Get documents about "Market Quality and Trader Behavior in a Manipulated Market
Document Sample
Market Quality and Trader Behavior in a
Manipulated Market: Anatomy of a Squeeze
John J. Merrick, Jr. (Baruch College)
Narayan Y. Naik (London Business School)
Pradeep K. Yadav (University of Strathclyde)
June 2003
A futures delivery squeeze example of
market manipulation
• We study the six-month period of an attempted delivery
squeeze in the March 1998 long-term UK government
bond futures contract (LIFFE Long Gilt futures).
• We employ a highly unusual data set of bond and futures
market trades by all major players provided to us by the
UK Financial Services Authority (FSA).
Previous empirical research on bond
market manipulation
• Treasury Bond squeeze of 1986 (Cornell and Shapiro,
1989).
• Salomon US Treasury 2-Year Note squeeze (Jegadeesh,
1993; Jordan and Jordan, 1996).
• These studies focus only on price distortions. Our paper
is the first to investigate both the price distortions and
the trading positions of major market participants.
Plan for today
• A brief look at normal pricing of futures at delivery.
• Delivery date futures pricing during a squeeze.
• “Squeeze Potential”
• Examine the cash market pricing distortions.
• Trace the positions taken by the major players
(dealers & customers) during the squeeze.
Some interesting insights
• We identify a key market microstructure
element that supports the squeezer.
• Provide two types of evidence for our
explanation:
* A shift in cash-futures arbitrage pricing.
* A “natural experiment” conducted by BOE.
Lessons
• Squeezes are accompanied by severe price distortions and
some erosion of market depth.
• Delivery nonperformance penalties in bond futures
markets create conditions that favor squeezers.
• Exchanges may need to "mark to market" the conversion
factors of their contracts more frequently.
• Regulatory reporting should flag forward term repo
positions to paint the full picture of cash market
positioning by individual firms.
March 1998 Gilt Futures Delivery Date “No-
Squeeze” Futures Pricing for 2 key issues @ 6%
Coupon Maturity CF Mod. Dur. Yield Delivery Date
9.00% 10/13/2008 0.9999442 7.02 6.00% 03/09/1998
Issue Price Forward Price Forward Price/CF Net Basis
123.28 123.27 123.27 0.00 Cheapest
---------------------------------------------------------------------------------------------------------------
Coupon Maturity CF Mod. Dur. Yield Delivery Date
8.00% 09/25/2009 0.9291579 7.59 6.00% 03/31/1998
Issue Price Forward Price Forward Price/CF Net Basis
116.50 116.40 125.28 1.86 2nd cheapest
March 1998 Gilt Futures Delivery Date “Full-Squeeze”
Futures Pricing: 9% 2008 Price Impact = +2% of par.
Coupon Maturity CF Mod. Dur. Yield Delivery Date
9.00% 10/13/2008 0.9999442 7.02 5.777% 03/09/1998
Issue Price Forward Price Forward Price/CF Net Basis
125.28 125.27 125.28 0.00 Co-Cheapest
---------------------------------------------------------------------------------------------------------------
Coupon Maturity CF Mod. Dur. Yield Delivery Date
8.00% 09/25/2009 0.9291579 7.59 6.00% 03/31/1998
Issue Price Forward Price Forward Price/CF Net Basis
116.50 116.40 125.28 0.00 Co-Cheapest
Market Yield Levels vs. the Contract’s 9% Notional Yield
Contract’s Squeeze Value (Fs - Fns) increased as yields fell
Squeeze causes significant price distortions
Value the 9% 2008 Gilt vs. the discount factors
derived from the BOE’s Gilt term structure model
(e.g., use a typical discounted cash flow bond
valuation approach).
But “fair” pricing in this typical no-arbitrage sense
means a “free squeeze option.”
Traders begin to bid this cdi1 higher in October 1997.
It begins to look “rich” in the market.
“Richness” of the 9% 2008 vs. Gilt term structure
Implied Squeeze Probability:
p = (F0 – Fns)/(Fs – Fns)
The participants and strategies
Squeezers – customers or dealers?
Use of forward term repo trades through the
delivery date
Contrarians
Par Value Positions (cdi1 & futures) of the First 3 Squeezers
3a. Inventories of First Three Squeezing Customers
1250 1.25
Par value of position (in £ mns)
1000 1.00
Price Distortion (%)
750 0.75
500 0.50
250 0.25
0 0.00
01- 16- 01- 16- 31- 15- 30- 15- 30- 14- 29- 13- 28-
Sep Sep Oct Oct Oct Nov Nov Dec Dec Jan Jan Feb Feb SC1
Time Period (Sep 1, 97 to Mar 4, 98) SC2
SC3
Price Distortion
Positions of 2 late entrant Squeezing Customers
3c. Inventories of Squeezing Customers
(Late Entrants)
1000 1.25
Par value of position (in £ mns)
1.00
750
Price Distortion (%)
0.75
500
0.50
250
0.25
0
0.00
01- 16- 01- 16- 31- 15- 30- 15- 30- 14- 29- 13- 28-
-250 Sep Sep Oct Oct Oct Nov Nov Dec Dec Jan Jan Feb Feb
-0.25
-500 SC5 -0.50
SC6
Time Period (Sep 1, 97 to Mar 4, 98) Price Distortion
A Dealer (#1) learns from trades with “smart” customers
4a. Inventory of Squeezing Dealer 1
1800 1.20
Par value of position (in £ mns)
1500 1.00
1200 0.80
Price Distortion (%)
900 0.60
600 0.40
300 0.20
0 0.00
01- 16- 01- 16- 31- 15- 30- 15- 30- 14- 29- 13- 28-
-300 Sep Sep Oct Oct Oct Nov Nov Dec Dec Jan Jan Feb Feb -0.20
SD1
Time Period (Sep 1, 97 to Mar 4, 98)
Price Distortion
Cash Gilt & Futures positions of Squeezing Dealer (#2)
4 b . In v e n to ry o f S q u e e zin g D e a le r 2
1730 1 .2 0
1480 1 .0 0
In v e n to r y le v e ls ( in £ m n s )
1230
0 .8 0
Pr ic e D is to r tio n ( % )
980
0 .6 0
730
0 .4 0
480
0 .2 0
230
-20 0 .0 0
01- 16- 01- 16- 31- 15- 30- 15- 30- 14- 29- 13- 28-
Sep Sep Oct Oct Oct Nov Nov Dec Dec Ja n Ja n Fe b Fe b
T im e Pe r io d ( 1 S e p 9 7 to 4 M a r 9 8 ) S D2
P ri c e D i sto rti o n
2 Squeezing Dealers exit early (take some profits in Dec.)
4c. Inventories of Squeezing Dealers
(Profit Takers)
1500 1.25
1200 1.00
Par value of position (in £ mns)
Price Distortion (%)
900 0.75
600 0.50
300 0.25
0 0.00
01- 16- 01- 16- 31- 15- 30- 15- 30- 14- 29- 13- 28-
-300 Sep Sep Oct Oct Oct Nov Nov Dec Dec Jan Jan Feb Feb -0.25
SD3
Time Period (Sep 1, 97 to Mar 4, 98) SD4
Price Distortion
Dealers Learn from Observing Trade Flows
• Cash Bond trades
One dealer “learns” from intermediating the position
accumulations of the 9% 2008 cash bond by the first two squeezing
customers. This dealer then begins to accumulate a long position in the
9% 2008 too.
• Forward Term Repo trades
One dealer is the “uninformed” counterparty to forward term
repo trades by squeezing customers (who buy for settlement date
February 20, 1998 and sell back for March 20, 1998). Soon afterwards,
this dealer goes the same way in both the repo market and then goes
long in both the cash and futures markets!
“Contrarians” appear in size after cdi1’s price shifts +0.60%
and bet heavily that the Squeeze will ultimately fail.
8b. Inventories of Contrarian Customers & Dealers
0 1.20
01- 16- 01- 16- 31- 15- 30- 15- 30- 14- 29- 13- 28-
Par value of position (in £ mns)
-1000 Sep Sep Oct Oct Oct Nov Nov Dec Dec Jan Jan Feb Feb 1.00
Price Distortion (%)
-2000 0.80
-3000 0.60
-4000 0.40
-5000 0.20
-6000 0.00
CC-ToT
Time Period (Sep 1, 97 to Mar 4, 98) CD-ToT
Price Distortion
The Squeeze and Market Depth
• We run regressions relating daily cdi1 price
changes to dealer inventory changes.
• Kyle’s Lambda is significantly positive during
squeeze phases III, IV and V on days when there
were net customer buys.
• Public traders attempting to buy cdi1 during these
times faced perceptibly higher market impact
costs.
Asymmetries in Settlement Nonperformance Penalties
• Cash Bond and Bond Repo Markets permit trades to
“fail” – the implicit cost is overnight interest charge.
• Futures Exchanges impose substantial fines on contract
shorts that fail to make a timely delivery.
• Squeeze’s endgame: In a delivery squeeze showdown
between a credible squeezer and contract shorts, the
contract shorts blink first.
Shift in Marginal Financing Rate for Cash-Futures
Arbitrage Pricing Relationship
The BOE’s Feb 16, 1998 “natural experiment”
• “Bank of England is prepared to make supplies of the stock available
from 23 February, on overnight repo only, to any gilt-edged market
maker (GEMM) who has been subject to a failed return or delivery of
stock, or has a customer who has been subject to a failed return or
delivery of stock.”
• This subtle action by the BOE removed the fear of failing
against the futures contract. The pricing distortions
collapsed immediately.
The reaction was consistent with the nonperformance
asymmetries explanation as the key market microstructure
support for the squeeze attempt.
For review
• Squeezes are accompanied by severe price distortions and
some erosion of market depth.
• Delivery nonperformance penalties in bond futures
markets create conditions that favor squeezers.
• Exchanges may need to "mark to market" the conversion
factors of their contracts more frequently.
• Regulatory reporting should flag forward term repo
positions to paint the full picture of cash market
positioning by individual firms.
Related docs
