Market Manipulation and Public Confidence in Prediction Markets. Please Comment

[John Delaney, CEO, Intrade.com]

Hello Everybody!

If we believe that market manipulation can occur in prediction markets
(aka "event markets"), I believe it is both appropriate and timely for
our industry to discuss this important issue again and issue guidance
where appropriate.

If you are reading this post, you are a Prediction Market stakeholder
and I heartily encourage you to comment.

Below are just a few jumbled questions to catalyze the discussion…

Does market manipulation (real or imagined) undermine public
confidence / participation in prediction markets?

Should prediction markets look to adopt anti-market manipulation
practices that originated in the 1930’s?

Should or can automated anti-manipulation protocols be effectively
applied to our markets?

What is the relationship between insider trading, market manipulation
and predictive accuracy in our markets?

If insider trading makes markets more efficient/predictive how does
that relate to anti-manipulation opinions?

Is defining and detecting manipulative practices so difficult to make
any guidance pointless?

Thanks to all in advance!

Regards,

John

[Jim]

I'm somewhat thinking out loud here and have not given a whole lot of
thought to the questions specifically proposed, but I do have one quick
response. That is.

My position is one of the favorable attributes of prediction markets is it
facilitates insider trading with the resulting benefit that knowledge as
applicable to the subject outcome can be shared without compromising the
underlying knowledge. Any attempt to constrain insider trading, as though
there is something inherently wrong with it, will defeat the intended
purpose of the market.

Jim Neiers

[John Delaney, CEO, Intrade.com]

Thanks for your comment and reply Jim.

Love to hear other opinions and perspectives.

[Thomas S Davis]

John -- "What is the relationship between insider trading, market manipulation

and predictive accuracy in our markets?"

To my mind, insider trading means an activity that will push a
prediction market towards more efficiency and more accuracy whereas
market manipulation does the opposite. Insider trading to me means
one or more individuals acting with some degree of independence,
typically selling together or buying together to take advantage of
"material non-public information." A liquid and otherwise well
functioning market has a pretty good chance of correcting itself in
the face of this kind of insider trading. Market manipulation, to me,
is more devious -- groups of traders buying and selling from one
another in order to give the appearance of higher prices, higher
volumes and higher value for the security in question. Market
manipulation could also work on the sell side, but I don't recall many
examples of this, perhaps because the ultimate profits from such a
scheme are limited. Buying pools, at least in theory, could be
maintained for a long time.

How does the recent naked short selling scandal fit into this
definition? To me this points up the difference between a futures
market and an equity market. The key difference is liquidity and the
duration of the security. Equities are designed to be nearly infinite
duration securities and compared to commodities or currencies they are
relatively illiquid. Futures have a much shorter duration and there
is a physical settlement which even for long dated futures tends to
keep things more honest. Event Contracts or prediction markets are
much more like futures, but can be illiquid. Their short duration,
however, minimizes chances for manipulation.

Having said all that I think it is essential that prediction markets,
especially public ones, be monitored for market manipulation schemes
and enforcement regimes developed. The amount of creatively a
community will spend in gaming a game can never be underestimated.
Tom

--
Thomas S. Davis
President
Cygnus Associates, LLC
11835 Roe Ave. #188
Leawood, KS 66211
tda...@cygnusassociates.com
www.cygnusassociates.com

[Jason Ruspini]

Hi John, let me offer a few scattered thoughts in response..

Aside from letting things take care of themselves, position limits are
both the easiest and most powerful way of combating price
manipulation. Beyond that the CFTC maintains a large trader reporting
infrastructure:
http://www.cftc.gov/industryoversight/marketsurveillance/ltrp.html
If Intrade wishes to come under CFTC jurisdiction, some of your
questions are basically answered for you.

But as the exchange currently operates, there may be more liberal ways
to address manipulation concerns. I am sure that Intrade maintains a
measure of market balance for margined contracts. That is, a market
is most off-balance if it consists of a single trader on one side of
the open interest and many traders on the other side. This end of the
spectrum is indicative of market power and possibly manipulation, but
rather than disallowing or scrutinizing large positions, Intrade could
instead publish or make available upon request some metric indicating
the "entropy" of positions such as the variance of position sizes. A
low balance metric would warn traders and price interpreters that
something may be fishy. Of course it could just indicate a well-
funded trader or an insider. We can begin to test the usefulness of
this metric though. Take all closed markets and for each day
calculate the balance metric, e.g. variance of position sizes. Then
for each market, sort all days by the metric. Did the closing prices
for the top 50% most off-balance days have more error than the prices
in the lower 50% based on the outcome of the contract? (This might
not work if markets tend to display a trend in balance as they
approach maturity.) How do some specific contracts that have elicited
manipulation speculation look according to this analysis? Gore
Nomination? Clinton Presidency? Giuliani nomination? (Let's just
assume they have already settled at zero.) I don't know if this is
going to work, but if it does, you have a very cheap, gentle way to
address potential manipulation. It's basically a warning label for
traders, and at the same time large traders are more likely to become
paranoid and resist exercising market power unless they really have
information, as they are more vulnerable to a "run" by other traders.
(The metric will not indicate what side of the market the large
traders are on, but this can be inferred by the price action.)

The CFTC generally does not regulate against insider trading, although
with event markets, they will regulate against outcome manipulation
which happens to be a type of insider trading. Outside of CFTC
jurisdiction, outcome manipulation may be desirable because it can act
as a prize mechanism for research questions, although such markets may
suffer from low volume.

Note that the familiar argument for not regulating against (price)
manipulation suggests that one *should* regulate against insider
trading. The argument roughly goes that manipulation ultimately
increases market accuracy because traders are attracted to markets
“subsidized” by manipulators, and this liquidity eventually leads to
greater accuracy. Therefore traders would be less likely to
participate in markets where insider trading, a negative subsidy by
this logic, is allowed, and this would ultimately decrease liquidity
and accuracy. You could argue that the decreased long-term information
due to discouraged liquidity is less significant than the initial
short-term information gains due to insider trading, but then wouldn’t
the increased long-term information due to encouraged liquidity be
less important than the initial short-term information loss due to
manipulation? I am not making a claim one way or the other.
Different assumptions will
lead to different answers, but there does seem to be some tension
here.

Thomas brings up a very important aspect of this question. In what
ways might (detection of) manipulation differ between traditional
futures and binary prediction markets? In one way, binary options are
less vulnerable to manipulation since they settle according to an
objective event at a specific time, whereas futures prices are open-
ended in one direction and their "meaning" is always open to
interpretation. In another way, manipulation may be more difficult to
detect in prediction markets, and some indicia that are available to
the CFTC for detecting corner-type manipulations do not have clear
analogs for all market power manipulations. It is highly suggestive
if someone is willing to buy december silver at 50 while march silver
sits at 30, or is willing to buy bullion at levels far above coin and
refining prices, or willing to bid silver up when gold has not moved,
etc. Prediction markets have similar checks but are still illiquid
enough that a
well-funded trader could overwhelm all related contracts. More
significantly, even with deep markets, it might be impossible to tell
whether a binary price is being manipulated within 10% or more,
especially if insiders are permitted. While this may not satisfy the
CFTC, letting these situations take care of themselves is probably the
best option for Intrade, perhaps alongside published metrics as
mentioned above.

Finally, the CFTC also disallows "wash" trades, which are trades
designed to give the impression of high volume at certain prices. For
instance, two friends could conspire to not only move the price of the
Iran Strike contract, but also trade back-and-forth with minimal
profit/loss implications to create the impression of extraordinary
volume in order to lend credence to their manipulation. I mention
this, because I'm sure we realize that if something like this happened
and CNBC picked-up on it leading to a temporary move in oil prices,
that would be very negative for the industry.

Hope this helps,
Jason Ruspini

[John Delaney, CEO, Intrade.com]

Thanks Jason and Thomas for the comprehensive and illuminating
replies.

They really are great.

Lets have some others comments please.

When posting the original questions above I wasn't so much thinking
about Intrade as I was about the broader industry.

So I will hold off replying on specific Intrade related questions for
now to ensure this stays an industry focused thread.

Happy to discuss our approach and ideas later.

Best,

John

[azman]

I think there can be no clearer sign of market manipulation than this
market, currently featured on Intrade.

2008.PRES.GORE
345 1.2 1.2 131 1.2 188.2k 0

That's right, one trader is posting a locked market at 1.2-1.2
345x131.

Should we believe that this trader is short 345 lots? Or is he long
131 lots? Or does John Delaney just not want to see traders get out of
their short Gore positions anywhere below 1.0 (the implied price where
he expires their winning positions).

On Aug 4, 11:31 pm, "John Delaney, CEO, Intrade.com"