WP/ 8 /2014 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Data Source Availability 4 16
17
C t x C t x T CI xy = 1 T [C t x C t y + (1 C t x )(1 C t y )] t=1 C t x = {0, if x is in the contractionary phase at time t; 1, if x is in the expansionary phase at time t} C t y = {0, if y is in the contractionary phase at time t; 1, if y is in the expansionary phase at time t} 1 35 [ 3,12,17,12, 3] 18
19
20
21
22
23
24
25
Frequency based filter Turning point analysis Narrow Credit Narrow Credit/GDP IHSG Common Credit/GDP IHSG IHPR IHSG IHPR IHPR Cycle Credit- 89% 84% 25% 80% 38% 36% Credit/GDP- IHSG 59% 22% 17% 26% 55% 61% Credit- Credit/GDP 26
Frequency based filter Turning point analysis Broad Broad Credit Credit/GDP Credit/GDP IHSG IHSG 77% 51% 52% 72% 2% 31% Common Cycle Credit- Credit/GDP Credit- Credit/GDP Cycle Business Cycle (PDB) Average Duration (Quarter) Financial Cycle (Narrow Credit) Financial Cycle (Broad Credit) 27
Peak to peak 19 38 40 Trough to trough 17 39 35 Cycle 18 39 37 Financial Cycle / Business Cycle 2.10 2.04 28
29
30
31
Aikman, D., A Haldane and B Nelson, Curbing the Credit Cycle, as presented at the Columbia University Center on Capitalism and Society Annual Conference, New York, November (Revised March 2011), 2010 Basel Committee on Banking Supervision, Guidance for national authorities operating the countercyclical capital buffer, Bank for International Settlement, BIS, 2010. Borio, C., The Financial Cycle and Macroeconomics: What Have We Learnt?, BIS Working Papers, 395, 2012 Bry, G., C. Boschan, Cyclical Analysis of Time Series: Selected Procedure and Computer Program, National Bureau of Economic Research, Technical Paper 20, 1971. Christiano, L.J., Fitzgerald, T.J., The Band Pass Filter, International Economic Review, volume 44, issue 2, pages 435 465, 2003. Claessens, S., Kose, M.A., Terrones, M., How Do Business and Financial Cycles Interact?, IMF Working Paper, WP/11/88, 2011 Comin, D., Gertler, M., Medium Term Business Cycle, American Economic Review, Volume 96 No. 3, 2006 Drehmann, M., C. Borio and K. Tsatsaronis, Characterizing The Financial Cycle: Don t Lose Sight of The Medium Term!, BIS Working Paper, 380, 2012 English, W., Tsatsaronis, K., Zoli, E., Assessing the Predictive Power of Measures of Financial Conditions for Macroeconomic Variables, BIS Paper No. 22, 2005 Everts, M., 2006, Duration of Business Cycles, Munich Personal RePEc Archive (MPRA), No.1219, 2006 Harding, D., Pagan, A., Dissecting the Cycle: A Methodological Investigation, Journal of Econometrics, volume 49, pages 365 381, 2002. Harding, D., Pagan, A., Synchronization of Cycles, Journal of Econometrics, 132, pages 59-79, 2006. Hatzius, J., Hooper, P., Mishkin, F.S., Schoenholts, K.L., Watson, M.W., Financial Condition Indexes: A Fresh Look after the Financial Crisis, National Bureau of Economic Research, 2010 Male, R. Louise, Developing Country Business Cycle: Characterizing the Cycle and Investigating the Output Persistence Problem, 2009 Ng, T., The Predictive Content of Financial Cycle Measures for Output Fluctuations, BIS Quarterly Review, June 2011 Utari, G.A.D., Arimurti, T., Financial Cycles In The Era of Free Capital Flows, Bank Indonesia Research Report, 2014 32
APPENDIX A. Interpolation of Private Sector Foreign Debt The recording of private sector foreign debt in quarterly format first started to be done in 1992-Q2, while previously the data was yearly. Because of that, data interpolation needs to be done from 1992-Q1 up to 1999-Q1. As a proxy, the growth pattern of PMTB investment (Gross Fixed Capital Formation) is chosen, either real or nominal growth with the following results: Panel 1. Interpolation results at Panel 2. Interpolation results in level growth (yoy) Chart A.1. Interpolation using Nominal Investment (DSTA Data) Panel 1. Interpolation results at Panel 2. Interpolation results in level growth (yoy) Chart A.2. Interpolation using Nominal Investment (SOFIE Data) 33
Panel 1. Interpolation results at Panel 2. Interpolation results in level growth (yoy) Chart A.3. Interpolation using Real Investment (SOFIE Data) The interpolation which is selected is the first one, using nominal SOFIE data, because its results are closest to yearly private sector foreign debt data. B. Bry-Boschan Coding Matlab code which was developed by Rand and Tarp (2002) already accommodates the use of monthly, quarterly, and yearly data. In the original code, the Bry-Boschan algorithm can only be used for the short term cycle with the wavelength (the distance from the peak to the peak or trough to trough) and phase (the distance from the trough to peak or peak to trough) which has already been specified for monthly, quarterly, and yearly data. Specifically for the purposes of this study, modifications were only done to quarterly data, so that the code can be used for a variety of different phases and cycles. In general, the main algorithm that is used is still the same as the original code, except in the weighting which is used in the Spencer curve. Following Everts (2006), the weighting in the Spencer curve that is used is 1/35 [-3,12,17,12,-3]. Everts (2006) stated that Harding and Pagan used 15-point smoothing of the Spencer curve, which is actually more suitable for monthly data. The complete program code is as follows: 34