Price discovery and persistent arbitrage violations in credit markets

AuthorChunchi Wu,Junbo Wang,Kasing Man,Hai Lin
Published date01 March 2020
DOIhttp://doi.org/10.1111/fima.12261
Date01 March 2020
DOI: 10.1111/fima.12261
ORIGINAL ARTICLE
Price discovery and persistent arbitrage violations
in credit markets
Hai Lin1Kasing Man2Junbo Wang3Chunchi Wu4
1School of Economics and Finance, Victoria
University of Wellington,Wellington, New
Zealand
2College of Business and Technology,Western
Illinois University, Macomb, IL
3School of Economics and Finance, City
University of Hong Kong,Kowloon Tong, Hong
Kong
4School of Management, State University of New
Yorkat Buffalo, Buffalo, NY
Correspondence
ChunchiWu, School of Management, State Uni-
versityof New York at Buffalo, Buffalo, NY.
Email:chunchiw@buffalo.edu
Abstract
This paper investigatesprice violations in credit markets using a data
samplespanning from 2002 to 2016. We find that price violations are
highlypersistent during the crisis period, particularly for speculative-
grade bonds. There is evidencethat price distortions and market dis-
integration are linked to market-wide and firm-level impediments
to arbitrage and limited capital provision. Higher firm-level impedi-
ments to arbitrage lead to less market integration, and more severe
and persistent pricing discrepancies. Moreover,we find that the neg-
ative CDS basis persists in the postcrisis period, which is attributable
to dealers’ lower capital commitment and deterioration in market-
making quality.
1INTRODUCTION
The arbitrage-free condition has been a building block in the development of modern financial and asset pricing theo-
ries. Forexample, standard term structure models build on the assumption of no-arbitrage conditions, and asset pricing
theory assumes that any temporary deviations of prices from efficient benchmarks can be arbitragedaway quickly by
rational traders.However, it has been shown that the arbitrage-free condition has often been violated in financial mar-
kets (see Kapadia & Pu, 2012) and that violations were especially severe during the subprime crisis (see Duffie, 2010;
Mitchell & Pulvino, 2012).1In particular, acute price violations in credit marketsduring the crisis attracted consider-
able attention. The credit default swap (CDS) basis, or the difference between the CDS rate and yield spreads of a par
bond with the same maturity as the CDS, widened to above 600 basis points after the collapse of LehmanBrothers. As
Duffie pointed out in his 2010 American Finance Association presidential address, “The extreme negative CDS basis
‘violations’ across broad portfolios of investment-grade bonds and high-yield bonds, respectively,is far too large to
be realistically explained byCDS counterparty risk or by other minor technical details.”
CDS and bond yield spreads both reflect a firm's credit risk premium. In a frictionless market, any discrepancy in
these two variables will be eliminated quickly by arbitrage. This suggests that the CDS basis should be close to zero if
arbitrageis perfect. However, arbitrage is rarely perfect, as cash flows of CDS and the reference obligation are typically
c
2019 Financial Management Association International
1Forexample, there were serious violations of covered interest rate parity, a negative spread between Treasurybond yields and London Inter-Bank Offered
Rate(LIBOR) swap rates, and a breakdown of the capital structure arbitrage across equity and credit markets.
Financial Management. 2020;49:207–233. wileyonlinelibrary.com/journal/fima 207
208 LIN ET AL.
not perfectly aligned. There are also other complications; for instance, physically settled CDS prices contain the value
of the cheapest-to-delivery (CTD) option, and arbitrage may require shorting the cash bond which can be costly or
sometimes infeasible. Past studies prior to the subprime crisis have shown that the CDS basis was close to zero, or
slightly positive, which can be attributed to the CTD option value or costly short selling. The huge negative CDS basis
that occurred during the subprime crisis is thus more difficult to explain, and posts significant challenges to rational
no-arbitrage pricing theory.
Violations of the arbitrage-based pricing relationship have profound implications for market efficiency and asset
pricing. In particular,persistent price violations can cause asset price distortions and market disintegration. The litera-
ture has suggested various sources of asset mispricing. Price violations are commonly attributed to limits-to-arbitrage
(see Brav,Heaton, & Li, 2010; Mitchell & Pulvino, 2012; Pontiff, 2006; Shleifer & Vishny, 1997). Arbitrage maynot be
feasible when transactioncosts and the risk of a firm's security are excessively high. Funding constraints or limited arbi-
trage capital provision can prevent arbitrageactivity (Brunnermeier & Pedersen, 2009; Mitchell & Pulvino, 2012) and
cause serious mispricing in similar securities.
Inthis paper, we examine the issues related to the violation of arbitrage-free pricing relations in credit markets using
a data sample spanning from 2002 to 2016, which covers both the crisis and normal periods. Weemploy two methods
to assess the severity of pricing violations in credit markets. First, using a vector error-correction model (VECM), we
examinethe dynamics of CDS and bond spreads, and conduct a test on the equivalence of the prices of credit risk across
the two markets overtime via cointegration analysis. Second, we perform a nonparametric model-free test for market
integration based on the concordance of price changes in the CDS and bond markets.
A unique character of the price distortion during the subprime crisis is that pricing discrepancies are not only large
but also highly persistent. Besides studying the magnitude of CDS-bond mispricing, we adopt a long memory model to
quantify the persistence in pricing discrepancies, an issue which is much less studied in the literature. The long memory
model provides a generic time-series measure of persistence or long-range dependency (see Hosking, 1981; Lo,1991;
Zivot & Wang, 2006), which characterizesthe slow-moving nature of arbitrage capital in times of stress very well. An
advantage of this model is that the persistence in pricing discrepancies due to arbitrage capital shortage or other fric-
tions can be nicely summarized by a parsimonious parameterthat measures how slowly moving the arbitrage capital is
and how persistent the resulting price violations are. After identifying the pricing discrepancies and their persistence
in credit markets, we exploretheir relations to impediments to arbitrage at both the market and firm levels.
We document several interesting findings that contribute to the current literature. First, both cointegration and
nonparametric tests show that pricing discrepancies across firms’ CDS and bond marketsare common and much more
serious in the crisis period. The level and volatility of CDS-bond pricing discrepancies can be quite persistent. The
persistence in price discrepancies is much stronger in the crisis period than during normal periods, and is higher for
noninvestment-grade(IG) bonds than for IG bonds.
Second, we find that both the magnitude and persistence of price discrepancies are closely related to firm-level and
market-wide impediments to arbitrage. Consistent with the theory of costly arbitrage(Shleifer & Vishny, 1997), firms
with high risk, illiquidity, leverage,and return volatility experience more severe and persistent pricing discrepancies.
Violations of price convergence are also more frequent, and price discrepancies are larger and more persistent in the
period when funding constraints are high and market liquidity dries up. The relative role of firm-levelimpediments to
systematic (market-wide) impediments depends on market conditions. We find that firm-level impediments to arbi-
trage play a nontrivial role in CDS-bond pricing violations during normal market periods, but that in times of stress,
systematic impediments become more important factors.
Third, there is evidence that slow capital movementaggravates the persistence of price violations in credit markets
during the subprime crisis. Using several proxies for arbitrage capital provision, we find that the sensitivity of firms
to funding availability has high explanatory power for the persistence of price violations in credit markets in times of
stress. In addition, the funding-related variables exhibit a strong persistence pattern during the crisis period which
coincides with the persistence in CDS-bond pricing discrepancies. These findings lend support to the contention that
persistence in the shortage of arbitrage capital causes persistence in pricing discrepancies in credit markets.

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT