This paper reports overwhelming evidence against the hypothesis that the federal funds rate follows a martingale over the 2-week reserve maintenance period, establishing that banks do not regard reserves held on different days of the week to be perfect substitutes. A theoretical model of the federal funds market is proposed that could account for these empirical regularities as the result of line limits, transaction costs, and weekend accounting conventions. The paper concludes that such transaction costs lie at the heart of the liquidity effect that enables the Federal Reserve to change the interest rate on a daily basis.

This paper reports overwhelming evidence against the hypothesis that the federal funds rate follows a martingale over the two-week reserve maintenance period, establishing that banks do not regard reserves held on different days of the week to be perfect substitutes. A theoretical model of the federal funds market is proposed that could account for these empirical regularities as the result of line limits, transaction costs, and weekend accounting conventions. The paper concludes that such transaction costs lie at the heart of the liquidity effect that enables the Federal Reserve to change the interest rate on a daily basis.