Guatemala: Selected Issues and Analytical Notes

Abstract

Guatemala: Selected Issues and Analytical Notes

Macro Financial Linkages: Assessing Financial Risks

A. Stress Testing the Financial System1

This section assesses the resilience of the banking system to a variety of shocks. We employ financial soundness indicators and a top-down stress test to evaluate the health of the banking system. The results suggest that the system could withstand a range of sizable shocks but moderately high dollarization of bank loans, relatively large exposure to the government, and growing bank foreign liabilities as well as maturity mismatches in U.S. dollar positions represent vulnerabilities.

Introduction

1. Guatemala’s financial system is dominated by banks, which operate as part of financial conglomerates. Financial system assets were 64 percent of GDP at end-2015, with banking system assets representing 83 percent of the total. Most financial institutions (representing about 90 percent of total assets) are part of the 10 financial conglomerates operating in Guatemala, and the majority of those that are part of conglomerates (2/3) are owned by the three largest domestic conglomerates. Financial conglomerates are organized under a local bank and comprising domestic and foreign subsidiaries as well as off-shore banks. Offshore banks represent about 8.4 percent of total financial system assets, and their role is to raise deposits in U.S. dollars in Guatemala and lend mainly to corporations in Guatemala and to a lesser extent to clients in other countries.2

2. The banking system has gone through consolidation and regionalization in recent years. Concentration of assets has increased and five banks hold 80 percent of total banking system assets. Latin American and U.S. banks hold 22 percent of assets while Guatemalan banks hold subsidiaries in Costa, Rica, El Salvador and Honduras, and lend abroad to corporates and banks in these three countries.

A04ufig1

Regulatory Capital to Risk-Weighted Assets (%)

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Source: IMF FSIs.

3. The financial system appears sound though capital is below the level of regional peers while dollarization and external rollover risks continue to represent vulnerabilities. At 14.1 percent,3 reported regulatory capital is above the minimum requirement (10 percent), for the system as a whole and for individual banks. However, capitalization is low by regional standards (including Tier 1 capital) and has been declining over the past few years. The NPL ratio increased slightly in 2015, but at less than 1½ percent of total loans remains low. The ratio of provisions to NPLs stood at about 67 percent, but would be lower if adjustments are made relatively generous rules for collateral valuation.

4. The banking system has become more dollarized, and over a third of financial intermediation is conducted in U.S. dollars. The degree of credit dollarization has increased from 27 percent of total loans at end-2010 to 36 percent in March 2016. The dollar loan-to-deposit ratio has been growing over time and reached 180 percent for the system as a whole in 2015 (it was above 200 percent on average for the three largest banks). Credit lines from abroad (mainly by the three largest banks) are used to fund FX loans that are not funded by FX deposits. While liquidity indicators are robust, with liquid assets accounting for 37 percent of total liabilities, rising bank foreign liabilities, currently at a decade high of about 15 percent of the total liabilities (comprising about half of all funding in foreign currency), represent a potential vulnerability. In addition, government bond holdings (11 percent of bank assets) are relatively large and about half of the bonds are in the available-for-sale portfolio used to manage liquidity, hence subject to market risk solely in case those government bonds are sold (excluding those held to maturity) and marked to market. Overall, private credit growth of 11 percent is moderate, and while FX loan credit growth exceeded domestic currency loan credit growth after the global financial crisis, the former has recently slowed sharply as large electricity projects financed by loans in foreign currency have been completed.4 Growth in FX loans has now come down to 9 percent, below the growth in domestic currency loans of 12 percent.

A04ufig2

Deposit and Loan Dollarization (%)

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Source: IMF FSIs.
Table 1.

Guatemala: Financial Soundness Heat Map

article image
Source: Fund staff estimates.

Scenarios for Solvency Stress Tests

5. Four scenarios were considered in the stress tests to assess the stability of the banking system. Full-fledged macroeconomic projections were quantified for an adverse macroeconomic scenario. An adverse scenario could have important consequences for Guatemala’s economy due to the combined shock: capital inflow shocks, exchange rate depreciation, surge in the domestic interest rate, and less investment. In this first scenario, the cumulative decline in economic growth relative to the baseline is 4 percentage points or 2 standard deviations based on the data from the 1980-2015 (i.e. zero percent), which is less severe than the recession experienced in the 1980s but more severe than the growth shock during the global financial crisis.5 Second, the shock to the nominal exchange rate assumes a 20 percent depreciation.6 Third, an interest rate shock assumes a 2 percentage point increase in the nominal interest rate on all assets. Fourth, a combined shock scenario was generated by the general equilibrium model to ensure consistency of macro variables. It assumes an initial shock of 20 percent exchange rate depreciation, the associated increase in the risk premium of 200 basis points, which captures an increased credit risk from unhedged borrowers. The model then suggests that a consistent impact on interest rates would be an increase by 2 percentage points and a decline in growth of 4 percentage points.

A04ufig3

Macroeconomic Scenarios Real GDP in year 0=100

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Source: IMF staff calculations.

Satellite Model for Credit Risk

6. To assess the impact of the shocks to growth, interest rates and exchange rate on bank-specific non-performing loan ratios a dynamic panel model was employed. The model was estimated based on a panel dataset including 18 banking institutions and quarterly observations for the period 2001:Q1 through 2013:Q3. The model specification was as follows:

LNPLi,t=μi+αLNPLi,t1+β1gt4+β2(δgt4)2+γ1riri,t4+γ2inflt1+γ3RDEPt1+ɛi,tj

where the indices i and t indicate, respectively, the banking institution and the time period.

LNPL denotes the logistic transformation of the NPL ratio: LNPL=In(NPL1NPL),g denotes real GDP growth (annual rate); rir is a real interest rate,7 infl is the four quarter rate of inflation based on the CPI, RDEP denotes the rate of growth of the rea1 (bilateral) exchange rate, where a positive value indicates a real depreciation of the Quetzal against the US dollar, and μi denotes bank-specific fixed effects. The estimated coefficients are presented in the following table:

Table 2.

Satellite Model for Credit Risk Dependent Variable: NPL Ratio (Logistic Transformation)

article image
Note: ***, **, * denotes statistical significance at the 1, 5, and 10 percent level.

The estimated model is non-linear in real GDP growth: the NPL ratio increases at an accelerated rate as real GDP growth declines further. Similar results were obtained using slightly different specifications of the equation above.

Overall, the NPL-ratio under stress was computed as:

NPLtstress=(NPLt1initial1NPLt1initial)exp{βΔXt}/[1+(NPLt1initial1NPLt1initial)exp{βΔXt}]

where Xt is the vector of macroeconomic factors used and β is the vector of coefficients (from the credit risk model). ΔXt represents the change in the levels of macro variables.

7. An adjustment of provisions and initial capital was also made. Specifically, effective provisioning rates were calculated as follows:

Adjusted effective provisioning rate=Nominal provisioning rate*(1Adjusted collateral valueLoan value)

The tests are undertaken under more rigorous rules for collateral valuation, given the difficulty encountered by the banks in recovering collateral. Collateral values are discounted 20 percent relative to the current market values, and a further adjustment is introduced to take account of the average time to recover collateral via lawsuits—about 3 years. Thus, the adjusted collateral value is calculated as follows:

Adjusted collateral value=Original collateral value*0.8*(1+0.07)-3

In the calculation, the discount rate is set at 7 percent. This value approximates the nominal interest rate at a 3year horizon corresponding to the baseline yield curve in domestic currency.

With respect to the capital used in the stress tests for the capital adequacy ratio, the capital considered is the equity of banks and does not include such instruments as subordinated debt and other hybrid capital, while the regulatory capital typically counts these instruments as capital.

8. Estimates from the credit risk model suggest that credit risk in the loan book is an important risk factor for the banking system. Loans to the economy represent more than half of total banking sector assets, of which a large share is directed towards the corporate sector. As a result of the decline in GDP growth, the depreciation of the Quetzal, and the rise in interest rates, the NPL ratios increase by 1 percentage point in the adverse scenario.

Results of Solvency Stress Tests

9. Banks appear quite resilient to high levels of stress, but there are areas of vulnerability.

  • After making adjustments to provisions to conduct the tests under more rigorous rules for collateral valuation and calculation of loan loss provisions, initial capital was adjusted downwards by 0.6 percentage points.

  • Under the adverse growth scenario (i.e. zero percent economic growth, see sections on “Scenarios for Solvency Stress Tests” and “Satellite Model for Credit Risk”), the CAR for the system falls by 0.9 percentage points.

A04ufig4

Impact of Stress on the CAR

(Percent)

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Source: IMF staff calculations.Note: Solvency stress test for FX risk assumes 20percent depreciation; test for interest rate risk assumes a 2 percentage points nominal interest rate increase; test for combined risk adds impact on the CAR from credit, FX, and interest rate risks.
  • Direct exchange rate risk is contained, given that most banks hold positive net open FX positions in foreign currency, with assets more than offsetting liabilities denominated in U.S. dollars. The net open position is equivalent to 1.3 percent of total assets, and it is negative in only three banks (equivalent to less than 1 percent of assets in 2 banks and 1.6 percent of the assets of a third bank). The situation implies that the banks would benefit from a depreciation of the Quetzal if the “indirect” credit risks associated with the depreciation are ignored. The CAR for the system would improve by 0.4 percentage points, with a depreciation of 20 percent vis-a-vis the U.S. dollar.

  • Indirect exchange rate risk, however, could trigger credit losses. A depreciation of the Quetzal by 20 percent would increase the debt burden and reduce the repayment capacity of un-hedged foreign currency borrowers. Based on the estimated model for credit risk (see section on “Satellite Model for Credit Risk”), having a 20 percent depreciation will make 13 percent of FX loans become NPLs, which would cause a 0.9 percentage points decline in the CAR. In a different and more adverse scenario, the assumption is that 40 percent of FX loans become NPLs, given that 40 percent of borrowers in foreign currency are un-hedged. This causes a 2.5 percentage points decline in the CAR. In this latter scenario, if the effects from both direct and indirect exchange rate risks are combined, the CAR falls by 2.1 percentage points.

  • Potential losses, in particular for larger banks, driven by interest rate risk, which materializes because of losses in interest income and valuation losses on government bonds, are notable. The valuation losses correspond to bonds held in the “available for sale” portfolios that were marked to market. As a result of securities losses, the CAR declines by 2.6 percentage points. If both the net interest income and repricing impacts are assessed, the CAR falls by 2.8 percentage points.

  • While the banking system can handle individual shocks well, a combined shock scenario based on a consistent set of assumptions from a general equilibrium macro model with the initial shock of 20 percent depreciation and an increase in interest rates of 2 percentage points would lead to an increase in credit risk from un-hedged borrowers and to valuation losses on government bonds, and will push the capital ratio below the regulatory minimum (i.e. the CAR falls by 7 percentage points).

Results of Liquidity Stress Tests

10. Banks appear quite resilient to severe liquidity shocks, but there are areas of weakness.

  • Liquidity stress tests assumed a combination of 10 percent run-off rates on domestic currency and FX deposits, and 5 percent on other FX liabilities and 3 percent on other domestic currency liabilities per month. Under this scenario most banks would remain liquid after 5 months. If 15 percent run-off rates per month are assumed on domestic currency and FX deposits (and the same run-off rates as above on other domestic currency and FX liabilities), more than half of the system becomes illiquid after 5 months.

  • Alternatively, if the impact from a 25 percent run-off rates on FX deposits, and 35 percent run-off rates on other FX liabilities (e.g. foreign credit lines) is assessed, one large and one middle-sized bank become illiquid after 5 months. Thus, reliance on foreign credit lines represents an area of vulnerability.

  • Maturity mismatches in U.S. dollar positions (which are presumably higher than in domestic currency positions, in particular in the short term) could represent a significant vulnerability though this could not be tested due to the lack of data on the maturity structure of assets and liabilities.

A04ufig5

Impact of Stress on Liquid Assets

(Percent)

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Source: IMF staff calculations.Note: Liquidity stress test assumes a 10 and 5 percent per month withdrawal of domestic demand deposits and other liabilities respectively, and a 10 and 3 percent per month withdrawal offoreign demand deposits and other foreign liabilities respectively.
Table 3.

Banking Sector Financial Soundness Indicators

article image

1=low risk, 2=increased risk, 3=high risk, 4=very high risk. The ratings are averaged across banks.

Table 4.

Results of Stress Tests

article image

Assumes 20 percent discount to initial collateral values, and another adjustment to take into account the average time to recover collateral is 3 years.

Assumes a 2 percentage points nominal interest rate increase.

Assumes a depreciation of the exchange rate of 20 percent, and 40 percent of FX loans become NPLs.

Adds aggregate losses caused by individual shocks (assumptions on individual shocks are maintained the same).

Assumes a 10 and 5 percent per month withdrawal of domestic demand deposits and other liabilities respectively, and a 10 and 3 percent per month withdrawal of foreign demand deposits and other foreign liabilities respectively.

B. Domestic Bank Network Analysis and Cross-Border Financial Spillovers8

The recent financial crisis has proven that stress events in individual institutions can spill over and undermine the stability of the entire financial system, highlighting the need to track direct and indirect financial interconnections among financial institutions. This note looks at domestic interbank linkages and cross-border bank interconnections in Guatemala to assess whether linkages across institutions within and beyond the country’s borders may have systematic implications. Results suggest that the Guatemalan banking sector is relatively resilient to shocks originating both domestically and abroad though some banks deserve higher scrutiny in terms of their vulnerability and the level of contagion they can generate.

Bank Network Analysis

11. To assess potential systemic implications of domestic interbank linkages we simulate the cascading effect upon the failure of selected banks due to credit and funding shocks. We do so by using the interbank exposure model developed by Espinosa and Sole (2010)9 which is designed to track the domino effects triggered by the hypothetical default of each bank in the system through the resulting credit losses and funding shortfalls (Figure 1).

12. The analysis is based on a stylized balance sheet identity that highlights the role of interbank exposures. In the identity, bank loans and other assets are funded by (i) capital, (ii) long-and short-term borrowing excluding interbank loans, (iii) deposits, and (iv) interbank borrowing. To analyze the effect of a credit shock, the model simulates the individual default of each institution’s interbank obligations in the network. For different assumptions of loss given default, it is assumed that the capital of creditor banks absorbs the losses on impact, triggering the creditor bank’s default if its capital is insufficient to fully cover losses. In addition to its direct credit exposures to other institutions, a bank’s vulnerability also stems from its inability to roll over all or part of its funding in the interbank market, and thus having to sell assets at a discount in order to re-establish its balance sheet identity. In this respect, a credit shock may be compounded by a funding shock and associated fire sales losses if default of an institution also leads to a liquidity squeeze. In the model, it is assumed that institutions are unable to replace all the funding previously granted by the defaulted institutions, which in turns generates a fire sale of assets. Default is again triggered if bank capital is not enough to absorb the funding-shortfall induced loss.10

Figure 1.
Figure 1.

Network Analysis Based on Interbank Exposures

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Source: Espinoza and Sole (2010)

13. Interbank exposures in Guatemala are small compared to banks’ capitalization. Interbank positions amount on average to 6 and 5 percent of lenders’ and borrowers’ bank capital respectively, hence the resulting capital and funding risks are contained. Only one bank (B3) has individual credit exposures to 2 other banks (B9 and B16) exceeding 100 percent of its capital (Figure 2). However, interbank exposure ratios mask important differences in bank capitalization, with bank capital ranging between 0.1 and 6 ½ billion quetzales (Figure 3).

Figure 2.
Figure 2.

Interbank Exposures

(Percent of pre-shock capital)

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Source: SIB. Note: The table shows downstream exposures of the banks in the first column to all other banks in percent of the lender bank’s capital. Note: Grey: No exposure; Green: Exposure <5%; 5%<= Yellow <10%; 10% < = Orange <20%; Red>=20%.
Figure 3.
Figure 3.

Bank Capital

(Billions Quetzales)

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Source: SIB.

14. As a consequence, capital impairment due to bilateral interbank exposures is limited. In particular, capital losses following individual bank defaults—a measure of how much the system would be weakened by the transmission of financial distress across institutions—reach significant levels only for one bank (Figure 4), the same that was identified in the bank exposure matrix. However, such downstream exposure does not translate in significant funding risk for the borrower banks (B9 and B16) given their higher capitalization.

Figure 4.
Figure 4.

Capital Impairment

(Percent of Pre-shock Capital)

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Note: Failed banks are listed in the first column, and the percentage capital loss for all other banks is listed in the corresponding row.

15. Accordingly, the degree of contagion and vulnerability across the system is contained and, where significant, it arises from credit risk. Figure 5 presents average capital losses in the network due to the default of individual banks, as opposed to bilateral vulnerabilities shown in the capital impairment matrix. Contagion and vulnerability indices are contained overall, and clustered around a few banks (banks 3, 9 and 16). The analysis also allows distinguishing between the credit and funding channel, with banks 9 and 16 being contagious, and bank 3 reciprocally vulnerable to credit risk due to the high credit exposure of the latter to the formers but not so much to the funding shocks. Hence, the default of banks 9 and 16 would trigger the default of bank 3 (Figure 6), although contagion would be limited to one round.

Figure 5.
Figure 5.

Index of Contagion and Vulnerability

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Figure 6.
Figure 6.

Domino Effect and Contagion Path

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Source: SIB, and staff calculations based on Espinoza and Sole (2010).

16. The analysis suggests that the domestic bank network is relatively resilient to contagion effects but some banks deserve higher scrutiny. While interbank exposures are limited on average, the system may be weakened by the transmission of financial stress across institutions. It is therefore important that the regulators track potential systemic linkages and maintain a higher scrutiny on those banks which are relatively more vulnerable (B3) or hazardous (B9 and B16) to the system in order to contain externalities in case of financial distress of individual institutions.

Cross-Border Financial Spillovers

17. We used the IMF Bank Contagion Module to assess the impact of financial spillovers to Guatemala from stress in international banks. Based on BIS banking statistics and bank-level data, the model estimates potential rollover risks for Guatemala stemming from both foreign banks’ affiliates operating in Guatemala and foreign banks’ direct cross-border lending to Guatemala borrowers. 11 Rollover risks were triggered in the scenarios analyzed here by assuming bank losses in the value of private and public sector assets in certain countries and/or regions. If the banks do not have sufficient capital buffers to cover the losses triggered in a given scenario, they have to deleverage (reduce their foreign and domestic assets) to restore their capital-to-asset ratios,12 thus squeezing credit lines to Guatemala and other countries. The estimated impact on losses in cross-border credit availability for Guatemala also incorporates the transmission of shocks through Panama, given its central financial role in the region. The assumption is that cross-border lending from Panama to Guatemala declines proportionally to the decline in cross-border lending to Panama from the banking systems where the shocks originate.13

18. Spillovers to Guatemala from stress in international banks are moderate and lower than in regional peers. The impact on foreign credit availability in Guatemala of the severe stress scenarios in asset values of BIS reporting banks, presented in the text figure and the table below, is lower than in other countries in the region, with the exception of Honduras. As of October 2015, the most sizable impact on claims on Guatemalan borrowers would stem from shocks in the US and Canada. Spillovers from a 10 percent shock to assets originating in the U.S. and Canada would reduce credit in Guatemala by 2.7 percent of GDP (or 5.5 percent of total domestic and cross-border credit to the public and private sectors). In contrast, a similar shock would reduce credit in Panama and El Salvador by 16 and 8 percent of GDP respectively. More generally, the level of upstream exposures of Guatemala to international banks14 implies an upper limit on the losses of about 7 percent of GDP (or 14 percent of total domestic and cross-border credit to the public and private sectors in Guatemala).15 This upper limit would correspond to a worst case scenario without any replacement, either domestic or external, of the loss of credit by BIS reporting banks to Guatemalan borrowers.

A04ufig6

Spillovers from International Banks’ Exposures as of October 2015: Effect on Credit of 10% Loss in All Bank Assets of BIS-Reporting Banks, RES Bank Contagion Module

(Percent of GDP)

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

1/ In Panama, the loss of credit includes credit by banks in the offshore center with minimal links to the domestic economy.
Table 5.

Spillovers from International Banks’ Exposures

(As of October 2015)

article image
Source: Research Department Macro-Financial Division Bank Contagion Module based on BIS, ECB, IFS,

Percent of on-balance sheet claims (all borrowing sectors) that default.

Reduction in foreign banks’ credit due to the impact of the shock on their balance sheet, assuming uniform

Greece, Ireland, and Portugal.

Greece, Ireland, Portugal, Italy, Spain, France, Germany, Netherlands, and the UK.

19. Spillovers from a shock originating in the U.S. assets only are relatively larger, and financial regional integration is important in the transmission of shocks. The impact of a 10 percent loss in U.S. assets value on cross-border credit availability in Guatemala would be 2.3 percent of GDP. 16 This effect stems from the large share of U.S. banks in total foreign bank claims on Guatemala, although the strengthening in international banks’ capital buffers and the cross-border deleveraging of assets after the global financial crisis is likely to have mitigated risks. As of October 2015, a 10 percent loss on European assets would result in a reduction in credit availability to Guatemala of about 1.1 percent of GDP.17 Increasing importance of financial integration with other countries in the region plays an important role in the transmission of shocks from Europe. Indeed, almost one third of the estimated credit losses in Guatemala (0.3 percent of GDP) resulting from a shock originating in Europe would be transmitted through cross-border lending from Panama, which is more dependent on European banks’ funding (Figure 7).

Figure 7.
Figure 7.

Foreign Bank Claims

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

C. Balance Sheet Analysis18

This section provides an update of the balance sheet analysis (BSA) of the Guatemalan economy presented in the 2013 Article IV report.19 The net external debtor position of the economy increased further, but this was again driven by continued FDI flows to the private sector, and the country maintained a small net debtor position excluding FDI, implying limited external risks. Risks from currency mismatches appear limited at the aggregate sectoral level, although unhedged borrowers in foreign currency (FX) present vulnerability. Private sector debt has increased moderately over the last decade.

20. External risks remain limited given the country’s small and stable net external debtor position excluding FDI liabilities. Guatemala had a total net external debtor position of about 21 percent of GDP in 2015, up from 18 percent in 2012, but this largely reflects continued increase in FDI liabilities, which are a sign of a strong capital structure at the country level, with stronger reliance on equity rather than debt. Excluding FDI liabilities, the economy has a small net debtor position of about 1½ percent of GDP, implying limited risks of a capital account crisis (Figure 8, left and Table 3). This position has been broadly unchanged since 2012, as a small increase in the net external debtor position of the financial sector—from 2 to 3 percent of GDP—and a small decline in the net external creditor position of the central bank were largely offset by the decline in the net external debtor position of the government as a share of GDP (Figure 8, right).20

Figure 8.
Figure 8.

Net External FDI and Debt Positions

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

21. Notwithstanding limited currency mismatches at the aggregate level, the rising share of FX bank credit highlights possible risks from unhedged borrowers. Currency mismatches by sector were little changed over the last few years. Despite a small decline in reserves of the central bank in percent of GDP, the net negative FX position of the consolidated public sector declined slightly from about 3 percent to about 2¼ percent of GPD since 2012, driven by the decline in external debt of the central government as a share of GDP (Table 6). The net positive FX position of the financial sector increased slightly from about ½ percent of GDP in 2012 to about 1 percent in 2015, as banks continued channeling additional external financing into domestic credit in FX. The non-financial private sector maintained a balanced net FX position—excluding FDI liabilities—with foreign assets broadly offsetting the sector’s small net FX debtor position vis-à-vis banks (Tables 6 and A1). At the same time, the rising share of credit in FX over the last few years highlights possible risks from unhedged borrowers (Section A of this note), although the concentration of such credit in the corporate sector, especially large corporations that are more likely to have natural hedges through export revenues, appears to be a risk-mitigating factor (Figure 9).

Figure 9.
Figure 9.

Bank Credit

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Source: National authorities and Fund staff estimates.

22. Private sector balance sheets appear healthy with limited accumulation of debt in the last decade. Private sector debt reached about 40 percent of GDP in 2015, up from a trough of 33 percent of GDP in 2010, but only slightly higher than the previous peak of 38 percent of GDP reached in 2007. More detailed bank credit and International Investment Position data by sector shows that credit to non-financial corporates (NFCs) in 2015 was still marginally below the peak reached in 2007, after recovering from 25 percent of GDP in 2010 to almost 30 percent.21 Credit to the household sector was not affected by the global credit crisis, remaining around 8 percent of GDP through 2007-11, and then gradually rising to 10½ percent of GDP by 2015. Most of the FX loans were on the corporate rather than household side.

A04ufig7

Guatemala, Non-Financial Private Sector Debt

(in percent of GDP)

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Sources: National authorities and Fund staff estimates.
Table 6.

External and Foreign Currency Positions

article image
Sources: Bank of Guatemala and Fund staff estimates.

Excluding net FDI position.

D. Dynamic Macro Financial Linkages22

This section simulates the effect of financial shocks on the real economy using the IMF Flexible System of Global Models. Results indicate that the monetary stimulus injected in 2015 could continue working through the economy this year or even in the next two years and peak in 2018 if the pass-through proves to be slow but persistent. Enhanced competition in the banking sector resulting in lower interest rate margin, as well as lower collateral requirements could positively and significantly contribute to economic growth, while an exchange rate shock could take a toll on growth through increased NPLs and market risk premia.

23. We used the IMF’s Flexible System of Global Models (FSGM) to simulate the effect of various financial shocks on macroeconomic variables. The model was developed by the Economic Modeling Division of the IMF’s Research department for policy analysis (Andrle and others, 2015). It comprises a system of multi-region globally consistent general equilibrium models combining micro-founded and reduced-form relationships for various economic sectors. The FSGM has a fully articulated demand side, and the supply side features are pinned down by Cobb-Douglas production technology. International linkages are modeled in aggregate for each country/region. The level of public debt in each country and the resulting implications for national savings determine the global real interest rate in the long run. The parameters of the model, except those determining the cost of adjustment in investment, have been largely estimated from the data using a range of empirical techniques. Real GDP is determined by the sum of the components of demand in the short run and the level of potential output in the long run. The households’ consumption-savings decisions are explicitly micro founded as are firms’ investment decisions. The OLG formulation of the consumption block gives the model important non-Ricardian properties, whereby national savings are endogenously determined given the level of government debt. Government absorption is determined exogenously, while imports and exports are specified with reduced-form models.

24. Transmission of financial shocks to the real economy takes place through changes in interest rates and risk premia. All interest rates are related to the risk-free interest rate, whose closest parallel is the monetary policy rate, from which they deviate because of risk premia or different maturities. The model includes several risk premia: one for the sovereign (which applies to the entire domestic economy), one that applies to both domestic households and firms, one that applies only to firms, and one that applies to the currency (or country). The expectations theory of the term structure determines the 10-year interest rates and there is an additional term premium. Interest rates related to consumption, investment, and holding of government debt and net foreign assets are weighted averages of the 1- and 10-year nominal interest rates. The exchange rate in the short run is determined via the uncovered interest parity condition, while in the long run it adjusts to ensure external stability given households desired holdings of net foreign assets.

25. We simulate 4 scenarios which capture the macro-economic impact of shocks to monetary policy rate, the exchange rate and market premia, lending margins, and collateral requirements. The first scenario simulates an expansionary monetary policy shock, modeled as a 100 basis points cut in the monetary policy rate, both under full pass-through and under a slower, but more persistent transmission to lending rates (Panels 1 and 2 in figure below); the second simulation looks at currency depreciations of 5 and 20 percent that assumes a non-linear reaction in market risk premia of 25 and 300 basis points (bp) respectively (Panels 3 and 4); the third scenario considers a reduction in bank lending margins resulting in a reduction in market risk premia of 100 basis points (Panel 5); and the fourth simulates a decrease in collateral requirements, which facilitates higher credit growth that is partially financed by higher foreign borrowing (Panel 6). Aside from nonlinearities that can arise owing to the zero lower bound on nominal interest rates, the model is symmetric and roughly linear; therefore, the simulated shocks can be scaled up or down to consider scenarios of different magnitudes.

26. Results indicate that the recent 100 basis points cut in the monetary policy rate could increase GDP substantially this year or even in the next two years. Specifically, if fully passed through to domestic lending rates, the monetary impulse provided during 2015 (a cumulative rate cut of 100 basis points to 3 percent amid declining inflation) could increase GDP in 2016 by almost 0.4 percentage points relative to the baseline in the absence of such stimulus due to both higher domestic absorption and export demand. The latter effect is induced by a depreciation of the quetzal following lower real interest rates compared to foreign rates. However, the policy rate signaling function in Guatemala is constrained by weaknesses in the transmission from the monetary policy rate to bank lending rates (and ultimately to prices and output). Under a slower pass-through scenario, where only 1/3 of the impulse is passed to lending rates, the resulting increase in GDP in 2016 would be only 0.1 percent, but the delayed, more persistent impulse would increase GDP by 0.2 percent in the medium term (2017-2019) with a peak impact in 2018.

27. Conversely, a large currency depreciation could decrease output by up to 1 ½ percent in the short and medium term, through its effect on NPLs and decreased risk appetite. In scenario 2, a depreciation of the quetzal triggers an increase in market premia due to higher NPLs in the context of the high degree of credit dollarization to un-hedged borrowers (40 percent of total credit). The risk premium reaction to the depreciation is modeled as nonlinear, since defaulting loans—and with them banks’ risk aversion—are likely to increase more than proportionally with the extent of the depreciation. Hence, a 5 percent depreciation/20 bp increase in risk premia results in a real GDP contraction of about 0.2 percent in the medium term. Conversely, under a 20 percent depreciation/300 bp increase in risk premia, GDP declines by up to 1½ percentage points three years after the FX shock as household consumption and private investment shrink. In both scenarios, the short-term effect is a one-year boost to real GDP as following the depreciation exports increase sufficiently to offset the adverse impact on domestic demand from higher risk premia.

28. Increasing efficiency and competition in the banking sector could raise GDP by up to 2 percent in the medium term. At about 8 percent, banks spreads are relatively high in Guatemala, which is a constraint for widespread access to credit and financial inclusion. In scenario 3, we simulate a decrease in banks’ net interest rate spreads which translates into a 100bp permanent reduction in market risk premia. Lower rates increase both private investment and household consumption expenditure, fostering domestic absorption. The resulting increase in inflation triggers a policy rate increase that produces an appreciation of the quetzal and lowers exports. The net effect on GDP is positive and persistent, with the increase ranging from less than ½ percent in the first year to more than 2 percent after six years.

Figure 10.
Figure 10.

FSGM Model Simulations of Macro Financial Linkages

(Percent difference from baseline)

Citation: IMF Staff Country Reports 2016, 282; 10.5089/9781475531534.002.A004

Source: Staff estimates based on IMF/RES FSGM.Note: All variables are expressed in real terms.

29. Reducing collateral requirements would yield a medium term boost to GDP similar to improvements in bank efficiency. Effective average collateral-to-loan ratio of 157 percent as of 2010 is moderate by regional standards.23 However, reducing collateral requirements in the longer run could help stimulate growth. In scenario 4, permanently halving collateral requirements is assumed to bring about a 5 percent of GDP increase in domestic credit, slightly less than half of which would be financed by domestic excess liquidity and the remaining by an increase in foreign liabilities. Similarly to the effect of lower margins, this would increase domestic absorption while slightly depressing exports. The increase in private investment would also accumulate into a higher stock of private capital, which contributes to a persistent increase in both actual and potential output. As the latter adjusts more gradually, output rises above potential and a moderate positive output gap opens up. The overall medium term increase in GDP ranges from 0.3 percent in the first year to over 1 ¾ percent in the sixth.

References

  • Allen, Mark, Christoph Rosenberg, Christian Keller, Brad Setser and Nouriel Roubini, 2002, “A Balance Sheet Approach to Financial Crisis,IMF Working Paper 02/210 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Amo-Yartey, Charles, 2012, “Barbados: Sectoral Balance Sheet Mismatches and Macroeconomic Vulnerabilities,IMF Working Paper 12/31 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund. (2013) “Guatemala Selected Issues and Analytical Notes,Analytical Note IV, IMF Country Report No. 13/248.

    • Search Google Scholar
    • Export Citation
  • Mathisen, Johan, and Anthony Pellechio, 2006, “Using the Balance Sheet Approach in Surveillance: Framework, Data Sources, and Data Availability,IMF Working Paper 06/100 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

Annex I. Net Intersectoral Asset and Liability Positions

Table A1.

Guatemala. Net Intersectoral Asset and Liability Positions, September 2015

(in percent of GDP)

article image
Sources: Standardized Report Forms for Monetary and Financial Data, International Investment Position, and IMF staff estimates.
Table A2.

Guatemala: Gross Financial Assets and Liabilities of Economic Sectors

(Percent of GDP)

article image
Sources: National authorities; and Fund staff estimates.
1

Prepared by Iulia Ruxandra Teodoru.

2

Offshore banks need to be part of a financial group legally authorized in Guatemala and are subject to the same supervision requirements as domestic banks. The main difference with domestic banks is that they cannot accept small deposits (i.e. less than $10,000), deposits are not covered by the deposit insurance fund, and interest earnings on deposits are not subject to taxes. Also, information submitted to SIB on deposits is aggregated without revealing depositors’ or investors’ names.

3

The stress tests below are using a CAR of 13.9 percent (as of September 2015). The definition of CAR used in the stress tests differs from the regulatory definition of capital.

4

Also, elimination of the electricity and housing sectors’ exemption from higher capital requirements on foreign-currency loans has been announced; this may have also contributed to the credit growth slowdown.

5

Based on the data from the 1990-2015, 2 standard deviations would imply a higher economic growth (i.e. 1 percent) for the macroeconomic scenario, slightly milder compared to the scenario assumed in the stress test. This period is also characterized by lower exchange rate volatility given improvements in exchange rate management. While this scenario could be considered for the stress test, we chose a scenario which assumes a stronger shock which includes, for instance, a disorderly unwinding of UMP.

6

The assumption of a 20 percent exchange rate depreciation is illustrative Large historical depreciations of between 25-50 percent happened in the early and late 1990s.

7

The real interest rate was measured as the lending rate minus the CPI inflation rate calculated over the previous four quarters.

8

Prepared by Valentina Flamini.

9

Espinosa-Vega, M. and J. Sole, 2012, “Cross-Border Financial Surveillance: A Network Perspective,” IMF WP/10/105, Washington, DC.

10

Calculations assume 100 percent loss given default and funding shortfall, and 10 percent of lost funding being non-replaceable. Interbank exposure and capital data are as of December 31, 2015.

11

For methodological details see Cerutti, Eugenio, Stijn Claessens, and Patrick McGuire, 2012, “Systemic Risks in Global Banking: What can Available Data Tell Us and What More Data Are Needed?” BIS Working Paper 376, Bank for International Settlements. Banks exposures and spillover estimates were provided by Camelia Minoiu and Paola Ganum (RES).

12

Bank recapitalizations as well as other remedial policy actions (e.g., ring fencing, monetary policy, etc.) at the host and/or home country level are not taken into account in this model.

13

Panamanian banks have a more limited integration in the network analysis as they merely transmit the stress in international banks, rather than also being subjected to stress scenarios of losses in their asset values.

14

Based on consolidated claims on Guatemala of BIS reporting banks—excluding domestic deposits of subsidiaries of these banks in Guatemala.

15

Total credit to the non-bank sectors in Guatemala is calculated by adding IFS local (both domestic and foreign owned) banks’ claims on non-bank borrowers and BIS reporting banks’ direct cross-border claims on non-bank sectors (BIS Locational Banking Statistics Table 6B).

16

Spillovers from exposures to the USA increased significantly compared to the earlier (2013Q3) estimates of 0.25 percent of GDP because the latest simulations require advanced economy banking system to hold 8.5 percent capital ratio to be considered as “adequately capitalized” (in line with Basel III) compared to 6 percent in previous estimates. For any given shock to their balance sheets, this higher required minimum capital leads to a greater deleveraging by the banking system that receives the shock, and therefore to a higher funding risk exposure of the borrower country, in this case Guatemala.

17

Spillovers from exposures to large European banks are somewhat lower compared to the earlier (2013Q3) estimates of 1.9 percent of GDP because foreign claims decreased significantly, more than offsetting the deleveraging effect caused by the higher minimum capital requirement.

18

Prepared by Jaume Puig-Forné.

19

Full intersectoral data required for the BSA analysis are currently available up to September 2015. The BSA analysis in the 2013 Article IV report was based on data as of end-2012 (see IMF (2013)).

20

While the net debtor position of the financial system is not large relative to the size of the economy, it is large by international standards relative to the size of the banking sector, implying significant but manageable rollover risks (Section B, Domestic Bank Network Analysis and Cross-Border Financial Spillovers).

21

Includes domestic bank credit to corporates and external borrowing by corporates from international banks or capital markets.

22

Prepared by Valentina Flamini, Benjamin Hunt, and Keiko Honjo.

23

Regulatory limits require that credits may not exceed 70% of the value of the collateral or 80% of the value of mortgage guarantees. However, the effective collateral-to-loan ratio reported by the companies in the Enterprise Survey is higher since the ratio reflects the remaining maturity of the loans (the ratio increases as the loan gets closer to maturity and a portion of the loan is repaid).

Guatemala: Selected Issues and Analytical Notes
Author: International Monetary Fund. Western Hemisphere Dept.