Oce Lifecycleofexchange

The Lifecycle of Exchange-traded Derivatives
Grant Cavanaugh Ph.D.

Praedicat
grant.cavanaugh@praedicat.com
Michael Penick
US Commodity Futures Trading Commission
mpenick@CFTC.gov
June 28, 2014
The research presented in this paper was co-authored by a Commodity
Futures Trading Commission (CFTC) economist Michael Penick with coauthor, Grant Cavanaugh, a former CFTC contractor who performed work
under CFTC OCE contract (contract CFOCE-12-CO-0194). The analyses
and conclusions expressed in this paper are those of the authors and do not
reflect the views of other members of the Office of Chief Economist, other
Commission staff, or the Commission itself. The CFTCs economists produce
original research on a broad range of topics relevant to the CFTCs mandate
to regulate commodity future markets, commodity options markets, and the
expanded mandate to regulate the swaps markets pursuant to the DoddFrank Wall Street Reform and Consumer Protection Act. These papers are
often presented at conferences and many of these papers are later published
by peer-review and other scholarly outlets.
Abstract
This article provides a statistical description of the lifecycle of exchangetraded derivatives in the United States. Using annual volumes for most
derivatives reported to US exchanges since 1954, we present distributional estimates of the rate at which derivative trading volumes rise
and fall. Our results suggest that the lifecycle of cleared derivatives
changed fundamentally in the 2000s. In that decade, derivatives with
low trading volumes moved to modest volumes with increased probability. Prior to the 2000s, low volume contracts were more likely to
remain stuck at low volumes or be delisted altogether. This additional
resilience from low levels of trading meant that the expected trading
volume for a new cleared derivative after ten years of trading actually
grew between the 1990s and 2000s. This is surprising given that many
new contracts were launched in the last decade and a historically large
percentage of contracts traded at low volume in any year. The results
suggest that trading volumes varied more decade to decade than from
exchange to exchange or product type to product type.
JEL classification: G12, N22, G17.

Keywords: Derivatives; Innovation; Financial History.
1.
Introduction
Every year, financial innovators launch new derivatives markets in the

hope of serving unmet risk management needs stemming stemming from ongoing economic activity. Many of those new markets fail to attract the liquid
trading that is necessary both to generate profits for derivatives exchanges
and to drive down transaction costs for their customers. A handful of new
markets will host large volumes of trading, enjoying network effects that will
sustain further trading in the future.
This article is intended to answer the fundamental question facing financial innovators as they consider whether or not to offer a new derivative
contract for trading on their platform: what are the chances that a new
derivative will reach a sustainable level of liquidity?
Silber (1981) and Carlton (1984) provided the first summary statistics
on the survival of new futures contracts and they suggested that most new
derivatives fail, usually soon after their launch. However, since their publication, technological innovation and organizational changes at derivatives
exchanges altered the economics of derivatives trading in ways that may have
upended long standing patterns in product lifecycles (Gorham and Singh,
2009).
Recently, Gorham and Kundu (2012) used a large dataset from the Futures Industry Association (FIA) to demonstrate a steep increase in the rate
at which new futures contracts are launched. They also provide point estimates of multiple metrics for the success of new futures contracts.
Here we extend the more recent work in Gorham and Kundu (2012),
providing distributional estimates of exchange-traded derivatives contracts
movement between states of annual trading volume using a dataset that
includes cleared derivatives and options as well as many historic futures contracts absent from most electronic databases. The analysis presented is comprehensive, covering all US exchange-traded derivatives contracts launched
since the mid-1950s. The analysis is also presented in a simple and novel
1
form, as a non-stationary Markov model estimated using Bayesian methods.
1.1.
Derivatives reform and lifecycle statistics
Basic statistics on the lifecycle of derivatives are particularly valuable now

because ongoing policy debates on derivatives regulation in the US and Europe have hinged on projections of how new regulations will impact liquidity
and trading patterns. Better baseline statistics of the lifecycle of derivatives,
particularly statistics that take into account recent shifts in the dynamics of
exchange reported derivatives trading, can inform that debate.
Title VII of the Dodd-Frank reforms focuses on swaps markets1 2 , the
hitherto unregulated derivatives markets that, since the first publicly disclosed swaps trade in 1981, had grown to a notional outstanding value of
USD 639 trillion by June 2012. (By contrast, options and futures had a
combined notional outstanding value of USD 60 trillion (Bank of International Settlements, 2012).) Title VII mandates that swaps markets adopt
practices related to many critical market functions, including information
dissemination, counter-party risk, and margining, that are comparable to
those of exchange-traded futures and options.
This regulatory change suggests that the coming years will see convergence between previously unregulated swaps markets and standard exchangetraded derivatives markets. This convergence, in turn, raises both normative
1
Swaps trades have generally been negotiated bilaterally, often over the phone through
or with large swap dealers, rather than via the central limit order book system used by
exchange-traded derivatives. In theory, markets using bilateral negotiation and central
order books trade contracts that imply equivalent cash flows (Mello and Reilly, 2012).
However, Duffie (2012) presents evidence that informational asymmetries persist in even
highly liquid bilateral markets.
2
In practice, the distinction between swaps and futures is often murky. For example,
some swaps trades are negotiated bilaterally and then converted into futures trades on
markets such as the CME Groups ClearPort. Those trades are reported to exchanges and
are consequently included in the dataset used in this article. The CME and ICE, the two
largest US futures exchanges, have recently announced plans to convert many of their most
popular swaps markets into futures markets with physical delivery of swaps contracts at
settlement (i.e. futures trades that become swaps), providing yet another hybrid model.
questions (How desirable is the move toward increased clearing, public disclosure of pricing information, and greater standardization of margins?) and
positive questions (What will the likely costs or regulation be in terms of
trading volume?) that would benefit from reliable statistical descriptions of
the lifecycle of derivatives.
The relative scarcity of basic statistics on the lifecycle of derivatives has
already introduced confusion into the policy debate surrounding Title VII.
In one prominent example, the International Swaps and Derivatives Association (ISDA) released a position paper on regulations mandating price transparency and clearing in swaps markets comparable to that in exchange-traded
derivatives markets in late 2011 (ISDA Research Staff and NERA Economic
Consulting, 2011). The paper highlights previous research showing high rates
of failure among exchange-traded derivatives. Assuming a connection between those failure rates and exchanges price transparency and clearing, the
paper goes on to argue that swaps contracts subject to proposed regulations
would subsequently lose their liquidity and begin to fail. That suggestion is
misleading. First, it ignore the comparable failure rates for bilateral swaps,
which are difficult to quantify. Second, it relies on the assumption tested
here - that derivatives continue to fail at the rates documented decades ago.
Our results suggest that assumption is not robust to recent changes in the
underlying structure of cleared derivatives markets.
In addition to providing common ground for policy debates, we hope that
the following analysis will inform the decisions of derivatives innovators. In
general, contracts are showing greater flexibility, moving up from low levels
of annual trading. This may have implications for how exchanges allocate
their limited budgets for marketing and education. Contracts previously
considered too uneven in their year-to-year trading to succeed may indeed
have substantial growth potential given proper marketing and educational
support.
2.
Data
Our analysis is based on annual volume figures for US exchange-traded

derivatives (primarily futures, options, and cleared swaps). These figures
are/were freely available to the public through trade publications, directly
from exchanges, in newspapers, and from the website of the Commodity
Futures Trading Commission (CFTC). For ease of access, we used:
An electronic database maintained by the CFTC aggregating basic,
market-level daily trade data (such as volume and open interest) regulated futures and options exchanges, called designated contract markets
(DCMs), This dataset covers all recent trading volumes reported to US
exchanges of futures, options and swaps, cleared pursuant to DCM
rules. Most contracts in that database have volume figures dating back
to the early 1980s.
We supplemented this basic dataset by adding in futures trading figures
compiled by hand from historical publications released by derivatives
exchanges. The resulting dataset includes many short-lived contracts
listed on now-defunct exchanges that are unlikely to appear in most
electronic databases of trading statistics.
The merger of these resources may represent the most comprehensive
dataset on derivatives trading volume to date.
3.
Markov model for the lifecycle of derivatives
We present our primary results in the form of a Markov model. That

model begins by imagining that a derivative contract moving between discrete
states (x) of trading volume at discrete times (t as in xt )according to a
discrete-time Markov chain, defined generally as in equation 1.
4
P (Xt = xt |Xt1 = xt1 , Xt2 = xt2 , . . . , X0 = x0 ) = P (Xt = xt |Xt1 = xt1 )

(1)
In the context of derivatives, the left side of equation 1 can be restated
as in equation 2.
P (Volume levelyear t+1 |Volume levelyear t )
(2)
Contract are assigned a martix (P as in equation 3) that describes the

probability of moving to any of a set of discrete states (time j) of annual
trading volume in the following year given their state of trading volume
today (time i). This is the transition matrix commonly used to describe a
Markov process (Page, 2012).
p1,1 p1,2 . . . p1,j
p2,1 p2,2 . . . p2,j
..
..
.
...
P = ..
.
.
pi,1 pi,2 . . . pi,j
..
..
..
..
.
.
.
.
...
. . .
..
.
. . .
..
.
(3)
Volume level for any given contract-year is equivalent to the common

logarithm of the annual trading, rounded down to the nearest integer. (For
example, annual trading of 10, 500 is assigned a volume level that groups it
with all contract-years with volume 10, 000 and < 100, 000.) We assigned
a special level for annual trading of 0.
For ease of estimation we work with the rows of the transition matrix P
which we denote as . Those rows sum to 1, so, assuming that row entries are
randomly distributed, each row can be assigned a Dirichlet distribution, commonly used for the probability of ending in an exhaustive set of categorical
states. That assignment is defined in equation 4.
Volume levelyear t+1 |Volume levelyear t Categorical()
(4)
Dirichlet(xvol level 0 , xvol level 1 , . . . , xvol level 108 )

We modeled these transition probabilities via Bayesian Gibbs sampling
through R and the Bayesian statistical package JAGS (Plummer, 2003). (We
used the rjags package (Plummer, 2013).) These methods treat the underlying probabilities of moving between states of trading volume as randomly
distributed parameters, as in equation 4.
After estimation, we combine the vectors to reconstruct the transition
matrix for a Markov model P . As with any Markov model, we can multiply
a vector, 0 describing the probability that a new derivative will start in
any given state (at time 0) by the transition matrix to produce a vector of
probabilities that a new market will be in any state over an arbitrary number
of periods (k) as in equation 5.
0P k = k .
(5)
We can multiply the vector k by yet another vector of annual trading

volumes corresponding to each possible state to get an approximation of the
expected trading volume in that arbitrary year. We present all our expected
trading volumes at a ten year horizon (setting k = 10), but the Markov
model is flexible in this regard.
Note that we do not assume that the transition matrix P is stationary
across time. We measure transition matrices for various contract groupings
including decades, product categories, and exchanges to test whether they
are distinct. Given that we do not assume stationarity our expected value
estimates do not describe an equilibrium, only the general direction of the
market.
3.1.
Prior probabilities on moving between states of annual volume
Our model presumes that the data on the volume level next year (Volume levelyear t+1 )
is segregated by the volume this year (Volume levelyear t ) and we assigned
each of those subsets prior probabilities (corresponding to parameter x in
equation 4) of moving to any volume level in the next year. Those priors
came from an informal survey of economists at the CFTC.
That survey found beliefs corresponding roughly to:
Pr(Volume levelyear t+1 = Volume levelyear
t1 ) = 0.16
t ) = 0.63
t+1 ) = 0.14
The probability of a contract jumping more than one order of magnitude

up or down was assigned a value of 0.01. In edge cases (Volume levelyear t =
0 and Volume levelyear t = 108 ) where a move up or down would take the
contract below annual trading of 0 or to annual trading 109 , we combined
the probabilities of moving up or down with the probability of remaining in
the same state. Table 2 at the end of this article shows the full matrix of
transition probability priors.
We chose to assign informative priors on transition probabilities because
flat priors (equal weighting to the probability of a transition to any state)
unfairly biased the estimation, giving exchanges or product subgroups with
few observations a relatively high probability of jumping to extraordinary
levels of trading.
Empirical CDF
1.00
0.75
0.50
0.25
0.00
0e+00
2e+08
4e+08
6e+08
Annual volume
Fig. 1. Empirical cumulative distribution function of annual trading volumes
by contract
4.
Derivatives volumes over time
4.1.
Concentration of trading volume over time
Figures 1 and 2 display the empirical cumulative distribution function

(ECDF) of annual trading volumes by contract for every year in the sample.
In each figure, individual lines represent the ECDFs for a single year, with
lines approaching a right angle showing greater concentration of trading volume in a few contracts. Figure 1 clearly shows that most contracts trade at
low volumes in any given year, with roughly 80 percent of contracts showing
little or no volume in any given year since 1954.
However, figure 1 obscures substantial variation in the concentration of

volume over time. Figure 2 zooms in on the same annual ECDFs displayed
in figure 1. The ECDF for each year is colored chronologically, with the lines
representing the oldest years in the sample in red and the most recent years
in purple. Each panel of figure 1 shows the same ECDFs, but the years in a
specific decade are highlighted (in black) to give a sense of how concentration
has varied from decade to decade.
In this graphic we see clear patterns in concentration over time. Markets grew steadily less concentrated between the 1950s and 1990s (perhaps
with some retrenchment between the 1980s and 1990s), shown by flattening
ECDFs for each succeeding decade. That trend reversed sharply in the 2000s,
with the annual ECDFs approaching a right angle. In the 1980s the range
of 15,000 to 30,000 roughly marked the 50th percentile for annual trading
volumes, with half of the listed contracts trading above that range and half
below. By the 2000s that range had fallen to between 300 and 8,000.
Figure 2 itself highlights one likely cause of this shift - the explosion
of innovation during the 2000s. The ECDFs for the 2000s are appreciably
smoother than those of previous decades, with 2011 looking almost like a
continuous function. This smoothness is due to the inclusion of additional
contracts. Figure 3 directly displays the number of contracts with annual
reported volume (which is allowed be zero) in the sample by year. It shows the
same explosive trend in innovation discussed in Gorham and Kundu (2012),
with over 3000 derivatives contracts reporting annual volume in 2011.
4.2.
Probability of individual contracts moving to different levels of trading by decade
Figures 4 and 5 give the probabilities of individual contracts moving between volume levels in a given year t (indicated by the row of estimates)
and volume levels in year t+1 (indicated by the column of estimates). These
probabilities, estimated separately for each decade in the sample via equation
9
1960's
1970's
Year
2010
1980's
Empirical CDF
1950's
1.0
0.9
0.8
0.7
0.6
0.5
1.0
0.9
0.8
0.7
0.6
0.5
1.0
0.9
0.8
0.7
0.6
0.5
1.0
0.9
0.8
0.7
0.6
0.5
1.0
0.9
0.8
0.7
0.6
0.5
1.0
0.9
0.8
0.7
0.6
0.5
1.0
0.9
0.8
0.7
0.6
0.5
2000
1990
1980
1970
1960
1990's
2000's
2011
0e+00
1e+06
2e+06
3e+06
4e+06
5e+06
Annual volume
Fig. 2. Empirical cumulative distribution function (ECDF) of annual trading

volumes by contract with scale adjusted to distinguish between decades Each line represents the ECDF for a different year. Each of the stacked
panels highlights the years in a particular decade in black. Note that an
ECDF approaching a right angle represents a year in which volume was
concentrated in a few contracts. Hence, with some exceptions in the 1990s
the market as a whole becomes less concentrated, until the 2000s when it
abruptly becomes highly concentrated.
10
Derivatives in sample
3000
2000
1000
0
1960 1970 1980 1990 2000 2010
Year
Fig. 3. Number of contracts in sample by year
11
4, combine to form the transition matrix for a Markov model of a contract

emerging over time.
The parameter estimates indicate that there is substantial inertia across
every decade keeping contracts with a given level of trading volume at that
same volume in the following year. In virtually all decades in the sample,
contracts trading at or above 1,000 in annual volume were more likely to
remain at their trading volume level than to move up or down. This dynamic
is particularly strong at higher levels of trading. In most decades where
relevant observations were available, contracts with annual volume of one
million or above remained in that range the following year with probabilities
between 80 and 90 percent (see lower right-hand corner of figure 5).
We also see substantial historical evidence of inertia at very low levels of
trading. From 1970 until 2000, the median probability that a contract with
trading volume of zero would remain at zero the next year, ranged between
80 and 95 percent (see upper left-hand corner of figure 4).
The transition matrix begins to depart from the prevailing story in Silber
(1981) and Carlton (1984) when you look at contracts at lower levels of
trading in the 2000s. (See the top rows of figure 4.) The inertia for those
contracts is lower than in previous decades, with the median probability of a
contract at an annual volume of zero remaining at zero falling to 70 percent
(figure 7). While zero volume contracts remained unlikely in absolute terms
to rise to higher volume levels, the 95 percent probability interval for the
transition probability for the 2000s does not overlap with those for recent
decades, meaning that the difference holds with high probability.
During the decade of the 2000s, contracts were substantially more likely
to jump from an annual trading volume of 0 to trading volumes between 10
and 1000 than in previous decades. (See top row of figure 4 and 8.) Combined
with the apparent trend toward maintaining rather than delisting contracts,
this suggests that there was less path-dependence for trading volumes in the
2000s. While more contracts traded at a volume of 0 in any given year (see
12
0.15
0.2
0.29
0.08
0.3
0.16
0.21
0.31
0.09
0.23
0.31
0.1
0.43
0.22
0.07
0.19
0.33
0.12
0.11
0.12
0.24
0.49
0.04
0.08
0.08
0.11
0.02
0
0
0.11
0.7
0
0.2
0.03
0.17
5
0.53
0
5
0.12
0.5
0.25
0.2
0.2
0.15
0.0
0.0
1.0
0.7
0.14
0
0.5
0.2
0.0
1.0
0.7
0.5
0.01
0.29
0.05
0.01
0.2
0.17
0.08
0.06
0.4
0.01
0.21
0.04
0.18
0.14
0.1
0.48
0.03
0.18
0.04
0.02
0.0
0.15
0.1
0.04
0.02
0.45
1980
0.25
0.14
0.07
0.01
0.01
vol 1000's, t
1990
0.09
0.1
0.56
1.0
0.02
0.03
0.05
0.02
0.06
0.27
0.2
0.11
0.13
0.29
0.15
0.2
0.05
0.02
0.09
0.02
0.17
1.0
0.13
0.7
0.06
0.7
0.25
0.5
0.29
1970
0.5
0.03
vol 100's, t
0.05
0.01
0.13
0.23
1980
0.02
0.0
1990
1.0
0.26
0.16
0.01
0.01
0.01
0.1
0.02
0.05
2010
0.01
0.05
0.06
0.03
0.62
0.55
0.13
0.12
0
0
0.16
0.05
0.7
Decade
vol 10's, t
0.11
0.58
1950
0.02
0.01
0.09
0.51
0.0
0.04
0.27
1.0
0.11
0.22
0.22
0.7
0.15
1960
1950
0.5
0.08
1970
1960
0.04
1980
2010
0.05
0.09
0.21
1970
0.09
0.1
0.06
0.32
1990
2000
0.28
0.47
1950
0.59
1960
0.02
2010
0.5
0.46
1950
0.06
0.14
0.2
0.14
0.1
0.08
0.2
1960
0.75
0.02
1970
0.08
0.03
0.04
0.0
0.09
0.12
0.08
0.45
0.2
0.02
1.0
0.15
vol 1's, t
1980
0.14
0.61
0.02
0.48
1990
0.06
0.0
0.15
0.18
1.0
0.58
0.7
0.2
0.5
0.08
0.1
0.02
0.2
0.02
0.04
0.53
2000
2000
0.06
0.13
2010
0.01
0.0
0.13
0.05
0.6
1950
2000
0.01
0.89 0.02
1960
0.01
0.01
0.7
0.03
0.02
0.03
0.01
0.5
1970
0.06
0.03
vol 10^8's, t+1
0.2
0.93
vol 10^7's, t+1
0.04
1980
0.04
vol 10^6's, t+1
0.0
0.01
0.04
0.82
0.08
vol 10^5's, t+1
vol 0, year t
1990
0.14
0.13
vol 10^4's, t+1
1.0
0.15
0.03
1.0
0.17
0.7
vol 1000's, t+1
0.7
0.46
2000
vol 100's, t+1
0.7
vol 10's, t+1
0.5
vol 1's, t+1
0.5
vol 0, year t+1
2010
Pr(yearonyear move)
Fig. 4. Transition matrix for Markov model of derivatives contract moving

between states of annual trading volume by decade - row represents state
in year t, column represents state in year t + 1, median estimate indicated
by dot, 95 percent probability interval
13indicated by line - part 1: transitions
given annual volumes 0 and < 10, 000
0.08
0.92
0.02
0.02
0.92
0.92
0.02
0.02
0.92
0.7
0.5
0.2
0.0
1.0
0.7
0.92
0.5
0.14
0.84
0.94
1.0
0.06
0.2
0.01
0.74
0.01
0.0
0.0
1.0
0.7
0.5
0.2
0.0
0.02
0.01
0.74
0.74
1.0
0.02
0.02
0.7
0.81
0.17
1950
0.91 0.04
0.5
0.02
1960
0.92 0.04
0.2
0.93
0.02
0.07
1970
0.0
0.15
1980
0.77
1.0
1990
0.7
0.83
vol 10^8's, t
2000
0.05
0.15
0.94 0.02
0.5
2010
0.8
0.91 0.02
0.2
1950
0.08
1960
0.01
0.02
0.0
0.09
0.78
1.0
0.81
0.7
1970
0.03
0.19
vol 10^7's, t
0.03
1980
0.5
1990
0.06
0.2
2000
0.04
0.0
2010
0.05
0.08
0.02
0
0.02
1950
1.0
0.01
1960
0.01
0.7
0.84
0.83
0.12
0.01
0.1
0.11
0.2
0.01
0.0
1970
Decade
1980
0.14
0.84
0.7
0.06
0.07
0.05
0.07
0.81
vol 10^6's, t
1990
0.09
0.01
0.67
0.0
2000
0.22
0.01
0.72
0.17
0.09
0.04
1.0
2010
1.0
0.01
0.85 0.05
0.7
1950
0.02
0.5
0.02
0.2
1960
0
0.01
0.15
0.24
0.7
0.01
0.58
0.04
0.13
0.1
0.01
0.61
vol 10^5's, t
0.09
0.13
0.07
0.09
0.01
1970
0.71
0.7
0.0
0.01
1980
0.5
0.01
vol 10^8's, t+1
0.13
vol 10^7's, t+1
0.09
vol 10^6's, t+1
0.1
0.04
0.56
0.61
0.01
0.04
0.01
1990
0.01
2000
0.02
0.25
vol 10^5's, t+1
0.21
0.02
1.0
0.01
2010
0.01
1950
0.01
0.04
0.03
1960
0.04
0.05
1970
0.01
0.7
0.05
1980
0.01
vol 10^4's, t+1
vol 10^4's, t
1990
vol 1000's, t+1
0.5
0.04
vol 100's, t+1
0.2
0.05
2000
vol 10's, t+1
0.5
vol 1's, t+1
0.2
vol 0, year t+1
2010
Pr(yearonyear move)

between states of annual trading volume by decade - row represents state
in year t, column represents state in year t + 1, median estimate indicated
by dot, 95 percent probability interval
14indicated by line - part 2: transitions
given annual volumes 10, 000
vol 0, year t+1

2010
0.46
0.7
1990
0.82
1980
0.93
1970
0.89
1960
vol 0, year t
Decade
2000
0.6
1950
0.53
0.00
0.25
0.50
0.75
1.00
Pr(yearonyear move)
Fig. 6. Probability of remaining at annual volume of zero from year to year
by decade
15
vol 0, year t+1

2010
0.46
0.7
1990
vol 0, year t
Decade
2000
0.82
1980
0.93
1970
0.89
1960
0.6
1950
0.53
0.00
0.25
0.50
0.75
1.00
Pr(yearonyear move)
Fig. 7. Probability of remaining at annual volume of zero from year to year
by decade
vol 1's, t+1

2010
0.01
0.02
1960
Decade
1970
0.03
0.13
1950
1980
0.02
0.03
1970
0.14
0.09
1950
0.1
0.00
1960
0.2
1990
0.25
0.50
0.75
1.00
0.21
0.00
Pr(yearonyear move)
vol 1's, t
1980
0.15
2000
vol 0, year t
Decade
2000
1990
vol 10's, t+1

2010
0.17
0.25
0.50
0.75
1.00
Pr(yearonyear move)
Fig. 8. Probability of transition from annual volume of 0 to annual volume in

the single digits (left) and from annual volume in the single digits to annual
volume 10, < 100 (right)
16
vol 100's, t+1

2010
0.11
1960
0.05
Decade
0.13
1980
0
0.08
0.00
0.25
0.50
0.75
1.00
0.12
1980
0.15
1970
1950
0.31
0.22
1990
1960
1950
0.1
0.29
0.06
0.09
0.00
Pr(yearonyear move)
vol 100's, t
1990
2000
0.27
vol 10's, t
Decade
2000
1970
vol 1000's, t+1

2010
0.25
0.50
0.75
1.00
Pr(yearonyear move)
Fig. 9. Probability of transition from annual volume in the hundreds to annual volume in the thousands (left) and from annual volume in the thousands
to annual volume tens of thousands (right)
figure 2), contracts were substantially more likely to jump up from such low
trading volumes in the 2000s.
Having reached annual trading volumes in the 10s or 100s (see 9), contracts in the decade of the 2000s were again substantially more likely to
continue increasing their trading volume in the 2000s than in the 1980s or
1990s. Only after reaching trading volumes in the 1000s (figure 10) did the
probability of an individual contract progressing to higher levels of annual
trading volume fall roughly back within the same range as those from previous decades. In the 2000s, contracts generally moved up to annual trading
in the thousands with an ease not seen in previous decades.
Contracts trading in the tens of thousands were 8 percent more likely
to fall back to lower levels of annual volumes in the 2000s than in previous
decades, a difference that holds with high probability. This indicates that
some of the flexibility gained for contracts at lower levels of trading may have
come at the expense of contracts at mid to high levels of trading. (However,
as we see in figure 13, discussed below, that retrenchment from trading in
the tens of thousands was not enough, on balance, to lower the prospects of
a new contract over the course of ten years.)
Annual trading in the 10,000s appears to represent an important milestone for contracts across the sample. Having reached this level of trading,
17
vol 10^4's, t+1

2010
0.1
0.18
1990
vol 1000's, t
Decade
2000
0.17
1980
0.14
1970
0.12
1960
0.11
1950
0.11
0.00
0.25
0.50
0.75
1.00
Pr(yearonyear move)
Fig. 10. Probability of transition from annual volume in the tens of thousands
to annual volume in the hundreds of thousands
18
volume level vol
vol 0s, t
10
100
1000
10
10
10
107
8
10
0s, t+1
0.66
0.56
0.42
0.20
0.10
0.04
0.01
0.01
0.01
0.00
1
0.08
0.16
0.12
0.05
0.01
0.00
0.00
0.00
0.00
0.00
10
0.11
0.16
0.21
0.17
0.05
0.01
0.00
0.00
0.00
0.00
100 1000 104 105 106 107 108
0.09 0.03 0.02 0.00 0.00 0.00 0.00

0.08 0.03 0.01 0.00 0.00 0.00 0.00
0.18 0.06 0.02 0.00 0.00 0.00 0.00
0.38 0.17 0.03 0.00 0.00 0.00 0.00
0.20 0.47 0.15 0.01 0.00 0.00 0.00
0.03 0.19 0.62 0.10 0.00 0.00 0.00
0.00 0.02 0.14 0.74 0.08 0.00 0.00
0.00 0.01 0.01 0.08 0.84 0.04 0.00
0.00 0.00 0.00 0.00 0.05 0.90 0.03

0.00 0.00 0.00 0.00 0.00 0.13 0.87
Table 1: [
180pt]Median estimates of transition matrix between volume states on full
sample - with annual trading volume state in year t denoted by row, trading
volume state in year t denoted by column
the likelihood of outright collapse (annual trading volume falling to 0 in the
next year) fell to very low levels and was largely indistinguishable across the
decades (figure 11). Table 1 presents the median estimates of transition probabilities estimated across the full sample (i.e. aggregating across decades).
They show clearly that having reached annual trading of in the 10,000s, a full
collapse becomes relatively unlikely (4 percent). In fact, for contracts that
achieve annual trading in the 10,000s, the probability of falling more than
one volume level is below 10 percent. (See the sixth row of table 1.) Note
that these full sample estimates are biased toward recent decades because
the sample contains more observations from recent decades.
As suggested above, one hypothesis regarding the recent shift in derivatives lifecycles is that the additional flexibility that low volume contracts
enjoyed in the 2000s came directly at the expense of mid-range to higher
volume contracts. In volatile markets, hedgers might be choosing niche contracts with lower basis risk over more liquid cross-hedges. What would that
mean for the overall outlook for lifetime trading of derivatives? We test this
19
vol 0, year t+1

2010
0.05
0.04
1990
0.05
1980
1970
vol 10^4's, t
Decade
2000
0.05
0.03
1960
1950
0.01
0.01
0.00
0.25
0.50
0.75
1.00
Pr(yearonyear move)
Fig. 11. Probability of transition from annual volume in the tens of thousands
to annual volume of 0
20
by looking at the expected trading volume of a new derivative over the course
of ten years.
Combining draws from the transition matrix in figures 4 and 5 with draws
from a vector representing the probability of a contract starting in each of
the available states of annual trading volume (estimated using the same basic
model presented in equation 4) we can get the probability that a new contract
will be in any given state of volume after ten years of trading. Those values
are displayed in figure 12. Figure 12 makes clear the resilience of contracts
trading at low levels in the 2000s. Only 32 percent of contract that debuted
with zero volume were still trading at zero volume after ten years in the
simulation representing the 2000s. Those probabilities were 46, 48, and 52
percent in the 1990s, 1980s and 1970s respectively (See the first column
of boxes in figure 12).3 Instead of languishing, contracts simulated from the
2000s were more likely to migrate over ten years to moderate levels of trading.
(See the columns of boxes in figure 12 corresponding to annual trading volume
between 100 and 10,000.) Those same contracts were, however, less likely to
reach the highest levels of trading ( 100, 000) than contracts from other
decades. The 1980s appears to be the best decade for such blockbuster
contracts, as suggested in Gorham and Kundu (2012).
Simply comparing the raw probabilities of reaching various levels of volume after ten years, it is difficult to discern which decade provided a better
overall environment for new contracts. To make that comparison, we normalize the probabilities in figure 12 by the lower bound of each trading range (i.e.
multiplying the probability of being in the trading state 100 and < 1, 000
by 100). This give an approximation of the expected trading volume of a
new contract after ten years, displayed in figure 13. Based on that graph, we
can conclude:
3
Note these simulated values simply describe the dynamics of the transition matrices
when compounded. They ignore delisting. If we accounted for delisting, a practice that
was more common in previous decades, the probabilities of failure would likely be higher
for those decades.
21
0.6
0.2
0.4
0.2
0.0
0.6
0.4
0.14
0.22
0.08
0.13
0.13
0.09
0.11
0.02
0.1
0.02
0.02
0.01
0.03
0.02
0.05
0.05
0.02
0.04
0.09
0.08
0.05
0.07
0.15
0.48
0.08
0.12
0.17
0.04
0.46
1990
0.0
0.6
0.15
0.1
0.04
0.01
0.01
0.32
0.03
0.1
0.16
0.17
0.13
0.05
0.02
2010
2000
0.0
0.6
0.4
0.2
0.2
0.01
0.52
0.0
0.6
0.4
0.1
0.2
0.4
0.1
1980
0.4
0.21
0.11
0.33
0.2
0.0
0.6
0.05
1970
Pr(Being in a given state after 10 years
1960
0.0
0.6
0.15
1950
0.4
0.35
0.2
0.0
0.12
0.14
0.15
0.1
0.07
0.04
0.02
l
vo
l
vo
l
vo
l
vo
l
vo
l
vo
l
vo
l
vo
l
vo
l
vo
's,
^8
10
's,
^7
10
's,
^6
10
's,
^5
10
's,
's,
^4
10
00
10
r1
a
ye
r1
a
ye
r1
a
ye
r1
a
ye
r1
a
ye
r1
a
ye
r1
r1
a
ye
ea
,y
0's
10
's,
10
r1
r1
a
ye
ea
,y
1's
0,
Volume level
Fig. 12. Box and whiskers plot of probability of a new contract being at
different levels of trading after 10 years by decade - median simulated probability marked in text, upper and lower
22 hinges of the box plot correspond to
the first and third quartiles (the 25th and 75th percentiles)
The expected trading volume after 10 years for a contract has varied
substantially from decade to decade;
There is no clear trend that emerges from these variations over time;
The expected trading volume at year ten for a contract in the 2000s
was firmly in the middle of the historical range - the 2000s were lower
than the 1980s, higher than the 1990s, and all three decades showed
substantial overlap with the earlier decades in the sample;
In the 2000s, low volume contracts tended to rise to modest levels of
trading, balancing any fall in the probability of reaching the highest
trading levels.
2000
Decade
1980
1960
1e+05
1e+06
Expected volume after 10 years

Fig. 13. Expected trading volume over ten years by decade
While a larger percentage of contracts were at low volumes in the 2000s
23
than in previous decades (figure 1), individual contracts were considerably

more likely to jump up from very low volumes to moderate volumes (figure
8). The net effect of these trends set the expected volume of contracts at
year ten well within the historical range of earlier decades (figure 13). This is
remarkable given the explosion in the number of contracts launched (figure 3).
It suggests that the marginal value of an innovative contract (approximated
by its expected trading volume at year ten) did not fall in the 2000s, despite
exponentially higher rates of innovation than in past decades.
This shift is consistent with the hypothesis that electronic trading made
trading activity more mobile across derivatives markets and substantially
cut the costs of launching and sustaining a derivatives contract. But changes
went above and beyond the introduction of electronic trading on US and
European exchanges in the 2000s, making it difficult to identify the causes of
product lifecycle shifts in aggregate statistics. For example, many of the new
contracts launched in the 2000s (and included in this sample) are bilaterallynegotiated, but centrally-cleared swaps. In the wake of Enrons collapse,
which threatened energy firms with counter-party defaults on their swaps
trades, exchanges launched popular new facilities devoted to these clearedswaps, including the CMEs ClearPort. While those contracts benefited from
a suite of tools associated with electronic trading, they were not subject to
electronic trading in the narrow sense of actually having buy and sell orders
matched on an electronic platform.
To isolate the influence of electronic trading, we look at contracts trading
on the New York Mercantile Exchange (NYMEX), where electronic trading
was introduced suddenly. The NYMEX does not offer an ideal natural experiment. Its trading patterns were likely influenced by the shift toward cleared
swaps throughout the 2000s. However, the abruptness of the exchanges
switch to electronic trading does offer some scope for teasing out the relative
import of electronic trading.
24
5.
Derivatives volumes by exchange
Differences in trading volume patterns over the life of a derivatives contract may be influenced by the exchange offering the contract. Carlton (1984)
hypothesized that economies of scale in designing and launching a contract
gave those on larger exchanges a relative advantage in terms of trading volumes. Similarly, there may be network effects stemming from an exchanges
ability to cross-margin trades.
Cuny (1993) and Holland and Fremault (1997) suggest that innovative
exchanges may enjoy a first-mover advantage, capturing a disproportionate
share of trading on those contracts that they launch. Gorham and Kundu
(2012) tests this hypothesis and finds little persistent advantage. In the
context of a Markov model of trading volumes, if indeed there is a firstmover advantage, then we would expect innovative exchanges to distinguish
themselves with higher expected trading in year ten.
Figure 14 presents expected volume in year ten for contracts on all exchanges in the sample. Contracts show greater distinction across decades (as
in figure 13) than across exchanges. It is possible to distinguish individual
exchanges from one another. For example, contracts on the Chicago Board of
Trade have an advantage over those on the NYMEX in expected value terms.
But no exchanges clearly distinguish themselves from the general tendency
with greater than 95 percent probability. Possible exceptions include:
the single-stock futures traded on OneChicago which show particularly
low expected trading volumes over ten years
the two registered exchanges in the IntercontinentalExchange group,
marked ICE and ICEU in figure 14, which likely have higher expected
trading volumes than most other exchanges. It is important to note
that these exchanges specialize in OTC markets, only a handful of
which have been reported to the CFTC as futures. Consequently, some
25
of their performance may represent selection bias.4
5.1.
CME acquisitions test the important of exchange to lifecycles
Recent exchange acquisitions offer the chance to test the effects of particular exchanges on trading volumes. Gorham and Kundu (2012) singled
out the CME as the exchange with a persistent advantage over its rivals leading other major exchanges in mean volume in the 5th year of trading,
mean lifetime volume, and their approximations of present value discounted
fee generation. In the late 2000s, the CME Group effectively5 took over both
the New York Mercantile Exchange (designated in the database as NYME
but commonly referred to as the NYMEX) and the Chicago Board of Trade
(CBT). After the acquisitions, the exchanges contracts continued to be reported as before (i.e. NYMEX contracts continued to be reported in the
dataset as NYMEX contracts).
If indeed the CME did enjoy a persistent advantage on multiple volume
metrics, then presumably the transition matrices for NYMEX and CBOT
contracts, calculated using the Markov models profiled here, would improve
following their acquisitions. These acquisitions could also test a weaker
form of that same hypothesis. If exchange management is important to
contract lifecycles, then the CBOT and NYMEXs contracts transition matrices should converge to the CMEs, regardless of whether the CME has an
advantage over other exchanges or not.
Figures 15 and 16 for the NYMEX and ??s figures ?? and ?? present
the transition matrices for each of the merged exchanges in the years before
4
In late 2012, the IntercontinentalExchange announced that many of its most popular
OTC contracts will begin trading as futures.
5
Technically, the CME and CBOT merged. However, the CME was the dominant firm
in the merger, initiating the transaction and retaining most of the key staff positions.
Olson (2010) provides an inside account of the fight between the CME and ICE for control
of the CBOT.
26
West Coast
Seattle
PCE
PBOT
PANY
NYPE
NYME
NYL2
NYL
NYBOT
NOCE
MilGX
Memphis
MGE
MESL
Exchange
MCE
KCBT
IEPA
ICUS
ICEX
ICEU
ICE
GE
EUS
EUR
ELX
27
CRCE
CMX
and after the merger.6

The CBTs transition matrices (??s figures ?? and ??) show no consistent
trends in post-merger years relative to the earlier years in the sample. Postmerger years with strong performance (contracts showing a high probability
of advancing to a higher level of liquidity - such as 2010, where many of the
contracts previously trading with annual volumes in the thousands advanced
to the tens of thousands) do not stand out relative to the pre-merger era. To
the extent that the CBT shows any post-merger trend, it stems from 2010,
an especially volatile year for the CBT, where many contracts advanced to
trading in the tens of thousands and a particularly large percentage fell back
from annual volumes in the tens of thousands.
Unlike the transition matrix for the CBT, the NYMEX shows a clear trend
in its transition probabilities. On the rows in figures 15 and 16 indicating
trading volume between 10 and < 1, 000, 000 (rows three through five in
figure 15 and rows one and two in figure 16), a gradual pattern in volume
level transitions emerges that is strong enough, by the end of the decade,
to hold with high probability. Starting roughly in 2006, the 10s, 100s, and
1000s became sinks (rows three through five in figures 15). The probability
of staying at these levels year on year increases gradually. The probability
of rising out of that range falls. At levels immediately above that sink (rows
one and two in figures 16), the probability of falling into the sink rises at the
clear expense of the probability of staying put or rising. This trend predates,
and is uninterrupted by, the CME merger.
Neither transition matrix support the hypothesis that exchange management is an important factor in lifecycle patterns, much less the hypothesis
that CMEs systems and network effects boost trading volumes substantially
relative to competing exchanges.
6
We chose to present the full transition matrix for the exchange-year comparisons rather
than the expected value figures because we believe that the former provide more robust
inference. Expected value calculations are sensitive to the initial trading volumes of the
contracts that happened to launch after the merger.
28
0.45
0.21
0.29
0.66
0.03
0.66
0.13
0.39
0.07
0.07
0.07
0.53
0.07
0.34
0.15
0.74
0.02
0.73
0.73
0.01
0.01
0.16
0.26
0.03
0.27
0.13
0.35
0.22
0.19
0.09
0.22
0.3
0.27
0.31
0.28
0.53
0.01
0.01
0.25
0.49
0.13
0.12
0.44
0.13
0.01
0.09
0.41
0.28
0.03
0.29
0.24
0.1
0.28
0.4
0.11
0.31
0.27
0.01
0.01
0.08
0.48
0.31
0.13
0.42
0.39
0.16
0.4
0.33
0.1
0.48
0.39
0.15
0.04
5
1.0
0
0.0
0
0.2
5
0.56
0.21
0
0
0.02
0.19
0.71
0.37
0.42
0.07
0.04
0.04
0.06
0.7
0.19
5
1.0
0
0.0
0
0.2
5
0.7
0.5
0.56
0.2
0.15
0.09
0.5
0.56
0.04
0.07
0.15
0.52
0.13
0.32
0.22
0.34
0.21
vol 1000's, t
0.1
0.07
0.39
0.05
0.13
0.0
0.21
0.13
0.08
0.06
0.16
0.41
0.11
0.04
0.33
0.16
0.29
0.07
0.37
0.02
0.21
0.11
0.01
0.06
0.01
0.19
vol 100's, t
0.3
0.5
0.11
0.3
0.18
0.5
0.25
0.5
0.25
0.25
0.04
0.07
0.1
0.01
0.11
5
1.0
0
0.0
0
0.2
5
0.5
0.09
0.44
0.25
0.04
0.7
0.5
0.05
vol 10's, t
0.06
0.36
0.29
0.34
0.04
0.32
0.01
0.44
0.02
0.29
0.15
0.77
0.36
vol 1's, t
0.18
0
0.55
0.5
0
0
0.35
1.0
0.7
0.94 0.01
0.5
5
1.0
0
0.0
0
0.2
5
0.7
0.5
0.98
0.5
5
1.0
0
0.0
0
0.2
5
0.7
0.18
0.5
0.08
0.46
0.5
0.69
0.11
0.5
0.05
0.03
0.5
0.13
5
1.0
0
0.0
0
0.2
5
0.81
0.22
vol 0, year t
5
1.0
0
0.0
0
0.2
5
5
1.0
0
0.0
0
0.2
5
0.7
0.02
0.03
5
1.0
0
0.0
0
0.2
5
0.11
0.7
0.15
0.58
ExchangeYear
0.14
0.23
NYME2010
NYME2009
NYME2008
NYME2007
NYME2006
NYME2005
NYME2004
NYME2003
NYME2002
NYME2001
0.02
0.07
NYME2010
NYME2009
NYME2008
NYME2007
NYME2006
NYME2005
NYME2004
NYME2003
NYME2002
NYME2001
0.2
0.02
NYME2010
NYME2009
NYME2008
NYME2007
NYME2006
NYME2005
NYME2004
NYME2003
NYME2002
NYME2001
vol 100's, t+1 vol 1000's, t+1 vol 10^4's, t+1 vol 10^5's, t+1 vol 10^6's, t+1 vol 10^7's, t+1 vol 10^8's, t+1
0.08
0.35
0.98
NYME2010
NYME2009
NYME2008
NYME2007
NYME2006
NYME2005
NYME2004
NYME2003
NYME2002
NYME2001
vol 10's, t+1
0.05
0.7
vol 1's, t+1
0.43
0.7
vol 0, year t+1
NYME2010
NYME2009
NYME2008
NYME2007
NYME2006
NYME2005
NYME2004
NYME2003
NYME2002
NYME2001
Pr(yearonyear move)
between states of annual trading volume by decade - NYMEX before and after
CME merger (announced March 2008, finalized September 2009) and before
and after switch to electronic trading29
(September 2006) - part 1: transitions
0.08
0.07
0
0
0
0.7
5
1.0
0
0.0
0
0.2
5
0
0.5
5
1.0
0
0.0
0
0.2
5
0.7
0.5
0.97
0.97
0.01
0.01
0.01
0.01
0.01
0.02
0.02
0.02
0.02
0.97
0.97
0.97
0.97
0.97
0.92
0.92
0.92
0.97
0.97
0
0.2
0.0
0.8
0.16
0.16
0.62
0.98
0.92
vol 10^8's, t
0.16
0.8
0.97
0.97
0.01
0.16
0.59
0.74
0.15
0.8
0.91
0.02
0.08
0.25
0.01
0.61
0.11
vol 10^7's, t
0.51
0.18
0.47
0.59
0.01
0.63
0.25
0.8
0.16
0.25
0.69
0.46
0.34
0.19
0.41
0.9
0.13
0.72
0.73
0.09
0.1
0.07
vol 10^6's, t
0.88
0.11
0.97 0.03
0.02
0.08
0.02
0.92
0
0.62
0.29
0.01
1.0
0.7
0.5
0.53
0.05
0.41
0.63
5
1.0
0
0.0
0
0.2
5
0.31
0.03
0.56
0.7
0.38
0.5
0.73
0.25
0.04
0.01
0.3
0.36
0.69
0.05
5
1.0
0
0.0
0
0.2
5
0.7
0.52
0.1
0.5
0.71
0.01
0.18
0.07
0.02
5
1.0
0
0.0
0
0.2
5
0.74
0.2
0.08
0.02
0.73
0.5
0.04
0.7
0.02
0.17
0.5
0.59
0.05
5
1.0
0
0.0
0
0.2
5
0.04
0.2
0.7
0.61
vol 10^5's, t
ExchangeYear
NYME2010
NYME2009
NYME2008
NYME2007
NYME2006
NYME2005
NYME2004
NYME2003
NYME2002
NYME2001
0.01
NYME2010
NYME2009
NYME2008
NYME2007
NYME2006
NYME2005
NYME2004
NYME2003
NYME2002
NYME2001
0.05
vol 10^4's, t
0.06
0.01
0.22
NYME2010
NYME2009
NYME2008
NYME2007
NYME2006
NYME2005
NYME2004
NYME2003
NYME2002
NYME2001
0.01
0.03
0.04
0.67
5
1.0
0
0.0
0
0.2
5
NYME2010
NYME2009
NYME2008
NYME2007
NYME2006
NYME2005
NYME2004
NYME2003
NYME2002
NYME2001
0.7
0.55
0.23
0.01
5
1.0
0
0.0
0
0.2
5
0.01
0.07
0.29
0.7
0.01
0.05
0.5
0.03
vol 100's, t+1 vol 1000's, t+1 vol 10^4's, t+1 vol 10^5's, t+1 vol 10^6's, t+1 vol 10^7's, t+1 vol 10^8's, t+1
0.01
5
1.0
0
0.0
0
0.2
5
vol 10's, t+1
0.5
vol 1's, t+1
0.7
0.06
0.5
vol 0, year t+1
NYME2010
NYME2009
NYME2008
NYME2007
NYME2006
NYME2005
NYME2004
NYME2003
NYME2002
NYME2001
Pr(yearonyear move)
between states of annual trading volume by decade - NYMEX before and after
CME merger (announced March 2008, finalized September 2009) and before
and after switch to electronic trading30
(September 2006) - part 2: transitions
given annual volumes 10, 000
5.2.
Recent trends in NYMEX lifecycles and the importance of

electronic trading
While they do not show a strong influence from the CME acquisition,
figures 15 and 16 may speak to the influence of electronic trading. Pronounced lifecycle trends on the NYMEX seem to begin in 2006, when the
exchange abruptly switched from open-outcry to electronic trading. These
trends mirror the more general tendency across derivatives markets over the
last decade, with more flexible trading at low volumes, more contracts moving up to modest volumes, and a small decline in the probability of trading
at high levels.
As mentioned above, it is difficult to separate out the effects of electronic
trading per se from those of the whole suite of new tools that arrived with
electronic trading, such as clear swaps platforms. NYMEX, within its specialization in energy contracts threatened by Enron, was a pioneer in cleared
swaps transactions. In 2003, it launched ClearPort, the platform now used
for all of the CMEs cleared swaps trades. ClearPort was marketed as an
electronic trading system because it disseminated information about specialized swaps trades via screens (Reuters News, 2003). However, most cleared
swaps transactions are negotiated bilaterally, over the phone. That means
that ClearPort trades are supported by electronic infrastructure, but they are
not fully electronic. So, while NYMEXs abrupt shift from pit-based to electronic trading offers a prime opportunity to isolate the influence of electronic
trading on derivatives volumes, the advent of cleared swaps complicates both
the analysis of NYMEX data and the definition of electronic trading.
To the extent that we can identify the influence of fully electronic trading
on its own, then 2006, the year that NYMEX abruptly closed pit trading,
should produce discontinuities in ongoing lifecycle trends. 2006 does indeed
show evidence of a discontinuity. That evidence is not overwhelming, but it
does support the hypothesis that the large changes in derivatives markets in
the 2000s were driven specifically by the switch to electronic trading.
31
6.
Derivatives volumes by product type
Figure 17 shows expected value estimates for trading at year ten for each
product type. Derivatives based on US treasuries, and, to a lesser extent,
derivatives based on natural gas and stock indexes enjoy higher expected
volumes than other product types.
The distinction between these strong performers and most of the other
product types in the sample is appreciable but does not hold with high probability. The 95 percent probability interval for each of those high expected
volume product types sits within the upper tails of the distributions for other
product types. The long upper tails that shadow the three top performers are
largely a function of uncertainty in estimating the parameter for relatively
uncommon product types rather than stellar historical performance. They
reflect the fact that we have relatively few observations of derivatives based
on wood products, for example, and so our model allows for the possibility
that out-of-sample wood products may show high trading volumes in the
future.
Major currencies, grains, precious metals, petroleum-related products,
and interest rates not derived from US treasuries define the middle of the pack
for expected year ten volumes. They are joined by a large group of product
types whose expected volumes are subject to great uncertainty, thanks to a
scarcity of data.
Among these average performers, plastics and chemicals may be promising niches for innovation. While their estimated expected volumes are subject
to considerable uncertainty, the data points we have indicate that they are
relatively strong performers.
On the low end of our expected year ten volume estimates are singlestock futures7 and weather derivatives. Both are relatively new product
7
This is consistent with figure 14 which shows OneChicago, the exchange specializing
in single-stock futures as a relative under-performer in expected trading volume at year
ten.
32
Yield Insurance
WOOD PRODUCTS
WEATHER
STOCK INDEX
SINGLE STOCK FUTURES
REAL ESTATE
PRECIOUS METALS
PLASTICS
PETROLEUM AND PRODUCTS
OTHER FINANCIAL INSTRUMENTS
OTHER AGRICULTURAL
OILSEED and PRODUCTS
Subgroup
NATURAL GAS AND PRODUCTS
NARROW BASED INDICES
LIVESTOCK/MEAT PRODUCTS
Interest Rates non U.S. Treasury
Interest Rates U.S. Treasury
GRAINS
Fertilizer
FOODSTUFFS/SOFTS
FIBER
EMISSIONS
ELECTRICITY AND SOURCES
DAIRY PRODUCTS
CURRENCY(NONMAJOR)
CURRENCY
CHEMICALS
33
BASE METALS
1e+05
1e+07
Expected volume after 10 years
vol 1's, t+1

Yield Insurance
WOOD PRODUCTS
STOCK INDEX
0
0
0.18
REAL ESTATE
PLASTICS

0
0.01
0.06
0.25
INT RATES U.S. TRES

INT RATES NON U.S. TRES
GRAINS
0.07
0.03
CURRENCY(NONMAJOR)
CURRENCY
CHEMICALS
BASE METALS
INT RATES U.S. TRES
FIBER
0.22
0.06
0.22
0.11
0.13
0.05
0.11
0.22
0.11
EMISSIONS
0.6
DAIRY PRODUCTS

FOODSTUFFS/SOFTS
FIBER
Fertilizer
EMISSIONS
GRAINS
Fertilizer
PLASTICS
0.03
FOODSTUFFS/SOFTS
0.44
vol 1's, t

vol 0, year t
Product subgroup
0.16
0.18
0.03
OTHER AGRICULTURAL
PRECIOUS METALS
0
0.14
0.06
OTHER AGRICULTURAL
REAL ESTATE
0.25
PRECIOUS METALS
WOOD PRODUCTS
STOCK INDEX
0.08
WEATHER
0.02
Product subgroup
WEATHER
vol 10's, t+1

Yield Insurance
DAIRY PRODUCTS
0.19
CURRENCY(NONMAJOR)
0.06
CURRENCY
0.03
0.25
0.19
0.06
0.05
CHEMICALS
BASE METALS
0.01
0
0.00 0.25 0.50 0.75 1.00
0.00 0.25 0.50 0.75 1.00
Pr(yearonyear move)
Pr(yearonyear move)
to annual volume tens of thousands (right) by product type
34
vol 100's, t+1
WOOD PRODUCTS
0.06
0
0.36
REAL ESTATE
0.05
REAL ESTATE
0.03
0.21
PLASTICS
0.59
0.25
0.13
0.17
Fertilizer
0.04
0.32
0.08
0.04
0.13
0
FIBER
0.17
0
FIBER
0
0.18
CHEMICALS
BASE METALS
0.28
0.12
0.17
0.07
0.08
0.05
0.03
0.04
0.2
0.2
0.2
DAIRY PRODUCTS
0.25
CURRENCY(NONMAJOR)
0.1
EMISSIONS
0.2
CURRENCY(NONMAJOR)
0.32
CURRENCY
INT RATES U.S. TRES
FOODSTUFFS/SOFTS
0.18
DAIRY PRODUCTS
0.18
Fertilizer
0.16
GRAINS
EMISSIONS
0.23
FOODSTUFFS/SOFTS
GRAINS
0.22
OTHER AGRICULTURAL
INT RATES U.S. TRES

CURRENCY
0.17
0.34
0.19
CHEMICALS
BASE METALS
0.04
vol 100's, t
0.16
vol 10's, t
Product subgroup
PLASTICS
OTHER AGRICULTURAL
0.1
PRECIOUS METALS
0.19
0.17
0.26
STOCK INDEX
PRECIOUS METALS
WEATHER
WOOD PRODUCTS
WEATHER
STOCK INDEX
vol 1000's, t+1

Yield Insurance
Product subgroup
Yield Insurance
0.76
0.1
0.00 0.25 0.50 0.75 1.00
0.00 0.25 0.50 0.75 1.00
Pr(yearonyear move)
Pr(yearonyear move)
to annual volume tens of thousands (right) by product type
35
types with many correlated contracts launched in recent years. Interestingly,

these contract types appear to under-perform relative to some product types
like yield insurance and emissions in which trading was effectively smothered
by external events (the proliferation of subsidized crop insurance in the US in
the case of yield insurance and the failure of the US to consistently promote
cap-and-trade legislation in the case of emissions).
Interestingly, single-stock futures and weather derivatives were more likely
than most other product types to climb up from low levels of trading volume.
Figures 18 and 19 present the probability of any contract moving to higher
levels of annual trading volume for each product type in the sample. Singlestock futures are particularly likely to recover from years of zero trading
volume and weather derivatives are particularly likely to move up to annual
trading volumes in the thousands. (See ??s figures 25 and 26 for additional
details.) Insofar as these product types move fluidly up and down from
low levels of annual trading volumes they are representative of recent trends
across derivatives markets.
7.
Conclusions
In this article we have presented a comprehensive analysis of trading

volumes for derivatives reported to exchanges in the United States. Looking
across decades, exchanges, and product types we see multiple trends that
challenge or significantly modify findings of existing studies.
While a larger percentage of contracts had little or no volume in any given
year of the 2000s, contracts did not fail at the high rates noted in previous
analyses. Instead, they remained at low levels of trading until they were
needed, transitioning back into active trading with greater probability than
in previous decades. Interestingly, this flexibility from low levels of trading
meant that the long term outlook for a new contract did not erode despite
remarkable levels of new contract innovation. During the 2000s the expected
36
volume of a new contract after ten years was above that of the 1990s and
within the range of previous decades. On balance, the explosion of innovation
catalyzed by electronic trading did not hurt the prospects for the marginal
contract.
We find that expected year ten trading volumes varied more decade to
decade than from exchange to exchange or product type to product type. In
particular, the lifecycle of a derivative on any given exchange was largely indistinguishable from that on any other, with the likely exception of OneChicago,
which specializes in single-stock futures.
We find evidence that the decadal changes in derivative lifecycles were
driven by the switch to electronic trading rather than the consolidation of
exchanges by looking at trends on the New York Mercantile Exchange. The
effects of electronic trading are difficult to separate from the the related innovation of cleared swaps. However, trends in NYMEX volumes following
the 2006 launch of widespread electronic trading tentatively support the hypothesis that electronic trading is indeed driving recent trends in derivatives
lifecycles across all sampled markets.
7.1.
The statistical characteristics of derivative volumes
In addition to facilitating quick distributional comparisons across various

contract groupings (decade, exchange, and product type), our framework
(Markov models) allows us to explore some basic questions about derivative
markets in general. For example, based on our Markov models it appears
that trading volumes do not follow a normal or log-normal random walk over
time. In figures 4 and 5 it is clear that the probability of remaining at a given
level of annual volume varies dramatically from one level to the next. These
differences hold with greater than 95 percent probability as do variations
in the volume dynamics across time (indicating that normal or log-normal
models of trading volume would suffer from stationarity problems as well).
Furthermore, switches to higher and lower levels of trading are often not
37
symmetric. In particular, an outright crash to zero trading volume appears

more likely than would be predicted by a symmetrically distributed random
walk.
However, our analysis does affirm the common observation that it is unusual for a contract to experience initial popularity and to crash subsequently
(Johnston and McConnell, 1989). After reaching a trading volume in the tens
of thousands, the probability that a contract will have annual trading volume
of zero in the subsequent year drops appreciably.
7.2.
Optimal contract innovation
Much of the literature on derivative innovation focuses on the problem of

choosing the optimal derivatives contract to launch next. This analysis does
not directly address that question, but it does present some trends relevant
to previous theoretical work which could inform further investigation.
One interpretation of Duffie and Jackson (1989) provides that revenue
maximizing marginal innovations are uncorrelated with existing contracts.
However, recent trends suggest that one of the key assumptions underlying
this finding only holds weakly. Historically, correlated contract innovations
have not shown diminishing marginal volumes. In general, innovation in
derivatives markets has exploded in the last decade seemingly without dragging down expected trading volumes at year ten. Indeed, some of the highest
volume product types (in expected volume terms) are highly correlated both
to other derivatives of their product category but also to the average returns
of the economy as a whole (US treasuries and stock indexes).
Tashjian and Weissman (1995) explains the proliferation of correlated
(and often redundant) contracts as a form of price discrimination. They assume, that an exchange can charge higher fees on the transaction for parties
with larger and more concentrated exposure to a given underlying index.
This framework for understanding product innovation holds up well in light
of recent trends. As we have discussed, many recently launched contracts are
38
cleared swaps, which tend to be more specialized than conventional futures

or options. As Tashjian and Weissman (1995) predicted, exchanges charge a
substantial premium on these specialized transactions. Fees on CMEs ClearPort platform were more than 350 percent those on conventional electronic
futures trades as of late 2012 (CME Group, 2012). However, Tashjian and
Weissman (1995) suggested that exchanges would charge more for singleasset-based derivatives (such as gold) than for derivatives that represented
the holding of multi-product firms (like the crack-spreads used to reproduce
petroleum refiners returns). In practice, cleared-swaps contracts, with their
relatively high fees, appear to be biased toward the latter.
A third explanation of recent patterns in derivative innovation is psychological. Based on Tversky, Kahneman, et al. (1981), Shiller (1994) suggests
that a hedge appears more attractive when its presented as the elimination
of risk rather than when it is described as the reduction of risk. This tendency to overvalue hedges tailored to the needs of specific firms may explain
the proliferation of correlated contracts (and their relative success) above and
beyond the price discrimination suggested in Tashjian and Weissman (1995).
7.3.
What kind of economic goods are derivatives?
The present analysis could also suggest new ways of understanding the
economic value of derivatives. If indeed derivatives are simply contingent
contracts that move cash flows across time and states of nature, then they
should derive all their value from the way that they mesh with hedgers risk
preferences. It follows from that idea, that if risk preferences remain stable
over time, then derivative trading patterns should also remain stable.
But trading patterns have not been stable over the last decade. Instead,
they bear a striking qualitative resemblance to those of information goods,
particularly media:
Each class of economic goods was, until recently, simple to classify:
normal goods with elastic demand and network effects.
39
Starting a new derivatives market, just like producing a new music album or launching a magazine, was a high risk, high reward proposition.
In the last decade both saw paradigm shifts in their marginal cost
structure (i.e. there were fundamental changes in supply).
At the same time, new technologies allowed consumers ubiquitous access to goods (i.e. there were also fundamental changes in demand).
After those twin revolutions, markets:
- Rewarded specialty products more than in the past;
- Hosted blockbusters as large as/larger than ever;
- Did not offer the same opportunities for strong but less-thanblockbuster products.
In media (and informational goods more generally) this transition has
upended many long-profitable business models and catalyzed a great deal of
innovation. In derivatives it has certainly opened up the door to many new
entrants like ICE, which now is one of two large futures exchanges in the US
today. But it is not clear whether those new entrants are using fundamentally
new business models.
How strong is the parallel? Should economists study derivatives alongside informational goods? What does the possible connection suggest for
the future of derivatives? We believe that these questions provide a solid
foundation for future research.
References
Bank of International Settlements, 2012. Statistical release: OTC derivatives
statistics at end-june 2012.
Carlton, D., 1984. Futures markets: Their purpose, their history, their
growth, their successes and failures. Journal of Futures Markets 4, 237271.
CME Group, 2012. Press release: CME Group volume averaged 11.9 million
40
contracts per day in september 2012, up 16 percent from august 2012.

Thanks to Silla Brush of Bloomberg for directing us to this press release.
Cuny, C., 1993. The role of liquidity in futures market innovations. Review
of Financial Studies 6, 5778.
Duffie, D., 2012. Dark Markets: Asset Pricing and Information Transmission
in Over-the-Counter Markets. Princeton University Press.
Duffie, D., Jackson, M., 1989. Optimal innovation of futures contracts. Review of Financial Studies 2, 275296.
Gorham, M., Kundu, P., 2012. A half-century of product innovation and
competition at U.S. futures exchanges. The Review of Futures Markets 20,
105140.
Gorham, M., Singh, N., 2009. Electronic Exchanges: The Global Transformation From Pits to Bits. Elsevier Science.
Holland, A., Fremault, A., 1997. Features of a successful contract: Financial
futures on LIFFE. Bank of England Quarterly Bulletin 37, 181186.
ISDA Research Staff, NERA Economic Consulting, 2011. Costs and benefits
of mandatory electronic execution requirements for interest rate products.
Discussion Paper 2, International Swaps and Derivatives Association.
Johnston, E., McConnell, J., 1989. Requiem for a market: An analysis of the
rise and fall of a financial futures contract. Review of Financial Studies 2,
123.
Mello, A., Reilly, J., 2012. Margins, liquidity and the cost of hedging. Tech.
rep., MITs Center for Energy and Environmental Policy Research.
Olson, E. S., 2010. Zero-Sum Game: The Rise of the Worlds Largest Derivatives Exchange. Wiley.
41
Page, S., 2012. Model thinking: Markov models. Video lectures.

Plummer, M., 2003. JAGS: A Program For Analysis of Bayesian Graphical
Models Using Gibbs Sampling.
Plummer, M., 2013. rjags: Bayesian graphical models using MCMC. R package version 3-10.
Reuters News, 2003. NYMEX launches Cleaport electronic trading system.
Shiller, R., 1994. Macro Markets: Creating Institutions for Managing Societys Largest Economic Risks: Creating Institutions for Managing Societys
Largest Economic Risks. Clarendon Press.
Silber, W., 1981. Innovation, competition, and new contract design in futures
markets. Journal of Futures Markets 1, 123155.
Tashjian, E., Weissman, M., 1995. Advantages to competing with yourself:
Why an exchange might design futures contracts with correlated payoffs.
Journal of Financial Intermediation 4, 133157.
Tversky, A., Kahneman, D., et al., 1981. The framing of decisions and the
psychology of choice. Science 211, 453458.
42
volume level 0
1
10 100 1000 104 105 106 107 108
0
0.78 0.14 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
0.16 0.63 0.14 0.01 0.01 0.01 0.01 0.01 0.01 0.01
1
0.01 0.16 0.63 0.14 0.01 0.01 0.01 0.01 0.01 0.01
10
0.01 0.01 0.16 0.63 0.14 0.01 0.01 0.01 0.01 0.01
100
0.01 0.01 0.01 0.16 0.63 0.14 0.01 0.01 0.01 0.01
1000
0.01 0.01 0.01 0.01 0.16 0.63 0.14 0.01 0.01 0.01
104
0.01 0.01 0.01 0.01 0.01 0.16 0.63 0.14 0.01 0.01
105
0.01 0.01 0.01 0.01 0.01 0.01 0.16 0.63 0.14 0.01
106
0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.16 0.63 0.14
107
108
0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.16 0.76
Table 2: Priors on transition between volume states - with annual trading
volume state in year t denoted by row, trading volume state in year t denoted
by column
43
0.6
0.21
0.77
0.14
0.38
0.28
0.31
0.08
0.42
0.09
0.13
0.08
0.74
0.07
0.61
0.25
0.07
0.34
0.25
0.7
0.16
0.32
0.15
0.09
0.14
0.14
0.1
0.77
CBT
0.13
0.02
0.52
0.18
0.18
0.09
0.03
0
0.0
1.0
0.7
0
0.0
1.0
0
0.5
0.02
0
0.2
1.0
0
0.7
0.5
0.2
0.0
0.5
0.25
0.61
0.01
0.5
0.73
0.01
0.2
0.02
0.08
0.06
0.01
0.15
0.64
0.59
0.25
0.77
0.01
0.74
0.02
0.01
0
0.2
0.01
0.74
0.74
0.02
0.02
0.01
0.02
0.11
0.74
0.9
0.07
0.61
0.11
0.02
0.02
0.05
0.6
0.24
0.07
0.06
CCX
0.05
0.05
CME
0.42
0.45
0.01
0
0.21
0.74
0.01
0
0.56
0.02
0.0
0.13
0.74
0.1
0.01
0
0.5
0.01
0.04
0.02
0.55
0.25
0.01
0.01
0
0
0.13
0.74
0.1
0.09
0.07
0.01
0.72
0.24
0.02
CMX
ACE
1.0
0.45
0.25
0.16
0.73
0.7
0.12
CRCE
BCC
0.02
0.43
0.5
0.01
1.0
0.2
0.74
0.01
0.01
0.73
0.02
0.74
0.02
0.02
0.01
0.13
0.0
0.02
0.05
0.11
0.01
0.01
0.05
0.16
EUR
ELX
1.0
0.06
0.7
EUS
0.01
vol 1's, t
GE
0.18
ICEX
ICE
0.06
KCBT
ICEU
0.15
0.59
0.94
IEPA
0.15
0.76
0.03
0.7
0.16
0.12
0.5
0.06
0
0.51
0.6
0.57
MCE
ICUS
MESL
0.1
0.01
0.35
MGE
0.94
Memphis
0.2
ACE
MilGX
0.01
0.94
NYME
0.16
0
0
0.88
BCC
0.94
0.5
NOCE
0.16
0.0
CBT
NYBOT
0.16
0.35
CCX
NYL
0
0.01
0.03
0.01
0.05
0.01
0.12
0.94
CME
NYL2
CRCE
NYPE
0.57
PANY
0.19
1.0
0.94
EUS
PBOT
0.06
0.05
0.79
GE
PCE
0
0.06
ICE
0.01
0.7
ICEU
CMX
0.13
ELX
0.28
0.96
EUR
0.13
ICEX
0.13
0.5
0.85
0.52
ICUS
0.13
vol 0, year t
IEPA
0.6
0.04
0.41
0.2
0.15
0.0
0.16
0.35
KCBT
0.6
0.65
MCE
Seattle
0.01
0.7
MESL
0.03
1.0
0.94
MGE
West Coast
0.7
0.01
0.97
Memphis
0.5
MilGX
0.01
0.94
0.2
0.94
0.01
0.0
0.94
NOCE
0.01
1.0
NYBOT
0.01
0.01
0.94
0.05
0.94
NYL
0.06
0.7
0.15
0.12
0.5
NYL2
0.04
0.2
0.56
0.76
0.0
NYME
0.01
0.21
0.94
NYPE
0.0
0.01
0.07
PANY
0.19
vol 10^8's, t+1
1.0
0.04
0.04
0.7
0.67
vol 10^7's, t+1
vol 10^6's, t+1
vol 10^5's, t+1
vol 10^4's, t+1
0.01
0.95
vol 1000's, t+1
0.2
vol 100's, t+1
0.01
0.94
PBOT
vol 10's, t+1
PCE
Exchange
vol 1's, t+1
0.94
Seattle
0.5
vol 0, year t+1

West Coast
Pr(yearonyear move)
between states of annual trading volume by exchange - part 1: transitions
given annual volumes 0 and < 10
44
0.14
0.22
0.17
0.02
0.02
0.74
0.13
0.19
0.04
0.19
0.28
0.23
0.01
0.35
0.15
0.11
0.4
0.3
0.3
0.19
0.07
0.19
0.52
0.55
0.19
0.53
0.17
0.07
0.13
0.05
0.39
0.01
0.05
0.05
0.25
0.09
0.37
0.22
0.14
ICEX
0.28
0.03
0.41
0.39
0.13
0.36
0.11
0.16
0.02
0.08
0.29
0.16
0.09
0.09
0.25
0.31
0.02
0.29
0.1
0.3
0.2
0.5
0.01
0.03
0.21
0.42
0.17
0.09
0.21
0.27
0.18
0.14
0.14
0.06
0.21
0.38
0.16
0.01
0.2
0
0.5
0.2
0.0
1.0
0.7
0.5
0
0.01
0.01
0.01
0.74
0.2
0
0.0
1.0
0.7
0.5
0.2
0.02
0.04
0.0
0.02
0.74
1.0
0
0.0
0.02
0.2
0.7
0.0
0.05
0.01
0.29
0.02
0.1
0.05
0.12
0.35
0.25
0.17
0.33
0.14
0.25
CBT
0.05
0.05
0.01
0.74
0.28
0.04
0.29
0.01
0.74
0.64
0.6
0.02
0.75
0.25
0.01
0.41
EUS
0.03
0.04
0.03
0.08
0.06
0.34
CCX
0.27
CME
0.06
vol 100's, t
0.08
0.14
CMX
0.15
0.12
0.51
EUR
0.15
0.3
GE
0.36
KCBT
0.81
MCE
0.19
0.11
0
0.3
0.08
0.16
0.15
0.01
0.03
0.74
0.21
0.02
0.07
0.01
0.31
0.74
0.13
0.19
0.02
0.14
0.05
0.11
0.13
0.25
0.13
0.61
0.05
0
0.58
0.01
0.42
0.01
0.06
0.01
0.74
MGE
0.01
0.4
0
0
0.08
0.19
0.15
1.0
0.7
0.03
0.5
0.2
0.22
0.03
0.01
0.2
0.48
0.05
0.25
0.0
0.21
0.08
0.56
Memphis
0.36
1.0
0.18
0.13
0.7
0.08
0.5
0.39
ACE
0.01
0.16
0.2
NOCE
BCC
0.74
0.0
0.07
0.02
NYBOT
0.01
0.11
NYL
CRCE
0.74
0.15
NYPE
0.02
PBOT
0.62
ELX
1.0
0.36
PCE
0.05
0.22
0.29
West Coast
ICE
ICEU
0.7
0.01
0.28
0.5
0.1
0.26
CBT
Exchange
0.01
0.74
0.56
0.74
0.02
0.2
0.02
CCX
ICUS
0.02
0.15
0.04
0.0
0.36
IEPA
0.11
0.19
0.16
CME
MESL
0.24
0.14
CMX
MilGX
0.26
0.17
NYL2
0.04
0.7
NYME
0.21
0.3
EUS
PANY
0.13
0.08
0.03
vol 10's, t
EUR
0.16
0.09
0.45
GE
Seattle
0.16
0.19
0.08
0.09
0.05
0.02
ICEX
ACE
ICE
0.19
0.04
0.41
0.58
ICUS
0.56
0.62
IEPA
MCE
BCC
0.13
0.23
1.0
MESL
0.43
0.36
0
0.09
0.58
0
0.04
KCBT
0.35
0.02
0.7
0.3
MGE
CRCE
0.15
MilGX
0.5
0.81
0.3
0.06
0.2
0.23
0.13
0.01
0.0
0.01
0.54
ELX
0.74
0.73
NOCE
0.02
0.02
0.06
0.02
NYBOT
ICEU
0.23
0.0
1.0
0.7
NYL
0.25
0.5
1.0
0.04
0.01
0.28
0.39
NYL2
0.74
0.22
1.0
NYME
0.02
0.44
0.7
0.7
0.16
0.5
0.3
0.2
0.08
NYPE
Memphis
0.45
vol 10^8's, t+1
vol 10^7's, t+1
0.36
vol 10^6's, t+1
0.25
vol 10^5's, t+1
vol 10^4's, t+1
0.58
vol 1000's, t+1
0.11
0.01
vol 100's, t+1
0.5
PBOT
vol 10's, t+1
PCE
PANY
vol 1's, t+1
0.79
Seattle
0.5
vol 0, year t+1

West Coast
Pr(yearonyear move)
45
0.05
0.01
0.03
0.13
0.42
0.32
0.02
0.07
0.34
0.47
0.06
0
5
0.02
0.56
0
0.83
0.36
5
0.14
0.1
0.64
0.17
0.12
0.0
0.62
0.03
0
0.0
1.0
0.7
0.5
0.16
0.15
0.01
0.04
0
0.2
0.73
0.68
0.08
0.02
0.01
0.11
0.45
0.03
0.12
0.38
0.95
0.23
0.18
0.5
0
0
0.04
0.16
0.19
0.53
0.01
0.34
0.2
0.05
0.76
0.16
0.01
0.25
0.15
0.74
0.02
0.0
0.37
0.61
0
1.0
0.08
0.75
0
0.44
0.1
0.01
1.0
0.2
0.07
0.66
0.8
0.11
0.11
0.05
0.7
0.02
0.01
0.05
0.7
0.76
0.12
0.74
0.5
0.05
0.69
0.5
0.17
0.01
0.2
0.01
0.02
0.0
0.02
0.68
0.01
0.01
0.26
0.74
1.0
0.04
0.02
0.28
0.04
0.29
0.82
0.76
0.19
0.11
0.18
0.2
0.53
0.7
0.25
0.16
0.92
0.11
0.63
0.01
0.18
0.0
0.29
0.5
0.07
0.83
1.0
0.07
0.02
0.0
0.01
0.76
0.19
0.2
0.04
0.4
0.46
0.0
0.08
0.7
0.53
0.2
1.0
0.01
0.11
0.73
0.2
0.45
0.02
0.5
0.16
0.23
ACE
0.3
0.02
0.0
0.03
0.01
0.2
0.09
1.0
BCC
0.16
0.74
1.0
0.02
0.7
0.7
CBT
0.2
vol 10^4's, t
CCX
0.34
EUR
CME
0.15
EUS
CMX
CRCE
0.08
0.13
ICEU
0.38
ELX
0.2
0.44
ICEX
GE
0.01
ICE
0.01
0.27
ICUS
0.1
0.03
0.09
0.56
0.05
0.03
0.26
NOCE
IEPA
0.4
0.17
0.07
KCBT
0.13
0.02
NYBOT
MCE
0.01
0.06
MESL
0.7
0.05
0.19
0.73
0.02
0.13
0.01
0.51
0.02
0.5
0.74
0.06
0.01
0.3
0.15
0.02
0.01
0.09
0.29
0.25
0.01
0.05
0.27
0.29
0.2
0.09
PANY
MGE
0.45
0.12
0.73
0.06
0.56
0.11
Memphis
0
0.02
0.01
0.53
0.1
0.01
0.27
0.26
0.56
0.41
0.12
0.11
0.09
0.05
PBOT
MilGX
0.0
PCE
NYL
0.09
0.39
NYL2
0.53
0.3
West Coast
NYME
0.13
ACE
NYPE
0.01
0.08
0.13
0.62
0.53
0.25
0.1
0.04
0.01
0.06
0.03
CBT
Seattle
0.08
0.5
0.01
0.72
0.29
0.12
vol 1000's, t
0.17
0.06
0.56
0.01
0.09
0.26
0.24
0.51
0.25
0.61
0.07
0.15
0.02
0.04
1.0
0.13
Exchange
EUR
BCC
EUS
CCX
0.44
CMX
0.05
0.01
0.25
0.05
0.15
0.02
0.74
0.36
0.11
0.02
0.09
0.01
0.07
0.21
0.49
0.15
0.35
0.43
0.3
0.15
0.39
0.7
ICE
CME
0.42
CRCE
0.01
ICEU
ELX
0.05
ICEX
GE
0.06
0.11
IEPA
ICUS
0.3
0.55
0.25
MGE
KCBT
MilGX
MCE
0.05
0.01
0.5
MESL
0.11
0.74
0.3
NYBOT
Memphis
0.11
0.06
NYL
NOCE
0.18
0.5
NYL2
0.52
0.2
NYME
NYPE
0
0.1
vol 10^8's, t+1
vol 10^7's, t+1
0.2
0.08
0.02
vol 10^6's, t+1
0.19
PBOT
vol 10^5's, t+1
0.46
vol 10^4's, t+1
0.0
vol 1000's, t+1
0.18
1.0
vol 100's, t+1
PCE
vol 10's, t+1
0.7
PANY
vol 1's, t+1
0.08
Seattle
0.5
vol 0, year t+1

West Coast
Pr(yearonyear move)
46
0.01
0
1.0
0
0
0.01
0.7
0.74
0.5
0.02
0.73
0.2
0.01
0.02
0.0
0.74
0
0.0
1.0
0.7
0.5
0.2
1.0
0.08
0.02
0.01
0.8
0.74
0.08
0.02
0.04
0.02
0.89
0.05
0.9
0.7
0.04
0.01
0.01
0.01
0.01
0.73
0.5
0.04
0.02
0.19
0.76
0.2
0.01
0.0
0.07
0.62
0.01
0.0
0.51
0.29
1.0
0.01
0.15
0.01
0.7
0.14
0.15
0.05
0.73
0.77
0.02
0.06
0.6
0.82
ACE
0.01
0.5
BCC
0.25
0.01
0.16
0.73
0.01
0.01
0.92
CMX
CBT
0.06
CCX
0.93
0.02
0.87
CRCE
0.83
0.11
0.01
0.04
0.88
ELX
CME
0.01
0.73
0.73
0.15
0.01
0.01
EUR
0.01
0.74
0.02
0.01
0.73
0.2
0.02
0.01
0.07
0.02
0.02
0.07
0.74
ICEU
EUS
0.02
0.74
0.03
ICEX
GE
0.02
0.0
0.02
0.01
0.59
ICE
0.39
0.56
0.01
0.72
0.36
IEPA
0.19
ICUS
KCBT
0.01
vol 10^6's, t
MCE
MESL
0.01
0.73
0.74
1.0
MGE
0.02
0.01
0.74
0.01
0.73
0.01
0.02
0.74
NOCE
Memphis
0.73
0.7
0.01
NYBOT
MilGX
0.02
0.01
0.02
NYL2
0.02
NYME
NYL
0.02
0.5
0.73
0.02
PANY
NYPE
PBOT
0.09
0.52
0.2
PCE
0.5
0.81
0.43
0.06
0.0
Seattle
0.05
1.0
0.02
0.43
West Coast
0.17
0.1
0.34
0.8
0.7
0.13
0.01
0.1
0.07
0.7
0.42
0.01
0.01
0.02
0.01
0.35
0.5
ACE
0.02
0.59
0.56
0.25
0.2
0.01
0.47
0.0
0.74
0.41
0.01
0.02
1.0
0.7
0.01
BCC
0.5
CBT
0.2
CCX
CME
0.0
CMX
0.21
0.44
0.91
0.73
0.11
0
0.54
0.01
0.01
0.25
0.11
0.07
0.1
1.0
0.25
0
0.12
0.73
0.06
0.02
0.7
0.76
0.76
0.09
0.14
0.12
0.03
0.75
0.12
0.02
0.01
0.5
0.74
0.01
0.06
0.82
0.9
ELX
CRCE
0.01
0.09
0.01
0.02
0.02
ICE
EUR
0.55
0.74
0.25
0.73
0.03
EUS
0.03
0.01
0.01
0.02
0.02
0.36
0.01
0.0
ICEU
GE
0
0.01
0.8
0.16
1.0
ICEX
0.95
vol 10^5's, t
0.42
MCE
ICUS
0.38
0.5
0.16
IEPA
0.05
0.01
MESL
KCBT
MGE
0.7
0.7
0.22
0.25
0.74
0.61
0.01
0.01
0.02
0.01
NOCE
Memphis
0.5
0.74
0.02
0.01
0.74
vol 10^8's, t+1
0.73
0.2
NYL
vol 10^7's, t+1
0.02
0.01
MilGX
NYL2
NYBOT
Exchange
0.02
vol 10^6's, t+1
0.0
NYPE
0.25
vol 10^5's, t+1
0.6
vol 10^4's, t+1
vol 1000's, t+1
PANY
NYME
1.0
PBOT
vol 100's, t+1
vol 10's, t+1
0.7
PCE
vol 1's, t+1
0.5
0.2
Seattle
0.2
vol 0, year t+1

West Coast
Pr(yearonyear move)
given annual volumes 10, 000 and < 1, 000, 000
47
ELX
CRCE
CMX
CCX
CBT
0
0
0
0.0
1.0
0.7
0.5
0
0.2
0.92
0.02
0.02
0.02
0.92
0.92
0.92
0.11
0.89
0.02
0.92
0.21
0.79
0.02
0.02
0.02
0.92
0.92
0.92
0.02
0.02
0.02
0.92
0.02
0.0
0.98
CME
0.92
0.92
0.02
0.02
0.98
0.92
0.92
0
0.02
0.02
0.92
0.92
0.92
0
EUR
0.92
0.02
EUS
0.02
0.92
1.0
0.92
0.02
0.92
0.92
0.02
0.02
ICE
0.02
0.7
ICEU
0.92
0.5
ICEX
0.92
0.92
0.02
0.2
0.02
0.02
0.92
0.0
0.99
0.02
IEPA
ICUS
1.0
0.92
0.92
0.02
KCBT
0.02
0.7
0.92
0.02
0.92
0.92
0.02
0.92
0.01
vol 10^8's, t
0.02
0.02
0.01
0.74
0.01
0.74
0.03
0.02
0.9
0.02
MCE
MESL
ACE
0.01
0.73
0.05
0.5
0.02
0.91
0.74
0.05
0.01
0.8
0.2
0.18
0.02
0.03
0.01
0.73
MilGX
Memphis
0.02
0.0
BCC
0.91
0.74
1.0
0.55
0.01
0.01
0.02
NOCE
0.01
0.73
0.7
NYBOT
0.6
0.74
0.5
NYL
0.02
NYL2
0.25
0.2
0.96
0.02
0.01
0.01
0.73
0.01
0.99
0.0
0.73
PANY
NYPE
NYME
GE
0.02
1.0
0.01
0.02
0.01
0.74
0.7
PBOT
0.01
0.74
0.02
0.5
PCE
MGE
0.01
0.74
0.02
0.2
0.02
Seattle
0.0
1.0
ACE
West Coast
0.74
0.01
0.01
0.74
0.7
0.01
0.5
CBT
BCC
0.01
0.73
0.01
0.02
0.2
CCX
0.0
CME
0.01
0.74
0.73
0.02
0.01
0.3
1.0
CMX
CRCE
0.7
0.5
0
0.01
0.02
0.01
0.74
0.74
0.25
0.61
ELX
0.02
EUR
0.2
0.02
0.0
GE
ICE
0.02
1.0
ICEU
0.01
0.94
0.02
1.0
0.01
0.74
0.7
ICUS
0.04
0.5
IEPA
0.2
0.01
0.74
0.02
vol 10^7's, t
MCE
EUS
0.7
0.02
0.01
MESL
ICEX
0.5
0.74
0.01
0.74
0.01
0.74
0.02
N
MilGX
0.02
NOCE
KCBT
0.02
vol 10^8's, t+1
0.74
vol 10^7's, t+1
0.02
vol 10^6's, t+1
MGE
NYL
vol 10^5's, t+1
0.2
NYBOT
Memphis
0.0
vol 10^4's, t+1
NYPE
vol 1000's, t+1
NYL2
PANY
NYME
Exchange
1.0
vol 100's, t+1
vol 10's, t+1
0.7
PCE
PBOT
vol 1's, t+1
0.5
0.2
Seattle
0.0
vol 0, year t+1

West Coast
Pr(yearonyear move)
given annual volumes 1, 000, 000
48
0.57
0.16
0.74
0.06
0.06
0.02
0.03
0.08
0.08
0.8
CHEMICALS
0.98
BASE METALS
0.73
0.25
0.65
0.3
0.65
0.08
0.17
0.14
0.13
0.14
0.05
0.18
0.18
0.09
0.03
0.4
0.44
0.51
0.17
0.01
0.58
0.51
0.06
0.02
0.22
0.28
0.22
0.08
0.08
0.11
0.6
0.21
0.13
0.11
0.25
0.6
0.08
0.19
0.75
0.09
0.06
0.13
0.09
0.02
0.02
0
5
0.1
0.2
1.0
0.7
0.5
0.2
0.15
0.06
0.0
0.01
1.0
0.7
0.05
0.09
0.74
0.5
0.02
0.05
0.91
0.54
BASE METALS
0.75
CURRENCY(NONMAJOR)
0.65
DAIRY PRODUCTS
0.01
0.11
0.1
0.05
0.84
0.02
0.79
FIBER
0.02
0.0
0.25
0.06
0.0
0.61
FOODSTUFFS/SOFTS
0.22
0.2
0.07
0.52
CURRENCY
1.0
0.03
0.1
Fertilizer
0.11
0.85
GRAINS
0.18
0.52
0.05
0.7
0.08
0.0
INT RATES U.S. TRES
0.85
1.0
0.5
0.25
0.2
0.61
vol 1's, t
0.7
0.31
0.2
0.52
0.06
0.22
0.03
0.02
0.5
0.89
0.05
0.1
0.2
0.25
0.02
0.52
0.16
0.09
0.14
PRECIOUS METALS
0.05
0.03
0.0
0.6
0.44
STOCK INDEX
0.02
0.04
0.04
0.04
0.13
0.11
0.11
0.19
0.06
0.5
CURRENCY
WOOD PRODUCTS
0.14
CURRENCY(NONMAJOR)
Yield Insurance
0.04
0.22
0.53
0.52
0
0
0.35
0.05
0.03
1.0
0.07
0.7
DAIRY PRODUCTS
0.91
EMISSIONS
0.09
0.9
FOODSTUFFS/SOFTS
0.07
0.5
0.07
Fertilizer
0.12
0.2
0.02
0.03
0.21
0.0
0.71
0.03
1.0
0.81
GRAINS
0.06
0.21
0.06
0.7
0.44
WEATHER
0.06
0.07
0.03
0.5
0.69
0.2
FIBER
0.03
0.0
INT RATES U.S. TRES
0.07
0.99
0.11
1.0
0.03
0.7
0.06
0.5
0.06
0.16
vol 0, year t
0.66
0.05
0.2
0.15
0.15
0.0
0.25
CHEMICALS
0.2
0.69
EMISSIONS
0.39
Commodity subgroup
0.03
0.06
0.06
1.0
0.17
0.15
0.7
0.63
0.03
0.03
0.06
OTHER AGRICULTURAL
PLASTICS
0.52
REAL ESTATE
0.01
0.07
0.5
0.94
0.83
0.02
0.21
PLASTICS
0.05
0.25
OTHER AGRICULTURAL
0.2
0.44
PRECIOUS METALS
0.16
0.18
0.04
0.0
0.58
0.08
1.0
0.06
0.02
0.08
0.25
0.7
0.72
0.17

REAL ESTATE
0.02
0.5
0.53
WEATHER
STOCK INDEX
0.2
0.98
vol 1000's, t+1 vol 10^4's, t+1 vol 10^5's, t+1 vol 10^6's, t+1 vol 10^7's, t+1 vol 10^8's, t+1
0.0
vol 100's, t+1
WOOD PRODUCTS
vol 10's, t+1
1.0
vol 1's, t+1
0.7
0.99
0.5
vol 0, year t+1

Yield Insurance
Pr(yearonyear move)
between states of annual trading volume by product type - part 1: transitions
49
vol 100's, t+1
0.19
0.2
0.51
0.06
0.07
0.22
0.18
0.11
0.18
0.03
0.47
0.18
0.2
0.22
0.32
0.08
0.2
0.25
0.4
CHEMICALS
0.06
0.32
0.17
0.17
0.07
STOCK INDEX
0.45
0.52
0.1
0.34
0.26
0.04
0.1
0.41
0.19
0.13
0.31
0.1
0.06
0.21
0.38
0.16
0.06
0.34
0.58
0.03
0.16
0.25
0.48
OTHER AGRICULTURAL
0.21
0.41
0.18
0.26
0.29
0.1
0.04
0.28
0.22
0.33
0.12
0.28
0.17
0.06
0.1
0.45
0.07
0.15
0.11
0.06
0.27
0.13
0.27
0.07
0.08
0.46
0.53
0.25
0.28
0.43
0.08
0.24
0.6
DAIRY PRODUCTS
0.03
0.1
0.36
0.2
0.09
0.29
0.2
0.09
0.13
0.06
0.45
0.2
0.03
CURRENCY
0.01
0.04
0.26
0.04
0.12
0.55
0.34
0.35
0.19
0
5
0.2
0.0
1.0
0.7
0.01
0.02
0.76
0.5
0.2
0.0
1.0
0.7
0.5
0.15
0.2
0.0
1.0
0.7
0.5
0.2
0.0
1.0
0.7
0.1
0.49
0.21
0
0.1
0.5
0.07
0.2
0.0
0.02
0.02
0.06
0.01
0.03
0.04
0.39
0
0.26
0.02
0.05
0.05
0.05
0.15
EMISSIONS
0.13
0.03
0.17
0.03
0.01
0.03
0
0.15
0.31
0.02
0.06
0.32
FOODSTUFFS/SOFTS
0.16
INT RATES U.S. TRES
Fertilizer
0.24

GRAINS
0.03
0.31
0.23
0.45
0.17
0.01
vol 100's, t
0.31
0.01
0.03
0.04
0.26
0.13
0.18
0.02
0.11
0.02
0.8
0.17
0.16
0.01
0.22
0.01
0.04
0.06
0.26
0.15
PRECIOUS METALS
0.03
0.34
0.01
0.02
0.17
0.09
0
0.04
0.3
0.33
0
0.2
0.4
WEATHER
0.03
0.22
0.04
0.34
0.06
0.5
BASE METALS
0.16
0.06
0.03
0.22
0.03
CURRENCY
0.24
0.02
0.48
0.02
EMISSIONS
0.64
0
0
0.46
FIBER
0.2
FOODSTUFFS/SOFTS
0.36
1.0
0.13
0.42
0.01
0.7
0.17
0.2
0.5
0.12
Fertilizer
0.2
0.48
0.11
0.05
0.0
0.09
0.77
0.13
1.0
0.3
0.7
0.06
0.4
0.04
0.15
0.5
0.17
0.61
0.2
0.08
0.13
0.06
0.0
0.08
0.15
0.64
0.32
1.0
vol 10's, t
0.11
GRAINS
0.04
0.7
0.0
0.22
0.5
0.09
0.17
0.32
Product subgroup
0.15
0.6
0.2
0.03
0.13
0.0
0.15
BASE METALS
0.07
0.13
CHEMICALS
0.25
0.04
0.28
0.52
0.61
CURRENCY(NONMAJOR)
0.59
0.05
OTHER AGRICULTURAL
INT RATES U.S. TRES
1.0
0.07
0.21
0.7
0.1
0.25
0.12
WOOD PRODUCTS
0.46
CURRENCY(NONMAJOR)
0.03
0.01
DAIRY PRODUCTS
0.02
FIBER
0.59
0.06
0.5
0.32
0.33
PLASTICS
0.17
REAL ESTATE
0.22
0.01
0.14
0.05
0.2
0.4
0.03
Yield Insurance
0.2
0.05
0.05
PRECIOUS METALS
PLASTICS
0.36
0.67
0.13
0.0
0.02
1.0
0.43
0.7
WEATHER
0.07
1.0
0.53
0.7
0.32
STOCK INDEX
REAL ESTATE
vol 10's, t+1
0.5
WOOD PRODUCTS
vol 1's, t+1
0.5
vol 0, year t+1

Yield Insurance
Pr(yearonyear move)
50
0.13
0.8
0.08
0.34
0.36
0.19
0.03
0.05
0.17
0.64
0.11
0.13
0.72
0.06
0.1
0.83
0.02
0.11
0.73
0.08
0
5
0.2
5
0.2
0.0
1.0
0.7
0.5
0.07
0.75
5
0.12
0.0
1.0
0.7
0.5
0.2
0.03
5
0.01
0.74
0.01
0.02
0.2
0.03
0
0
0.71
0.08
0.14
0.01
0.01
0.68
0.01
0.13
0.14
0.2
0.71
0.0
0.11
0.01
0.1
1.0
0.49
0.17
0.7
0.06
0.03
0.09
0.67
0.19
0.01
1.0
0.41
0.07
0.7
0.12
0.6
0.48
0.0
1.0
0.7
0.21
0.25
0.5
0.09
0.69
0.16
0.2
0.02
0.86
0.01
0.11
0.16
0.17
0.05
0.71
0.04
0.1
0.02
0.01
0.16
0.05
0.67
0.0
0.12
0.01
0.01
0.01
0.04
0.74
0.03
0.15
0.01
0.01
0.62
0.14
0.01
0.01
0.74
0.02
0.05
0.0
0.13
0.11
0.63
0.15
0.02
0.03
0.03
Fertilizer
1.0
0.13
0.56
0.43
0.22
0.4
0.13
0.1
0.06
0.0
INT RATES U.S. TRES

0.7
0.01
0.46
0.02
0.74
0.02
vol 10^4's, t
0.04
0.34
0
0.32
0.09
0.53
0.08
0.02
0.24
0.01
0.04
0.62
0.02
0.16
0.01
0.2
0.49
0.12
0.06
0.01
0.27
0.02
0.43
0.0
0.06
0.1
0.03
0.04
0.02
0.57
0.21
1.0
0.7
0.48
0.2
0.13
0.5
0.06
0.01
0.04
0.58
0.24
0.04
0.32
0.02
0.15
0.5
Product subgroup
0.4
0.05
BASE METALS
0.2
0.06
CURRENCY
CHEMICALS
0.03
0.2
CURRENCY(NONMAJOR)
WOOD PRODUCTS
DAIRY PRODUCTS
0.1
0.68
0.29
0.5
0.01
0.02
0.02
0.08
0.21
0.66
0.02
0.13
0.08
0.01
0.13
0.11
0.12
CHEMICALS
0.29
0.09
0.15
CURRENCY
0.04
0.02
1.0
0.07
0.08
0.7
0.28
0.04
0.5
0.13
0.2
vol 1000's, t
0.17
0.01
0.35
0.05
0.14
0.21
0.03
0.02
0.03
0.25
0.07
0.26
0.08
0.03
CURRENCY(NONMAJOR)
EMISSIONS
0.01
0.43
0.11
FIBER
0.13
0.02
0.19
FOODSTUFFS/SOFTS
EMISSIONS
GRAINS
0.01
0.05
0.03
0.21
0.06
0.52
OTHER AGRICULTURAL
0.16
0.02
0.58
0.07
0.42
0.34
0.01
0.5
0.22
0.55
PLASTICS
0.13
0.13
0.37
PRECIOUS METALS
0.32
REAL ESTATE
0.01
0.11
0.25
0.18
INT RATES U.S. TRES
STOCK INDEX
0.61
0.5
WEATHER
0.07
Yield Insurance
0.01
0.14
BASE METALS
0.06
0.1
DAIRY PRODUCTS
0.22
0.44
0.14
0.17
FIBER
FOODSTUFFS/SOFTS
0.01
0.05
Fertilizer
0.74
0.13
0.07
0.44
0.02
0.02
0.27
0.03
0.07
GRAINS
0.1
0.47
0.5
0.05
0.07
0.65
0.2
0.02
0.17
0.07
OTHER AGRICULTURAL
0.08
0.2
PLASTICS
PRECIOUS METALS
0.18
0.07
0.03
0.0
0.17
0.4
0.0
0.09
STOCK INDEX
0.25
0.29
WEATHER
0.03
1.0
0.03
0.61
1.0
0.12
vol 100's, t+1
0.7
WOOD PRODUCTS
REAL ESTATE
vol 10's, t+1
0.7
vol 1's, t+1
0.5
0.5
vol 0, year t+1

Yield Insurance
Pr(yearonyear move)
51
OTHER AGRICULTURAL
0
0
0.01
INT RATES U.S. TRES
GRAINS
0
0
0.02
Fertilizer
FOODSTUFFS/SOFTS
FIBER
DAIRY PRODUCTS
CURRENCY(NONMAJOR)
CURRENCY
0.01
0.01
BASE METALS
0.02
0.76
0.06
0.74
0.01
0.03
0.88
0.01
0.99
0.02
0.74
0.02
0
0.01
0.24
0.01
0.73
0.01
0.94
0.02
0.02
0.06
0.04
0.01
0.92
0
5
0.74
0.06
0.2
0.0
1.0
0.7
0.5
0.04
0.89
0.2
0.0
0.02
0.06
0.02
0.6
0.26
0.01
0.8
0.04
CHEMICALS
0.04
0.01
0.94
0.11
0.73
0.02
0.01
0.02
0.04
0.02
0.78
0.02
0.01
0.18
0.01
0.06
EMISSIONS
0.81
0.74
0.1
0.07
0.01
0.91
0.01
0.01
0.13
0.01
0.8
0.02
0.01
0.73
0.03
0.04
0.78
0.13
0.02
0.01
0.91
vol 10^6's, t
0.74
0.01
0.74
0.03
0.02
0.01
0.02
0.01
0.06
0.81
0.02
0
0
0.02
PLASTICS
0.08
PRECIOUS METALS
0.01
1.0
0.74
REAL ESTATE
0.02
0.7
0.01
0.02
0.74
0.5
0.01
0.74
0.2
0.01
0.02
STOCK INDEX
0.25
0.0
0.56
0.02
0.01
0.16
0.73
0.07
1.0
0.71
WEATHER
0.15
0.7
WOOD PRODUCTS
0.02
0.08
0.73
0.17
0.5
Yield Insurance
0.03
0.01
0.2
0.99
0.12
BASE METALS
0.61
0.0
0.01
CHEMICALS
0.23
0.03
1.0
0.52
0.7
0.16
CURRENCY
0.29
0.65
0.5
0.11
0.16
0.05
CURRENCY(NONMAJOR)
0.2
0.01
DAIRY PRODUCTS
0.03
0.0
0.01
1.0
0.86
EMISSIONS
0.08
0.7
0.21
0.69
0.73
0.01
0.12
0.02
0.5
FIBER
0.91
0.06
0.85
0.05
0.01
0.05
0.2
0.54
FOODSTUFFS/SOFTS
0.16
0.0
0.08
0.06
0.01
0.09
1.0
Fertilizer
0.16
0.69
0.01
0.82
0.74
0.02
0.01
0.01
0.21
0.02
GRAINS
0.07
0.7
INT RATES U.S. TRES
0.02
0.03
0.74
0.5
0.06
0.02
vol 10^5's, t
0.75
0.87
0.01
0.18
0.2
0.05
0.11
0.69
0.74
0.01
0.11
0.01
0.0
OTHER AGRICULTURAL
0.02
0.01
0.74
0.0
0.01
0.14
0
0.5
0.02
0.05
1.0
0.01
1.0
0.42
0.7
0.01
0.02
0.64
0.06
0.2
PRECIOUS METALS
0.12
0.01
0.0
PLASTICS
Product subgroup
REAL ESTATE
0.02
0.02
0.01
0.74
0.02
0.01
0.92
1.0
0.74
0.01
0.05
0.7
0.04
0.02
1.0
0
0.02
0.5
0.7
0.7
WEATHER
STOCK INDEX
0.2
0.5
vol 100's, t+1
0.5
WOOD PRODUCTS
vol 10's, t+1
0.5
vol 1's, t+1
0.2
0.0
vol 0, year t+1

Yield Insurance
Pr(yearonyear move)
given annual volumes 10, 000 and < 1, 000, 000
52
INT RATES U.S. TRES
GRAINS
Fertilizer
FIBER
EMISSIONS
DAIRY PRODUCTS
CURRENCY(NONMAJOR)
CURRENCY
0
1.0
0.7
0.5
0.87
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.92
0.92
0.92
0.92
0.92
0.92
0.92
0.92
0.92
0.13
0.79
0.02
0.92
0.21
0
0.2
0.0
0.0
BASE METALS
0.02
CHEMICALS
0.99
0.92
0
0.02
0.92
0.92
0.02
0.02
FOODSTUFFS/SOFTS
0.92
0.99
0.02
0.92
vol 10^8's, t
0.02
0.92
0.92
0
OTHER AGRICULTURAL
0.92
0.02
0.92
1.0
0.92

0.02
0.7
0.92
0.02
0.08
0.92
0.5
PLASTICS
0.02
0.92
0.2
0.92
0.02
PRECIOUS METALS
0.02
0.0
0.95
0.01
1.0
REAL ESTATE
0.73
0.01
0.95
0.01
0.73
0.7
0.02
0.01
0.74
0.5
STOCK INDEX
0.01
0.04
0.02
0.74
0.01
0.74
0.2
0.02
WEATHER
0.02
0.74
0.0
0.02
0.98
1.0
WOOD PRODUCTS
0.02
0.01
0.74
0.7
0.02
Yield Insurance
0.91
0.5
0.06
0.13
0.2
Product subgroup
CHEMICALS
BASE METALS
0.76
0.09
0.09
0.0
0.89
CURRENCY
0.01
0.95
0.74
1.0
CURRENCY(NONMAJOR)
0.03
0.7
DAIRY PRODUCTS
0.9
0.02
0.01
0.92
0.5
0.74
0.2
0.03
EMISSIONS
0.07
0.0
0.02
0.91
1.0
FIBER
0.7
0.02
0.93
0.01
0.73
0.5
Fertilizer
FOODSTUFFS/SOFTS
0.01
0.04
0.2
0.01
0.03
0.01
0.74
0.77
0.0
GRAINS
1.0
0.7
0.5
INT RATES U.S. TRES
0.01
0.03
0.73
0.21
0.02
0.01
0.92
vol 10^7's, t
0.02
0.01

0.01
0.74
0.0
0.02
OTHER AGRICULTURAL
0.74
0.01
0.01
0.02
1.0
1.0
0.74
0.02
0.0
PLASTICS
0.01
0.02
1.0
REAL ESTATE
PRECIOUS METALS
0.7
0.7
0.5
0.7
0.01
0.5
WEATHER
STOCK INDEX
0.2
0.5
vol 100's, t+1
0.2
WOOD PRODUCTS
vol 10's, t+1
0.2
vol 1's, t+1
0.2
0.0
vol 0, year t+1

Yield Insurance
Pr(yearonyear move)
given annual volumes 1, 000, 000
53

Oce Lifecycleofexchange

Uploaded by

Copyright:

Available Formats

Oce Lifecycleofexchange

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Oce Lifecycleofexchange

Uploaded by

Copyright:

Available Formats

The Lifecycle of Exchange-traded Derivatives

Grant Cavanaugh Ph.D.

JEL classification: G12, N22, G17.

Every year, financial innovators launch new derivatives markets in the

form, as a non-stationary Markov model estimated using Bayesian methods.

Derivatives reform and lifecycle statistics

Basic statistics on the lifecycle of derivatives are particularly valuable now

Our analysis is based on annual volume figures for US exchange-traded

Markov model for the lifecycle of derivatives

We present our primary results in the form of a Markov model. That

P (Xt = xt |Xt1 = xt1 , Xt2 = xt2 , . . . , X0 = x0 ) = P (Xt = xt |Xt1 = xt1 )

Contract are assigned a martix (P as in equation 3) that describes the

p1,1 p1,2 . . . p1,j

p2,1 p2,2 . . . p2,j

pi,1 pi,2 . . . pi,j

Volume level for any given contract-year is equivalent to the common

Volume levelyear t+1 |Volume levelyear t Categorical()

Dirichlet(xvol level 0 , xvol level 1 , . . . , xvol level 108 )

We can multiply the vector k by yet another vector of annual trading

Prior probabilities on moving between states of annual volume

The probability of a contract jumping more than one order of magnitude

Derivatives volumes over time

Concentration of trading volume over time

Figures 1 and 2 display the empirical cumulative distribution function

However, figure 1 obscures substantial variation in the concentration of

Probability of individual contracts moving to different levels of trading by decade

Fig. 2. Empirical cumulative distribution function (ECDF) of annual trading

4, combine to form the transition matrix for a Markov model of a contract

vol 10^8's, t+1

vol 10^7's, t+1

vol 10^6's, t+1

vol 10^5's, t+1

vol 10^4's, t+1

vol 1000's, t+1

vol 100's, t+1

vol 10's, t+1

vol 1's, t+1

vol 0, year t+1

Fig. 4. Transition matrix for Markov model of derivatives contract moving

vol 10^8's, t+1

vol 10^7's, t+1

vol 10^6's, t+1

vol 10^5's, t+1

vol 10^4's, t+1

vol 1000's, t+1

vol 100's, t+1

vol 10's, t+1

vol 1's, t+1

vol 0, year t+1

Fig. 5. Transition matrix for Markov model of derivatives contract moving

vol 0, year t+1

vol 0, year t+1

vol 1's, t+1

vol 10's, t+1

Fig. 8. Probability of transition from annual volume of 0 to annual volume in

vol 100's, t+1

vol 1000's, t+1

vol 10^4's, t+1

volume level vol