Scribd
Upload
38 views4 pages

Lab Session 3: 1 Wooldridge, IE, 7.10

This document summarizes a lab session on loan approval modeling. The session focused on using an Excel file with loan application data to build and compare linear probability models for predicting whether applicants will be approved. Students were tasked with 1) constructing a baseline model regressing approvals on available variables, 2) estimating and commenting on results, 3) considering alternative models and choosing the best, and 4) using the chosen model to predict approvals for new applicants and identify the top 500 candidates to receive loans.

Uploaded by

giga sesitashvili
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
38 views4 pages

Lab Session 3: 1 Wooldridge, IE, 7.10

This document summarizes a lab session on loan approval modeling. The session focused on using an Excel file with loan application data to build and compare linear probability models for predicting whether applicants will be approved. Students were tasked with 1) constructing a baseline model regressing approvals on available variables, 2) estimating and commenting on results, 3) considering alternative models and choosing the best, and 4) using the chosen model to predict approvals for new applicants and identify the top 500 candidates to receive loans.

Uploaded by

giga sesitashvili
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Lab session 3

Lecturer: Nataliia Ostapenko

15/12/2017

1 Wooldridge, IE, 7.10

1
2 Wooldridge, IE, 7.12

3 Wooldridge, IE, 7.13

2
4 Wooldridge, IE, 7.16

5 Exercise on loan approval: who should we choose to


give loan to?
Attached is an excel …le where there are two worksheets: loan applications which are already
decided on and new loan applications. In the excel …le, there is the client id numbers
client_no, data on monthly income monthly_income, data on years worked years_worked,
if the clients have other loan obligations obligations, and if they received approval to their
loan application got_approval: The obligations and got_approval variables are binary: they
receive value 1 if the data entry is "yes".
For the new applications, we have similar data except the got_approval variable, because

3
this is what we have to decide: if the client should receive a loan or not. The bank wants
to give 500 new loans but there are 1000 applications. So we should …nd out which clients
(using client_id) should receive loans.
Given that the dependent variable is a binary variable, the model we will use here is a
linear probability model. (In fact using a probit or logit model would be more accurate for
the reasons we saw in class.) The idea of the exercise is that we should …nd the 500 clients
who have the highest predicted probability of receiving a loan.
The exercise is as follows:

1. Construct a model that regresses got_approval on the available regressors.

2. Estimate the model and comment on the results.

3. Consider reasonable alternative models (possibly with interactions between regressors


or other functional forms) and choose the model that best explains the data. Make
sure you use a correct way of comparison between models.

4. Using the estimation results of the model that you choose in (3), predict if the new
clients shall receive the credit. Identify the 500 clients who should receive the new
loans.

You might also like