What factors drive public rental housing fraud? Evidence from Hangzhou, China

In China, considerable importance has been attached to public rental housing fraud since the central government decided on the large-scale construction of affordable housing in 2010. In recent years, this problem has also become a concern in countries such as Britain and the United States. In this paper, a multinomial logistic regression (MLR) model comprising 10 independent variables is developed to examine the driving factors of public rental housing fraud. The parameters of the model are estimated via maximum likelihood estimation based on the Hangzhou public housing household survey. Results from this study suggest that seven factors, namely, family size, education, PCDI, CSC, ADC, PMC and Occupation = 1(servants), are statistically significant under the 5% level. The coefficients of these five factors are −0.847, −0.601, −0.732, −1.475, 0.987,-1.106 and 1.669, respectively. The positive (negative) coefficients mean that the variables will increase (decrease) the probability of fraud. At the end of this paper, policy recommendations are proposed for relevant government departments based on the results of the regression analysis.