To examine whether survey measures of inflation expectations have an effect on the dynamics of actual inflation, we use VAR (Vector Auto. Regression) analysis. SPPs have information about next quarter inflation expectations (all releases), next quarter inflation expectations (in second quarter releases), next quarter inflation expectations (in second quarter releases). first) and the inflation expectation of the next quarter (in the fourth quarter releases). In this section we will examine the heterogeneity of inflationary expectations measurements by different economic agents.
Based on the same argument, we also expect retailers' inflation expectations to affect consumers' inflation expectations. Survei Persepsi Pasar (SPP) reports the mode of the next quarter and year-end inflation expectations series. Correlation of inflation expectations from Business Survey (SKDU) and Professional Forecasters Survey (SPP and CF).
From the analysis in this section we can conclude that for the forecast horizon of a first quarter, the level of heterogeneity of inflation expectations of firms (SKDU), consumers (SK), retailers (SPE) and professional forecasters (SPP and CF ) is more. low.
Relationship with Past, Current and Future Inflation
However, it should be noted that consumers still take relatively high account of the current level of inflation when forming inflation expectations. Compared to consumers, retailers are more future-oriented and take less into account the current level of inflation when forming inflation expectations. Similar to what we find when using the entire sample period of both surveys, retailers, compared to consumers, show more forward-looking behavior and give less weight to the current level of inflation when forming inflation expectations for the next three months.
The highest correlation is found between corporate inflation expectations and inflation for the next quarter, which is the target forecast horizon. From this we can conclude that compared to consumers and retailers, companies have a more ≈rational∆ inflation expectation. Since SKDU balance scores need to be compatible with next quarter's price movements, we add 3 previous quarterly balance scores to the current balance score to get a measure of the year-over-year inflation expectation.
From Figure 10, we can see that the movement in corporate inflation expectations for the next quarter can mimic the movement in actual inflation relatively well. This confirms the fact that we can use CF data as a very good indication of quarter-on-year inflation expectations at the end of the period. The Consensus Forecast inflation forecast shows similar performance to the inflation expectation series produced by the SSMX model.
From the table, we can conclude that of all surveys reporting inflation expectations (not balance sheet scores), Consensus Forecast has the best performance in terms of accuracy in predicting the level of inflation over the target forecast horizon. Accuracy of CF measures of inflation expectations (compared to each quarter's average yoy inflation). From the analysis in this section, we can conclude that survey measures of inflation expectations appear to be forward-looking, but only over a fairly short horizon (usually shorter than the intended forecast horizon).
For the survey data that measures the expectation of inflation two quarters ahead, the expectation of inflation from the Consensus Forecast shows a higher correlation with actual inflation than the modified 6-month SK and the 6-month SPE. Among all surveys that report rates of inflation expectations (not balance points), the consensus forecast performs best in terms of accuracy in forecasting the level of inflation over the projected forecast horizon.
The Effect on Actual Inflation
It is important to note that we found a different response to shocks from the 1-quarter Consensus Forecast and changed the SKDU balance score to the dynamics of actual inflations. From previous sections, we found that both measures have a similar correlation with past, current and future inflation levels. From figure 15 we can see that shocks to changes in the Consensus Forecast's inflation expectation have an effect on the changes in the inflation level in horizons 1 and 2.
While shocks to changes in business expectations affect changes in inflation only at horizon 2. The graph also shows that shocks to changes in Consensus Forecast inflation expectations (of different forecast periods) affect changes in the level of inflation starting from horizon 1. The magnitude and length of the effects decrease as the forecast period increases.
As shown in Figure 16, compared to the other survey measures of inflation expectations, shocks to the modified SK and SPE inflation expectations have relatively less effect on the dynamics of actual inflation. Granger causality tests in Table 9 show that almost all the inflation expectations indicators granger cause inflation, except for SPE6m (SurveiPenjualanEcerean √ 6 months). On the other hand, inflation does not cause all indicators, except for SPE3m (SurveiPenjualanEceran- 3 months) and CF_2Q( Consensus Forecast √ 2 Quarter).√Based on Impulse Response Function and granger causality test results, we can conclude that almost all inflation expectations indicators have an effect on the dynamics of actual inflation.
The result of the inflation variance breakdown for each VAR model is shown in Table 10. From this table, we can see that according to the indicator used, shocks to inflation expectations represent up to 75.76%. SK3m_yoy, SPE3m_yoy, SK6m_yoy, SPE6m_yoy vs Inflasi_yoy All variables are in differences.
Directional Information in Predicting Current and Future Inflation
From Table 11, we can see that the addition of CF, SKDU_years, and SPE3 billion_years measures of inflation expectations add significant predictive power to the simple model of next-quarter inflation. SK6bln_years produces similar additional predictive power as SKDU_years, but this series actually aims to measure inflation expectations for the next quarter. For the inflation model of the next 2 quarters, the highest additional predictive power is provided by the inclusion of the CF_2Q series.
SSMX-generated inflation expectations series show significant additional predictive power if added to the inflation model at time t. For modeling purposes, CF_1Q series can be used as an alternative to inflation expectations series in a macroeconometric model. Compared to the one used by SSMX, this series can be considered a more ≈rational∆ (since more accurate in predicting future inflation level), but based on an imperfect foresight process (unlike SSMX, which uses actual inflation to estimate the inflation expectation ).
Different results that we observed between CF and SSMX series in terms of their predictive ability to determine current and future inflation levels arose from the different assumptions applied to their data generation process. However, the choice to implement this series in a macroeconomic model should be based on its accuracy in forecasting inflation and other macroeconomic variables. A few criteria for information about ≈ideal∆ inflation expectations needed for policy analysis and modeling purposes are.
Each survey must have a term structure of inflation expectations dynamics from 1 to 4 quarters per head horizons. If it is not possible to have reliable information from the respondents about their end of period yy quarterly inflation expectation rate, then the survey should have information about price movement expectation that includes next 1 to 4 quarter horizons (measured by the balance sheet method). By having both expectation and perception balance sheets, we will have enough information to transform the balance sheet data into inflation expectation rates using the Carlson-Parkin approach (as discussed in Millet, 2006).
CONCLUSION
In order to predict the movement of the future inflation rate, this data can be supported by other survey measures of inflation expectation, especially from Survei Kegiatan Dunia Usaha. For modeling purposes, a 1 quarter main consensus forecast (CF) series can be used as an alternative to inflation expectations series in a macroeconometric model, as this series provides directional information about current and next quarter inflation. This article also finds that inflation expectations from observed surveys appear to be forward-looking, but only for a relatively short horizon (mostly less than the intended forecast horizon).
Although the magnitude and length vary across measures of inflation expectations, the shock to inflation expectations significantly affects the dynamics of the actual inflation rate. Among all surveys that report the inflation expectation rate (not balance score), Consensus Forecast has the best performance in terms of accuracy in forecasting the level of inflation at the intended forecast horizon. To increase their usefulness in supporting monetary policy analysis, inflation expectation information from various surveys currently conducted by Bank Indonesia should be improved by increasing the compatibility of expectation measurement from different surveys and expanding the availability of time horizons.
Alamsyah, Halim, 2008, ≈Persistent Inflation and Its Impact on Monetary Policy Choices and Responses in Indonesia∆, University of Indonesia, September 2008. Curtin, R., 2006, ≈Inflations for ventninger: empiricalke testmodeller og empiricalke testmodeller. Bank Polskis workshop om ≈The Role of Inflation Expectations in Modeling and Monetary Policy Making∆ den 9. og 10. februar 2006. Gnan, Ernest, Johann Scharlerdan Maria Antoinette Silgoner, 2009, ≈Inflations forventninger-Role and Measurement for Monetary Policy,∆ Monetary Policy & the Economy, Oesterreichische National Bank (Austrian Central Bank), number 2, side 41-67, august 2009.
Harmanta, 2009, ≈Kredibilitas Kebijakan Moneter dan Dampaknya Terhadap Persistensi Inflasi dan Strategi Disinflasi di Indonesia: Dengan Model DSGE∆, Universitas Indonesia, Oktober 2009. Millet, F.C., 2006, ≈Menemukan Metode Optimal Kuantifikasi Ekspektasi Inflasi Pada Dasar Data Survei Kualitatif∆, manuskrip, Balliol College, University of Oxford. SatishRanchhod, 2003, ≈Hubungan Antara Data Survei Ekspektasi Inflasi dan Inflasi,∆ Buletin Bank Sentral Selandia Baru, Bank Cadangan Selandia Baru, vol.
Anwar, Muslimin en TevyChawwa, 2009, ≈Analisis Ekspektasi Inflasi Indonesia Pasca-ITF∆, WP No.9, Bank Indonesia, mei 2009. Dewati, Wahyu, Ndari Suryaningsih danTevy Chawwa, 2009, ≈Meninjau Transmisi Kebijakan Moneter: Pendekatan dan Panel VAR Gegevens ∆, Bank Indonesia, LHP nr. 19 desember 2009. Kenny, 2004 , ≈Survei Ekspektasi, Rasionalitas dan Dinamika Inflasi Kawasan Euro∆, Jurnal Pengukuran dan Analisis Siklus Bisnis.
Kurniati, Yati dan Sahminan, 2008, ≈Menurunkan ekspektasi inflasi dari imbal hasil obligasi pemerintah∆, Catatan Riset No. Laksmono R, Didi et al., 2000, ≈ Suku Bunga Sebagai Indikator Ekspektasi Inflasi∆, Buletin Ekonomi Moneter dan Perbankan, Bank Indonesia, Maret 2000.
