Mizrach/Neely Tatonnement 1. 2 Overview of Papers Discussed Bruce Mizrach and Chris Neely (2006),...
-
Upload
alyson-price -
Category
Documents
-
view
216 -
download
2
Transcript of Mizrach/Neely Tatonnement 1. 2 Overview of Papers Discussed Bruce Mizrach and Chris Neely (2006),...
Mizrach/Neely Tatonnement1
Mizrach/Neely Tatonnement2
Overview of Papers Discussed
Bruce Mizrach and Chris Neely (2006), “The Microstructure of Bond Market Tatonnement,” St. Louis Federal Reserve Working Paper #2005-70a.
Bruce Mizrach and Chris Neely (2006), “The Transition to ECNs In the Secondary Treasury Market.” Federal Reserve Bank Review, Nov/Dec., forthcoming.
Oleg Korenok, Bruce Mizrach, and Stan Radchenko (2006), “Structural Estimation of Information Shares.”
Michael Fleming, Bruce Mizrach, and Chris Neely (2006): “The On-The Run U.S. Treasury Market: A Complete Snapshot.”
Mizrach/Neely Tatonnement3
Outline
I. Conceptual Material on Market Microstructure
II. Structural Model
III. Description of Treasury Market Data and Architecture
IV. Estimates of Treasury Market Information
V. Alternative Bayesian Structural Approach
VI. Conclusion
Mizrach/Neely Tatonnement4
I. Market Microstructure Concepts
Mizrach/Neely Tatonnement5
Market Microstructure
Hasbrouck: “Market microstructure is the study of trading mechanisms used for financial securities.”
The growing availability of transactions databases has facilitated the study of high frequency phenomena in various markets. (Equities: TAQ; NASTRAQ. FX: Olsen;)
The majority of research has been on equities and foreign exchange, much less on fixed income.
Mizrach/Neely Tatonnement6
Market Fragmentation
Similar or identical securities often trade in multiple venues.
This paper: IDB spot versus futures markets in Treasuries.Mizrach and Neely (2006): IDB voice markets versus electronic markets Treasuries.Korenok, Mizrach and Radchenko (2006): 6 stocks that have dual listings on NYSE and Nasdaq.Fleming, Mizrach and Neely: BrokerTec versus eSpeed ECNs.
Hasbrouck: “In all security markets there is a trade-off between consolidation and fragmentation. Consolidation… brings all trading interest together in one place, thereby lessening the need for intermediaries, but as a regulatory principle it favors the establishment and perpetuation of a single market venue with consequent concern for monopoly power. Allowing new market entrants…maximizes competition among trading venues, but at any given time the trading interest in a security is likely to be dispersed (fragmented) among the venues, leading to increased intermediation and price discrepancies among markets.” (Italics added).
Mizrach/Neely Tatonnement7
Market Architecture
Securities also trade in a hybrid environment of market designs or architectures.
NYSE: Floor based auction organized by a specialist
Nasdaq: Interdealer electronic network
ECNs (ATS): Electronic networks with no dealer intermediaries (e.g. Archipelago, Instinet for
equities, BrokerTec, eSpeed for U.S. Treasuries)
Open outcry: CME, CBOT futures pits, Treasury phone based market
This is a rich area for empirical industrial organization research.
Mizrach/Neely Tatonnement8
Fundamental Concepts - Liquidity
Hasbrouck: “Liquidity is a summary quality or attribute of a security or asset market.
• Depth: The quantity available for sale or purchase away from the current market price.
• Breadth: The market has many participants.
• Resilience: Price impacts caused by the trading are small and quickly die out.
• Spreads: Difference between buyer and seller initiated prices.
Mizrach/Neely Tatonnement9
Fundamental Concepts – Price Discovery
Madhavan (2002, FAJ): Price discovery is the process by which prices incorporate new information.
In the papers discussed today, we focus on the dimension of which market leads other markets in the price discovery process. This concept is called information share. This paper is the first to compare spot and derivatives markets in Treasuries.
Hasbrouck (1995): “The information share associated with a particular market is defined as the proportional contribution of that market's innovations to the innovation in the common efficient price.” Lehmann (2002): “a decomposition of the variance of innovations to the long run price.”
Enough chatting, let’s get to the model!!
Mizrach/Neely Tatonnement10
II. Hasbrouck Model
Mizrach/Neely Tatonnement11
Multimarket Security Model - Hasbrouck (1995)
pi,t ip t i,t .
The price in security market i differs from the fundamental price p* only transiently. The coefficient β is there because futures and cash markets may have a slightly different basis.
pt p t 1
t .
The fundamental price itself follows a random walk. Error terms ξ and η can be contemporaneously and serially correlated.
This is called an unobserved components model because we don’t observe the efficient price directly.
Mizrach/Neely Tatonnement12
Permanent Component
p t zt 1 A1 p t 1 Ar pt r 1 t ,
N 1 1zt 1
p1,t 1 2,t 1p2,t 1
p1,t 1 N,t 1pN,t 1
.
1
1 N
1 N
,
If we assume the individual prices are I(1), have a VAR(r) representation, and that markets are cointegrated, the price vector has the Engle-Granger error correction form:
Matrix of long run multipliers
Note that this approach uses the reduced form VAR rather than the structural model (as in Korenok, Mizrach and Radchenko (2006)).
Note daily coefficient to adjust for futures basis and/or cheapest to deliver.
Mizrach/Neely Tatonnement13
Effect of Shocks In the Long Run
E t t
In computing the long-run effects of a shock, we need to take into account contemporaneous correlation
by taking a Choleski decomposition, finding
M i 1N j 1
i mij such that MM
Now, of course, we have all the same problems that the macroeconomists do. The Choleski decomposition is not unique. (1) This paper reports upper bound estimates; (2) An argument in favor of working directly with the structural model.
Mizrach/Neely Tatonnement14
Information Shares
Hj i j
n im ij
2
i 1n im i1
2 i 2
n im i2
2 nmnn 2
.
Hasbrouck
GGj j
i 1N i
.
Gonzalo-Granger (Harris, McInish and Wood (2002))
Other estimates include deJong and Schotman (2004), Yan and Zivot (2005). Lehmann (2002) attempts to reconcile these.
Mizrach/Neely Tatonnement15
III. Description of Treasury Market Data and Architecture
Mizrach/Neely Tatonnement16
U.S. Treasury Market
Stage Factoid How Traded Database
When Issued Before auction ECN, Voice eSpeed, GovPX, BrokerTec
New Issues Discrete issues Auction Treasury Department
On The Run Commoditized ECNs eSpeed, BrokerTec
Off the Run Illiquid Voice GovPX, BrokerTec
Futures Liquid Trading pits CBOT, CME
Cisco futures, TickData
This paper focuses on IDB voice transactions in GovPX, and futures market transactions from the CME and CBOT,
Stage Factoid How Traded Database
IDBs in GovPX This paper
On The Run ECNs Mizrach and Neely (2006) Fleming, Mizrach, Neely (2006)
Futures Trading pits CME, CBOT
This paper
Cantor’s eSpeed is discussed in Mizrach and Neely (2006). BrokerTec and eSpeed are discussed in Fleming, Mizrach and Neely (2006).
Mizrach/Neely Tatonnement17
On-The Run Treasury Market in 2005
Mizrach/Neely (2006): On-the-run volume nearly 100% electronic, split between eSpeed and BrokerTec, two ECNs.
61%
39%
BrokerTec
eSpeed
Momentum is with BrokerTec. Cantor had 70% share in 2001.
Mizrach/Neely Tatonnement18
Mizrach/Neely (2005) - Sample
A voice market with screen based display of quotes. This is NOT an ECN.
Trading is by large institutions. > 1mn$ per trade
Traders have a workup facility. Boni and Leach (2002).
2-year, 5-year and 10-year on the run Treasuries, October 1, 1995 to March 31, 2001, from GovPX. Open outcry trades from 3 major firms: Garban-Intercapital; Hilliard Farber; and Tullett Liberty.
3-month Eurodollar, 10-year and 30-year Treasury futures. The bonds trade on the CBOT, the Eurodollar on the CME.
Spot Market
Futures Market
We use the hours that the futures markets are open: 8:20 to 15:00 CST.
Mizrach/Neely Tatonnement19
Trading Activity – Spot Markets
Mizrach/Neely Tatonnement20
Trading Activity on ECNs
Notes: The data are daily averages for the period October 12 - December 31, 2004.
Mizrach and Neely (2006) show a 635% growth in the 2-year and a 2000% increase in the 5-year just on the eSpeed platform.
The Treasury market has joined the equity and foreign exchange markets in attracting computerized trading. More than 50^ of BrokerTec quotes are from computerized trades (Safarik, 2005) rather than from the primary dealers.
2Y 5Y 10Y 30Y eSpeed Btec eSpeed Btec eSpeed Btec eSpeed Btec Trades 851.88 1,805.92 2,220.32 3,644.49 2,541.09 3,913.39 748.34 612.42 Volume 7,491.16 18,053.63 10,946.95 15,262.98 11,163.86 13,400.46 1,911.66 1,258.02
Mizrach/Neely Tatonnement21
Trading Activity – Futures Markets
Mizrach/Neely Tatonnement22
Liquidity – IDB Spot Market
Maturity Bid/Ask Spread (bp) Depth (mn$) Market Impact (bp)Two-year note 0.35 24.5 0.4235Five-year note 0.29 10.7 0.9368Ten-year note 0.33 7.9 0.9066
Depth and spreads: Fleming (EPR, 2003). Sample period 1997-2000.
Market impact: Mizrach and Neely (2006) for January 1999.
Mizrach/Neely Tatonnement23
ECN Liquidity
2Y 5Y 10Y 30Y Liq. Measure Cantor ICAP Cantor ICAP Cantor ICAP Cantor ICAP Ticks 26,934 60,152 70,105 113,887 57,036 127,138 68,621 54,308 Inside Spr (bp) 0.091 0.081 0.103 0.090 0.187 0.166 0.296 0.328 Inside Bid Depth 71.77 105.48 26.69 28.65 35.31 33.27 8.57 6.72 Inside Ask Depth 71.55 99.38 26.16 28.51 34.96 33.28 8.76 6.61 Inside #Bids 8.39 7.60 6.07 5.28 8.20 7.44 3.55 2.42 Inside #Asks 8.24 7.25 6.07 5.64 8.15 7.47 3.64 2.41 Market Impact 0.0691 0.0361 0.2192 0.1054 0.0924 0.1611 0.7211 0.9416
Spreads and market impact have fallen by 75%.
Mizrach/Neely Tatonnement24
IV. Information Shares
Mizrach/Neely Tatonnement25
Comparison of Information Shares
Mizrach/Neely Tatonnement26
Summary of Information Share Results
These are upper bound estimates (placing spot market first in orthogonalization) of bimarket comparisons between spot and the closest maturity future.
The spot markets all have information shares below 50% by 2001, with the decline really accelerating after 1999. Futures vital to price discovery process.
H and GG measures have very similar qualitative results.
Mizrach/Neely Tatonnement27
Information Share in ECNs
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
12-O
ct-0
4
19-O
ct-0
4
26-O
ct-0
4
2-N
ov-0
4
9-N
ov-0
4
16-N
ov-0
4
23-N
ov-0
4
30-N
ov-0
4
7-D
ec-0
4
14-D
ec-0
4
21-D
ec-0
4
28-D
ec-0
4
Shar
e
HMW
Havg.
BrokerTec 10-year information share. BTec > 50% except for 30-year.
Mizrach/Neely Tatonnement28
What Explains Information Share
We model IS share as a function of:
lnIS1,t c b1 lnN1,t/N1,t N2,t b2 lnS1,t/S1,t S2,t b3Di,t .
Where N1 is the number of trades in market 1, S1 is the percentage spread in market 1, and Di is a dummy variable for 8 different sets of macroeconomic announcements.
Trades and spreads explain 10-15% of the differences in information shares. (Not bad for microstructure studies).
Macroeconomic announcements are rarely significant. Only the employment report (on 2 occasions) is significant.
Mizrach/Neely Tatonnement29
Full System Estimation
The GG story is a little cleaner: by 2001, the 10-year and 30-year futures have the dominant information shares.
Mizrach/Neely Tatonnement30
V. Bayesian Structural Estimation
Mizrach/Neely Tatonnement31
State Space Representation
pt Hx t
x t Fx t 1 v t ,
H IN N , x t pt
ut
We are interested in estimation of the structural parameters α, σ², Ω. Parameters are estimated by MCMC, drawing the variance-covariance matrix of vt and computing α, σ² and Ω using this matrix.
We also obtain confidence measures on these estimates from the Markov chain Monte Carlo iterations. These are much less ad hoc than sample averages of daily estimates and/or the upper lower bound estimates from the Hasbrouck orthogonalization.
F 1 01 N
0N 1 0N N
, v t 1 01 N
IN N
t
e t
, Ev tv t
2
2
2
2
For the HUC model:
Mizrach/Neely Tatonnement32
Information Shares – Mapping From Structural Model
0 E p t p t
2 2 ,
1 E p t p t 1
2 .
#
#
p t t C t 1 ,
C I .
0 C C ,
1 C .
#
#
Var t 2 ,
Cov p t , t 2 ,
Cov p t , p t 1 I 2 .
#
#
#
2 ,
1 2 .
#
#
Structural autocovariances: Reduced form:
Moments matched:
Solution:
IS derived from these:
Mizrach/Neely Tatonnement33
Information Shares
All of the information share measures in the literature: (1) Hasbrouck (H); (2) Gonzalo-Granger (GG); (3) deJong and Schotman (DS); (4) Yan and Zivot (YZ) can be redefined in terms of parameters of the structural model.
We also obtain confidence measures on these estimates from the Markov chain Monte Carlo iterations. These are much less ad hoc than sample averages of daily estimates and/or the upper lower bound estimates from the Hasbrouck orthogonalization.
Caveats:(1) GG Information shares can be negative.
(2) Hasbrouck shares are positive by construction, but can give the largest IS to a market which moves prices away from the efficient price.
Mizrach/Neely Tatonnement34
Conclusion
Information shares are a useful summary statistic of the relative importance of market structures that are fragmented.
Treasury market: Ignore the futures at your peril. ECNs and black boxes makes this market potentially more volatile. Supply constraints (reopenings, off-the-run, close substitutes like corporates) make these concerns second order in my view.
Direct estimation of the structural model seems to be the best way to go forward in this literature.
Despite strong identification assumptions, these measures correlate well with observable liquidity.