1995:Q4; 1996:Q1–1998:Q4; 1999:Q1–2004:Q1. This constant path is the average inflation rate in every sub period mentioned before. We construct the inflation
target in this manner because there are no officially announced inflation targets prior to the recent commitments of Bank Indonesia to move into inflation-targeting regime. The
money supply data used in our SSMM is the M1 measure.
In constructing the foreign income index, we use a geometric average of the GDPs of Indonesia’s major trading partners, weighted by nominal exports. These include
10 countries and 1 region: Malaysia, the Philippines, Thailand, Hong Kong, Singapore, South Korea, Taiwan, China, Japan, the U.S. and the rest of the OECD ROECD.
ROECD includes all members of the OECD except Japan, South Korea and the United States. Statistics for the ROECD are calculated as the weighted-average effect of all
countries in the group, so that the ROECD can be interpreted as one large “country”.
3
The construction of the nominal effective exchange rate follows the same principle as that of the foreign income index. Most, if not all, of Indonesia’s foreign debt is
denominated in US dollars, so we use the US 3-month Treasury Bill rate as the proxy for the foreign interest rate. Finally, we use the World CPI index from the International
Financial Statistics as the proxy for the foreign price level, with the base year rebased to
1993.
3
I thank the Econometric Studies Unit ESU of the Department of Economics in the National University of Singapore for making the data available for the purpose of this research.
119 Unit root tests are carried out on all the variables employed and the results are
shown in the Data Appendix. Where they are found to be integrated of order one i.e. I1, their changes are used in the estimation of the SSMM, except in the case of the LM
equation. The LM equation is specified and estimated in log levels because the variables in it are found to be cointegrated. The other exception to this rule is made for the non-oil
and oil output series, both of which are taken as deviations from their potential, which is measured by the Hodrick-Prescott filter, as done in most of the other SSMM studies cited
in Chapter 3.
5.2 Estimation Methods
