Philippines: Selected Issues
Author: International Monetary Fund
Publication Date:
27 Mar 2007
eISBN:
9781451831405
Language:
English
Keywords:
ISCR; CR; bank; spending; Philippines; lending; bank assets; bank w hich; capitalized bank; bank balance sheet quality; bank capital; Loans; Credit; Commercial banks
Download PDF (1.3 MB)

This Selected Issues paper highlights the Philippine growth performance led by the services sector. Average GDP growth is higher in the post-Asian crisis period in the Philippines, while the majority of the Philippines’s regional peers have experienced substantially lower growth in the post-Asian crisis period compared with the pre-crisis period. Trade and transport, storage, and communications services have been growth drivers while private and financial services have started to add new momentum. Various transfer programs are identified that would be much better targeted than across-the-board energy tax cuts.

Abstract

This Selected Issues paper highlights the Philippine growth performance led by the services sector. Average GDP growth is higher in the post-Asian crisis period in the Philippines, while the majority of the Philippines’s regional peers have experienced substantially lower growth in the post-Asian crisis period compared with the pre-crisis period. Trade and transport, storage, and communications services have been growth drivers while private and financial services have started to add new momentum. Various transfer programs are identified that would be much better targeted than across-the-board energy tax cuts.

III. Credit Growth and Bank Balance Sheets in the Philippines1

A. Introduction

1. Bank lending growth has been sluggish in the Philippines for a number of years. Following the Asian crisis, a mix of weak credit demand and reluctance by banks to extend credit has been argued to have contributed to the significant decline in lending relative to GDP observed in the Philippines. In recent years, however, a number of steps have been taken to strengthen the banking sector. Yet, credit growth has not recovered.

2. This chapter attempts to shed more light on the factors hampering credit growth. Demand side factors are more fully modeled than in previous studies on the Philippines (Lamberte, 1999), allowing the relative roles that demand and supply side factors are playing in credit growth to be more clearly identified. In addition, the longer sample period used permits testing for the stability of these relationships over time. The results suggest that balance sheet variables account for a large part of the explained credit growth variation and increasingly so since late 2004. While demand factors are important as well, this finding suggests that decisive resolution of the distressed asset problem will be necessary if the banking system is to contribute fully to an investment revival.

B. Background

3. The banking sector problems in the Philippines originated in the run-up to the Asian crisis. Until 1997, bank loans expanded significantly, partly driven by a real estate boom. Following the crisis, as asset prices collapsed and growth slowed sharply, the quality of bank assets deteriorated and bank credit declined substantially. Commercial bank loans in percent of GDP contracted from a peak of 58 percent in 1997 to 28 percent in 2006 (Figure 1).2

Figure 1:
Figure 1:

Comercial Bank loans

(In percent of GDP)

Citation: IMF Staff Country Reports 2007, 131; 10.5089/9781451831405.002.A003

Sources: CEIC Data Company Ltd, and IMF staff calculations.

4. The decline in credit could be traced to both demand and supply factors. Guinigundo (2005) argues that low bank lending reflects both the deterioration in asset quality as well as the slowdown in economic activity. Demand for credit declined as investment was sharply reduced because of uncertainty and some over-capacity achieved in pre-crisis years.3 At the same time, the erosion of borrowers’ creditworthiness, combined with weakened bank balance sheets, resulted in a more cautious lending stance by banks that made them highly selective in their lending behavior even at higher interest rates.4

5. Progress in strengthening bank balance sheets has been relatively slow. The banking system in the Philippines did not exhibit the same degree of weakness as some of its neighbors in the direct aftermath of the Asian crisis (Figure 2). However, the large amount of nonperforming assets (NPAs) that accumulated on the banks’ books during the crisis has not until recently been aggressively addressed, a result of the country’s fiscal position constraining the possible use of public money, and of weak legal protection for bank supervisors. Over the past two years, the Bangko Sentral ng Pilipinas (BSP) has carried out a number of reforms aimed at strengthening the supervisory framework, cleaning up the balance sheet of banks, and improving financial transparency. As a result, the amount of NPAs on bank balance sheets has declined and banks have started to raise new capital (Figure 3).

Figure 2:
Figure 2:

Non performing Loans

(Ratio to total loans)

Citation: IMF Staff Country Reports 2007, 131; 10.5089/9781451831405.002.A003

1/ NPL plus “real and other properties owned or acquired” (ROPOA) over TL plus ROPOA.2/ Average 1998-1999 only contains 1999 dataSources: CEIC Data Company Ltd; and IMF staff calculations.
Figure 3:
Figure 3:

Banking Distress Indicators

(In percent) 1/

Citation: IMF Staff Country Reports 2007, 131; 10.5089/9781451831405.002.A003

1/ 2006 Data refers to March 2006

6. The authorities have implemented a framework for the disposal of bad assets. Part of the NPA problem has been resolved under the Special Purpose Vehicle (SPV) Act adopted in 2002. While in other Asian economies, asset management companies have been funded by the government, the process in the Philippines has been led by the private sector. The SPV Act offers tax incentives – such as exemption from documentary stamp tax and capital gains tax – and regulatory relief for banks selling their distressed assets. Sales of P 100 billion of NPAs (one fifth of the stock) have been concluded. The SPV framework expired in April 2005, but was extended in March 2006 for another two years.

7. Financial transparency has also improved. The extent of the deterioration in bank asset quality following the Asian crisis has not been always transparently reported. Guinigundo (2005) observes that asset valuations were often unreliable, due to poor appraisal standards, leading to potentially overstated financial statements. Starting from end-2005, the financial statements of banks have to be prepared in line with International Financial Reporting Standards (IFRS), new accounting standards with stricter valuation requirements. New rules and standards have also been issued to clarify the role of external and internal auditors (in February 2005 and November 2005, respectively).

8. Furthermore, the authorities have taken steps to strengthen the regulatory framework and tighten enforcement. Prudential regulations governing lending to related interests (DOSRI) were expanded in March 2004 to include subsidiaries and affiliates. Exposure limits have also been introduced, as well as stiffer penalties for noncompliance. A new prompt corrective action (PCA) framework was adopted in February 2006, allowing banking supervisors to intervene before a significant decline in capital occurs and introducing specific sanctions if banks do not comply with their agreed capital restoration plan.

9. Nonetheless, weak balance sheets seem still to be constraining bank lending and investment. Lending to households is growing rapidly, albeit from a small base, supported by a pick-up in residential real estate loans and automobile and credit card financing. Corporate loans have, by contrast, remained sluggish. The stock of distressed assets in the banking sector – although declining – remains high, and tighter regulations and supervision may have amplified its effect on bank lending.5 A weak capital base can lead banks to avoid risky assets (loans) in favor of safer investments (government securities). Such portfolio shifts on the part of banks can restrict the supply of credit to the economy. On both accounts, preliminary evidence suggests that balance sheet variables seem still to influence the behavior of banks. Weaker banks – as measured by Fitch ratings – have tended to contract their loan book in favor of higher securities holdings (Figure 4 and Figure 5).

Figure 4:
Figure 4:

Securities Holdings by Bank Ratings

(in percent of assets)

Citation: IMF Staff Country Reports 2007, 131; 10.5089/9781451831405.002.A003

Ratings are from Moody’s and Fitch. C denotes an adequate bank with one or more troublesome aspects; D denotes a bank w hich has w eaknesses of internal and/or external origin; and E denotes a bank with very serious problems.
Figure 5:
Figure 5:

Loan Developments By Bank Ratings 1/

(YoY increase, in percent)

Citation: IMF Staff Country Reports 2007, 131; 10.5089/9781451831405.002.A003

1/ Ratings are from Moody’s and Fitch. C denotes an adequate bank with one or more troublesome aspects; D denotes a bank which has weaknesses of internal and/or external origin; and E denotes a bank with very serious problems.

C. Data and Methodology

10. The empirical analysis presented in this chapter is based on bank level data. Quarterly balance sheet data have been compiled for the largest 23 commercial banks continuously active over the period 1998 Q1 – 2006 Q2 from data published by the BSP. Commercial banks are the main provider of corporate loans in the Philippines and this sample covers 80 percent of total assets of the financial system as of end-June 2006. Since foreign banks may have the implicit backing of their parent company, the relationship between balance sheet information and lending may not be the same as for local banks and for this reason foreign banks were not included in the dataset.6

11. The framework used focuses on balance sheet variables in explaining bank lending. This chapter draws upon the theoretical contribution of Bernanke and Gertler (1987, 1989) on the relationship between bank capital and lending. This framework was used by a number of researchers to study the significant credit slowdown in the U.S. before and during the 1990-91 recession, as it was argued that demand factors alone could not explain its magnitude. The resulting empirical body of work has singled out declines in bank capital, tightened bank regulatory standards, and heightened market scrutiny of bank capital as the major factors behind the curtailment of bank lending (Sharpe, 1995). The framework has also been used to study credit growth in Japan (e.g., Woo, 2003).

12. The regression analysis explores the role balance sheet problems have played in the banks’ decisions to slow credit growth and invest more in securities. Results from three specifications are reported. The first specification reports the results from an OLS regression in the pooled sample. The second includes a complete set of time effects. The third regression adds a complete set of bank fixed effects. More specifically, panel regressions of the following form were estimated:

Ci⁢t=αi+βt+γBi⁢t−1+ɛi⁢t⁢(1)

where Cit denotes a measure of the policy choice for bank i in period t (growth in bank lending or shift in the portfolio allocation between loans and less risky assets), Bit-1 denotes a measure of bank balance sheet quality lagged one quarter, αi denotes a bank fixed effect, and βt denotes a time effect. Balance sheet strength was measured in three alternative ways: capital-to-asset ratio, nonperforming loans as percent of total loans, and nonperforming loans and foreclosed assets as percent of total loans.7 Bank lending is measured by growth in total loans over the previous quarter in percentage terms and shifts in portfolio allocation by the change in the share of loans in loans plus securities.

13. Two approaches are taken to assess demand side effects. In addition to the explicit inclusion of balance sheet variables in the estimation, an attempt is made to assess the relative role of demand factors in credit growth. In the first stage, the set of time dummies were interpreted as proxies for general credit demand conditions affecting all banks irrespective of their balance sheet strength. These dummies could, however, also capture economy-wide factors other than demand conditions. Hence, following Woo (2003), the time dummies were replaced in the second stage by an explicit set of lagged demand side factors Dt-1 (annualized nominal GDP growth or nominal fixed investment growth), and nominal lending rates, LRt-1:

Ci⁢t=αi+γBit−1+δ⁢Dt−1+ηL⁢Rt−1+ɛi⁢t⁢(2)

D. Results

14. The results suggest that balance sheet strength significantly influences credit growth. Table 1 reports the results for regressions for (quarter-on-quarter) loan growth and shows that the estimates for the coefficients on the capital ratio variable are consistently positive and statistically significant. Confirming the preliminary evidence shown in Figure 5, these results suggest that banks with weaker capital positions tended to expand their loan portfolio more slowly than stronger capitalized banks. Similarly, banks which carried less problem assets on their books were able to lend more in the ensuing period.8

15. Balance sheet variables seem to have gained increased traction recently. Overall, the measures of fit of the regressions are somewhat lower than in previous studies using shorter sample periods. One explanation could be that there are structural breaks in the data that render the pooled results less robust. To explore this possibility, model (3) in Table 1 was estimated over a 12 quarter rolling window. Figure 6 charts the estimated coefficient on the capital ratio. The results suggest that the role played by balance sheet variables, while stable until late 2004, has increased recently. At the same time, the standard error of the estimation has widened.9 One reason could be that banks may have responded differently to, the recent tighter prudential regulations and strengthened enforcement, with weaker banks increasingly curtailing their lending activities.

Table 1.

Philippines: Regressions for Quartely Bank Loan Growth, 1Q 1998 - 2Q 2006

article image
Source: IMF staff estimates.Notes: The dependent variable is quarter on quarter loan growth in percentage points. Standard errors are in parentheses. “*” indicates significance at the 10 percent level, “**” and “***” at the 5 and 1 percent level, respectively.
Figure 6.
Figure 6.

Coefficient estimates (capital-to-assets) using a 12-quarter moving data window

Citation: IMF Staff Country Reports 2007, 131; 10.5089/9781451831405.002.A003

Source: IMF staff calculation.

16. Demand factors appear also to have played a role in explaining credit growth. The estimates for the coefficients on the time dummies are statistically significant and their inclusion enhances the explanatory power of the model. The estimated proportion of explained loan growth attributable to these dummy variables has, however, decreased over time (Figure 7). These results would suggest that economy-wide factors such as weak credit demand now account for about a third of the explained variation in credit growth.10 Modeling more explicitly demand variables in equation (2) confirms these results, yielding statistically significant coefficient estimates with the expected signs. The level of lending rates is negatively associated with increases in credit growth, while nominal GDP and investment growth are positively linked with bank lending.

Figure 7.
Figure 7.

Proportion of explained loan growth accounted for by demand factors

Citation: IMF Staff Country Reports 2007, 131; 10.5089/9781451831405.002.A003

Source: IMF staff calculations.

17. Finally, the regression results also confirm that weaker banks have shifted their portfolio towards safer assets. Table 2 reports the results for regressions for (quarter-on-quarter) changes in banks’ portfolio as measured by the share of total loans in total loans plus securities. The estimates of the coefficient on the capital ratio variable are consistently positive and statistically significant, suggesting that weaker capitalized banks have decreased the share of their portfolio invested in loans. The results also show that banks with more problem assets, ROPOAs and NPLs, on their books shifted their portfolio away from loans and towards securities. These results holds even after controlling for time and fixed effects.

Table 2.

Philippines: Regression for changes in Bank Portfolio Choice 1Q 1998-2Q 2006

(change in loans/(loans+securities) as dependent variable)

article image
Source: IMF staff estimates.Notes: The dependent variable is the ratio of loans to loans and securities. Standard errors are in parentheses. “*” indicates significance at the 10 percent level, “**” and “***” at the 5 and 1 percent level, respectively.

References

  • Baldwin, Barbara, 2006, “Enhancing Financial Stability through Strengthening Consolidated Supervision: The Case of the Philippines,” unpublished mimeo (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Bernanke, Ben S. and Mark Gertler, 1987, “Banking and Macroeconomic Equilibrium,” in New Approaches to Monetary Economics: Proceedings of the Second International Symposium in Economics Theory and Econometrics, edited by William A. Barnett and Kenneth J. Singleton, pp. 89 -111 (Cambridge, Cambridge University Press).

    • Search Google Scholar
    • Export Citation

  • Bernanke, Ben S. and Mark Gertler, 1989, “Agency Costs, Net Worth and Business Fluctuations,” American Economic Review, Vol 79, pp. 14 -31.

    • Search Google Scholar
    • Export Citation

  • Guinigundo, Diwa, 2005, “The Philippine Financial System: Issues and Challenges”, BIS Papers, No 28, pp. 295 -311.

  • Lamberte, Mario B., 1999, “A Second Look at Credit Crunch: The Philippines Case,” Philippine Institute for Development Studies Discussion Paper Series No. 99-23 (Manila, Philippines: Philippine Institute for Development Studies).

    • Search Google Scholar
    • Export Citation

  • Sharpe, Steven A., 1995, “Bank Capitalization, Regulations, and the Credit Crunch: A Critical Review of the Research Findings,” Board of Governors of the Federal Reserve System Finance and Economics Discussion Series 95/20

    • Search Google Scholar
    • Export Citation

  • Woo, David, 2003, “In Search of “Capital Crunch:” Supply Factors Behind the Credit Slowdown in Japan,” Journal of Money, Credit and Banking, Vol. 35(6), pp. 1019 -1037.

    • Search Google Scholar
    • Export Citation
1

Prepared by Richard Podpiera and Raju Singh.

2

While credit growth accelerated significantly in 2006, it remained slightly lower than nominal GDP growth.

3

Lower demand has been also likely driven by regional developments, as some multinational corporations have relocated from the Philippines to other regional hubs with lower costs. Furthermore, highly rated borrowers have been able to secure capital market financing.

4

Other supply side factors have likely included the impact of ongoing regulatory reforms—as expectations of stricter regulations on non-performing assets likely made banks more cautious in lending—and a lack of centralized credit information bureau that would help banks limit risk in credit decisions.

5

The distressed assets ratio stood at 20 percent at end-November 2006 according to staff calculations, which exclude interbank loans. The authorities definition, which includes interbank loans, points to a distressed assets ratio of 17 percent at end-November.

6

In addition, all data were carefully examined for unexplained outliers.

7

Foreclosed assets (ROPOAs), mainly real estate, form a significant part of bank assets in the Philippines. These assets are illiquid and their valuation is problematic. They are therefore an important source of potential weakness for bank balance sheets.

8

Including both the capital ratio and problem assets in the equation does not change the results qualitatively.

9

These results were robust even once loan growth was adjusted for the sale of NPAs to asset management companies. Failing to make these adjustments may lead to the erroneous conclusion that the lending of a troubled bank has been curtailed, when in fact the decline in its outstanding loans is due to loan sales.

10

This attribution is based on comparing the fit of model (3) in Table 1 with the estimation of the same model without the set of time effects. The difference in explained variability is interpreted as the lower bound for the explanatory contribution of demand factors.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
Philippines: Selected Issues
Author: International Monetary Fund
Volume/Issue:
Volume 2007: Issue 131
Publisher:
International Monetary Fund
ISBN:
9781451831405
ISSN:
1934-7685
Pages:
37
DOI:
https://doi.org/10.5089/9781451831405.002

Comercial Bank loans

(In percent of GDP)

Non performing Loans

(Ratio to total loans)

Banking Distress Indicators

(In percent) 1/

Securities Holdings by Bank Ratings

(in percent of assets)

Loan Developments By Bank Ratings 1/

(YoY increase, in percent)

Coefficient estimates (capital-to-assets) using a 12-quarter moving data window

Proportion of explained loan growth accounted for by demand factors