A. Introduction

1. Uruguay’s financial system has changed profoundly since the crisis in 2002. Several of these changes stand out. First, the level of financial intermediation declined significantly as a result of the crisis, and is low by regional and historical standards: the private-sector-credit-to-GDP ratio fell from 29 percent in 1998 (closely matching the LA5 average of 30 percent) to 19 percent in 2010 (below the LA5 average of 33 percent).^2,3 Although credit to households has increased in recent years (the ratio of household credit to private consumption nearly doubled, rising from 9.8 in 2006 to 18.2 percent in 2010), the ratio of corporate credit to gross fixed capital formation declined from 86 to 83 percent. Second, the system is very robust: banks are liquid, well capitalized, and have a low share of non-performing loans (NPLs), although deposit and credit dollarization remain relatively high, at 74 and 68 percent of total deposits and credit, respectively. Third, Uruguay has a peculiar market structure in which a state-bank holds roughly half of the market and subsidiaries of foreign banks hold the other half.

2. To help promote the greater use of financial services, the government has formulated a broad strategy on “bancarización e inclusión financiera”. A major focus is on improving access to finance for low-income families and small companies, raise financial awareness, promote savings and a greater use of financial services, and strengthen consumer protection. Greater financial literacy is expected to enhance and widen the access of the population to financial services while at the same time promote their responsible use and healthy growth.

3. This paper looks at two aspects of Uruguay’s banking system that are also relevant to the issue of financial intermediation: market structure and profitability.

• Market structure. Available studies (e.g., Claessens and Laeven, 2004) suggest that competition in the banking sector is associated with greater financial intermediation and economic growth. In Uruguay, the 2002 crisis led to an increase in the concentration in the banking sector that has remained throughout the last decade. This paper finds that the financial system in Uruguay is more concentrated and has a somewhat lower degree of competition than those of peer countries. The findings on market competition in Uruguay are in line with those of Gelos and Piñón (2008).

• Profitability. The profitability of Uruguayan banks is slightly below the regional average according to the official data, and there are big differences across banks. This paper looks at three aspects and their influence on profits. First, the role of the market structure—and it finds that a bank’s market share matters for its profitability. Second, the differences in accounting between Uruguay and other countries-and it finds that the inflation adjustment used in Uruguay affects reported profitability. Third, the provisioning system in Uruguay—and it finds that Uruguay’s pioneering (in Latin America) dynamic provisioning framework might have led to over provisioning in some banks, which might have negative effects on profits.

B. Market Structure

4. Uruguay’s banking sector became more concentrated in the aftermath of the 2002 crisis and has remained concentrated since then. There were 17 banks back in 2003 and, after a modest number of entries and acquisitions, 13 banks operated by mid-2011. Although the sector’s concentration, measured by the Herfindahl-Hirschman index (HHI), decreased somewhat from 2,600 in 2003 to 2,403 in mid-2011 (see Figure 1), it remains high by the standards of the Horizontal Merger Guideline (2010).^{4, 5} Banking concentration also remains above the regional average of 1,500 (Chortareas et al., 2010) and exceeds its pre-crisis level of 1,226 in 2000. The increase in concentration during the 2002 crisis was significant (1,300 points) and had the potential to adversely affect market competition.⁶

Uruguay: Herfindahl-Hirschman Index

Using total banking sector assets, excluding BHU

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

Sources: Banco Central del Uruguay and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Uruguay: Herfindahl-Hirschman Index

Using total banking sector assets, excluding BHU

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

Sources: Banco Central del Uruguay and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

Uruguay: Herfindahl-Hirschman Index

Using total banking sector assets, excluding BHU

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

Sources: Banco Central del Uruguay and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

5. The market structure has changed only moderately since 2003, despite several mergers. The largest bank—Banco de la República Oriental del Uruguay (BROU), a state-owned bank—has maintained its market share of almost half the total banking assets. Banco Santander S. A. became the second largest bank in 2008 when it acquired ABN Amro Bank N.V, and it now has close to one-fifth of total banking assets. The rest of the market has remained fragmented, even though some mergers have taken place, with only three banks having a market share over five percent (BBVA, Itau, and, Nuevo Banco Comercial).⁷

6. The number and total size of non-banking financial institutions is not negligible. The non-banking sector consists of regulated and non-regulated lenders. Non-banks are regulated as long as they operate with borrowed money or are credit card providers; these are called administradoras de crédito. According to the report by AEBU (2011), there are 13 regulated non-banks, with total loans equal to eight percent of total banking sector’s loans. This segment has been recently very dynamic and concentrates on consumer loans. The non-regulated part represents private money lenders. Several non-banks—particularly the largest—have been acquired by banks (e.g., OCA, Pronto, and Creditel) and sell all or part of their loan portfolio to them.⁸

Uruguay: Market Shares

Percent of total banking sector assets, excluding BHU

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Uruguay: Market Shares

Percent of total banking sector assets, excluding BHU

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

• Download Figure
• Download figure as PowerPoint slide

Uruguay: Market Shares

Percent of total banking sector assets, excluding BHU

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

• Download Figure
• Download figure as PowerPoint slide

7. A measure of market structure suggests the presence of monopolistic competition in the banking market. A standard yardstick for the degree of market competition is the H-statistic, which follows from the Panzar-Rosse (1987) methodology (Box 1) and the empirical specification by Claessens and Laeven (2004). It is obtained by estimating the following reduced revenue equation:

where r is log of total revenues (from interest and services), cof is log of cost of funds (interest expenses over total borrowing), col is log of wage (wages over total assets), copc is log of cost of physical capital (other operating expenses over total assets), lta is the share of loans in total assets, capta is the ratio of equity capital over total assets, and ta is log of total assets. Parameters α, β, δ, θ, ρ, and λ are to be estimated.

Uruguay: Loans to Total Assets Ratio

(In percent)

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

Sources: Banco Central del Uruguay and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Uruguay: Loans to Total Assets Ratio

(In percent)

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

Sources: Banco Central del Uruguay and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

Uruguay: Loans to Total Assets Ratio

(In percent)

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

Sources: Banco Central del Uruguay and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

Table 1.

Uruguay: Banks’ Profitability

(In percent)

Source: BCU; First half of 2011.

Table 1.

Uruguay: Banks’ Profitability

(In percent)

	Return on assets (ROA)	Return on equity (ROE)
Public Bank
B.R.O.U	1.4	12.5
Private Banks
N.B. comercial	-0.1	-1.0
Banco Itau	-0.4	-7.1
Banco Santander	0.7	7.3
Credit Uruguay	0.8	11.3
B.B.V.A.	-1.2	-12.5
Citibank	1.0	14.3
Discount Bank	-0.2	-2.1
Lloyds TBS	-0.6	-7.3
HSBC Bank	-2.1	-37.7
Bandes Uruguay	-14.5	-83.6
Banco Surinvest	-2.0	-14.5
Bco. Nacion Argentina	-1.7	-9.6

Source: BCU; First half of 2011.

View Table

Table 1.

Uruguay: Banks’ Profitability

(In percent)

	Return on assets (ROA)	Return on equity (ROE)
Public Bank
B.R.O.U	1.4	12.5
Private Banks
N.B. comercial	-0.1	-1.0
Banco Itau	-0.4	-7.1
Banco Santander	0.7	7.3
Credit Uruguay	0.8	11.3
B.B.V.A.	-1.2	-12.5
Citibank	1.0	14.3
Discount Bank	-0.2	-2.1
Lloyds TBS	-0.6	-7.3
HSBC Bank	-2.1	-37.7
Bandes Uruguay	-14.5	-83.6
Banco Surinvest	-2.0	-14.5
Bco. Nacion Argentina	-1.7	-9.6

Source: BCU; First half of 2011.

8. The H-statistic is computed as the sum of β, γ and δ and conceptually measures the response of revenues to changes in input prices. The results, shown in Table 2, represent estimates for the period 2003-10. The H-statistics of 0.66 suggests monopolistic competition. The insignificant E-statistics confirm the validity of the Panzar-Rosse model⁹. The size of the H-statistic estimated here is broadly in line with previous studies on market competition in Uruguay. For example, Gelos and Piñón (2008) found a decline of the H-statistic from 0.75 to 0.5 between 2003 and 2006.¹⁰

9. Excluding the largest bank (BROU) and other possible outliers from the panel regressions does not change the results. Several banks could be considered outliers due to the fact that they have different business models. BROU is a state bank and might thus exhibit different behavior from that of private banks. There are also three small private banks—Banco Surinvest, Banco de la Nacion Argentina, and Discount bank—that have unique business models, in the sense that they are significantly less focused on loans. Excluding these banks does not change the results of the degree of competition (Table 2).

Table 2.

Uruguay: Market Structure

Note: Yearly data spans over 2003-10. Stars denote significance level as follows: ^*** at 1, ^** at 5, and ^* at 10 percent level.Data source: Banco Central del Uruguay.

Table 2.

Uruguay: Market Structure

	In total revenues / total assets)
	full sample	excl. BROU	excl. BROU, Bco. Surinvest, Bco. Nacion Argentina, and Discount bank
const	0.17	0.31	-0.33
	(1.01)	(1.02)	(1.01)
In(wages / total assets)	0.31***	0.3***	0.31***
	(0.11)	(0.11)	(0.11)
In(interest expenses / total borrowings)	0.16***	0.16***	0.18***
	(0.04)	(0.04)	(0.05)
In(other operating expenses / total assets)	0.19**	0.19**	0.16*
	(0.1)	(0.1)	(0.1)
loans / total assets	-0.07	-0.05	0.55*
	(0.22)	(0.23)	(0.33)
equity capital / total assets	-0.36	-0.45	-0.65
	(0.57)	(0.59)	(0.92)
In(total assets)	-0.01	-0.02	-0.01
	(0.07)	(0.07)	(0.06)
R2	0.43	0.41	0.43
Nobs / banks	107 / 17	99 / 16	75 / 13
H-statistic	0.66***	0.65***	0.65***
	(0.13)	(0.13)	(0.12)
E-statistic	-0.03	-0.025	-0.007
	(0.02)	(0.02)	(0.02)

Note: Yearly data spans over 2003-10. Stars denote significance level as follows: ^*** at 1, ^** at 5, and ^* at 10 percent level.Data source: Banco Central del Uruguay.

View Table

Table 2.

Uruguay: Market Structure

	In total revenues / total assets)
	full sample	excl. BROU	excl. BROU, Bco. Surinvest, Bco. Nacion Argentina, and Discount bank
const	0.17	0.31	-0.33
	(1.01)	(1.02)	(1.01)
In(wages / total assets)	0.31***	0.3***	0.31***
	(0.11)	(0.11)	(0.11)
In(interest expenses / total borrowings)	0.16***	0.16***	0.18***
	(0.04)	(0.04)	(0.05)
In(other operating expenses / total assets)	0.19**	0.19**	0.16*
	(0.1)	(0.1)	(0.1)
loans / total assets	-0.07	-0.05	0.55*
	(0.22)	(0.23)	(0.33)
equity capital / total assets	-0.36	-0.45	-0.65
	(0.57)	(0.59)	(0.92)
In(total assets)	-0.01	-0.02	-0.01
	(0.07)	(0.07)	(0.06)
R2	0.43	0.41	0.43
Nobs / banks	107 / 17	99 / 16	75 / 13
H-statistic	0.66***	0.65***	0.65***
	(0.13)	(0.13)	(0.12)
E-statistic	-0.03	-0.025	-0.007
	(0.02)	(0.02)	(0.02)

Note: Yearly data spans over 2003-10. Stars denote significance level as follows: ^*** at 1, ^** at 5, and ^* at 10 percent level.Data source: Banco Central del Uruguay.

10. Banking competition in Uruguay appears to be somewhat lower than the regional average. According to the estimates by Anzoategui et al. (2010), using exactly the same specification of variables as above, the average H-statistic for Latin America equals 0.77 (during 2002-08).¹¹ Thus, the Uruguayan banking market, with H-statistic of 0.66, exhibits a competition level that is somewhat below the regional average, although the difference is not statistically significant.

C. Profitability

11. Bank profitability is slightly below the regional average. The average of the reported banking sector’s ROA is 1.4 percent (2006-2010 average, weighted by total assets). However, as discussed in the next section, profitability is not directly comparable across countries due to differences in accounting standards. Uruguay applies a less frequently used version of the International Financial Reporting Standards (IFRS), see Section B. Once accounting adjustments are made to Uruguay’s ROA so that it can be compared internationally, this would increase to about 1.7 percent, which is still below the average (2.4 percent) of the regional distribution of ROAs.¹²

12. There are sizable differences in profitability across banks. Over the last five years the standard deviation of the distribution of ROA across banks was 3.8 percentage points (using official data). The highest ROA attained was 4.1 percent and the lowest -20.9 percent (Table 1 presents a cross-section of bank profitability in the first half of 2011). While some banks consistently report positive profits—usually the largest banks—several small banks (and some of them persistently) operate with losses.

Profits and market structure

13. This section analyzes the effects of the market structure on banking sector’sprofitability.¹³ In particular, it tests the two hypotheses, described in Box 2, for the ROA’s positive relationship with market structure and a cost efficient structure.

Box 2.

Profitability and Market Structure

Findings of a positive statistical relationship between profitability and market share could be interpreted in two different (although not mutually exclusive) ways (Berger, 1995). Extra profits of larger banks could stem either from using their market position vis a vis consumers or from their higher efficiency.

The relative-market power hypothesis (RMP) asserts that only firms with large market shares and well-differentiated products are able to exercise market power and earn super normal profits. A greater market power due to increasing market shares might lead to a price setting that is less favorable to consumers (lower deposit rates and higher loan rates) in more concentrated markets.

In addition to the market-power theory, there are efficiency explanations of the positive relationship between profits and market shares. Under the efficient-structure hypothesis (ES), firms with superior management or production technologies have lower costs and therefore higher market share and profits.

14. There is a statistically significant profit premium for market share in Uruguay, after controlling for differences in cost efficiency. A panel regression of the log of (1+ROA) on the log of market shares, and controlling for the effect of NPLs and the efficiency ratio (ratio of operating costs to revenues), reveals a robust relationship between the return and market share.¹⁴ As shown in Table 3, the profitability increased with the size of the market share with an elasticity of 0.02, which lends support to the RMP hypothesis.¹⁵ This finding is consistent with the measures of market concentration and competition (HHI and H-statistic), which suggest relatively higher concentration and lower competition in Uruguay vis-à-vis peer countries.

15. Cost efficiency and loan book quality play important roles as determinants of bank profitability. The estimates in Table 3 show that a higher ratio of operating expenses over revenues (a proxy for cost efficiency) lowers profitability—thereby providing support for the efficient-structure hypothesis (ES). This is consistent with the findings of a previous analysis of banking profitability by Wezel (2011) and findings of economies of scale by Mello Costa (2009). The share of NPLs in total loans or net provisioning costs over total assets (a proxy for the loan book quality) is also an important factor for profitability (see also Section C). Excluding outliers (BROU, Banco Surinvest, Banco de la Nacion Argentina, and Discount bank) does not change the results for the RMP and ES hypotheses.

16. The higher profitability of the largest bank (BROU) appears to be based on a high loan book quality, high collections, and a large market share. BROU leads the market in profitability by a difference of 1.5 percentage points of reported ROA on average over 2003-10. The bank has two specifics compared with other banks in Uruguay: it is charged with channeling salaries to public employees and it has preferred creditor status.¹⁶

17. The estimation also suggests that BROU may have higher fixed costs than other banks. In particular, the fixed effect for BROU in the regressions in Table 3 is lower (by 0.03) than the average and the difference is statistically significant at the five percent confidence level. This suggests that there might be high fixed costs for BROU that could stem from its unique operations (e.g., from administrating disbursement of public sector salaries, from operating a large branch network, etc.). At the same time, the large market share, the automatic deduction of installments from debtor’s salaries, and the preferred creditor status, generate benefits that seem to outweigh these costs. In the preparation of the government’s bancarización project, it is envisaged that the right to automatically deduct installments from debtor’s salaries will be granted to all commercial banks, hereby making the playing field more equal.

Table 3.

Uruguay: Market Power and Efficient Structure

Note: Yearly data spans over 2003-2010. Stars denote significance level as follows: ^*** at 1, ^** at 5, and ^* at 10 percent level. Data source: Banco Central del Uruguay.

Table 3.

Uruguay: Market Power and Efficient Structure

	ln(1+ROA)
	full sample	excl. BROU	excl. BROU, Bco. Surinvest, Bco. Nacion Argentina, and Discount bank
const	0.08**	0.08**	0.12***
	(0.03)	(0.04)	(0.03)
ln(mkts)	0.02**	0.02**	0.02**
	(0.01)	(0.01)	(0.01)
ln(nonperforming loans / total loans)	-0.003**	-0.003**
	(0.002)	(0.002)
operating costs / revenues	-0.05***	-0.056***	-0.07***
	(0.01)	(0.02)	(0.02)
net provisioning / total assets			-0.67***
			(0.25)
R2	0.41	0.43	0.62
Nobs	100 / 17	92 / 16	75 / 13

Note: Yearly data spans over 2003-2010. Stars denote significance level as follows: ^*** at 1, ^** at 5, and ^* at 10 percent level. Data source: Banco Central del Uruguay.

View Table

Table 3.

Uruguay: Market Power and Efficient Structure

	ln(1+ROA)
	full sample	excl. BROU	excl. BROU, Bco. Surinvest, Bco. Nacion Argentina, and Discount bank
const	0.08**	0.08**	0.12***
	(0.03)	(0.04)	(0.03)
ln(mkts)	0.02**	0.02**	0.02**
	(0.01)	(0.01)	(0.01)
ln(nonperforming loans / total loans)	-0.003**	-0.003**
	(0.002)	(0.002)
operating costs / revenues	-0.05***	-0.056***	-0.07***
	(0.01)	(0.02)	(0.02)
net provisioning / total assets			-0.67***
			(0.25)
R2	0.41	0.43	0.62
Nobs	100 / 17	92 / 16	75 / 13

Note: Yearly data spans over 2003-2010. Stars denote significance level as follows: ^*** at 1, ^** at 5, and ^* at 10 percent level. Data source: Banco Central del Uruguay.

Profit and Loss (P&L) accounting specifics

18. Officially recorded bank profits in Uruguay are affected by a particular accounting treatment for inflation. The IFRS allows for the use of inflation adjustments in high-inflation environments.¹⁷ Uruguay allows for such an adjustment, whereas most low inflation countries use a version of the IFRS based on nominal accounting. To make the Uruguayan P&L comparable with the more commonly used IFRS, inflation adjustments need to be removed from the reported P&L in Uruguay.

19. The inflation adjustments can overstate expenses, although their effect has not been consistent over time. For example, in some years, the adjustment did not have an impact (as in 2007-08) or even overestimated profits (2005). In recent years, however, the adjustment led to an overstatement of expenses and an understatement of profits. Foreign bank subsidiaries report P&L to their parent banks without inflation adjustments. For instance, banks Santander and Itau reported ROA of 1.99 and 0.95 percent in 2010 to their respective groups under the most commonly used, non-hyperinflation IFRS, while under Uruguay’s official accounting their reported ROA is lower by 0.44 and 0.31 percentage points.¹⁸ The difference is entirely due to inflation adjustments.

20. Thus, in recent years, internationally comparable profitability has been somewhat higher than what is reported under local accounting standards. Uruguayan banking P&L accounting standards prescribe to adjust assets and liabilities for inflation if annual inflation exceeds 12 percent or if accumulated inflation over three years exceeds 2.5 percent.¹⁹ In 2009, the overall sector’s ROA would have been 0.9 percentage points higher than reported without the inflation adjustment. Similarly, in 2010, the ROA would have been 1.6 instead of the reported 1.1 percent.

Uruguay: Inflation Adjustment

Net expenses on inflation as percent of total assets

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

Sources: Banco Central del Uruguay and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Uruguay: Inflation Adjustment

Net expenses on inflation as percent of total assets

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

Sources: Banco Central del Uruguay and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

Uruguay: Inflation Adjustment

Net expenses on inflation as percent of total assets

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

Sources: Banco Central del Uruguay and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

Uruguay: Return on Assets

Adjustments to align ROA with usual IFRS

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Uruguay: Return on Assets

Adjustments to align ROA with usual IFRS

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

• Download Figure
• Download figure as PowerPoint slide

Uruguay: Return on Assets

Adjustments to align ROA with usual IFRS

Citation: IMF Staff Country Reports 2011, 376; 10.5089/9781463926601.002.A004

• Download Figure
• Download figure as PowerPoint slide

Provisioning

21. Uruguayan banks have accumulated an ample cushion of loan loss provisions, presently amounted to about six times non-performing loans (NPL).²⁰ There are two main reasons for such high coverage: First, NPLs have been on a long-term downward trend after the 2002 financial crisis, not least thanks to a stronger economy leading to better loan portfolio quality and considerable improvements in bank regulation and supervision. Second, dynamic provisioning (DP), which Uruguay pioneered in Latin America a decade ago, has yielded a large reserve buffer, which turned out to be much higher than the loan defaults during the economic slowdown of 2008-09. More specifically, the accumulation of dynamic reserves until late 2008, an equivalent of US$100 million (2.6 percent of total loans), greatly exceeded the maximum drawdown of such reserves recorded by February 2010 (US$12 million or 0.4 percent of total loans).

Box 3.

Dynamic Provisioning in Uruguay

Dynamic provisioning (DP) requires banks to build reserves for anticipated, but not yet realized, loan losses. Based on the loss experience of the past credit cycle and the pioneering Spanish model, Uruguayan banks in 2001 began constituting general provisions for current loan portfolios as well as new loans. During the upswing, which lasted until 2007, banks continuously put aside the difference between the average monthly specific provisions, as recorded during the previous cycle, and the lower actual provisions of the curent period. Subsequently, in the downturn of 2008-09, banks tapped these general loan loss reserves by using that difference to cover the cost of rising loan delinquencies.

The system of specific and dynamic provisions will be modified in mid-2012 in response to the average loan defaults during 1999-2007, which notwithstanding the 2002-03 crisis were lower than in the former reference period of the 1990s. In particular, the dynamic provisions will no longer be applied to non-performing loans—which are already covered by specific provisions—and the contribution rates will be set in line with updated default expectations for high-quality loans in the upper three classification categories (including “special mention” loans). In addition, dynamic provisions will also be made on the increment in loan volumes in these categories. Specifically, the modified regulation requires banks to contribute to their individual dynamic provisioning funds, DP_t, the difference between the monthly statistical net losses on higher-quality loans to the non-financial private sector and the realized net loan loss in that month:

$\mathrm{\Delta }{\mathit{DP}}_t=[\underset{i=1}{\overset{n}{\mathrm{\Sigma }}}\frac{1}{12}{\alpha }_i\mathrm{\Delta }{c}_{i,t}+\frac{1}{12}{\beta }_i{c}_{i,t}]-\underset{i=1}{\overset{m}{\mathrm{\Sigma }}}(\mathrm{\Delta }{E}_{i,t}-{\mathit{R}}_t)$

These statistical losses are derived by multiplying 1/12 of the expected annual rates of loss (β_i and α_i, with α_i>β_i) of three upper loan categories (n) by the respective loan volumes, C_i,t and their increment ΔC_i,t. The net loan loss is calculated as the total cost of additional specific provisions ΔE_i,t in all loan categories (m) net of recoveries of written-off loans, R_t Lastly, the current maximum limit to banks’ DP funds of 3 percent of loans will be replaced by a bank-specific limit that is determined by the product of the share of loans in each of the three categories and their respective expected delinquency rates. The BCU has estimated that, based on May 2011 data, the average fund size will drop to about 2.1 percent of total loans (from an actual 2.3 percent in that month).

This forward-looking approach of loss recognition has the important merit of smoothing provisioning costs over the cycle but it also has a flipside. During the 2008-09 downturn, banks’ income statements were largely shielded from the (moderate) deterioration in their loan exposures and thus were in a good position to keep granting credit and support the economy. Yet, the correct calibration of the DP parameters is not easy, as it by necessity relies on past information, and any future change in expected loss on account of economic and regulatory factors is likely to lead to suboptimal provisioning.

22. In view of the muted loan losses during the downturn, it might be argued that the system as a whole, and some banks in particular, have accumulated excess loan loss reserves. Considering that provisions should cover expected losses (and capital unexpected losses), Uruguayan banks could currently sustain a multiple of the loan losses incurred during the relatively mild dowturn of 2008-09. Typically, the expected loss (EL) is defined as the exposure at default (EAD) multiplied by the probability of default (PD) over one year and the loss given default (LGD). As the PD and LGD were not available to the authors, the EL is proxied by the net provisioning flow (NPF) during a 12-month period. Specifically:

$\mathit{N}\mathit{e}\mathit{t}\mathit{\text{provisioning}}\mathit{\text{flow}}\mathit{\text{=}}\mathit{\Delta }\mathit{\text{stock of}}\mathit{\text{specific provisions}}\mathit{\text{+ loan}}\mathit{\text{writeoffs}}\mathit{\text{-}}\mathit{\text{loan}}\mathit{\text{recoveries.}}$

The NPF is then set in relation to the stock of dynamic provisions (DP/NPF), see Table 4.

Table 4.

Uruguay: Provisions Coverage of Expected Loss

(In percent)

Table 4.

Uruguay: Provisions Coverage of Expected Loss

(In percent)

Bank		Downturn DP/Downturn NPF	Current DP/Downturn NPF	Total Current Provisions/ Downturn NPF
	1	103	110	370
	2	n/a	n/a	n/a
	3	n/a	n/a	n/a
	4	1726	1625	2733
	5	610	554	1062
	6	99	364	633
	7	1953	1796	3170
	8	60	227	387
	9	334	795	5458
	10	n/a	n/a	n/a
	11	n/a	n/a	n/a
	12	n/a	n/a	n/a
	13	488	326	655
	Average	672	725	1809

View Table

Table 4.

Uruguay: Provisions Coverage of Expected Loss

(In percent)

Bank		Downturn DP/Downturn NPF	Current DP/Downturn NPF	Total Current Provisions/ Downturn NPF
	1	103	110	370
	2	n/a	n/a	n/a
	3	n/a	n/a	n/a
	4	1726	1625	2733
	5	610	554	1062
	6	99	364	633
	7	1953	1796	3170
	8	60	227	387
	9	334	795	5458
	10	n/a	n/a	n/a
	11	n/a	n/a	n/a
	12	n/a	n/a	n/a
	13	488	326	655
	Average	672	725	1809

23. During the downturn the un-weighted average coverage was more than six times the net provisioning flow. ^{21, 22} This number has increased slightly since then, due to the still-growing DP funds. The total provisions coverage (i.e., dynamic, specific, and other general provisions) currently amounts to 18 times downturn losses in 2008-09. During the downturn, three banks had insufficient or barely sufficient coverage by dynamic provisions (second column), but even these banks now show coverage ratios of between 110 and 364 percent of downturn NPF (third column) and their total provisions coverage ranges between 370 and 633 percent (fourth column). Five banks did not record any positive NPF during 2006-10 on account of releases of provisions and loan recoveries and are consequently excluded from the calculation of averages.

24. Generally, there is a large dispersion of DP coverage ratios, ranging from about 100 percent to around 20 times expected loss. This is because the statutory rates for accumulating dynamic provisions follows an approach based on uniform (average) rates for all banks, regardless of their diverging risk profiles of loan portfolios. For some banks showing lower coverage ratios, the statutory DP rates were about adequate, while for safer banks with a lower expected loss, these rates seem to have been larger than needed and led to high coverage ratios.

25. Owing to the divergence of actual loan losses from expected losses, there has been a net cost of dynamic provisioning through the cycle. During 2006-09, encompassing both upswing and downturn years, the pre-tax return on assets (ROA) of private banks (i.e., the entire banking system, excluding BROU) would have been about one-fifth higher, if dynamic provisioning had not been in place (this assessment supposes that lending conditions would not have been different in the absence of DP). ^{23, 24} The high costs have also resulted from the fact that most banks had not reached the upper limit of 3 percent of loans and thus could not completely offset the cost of rising specific provisions during the downturn.

26. The authorities have recently announced changes to the provisioning system to align it with updated default expectations (see Box 3). ²⁵. Effective end-June 2012 the rates for specific provisions in the loan classification system will be adjusted as will be the classification criteria for certain loans. The dynamic provisioning system will undergo fundamental changes: banks will have to provision only for higher-quality loans, the dynamic provisioning formula will be modified in line with the most recent calculations of expected loss, and the limit of the DP funds will be become bank-specific and presumably decrease on average. Based on recent BCU calculations the average size of DP funds would intially drop by only about ¼ percentage point (with larger reductions for individual banks), but the changes to the contribution rates of the DP formula may allow a larger offset of loan losses in the non-expansionary phase of the credit cycle going forward. In sum, the regulatory changes are certainly a step in the right direction, although their overall effect remains to be seen.

D. Conclusions

27. This paper analyzed market structure and provisioning framework as two aspects relevant for financial intermediation in Uruguay and found that:

• The market is more concentrated and competition is estimated to be somewhat weaker, compared to pre-2002-crisis levels and peer countries. The concentration in the banking market increased in the aftermath of the 2002 crisis, which could have set a lower standard of market competition. In addition, there is some evidence that market power affects profits positively. Since research suggests that greater competition in the financial sector is associated with greater financial intermediation and economic growth (Claessens and Laeven, 2005), this situation warrants further analysis to understand challenges for achieving deeper financial intermediation in Uruguay.

• Profits are also affected by the accounting treatment of inflation. The accounting standards in Uruguay differ from the usual international practice under low inflation environments. While in some years this accounting practice may have no effect over recorded profits, in others it may overstate or understate banks’ return on assets in international comparisons. To make accounting internationally comparable, the operating results need adjustments for net expenses on inflation. Once such an adjustment is made, Uruguay’s banking sector profitability in the last two years improves.

• Provisioning requirements are strict, and appear to have generated an ample cushion of loan loss reserves. This cushion is a strength in turbulent times (especially under tail risks), but it is also costly. Setting appropriate parameters for the provisioning framework is not straightforward since the size of the future shocks is not known, and cannot always be inferred from the past, especially if large structural changes are taking place. That said, the recently-announced recalibration of the system’s parameters is a welcome step in addressing the cost of very high reserve levels.