avoid alternative facts about alternative data

Avoid alternative facts about alternative data

Date: June 24, 2020

Dear Colleague:

There’s quite a lot of misinformation and misunderstanding surrounding “alternative data” with respect to credit scoring and consumer credit access. So much so, that the Consumer Financial Protection Bureau (CFPB) recently issued an open request for information on the impact of such data on “credit invisibles”—consumers whose credit histories are insufficient to obtain credit scores when using conventional credit scoring models.

In support of our goal of accurately scoring as many consumers as possible, VantageScore Solutions has embraced “alternative data” from the very beginning, and I feel qualified to offer some thoughts on the matter.

First of all, what is “alternative data?” Most definitions characterize alternative data as data other than the debt-payment records lenders report to the national credit reporting companies (CRCs). Other definitions consider alternative data to be any data that resides outside of a consumer’s credit file.

For the purposes of this article, let’s use the former definition and level-set a number of distinctions.

Rent, cellphone, and utility payments are good examples of alternative payment data under our definition. What’s more, they are highly predictive, and credit scores generated using alternative data are highly predictive. As revealed in a pair of Experian studies, alternative data can have a significant positive impact on consumer credit scores and access to credit. The first study, Let There Be Light (2015), examined a sample of Experian credit files that contained positive utility-payment data, while the second, Let the Light Shine Down (2016), looked at the score impacts of positive-payment data on customers at a large Northeastern energy utility. The studies found that a vast majority of consumers (97 percent in the 2015 paper and 96 percent in the 2016 study) saw increases in their VantageScore 3.0 credit scores thanks to positive utility-payment data.

In addition, substantial numbers of consumers (nine percent in the 2015 study and seven percent in the 2016 edition) increased their credit-file “thickness”—migrating from no-hit (zero trades on their credit files) to thin-file (four trades or fewer), or from thin-file to thick-file (five or more trades) when rent- and utility-payment data were taken into account. Greater file thickness allows consumers broader access to credit, and potentially better borrowing terms.

VantageScore 3.0, like all preceding VantageScore models, considers positive-payment activity, such as on-time payments of rent, phone and utility bills, whenever that data is available in a consumer’s credit file. (VantageScore models do not pull in this data, or any other data, that resides outside of consumer credit files.) VantageScore was the first to include this data in a generic credit scoring model, and our established rival eventually followed suit in its later models.

We believe our models’ consideration of this more comprehensive data corrects a historic imbalance, under which consumer credit scores could be penalized for negative utility-payment behavior but received no credit for positive rent or utility payments: Years, or even decades, of steady, timely payments yielded no benefit, but if an unpaid rent or utility payment were sent to collections, that event would appear as a derogatory entry on the consumer credit file, with a corresponding reduction in credit score.

So how much of this data is currently impacting consumers? Admittedly, not much.

Relatively few landlords and utilities currently report payment data to the national CRCs, but we applaud forward-thinking organizations such as the Policy and Economic Research Council (PERC), which advocate for more widespread reporting of these data. We support efforts to report a greater amount of that alternative data to the national CRCs, and believe doing so will expand credit access for many deserving U.S. consumers.

VantageScore has been committed from the start to the use of alternative data found within consumers’ credit files, which is subject to thorough data accountability standards under the federal Fair Credit Reporting Act (FCRA). Because our models accept this alternative data, some refer to them as “alternative models.” Others refer to VantageScore models as “alternative models” in the context of being an “alternative” to another major brand. We prefer to think of VantageScore models as competitive models that incorporate comprehensive, regulated, FCRA-compliant credit bureau data.

All the best,

Barrett Burns