Listed Government-owned Company Debt Default Risk Measurement

Upon the rapid evolution in developing economies,financial market undoubtedly plays a core role in the economic system in terms of its high liquidity and speedy capital growth.High yield,as one of the main characteristics of the financial market,brings up high risks consequently,which includes credit risk,liquidity risk,market risk,etc.Among the miscellaneous risks,credit risk becomes an unignorable topic nowadays that is being considered by banks and exchanges for loans and bonds.It not only affects the benefits for both parties of the financial instruments but as well even influences the stability of the economic system.Risk management which depends on reliable risk measurements is a major approach to reduce financing risk and refrain from financial crisis.The volume of custodian bonds reached 65 trillion since the bond market was established at 2002.This fact has gradually highlighted problems on risk management awareness,lack of managing experience and outdated risk measurements.Frequent debt defaults in 2016 forced us to search for a more targeted and matchable credit risk measurement:KMV model.KMV model is a measurement that based on option pricing theory and regressed by thousands of debt default cases occurred in capital markets in multiple developed countries.It evaluates shareholders’ equity as a call option where the underlying asset is substituted with candidate corporation capital and strike price with corporation’s book value.At the expiration date of the debt,one can conclude whether the corporation will default or not by comparing the total amount of their liabilities and assets.The corporation is very likely to default if their liabilities exceed assets.However,the critical value for default in KMV model was defined depending on cases in developed counties,which does not precisely reflect the development situation in our capital market.Hence,the default critical value will be adjusted with Genetic Algorithm later in this thesis.The approach of optimization has been broken down into five parts.The first part gave a brief introduction on the considerations of selecting topic,the benefits of improving KMV model and development of that model in the past.Then,in the second part,we considered the advantages and disadvantages on varies in-used risk measurements,which laid a foundation on model optimization in following part.A thoroughly illustration on KMV model,Genetic Algorithm and the implementation of Genetic Algorithm in KMV model in terms of modifying default critical value was presented in part three.And in part four,an empirical research on 2013-2016 domestic A-share market data has been provided.Within the research,I gathered companies’ financial data and performances in stock market;coded with Matlab to evaluate their capital value and volatilities of the capital value;determined the best default critical value with Generic Algorithm;calculated default distances of the sample companies;and concluded predictions on debt default.Finally,part five summarized the application and still-existing pitfalls of the model.Three critical parameters in the model are developed in the thesis,which are stock price volatility,equity value and default critical value.Instead of developing based on the daily stock prices,the volatility for yearly stock price was calculated as a precise value with all valuation data for companies which were specially treated in a specific year.In terms of the equity value,outstanding shares,non-tradable shares and restricted shares are brought up to be considered.Based on the fact that the Split Share Reform is coming to closure,restricted shares would be a core part in pricing.Hence,the pricing of market outstanding shares and restricted shares are separated and evaluated more relying on the reality.Default critical value plays a crucial rule in the KMV model.Unlike previous research using a given critical value,I use Genetic Algorithm to optimize the critical value based on the financial data from selected domestic companies to get a more accurate evaluating result.