Weathering the Economic Storm: Mexican Crisis Experience Illustrates Need for Advanced Analytics, Greater Portfolio Intelligence
Alejandra Martinez; Signals, Winter 2001
The job of managing a consumer credit portfolio through a period of significant economic change can be likened to that of the captain of a ship caught in a storm. The ability to bring the ship safely to shore with minimal damage depends not only on the vessel's strength, but it's positioning relative to the waves buffeting it. The captain benefits from knowing, as early as possible, the expected duration of the storm, the direction from which it approaches, and its severity. Having some idea of these, the captain can work to best position the ship relative to the waves, and with skill and experience bring her to safe harbor.
Today's U.S. portfolio managers looking for lessons in managing through economic change need only look to the Mexican consumer credit market of the 1990's. Taking stock of the damage resulting from the volatile economic change in Mexico in the last decade helps one develop new ways for managing in times of volatility and uncertainty, such as those currently being experienced in the U.S.
The economic seas were indeed stormy in Mexico over the last decade. A true crisis emerged when between December 1994 and March 1995 interest rates soared from 14-75%. The average Peso/Dollar exchange rate for 1995 was 6.37—roughly double the rate of 3.27 for 1994. GDP dropped from a growth of 3.5% in 1994 to a decrease of 6.9% in 1995.
The waves of change hitting the economy during the decade were characterized by unprecedented volatility: movements on interest rates from 10 to 60 percentage points; exchange rates that yielded 15-50% depreciation of the local currency, inflation rates that varied from 10% to more than 20%. During August and September 1998 alone interest rates rose from 17-39% and then returned to normal levels. However short-lived, that episode caused financial institutions to make radical decisions related to their consumer credit portfolios damaging their long-term value and development.
Given concern for the impacts of such changes on credit quality, participants in the industry (then dominated by banks) invested in predictive technology during the 1990's in order to better manage their consumer portfolios. These investments focused on account level predictive technology for consumer portfolio management. Banks adopted world-class (at the time) technologies for establishing risk-based credit decision processes and profitability-based market segmentation. Such technologies have proven robust in rank ordering projected consumer-level risk. But projections are subject to considerable drift because they are hardwired to the historical environment. Since they do not specifically quantify the impact of exogenous impacts in the history, they implicitly project that the environment of the past will repeat in the future. In periods of change in the environment, this reduces the utility of the tools significantly.
Experience with such tools during those tumultuous years taught two important lessons: Success with predictive models used in markets experiencing economic change depends largely on the ability of the portfolio manager to arrive at and implement the right adjustments—major or minor. And, the policy changes and adjustments need to vary in size and direction as swings of the economy occur. Unfortunately, the adjustments tend to lag in time—depending on how rapidly the portfolio manager is able to identify the direction and magnitude of the swing. One is left to employ too much art and not enough science—often late in the game.
Ineffective adjustments can adversely affect customer relationships, leaving little hope of reconstructing them later. To return to our analogy, the ship may not sink completely, but it may sustain long-lasting and expensive damage if it is steered incorrectly. Indeed, this is precisely what occurred in the consumer credit market in Mexico in the 1990's. With this experience in mind, it is now time to invest in strategic decisioning platforms that reflect and address these issues.
Portfolio management decisions taken to cope with the 1994-1995 Crisis in Mexico would have been approached very differently if the analytical tools capable of distinguishing intrinsic behavior from exogenous effects had been available. The ships would have been steered more skillfully in the face of the heavy seas. A few examples:
• Identification of portfolio segments projected to deteriorate no matter what restructuring efforts took place would have enhanced the effectiveness of exit strategies from the outset. Segments showing high sensitivity to exogenous impacts and more certain deterioration in performance could have received the focus of targeted policy changes. On the other hand, identification of segments projected to continue to perform well regardless of economic swings, even large ones, would have avoided the damage to relationships with credit card customers inappropriately affected by credit line decreases and blocked authorizations—sometimes implemented even for current accounts.
- Segmentation based on lifetime value as opposed to short-term risk and profitability would have yielded major benefits. In particular, such criteria would identify segments capable of paying for the short term decreases in revenue or marginal increases in losses given a long term view of the relationship modeled through various future economic scenarios.
- Model calibration and redesign would have been less intense given that much of the calibration required accounted for exogenous impacts—impacts that vary over time and occur more sporadically than maturation (intrinsic portfolio change) effects.
- Policy establishment would have been less erratic in reaction to increasing uncertainty. The policies would have recognized credit quality and profitability generation capability in a range of potential scenarios.
After the crisis ended in 1996/1997, important questions arose: Is this value now seen coming from good economic times or from intrinsically good portfolio constitution? Is it time to extend more credit (within projected loss constraints) or is it better to be cautious? The critical (but unaddressed) need was for reliable measures of segment-level sensitivity to exogenous impacts. Large up-side opportunities could have been realized if such measures had been available.
The post-crisis environment brought many significant market developments—burgeoning consumer loan portfolio purchase and sales, as well as mortgage portfolio valuations and investments by foreign banks. All would have benefited from the capability to distinguish a portfolio's intrinsic value from the potential generation or destruction of value coming from the environment.
In the later part of that decade, newer industry participants increased their presence as well, most notably finance companies (including retailers and non-bank mortgage lenders (i.e. the sofoles), and the captive auto finance companies. These companies in general have not incorporated the tools to help cope with economic change. Their portfolios are increasing in size and becoming more complex to manage, reflecting new products and customer expectations—are reshaping the use of credit and the performance of existing portfolios.
The Mexican consumer credit industry has always coped through difficult times even when measured with international yardsticks. The financial community requires world-class standards of evaluation in order to maintain access capital and debt markets. Today the rating agencies, financial markets and potential investors find an industry that has rapidly adopted all the standard U.S.-developed technologies and processes. For example, most Mexican credit card issuers use the major credit card processors, as well as more general methods offered by U.S. risk management and general consulting companies. The financial arms of U.S. automakers have established operations in Mexico similar to those they have in the States, bringing their traditional account-level approaches with them.
Changes in account-level approaches are now under way. Spanish global banks made several Mexican acquisitions in 2000/2001, and have quickly switched from the traditional US-based analytical methods to a more intuitive day-to-day approach. Before acquisition, Mexican banks would pay for sophisticated account level tools that, given the size of the market, were difficult to cost-justify. Now they consider block level segmentation for differentiated treatment adequate. As a result, there is less perceived need for the account level differentiation treatment—it is too expensive when a strategic block segmentation based on value is in place.
Mexican finance companies (sofoles, retailers) have not experienced the major economic swings because most of them flourished after 1995 when banks were building reserves for losses and recovering from the Crisis. As a consequence, today's finance companies need to prepare to cope with difficult times. With their huge portfolios, they are primed to establish the analytical platforms necessary to better understand the impacts on their consumer and mortgage portfolios, and achieve the world-class standards required of them to access the international debt markets.
In 2001 the seas are much calmer, but not without waves capable of generating significant drift and dislocation. Macroeconomic indicators are now stable in Mexico with single digit interest rates and stable foreign exchange rates. However, the market is feeling the effects of the U.S. economic downturn, since international trade depends almost 90% on exports to the U.S. Concerns of portfolio managers have turned to growth in uncertain times. They may have experience in keeping a steady hand on the helm through negative economic conditions, but this does not assure their success unless new tools for portfolio navigation are deployed. Without these tools, precise forecasts remain elusive, because the history of the portfolio is loaded with a combination of negative exogenous effects and recent optimistic numbers resulting from good economic times.
While another economic storm of the scale experienced in 1995 is considered a remote possibility, Mexico is not insured for environmental effects that may partially destroy the value of existing consumer portfolios. Now is the time to learn from past experience. We must study the recent history available using more sophisticated approaches that distinguish and quantify intrinsic portfolio effects and the exogenous impacts in a consumer credit portfolio. Understanding with new clarity our position, our direction, and the forces affecting our movement, we will be much better prepared to weather the storms, however severe, as we manage credit portfolios into the future.
Alejandra Martinez is a pioneer in the use of risk, profitability and value forecasting technologies for consumer credit portfolios in Mexico, successfully customizing and incorporating them into management decision practices. She holds an MBA from Stanford University and a BS in Industrial Engineering from Universidad Iberoamericana. Currently, she is CEO of Analyse, S.A. de C.V., a consulting company focused on consumer portfolios in Mexico. |
Top of page