Tulyakova, Marina (2019) A Dynamic Model for Optimal Covenants in Loan Contracts. PhD thesis, University of Trento.
|PDF - Doctoral Thesis|
|PDF - Disclaimer |
Restricted to Repository staff only until 9999.
Covenants are an important part of financial contracts, that are used for resolving the conflicts of interest between borrowers and lenders. In more formal way covenants can be determined as special provisions in loans that give lenders the possibility of putting certain actions in force (normally early repayment) when covenants are violated. For instance, a covenant may restrict the company in taking additional credit, or require a firm to maintain certain financial ratios, such as leverage, coverage, liquidity ratios, etc. Our study develops a theoretical framework that allows to determine the covenant strength index that should be included in a debt contract in a way that minimizes expected losses for a bank subject to the rising restructuring costs. This optimal covenant is found in order to better allocate control rights ex ante and to minimize the costs of renegotiations for both parties. The approach that explores dynamic contingent claim models is applied to the prob- lem. This approach was pioneered by Black and Scholes (1973) and Merton (1974), and extended by Black and Cox (1976). The dynamics of the optimal covenant strength with respect to various model parameters is investigated. Different modifications to the initial model are considered which are important in exploring more realistic model setting. First of all, we introduce the concept of deadweight costs of distress or firesale price. The concept of deadweight losses imply that the debt holder gets some fraction of the asset value on default instead of the fundamental asset value (Das and Kim, 2015). Along with the concept of deadweight costs, the notion of firesale price is used, that represents the price at which the asset can be sold before the contract maturity. We explore how this extension of our baseline dynamic model influence the optimal level of covenant strictness in debt contracts. We further develop a model of an optimal covenant in bank loans with information asymmetry. Asymmetric information as a source of agency problems is very important in studying control rights in financial contracting. The conclusions of the papers on information asymmetry regarding control rights allocation and covenant strictness are often ambiguous. Different papers demonstrate more or less control rights of lenders or greater or lesser strictness of covenants depending on the setting and model parameters. Our model is unique in a sense that it unites different implications of empirical and theoretical models with information asymmetry and reflects both perspectives. We also introduce a framework for accessing the consequences of covenant violation in Monte Carlo simulation. Our simulation model allows us to measure different risk- parameters of a project, such as the probability of covenant violation and the probability of repayment of the loan. The measurements (average number of covenant violations per contract, frequency of covenant violation, frequency of loan repayment) can be used in implementation of different rules for a bank that extends the traditional risk-analysis of a project. Moreover, we implement a recursive technique for determining the level of covenant strength that allows the bank to maintain the performance of a specific risk- parameter. We employ a dynamic approach in the spirit of Borgonovo and Gatti (2013); Chang and Lee (2013); Liang et al. (2014) by simulating project value paths over time.
|Item Type:||Doctoral Thesis (PhD)|
|Doctoral School:||Economics and Management (within the School in Social Sciences, till the a.y. 2010-11)|
|Subjects:||Area 13 - Scienze economiche e statistiche > SECS-S/06 METODI MATEMATICI DELL'ECONOMIA E DELLE SCIENZE ATTUARIALI E FINANZIARIE|
Area 13 - Scienze economiche e statistiche > SECS-P/03 SCIENZA DELLE FINANZE
|Repository Staff approval on:||07 May 2019 10:36|
Repository Staff Only: item control page