Hybrid Approach for Quantifying Company Assets Using Structural Credit Risk Models

1. Introduction

Credit risk modelling is a systematic approach towards assessment of credit risk and computation of cost of risk (Duffie and Singleton, 2012). It helps creditors to analyse different factors which are contributing to the overall risk in their portfolio. Credit risk modelling helps in predicting the probability of loss in unforeseen circumstances. Thus over the decades credit risk modelling has emerged as one of the major tool for investors to make informed decisions. Banks evaluate the borrowing company’s credit risk to assess their creditworthiness, to price their loans and to protect their investments appropriately. Creditors analyse the risk of the financial products to diversify their portfolios by identifying the risk and return of their investments. Credit default is a rare occurrence with low odds. Yet it demands special attention among the many risk components, such as the market risk or operational risk, because of the narrow window of predictability and the immense magnitude of the loss acquired in case of defaults (Jiménez and Saurina, 2004). The global crisis of 2008-2009 provides testimony to this. All creditors, regardless of their geographic locations, devote the majority of their resources towards controlling credit risk and developing models to predict future likely defaults or losses. Though the financial institutions have standardized and revamped their credit models much better after the 2008 crisis, due to the pandemic, high uncertainty has been introduced in the market. Existing model factors have to be studied taking into account the contractions in the GDP and deviations in the exchange reflecting the impact of the pandemic and its variation across different countries and business sub-sectors. Financial institutions are in a situation to re-assess the credit models of their portfolio and be pro-active in handling the products or services that do not meet their payment terms. With the adoption of digital transformation over the last decade, the financial institutions have access to large amount of data in order to assess risk profiles more accurately (Rajini, Ramamoorthy, Rammohan, Rajakumar and Niveditha, 2020).

Credit risk models play a key role in improving the credit score of a bank. Generally, creditors adopt different models for corporate as well as retail segments. Some of the current applications follow a risk based pricing, aiming to improve returns, reducing risk of their portfolios and achieving better performance of various businesses by following appropriate risk adjustments. Credit risk models include the possibility that investors may fail to pay on demand and allows estimation of the cost of loss, so that funds may be allocated appropriately. The overall ability of the borrower to repay the loan under the original terms is used to measure the credit risk. Merton models were traditionally developed and implemented for various practical applications including financial product pricing (Altman and Saunders, 1997). In this paper we propose a hybrid approach for enhancing widely used Merton Model after studying its evolution over years. The Merton Model has been applied to represent the asset structure of Tesla stocks by extracting the required parameters from the live stock price of Tesla from S&P 500 Index after the onset of pandemic.

We propose a unique methodology to evaluate the market value of the firm’s assets by assuming Merton Model based asset structure, whose parameters have been obtained by an Annualized Historical Volatility Model (AHVM).

The paper is classified as follows: The technical factors in the credit risk domain and the traditional models with interesting adaptations have been discussed in the first two sections. The following sections explore the data set and the proposed adaptations to the chosen model. Finally, we discuss our results insights and possible improvements.