Using 2012 National Financial Capability Studies (NFCS) state-by-state data, we examined the racial difference in responding with correct, incorrect, and DK responses to a 5-item financial knowledge scale. The NFCS is a large national survey of U.S. adults and households, and its state-by-state survey was conducted in 2012 among a nationally-representative sample of 25,509 American adults. The financial knowledge scale is composed of five factual questions concerning knowledge about compound interest, inflation, bond price, mortgage payment, and financial risk. Questions are in a multiple-choice format, and each has one correct answer, two incorrect answers, a DK answer, and “prefer not to say” answer. We employed grouped-data multinomial logistic regression to examine gender difference in choosing DK response and incorrect answers, relative to answering with a correct answer.
Results show that, controlling other socio-demographic variables, racial minorities were more likely than the white to have incorrect answers on all five questions: compound interest (β=.34, p<.001), inflation (β=.43, p<.001), bond price (β=.32, p<.001), mortgage payment (β=.57, p<.001), and financial risk (β=.59, p<.001). Results also show that racial minorities are more likely to respond DK than whites across four questions: compound interest (β=.24, p<.001), inflation (β=.32, p<.001), mortgage payment (β=.49, p<.001), and financial risk question (β=.26, p<.001). The race difference in giving DK response to the bond price question was not significant (β=.02, z=.44, p=.66).
Our study challenges the conventional measures of financial literacy by demonstrating the racial disparity in DK response and incorrect answers. The findings suggest that financial knowledge levels of racial minorities can be substantially higher than most studies have shown, and the racial gap may not be as large as it seems. In addition, racial disparity in choosing incorrect answers appears domain-specific, with relatively big impact on mortgage payment and financial risk questions, and small impact on compound interest and bond price questions. Our study demonstrates that financial literacy measures are capturing not just knowledge, but other things (such as risk propensity, self-confidence). To measure financial knowledge precisely, future research should consider constructing measures that discourage DK response.