Uruguay: Selected Issues

Abstract

Uruguay: Selected Issues

Stress-Testing The Banking Sector1

This paper looks at the stability of Uruguayan banking sector. After a brief description of the structure of the sector and its potential vulnerabilities, it presents the results of solvency and liquidity stress tests covering all public and private banks. These tests show banks have enough capital to withstand adverse credit, interest rate, sovereign, and exchange rate shocks. Banks’ structurally low profitability, however, constrains their ability to absorb more severe shocks.

A. Banking Sector Structure and Recent Developments

1. The Uruguayan banking sector is relatively small. The total assets of the 11 banks considered in this paper (2 public banks and 9 private, foreign-owned, banks) represent 80 percent of GDP. Credit is particularly low. Bank credit to the non-financial sector represents around 32 percent of GDP. This is well below the level observed in peer income-level countries and in other countries in the region. Credit to the non-financial sector also represents less than 41 percent of total banking assets. The average, however, masks significant variations across banks, with loans representing more than 88 percent of BHU’s assets, but only 10 percent of the assets of the Uruguayan affiliate of Banco de la Nación Argentina.

2. Dollarization is high, with 55 percent of the loans and 78 percent of deposits at end-June 2016 denominated in foreign currency. The dollarization of credit varies by borrowing sector. Nearly all loans to the mostly export-oriented agriculture sector and more than 90 percent of the loans to the manufacturing sector are denominated in dollar. On the contrary, less than 5 percent of the loans to households are in foreign currency (13 percent for private banks). About 30 percent of foreign-currency loans go to the non-tradable sector.

Figure 1.
Figure 1.

Uruguay: Dollarization of the Credit Market

Citation: IMF Staff Country Reports 2017, 029; 10.5089/9781475573626.002.A002

Sources: BCU; IMF staff calculations.
Figure 2.
Figure 2.

Uruguay: Structure of the Banking Sector

Citation: IMF Staff Country Reports 2017, 029; 10.5089/9781475573626.002.A002

Sources: BCU; IMF staff calculations.

3. The sector is segmented between public banks and private banks. On the one hand, public banks (BROU and BHU) dominate the peso market, with around 60 percent of peso deposits and loans. BROU itself receives 48 percent of all peso deposits and provides 40 percent of all peso loans. On the other hand, private banks hold more than 70 percent of the foreign-currency credit market and 60 percent of the deposits in foreign currency (mostly, if not exclusively, U.S. dollar).

4. Banking sector vulnerabilities have increased with the growth slowdown but remain limited. The spidergram (Figure 3) plots the ranking of the z-score of six variables capturing different dimensions of banking sector soundness: capitalization level (assessed through the capital adequacy ratio), asset quality (ratio of non-performing loans to total loans), profitability (return on assets), liquidity (ratio of liquid assets to short-term liabilities), FX risk (share of foreign currency loans to the non-tradable sector), and credit growth prospects (bank credit growth). The further the value of each variable in a given period is from its historical mean, the higher the z-score and the numerical ranking assigned to it (see Cervantes and others, 2014). Based on 2016Q2 data, the chart points to increased vulnerabilities in terms of asset quality and profitability. The NPL ratio has been historically very low in Uruguay and has doubled over the past two years. This explains the ranking of 10 associated with that variable, even if the current level does not, at this stage, seem a cause for significant concern, and provisions for loan losses remain high. Similarly, the drop in the return on assets over the last year has pushed the numerical ranking for that variable to 7, a value typically indicative of medium-high risk.

Figure 3.
Figure 3.

Uruguay: Banking Sector Vulnerabilities

Citation: IMF Staff Country Reports 2017, 029; 10.5089/9781475573626.002.A002

Note: Aggregates exclude BHU. Sources: BCU; IMF staff calculations.

5. While the sector’s capital adequacy ratio has stabilized after several years of decline, banks’ structurally low profitability constrains their future loss-absorption capacity. Because of slowing loan growth, the system-wide CAR (14.7 percent at end-June 2016, including BHU) has slightly increased over the past year, ending several years of decline. Public banks continue to exhibit higher capitalization levels than their private counterparts. At the same time, the still low level of dollar interest rates combined with the slowdown in lending weighs on the profitability of banks’ foreign exchange activity, given that banks keep their excess dollar assets in the form of deposits or invest them in sovereign (domestic or foreign) securities. The segmentation of the sector and the lack of competition on the peso deposit and credit markets—reflected by the large market share of the two public banks—are also sources of inefficiencies.

B. Solvency Stress Tests

6. Banks can withstand a range of adverse credit, interest rate, sovereign risk and exchange rate shocks. The stress tests consider the impact of different types of shocks on capital and risk-weighted assets in a static way. In particular, they do not include any feedback effect of the shocks on credit growth. The tests were performed on individual bank data at end-June 2016.

  • Credit shocks include (i) an aggregate NPL shock equal to 5 percent of currently performing loans (which corresponds to a tripling of the current aggregate NPL ratio, whereas the current slowdown has so far resulted in a doubling of the NPL ratio), and (ii) sectoral increases of NPLs equal to respectively 7, 12, and 10 percent of currently performing loans in the more export-oriented agricultural, manufacturing and tourism sectors (a fourfold increase from current sectoral NPL ratios).2

  • The interest rate shock considers the effect of a 208 bp increase in both dollar and peso interest rates across all maturities (parallel shift of the yield curves). This increase corresponds to the maximum rise in the federal funds target rate observed over a twelve-month period since 2000, and lies between the increases considered by the Superintendent for Financial Services (SSF) in the adverse and crisis scenarios of its May 2016 stress tests (+156bp and +275bp respectively). The shock affects banks’ net interest income due to the maturity gap between assets and liabilities,3 as well as the value of fixed-income securities on banks’ balance sheets (repricing impact).

  • The sovereign risk shock scenario shows the repricing effect of a 600bp increase in the risk premium on Uruguayan securities, corresponding to two standard-deviations of the monthly average EMBI spread over the period June 2001–June 2016. In its May 2016 stress tests, the SSF considered country risk shocks of +483bp and +1000bp in its adverse and crisis scenarios. A sovereign risk shock affects the value of both peso and dollar-denominated sovereign securities.

  • The exchange rate shock assumes a 30 percent depreciation of the peso (equivalent to the depreciation of the peso observed in the last depreciation period, between end-2015 and end-March 2016). This is a bit less than the 35 percent depreciation assumed by the SSF in its crisis scenario (the adverse scenario assumed a 23 percent depreciation). The shock has both a direct effect on the value of banks’ assets denominated in U.S. dollar (with a depreciation actually increasing the peso value of those assets, thereby depressing the capital-adequacy ratio) and an indirect effect on asset quality with 20 percent of the FX-denominated loans assumed to become non-performing following a 100 percent depreciation.4 The latter assumption is based on estimates from 2011 indicating that roughly 17 percent of corporate borrowers have a significant net open short-term FX position (see the companion paper on balance sheet analysis).

7. The low level of credit limits the effect of credit shocks. A proportional increase in NPLs across all sectors implies a reduction in the system-wide capital-adequacy ratio (CAR) by 0.7 percent. A more severe shock affecting export-oriented sectors results in an even smaller deterioration of 0.4 percent. Reverse stress testing indicates that NPLs would have to rise by 19 to 22 percent of currently performing loans for the CAR of either a third of the banks or banks representing a third of total assets to fall below the regulatory minimum of 8 percent—an extreme event.

8. Maturity mismatches imply a significant effect of a rise in interest rates on net interest income. The impact on net interest income is sizable, reducing the CAR of one bank below the regulatory threshold.5 The repricing effect is also significant, although the impact on the CAR is mitigated by a decrease in risk-weighted assets as FX-denominated bonds make up 92 percent of banks’ bond portfolio, or 8 percent of total assets (peso-denominated sovereign bonds are zero risk-weighted). Taking both effects into account, two banks, representing close to 10 percent the system’s assets and 7 percent of its capital, would register CARs below the regulatory minimum.

9. A sovereign risk premium shock would mostly affect public banks as private banks’ holdings of public debt securities is on average small. The aggregate CAR would drop by 1.6 percent, but by 3.5 percent for public banks only.

10. Given banks’ aggregate long position in foreign exchange, a depreciation of the peso potentially creates large direct gains. Those, however, do not translate into higher CARs as the risk-weighted assets increase more than capital. In addition, a large depreciation implies an increase in NPLs to non-hedged borrowers. Overall the system-wide CAR decreases moderately, by 1.4 percent, with no bank falling under 8 percent.

11. A combined credit risk, interest rate, sovereign risk and FX risk shock could have larger effects, that could be cushioned by countercyclical buffers.6 The system-wide CAR would remain close to 9 percent with however some large variations across banks. Five banks, representing half of the capital of the whole system, would register CARs (not including countercyclical buffers) below the 8 percent regulatory minimum. Countercyclical buffers could be used to support CARs but would be insufficient to lift all of them above 8 percent. Yet, the combined shock would not trigger any bank failure even if balance sheet adjustments would be required.

12. Contagion effects are limited by the small interbank exposures. The interbank exposure analysis in Figure 4 uses interbank exposure data on both the asset and liability sides to determine potential second round effects of the previously described macro-shocks.7 The results show very low contagion and vulnerability levels, as banks mostly fund themselves through deposits and interbank exposures represent less than one percent of total banking assets at end-June 2016.

Figure 4.
Figure 4.

Uruguay: Solvency Stress Test Results

Citation: IMF Staff Country Reports 2017, 029; 10.5089/9781475573626.002.A002

Sources: BCU; IMF staff calculations.
Figure 5.
Figure 5.

Uruguay: Interbank Exposure Analysis

Citation: IMF Staff Country Reports 2017, 029; 10.5089/9781475573626.002.A002

Sources: BCU; IMF staff calculations.

C. Liquidity Stress Test

13. Deposits represent more than 85 percent of banks’ liabilities. Two-thirds of those deposits are in foreign currency and redeemable at demand. Public banks have a higher share of peso deposits in their total liabilities than private banks (24 percent versus 14 percent), in particular because BHU hardly holds any foreign currency deposit. Given banks’ reliance on deposit funding and low level of credit, the aggregate FX loans-to-FX deposit ratio is relatively low, around one-third. However, the peso loans-to peso deposits ratio is close to 100 percent. Bank liquidity is further supported by high reserve requirements (increased to 28 percent of both peso and dollar deposits on April 1st). Liquid assets cover around 35 percent of demand deposits, with an even higher ratio for deposits in foreign currency.

14. As a result, the system could withstand a severe liquidity drain both in peso and foreign currency for at least five days. The liquidity stress test considers a simple liquidity drain on all banks, with 10 and 5 percent of the demand deposits in respectively pesos and foreign currency, and 2 percent of the time deposits in either currency, withdrawn every day for five consecutive days. Assuming that 95 percent of the liquid assets and 5 percent of the other assets are available in a day, no bank is found illiquid after five days.

Figure 6.
Figure 6.

Uruguay: Bank Funding and Liquidity

Citation: IMF Staff Country Reports 2017, 029; 10.5089/9781475573626.002.A002

Sources: BCU; IMF staff calculations.

D. Conclusions

15. Banks in Uruguay appear sufficiently solid to withstand adverse solvency shocks. This paper considered simple stress tests, looking at the static effects on banks’ CAR of credit, interest rate, sovereign, and exchange rate shocks, separately and in a combined way. While several banks would see their capital-adequacy ratio fall under the regulatory minimum in case of combined shocks, requiring balance sheet adjustments, the system-wide CAR would remain above 9 percent after including countercyclical capital buffers. Contagion effects would be limited by the very small interbank exposures. The implementation of Basel III requirements in 2017 will increase the capital of the five systematically important banks, thereby enhancing their resilience to shocks further.

16. The liquidity of the system is high. The funding structure of banks and their low level of credit result in a high level of liquidity, supported by high reserve requirements. Banks’ liquidity is higher in dollar than in peso, consistent with the higher dollarization of deposits than loans.

17. The declining profitability of banks constrains their future loss absorption capacity. The low profitability likely reflects high labor costs, a lack of efficiency in some banks, and a conservative business model. As the 2014 financial inclusion law expands the potential client base for private banks, these could be encouraged to compete more aggressively in the peso lending market, without jeopardizing their capitalization levels.

Reference

Cervantes, Ricardo, Phakawa Jeasakul, Joseph Maloney and Li Lian Ong, 2014, Ms. Muffet, the Spider(gram) and the Web of Macro-Financial Linkages, IMF Working Paper, WP/14/99.

  • Search Google Scholar
  • Export Citation
1

Prepared by Frederic Lambert.

2

A 25 percent provisioning rate on new NPLs is assumed. The effect of the shock on risk-weighted assets is assumed to be of the same magnitude as the effect on capital.

3

The test considers the cumulative gap at a 12-month horizon. It assumes the increase in interest rate is passed through symmetrically to deposit and lending rates.

4

This means that 6 percent of FX loans currently performing would become non-performing following the assumed 30 percent depreciation shock. The provisioning rate on new NPLs is set at 25 percent.

5

This effect is mechanic given the maturity gap between assets and liabilities and the assumed symmetric pass-through to deposit and lending rates in the model. In reality, banks expect a positive effect on their income from a rise in interest rates as the pass-through to deposit rates is projected to be much smaller than the increase in lending rates.

6

The combined shock scenario combines all the shocks that were previously examined separately. While somehow extreme, it may represent a more realistic crisis scenario, in which a large exchange rate depreciation accompanied by an increase in global interest rates would force an increase in domestic interest rates that would negatively affect economic activity and trigger an increase in domestic credit risk.

7

The exercise assumes a loss given default of 100 percent following a credit shock and that 35 percent of the funding lost as a result of a bank failure is not replaceable, with a funding-shortfall induced loss factor of 1.

Uruguay: Selected Issues
Author: International Monetary Fund. Western Hemisphere Dept.