Working Paper: NBER ID: w30561
Authors: Emil Siriwardane; Adi Sunderam; Jonathan L. Wallen
Abstract: We use arbitrage activity in equity, fixed income, and foreign exchange markets to characterize the frictions and constraints facing intermediaries. The average pairwise correlation between the 29 arbitrage spreads that we study is 21%. These low correlations are inconsistent with canonical intermediary asset pricing models. We show that at least two types of segmentation drive arbitrage dynamics. First, funding is segmented—certain trades rely on specific funding sources, making their arbitrage spreads sensitive to localized funding shocks. Second, balance sheets are segmented—intermediaries specialize in certain trades, so arbitrage spreads are sensitive to idiosyncratic balance sheet shocks.
Keywords: Arbitrage; Financial Intermediaries; Asset Pricing
JEL Codes: G12; G13; G2
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| specific hedge funds' balance sheet conditions (G23) | certain trades (F19) |
| funding segmentation (I22) | varying correlations among arbitrage spreads (C10) |
| balance sheet segmentation (G32) | arbitrage spreads (G19) |
| tightening of risk limits at JP Morgan (F65) | equity spot-futures arbitrage spreads (G13) |
| shocks to funding availability (F35) | arbitrage outcomes (G19) |
| unsecured arbitrage spreads (G19) | sensitivity to movements in the TED spread (E43) |