When real sector ROIC begins to slip, an increasing number of companies have difficulty finding the cash to repay their debts. With all the doubts about company accounts, cash is one of the most telling indicators of trouble ahead. This, in turn, puts tremendous pressure on banks, whose per- formance begins to slide as nonperforming loans begin to rise, as was the representative case of Mexico (Figure 3.6).
Interpreting the Warning Signals for Banks
The potential impact on banks, whose loan portfolios are dominated by fail- ing real sector companies, is intuitively obvious. Nonetheless, actually tracking and measuring the impact of the real sector’s problems on banks is another matter altogether. Unfortunately, understanding the full extent of the problems in banks’ balance sheets is obscured by accounting practices and transparency issues. Moreover, to interpret the full impact of the real sector’s problems on the loan book requires them to be marked-to-market, which is almost never the case until a crisis occurs. The size and breadth of banks’ problems generally cannot be known until after the fact, when a full accounting forces them to the surface. Hence, estimating the true extent of banks’ problems before the fact requires interpreting trends and accounting information.
Tracking Nonperforming Loans11 The trend ratio of NPLs to total loans is a powerful sign of bank problems when credible and timely data are avail- able. An NPL portfolio of 1 percent of total assets signifies probable write- offs of about 5 percent of bank capital, depending on the bank’s provisions.
Using this as a yardstick, NPLs of 1 percent and rising12can signal trouble, whereas levels of 5 percent invariably signal significant problems. The trend is also important: Rising NPLs, regardless of their level, signal problems.
The good news is that regulators and bankers carefully monitor NPLs. The bad news, however, is that unless disclosure practices are strongly enforced by regulators and accounting firms, the NPL levels that are reported are usu- ally unreliable. The reason for this is that banks have a strong incentive to underreport their NPLs. Reporting them accurately, after all, would force the bank to increase the funding of its bad-loan reserves, which would then lower its earnings. Accurate reporting might also shrink the bank’s loan portfolio, which would then compel the bank to raise additional capital to replace its loan losses. It could also trigger supervisory intervention. Hence, banks have strong incentives to underreport the true extent of the problems on their portfolios; for this reason, too, some countries’ banking systems manipulate the meaning of an NPL to remove some of its sting.
Investors and managers cannot necessarily count on external auditors and regulators either. They can be fooled, and their accounting invariably comes too late. Thus, even in financial systems where regulators are respected
Figure 3-6
AS PERFORMANCE SLIDES, NPLS RISE Mexico, ICR percent NPLs*
4.7 4.3 4.4
3.9 3.8
1990 1991 1992 1993 1994
9.0 9.2 6.6
3.7 2.3
Crisis
NONEXHAUSTIVE
NPLs
Real sector performance ICR
*NPLs are calculated as percentage of total loans Source: IFS; Bloomberg; McKinsey analysis
FIGURE 3.6 As performance slides, NPLs rise.
for their skills and perspicacity, accounting of probable loan losses should be suspect.
To make matters worse, we have found that the greater the problems in the loan portfolio, the greater the incentive to lie about their quality. This has led us to take this kind of information with a grain of salt. For example, we routinely use a modeling technique that stress tests the health of banks by using loan-loss scenarios. In these analyses we assume that NPLs are understated by 25, 50, and 100 percent—and ex post audits typically bear out this skepticism.
In sum, what should be an important indicator of the health of banks is also sometimes the least reliable. Reported NPLs in countries with known problems, therefore, should be viewed suspiciously until all of the facts can be assembled and analyzed. While we maintain a healthy skepticism about the accuracy of NPL reporting, we also invariably track it, looking for pat- terns and trends. We read the published reports and listen to the opinions of the experts, even when their opinions are based on little more than impres- sions. In addition, we also track other indicators.
Other Indicators of Banking System Health To track the health of the banking sys- tem, we recommend four other indicators:
1. High rates of growth of the loan portfolio are used frequently to illus- trate the health and vitality of a banking system. This measurement must be applied cautiously, however. Sustained credit booms with high growth rates put systems under serious strain and frequently can lead to banking system
10
40 22
23 4
9 17
Agriculture, forestry/fisheries Mining
Utilities: electricity, water, gas
Manufacturing
Construction
Services: retail, restaurants/hotels Transportation, communications
Loan portfolio growth
-3 -1 -1 -1
2 6
-9
Sector economic growth
Percent 1996–98
Source: Asociación Nacional de Instituciones Financieras (ANIF); Asobancaria; McKinsey analysis
FIGURE 3.7 Colombian banks’ loan growth outpaces sector economic growth.
failures. Indeed, a common pattern in the period leading up to a financial crisis is rapid loan growth. Figure 3.7 shows one such case: The rates of growth of the loan portfolio to the least healthy sectors of the Colombian economy outpaced the growth of credit to healthier sectors during the two years prior to that nation’s crisis.
2. ROA (return on assets) calculations also are used commonly as met- rics of banking system health. In our judgment, levels close to or below 1 per- cent indicate that the system is barely matching the normal profit levels of developed economies. Figure 3.8 shows data for Colombia’s banking system during the period leading up to the 1998 crisis. Similar information for Korea’s banks shows that the ROA fell from 0.62 percent in 1994 to nega- tive 1.06 percent in 1997. Moreover, asset values are questionable in many developing countries for the reasons noted above.
3. Capital adequacy ratios that fall below 10 percent are also worri- some, especially in emerging markets. It is important to examine the defini- tions and requirements for tier-1 and tier-2 capital (and how these relate to NPLs), since these can vary across countries. In 2001, for instance, Mexico’s regulators still allowed tax losses to carry forward and be counted as bank capital. This definition would have been acceptable if the Mexican banks had somehow been guaranteed future profits. Since the capital had yet to be earned, however, this was a dubious practice.
4. Net interest margins are also useful indicators. When this indicator drops under 200 basis points, we become concerned. When they fall below this level, it is very difficult for banks to earn sufficient returns, given their cost structures and capital investments, which are other indicators.
1.0 0.68
-1.28
-1.03 1.24
Dec 96 Dec 97 Dec 98 Dec 99
Minimum expected ROA
g
COLOMBIA’S DECLINING BANK PROFITABILITY SIGNALS PROBLEMS ROA, percent
Source: Asobancaria; Superbancaria; McKinsey analysis
FIGURE 3.8 Colombia’s declining bank profitability signals problems.
Banking system indicators provide less timely warning signs than do those of the real sector of the economy. Consequently, predictions of financial crises that are based solely on them are less insightful and robust than predic- tions based on a combination of banking and corporate sector indicators.
Why Banks Fail
After many years of working to counter the effects of financial crises, we have developed a point of view about why banks fail. The simple causes, in our opinion, are traceable to errors made while underwriting loans.
However, this answer is not very satisfying in itself and it leads to more questions, such as: Why were these underwriting errors made in the first place? What impacts do special conditions that are generally not present in more developed and stable markets have in these bank failures? Do these banks operate in dangerous markets?
Bankers operating in economies hit by financial crisis recognize that corporate sector underperformance places enormous pressure on banks.
They can trace the effects of real sector underperformance on their loan portfolios. Since loan portfolios in most countries are kept on the balance sheet, without being marked-to-market, it is nearly impossible for bankers (or regulators) to fully and fairly assess the value of their portfolios—and the state of health of their banks. Moreover, because the banks themselves provide the lion’s share of financial intermediation in these countries, the capital market trading that would force visible repricing in a more advanced economy generally does not exist. In crisis conditions, many banks simply will not recognize the extent to which their asset values have declined, since they fear a collapse of confidence.
Even those who knew before the crisis that they were on perilous ground, and who could have possibly reshaped their portfolios, generally did not. As we showed in Figure 3.7 above, Colombian bankers deepened their commit- ments to the very companies whose declining performance was threatening them. This phenomenon is common in many developed economies as well.
Old Game, New Game
Before the globalization of economies, the old banking game had been about channeling funds to businesses in a closed economy. In most of the emerging market countries we studied, banks historically had relatively little choice regarding to whom they lent, for which projects, or at what rates.
Their role was to create a deposit base, then lend the funds (as well as those borrowed or granted from overseas) to the government and strategic sectors of the economy, often as directed by the government or the banks’ conglom- erate owners. Banks and financial intermediaries were not expected to make
much money lending. Nor were they considered risk takers. After all, their debtors faced little or no competition, and in any case would not be
“allowed” to fail. Given these parameters, banks, not surprisingly, built enormous systems for taking deposits based on numerous local branches.
They also built large bureaucracies for working with government—in fact, many banks took on the feel of a government agency themselves.
In countries dominated by high inflation and frequent currency devalu- ation, banks also developed skills in buying and selling foreign exchange (often an exclusive legal right) and taking advantage of the spread between these treasury activities and the negative real interest rates they paid out on their deposit base. Banks, like their real economy cousins, were products of their controlled-economy environment.
Reform, however, changed this environment radically. First, as barriers to entry were removed and ownership laws restructured, new domestic and international banks and nonbank lenders entered the sector, intensifying competition for the most attractive borrowers. Second, as deposit and loan rates were deregulated, banks in many countries were allowed to price risk for the first time, which required a whole new set of skills.
Third, lending practices were significantly liberalized, allowing banks to decide to whom to lend, and for how much, again requiring a new set of underwriting skills. Fourth, BIS capital adequacy regulations were broadly adopted, focusing attention on risk assets. In most emerging markets this caused increases in the resources available to lend, which in the past had been tied to other balance sheet indicators (especially cash reserves).
Fifth, intermediaries were allowed to borrow money from abroad, in foreign currency, providing significant additional liquidity and opening up a new area of currency risk. Sixth, in some cases, lower fiscal deficits and access to low-cost funds through international capital markets and from foreign market entrants caused a “crowding in” effect. As government took a smaller proportion of their domestic loans, banks were left with excess liq- uidity to deploy, including times when lax monetary policy may be in place to give a country’s liberalization a chance to flourish.
Almost overnight, banks were playing a very different game. Now they were free to grow their loan books for scale, while simultaneously ensuring that they could recover their funds. They were flush with resources bor- rowed overseas, much of it loaned on a short-term basis in foreign currency.
They also experienced lower reserve requirements that effectively increased their lending liquidity yet again. They faced competition for their traditional
“safe” government and conglomerate customers, many of whom were no longer “safe” due to real economy reform.
Banks had to broaden their base to include new customer segments, and were required to choose for themselves which customers were good bets, at what rates, and with which product instruments. As competition increased,
they were also required to run their own operations for profit, but faced dif- ficulties in streamlining their bloated structures dramatically. In sum, they needed new rules and new skills.
The Failure to Adapt to Market Conditions
Most banks were not up to the new game. Four reasons help to explain their failure: weak credit skills, cozy and directed lending practices, rational expectations built on the old game, and the drive for new earnings.
Weak Credit Skills Credit skills had never been important to banks in most of these countries, and thus they were never developed or rewarded. As a result, when the underlying conditions changed, local banks were starved for talent and forced to place increasingly important lending decisions in inexperienced, unskilled, and unsupervised hands. The requisite credit cul- ture simply did not exist.
In many economies, deficient credit skills were among the biggest issues leading to financial crisis. In the case of a Mexican bank, for example (Fig- ure 3.9), much of its bad loan portfolio was flawed at origination. Only 5 percent of its NPLs could be traced to events in the macroeconomy beyond the bank’s control.
Directed Lending Many banks in these countries had been either part of a conglomerate, controlled directly by the government, or established by gov- ernment with a charter to foster the growth of some sector of the economy.
POOR CREDIT SKILLS LEAD TO DESTABILIZING NPLS Pesos, percent NPLs
Insufficient/flawed analyses
Out of bank’s control
Number of cases
Amount in millions of pesos Poor monitoring
of client performance Inadequate structuring Policy violations Origination
Monitoring
Mexican bank
100% = 123 571
24 26
27 24
19 20
26 25
4 5
Source: McKinsey client analysis—Mexico
FIGURE 3.9 Poor credit skills lead to destabilizing NPLs.
They were set up as facilitators of development, not as profit generators, and their lending practices reflected these relationships and objectives.
Conglomerate-owned banks in Ecuador and Jamaica, for example, existed solely to further the conglomerate’s growth objectives. For these rea- sons, assets in one company were used to guarantee loans in other companies, often regardless of the creditworthiness of the companies or the expected investment returns. Banks within the group made loans based on these guar- antees, or in some cases with little or no guarantee at all. In such cases, loans were based on group-level strategy decisions and government policy objec- tives, rather than more legitimate underwriting guidelines.
State-owned development banks, such as Bancafé in Colombia, experi- enced similar pressures and made loans based on the needs of economic development and industrial policy, and often to state-owned or controlled natural resource and heavy industry entities.
Not surprisingly, conglomerate-owned banks tend to be poor perform- ers. Even when deregulation freed the banks to lend to other parts of the economy, longstanding (often family) relationships continued to exert hid- den pressure. A combination of coercion (such as the withholding of key licenses) and the inertia resulting from years of habit made these lending pat- terns difficult to break. In one case in Russia, for example, more than 75 per- cent of all bank loans went at preferential rates to bank-related companies.13 Rational Expectations Built on the Old Game Ironically, even if a more rational credit policy had been established in Russia and elsewhere, the old corpo- rate behemoths would have still received the lion’s share of funds. After all, these old-line borrowers still dominated virtually the entire economy. They had long-established relationships and favorite “cronies” to substitute for their poor or nonexistent credit histories. From a banker’s perspective, they presented a better risk than the new market entrants, which relied on unproven reforms and had yet to prove themselves.
The Frenetic Drive into New Markets Even these behemoths could not absorb all the liquidity flowing into local banks, particularly when access to foreign capital was also available. Thus, as the corporate and government borrow- ers gained access to foreign bank lending, foreign direct investment (FDI), and capital markets, local banks began to be crowded out, and therefore began to look for new lending targets, including individuals and small and mid-sized businesses. The opportunities in these segments were large, as they had typically been previously underserved and underdeveloped. Now, with the advent of reform, they were growing rapidly. However, these seg- ments were unsafe, at least for local banks.
The corporate middle-market is difficult for banks to assess, even in devel- oped economies. In emerging markets, these difficulties are compounded:
Information on debtors is unreliable or nonexistent; virtually all relationships are new, with little credit history to help discern potential risk. Furthermore, the financial statements of such firms frequently lagged performance, are rarely audited, and are often inaccurate. In addition, it is often impossible to differentiate between the company and the individual owner. Thus, special skills are required to lend effectively to customers in this segment—and banks sometimes do not possess those skills.
In addition to commercial loans, there was a pent-up demand for con- sumer lending—in the form of auto loans, mortgages, credit cards, and other household durable loans. Most of the loans in these areas were too small to warrant individual analysis. Rather, they relied on underwriting techniques that depend on the historic performance of the borrower, not on his or her future cash flow. Unfortunately, most consumers in emerging economies do not have a consumer-lending history in any organized, commercially avail- able form. There are no credit bureaus, and those that do exist tend to be woefully inadequate. Bankruptcy and claims information is inaccessible. All of this makes risk assessments and write-offs hard to manage.
Regardless of the risk, banks felt the need to find a home for the grow- ing amount of resources on their balance sheets. Banks began a “race for the bottom” in their quest for market share, in which middle-market and con- sumer loan portfolios were often the fastest growing components of the total portfolio. In Colombia, for example, during the five years prior to the crisis, mortgage loans grew at an annual compounded rate of 19 percent, and commercial loans grew at 15 percent per annum.14
Watching the Descent into Unmanaged Risk
Banks—bewildered by the pace of change around them and determined to remain relevant in the evolving economy—often built their loan books to the sky. In so doing, they exposed themselves to a number of significant risks which varied in intensity by country. Chief among these, of course, was the risk of lending without the requisite borrower information or credit skills to evaluate individual loans (or to effectively monitor those loans once made).
Also critical were the liquidity and currency risks inherent in the kinds of borrowing and lending that banks were doing.
Banks were borrowing short-term funds from abroad, often in foreign cur- rency, and lending those funds domestically in longer-term instruments denom- inated in the local currency and sensitive to local economic conditions.
Sometimes they loaned in foreign currency and the borrowers took the exchange rate risk, with disastrous results for them and the bank. Thailand’s massive and ultimately disastrous move into property loans is just one example.
Mortgages, with their long term structures, were being mismatched with deposits and other sources of bank funding with much shorter maturity. The