Income Estimation Models and A Dream

As everybody knows, the concern of current deficit, overindebtedness of households and low saving rates have been led to some policy intervention by the Turkish Government for the last two years. Turkey has developed and implemented a macro prudential policy framework, which has had success in mitigating financial stability concern.

Banking Regulation and Supervision Agency (BRSA) has the major role of tackling this problem. According to their regulations:

• Number of installments on credit card payments has been limited
• Credit card limits have been linked to net income,
• They increased the risk weights for the calculation of capital adequacy and extended loan provisioning regulations to credit cards, overdrafts and vehicles loans, so allocated more general reserves for these loans by bank.
• They increased the minimum payment of credit card balance.
• The maturity of consumer and vehicle loans limited to 36 months.

Although these regulations decreased the volume of the consumer lending and credit card spending, they also have caused the slowing of economic activity, regulation arbitrage (underground economy) and the increase of problem loans. The increase in NPL may be the most critical results of these regulation. I do not want to focus on the details of causes and results of this problem. My colleagues and I are working on an article and it will be published soon.

The increasing debt problems of households has led countries all over the world to seek new solutions. One of our regulation is to tie the credit card limit with net income. According to the regulation, credit card holders will not be able to exceed the upper limit, which will be set by regulation of their monthly income. The consumer should present his monthly income in order to increase his credit card limit. The concept of income is always controversial for Turkey because of underground economy. Although the percentage of underground economy to total economy has been increasing for the last ten years, the real income of households is not provable enough. Therefore, banks have improved their income estimation models to decrease the problem of adverse selection for the last five years. In these models, many different data could be used such as, KKB (credit bureau) scoring and loan, demographic, credit card-loan, asset and cash flow information of customers.

Generally, multi regression linear models and dynamic regression models are used in these models by banks when defining algorithmic and measuring relationship between variables. These models are supported by VIF (variance inflation factors) or kappa statistics to measure multicollinearity problem.

Apart from this technical detail, we can understand that income models are not only to measure the real income of households, but also they can be powerful tools for the prevention of the adverse selection problem.

These models can also provide a range of benefits including better fraud mitigation, stronger risk management, and responsible provision of credit.  Using income estimation models to understand banks complete financial picture is valuable in all phases of the customer lifecycle, including:

• Loan Origination – use as a best practice for determining income capacity
• Prospecting – target customers within a specific income range
• Acquisitions – set line assignments for approved customers
• Account Management – assess repayment ability before approving line increases
• Collections – optimize valuation and recovery efforts

I would like to mention a possible outcome of this tool. In my view, Income estimation model can be developed for all financial system in Turkey. Motto; one model, one country!

According to my dreams, KKB, BKM or TBB will create a dynamic income estimation model. It will be validated by an independent firm or BDDK.

• This model will integrate into current processes of banks, such as collection scores from KKB.

• It will assess all income sources including wages, investments, alimony rent and maybe cash flows.

• Credit card limits or all indebtedness of retailers will be linked the model. Therefore, banks will adapt this model into their internal process.

• In order to get a loan, individual customers should prove their income. With the exception such as cashable collateral, customers will not borrow money from financial institutions.

What will benefits of this model be?

• Although increased consumer debt level can in part be an important stimulus of growing economies, these debt levels in several markets may be leading large groups of consumers towards overindebtedness, and could pose a risk to economic growth, household well-being, and financial institutions with significant levels of consumer debt in their portfolios.With this model, economy administration can easily control this figure. For example, they can reduce the prevalence of over-indebtedness to reasonable levels without any side-effect.

• Consuming is always and always indispensable for our country. Even economy administration prohibits deferring paymentson any things (such as mobile phones), people will find the way to buy the thing (The import of mobile phone has still been increasing continuously as we see). Then, our administration can control and limit the underground economy via an income estimation model.

Either prove your income or give up loan application!

Demanding loan will require you to submit the proof of earnings or assets. This model will be a powerful incentive for people to prove their real income. Otherwise, they will not be able to get their loan.  In the medium term, income estimation model can be an effective tool for tackling the underground economy in this way.