Uma abordagem integrada para a gestão da liquidez e alocação de recursos no curto prazo em bancos comerciais

• Ano de defesa: 2007
• Autor(a) principal: Alcântara, Wenersamy Ramos de
• Orientador(a): Douat, João Carlos
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Administração contábil e financeira; Alocação de ativos; Lucro econômico; EVA; Gestão de liquidez; Risco de liquidez
• Palavras-chave em Inglês: Liquidity risk; Liquidity management; Asset allocation; Economic profit; RAROC
• Link de acesso: https://hdl.handle.net/10438/2545

This work develops an integrated model for optimal asset allocation in commercial banks that incorporates uncertain liquidity constraints that are currently ignored by RAROC and EVA models. While the economic profit accounts for the opportunity cost of risky assets, what may even incorporate a liquidity discount, it neglects the risk of failure due to the lack of sufficient funds to cope with unexpected cash demands arising from bank runs, drawdowns, or market, credit and operational losses, what may happen along with credit rationing episodes or systemic level dry ups. Given a liquidity constraint that can incorporate those factors, there is a probability Pf that there will be a fail and the liquidity constraint will not hold, resulting in a value loss for the bank, represented by a stochastic failure loss Lf. The total economic profit, given the possibility of loss due to the lack of liquidity, is then optimized, resulting in a short-term asset allocation scheme that integrates market, credit and operational risks in the liquidity management of banks. Even though a general approach is suggested through simulation, it is provided a closed form solution, under some simplifying assumptions, that is thoroughly discussed. An analysis of stylized facts about liquidity in Brazil suggests that the current decreasing trend in interest rates have some influence in the reduction of liquid assets as a proportion of deposits, increasing the relevance of liquidity management models such as the one proposed in this work. The model was then applied to Brazilian banks data resulting in an estimated 8.5% yearly gain over the optimization without liquidity considerations. Even though it is not possible to establish the significance of this result due to the approximations used, its sensibility to changes in the parameters is not high. After increasing and decreasing all parameters by 20%, the gain change ranged from 6.3% to 8.8%. Gains may reach 11.1% if the amount of liquid resources available for allocation is multiplied by four, and even if the loss given failure is reduced by 8.6 times there still are gains of about 0.5% a year in the return on equity, giving empirical indications that the model may have a relevant impact over the value creation in banks.