| With the wide usage of credit card and the development of credit card market, new developed credit card borrowing behaviors and payment attitude present unfamiliar challenges to consumer finance researchers, card issuers, and regulators. As the leading unsecured debts, credit card debts become the focus of academic and public policy in recent years with the growth of the number of households carrying credit card debts and the increase of the level of outstanding credit card balance.;The borrowing behaviors of households vary dramatically among different generations due to different social attitudes, varying consumption-savings habits, and enduring effects of historical events, such as early education and peer effects. In addition, the borrowing behaviors of households differ in different economic environments. For example, development of credit card market, availability of card facilities, financial environments, policy, and legislations all impact borrowing behaviors of households. Thus, the differences among generations and changes of economic environments which can be captured by Age (A), Period (P), and Cohort (C) Estimation Method have implications to estimate and forecast future credit card borrowing patterns.;Due to data limitation, most previous researches on consumer debts are on cross-sectional data. One limitation of these studies is that they overlooked the importance of cyclical influence (period effects) tied to the business cycle and social structural transformation because of population shifts (cohort effects). Applying the Age-Period-Cohort (APC) analysis and utilizing the series of Survey data (1989-2004 Survey of Consumer Finance), this study identifies the existence of Age, Period, and Cohort effects in households' credit card borrowing behavior, and how Period/Cohorts effects are taking into account to modify previous understanding of consumers' credit card borrowing behavior over life cycle. Specifically, two related questions are examined: (1) what is the credit card borrowing pattern across different age groups with and without cohort adjustments (2) what are the underlying factors accounting for consumers' revolving behaviors (propensity to borrow and magnitude of debts) on credit cards.;In order to answer the above two questions, aggregate level APC conventional models are used to identify the existence and pattern of A, P, and C effects. Age profile of credit card borrowing is studied using a two-way fixed effects model. A cross-classified two-level mixed effects model is proposed to disentangle the variance of borrowing behavior between different cohorts and periods and to investigate the determinants of credit card borrowing behaviors. The analysis result shows that younger generations are more inclined to borrow money on their credit cards and to carry more outstanding balance than other generations. Consistent with previous researches, demographic variables are not significant once A, P, and C factors are included in the model. In addition, the residual variation between Cohort variables, or Period variables are significant after controlling the individual-level explanatory variables, which provides the further evidence of the existence of Period and Cohort effects on households' credit card borrowing behavior.;As a complementary analysis to the APC analysis on credit card borrowing behavior using U.S data, a study on credit card borrowing behavior in Taiwan-one of the largest Asia emerging markets-is performed. There is a famous 'puzzle' in both Asia and U.S credit market, that is, consumers may revolve on credit cards while carrying sizable liquidity assets at the same time. In this research, those revolvers on credit card carrying sizable liquidity assets are defined as revolvers 'who do not want to pay' and those without liquidity assets are revolvers 'who cannot pay'. With the availability of a unique credit union panel data from 2008 to 2009, I perform an empirically comparison between these two groups on the underlying determinants of revolving decisions and amount, and try to differentiate those 'who do not want to pay' from 'who cannot pay'.;The findings of this study help to clarify the understanding on credit card debt accumulation pattern and the related determinants of credit card borrowing. The results are useful to researchers, policy makers, and practitioners who need to evaluate or formulate regulations in credit card market. |
