The Selected Issues paper discusses external shocks and its effects on Chile. The economy of Chile is susceptible to global financial predicaments, external demands, and commodity rates. This paper reports on financial spillovers from 2008–12, its methodologies, and the pressures on bank funding markets. The paper also examines performance of nonfinancial sector during the 2008–09 crisis. The Executive Board sees the document as an analytical description of Chile in the global scene.

Abstract

The Selected Issues paper discusses external shocks and its effects on Chile. The economy of Chile is susceptible to global financial predicaments, external demands, and commodity rates. This paper reports on financial spillovers from 2008–12, its methodologies, and the pressures on bank funding markets. The paper also examines performance of nonfinancial sector during the 2008–09 crisis. The Executive Board sees the document as an analytical description of Chile in the global scene.

II. Financial Spillovers to Chile 1

A. Introduction

1. Chile’s economy is well integreated into the global financial system and is therefore exposed to changes in external financial conditions. Changes in global risk aversion and liquidity directly affect costs and availability of banks’ external borrowing (accounting for about 10 percent of liabilities). External pressures also transmit via arbitrage to domestic interest rates. In addition, changes in credit ratings of parent banks may affect foreign bank subsidiaries, which account for nearly forty percent of the banking sector (see Figure 1).

Figure 1.
Figure 1.

Chile: Stylized Facts, 2008-12

Citation: IMF Staff Country Reports 2012, 266; 10.5089/9781475510492.002.A002

Source: Banco Central de Chile, SBIF, Bloomberg, and own calculations.

2. This paper quantifies the spillover of global credit and liquidity risks to Chilean banks’ funding costs. For the bond market, a model of bank credit risk is used to study the effects of bank-specific factors and global credit risk factors on banks’ bond credit spread. For the interbank market, the paper updates and extends the analysis of FSR (2010) by adding proxies for global risk as explanatory variables.

3. The results suggest that global spillovers played an important role in the dynamics of funding spreads. Spillovers on average accounted for 40 percent of the bond market spread and 60 percent of the interbank market spread. Until mid-2010, banks’ bond credit spread was mostly driven by changes in banks’ fundamentals and thereafter by global risk factors. Changes in the U.S. interbank market spread accounted for most of the movements in Chile’s interbank market spread in 2008, while more recently spillovers from the euro area played a dominant role. Policy measures to increase liquidity implemented in Novemeber 2008 helped reduce the interbank market spread.

4. Nevertheless, spillover effects after 2009 have been moderate and financial intermediation does not seem to be impaired. The estimates suggest that spillovers elevated banks’ bond credit and interbank market spreads in Chile by only about 50 basis points on average between mid-2010 and early-2012. Credit growth has been very strong over the last two years in all sectors.

5. The rest of the paper is organized as follows. Next section describes the data and methodology used to decompose the funding spreads. The section C reports the results of the analysis, and section D concludes.

B. Methodology and Data

6. The effects of domestic and external variables on banks’ bond credit spread and interbank market spread are examined using least squares estimation, with standard errors adjusted for heteroscedasticity and autocorrelation. The interbank and bond markets are important bank funding sources. Interbank lending and wholesale deposits represent 20 percent of total banks’ liabilities. Bonds account for about 15 percent of banks’ funding. Changes in the wholesale funding rates should also affect retail deposit rates, possibly with some lag.

The Bond Market

7. Banks’ bond credit spread is defined as the difference between the yield on these bonds and the risk-free yield (government bonds) of similar maturity. Although this spread is affected by liquidity premia and tax issues, it mostly measures the premium for credit default risk that investors charge for lending long-term funds to banks (such as subordinated debt). The series is compiled by the Banco Central de Chile.

8. The bond credit spread has been difficult to explain using standard structural models. Modeling credit default risk is usually based on the value of a firm relative to its debt–the more the value of a company approaches the value of its debt, the more risky the company becomes, and vice versa (that is, measuring the distance to default). Since Merton (1974), the equity is viewed as a call option on a firm’s assets with maturity T; the equity price is the spot price and the maturing debt at time T per share is the strike price. Using equities as proxy for a company’s value, the credit default risk (corporate credit spread) is a function of the debt per share, volatility of equity price, and the risk-free interest rate. However, these variables explain only a fraction of credit spread variability. This is known in the literature as the credit spread puzzle – see Duffee (1998).

9. This paper uses a semi-structural model to decompose banks’ bond credit spread into a fundamental part and a global risk spillover part. The methodology is based on Otker-Robe and Podpiera (2012), who derive pricing of bank credit risk from a leveraged portfolio model. Banks are viewed as leveraged portfolios, since they borrow funds and invest them into a portfolio of risky projects. Therefore, portfolio theory could be applied to banks. In particular, there exists a risk-return efficient frontier that is the yardstick for pricing the credit risk of banks. While fundamentals anchor the long-term level of the spread, short-term volatility tends to be connected with periods of high market uncertainty and risk aversion.

10. The structural part of the model is based on the assumption that banks try to minimize risk and maximize profit. The spread is modelled as a funcion of a set of fundamentals, including banks’ net interest margin, operating expenses, return on assets, and the slope of the yield curve. Banks balance risk and return and thus optimize along a risk frontier. Following Otker-Robe and Podpiera (2012), the banks’ bond credit spread (CS) is modeled as:

CSt=c+αNIMt+βEFFtγROAtδSLOPEt+ϑt,
  • NIM denotes the net interest margin, which is the difference between the interest received from lending and paid for cost of funds; expressed as a ratio of interest bearing assets. It could be viewed as a measure of risk-taking, since loans are priced according to their risk score. In a competitive market, banks with more risky portfolios would have a higher net interest margin and would have to pay a higher interest rates on their bonds.

  • EFF is the efficiency ratio, calculated as the ratio of operating expenses to total revenues. It could be viewed as a measure of operational risk: a strong management would allocate resources well and maintain a low ratio of operating expenses to revenue. Thus, an increasing efficiency ratio signalizes higher operational risk, which would lead to a higher bond spread.

  • ROA is the return on assets, which measures profits the banking sector generates with given assets. Increasing profitability allows for strengthening capital and reserve buffers and thus increases resilience of the banking sector. As such, a higher return lowers default risks in the banking sector and the bond spread.

  • Slope is the slope of the yield curve, which is the difference between the yields on four-year and one-year inflation-indexed bonds issued by the Chilean government. Changes in the slope indicate changes in expected growth prospects of the Chilean economy and have implications for the future profitability of the banking sector. An increase in the slope signalizes improving economic conditions and lower clients’ default rates, hence better profitability of the banking sector and lower bond spread.

The data for all the above explanatory variables are from the SBIF, except for SLOPE, which is from Bloomberg.

11. The remaining part of the model includes global risks measures and local liquidity factors. In particular, the structural model is enriched by adding a global volatility index (VIX), CDS spread of European banks (both data from Bloomberg), and domestic liquidity factors (data from Banco Central de Chile):

CSt=c+αNIMt+βEFFtγROAtδSLOPEt+θVIXt+φCDSt+ϑΔMFt++ρΔPFt+ξt,

where ξt ~N(0,σ).

  • VIX is a volatility index based on S&P 500 and is often used as a proxy for global investors’ risk aversion. However, since it is measured on the U.S. stock market, it does not necessarily capture the risk premia in other markets (such as Europe) and submarkets (banking industry, in particular). Increasing risk aversion increases credit risk premia on banks’ long-term borrowings.

  • CDS of European banks captures stress in the European banking system. European banks have substantial presence in Chile, so an increase in the European banks’ CDS could have spillover effects.

  • MF and PF stand for the stock of time deposits by mutual funds and pension funds, respectively. These funds are the major providers of wholesale deposits for Chilean banks, and the amount of these deposits varies over time as funds change their portfolios.

Financial market data is at daily frequency, while banking sector’s fundamentals are interpolated to daily frequency from quarterly data. The regression analysis uses daily data from July 1, 2008 to January 6, 2012.

The Interbank Market

12. Interbank market spread reflects risk premia on short-term funding. In this paper the spread is proxied by the difference between the 90-day peso TAB rate and the overnight interest rate swap for the same maturity. The interbank market is a platform for unsecured lending among banks and thus quoted rates incorporate liquidity and credit risk premia. The interest rate swap contains expectations about the future path of the policy interest rate but practically no credit and liquidity premia, since the swap transaction does not involve transfer of funds. Therefore the spread reflects the two risk premia. While the liquidity premium is driven by the needs and availability of funds, credit risk is linked to the counterparty risk. Under normal market conditions, the spread is positive but close to zero as the credit and liquidity risk premia are small. An increase in the spread indicates rising market pressures. Both series are downloaded from Bloomberg.

13. The liquidity premium is identified through a set of proxy variables. In the literature, liquidity premia are only indirectly or partially identified. In its indicative decomposition of interbank rates, BoE (2007) identifies the liquidity premium as the residual (the so called non-credit risk premium) after accounting for credit risk. Michaud and Upper (2008) quantify market liquidity, while the liquidity of borrowing banks and technical factors of the market remain unobserved. This paper uses several proxies for market liquidity premia, including deposits of institutional investors, short-term central bank’s instruments and the central bank’s temporary extended liquidity facility (see also FSR, 2010). Market premia in the U.S. and Euro interbank markets are also inlcuded to control for spillover effects.

14. Banks’ counterparty risk can be approximated by credit spreads. Counterparty risk is essentially the risk that the unsecured loan will not be repaid due to a default of the debtor. Such a risk is embedded in banks’ bond credit spreads and credit default swaps, so they are often used as proxy variables for credit risk. For instance, BoE (2007) uses CDS spreads to identify the credit-risk component of the interbank market spread, while FSR (2010) uses banks’ bond credit spreads for that purpose.

15. The specification of the Chilean interbank market spread (IMS ) includes domestic and global risk factors:

IMSt=c+αΔMFt+βΔPFt+γΔCBt+δIMSEUt+θIMSUSt++ωDt+ρCSt+φCDSt+ϵt,
  • MF PF and for the stock of time deposits by mutual funds and pension funds.

  • CB denotes the stock of central bank’s short-term instruments. It accounts for the regular liquidity operations by the central bank.

  • D is a dummy for the period of expanded liquidity operations by the central bank. Since October 2008, the central bank accepted bank deposits as collateral for the 7-day repo operations. This measure, initially introduced for six months, was subsequently extended through the end of 2009, and the transaction tenor was prolonged up to 28 days. In December 2008, the central bank introduced a collateralized line of credit for transactions exceeding 28 days, in which it accepted General Treasury bonds, among others, as collateral. And since mid-2009, a new facility was established (tenors of 90 and 180 days), through which banks accessed funding from the central bank at prevailing monetary policy rate. Further, the central bank introduced 28-day dollar swap auctions. The Ministry of Finance transferred government’s dollar funds from abroad and deposited them as term deposits in local banks, and also auctioned dollar deposits. In the regressions, the effects of these policy measures are accounted for by a dummy variable, which equals one from November 2008 to mid-2010 and zero otherwise.

  • IMSEU denotes Euro interbank market spread, which is the difference between the three-month euro interbank market rate and the overnight euro interest rate swap for the same maturity. It measures liquidity and credit risk pressures in the euro interbank market.

  • IMSUS denotes dollar interbank market spread, which is the difference between the three-month dollar federal funds rate and the overnight dollar interest rate swap at the same maturity. It measures both the liquidity and credit risk pressures in the dollar market.

  • The credit risk premium is measured by banks’ bond credit spread, denoted by CS, and CDS of European banks, labeled as CDS. And εt~N(0,σ).

Data for the domestic variables is from the Banco Central de Chile, while IMSEU and IMSUS are from Bloomberg.

C. Results

16. The results point to moderate spillovers from global financial stress. Although spillovers were clearly one of the driving factors of domestic funding spreads, the magnitude of the effects is relatively small, especially after 2009. The estimates suggest that global spillovers pushed up funding cost in Chile by about 50 basis points on average from mid-2010 to January 2012. Pressures in the U.S. interbank market were the key driver of changes in Chile’s interbank market spreads in 2008–09. More recently, financial tensions in the euro area have been the main source of spillovers. Both bank fundamentals and global factors have been important determinants of changes in the bond spread.

The Bond Market

17. The bank bond spread has been driven by banks’ fundamentals as well as global risk factors. Table 1 shows the regression results for the bond spread. All coefficients are correctly signed and statistically significant. A decomposition of the spread shows that fundamental factors accounted for the largest portion of the spread until mid-2010 (see Figure 2). In the period since then, spillovers from global risk factors (proxied by VIX and CDS of European banks) have became more important.

Table 1:

Bond Market Spread

article image
Note: Standard errors have been adjusted for autocorrelation and heteroscedasticity; Nobs = 1254.
Figure 2.
Figure 2.

Chile: Funding Markets, 2008-12

Citation: IMF Staff Country Reports 2012, 266; 10.5089/9781475510492.002.A002

Source: Banco Central de Chile, SBIF, Bloomberg, and own calculations.

18. Domestic liquidity factors have also played a role. Changes in the stock of time deposits of pension funds correlate negatively with bond market spreads, which suggests that when pension funds increase the share of domestic assets in their portfolio, they invest in both deposits and bank bonds. On the other hand, an increase in the time deposits of mutual funds increases bond market spreads, since mutual funds invest mostly in domestic assets and over time shift from deposits to bonds and vice versa.

19. The evidence from partial regressions confirms the robustness of the results. Isolating the effect of banks’ fundamental factors and global risk factors individually (Table 1, last two columns) shows that both sets of variables are robust explanatory variables of the bond credit spread.

The Interbank Market

20. The interbank market spread contains both domestic and external risk premia. As shown in Table 2, the interbank market spread has been driven by domestic liquidity and credit risk factors as well as global spillovers. These factors together explain 70 percent of the variation in the spread.

Table 2:

Interbank Market Spread

article image
Note: Standard errors have been adjusted for autocorrelation and heteroscedasticity; Nobs = 1255.

21. Among domestic factors are local liquidity shocks, policy measures, and counterparty risk. Activities of institutional investors, such as shifts in time deposits of pension and mutual funds, affect banks’ liquidity. Reduction of institutional time deposits reduces liquidity and increases the interbank spread. In addition, the interbank spread has been affected by the central bank’s extended liquidity operations from November 2008 till mid-2010. The results suggest that these operations reduced the interbank market spread by about 24 basis points. The counterparty risk, which is proxied by the banks’ bond credit spread, is also a significant factor. One percentage point increase in the bond spread leads to about 50 basis points increase in the interbank market spread.

22. Spillovers from financial tensions abroad also afected risk premia in the interbank market. The interbank spread has been affected by changes in global risk factors (see Figure 2). In particular, one percentage point increase in the interbank market spread in the U.S. or in the CDS of European banks triggers about 30 basis points rise in Chile’s interbank market spread. Pressures in the U.S. interbank market were the key determinant of the Chile’s interbank market spread until early-2009. In the reminder of 2009 and until mid-2010, the spread fell as external pressures dissipated and domestic credit risk premium delined. Since mid-2010, however, spillovers from hightened financial tensions in Europe have played a prominent role. The resutls remain similar if the CDS of Citibank (a proxy for CDS of credit risk in the U.S. banks, data from the Bloomberg) and the VIX are added to the regression (both variables turn out to be insignificant, see Table 2, last two columns).

D. Concluding Remarks

23. This paper analyzed pressures in the bank funding markets in Chile from mid-2008 to early 2012 with particular focus on spillovers from global risk factors. The main findings are the following:

  • The interbank market in Chile had been severely affected by tensions in the U.S. interbank market after the Lehman crisis. As a result of aggressive policy responses in Chile and abroad, pressures dissipated by mid-2009.

  • Between mid-2008 and mid-2009, banks’ funding cost were also driven up by deteriorating fundamentals of the banking sector. Subsequent improvements in fundamentals significantly lowered spreads between mid-2009 and mid-2010.

  • Since mid-2010, funding market spreads have been driven mainly by spillovers from financial tensions in Europe. However, spillovers so far have been moderate and have not affected credit intermediation.

References

  • BoE, 2007, “An Indicative Decomposition of Libor Spreads,” Markets and Operations, Quarterly Bulletin, Q4, Bank of England.

  • Duffee, G. R., 1998, “The Relation Between Treasury Yields and Corporate Bond Yield Spreads,Journal of Finance, Issue 53, pp. 222541.

    • Search Google Scholar
    • Export Citation
  • FSR, 2010, “Determinants of the Prime-swap Spread,” Financial Stability Report, Second Half, Banco Central de Chile.

  • Merton, R. C., 1974, “On the Pricing of Corporate Debt: The Risk Structure of Interest Rates,Journal of Finance, Issue 29, pp. 44970.

    • Search Google Scholar
    • Export Citation
  • Michaud, F.-L., Ch. Upper, 2008, “What Drives Interbank Rates: Evidence from the Libor Panel,” BIS Quarterly Review, March, Bank for International Settlements.

    • Search Google Scholar
    • Export Citation
  • Otker-Robe, I., J. Podpiera, 2012, “Explaining Credit Default Swaps Pricing for Large Banks,The Capco Institute Journal of Financial Transformation, Series on Risk, Issue 34, pp. 6375.

    • Search Google Scholar
    • Export Citation
1

Prepared by Jiri Podpiera.

Chile: Selected Issues
Author: International Monetary Fund