Where: Leg denotes logarithm E1: household earnings A1: Savings account balance (Rands} Hi 2 1 job has a housing allowance and zero otherwise Ti 2 Number years of education of the household. Hi: 1 it bad credit rating assessment and zero otherwise. 2.1 Interpret the estimated coefcients in the model. {5} 2.2 Assuming all other factors in the model remain constant (ceteris paribus), calculate: i. The probability that a household with earnings of RM DUB will own a house? {4} ii. The rate of change of probability at the earnings level of RH] UDD? {2} 2.3 Statistically determine whether all variables jointly are important determinants for the purchase of a house. |Clearly outline the steps {4} 2.4 Explain how you can use regression restrictions to determine the impact of explanatory variables on the purchase of a house. You can use any of the variables given. Clearly outline the steps 55} Question 2: (20 marks) A binomial logit model is estimated to determine the purchase of a house using a bank loan in a selected suburb in South Africa. The purchase of a house using a bank loan is a binary variable with Y=1 for purchasing and zero otherwise. The number of households is 100. The estimated binomial logit model is given by L; Li = 0.55 + 0.57LogE; + 0.112LogA; + 1.352H; + 0,452T; - 1.452R; Z = (-0.73) (2.97) (1.97) (12.567) (-1.98) (-1.98) Mcfadden R2 = 0.3507 LR statistics = 9.6073 Prob (LR statistics) = 0.027