Most Used Payment
Evidence of substitution from credit cards to debit cards for payments is further supported when we examine the payment instrument individuals cite as the one most frequently used at the point of sale. In Table 6 we present results from logistic regressions with robust standard errors of the following form: MostF requentij = β0 + β1RevolvingBalancesi + γxi + i . (3) MostF requentij is a dummy variable equal to 1 if consumer i reports using payment j most frequently. ( credit card)
As before, RevolvingBalancesi is a binary variable indicating whether an individual regularly carries a credit card balance, and xi is a vector of socio-demographic and other characteristics of consumer i. Extended results of these regressions are displayed in Table 7. The results in Table 6 largely confirm the evidence presented in Table 4. Individuals with revolving balances are significantly less likely to cite credit card as the payment instrument most frequently used at the point of sale, controlling for socio-demographic characteristics and rewards program participation. ( credit card)
The calculated odds ratio indicates that revolvers are around half as likely as convenience users to use credit cards most frequently. Individuals with revolving balances are, however, significantly more likely than convenience users to cite debit card as the payment instrument most frequently used. Calculated odds ratios indicate that revolvers are nearly one-and-a-half times more likely to use debit cards most frequently. ( credit card)
There is no apparent difference between the two groups in the likelihood of citing cash as the most frequently used payment instrument. Revolvers are more likely to cite check as their most frequently used 12 payment instrument, although this effect becomes insignificant after controlling for credit card and debit card rewards. ( credit card)
Credit scores Throughout this report, we refer to consumer credit scores. Lenders use these scores to predict a consumer’s relative likelihood of repaying a debt compared to other consumers. Credit scores are used by most credit card issuers to determine consumers’ eligibility for credit and to set pricing for the credit lines they offer.
Data we rely on in this report utilize widely-used, commercially-available credit scores. There are two important limitations to the way we use credit scores in this report. Different credit score models, while fundamentally similar, may include or exclude different data points or weight them differently. This means, first, that we are aggregating data on the basis of credit score even though not all consumer credit scores are computed using identical methodologies. Second, it means that, when reporting certain metrics over longer time horizons, the introduction of new models and changes in models complicates comparisons between different points in time.
In some cases, one or both of those two issues could change which “credit score tier” a certain account or consumer falls in. (We explain the credit score tiers that we use in this report below.) The Bureau believes that different credit scoring methodologies, over the time 20 See Consumer Fin. Prot. Bureau, Monthly Complaint Report Vol. 25, at 16 (July 2017), available at https://www.consumerfinance.gov/data-research/research-reports/monthly-complaint-report-vol-25. 23
CONSUMER FINANCIAL PROTECTION BUREAU — CONSUMER CREDIT CARD MARKET REPORT
periods and set of market participants we examine in this report, are sufficiently consistent that it remains informative and useful to report aggregate results and changes over time by credit score. We nevertheless proceed with caution when assigning precision, beyond a reasonable degree, to certain results. When reporting results by credit score in this report, we group scores into five tiers. This fivetier grouping differs slightly from the four-tier grouping the Bureau used in past reports, but it aligns with the groupings used in the Bureau’s more recent Consumer Credit Trends reporting.21 We define the ranges encompassed by the five tiers, as well as each tier’s share of the creditscored population and the cardholder population, in Table 1 below. The only substantive difference between these tiers and those we relied on in our past reports is that we now subdivide consumers with scores below 620, which we previously referred to as “deep subprime” scores, into two tiers. These we refer to as “deep subprime” and “subprime,” respectively. Nonsubstantively, the range we referred to previously as “core subprime” we now refer to as “nearprime.” The “prime” and “superprime” ranges are unchanged from our previous reports. Throughout this report, we refer to scores below 660 collectively as “lower credit scores.” We intend this change in credit score tiering to allow us to examine outcomes for consumers with lower credit scores in more detail. 21 See Consumer Fin. Prot. Bureau, Credit Cards: Data Snapshot for September 2017, (Sept. 2017), available at https://www.consumerfinance.gov/data-research/consumer-credit-trends/credit-cards. 24
CONSUMER FINANCIAL PROTECTION BUREAU — CONSUMER CREDIT CARD MARKET REPORT TABLE 1: CREDIT SCORE RANGE SHARES
AS OF Q2 2017 (CCP) Credit score tiers % of U.S. population with a credit score % of U.S. scored credit cardholding population Superprime (scores of 720 or greater) 56% 62% Prime (scores from 660 to 719) 17% 17% Near-prime (scores from 620 to 659) 9% 9% Subprime (scores from 580 to 619) 7% 6% Deep subprime (scores of 579 or less) 11% 6% Credit scores in the CCP and Y-14 are refreshed regularly. When not otherwise specified, refreshed scores are used throughout this report to assign accounts and consumers into credit score tiers. As a result, when analyzing trends over time within a particular credit score tier, we are not analyzing a constant set of accounts or consumers, but rather accounts or consumers which fall into that tier at each point in time under analysis, unless otherwise specified. This clarification is especially important to note in light of the substantial propensity of consumers to experience changes in their credit score that are large enough to move them from one credit tier to another. 22 Just as individual consumers’ scores can fluctuate over time, the distribution of consumers (including both cardholders and non-cardholders) in aggregate across these tiers is not static. We touched on this briefly in our 2015 Report, but it is worth elaborating in greater detail here.23 The share of consumers with a credit score represented by each of the various tiers has fluctuated over the past decade by as much as four percentage points, as shown in Figures 1 and 2 below. (Figure 2 shows identical results to Figure 1, but highlights the section of the graph between 50% and 100% to better illustrate fluctuations.)