Credit scores might seem an odd relation to libertarianism at first glance, but if you can hang with me through this article, I will explain why they tie in and are a very applicable microcosm of libertarianism applied in practice.
The world economy relies heavily on credit scores. They are used as a thumbnail determinant for everything from creditworthiness to trustworthiness to character flaws. Each of us who participates in business in the heaviest parts of the world economy carries a credit score of some sort with us throughout our lifetime, though it constantly changes. Statisticians work hard to come up with ways to make the scores more accurate and better predictors of human behavior.
The most common misconception about credit scores is that they are a way of determining the future likelihood that an individual will perform in a non-desirable manner. In most cases, that means failure to repay a debt. In some cases (where it might be used in employment) it might be the likelihood of theft or mismanagement. The problem is that these scores are relatively good determinants of a population of people, but are a very poor determinant of an individual outcomes. In other words, they aren’t designed to tell a business if a single person is likely to act in a undesirable fashion, but rather if a high percentage within a given group of people would be likely to do so.
Let’s say we have a person with a house they used to live in but are now using as rental property. They are trying to determine a potential renter’s likelihood to continue rent payments. Since we aren’t really looking for whether a population of people are likely to continue rent payment and we are only looking at whether this one person is likely to repay, a credit score is not a very good indicator. References, income, work stability, and a personal interview would be better. If we have a large company with thousands of apartment units, would a credit score be a good indicator for potential renters? I would say both yes and no.
Credit scores have become somewhat accurate over the years at being predictive of populations, but they are far from being without biases and flaws. One problem is something that the gambling industry relies upon to make a lot of money. If you’ve ever played roulette, you might notice that oftentimes there is a electronic display next to the table that shows several of the last outcomes from previous spins. For example, I might notice that the last 10 spins have resulted in a black number. So, it might be assumed that chances are greater that it will land on red next. However, it’s really just an illusion. Because we aren’t gambling on whether there will be a red result in a series of outcomes, we are rather gambling that the single spin will land on a red, so the chances are the same as any other spin.
The same applies to a potential renter, debtor, or employee. We aren’t really trying to predict whether a population of people will result poorly, but we are actually trying to predict whether an individual person will be a good bet. As I said, a credit score can often predict that if we give loans to a population of people with the same credit score that the vast majority of predictions will be true, there’s going to be a significant margin of error, partially based on the fact that each individual outcome is not directly related to the whole.
In other words, we can think of each individual decision as a single outcome as easily as we can an outcome across a group of people. It’s just that if we view each person as an individual, we can begin to assess the sorts of factors that can’t be taken into account by a credit score. We can assess an individual’s reasons for having poor credit. Maybe there are legitimate excuses or maybe not. We can assess the story provided by a credit history rather than the score of that history. We might see, for example, that this person was paying for a loan on a television when they weren’t paying their credit cards. Maybe we are also making a loan on a television and should take that into account. We can see the story of the person as a whole rather than an assumption based on a score across a population.
As an example of this, I’ll defer to some personal experience. I once established a small finance company that made loans on merchandise purchased at a retailer that had been rejected by every other possible option for a lender. Everyone had a bad credit score, so I relied on other factors to determine whether to make a loan and how to structure it. The bad debt experienced on the portfolio was lower than the average bad debt by any of the lenders who had rejected these very customers. The entire portfolio was performing better than a portfolio of people who all had above average credit scores. By thinking of each loan decision as an individual transaction with an individual person, things were handled in such a way as to produce different results.
How does any of this apply to libertarianism?
Firstly, because credit scores have been either directly or indirectly affected artificially by the influence of government. The government demands that credit guidelines and scoring must not be biased against any group of people based on sex, race, religion, age, income level, class, etc. This noble goal has had an ill effect. In seeking to unbias a process, we have added biases to correct for it. Adding bias for bias doesn’t typically help statistically.
Is it possible that information about people can be biased through credit history? Absolutely, but we can’t really fix those biases by adding more. We only compound the amount of bias we have. The vast majority of people don’t want these kinds of biases present in such important decisions, but government manipulation of how the data predicts future outcomes isn’t helping in that it doesn’t allow the statistics to work properly.
The incredible popularity and ubiquitous use of credit scores is primarily the result of government pressure on people to illustrate that their decisions are unbiased by using a score that is believed to hold no bias. If you don’t make a loan to a minority or don’t rent to a minority, you can point to a score rather than your own individual perceptions about the individual, thus saving yourself from potential fines and/or other financial exposures.
Do people have potential bad biases, such as racism, on their own when making these decisions? Of course they do, but probably less so than a score that is not designed to predict the outcome of an individual transaction. We might not be so heavily reliant on credit scores (and the unfairness of using them) if it weren’t for government regulations insisting that we be able to illustrate non-bias.
As libertarians, we believe in treating people as individuals rather than collectives. Thinking of people as individuals is a much better determinant of potential future outcomes and behaviors. It doesn’t matter whether you are trying to plan a society or trying to make good credit decisions. Statistics aren’t designed to view people as individuals, and making decisions not based on people as individuals is bad for business. While statistics might be useful in explaining collective behavior, they aren’t very useful in predicting individual outcomes with individual information and variables. Statistics only work well with good data. The best data can only be gathered by looking into individual circumstances. That can’t be determined well through a credit score.
I would never argue that there should be a ban on credit scores. Instead, I would argue that there should be a ban on government involvement in private transactions. There will always be negative bias toward certain groups of people in ways that are not desirable, but it is a question of whether we want a process that we know sometimes fails or do we want a process we know will always fail. Government involvement guarantees bias in credit scores. Individuals making decisions about individuals most typically has more positive outcomes with less bias. Instinctively, most people know that making assumptions about individuals based on a collective that individual might be a part of is inherently wrong. It makes a lot more sense to make decisions on an individual basis, and that will always produce better results overall. By their nature, credit scores or other statistical tools are very bad at predicting what sort of people individuals are.