Measuring the quality of the Credit Scoring Model is an important factor to determine how well your model able to predict the probability of client default. Various Quantitative indices are used to measure the quality of such scoring models like- Gini Index, Lift chart, KS-statistics, GainChart, AUC etc. In our previous articles, we have individually covered these topics. In this article, we try to cover the difference between the Gini Index and Lift chart.

Gini Coefficient

Gini coefficient is a measure of goodness of a binary regression model, The Gini coefficient is a ratio of two areas – calculated by sorting scores from a distribution from low to high, determining the cumulative lift (Lorenz Curve), the area between cumulative uniform distribution and the Lorenz Curve on the interval [0,1] and dividing the area result by the area under the cumulative uniform distribution on the same interval.

Gini Coefficient above 60% is a good Model.

Gini= A^{U} -A^{L }/ A^{U }

Gini checks the predictive power of a credit risk model, the degree at which the model has a better discrimination power that random values.

Lift Chart

Lift curve is the plot between total lift and %population. It’s important to note that for a random model, this always stays flat at 100%.

Lift Chart helps us to determine till what level we target customers that will give us better results as compared when compared to a strategy without any model. Lift chart tells us how much better we can expect with the predictive model as compared to without any model.

Lift= Gain% X %population in the decile