Information about Asia and the Pacific Asia y el Pacífico
Journal Issue
Share
Article

Indonesia: Selected Issues

Author(s):
International Monetary Fund
Published Date:
August 2006
Share
  • ShareShare
Information about Asia and the Pacific Asia y el Pacífico
Show Summary Details

VI. Explaining Higher Inflation in Indonesia—A Regional Comparison1

A. Introduction

1. Over the past 15 years inflation in Indonesia has been consistently higher than elsewhere in the region.2 Indeed, inflation has often exceeded the 7–11 percent threshold above which it is estimated to adversely affect growth (Khan and Senhadji, 2000). At these rates, inflation may also make the poor significantly worse off by reducing real minimum wages and the income share of the lowest quintile (Easterly and Fischer, 2001). The higher inflation rate and its potential adverse effects raise the question: what is driving the Indonesian inflation differential vis-à-vis its neighbors?

2. This chapter reviews a number of stylized facts comparing inflation in Indonesia with other Asian countries. It uses econometric techniques to assess various hypotheses to explain the Indonesian inflation differential vis-à-vis neighboring countries. Finally, it discusses the policy implications for reducing inflation in Indonesia towards rates prevailing in the region.

B. Indonesian Inflation in the Asian Context

3. Inflation in Indonesia has been substantially higher than in the other countries of the region, both before and after the crisis. Over the past one and a half decades, the annual inflation rate in Indonesia has averaged about 12.5 percent, or 9 percentage points higher than in neighboring countries, and has shown significantly higher variance (Figure 1 and Table 1). While the magnitude of the inflation differential reflects in part the high inflation rates registered in Indonesia during the Asian financial crisis, the differential widened in the period after the crisis as inflation in neighboring countries declined, while remaining broadly unchanged in Indonesia (Figure 2).

Table 1.Comparing Headline Inflation in Selected Asian Economies (1991–2005)
IndonesiaKoreaMalaysiaPhilipppinesSingaporeThailandDifferential
Annual average
1991-200512.54.53.07.51.43.89.0
Precrisis 1/8.25.83.89.42.35.13.5
Post crisis 1/8.43.21.75.10.82.26.2
19919.49.34.318.73.45.72.5
19927.56.24.88.92.34.22.7
19939.74.83.67.62.33.35.7
19948.56.34.19.13.15.13.4
19959.44.53.56.91.75.85.4
19967.04.93.59.01.45.92.8
19976.24.42.65.92.05.62.4
199858.07.55.19.7−0.38.1
199920.70.82.86.70.00.320.1
20003.82.31.64.31.31.61.8
200111.54.11.46.11.01.79.2
200211.82.81.83.0−0.40.6
20036.83.51.13.50.51.85.1
20046.13.61.46.01.72.83.4
200510.52.83.07.60.54.57.8
End-Period, 12-month percent change
1991-200513.34.22.97.11.33.79.9
Crisis, 199877.54.05.210.3–1.44.372.0
Precrisis 1/8.45.93.78.52.15.33.8
Post crisis 1/8.92.51.95.20.82.16.8
19919.99.34.313.12.94.64.1
19925.04.54.88.21.83.11.1
199310.25.83.48.42.64.45.7
19949.65.64.07.22.94.85.0
19959.04.83.18.60.87.45.2
19965.14.93.47.12.04.81.0
199710.36.62.77.32.07.65.7
199877.54.05.210.3−1.44.3
19992.01.42.54.30.70.70.5
20009.30.01.36.72.11.4
200112.53.21.24.1−0.60.8
20029.93.71.72.50.41.68.3
20035.23.41.23.90.71.83.3
20046.43.02.18.61.32.93.5
200517.12.63.26.61.35.813.8
Annual standard deviation
1991-20054.70.60.41.20.50.94.0
Crisis, 1998-9923.80.90.91.80.81.922.5
Precrisis 1/1.20.60.41.30.40.90.6
Post crisis 1/2.30.50.30.90.50.71.7
19910.50.50.62.20.60.6−0.2
19922.11.10.40.70.30.81.6
19930.70.50.60.60.20.60.3
19940.90.50.41.10.40.40.4
19950.70.60.31.30.60.90.0
19961.70.30.22.00.31.21.2
19971.90.80.50.90.31.51.2
199823.31.40.81.31.12.222.0
199924.20.41.02.30.51.623.3
20003.90.70.21.20.50.53.4
20011.60.80.21.00.90.61.0
20021.90.50.40.50.50.51.5
20031.00.40.30.40.40.30.6
20040.80.50.51.70.50.70.1
20054.50.40.40.70.51.43.9
Source: IMF, International Financial Statistics database.

Precrisis includes 1991–97. Post crisis include 2000–05, except for end-year inflation that also includes 1999.

Source: IMF, International Financial Statistics database.

Precrisis includes 1991–97. Post crisis include 2000–05, except for end-year inflation that also includes 1999.

Figure 1.Average Annual Inflation Rates Across Selected Asian Countries, 1991-2005

Source: CEIC database.

Figure 2.Average Annual Inflation Rates Across Selected Asian Countries

Sources: CEIC database and Fund staff calculations.

4. Inflation in Indonesia has been higher across a wide range of products. Indonesia has registered a positive inflation differential vis-à-vis the other countries in the sample across all the main components of the CPI. After the crisis, the differential widened for most of the CPI components, with the exception of food items. Indeed, contrary to the commonly held belief in Indonesia that high inflation is the result of distortions in the agricultural sector and weak rural infrastructure, the inflation differential for food items has decreased after the crisis, while widening for all the other categories in the CPI, particularly housing, education, and transportation and communications (Figure 3). In sum, no single CPI component can alone explain the Indonesian inflation differential with respect to other Asian countries; therefore, other factors need to be explored.

Figure 3.Indonesia Inflation Differential Across CPI Components 1/

1/ Other Asian countries include: Korea, Malaysia, The Philippines, Singapore and Thailand.

Source: Fund staff calculations.

C. What Can explain Inflation in Indonesia and in Neighboring Countries?

Theories explaining inflation across countries

5. Different factors have been analyzed in the literature to explain the sources of inflation and inflation differentials across countries. It is generally accepted that inflation is:

  • A phenomenon with some degree of inertia due, for example, to the way expectationsare formed. Inflation expectations are in part adaptive or backward looking, particularly in countries that, like Indonesia, have experienced long periods of high inflation (Mankiw and others, 2003).
  • Closely related to country-specific shocks. These include demand and supply shocks associated, for example, with the pace of economic activity (Coe and McDermott, 1997), the stance of monetary policy, and exchange rate fluctuations (Siregar, 2002). High demand pressures, expansionary monetary policy (e.g., rapid growth in monetary aggregates) and significant currency depreciations have been found to be positively correlated to inflation across countries (Anglingkusumo, 2005; Campillo and Miron, 1996).
  • Dependent on the structural features of the economy. For example, the degree of central bank independence (Berger and others, 2001), economic openness (Romer, 1993), the public debt burden (Campillo and Miron, 1996), and the type of exchange rate regime (Loungani and Swagel, 2001) are found to have an impact on the rate of inflation across countries. Central bank independence is generally found to reduce inflation, especially in less developed countries, as it helps to insulate monetary policy from political influences. Inflation is often found to be negatively associated with economic openness, as this increases the costs of unanticipated monetary expansion and allows for additional productivity gains and price competition (IMF, 2006). Fiscal imbalances may also lead to higher inflation either by triggering higher money growth or forcing currency depreciation.
  • Related to the degree of political stability and institutional development. For example, frequent cabinet changes and weak institutions shorten the time horizon of governments and make difficult the pursuit of consistent and sound policies to maintain low inflation. A number of political and institutional variables have been found to affect inflation, particularly in developing countries (Aisen and Veiga, 2005; Cukierman, Edwards and Tabellini, 1991).

Some simple facts to explain inflation in Indonesia and neighboring countries

Some factors explaining inflation seem to play a more important role in Indonesia than in neighboring countries.

  • Inflation inertia appears to be stronger in Indonesia than in the other countries of the sample. Countries that have had relatively high inflation rates in the past (e.g., in the pre-crisis period) have registered higher inflation rates in more recent years. In this respect, Indonesia seems to face stronger inflation inertia than the other countries as inflation has been higher after the crisis than before the crisis (Figure 4).
  • The positive relationship between inflation and the output gap is stronger in Indonesia. As expected, smaller output gaps are associated with higher inflation rates across all the countries of the sample (Figure 5). Once again, Indonesia appears to be an outlier with much higher inflation for a given output gap (Table 2).3
  • Political and institutional factors affect inflation differently across countries. Political risks and government instability (measured by the International Country Risk Guide (ICRG)’s indices) are positively correlated with inflation across selected Asian countries. However, Indonesia is once again an outlier in the region with high political risks associated with relatively higher inflation (Figures 67).4

Figure 4.Pre-Crisis and Post-Crisis Average Inflation in Selected Asian Countries

Sources: CEIC database and Fund staff calculations.

Figure 5.Average Inflation Versus Average Output Gap (HP filter) in Selected Asian Countries, 1991-2005

Sources: CEIC database and Fund staff calculations.

Table 2.Inflation Correlations in Selected Asian Countries and Indonesia 1/
Asian CountriesIndonesia
Output gap−0.16−0.47
Political risks0.430.39
M2 growth0.620.79
NEER perc. change−0.58−0.69
Source: Fund staff calculations.

Asian countries include: Korea, Malaysia, Philippines, Singapore, Thailand, and Indonesia

Source: Fund staff calculations.

Asian countries include: Korea, Malaysia, Philippines, Singapore, Thailand, and Indonesia

Figure 6.Average Inflation Versus Average Government Instability (ICRG) in Selected Asian Countries

Source: Government stability from ICRG re-based. Higher index indicates government instability.

Figure 7.Average Inflation Versus Average Political Risk (ICRG) in Selected Asian Countries

Source: Political risk from ICRG re-based. Higher index indicates higher risk.

In other respects, Indonesia is similar to other Asian countries…

  • The influence of changes in monetary aggregates and exchange rates on inflation in Indonesia is broadly similar to that in neighboring countries (Figures 89). In Indonesia, monetary growth and currency depreciation (NEER) are on average higher, but the degree of correlation with inflation appears to be similar to that in other countries (Table 2). Importantly, the more expansionary monetary policy and the higher average inflation rate in Indonesia suggest that the country’s inflation differential may in part be a monetary phenomenon.

Figure 8.Average Inflation Versus Average Exchange’Rate (NEER) in Selected Asian Countries

1/ Lower NEER indicates depreciation.

Sources: CEIC database and IMF, Information Notice System database

Figure 9.Average Inflation Versus Average Money Growth (M2) in Selected Asian Countries

Sources: CEIC database and IMF, International Financial Statistics database.

6. Overall, Indonesia appears to differ from neighboring countries in the way some structural factors relate to inflation while being similar in other respects. An econometric analysis is therefore needed to clarify the differences and similarities and to explore the reasons underlying the Indonesia inflation differential vis-à-vis its neighboring countries.

D. An Econometric Analysis

7. In this section, econometric analysis is used to explain why inflation in Indonesia has been higher than in the region. The analysis is carried out in three steps. First, a cross–country empirical model identifying the main inflation determinants across our sample is estimated (Box 1).5 Second, a set of Indonesia-related slope dummies is employed to investigate whether the role of the inflation determinants differs in Indonesia from the average of the selected countries.6 Finally, the causes of Indonesia’s higher inflation are examined by looking at the combined effect of two elements: how inflation determinants have evolved in Indonesia compared to other countries in the sample and how they have differently affected inflation in Indonesia compared with other countries (using the coefficients of the dummy variables). In this framework, the basic econometric model sets a benchmark against which to test different hypotheses about the reasons for the higher inflation in Indonesia.

Step 1—Explaining Cross-Country Inflation in Selected Asian Countries

8. A simple model explains inflation well in the selected Asian countries. Across the sample, the rate of inflation depends positively on past inflation, the output gap, currency depreciation, and growth in M2 (Table 3). Inflation also depends on institutional factors as measured, for example, by the ICRG’s political risk index. In particular, government instability and the quality of national bureaucracies are the two institutional factors that make the strongest contribution to inflation.7 Structural factors such as the degree of economic openness, the public debt burden, and the level of price regulation (as measured by an index from the Heritage Foundation) play no role in shaping inflation dynamics across the selected countries (these results are not reported).

Table 3.Determinants of Cross-Country Inflation in Selected Asian Countries
Model (1)Model (2)
Previous inflation0.64***0.6***
Output gap0.12*0.11**
M2 Growth0.03***0.03***
NEER growth 1/−0.2**−0.2**
Political risk (ICRG)0.1**
Government instability (ICRG)0.14***
Lack of bureaucracy quality (ICRG)0.59*
Chi-square0.660.00
Obs.348348
Notes: Difference GMM estimations using quarterly data 1990Q4-2005Q4.Independent variable: End-period inflation rate.Countries: Indonesia, Korea, Malaysia, Philippines, Singapore, Thailand.*,**,*** denote significance at 10, 5, 1 percent, respectively.

Negative changes denote depreciation.

Notes: Difference GMM estimations using quarterly data 1990Q4-2005Q4.Independent variable: End-period inflation rate.Countries: Indonesia, Korea, Malaysia, Philippines, Singapore, Thailand.*,**,*** denote significance at 10, 5, 1 percent, respectively.

Negative changes denote depreciation.

Step 2—What is special about inflation in Indonesia?

9. Inflation in Indonesia is more persistent and more sensitive to country-specific shocks and political risks than in other Asian countries. The estimates of the slope dummies suggest that inflation in Indonesia is more sensitive to past inflation, the output gap, exchange rate fluctuations and political risks than in the other countries of the sample. The estimated magnitude of these effects appears significant. For example, using a modified version of our basic model (Table 4, model 3), an additional one percentage point in past inflation is associated with an increase in the inflation rate of about 0.65 percent in Indonesia, 0.2 percentage points higher than the sample average. This suggests that the historically high inflation rates in Indonesia have generated strong inflation inertia. At the same time, an additional one percentage point change in either the output gap or in the depreciation of the currency in Indonesia is associated with an increase in the inflation rate of about 0.36 percent, 0.3 percentage points higher than in other countries (Table 2, models 4 and 6). Finally, a one percentage point increase in the overall political risk index increases inflation in Indonesia by 0.6 percentage points compared to an increase of 0.4 percentage points for the average of the sample (Table 2, model 7). The results are robust to structural changes due to the crisis as the model estimated for the whole sample is broadly similar to the model estimated for the post crisis period (Box 1).

Table 4.Determinants of Cross-Country Inflation : The Case of Indonesia
Model (3)Model (4)Model (5)Model (6)Model (7)
Previous inflation0.45***0.67***0.60***0.53***0.63***
Output gap0.13*0.08*0.12**0.09**0.12**
M2 Growth0.03***0.03***0.02***0.01***0.03***
NEER growth 1/–0.2***–0.19**–0.18**–0.06***–0.19**
Political risk0.12***0.1*0.11***0.9***0.07**
Previous inflation (Ind. dummy)0.20**
Output gap (Idn dummy)0.28***
M2 Growth (Idn dummy)0.07
NEER growth (Idn dummy)–0.3***
Political risk (Idn dummy)0.10***
Chi-square0.000.000.000.000.00
Obs.348348348348348
Notes: Difference GMM estimations using quarterly data 1990Q4-2005Q4.Independent variable: End-period inflation rateCountry coverage: Indonesia, Korea, Malaysia, Philippines, Singapore, Thailand.*,**,*** denote significance at 10, 5, 1 percent, respectively.

Negative changes denote depreciation.

Notes: Difference GMM estimations using quarterly data 1990Q4-2005Q4.Independent variable: End-period inflation rateCountry coverage: Indonesia, Korea, Malaysia, Philippines, Singapore, Thailand.*,**,*** denote significance at 10, 5, 1 percent, respectively.

Negative changes denote depreciation.

10. Monetary policy has similar effects on inflation in Indonesia and other Asian countries. The coefficient of the Indonesia-related slope dummy for money growth is not significant (Table 4, model 5). This suggests that an additional one percentage point increase in the growth of M2 is associated with an increase in inflation of a similar magnitude as in neighboring countries.

Step 3—Explaining the Indonesian inflation differential

11. The Indonesian inflation differential vis-à-vis other neighboring countries is largely explained by inflation inertia and political risks, in addition to rapid monetary expansion and currency depreciation. In general terms, the inflation differential can be explained by a combination of two elements: how the different factors influencing inflation have evolved over time in each country and how differently they have affected inflation in Indonesia relative to other countries (in our model, the coefficients inflation inertia and political risks explain on average about 75 percent of the Indonesian inflation differential vis-à-vis selected Asian countries (Table 5).8 Monetary policy and exchange rate depreciation are also seen to contribute to the inflation differential, although to a smaller extent (about 25 percent), with the additional money-generated inflation coming from the expansionary monetary policy in Indonesia compared with other countries. In contrast, the output gap has played little role in determining the Indonesian inflation differential, as its stronger effect on inflation (significant and large slope dummy coefficient) has been largely offset by the lower average output gap in Indonesia compared with the other countries in the sample.

Table 5.Sources of the Average Indonesia Inflation Differential
Mean variablesAverage additional inflation effect 2/
Dummies CoefficientsSample 1/IndonesiaPercentage difference
Previous inflation0.25.512.7132%5.242.6%
Output gap0.28−0.2−0.1−19%0.00%
M2 Growth 3/0.0712.920.458%1.19.0%
NEER growth−0.3−2.5−6.9177%2.016.0%
Political risk0.1031.845.844%4.032.4%
Total12.3100.0%
1/Country coverage: Indonesia, Korea, Malaysia, Philippines, Singapore, Thailand, and Indonesia.
2/Effects calcualted using different models and average values.
3/Coefficient not significant.
1/Country coverage: Indonesia, Korea, Malaysia, Philippines, Singapore, Thailand, and Indonesia.
2/Effects calcualted using different models and average values.
3/Coefficient not significant.

E. Conclusions and Policy Issues

12. Over the past one and a half decades, Indonesia has consistently recorded higher inflation than its neighbors. The analysis suggests that the causes of the Indonesia inflation differential vis-à-vis other countries in the region include various structural factors, such as strong inflation inertia and political instability, combined with expansionary monetary policy and currency depreciation. On the other hand, structural factors such as the degree of economic openness, the public debt burden, and the level of price regulation play no role in explaining inflation across Asian countries.

13. In light of the strong inflation persistence, reducing inflation requires maintaining a consistent monetary framework and asserting the credibility of the policy framework. The strong persistence of inflation in Indonesia could imply that the convergence process to lower regional inflation rates might be slow and costly in terms of economic growth. To reduce this cost and accelerate the convergence process, the central bank has an important role to play, building its credibility and thus affecting the formation of inflation expectations. In this respect, Bank Indonesia’s recently adopted inflation targeting framework could play a helpful role.

Box 1.An Empirical Analysis of the Indonesia Inflation Differential Compared to Other Asian Countries

A cross-country econometric model of inflation determinants is estimated in order to examine various hypotheses about the sources of the Indonesia inflation differential. The basic model regresses current inflation on a set of possible inflation determinants and takes the following form:

where, i denotes the country and t time. Past inflation (πit-1) captures the degree of backward looking inflation expectations (adaptive expectations or inflation inertia), the output gap (YGAPlt-1), measured using an HP filter, summarizes demand pressures, M2 and NEER growth rates (M2git-1, NEERgit-1) encapsulate the role of monetary policy and exchange rate developments, and political risk (Polriskit-1) captures the role of institutional features. This latter variable (from ICRG) is a composite index including measures of socioeconomic conditions, government stability, internal and external conflicts, corruption, and bureaucracy quality. Difference GMM estimates are used to address possible endogeneity problems in this dynamic panel analysis.

A set of Indonesia-related slope dummies is introduced to investigate whether the role of the inflation determinants differs in Indonesia compared to the average of the selected countries. The coefficients of the slope dummies measure how much the impact of a given variable on inflation differs in Indonesia compared to the country sample.

The data consist of a quarterly panel covering the period 1990Q4-2005Q4 and including six countries: Korea, Malaysia, The Philippines, Thailand, Singapore, and Indonesia.

Results from the model are stable. The main results are presented in Section D using difference GMM estimates. They do not change substantially with alternative specifications of the model (e.g., using average inflation rates, reserve money, and price regulation indexes), and are robust to heteroskedasticity and correlated disturbances. Moreover, in-sample analysis suggests that the model is robust to structural changes due to the crisis as the model estimated for the whole sample is broadly consistent with the model estimated post crisis for inflation inertia, output gap and NEER depreciation. Finally, the model has been tested using Singapore slope dummies and results suggest that, unlike Indonesia, this country is not substantially differ from the sample average (i.e., Singapore slope dummies are not significant).

This model should not been seen as explaining inflation in any particular country, but setting a benchmark against which to examine different hypotheses about the causes of the Indonesia inflation differential. In this framework, the Indonesia inflation differential represents the difference between Indonesia and the average of the selected Asian countries.

References

    AisenA. and F.J.Veiga(2005)“Does Political Instability Lead to Higher Inflation? A Panel Data Analysis,”IMF Working Paper 05/49.

    AnglingkusumoR.(2005)“Money inflation Nexus in Indonesia: Evidence from a P-Start Analysis,”Tinbergen Institute Discussion Paper 54/4.

    • Search Google Scholar
    • Export Citation

    ArellanoM. and BondS.(1991)“Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations,”Review of Economic Studies 58.

    • Search Google Scholar
    • Export Citation

    CampilloM. and MironJ.A.(1996)Why Does Inflation Differ Across Countries?” NBER Working Paper 5540.

    CoeD.T. and McDermottJ.C.(1997)“Does the Gap Model Work in Asia?”IMF Staff Papers 44 1.

    CukiermanA.EdwardsS. and TabelliniG.(1991)“Explaining Fiscal Policy and Inflation in Developing Countries,”Journal of International Money and Finance 10.

    • Search Google Scholar
    • Export Citation

    EasterlyW. and FischerS.(2001)“Inflation and the Poor,”Journal of Money Credit and Banking 33 22.

    International Monetary Fund(2006)“Globalization and Inflation,”World Economic OutlookApril.

    KhanM.S. and SenhadjiA.S.(2000)“Threshold Effects in the Relationship Between Inflation and Growth,”IMF Working Paper 00/110.

    LounganiP. and SwagelP.(2001)“Sources of Inflation in developing Countries,”IMF Working Paper 01/198.

    MankiwN.G.ReisR. and WolfersJ.(2003)“Disagreement about Inflation Expectations,”NBER Working Paper 9796.

    RajaguruG. and SiregarR.(2002)“Sources of Variations between the Inflation Rates of Korea, Thailand and Indonesia during the Post-1996 Crisis,”CIES Discussion Paper 299.

    • Search Google Scholar
    • Export Citation

    RomerD.(1993)“Openness and Inflation: Theory and Evidence,”Quarterly Journal of Economics CVII (4).

1Prepared by Geremia Palomba (APD).
2The analysis is based on a panel of quarterly data covering the period 1990Q4–2005Q4, and including, in addition to Indonesia, five neighboring countries: Korea, Malaysia, Philippines, Singapore, and Thailand.
3Results remain broadly unchanged if the average inflation rate excludes the crisis period.
4Indonesia would be less of an outlier if the average annual inflation rate excluded the crisis period. It is worth noting that despite the similarity in the average level of political risk between Indonesia and Korea (and the lower average government instability in Indonesia), Indonesia has recently been falling behind Korea on both accounts. However, before the crisis Indonesia had lower political risk and instability, thus showing a better average rating than Korea for the period as a whole.
5The basic model is a dynamic panel in which the explanatory variables include the lagged dependent variable. To address possible endogeneity problems, difference GMM estimators are used (Arellano and Bond, 1991).
6The coefficients of the slope dummies measure how much the impact of a given variable on inflation is different in Indonesia compared to the sample of selected countries.
7The ICRG political risk variable aggregates twelve different subcomponents, including government stability and bureaucracy quality. The indices have been re-based, so that the greater the political risks (instability, lack of quality of the bureaucracy) the higher are the indices.
8These values are only indicative as they are obtained by using coefficients from different models.

Other Resources Citing This Publication