A house built on sand

The troubles of Mexico’s bankers are only just beginning. In the months ahead their past mistakes will come under renewed scutiny in a major investigation demanded by the Mexican congress. Some bankers could lose their jobs, a few could wind up in jail, and their banks – still suffering the effects of the industry’s near collapse following the 1995 peso crisis – will be put under new pressures.

With political point scoring a major factor in the investigation, predicting the final outcome is tough. The bleakest scripts involve a systemic collapse of the banking system and maybe renationalization of the banks.

For the first time in Mexico’s modern history opposition politicians have the power of a congressional majority, and they are determined to use it. They want the bankers, civil servants and ministers who they judge acted unprofessionally, and in some cases fraudulently, before and after costly bank rescues, to be brought to account.

Their attacks on the banks, some of which have links to the ruling Partido Revolucionario Institucional (PRI), may be politically motivated, but the signs are that a searching inquiry will find plenty of genuine dirt.

The focus of attention is the Mexican bail-out vehicle, the Fondo Bancario de Protección al Ahorro (Fobaproa), which in the wake of the crisis took over the worst banks and recapitalized the better ones. But the net is likely to be cast wider, drawing in politicians involved in bank privatization in 1991 and 1992 when some banks ended up in the hands of crooks, and bankers who used the bail-outs to cover up lending errors and sloppy practices. Even foreign bankers who bought distressed banks cannot rest easy. While not accused of unprofessional behaviour or fraud, they did get some very favourable deals whose terms have not been fully disclosed. When these are revealed questions may be asked about why such deals were cut with foreigners.

“[Fobaproa] will be one of the major issues over the next three years and could cost the ruling PRI party the presidency in the year 2000,” says leading Mexican economics commentator Sergio Sarmiento. “Although most people don’t fully understand the subject, it’s easy [for the opposition] to say that the government rescued the banks for the bankers and gave them $65 billion [the estimated costs of the bail-out roughly equivalent to 14% of GDP].”

Weakest in Latin America

Fobaproa resolved the crisis by swapping banks’ bad assets for a government-guaranteed note with a 10-year maturity which capitalizes interest. The debts remain administered by the banks and they retain a contingent liability of roughly 25% on the uncollected portion. Three years after the crisis little progress has been made in resolving the debts and now the government wants to turn Fobaproa’s liabilities into public debt. But opposition parties have cried foul and demanded an audit of Fobaproa before agreeing to this and other reforms aimed at modernizing Mexico’s financial system. US charges that leading Mexican banks have been involved in laundering drug money (see separate story, page 48) have added weight to their case for an inquiry but, ironically, have slowed reforms designed to improve financial regulation and supervision.

The threat to the Mexican banking system comes if the opposition refuses to accept Fobaproa’s liabilities as public debt. The opposition claims the government had no legal authority to add to the public debt in this way. Others have pointed out that the government was also legally compelled to rescue the banks. But if the exchange of assets for notes is found to be illegal and the bad debts went back to the banks the system would collapse. Total system capital at end-1997 was roughly $10 billion while the Fobaproa notes issued for loan acquisition amount to $18 billion.

Mexican banks are in any case very weak. Moody’s Investors Service gives them an average country financial strength rating (a measure that excludes the possibility of government intervention in a crisis) of a low E+, the weakest in Latin America and the third lowest in the world, ahead of only Thailand and Pakistan. On this basis Mexican banks are even weaker than Korean and Chinese banks.

Analysts further complain that while Mexico introduced US GAAP accounting standards last year they are not applied rigorously. When analysts start making capital deductions to take account of the differences, as well as the fact that average reserving of non-performing loans is only 60%, the capital of many banks comes out as “thin to non-existent”, to quote one researcher. Some banks even have negative equity measured this way. And Mexican banks have not yet begun reserving for their contingent losses arising from the loans sold to Fobaproa.

The other bogeyman of Mexican banking is nationalization. The two opposition parties are the market-oriented Partido Acción Nacional (PAN) and the left-wing Partido de la Revolución Democrática (PRD). Some elements of the latter believe that nationalization of the banks is the best solution to the problems. But Sarmiento says that when the PRD is in power – it runs Mexico City – it has proved less radical than its rhetoric. “The real risk of having the banks nationalized is very small. Not even the PRD would do that in my opinion,” he says.

To others the whole Fobaproa issue is a political storm that will eventually subside. Says JP Morgan managing director and general manager in Mexico, Eduardo Cepeda: “We have to separate the economic handling of the banking crisis from the political deadlock. Congress understands that the banking system had to be rescued, understands that public debt [the Fobaproa notes] is public debt and that the reforms have to be passed. But they want a political price for it.”

That political price could be very high. It may include the resignations of central bank governor Guillermo Ortiz, who is under pressure for his role in bank privatization while in the finance ministry, as well as of Eduardo Fernandez, the president of Mexico’s financial supervisor Comisión Nacional Bancaria y de Valores (CNBV). Ortiz did not respond to Euromoney‘s requests for an interview.

If they go, a clutch of top bankers are likely to follow. The Mexican public feels that many bankers have come through the troubles unscathed while ordinary people lost out. “The feeling is ‘why should the bankers get off scot free by selling loans to Fobaproa whereas Mr García [the ordinary Mexican] lost his house?'” says Felix Boni, head of research for WestMerchant Bank in Mexico City.

The review of Fobaproa is likely to provide critics with plenty of evidence of malpractice. “There has been much speculation as to what the auditors may find in the portfolio of loans assumed by Fobaproa,” says Paul Warme, a banking analyst with Paribas in New York. “Many suspect that fraudulent and illegal lending on the part of some banks may be discovered. In fact Ortiz has publicly conceded that many loans in the Fobaproa pool were granted under suspicious circumstances and conditions.

“In addition,” continues Warme, “the auditors may find that some banks were careless, and perhaps even negligent in the granting of many credits prior to the crisis. It is no secret that many loans assumed by Fobaproa are poorly documented. In many cases promissory notes were not properly executed, loan files were incomplete and collateral was not properly registered. This effectively means that many of these loans are unenforceable in a court of law. The auditors may also find that many loans were granted without the banks conducting a proper credit analysis and without conducting a thorough due diligence.” He adds: “If the auditors find this to be the case, public sentiment against the banks could worsen and demands that the banks – or the shareholders – take responsibility for their carelessness could heighten.”

Let’s eat steak for lunch

The origin of low professional standards in Mexican banking goes back to nationalization in 1982 and subsequent privatization of the banks in 1991 and 1992. Over a decade of state ownership the Mexican banks lent mainly to the government and the skills of granting private credit were lost.

Then came privatization with 18 banks being sold to the highest bidders, some of whom had little or no experience in managing financial institutions. During the economic growth period prior to the peso crisis, lending accelerated and much of it appears to have been done carelessly and some fraudulently. Another area of controversy is where some bank owners got the funds to buy the banks in the first place. There is a view that banks lent each other the money and that these loans have ended up in Fobaproa.

Bankers may blame their troubles on the peso crisis, but analysts say that poor performance and lax supervision meant that the banks were heading for a fall anyway.

A few bankers have already been charged in connection with the banking crisis. Jorge Lankenau, who ran Banca Conía, is in prison awaiting trial. Banking authorities accuse him of covering up $2 billion of losses through fund transfers between Conía, its brokerage house and offshore companies. Originally placed under house arrest in Monterrey last October, he gave a 24-man police cordon the slip but turned himself in a few days later.

Others have been even more elusive. Angel Rodríguez Sáez, former chairman and president of Banpaís, fled Mexico in 1996. He fought extradition from Spain for two years before being returned last month to face charges of fraud and embezzlement. Officers of Interpol had to fish him out of the Mediterranean after he dived off his yacht in Ibiza in a last ditch effort to evade capture. Another banker, Carlos Cabal Peniche, who allegedly plundered Unión and Cremi, remains a fugitive. Unión and Cremi, which merged, were supported by Fobaproa before the peso crisis, adding weight to the argument that Mexican banking problems predate the crash.

Before the crisis Fobaproa was a tiny department of only 30 employees in the Banco de Mexico, home to career central bankers serving out their last few years before retirement. It was set up to insure bank deposits, and senior central bankers barely knew where its offices were. Then, almost overnight, this trust fund became the vehicle for organizing bank bail-outs, and, as one Mexican banker puts it: “The tiny animal became a monster.”

With assets of Ps218.7 billion ($24 billion) – consisting mainly of the bad assets of recapitalized and intervened banks – and liabilities of Ps552.3 billion, Fobaproa is now nicknamed “one of Mexico’s largest banks”, albeit one with a sizeable hole in the balance sheet. Fobaproa now has 300 employees and its main offices are four floors of the old Banca Cremi headquarters, one of the banks being liquidated. The plush surroundings and elaborate dining service inherited from Cremi – turning down steak for lunch is considered bad form by the butlers – suggest that in their short heyday Mexico’s newly privatized banks enjoyed a lavish lifestyle.

Something for the rich

Fobaproa itself is regarded by ordinary Mexicans as a vehicle for saving the wealthy. “It is something for the rich,” says a taxi driver who lost his furniture store in the crisis. Intellectual observers agree that bankers exploited the opportunity Foboproa presented to them to offload bad loans. Sarmiento says that Fobaproa started off by rescuing banks in deep trouble but soon had a line of bankers outside its door to see what was on offer. “Fobaproa ended up rescuing banks that could have survived on their own,” he says.

Since the crisis neither Fobaproa nor the banks have been very successful in either selling or recovering bad debts. Mexican investor Jose Luis de la Mora was one of six partners in Wyss Carrasco, an investment fund set up after the peso crisis to buy distressed companies and turn them around. It bought several direct from banks and shareholders but he recalls that the process became a good deal harder after Fobaproa became involved. “Fobaproa was overwhelmed,” he says. “It couldn’t solve the problem of so many loans all at the same time. It didn’t even have sufficient basic resources such as staff, telephones and computers.” In one case a deal was agreed with bankers but Fobaproa executives did not give the go ahead probably through fear of being criticized for selling the assets too cheap.

Bankers’ own collection efforts are not helped by a strong borrowers’ lobby that is challenging repayment and has been fighting the issue of whether banks can charge interest on back interest – standard banking procedure – in the Mexican courts.

Says one banker:”We have a society where no-one pays. Borrowers are happy today because no-one is knocking on their door saying ‘pay me’.” Under Mexico’s complex legal system recovering money through the courts is extremely arduous and long-winded which helps debtors to escape repayment. When borrowers discover that their loans come under Fobaproa, as most have, they are even more reluctant to repay believing that the government will eventually let them off.

The other side of the coin from borrowers escaping legal sanctions is that white-collar crime, such as banking fraud, is rarely a matter for prosecution in Mexico. Says a Mexican consultant: “This country is a haven for educated crooks. No-one is ever guilty.”

Rogelio Sada, the congressman responsible for the PAN’s economic policy, says that the moral credentials of some of those who bought Mexican banks are in doubt. “We know that there were a lot of credits given to important companies with wealthy owners – that needs some clarification,” he says. “Debt was bought from the banks on the condition that shareholders would raise capital but in some cases the increases didn’t come through.”

Sada says that one of the weakest points of the Fobaproa programme was that when the banks’ bad loans were purchased many different deals were made. “Some banks got huge portions bought by Fobaproa, some got less. When a big chunk of [government] money is used like this it needs an explanation,” he says.

The government’s defence of its handling of the economy and the role of Fobaproa is simple: the banking system did not collapse, there were no runs on any banks and total costs, even at 14% of GDP, are less than those of comparable crises. The Chilean crisis during the 1980s cost 30% of GDP.

Mindful of the slow pace of progress and the high public profile Fobaproa has attained, the central bank appointed one of its hottest properties, the 34-year-old lawyer Javier Arrigunaga, to the post of director general last year. The government proposes splitting Fobaproa into two companies, a deposit guarantee fund (Fogade) and an asset recovery commission (CRB), modelled along the lines of the US Resolution Trust Corporation, with the aim of disposing of the current assets of Fobaproa in six years. New assets would be managed by Fogade which would provide limited guarantees to depositors in an attempt to reduce moral hazard. The old arrangements whereby Fobaproa provided 100% guarantees to depositors have been criticized for promoting moral hazard.

Poor asset recovery

Arrigunaga admits that Fobaproa was initially unprepared for its task and that asset recovery and disposal have been slow. “The banks haven’t done a very good job in the recovery of assets,” he says. “It’s difficult to say how much is due to the quality of the loans and how much due to the quality of the banks but we have to make sure they have the right incentives to collect.”

Arrigunaga’s view is that Fobaproa doesn’t have the resources to make credit decisions and that this must be left to the banks especially as Mexico has no third-party service industry to turn to. But Fobaproa is starting to hire foreign banks to manage loans of the supported banks and may eventually move on to the loans of surviving banks if it is unhappy with their collection efforts. The contract to manage a small portion of the loans of Oriente, one of the supported banks, was recently given to the US company Amresco with a base valuation put on the loans of 18 cents on the dollar. Fobaproa executives say the valuation is in line with expectations since this was one of the worst parts of the portfolio and the deal has upside.

Current government estimates are that 30% of the value of bad assets will be recovered although analysts put it lower at 15% to 20% and say that the longer it takes the less will come back.

Says Philip Guarco, an analyst with Moody’s Investors Service: “It’s a big mess. This is starting to look like Chile in the mid-1980s [after a major banking crisis]. Fobaproa was modelled after the Chilean programme and it took Chile 12 years to dig itself out.

“The problem of recovering any value in these loans is that the value dissipates over time. If they had sold them immediately they would have got 30% to 40%. Now 15% to 20% is more realistic.”

The other change Arrigunaga has brought to Fobaproa is greater transparency. While figures on recovery rates are still not available, a full report and balance sheet has been published for the first time. Delivered in a Banca Cremi envelope – showing how Fobaproa’s new management intends to realize the full value of distressed assets – the report contains figures supporting the view that many individual deals were cut.

The loss-sharing arrangements on debt swapped for Fobaproa notes appear to have been flexible as is the interest rate, which analysts had assumed was a standard 91-day cetes rate but ranges as high as Libor plus 5%. For example Serfin, which is now 19.9% owned by HSBC, received a Ps4.4 billion note with a value date of January 2, 1996 with a 25% loss-share paying Libor plus 4% to maturity. Spain’s BBV which took over Probursa received three notes with no loss-sharing.

These differences in interest rates would make it difficult to convert the Fobaproa notes into standard government debt. Some kind of securitization structure might have to be set up with cash flows passed through to the banks, or different values of government debt swapped for the notes depending on their interest rate.

Then there is the assistance provided to supported or problem banks. Again it was assumed that loans were swapped for notes. Now it appears that nearly all assistance provided to Obrero, Cremi, Unión and Oriente was in the form of cash. Serfin and Inverlat received a mixture of cash and notes and Banpaís (now sold to Mexico’s Banorte) received at least one cash injection.

Now analysts are asking what other kinds of special deals are going to be revealed. In the later deals, for example, is all interest being capitalized or is some being paid in cash? How much has the formula of two pesos of loans for one peso of capital and 25% loss-sharing been stretched? Some analysts believe it could have gone as high as six to one and that when recovered Mexican banks – such as Atlántico, Promex, Bancen and Banpaís – have been sold on, the original terms may have been reworked.

Foreign bankers are especially sensitive about the terms of their deals. Most of these deals concern banks that were either supported or required great assistance from Fobaproa such as Santander’s acquisition of Mexicano and Citibank’s purchase of Confía. While each deal is different the general approach has been to encourage acquiring banks to keep the bad assets and for Fobaproa to provide the reserves for provisioning. In the case of Inverlat, managed by Bank of Nova Scotia – which has a 10% stake, with an option to convert a bond to obtain majority control – there is a commitment by Fobaproa to maintain the tier one capital ratio at 9% to make good any asset recovery problems.

William Sutton, president of the executive committee of Grupo Financiero Inverlat, says: “The local press criticizes us for getting the deal of a lifetime but back in mid-1995 when we started negotiations the outlook was very bleak. Our minority interest in Inverlat [5% due to go to 8% through a convertible bond] was wiped out in the crisis and written off so we could have invested in any Mexican bank. No other foreign banks were looking at Mexico at the time.”

While PAN’s Sada says that none of the current deals will be reneged on, he admits that foreign ownership of banks is a sensitive issue. His party has not yet developed a firm policy on the issue but says that whatever evolves will have to take account of the need for Mexican banks to be ready for new competition. Included in the reform proposals before congress is one allowing greater ownership of Mexican banks by foreigners, currently limited to those banks with 6% or less of total system capital. Also on the reform agenda is a proposal to make the CNBV more independent and to give the central bank control over foreign exchange policy.

Tex-Mex GAAP

Right now, however, Mexican banks are some way from being ready for further foreign competition. One of the problems they face is the amount of space on their balance sheets occupied by Fobaproa notes – 60% of assets in the case of Banorte when its acquisitions of Banpaís and Banco del Centro are taken into account, 41% in the case of Santander Mexicano and 31% for Serfin. Since these notes capitalize interest they are an expanding asset class that has to be funded by deposits on which interest has to be paid in cash. There is no liquid market for the Fobaproa notes which is why turning them into government debt would be very attractive from the banks’ viewpoint. Currently, Mexican banks have severe cash flow problems and are unable to do much new lending, which inhibits their prospects for growing out of the crisis and contributing to Mexico’s economic recovery.

But digging deeper into the balance sheets of Mexican banks reveals other problems. While the average system-wide capital ratio is ostensibly 18%, making allowances for such things as the Fobaproa contingency and a fuller coverage of past due loans (PDLs) takes the figure a lot lower and into negative territory in the case of some banks. “I call it Tex-Mex GAAP because the banks are still doing a lot of non-traditional things in their accounts,” says Patrick Boucher, co-head of Latin American bank research for Santander Investment. “For example Mexican GAAP is flexible in allowing banks to account for past due loans (PDLs) as seized assets. In North America seized assets are part of PDLs until the asset is sold. Seized assets should also not be kept for more than 360 days but, in Mexico, banks are allowed to create off-balance-sheet trusts, to sell the asset into the trust and receive a note that is included in miscellaneous notes and securities. The actual balance of these trusts is not disclosed.”

Boucher says he puts these notes and seized assets back into PDLs. But the work doesn’t stop there. Restructurings are also a concern to Boucher and other analysts as they fear many will come back into PDLs when grace periods expire. In the mortgage sector there is a view that lease-to-own programmes are covering up heavy subsidizing of borrowers by banks as a way of keeping the loans current. The low average reserving of past-due loans, about 60% system wide, is another worry while provisioning for Fobaproa-related losses does not begin until later this year.

Analysts have also queried the inclusion of tax credits in Mexican banks’ assets which is non-tangible equity that cannot be called upon in a crisis; the introduction of inflation accounting which by restating amounts in current pesos skews figures; and the inclusion of subordinated convertible debt, of which Mexican banks have a great deal, as tier one capital. Mexico doesn’t subscribe to the BIS capital adequacy guidelines and goes its own way on such issues.

“With tax credits they are only as good as expectations of future earnings,” says Boucher. “If you put them on the balance sheet you are assuming within the next 10 years you will generate pre-tax income to fully amortize those credits. That is a big if.”

Warme of Paribas feels the situation with Mexican banks is so serious that measuring Mexican banks by the traditional yardstick – price to book value – is no longer relevant as book value is unquantifiable. One part of his evidence is contained in recent M&A transactions that have wiped out shareholder value.

In a recent report he says: “Following Bital’s acquisition of Atlántico (fourth quarter, 1997), Citibank’s acquisition of Confía (May 1998) and Bancomer’s acquisition of Promex (May 1998), the original shareholders in the acquired banks had their equity stakes in these banks wiped out. These transactions…confirm our view that the banks’ reported book values are overstated, and that further unexpected losses in the banks’ loan portfolios are possible.”

As an alternative measure of value, Warme uses economic value added (EVA) which shows a firm’s operating profitability after subtracting the cost of capital. Warme’s findings on Mexico’s two leadings banks using this method – Bancomer and Banacci which is the parent company of Banamex – are startling. “Our EVA analysis indicates that both groups’ return on capital falls short of their cost of capital by a significant margin. This suggests that the groups are destroying rather than adding shareholder value. Our projections also suggest that going forward, the groups will continue to destroy shareholder value over the next five years.”

Warme puts this down to the banks carrying of large volumes of NPLs which is “literally choking these banks and destroying any value being created by their core businesses”. PDLs as a percentage of total loans in the first quarter were 15% for Bancomer and 21% for Banamex. “Unreserved past due loans represent 49% of equity for Bancomer and 46% for Banamex, well beyond levels that we would consider prudent. It remains uncertain how successful the banks will be in restructuring and collecting their enormous portfolios of non-performing loans, and whether additional provision will have to be taken to absorb future losses.”

Despite all these negative factors a number of analysts are bullish on Mexican banks, believing that the bad news is out and there are good buying opportunities. In particular, they think Fobaproa is something of a diversion. “We recommend that investors ignore the politically driven noise affecting Mexican banks, focus on the operating fundamentals, which remain strong, and take advantage of this noise to buy on weakness,” says Daniel Abut of Goldman Sachs in a recent report.

Says Javier Fernández, chief financial officer of Bancomer: “There is no question that past-due loan ratios for Mexican banks are very high when compared with developed banking markets such as the US, Canada and some European countries. The absolute level of non-performing loans is the cost of the 1994-1995 crisis, but it does not mean that value will be destroyed in the next five years. In fact our bank’s successful workout results of the past two years indicate otherwise. The analyst [Paul Warme] reached a different conclusion only because of the specific scenario [EVA] he is using. Other highly respected analysts disagree with this conclusion and have buy and strong buy recommendations on the shares of major Mexican banks.”

But whichever way share prices go Mexican banks will have to deal with the bad debt problem. There is talk now of the banks buying back some of the bad assets that went to Fobaproa although Arrigunaga stresses that it would have to be after loan restructuring so that benefits pass to the borrower. If the banks were allowed to buy back the old loans at huge discounts without any restructuring the political furore would be enormous.

Jorge Hierro, deputy president of Grupo Financiero Banamex-Accival (Banacci), says that the bank would be interested in buying back loans at the right price. He says the bank has a team of account executives dedicated to recovering bad debts and that the individuals do not know whether a debt is held by the bank or has been sold to Fobaproa.

Hierro says that 12% of Banamex’s assets are made up of Fobaproa notes and that both it and Bancomer have the smallest percentages in the system. Capital resources at Banamex have increased two and a half times since 1994, say bank executives. “Bancomer and Banamex were not rescued [by Fobaproa],” says Hierro. “The government decided to set up a general programme available to all the banks and we decided to participate. We sold relatively small proportions [of bad loans].” Hierro adds that the Fobaproa issue has become very politicized. “It’s a mistake to generalize about Mexican banks,” he says. “No-one at Banamex is worried about the audit.”

Serfín investor relations manager Roberto Cejudo agrees that Mexican banks including Serfín have made past mistakes in granting loans without proper guarantees. “The audit is going to be a challenging issue for Mexican banks in the months ahead. No Mexican bank can feel entirely comfortable while this investigation is going on,” he says. “Serfín has been doing great in its collection [of bad debts] efforts, both with those loans in our porfolio and those held by Fobaproa. We have 1,000 people in our work-out area and last year we exceeded our goals which were in any case very aggressive. We have learned from past mistakes and now have very conservative loan growth.” Of the analysts’ comments about capital shortage, Cejudo says: “They may be theoretically correct but their claims are exaggerated. Of course we would like to have 100% coverage of non-performing loans but 60% is more than enough.”

Isaac Volin, director general of financial analysis with CNBV, which has also come under pressure for its role in the banking crisis, says that the use of tax credits by Mexican banks is in line with standard international practice; that Mexican banks don’t need 100% provisioning of NPLs because they can expect to recover some debts; that subordinated debt is only counted as tier one capital if it is mandatorily convertible within three years; and that under Mexican capital regulations 80% of the revaluation surplus can count as tier one capital because “otherwise inflation would create an undesirable redistribution effect by shifting what was originally tier one to tier two capital”.

Improvements there may have been but the Fobaproa audit will be more interested to understand how Mexican banking first got into such a sorry state. Whether the audit is a piece of politics or not doesn’t alter the fact that serious mistakes were made. Since then Mexican banks have travelled only a small way towards putting their house in order.