CONSUMPTION CREDIT DEFAULT PREDICTIONS
AbstractConsumption credit plays an increasingly important role in facilitating consumption and enables consumers to smooth consumption. Today, as much as 26% of all card transaction volume in Sweden is made with credit cards. In addition, many retailers themselves offer different types of credit alternatives. However, lending is associated with risks and it is therefore important to be able to correctly predict credit defaults. This paper investigates what factors are important to take into consideration when making credit default predictions by estimating a probit regression model using 170.000 approved consumption credits. While most traditional scoring methods mainly look at financial and demographic variables this paper shows that behavioural variables are at least as important when making default predictions.
From the beginning of time credit has been used to allow for smoothing of consumption. Lending, borrowing, instalments, payment after or before delivery, consumption credit, all kinds of forms have existed and have been a vital part to smooth transactions and enable economic growth. Today as well, consumer credit alternatives play an important role in the economy to facilitate consumption. In Sweden, where credit card penetration is much lower than in many other industrialized countries, invoicing and purchase by instalment plays an even more important role. In Sweden about 12% of all card transactions are credit card transactions compared to 26% in Germany and 50% in the US. When looking at the total volume of money rather than the number of transactions the findings are even more convincing, only 21% of all card volume in Sweden is in the form of credit transactions, in Germany it is 27% and in the US 68%. Credit allows consumers to smooth consumption in both the long and short term. In the short term consumers can purchase and pay after they receive their salary. In the longer term, younger people for example, might want to maintain consumption at a higher level than their current income allows, in the expectance of increasing income in the near future. Credit rather than prepayment is also often associated with the transfer of transaction risk from the buyer to the seller.
While the facilitation of consumption credits increases purchasing power and hence sales, it also includes risk taking, the risk of not getting paid in time, or not at all. Lenders, be they credit institutions or retailers, minimize risks by trying to predict defaults. Considering the vast amount of credit provided to Swedish consumers, thus enabling them to smooth consumption, it is of great importance for social welfare to improve the lenders’ ability to predict defaults. Better default predictions mean that more people can be provided with credit at a lower cost. Many lenders use some type of scoring model to try to predict who will default on their loan. The most commonly used models are developed by external credit reporting agencies and based on primarily public data sources. However, many of the larger lenders have also developed internal credit scoring models.
Credit has always been a vital part of
commercial transactions, and important for a well functioning economy. People
have become more and more dependent on credit and credit is used not only to
finance large personal investments such as house purchases but also to finance
other kinds of investment and even consumption. For example credit card
penetration which can be seen as a good indicator of our dependence on credit,
increased by 43% from 1998 to 2005 in Sweden.
However, things have changed since the days when credit was personal, like the
one between the local bank and a well known client. Nowadays, lending has
become more anonymous and the debtor is rarely known to the party that takes
the credit risk. This development has been enabled by the standardisation of
transactions and different methods have been developed to control the risk
involved. When one extends a loan, the lender has to have some way of
estimating the risk of default and account for this risk. The method used when
estimating the risk of default for personal loans is called credit scoring, and
the importance of credit scoring has increased with the development of
different securitisation-techniques. Securitisation has not only led to an even
further increase in the distance between the debtor and the lender, but credit
scoring is also used in the pricing of the security.