Recent Developments in Banking System Liquidity1
1. This paper discusses recent developments in banking system liquidity in Indonesia. At the outset, it highlights several structural characteristics that have helped shape liquidity developments in recent years, notably the microstructure underlying the distribution of liquidity and limitations in the money markets. Taking these into account, it then attempts to answer several questions. First, what has driven the decline in available liquidity since late 2011, in particular, why did deposit growth fall sharply over the past three years? Second, how have banks been affected by the decline—specifically, why has it been more acute at smaller banks? Third, what has been the effect of policy changes, particularly since mid 2013, on banking system liquidity; i.e., have they exacerbated structural liquidity problems or been effective in helping banks manage liquidity needs in a period of adjustment to tighter funding conditions? Fourth, what are the near-term prospects for banking system liquidity, in particular with a possible tightening in global financial conditions? Finally, the paper offers views on the implications of these factors on the monetary policy transmission mechanism in the current environment.
B. Characteristics of Banking System Liquidity in Indonesia
2. Several aspects of agents’ behavior characterize the liquidity situation in the Indonesian banking system, which has led to a structural liquidity problem. This problem can be best described as the concentration of excess liquidity in larger banks, which exists beyond normal liquid asset requirements and has not been absorbed through instruments used by Bank Indonesia (BI) to conduct open market operations (Figure 1). It can be traced back to BI’s liquidity support operations during the late 1990s, which resulted in a major expansion in the monetary base.2 Large portfolio inflows in 2010–11 also contributed substantially to this liquidity, as a result of partial sterilization of the inflows.3 The structural liquidity problem is symptomatic of the undeveloped money markets in Indonesia, which are thin, segmented, and generally inaccessible to the smaller banks. As a result, monetary policy transmission is hampered and development of a functional term structure of interest rates is constrained.
Figure 1.Banking System Liquid Assets and Concentration
Source: Bank Indonesia.
1/ Liquid assets include cash, placement to the central bank, interbanklending, securities, and acceptance claims. Reserve requirements (RR) include the primary RR, secondary RR on security holdings, and loan-to-depositratiolinked RR.
2/ Total sample of 108 banks.
3. The main characteristics of banking system liquidity in Indonesia are as follows:
Banks’ funding structure is generally very short term, with dependence on large deposits, and highly skewed in favor of large banks. Reflecting depositor preferences, more than 90 percent of maturities are of one month or less or at call, although they are typically rolled over. Current and savings accounts (CASA deposits) represent over half of the total funding base and term deposits4 the rest. Large corporate-related deposits tend to be a dominant source of funding5 and come from the banks’ premium customers. The top four largest banks (i.e., Buku 4 banks, or those with core capital above Rp 30 trillion) have access to almost three-fifths of low-cost and stable CASA deposits of the system, reflecting their larger branch and ATM network and name recognition, while mid-sized and small banks rely predominantly on higher-cost term deposits for funding (Figure 2).
Demand for precautionary liquidity buffers is high, reflecting banks’ self-insurance for liquidity management. Given the history and funding structure of the Indonesian banking system, deposit flight remains an element taken into account in banks’ behavior.6 Banks have tended to place sizable excess liquidity with BI in a range of short-term facilities. To manage available liquidity, some have also tended to keep their credit extension in check relative to their deposit-taking activities,7 with low loan-to-deposit (LDR) ratios at some banks precipitating BI to introduce a floor on them in 2010 to spur credit growth. The shallowness of the money market, perceptions of supervisory stigma in accessing BI’s standing facility for lending, uncertainties surrounding the availability of a fully functioning lender of last resort, lack of averaging in reserve requirements, and capacity limitations of the deposit insurance scheme also contribute to the preference by banks to hold excess liquidity.
The money market lacks depth, reinforcing the concentration of bank liquidity and diminishing its influence on broader liquidity conditions. Incentives to trade in the interbank money market are weak, given banks’ desire to hold precautionary liquidity, while access by smaller banks is limited due to concerns about counterparty risks. Given the small interbank volumes relative to system-wide activity, money market rates do not give an accurate picture of funding conditions and have a limited impact on banks’ deposit rate setting.
C. Recent Trends in Bank Liquidity Conditions
4. Banking system liquidity in Indonesia has generally been on a declining trend since late 2011, driven primarily by a slowdown in deposit growth. This trend is reflected in a slowing of growth in both term and CASA deposits, which peaked by early 2012 and declined noticeably thereafter (Figures 3 and 4). The deposit growth slowdown can be attributable to several factors: (i) constrained corporate cash generation following the downturn in the commodity cycle since 2012; (ii) more volatile capital flows since mid 2012, lowering the net foreign assets (NFA) contribution to M2; and (iii) the drawdown of deposits given the opportunity cost imposed by low bank deposit rates prior to mid 2013.
Figure 3.Loan Versus Deposit Growth
Sources: CEIC Data Co. Ltd.; and IMF staff estimates.
Figure 4.Contribution to Customer Deposit Growth
Sources: CEIC Data Co. Ltd.; and IMF staff estimates.
5. At the same time, bank credit growth picked up as the effects of the global financial crisis (GFC) waned, peaking in May 2012 at around 26 percent (y/y). Consequently, the average LDR of banks rose sharply from 72 percent at the beginning of 2010 to more than 92 percent by July 2014. The trend of higher LDRs from slowing deposit growth relative to credit growth was not unique to Indonesia; this was also seen in Malaysia, Singapore and Thailand (Figure 5), driven in part by a combination of credit-supportive policies adopted after the GFC and shifts from bank deposits into alternative financial products amidst relatively low levels of local interest rates.
Figure 5.Selected ASEAN Countries: Loan-to-Deposit Ratios
Sources: CEIC Data Co. Ltd.; and IMF staff estimates.
6. Prior to mid 2013, however, large banks in Indonesia still viewed liquidity buffers as comfortable. At the time, they continued to hold a large amount of structural liquidity, predominantly in the form of short-dated nontradable term deposits with BI. Banks’ reliance on them for liquidity management was reinforced over time by BI’s gradual reduction starting in mid 2010 of its issuance of SBIs for liquidity absorption (Figure 6), due to the tendency for these instruments to attract volatile capital flows, which complicated monetary management. In contrast, due to the constraints in their funding structure, smaller banks were more sensitive to liquidity pressures, with many offering term deposit rates above the deposit insurance corporation (LPS) guaranteed ceiling rates.
Figure 6.Bank Indonesia’s Liquidity Absorption by Instrument
Sources: Bank Indonesia; CEIC Data Co. Ltd.; Bloomberg L.P.; and IMF staff estimates.
7. From mid 2013 to mid 2014, a reassessment of liquidity conditions among large banks occurred. This change was prompted by the shock from the May 2013 Fed tapering talk combined with several related developments, which tightened bank funding conditions, triggering a period of adjustment.
Liquidity conditions tightened further as deposit growth continued to slow noticeably relative to credit growth, pushing up LDRs further (Figure 7). Initially this was due to foreign portfolio outflows, mainly in 2013:Q3, but as these subsequently normalized, deposit growth remained subdued into 2014:H1 by the broader slowdown in economic activity. Liquidity conditions in 2014 were also affected by substantial frontloading of government financing needs8 and large private external amortization payments. Bank Indonesia signaled its guidance for banks to adjust liquidity positions by increasing from September 2013 the secondary reserves requirement (on liquid assets) and lowering the upper LDR limit on which banks would incur additional reserve requirements.9
Attempts to slow private credit growth could not initially keep pace with declining deposit growth, particularly for large banks. Large banks mainly retrenched new loan approvals, particularly for working capital; outstanding loans could not be adjusted down as quickly, as customers drew on previously unutilized credit lines. Faced with more severe funding constraints, smaller banks cut credit lines forcefully and mobilized funding from equity owners. Several of them also borrowed from the interbank repo market, which was temporarily boosted by a new Mini Master Repo Agreement rolled out at end 2013. As a result, even though the LDRs of smaller banks (Buku 1 and 2) had begun to decline, the LDR of larger banks (Buku 3 and 4) generally rose further in 2014:H1 (Figure 8), indicative of the tighter funding pressures they faced.
These developments, coupled with a reassessment of depleted liquidity buffers, prompted a surge in large banks’ competition for term deposits. The onset of acute funding pressures in this period was reflected in a steep decline in net available funding (customer deposits subtracted by credit, adjusted for statutory reserve holdings), which led to a marked rise in deposit rates due to aggressive competition for term deposits (Figure 9).
Figure 7.Monetary Aggregates and Loan-to-Deposit Ratios
Sources: CEIC Data Company Ltd.; and IMF staff estimates.
Figure 8.Loan-to-Deposit Ratios by Bank Size 1/
Sources: Bloomberg L.P.; and IMF staff estimates.
1/ Based on data from publicly listed banks.
Figure 9.Funding, Deposit and Lending Rates, and Interest Margins
Source: Bank Indonesia.
1/ Net available funding is equal to third-party funding net subtracted by third-party credit and statutory reserves.
These system-wide average rates likely understate the extent of the “deposit war” among banks, as banks were widely reported to offer special rates to their premium customers that were higher than at-the-counter rates. Lending rates were also adjusted upward, but less rapidly and only with a lag, as banks’ wide net interest margins (NIMs) provided a cushion to absorb increases in deposit rates for a time, as banks were reluctant to lose market share with their premium customers.
D. Role of Policy Actions
8. Actions by BI have been aimed at smoothing adjustment in rupiah liquidity conditions. When the tapering talk pressures hit, BI rapidly unwound the term deposit facilities with banks to offset the sharp liquidity contraction that would have resulted due to its large FX sales interventions, particularly in the June-July 2013 period. As a result, the amount outstanding in such facilities was brought down to virtually zero by September 2013. Purchases of rupiah government bonds in line with BI’s dual intervention (FX and bonds) strategy also provided a secondary source of liquidity injections. The combination of these injections more than offset the contractionary influences, resulting in a buildup of cash parked by banks in BI’s overnight deposit facility (FASBI), which more than doubled between May and August 2013, and which has since remained large. This buildup also kept overnight interbank rates close to the FASBI rate, notwithstanding increases in the BI benchmark rate (Figure 10). Meanwhile, the use of swaps by BI, including via FX swap auctions, provided rupiah liquidity to segments of banks experiencing acute mismatches between FX and rupiah funding.
Figure 10.Policy and Interbank Rates and Deposit Facility
Source: Bloomberg L.P.
9. Policy actions outside of BI, however, have tended to exacerbate the tightness in bank liquidity conditions up to mid 2014. First, the government’s sizable issuance of rupiah securities added to liquidity demands on larger banks. These banks, as primary dealers, have had to commit liquidity to securities auctions at an elevated level since mid 2013, as a consequence of the government’s catch-up issuance in 2013:H2 and frontloading strategy in 2014:H1. Second, slower-than-usual execution of the government capital budget in 2014:H1 added to liquidity pressures, particularly at smaller regional banks, which tend to be more reliant on this type of funding. Third, deposit rates, particularly those at the counter, rose further following an increase in deposit insurance (LPS) guaranteed rates in May 2014, which occurred at a time of peak seasonal cash demand pressures before Ramadan.
10. Recent supervisory action to introduce deposit rate caps could create additional distortions. The measure, undertaken by the Financial Services Agency (OJK), aimed to alleviate competition among large banks for term deposits and bring down lending rates. Effective October 1, 2014, deposit rates offered on all bank deposits up to Rp 2 billion were capped at no more than the maximum LPS-guaranteed rate (at 7.75 percent). Deposit rates for large deposits (above Rp 2 billion) were capped at (i) 200 bps above the current BI rate for the largest banks (i.e., Buku 4 banks) and (ii) 225 bps above the current BI rate for medium to large banks (i.e., Buku 3 banks, with core capital between Rp 5 trillion and Rp 30 trillion). While the full effects of these measures have yet to be seen,10 it could potentially send confusing signals about BI’s monetary policy stance and encroach on BI’s monetary policy operations, further distorting policy transmission by introducing an interest rate control with a fixed spread around the BI rate. The caps also limit the ability of individual banks in the system to adjust to shocks on liquidity via interest rate flexibility on deposits, which could not only affect their ability to mobilize large deposits, but even small ones (i.e. less than Rp 2 billion) given the lower level of rate caps being applied to these. Under these conditions, banks may resort to keeping more liquidity for self-insurance purposes.
E. Near-Term Prospects and Risks
11. Even though bank liquidity conditions remain relatively tight, there has been some easing since the second half of 2014 on a combination of adjustments and seasonal factors. First, credit growth has decelerated below deposit growth in 2014:H2. Second, portfolio inflows have been sizable since late 2013, helping replenish deposits, as reflected in increases in the NFA component of M2 (Figure 7) and the level of BI’s gross international reserves. Third, government budget execution improved in 2014:H2, while seasonal deposit outflows also eased, following the surge in profit repatriation abroad, external amortization payments, and imports in 2014:Q2. Large banks have lowered their special deposit rates to premium customers. However, system-wide LDR remained relatively elevated, registering at around 88 percent as of November 2014.
12. The banking system remains vulnerable to deterioration in external financing conditions, or at the extreme, a sudden reversal of portfolio flows. This could occur in the event of unexpected sharp tightening in global financial conditions in 2015. Buoyant conditions in 2014 had facilitated record sovereign bond issuance and portfolio inflows into rupiah government bonds, with foreign bidders taking up about a third of new rupiah treasury issuances. Vulnerability to a reversal is now heightened as large foreign holdings have risen further to record highs (Figure 11).11 Domestically, the gross fiscal financing requirement in 2015 is expected to remain large relative to 2014. If the government were to face challenges borrowing abroad, it would have to rely more on the domestic market, potentially putting a further squeeze on domestic liquidity and driving up private borrowing costs.
Figure 11.Foreign Ownership Share of Rupiah Government Bonds
Sources: Bloomberg LP.; CEIC Data Co., Ltd; and IMF staff estimates.
F. Implications for the Monetary Transmission Mechanism in the Current Environment
13. Monetary policy operations tended to have a more limited impact on the banking system’s marginal cost of funds, which appears driven more by rates on premium deposits. As noted, the rates paid by banks on these deposits tend to be less reflective of money market conditions, due to banks’ heightened attention to their LDR. However, changes in the BI policy rate and LPS guarantee rate affect banks’ average cost of funds, as at-the-counter deposit rates are benchmarked off these two policy rates (even before the recent OJK action to cap deposit rates).
14. Lending rates appear relatively insensitive to the increase in the marginal cost of funds, and tend to be based more on the average cost of funds of the largest banks, which set the market reference for lending rates and are the dominant credit providers in the system. Lending rate adjustments have been sluggish, as these banks preferred to compress their large NIMs as system-wide deposit rates increased, rather than raise lending rates sharply and unnecessarily risk the loss of their premium customers or a rise in nonperforming loans.
15. The main impact from an interest margin squeeze tends to occur through loan portfolios of the small and medium sized banks, reflected in either aggressively cutting credit lines to noncore customers or attempting to shift the loan book towards high-yielding credits to preserve their NIMs.
AffandiY. and S.Peiris2012 “Interest Rate Transmission, Lending Conditions, and Monetary Policy” in Philippines: Selected IssuesIMF Country Report No. 12/50 (Washington: International Monetary Fund). Available via the Internet: www.imf.org/external/pubs/ft/scr/2012/cr1250.pdf
Bank Indonesia2002 “Transmission Mechanisms of Monetary Policy in Indonesia” eds. by P.Warjiyo and J.AgungDirectorate of Economic Research and Monetary Policy (Jakarta).
EnochC.B.BaldwinO.Frecaut and A.Kovanen2001 “Indonesia: Anatomy of a Banking Crisis—Two Years of Living Dangerously 1997–99” IMF Working Paper No. 01/52 (Washington: International Monetary Fund). Available via the Internet: www.imf.org/external/pubs/ft/wp/2001/wp0152.pdf
HeenanG.2011 “Monetary Operations, Liquidity Management and Money Market Development” in Indonesia: Selected IssuesIMF Country Report No. 11/310. (Washington: International Monetary Fund). Available via the Internet: www.imf.org/external/pubs/ft/scr/2011/cr11310.pdf
International Monetary Fund2010Indonesia: Financial System Stability AssessmentIMF Country Report No. 10/288 (Washington). Available via the Internet: www.imf.org/external/pubs/ft/scr/2010/cr10288.pdf
International Monetary Fund2013Regional Economic Outlook: Asia and PacificApril 2013 (Washington). Available via the Internet: http://www.imf.org/external/pubs/ft/reo/2013/APD/eng/areo0413.htm
Moody’s Investors Service2014 “Indonesian Banks: Rising Cost of Deposits Will Introduce Credit Differentiation” in Banking: Special Comment. Available via the Internet: https://www.moodys.com/research/Moodys-Indonesian-banks-rising-cost-of-deposits-will-introduce-credit—PR_303362
Prepared by Seng Guan Toh.
Following the onset of the Asian Financial Crisis, banks representing more than half of total assets of the system experienced deposit runs to varying degrees. In response, BI injected a large amount of liquidity into the banking system and in return received nontradable government bonds carrying little interest. Since then, BI has relied on use of its own securities (SBIs) to mop up the resulting excess liquidity (IMF, 2010).
The challenge to monetary authorities of sterilizing liquidity from surging inflows to the region was not confined to Indonesia. For example, during 2010–11, the Philippines’s central bank nearly exhausted its holdings of government securities for use as collateral in reverse repo transactions. Given constraints on issuance of its own securities, Bangko Sentral ng Pilipinas had to rely increasingly on other instruments such as nontradable special deposit accounts. Furthermore, the emergence of excess reserves in the money market in the Philippines contributed to market interest rates falling below the policy rate, affecting monetary policy transmission (Affandi and Peiris, 2012).
At end 2013, almost half of term deposits had a maturity of one month or less.
The heavy concentration of the deposit base was such that the Deposit Insurance Corporation (LPS) noted that at end 2013, a mere 0.3 percent of depositors owned 61 percent of deposits by value. This is also why the LPS did not guarantee approximately 55 percent of the nominal values of total deposits (despite having an ample ceiling equivalent to almost US$200,000 per bank depositor).
This also helps explain why the maximum deposit guarantee rates set by LPS have some influence on banks’ at-the-counter deposit rates.
Small banks’ LDRs are typically capped by funding constraints.
In light of market uncertainties, the government frontloaded its issuance of rupiah securities in 2014:H1. As a result, it accumulated surplus financing estimated in excess of 0.6 percent of GDP in 2014:H1 compared to the same period in the previous year, which resulted in a large buildup in government deposits placed in BI.
Bank Indonesia raised the secondary reserve requirement (RR) in September 2013 (fulfilled by banks’ holding of treasury and BI securities) from 2.5 percent to 4.0 percent, to be phased in by December 2013, and also tightened the LDR-linked RR by applying it to banks with an LDR in excess of 92 percent (previously 100 percent) and with a capital adequacy ratio under 14 percent.
A similar measure to cap deposit rates was attempted by BI in 2009 through suasion, with mixed results, notably with lending rates remaining relatively high while tending to bolster banks’ net interest margins.
Notably, the foreign share of rupiah government bond holdings rose from about 30 percent in October 2013 to 40 percent by January 2015. However, for Indonesia, the predominant share of institutional investor holdings (relative to retail investors) historically provides a partial stabilizing factor during outflow episodes due to global volatility.