Round Number Preferences and Left-Digit Bias: Evidence from Credit Card Repayments

We examine round number preferences and left-digit bias in credit card repayments. We show half of all manual credit card payments are at a small set of round-number values, although all values between the statement minimum and balance were permissible. We use minimum payments, which place a floor on the payment value, to measure the strength of rounding behavior in a natural experiment. We find that when the minimum payment level just rules out a round number from the bottom of the range of possible repayment (e.g., a minimum payment of £50.01 just rules out a payment of £50.00), payments at the next round number jump by 15%–20% (e.g., payments at £60.00). That is, in response to a smooth increase in minimum payments, the selected payments jump up a ladder of round numbers when each lower rung is removed. We also see a classic—although nuanced—left digit bias. When the balance increases such that the left digit changes (e.g., from £1,999 to £2,000), payments in full drop by 3%–8% as if an increase in the left digit makes the balance seem much larger. But average repayments increase by 10%–15%, as smaller partial repayments are all dragged up the ladder rungs. This preference for a small set of round numbers as payment values and this effect of left digit bias on the level of repayment have policy implications. Rounding could be used by policy makers to encourage faster paydown of credit card debt, using one-click “rounds-ups,” defaults, or shrouding. These policy options could deliver interest savings of up to 40%.