Gustavo Lasala, MBA ’04, created a credit-scoring system that transformed an industry. His work is helping overlooked borrowers realize their dreams—one microloan at a time.
By Dashka Slater
Photographs by Ryan Heffernan
Sitting in the offices of Acción Texas, a San Antonio–based microlender, Noelia Gutierrez posed a familiar problem: She had a nice face, a good story and a lousy credit history. The daughter of migrant field workers, she’d come to the United States with her family at age 4 and had spent her childhood alternating between school during the week and work on the weekends, sometimes laboring in the fields, sometimes selling tamales with her mother. At 14, under the tutelage of a neighbor who was a notary public, she began helping other farmworkers with their tax returns, hoping to build the skills to have a different kind of life than the one her parents led. Now, she was hoping for a $2,000 loan so she could buy a computer for her expanding business, Electronic Tax Refund.
Unfortunately, Gutierrez’s credit history was a mess. But she blamed her ex-husband for her low credit score, and her commitment to her flourishing business indicated the kind of discipline that leads to prompt repayment. Was that enough to go on? On the one hand, the mission of Acción Texas is to serve people like Gutierrez—the 21 million Americans who rely on dodgy credit sources like payday lenders and pawn shops, and the nine million people without a bank account. On the other, if Gutierrez defaulted, that would mean 2,000 fewer dollars available to other worthy loan applicants and a $2,000 loss to explain to the organization’s funders.
Microfinance organizations around the country grapple with the problem posed by applicants like Gutierrez every day. The sums of money involved are generally small—loans range from $500 to $150,000, with most at the low end of the range. But the amount of available information is often small, as well. The people who turn to microlenders have often spent most of their lives in the cash economy. Figuring out whether an applicant is likely to repay a loan can take weeks, and the risk of guessing wrong is high.
In 2003, a 32-year-old MBA student named Gustavo Lasala went to work as a Citibank Community Summer Intern at Acción Texas and ended up finding a way to reduce the risk of a wrong guess, while shortening the time required to make a decision. At the time, Lasala’s plans for the future had absolutely nothing to do with microfinance. A native of Uruguay, he came to McCombs planning to work in international commerce in his home country. He applied for the internship at Acción Texas solely because he was interested in getting experience in emerging markets and figured microfinance fit the bill.
But while Lasala would turn out to have a knack for predicting the future, in this case his predictive abilities failed him. He didn’t go back to Uruguay, nor did he end up working in international commerce. Instead, he developed the nation’s first microlending credit scoring tool, an innovation that would allow Acción Texas (now Acción Texas-Louisiana) to make quick and accurate decisions when presented with a case like that of Noelia Gutierrez. In the process, he would catapult Acción Texas to the top of the microfinance world, making it the largest and best-performing microfinance institution in the U.S.
Cracking the Code
Gutierrez ended up getting the loan from Acción Texas, repaid it and then came back for another, which she used to move her expanding business into larger quarters. Success stories like hers are, to Lasala’s way of thinking, the reason microlenders need ways to make accurate predictions about a borrower’s probability of repayment.
“The mission of most of these [microlending] organizations is to put the money in the hands of those that need it, not to burn it in the process,” Lasala explains. “If they achieve a positive result from their operations, these organizations fulfill their mission and they will be around tomorrow to help someone else.”
With his large blue eyes and baby doll eyelashes, Lasala looks younger than his 39 years. He is intense and charming, with an air of seriousness that nearly everyone he works with remarks upon. His father was the chief financial officer for the Banco República, the largest bank in Uruguay, and Lasala grew up steeped in the world of finance, convinced that commerce is the engine of progress. He started out working for Hydro Agri, the world’s largest producer of mineral fertilizers, and then went on to form his own consulting firm. But he soon decided he needed an MBA to compete in the global arena, so he came to the U.S. with his wife in 2002, planning to return to Uruguay once he’d secured his degree.
The fact that Lasala ended up working for an organization where the average loan is $17,000 rather than presiding over multi-million-dollar deals was pure chance.
“I could have gone to any other microfinance institution and I wouldn’t have ended up working in microfinance,” he says. But at Acción Texas all the elements were in place—a problem that needed solving, an intern with an idea how to solve it and management that was willing to commit the resources necessary to try something unorthodox.
At the time of Lasala’s internship in 2003, Acción Texas was 9 years old and already a dominant player in microfinance, with eight statewide offices and an $8.2 million portfolio that was growing 20 percent each year. But growth brought its own problems. Nearly a quarter of its portfolio was in loans that either defaulted or saw payments go more than 31 days past due, and expenses were growing faster than revenues. Janie Barrera, Acción Texas’ founding CEO, believed that if her organization could balance growth with stability, it could reach something no U.S. microfinance organization had yet achieved: self-sufficiency.
“The way you become self-sufficient is to scale—keep your costs low and make as many loans as possible,” Barrera explains. “That’s my dream. So we had to have some way to streamline.”
When Lasala began working as an intern, he heard from the organization’s staff that they needed a quicker way to screen applicants. The Fair Isaac Co., or FICO, credit score used by banks was of no use— Acción’s clients had an average credit score of 575, well below the 680 threshold for borrowing at most banks. Still, Barrera had long had an interest in finding some other way of predicting which applicants were likely to repay their loans. She had even tried to find someone to build a predictive model but had been told it wasn’t possible. So for now Acción Texas underwriters used a grid of yes or no questions to make their decisions. It was cumbersome and time consuming and, most importantly, it wasn’t giving them the answers they needed.
As it happened, Lasala had just built a credit scoring tool in a data mining class he’d taken from Professor Maytal Saar-Tsechansky at McCombs. Data mining uses evidence from the past to predict the future, extracting and analyzing information to find patterns that can be used to make decisions. In its nine-year history, Acción had gathered lots of data, mostly for the purpose of reporting outcomes to its funders. When Lasala saw the cache of available information, he wondered if he could use it to build a credit scoring tool for microfinance.
“Nobody had done this,” explains Saar-Tsechansky. “The little institutions that were handling microfinancing didn’t have the resources, didn’t have the human capital, didn’t have the expertise.”
There was another factor as well. “The whole idea of estimating risk is a business angle,” Saar-Tsechansky says. In her nine years teaching at McCombs, she could only recall one other student who applied her modeling techniques outside of the corporate world—an Army major who helped build a model for predicting the location of roadside bombs in Iraq. In the nonprofit sector, there was a concern that being data-driven might mean being too hard-nosed, too focused on the bottom line.
“The argument was, ‘Why use a credit scoring tool for our clients? Then we will become a bank and we will not do what we are supposed to do,’” Lasala explains.
As Lasala sorted through the data from a sample of 500 loans, he tracked 35 different criteria to see which ones correlated with default. He was looking for factors that might indicate stability and character, factors similar to the commonly used indicator of how long a person has lived at their current address. Fewer than 10 of those criteria ended up being useful (Lasala won’t say which ones), but they did the job. When he tested the model to see if it accurately predicted the outcomes for real clients, he found that it did.
“It was a very nice feeling,” he recalls with characteristic understatement, “to find that I cracked the code.”
At that point he had a month left in his internship. Barrera assigned him four interns, working in shifts to pull files from the cabinets and load the data into spreadsheets that would feed the model. Lasala often worked from 9 a.m. to 2 a.m. refining it. By the time he had finished his internship, he told Barrera that the model could cut Acción Texas’ bad loans in half, increase net income by 15 percent and increase the organization’s capacity by 50 percent. Barrera’s parting words to Lasala were, “Make sure you come back to me when you graduate.”
He did, becoming chief financial officer for Acción Texas the following spring. Within weeks, in the summer of 2004, the scoring engine, based on data from about 1,000 loans, was ready to be deployed.
Scaling the Solution
Enter an applicant’s data and the model spits out a scorecard that indicates whether or not an applicant qualifies, or advises the underwriter to investigate further. While it used to take a week to 10 days to give an applicant preliminary approval, it now takes 48 hours.
“That allowed our underwriters to do more in less time, without losing the touch, the eye contact with the person, without losing the human making the final decision,” Lasala explains. “And that was the strongest case for using a scoring tool—you can do more with less time and you can make better decisions. It’s not what side of the bed the underwriter woke up on. It gave Acción Texas a huge competitive advantage.”
In 2008, Acción Texas began offering that advantage to other microlenders, helping other organizations achieve economies of scale while diversifying their own income stream. Microloan Management Services™ is now a complete microfinance service center offering a menu that includes its proprietary credit scoring tool, as well as underwriting, closing documents and portfolio management to 11 nonprofit finance institutions around the country.
“Instead of us trying to have our own sophistication, we bought their sophistication,” explains Barbara Eckblad, director of lending at the Wisconsin Women’s Business Initiative Corp. WWBIC is a classic microlender, with a $6 million portfolio and an average loan size of $28,000. In late 2009, when the organization began working with MMS, they were grappling with a surprising number of defaults.
“We really couldn’t know where the next defaults were coming from,” Eckblad recalls. “There were a lot of surprises—businesses that were paying, paying, paying and then suddenly were closing their doors.”
They needed better predictive abilities, and they needed to become more disciplined in their underwriting. MMS allowed them to achieve both goals.
“Working within the parameters of MMS, it’s inherently more disciplined,” Eckblad says. “We do at times override MMS, but it’s a big deal when we do it, and it’s very conscious.” Defaults are down, and many of those businesses that do close their doors still repay their loans.
By January 2008, Acción’s portfolio had doubled, from $8.2 million to $17 million, and the percentage of its loans that were delinquent or in default had fallen below 10 percent. This success attracted the attention of Citigroup, which committed to buying up to $30 million worth of loans, while paying the nonprofit to service them. That, in turn, provided Acción with more lending capital. Now the largest and best-performing microlender in the U.S., Acción Texas-Louisiana has opened four Louisiana offices since 2009 and is considering moving into other states in 2011.
But the organization still hasn’t reached self-sufficiency—income generated from its portfolio covers around half of its expenses, while the rest comes from grants. But Barrera believes it can be done. “If we can keep our costs down and grow our portfolio another $15 million, we could reach there,” she says.
Yet many remain skeptical that microfinance can ever achieve scale in the United States.
Part of the problem is mission confusion—microlenders themselves are often unsure if they are a business or a charity. But there are other limitations, as well. In the developing world, roughly half of the population is self-employed. In the U.S., the figure is just 10 percent. Overhead costs are higher here, and anti-usury laws limit the interest that can be charged even to the riskiest borrowers. (Acción Texas-Louisiana charges a 15 percent annual percentage rate, or APR.)
Still, Lasala believes the commercial model of microfinancing serves a critical role.
“The alternative is to depend on grants, and grants are getting more competitive every day,” he observes. There are 25 million small businesses in the U.S., and he calculates that if even 3 percent of those need a $7,500 loan, the potential microlending market is $5 billion.
Lasala’s interest in seeing a microlender achieve self-sufficiency has led him to the for-profit sector. Last summer, he left Acción to become director of retail operations at Progreso Financiero, a Mountain View, Calif.–startup that is positioned to be the nation’s first microlender to go public.
“You can feel the entrepreneurial spirit here,” he said last October, standing on the company’s humming second floor, surrounded by employees typing furiously on laptops in row after row of glass-topped tables. The company, founded in 2005, has 210 employees and no conference room—staff members meet at a local coffee shop.
“What’s interesting about Progreso is that we serve a client that is much lower on the financial ladder,” he explained over a mocha and an almond croissant. “Acción’s average loan size was $17,000, and the main purpose was to build or expand a business. The average Progreso loan is $900 or $1,000, and the vast majority of our clients want that loan to build their credit. These are people who are totally out of the system. Everything is more difficult and more expensive because they don’t have a credit history. We are the door through which they enter financial inclusion.”
Progreso Financiero’s target market is low-income Latinos in California and Texas. This is the population that typically turns to payday lenders, which charge astronomical fees for short-term loans.
“Payday lenders don’t deliver financial freedom,” Lasala says. “They don’t record your history of payments, so you don’t build a credit history.”
Progreso’s model is to rent space in pharmacy and grocery chains with large numbers of Latino customers and set up a booth with a couple of computers and three agents to take payments and provide loan services. The company charges 36 percent APR for its loans.
“Our interest rate is expensive for me, for you,” Lasala says. “But it’s not expensive when you consider the 400 percent charged by payday lenders. And it also reflects the risk we incur.”
Progreso is expanding rapidly, having already loaned out more than $100 million through 43 stores in California and seven in Texas as of the end of 2010. (In December 2010 alone, the company closed more than 10,000 loans—all to individual consumers. By comparison, Acción closed roughly that many loans—mostly to small businesses—in seven years.) Lasala is helping lead that expansion, managing the relationship with pharmacy and grocery chains while also working as a liaison with community development groups that can send them customers who might otherwise turn to payday lenders.
“Gustavo brings credibility,” says James Gutierrez, Progreso’s CEO. “He has a great reputation with a lot of the nonprofits and community development institutions, and we thought our system—[in a way that’s] similar to MMS—could be exported to other organizations, as well.”
While energized by his new position, Lasala hasn’t entirely turned away from the nonprofit sector, having recently been elected to the board of directors of Foundation for Women, a small San Diego–based microfinance organization. And he continues to watch MMS from afar, likening his role to that of a proud parent whose baby can now walk on its own.
For Saar-Tsechansky, who has watched Lasala with a similar kind of pride, his story proves the value of setting out to make a difference, instead of setting out to make a buck: “He had an opportunity to contribute to an important change in people’s lives and to change an entire industry,” she says. “This is something very few graduates, particularly in the first few years after business school, have an opportunity to do.”
Microlenders play a key role for traditionally overlooked borrowers, providing access to money but also a sense of financial responsibility, both of which are generally unavailable from banks or payday lenders. “The product of a microlender is not just the loan,” Lasala explains, it’s an entire package of services, which may include basic business education and more frequent client meetings to help promote good “financial behaviors.”