Quantitative Analysis of Haircuts: Evidence from the Japanese Repo and Securities Lending Markets

Given the absence of comprehensive studies on market structure and haircuts for repo and securities lending transactions, this study provides a quantitative analysis of the subject using government bonds and equities transaction data covering most of the Japanese market. Specifically, we conducted a panel data regression analysis of government bond repo transactions, controlling for factors such as transaction entities and transaction types, and provided a detailed analysis of the haircut-setting mechanism. Accordingly, we determined that explanatory variables affecting credit risk, market risk, and liquidity risk, such as the credit quality of government bonds, the residual maturity of government bonds, and the presence of foreign exchange risk, significantly impact haircut setting. Furthermore, financial institutions closer to the center of the network, which engage in transactions with additional financial institutions, tend to set lower haircut rates through more efficient matching of borrowing and lending needs for cash and securities. Thus, the credit quality of government bonds transacted, exchange rate stability, and the presence of intermediaries important to the trading network significantly impact the degree of market functioning. The results were robust, paving the way for further discussions on trends and risk management of securities financing transactions, which are essential to financial markets.