Tps spline, no set lambda, 10% US stations used for training 90% for testing

This analysis examines bank borrowing choices during the 2023 banking crisis using three complementary approaches: Model Framework: Linear Probability Model (LPM): BTFP vs DW among borrowers Logit Model: BTFP vs DW among borrowers Multinomial Logit: Choice among {No participation, BTFP only, DW only, Both BTFP+DW} Trivariate Probit: Joint modeling of BTFP, DW, and FHLB decisions with error correlations.

Tps spline, no set lambda, 10% US stations used for training 90% for testing

This analysis examines bank borrowing behavior across Federal Reserve emergency facilities (BTFP and Discount Window) during the March 2023 banking crisis. We exploit the institutional differences between facilities—particularly BTFP’s par valuation of eligible collateral—to test whether banks strategically selected facilities based on balance sheet vulnerabilities.

Tps spline, no set lambda, 10% US stations used for training 90% for testing

Tps spline, no set lambda, 10% US stations used for training 90% for testing

Tps spline, no set lambda, 10% US stations used for training 90% for testing

Tps spline, no set lambda, 10% US stations used for training 90% for testing