Why, at a time when Mark Zuckerberg is being pilloried by members of Congress for putting his company’s interests ahead of his users, are policymakers creating an opening for financial companies to try the same?
People frame the question in a zero-sum context: either we allow companies to innovate in the spirit of bringing new solutions to the marketplace, or we handcuff progress in the name of protecting the rights of consumers.
The idea of a regulatory sandbox, currently being applied in DC, sides with the proponents of innovation. Unfortunately, they see the question as one or the other. Their support of innovation is limited entirely to reducing protections. On other approaches, they are silent. The best tools in economic development - a subsidized loan fund, workforce training, or technical advice – are missing. The answer, it seems, is to take from consumers to give to businesses.
I am not sure why we need government to do that, and equally, I do not understand why policymakers would perceive that the public benefits when government eliminates protections.
Sometimes a company with the best of intentions unwittingly creates a product with side effects that harm people. In my opinion, our technology community places too much faith in the power of data to enhance our communities. The truth is more complicated. Data can be a force for good, but without considering the context of where it comes from and how it will be applied, it is neither good nor bad.
“The real lesson we learn over and over again in banking, wrote Karen Shaw Petrou in an American Banker editorial is that retroactive consumer protection leaves a lot of badly hurt, vulnerable households in the ditch. It needs to be thought of now before these products become even larger and more dominant in the financial system.”
About ten years ago, during a trip to the Bay Area, I had the opportunity to share a meal with a group of youngish start-up types. Most worked at a new company that wanted to aggregate data on local schools and then white label it to consumer-facing real estate brokers. The idea was to make a one-stop method of grading school quality so that people could quickly evaluate the quality of nearby schools when they were contemplating the purchase or rental of a home.
They succeeded in their goal. The company built a system that could capture education data, feed it into an algorithm, and produce a pure 1-10 numerical grade for any school. There was no human judgment – it was an entirely objective process.
Unfortunately, they created a system whose outputs correlated strongly with local socioeconomics. Statistics on student performance correlate highly with the socioeconomic status of the pupils. In the abstract to its hosted discussion of the question, the American Psychological Association comments:
Research indicates that children from low-SES households and communities develop academic skills slower than children from higher SES groups. For instance, low SES in childhood is related to poor cognitive development, language, memory, socioemotional processing, and consequently poor income and health in adulthood. The school systems in low-SES communities are often under-resourced, negatively affecting students’ academic progress and outcomes. Inadequate education and increased dropout rates affect children’s academic achievement, perpetuating the low-SES status of the community.
By blindly collecting inputs out of context and then amplifying them to realtors, an effort at informing consumers had the unintended effect of reinforcing neighborhood inequality.
The use of data can sharpen our thinking, clarify our decision-making, and expose close-mindedness for its shortcomings. Nonetheless, everyone should be at least partially skeptical of data’s benefits, because all data is vulnerable. The garbage-in garbage-out hypothesis, whose expression originated in England in the 19th century, remains as accurate today as it was then.
The Fintechs won’t be the first players in banking who make a mistake in how they use data to make decisions. We have fair lending laws because many banks refused to extend credit to applicants from protected classes for reasons that had nothing to do with their credit-worthiness. Red-lining, the practice of drawing red lines across maps to indicate where loans could and could not be made, was initially an opaque means to limit mortgage lending to “desirable” (not black) neighborhoods in cities in the 30s. We addressed those mistakes, but it took time. Can we be more responsive this time? I believe we must, if only because the pace of change is far greater this time.