Posted by Jim Lindgren:
Brewing Scandal at CNBC: Allegations of Cheating in Stock Contest.--
http://volokh.com/archives/archive_2007_05_27-2007_06_02.shtml#1180718507
There is a fascinating scandal brewing over at [1]CNBC.
CNBC has [2]delayed announcing the winner of its Million Dollar
Portfolio Challenge contest, which ended a week ago, because of
allegations of cheating in the form of a version of after-market
trading. With a $1,000,000 annuity as the prize, the contest proceeded
in two stages: a 10-week contest followed by a 2-week finals for the
10 weekly winners and 10 others with the best overall portfolio
results. There was also a 2-week consolation contest for those who
didn�t make the finals. Although CNBC and its corporate blog have not
explained the nature of the problem, commenters to other blogs have
been posting some superficially plausible speculations about the
complaints of the losing finalists.
So far we have:
(1) a string of highly suspicious (and almost impossibly savvy) trades
by some of the 20 contestants in the finals,
(2) an unsubstantiated claim by an anonymous blog commenter that he
knows how such trades were accomplished because he found a software
glitch that he used to cheat his way to a near-winning performance in
the consolation round contest, and
(3) the announcement by CNBC that patterns of suspicious trading and
allegations of irregularities have caused them to launch an internal
investigation before declaring a winner in the contest that ended a
week ago.
It will be interesting to see how the journalists at CNBC handle this
CNBC scandal, in particular what they disclose about when they first
heard about any problems and what they tried to do to prevent any
violations. So far Mark Koba, the [3]corporate blogger assigned to the
contest, has performed shamefully. If Koba has been muzzled by his
superiors, he should disclose this fact in a matter of fact way. If he
has been muzzled and ordered not to disclose that he has been muzzled,
then he's in a tougher spot (journalism jobs are hard to come by; not
every outrage is a quitting offense). Will CNBC sweep any
irregularities under the rug or instead engage in serious financial
journalism and be tough on anyone that they find to have used a
software glitch to violate the rules of the contest.
In the absence of meaningful disclosure by CNBC, here is what has been
alleged about how after-market trading might have been accomplished.
Some companies release earnings before the market opens or during the
trading day, but some release just after the market�s 4pm close.
Contestants could make only one set trades every day, with their
choices supposedly made by 3:59pm ET each afternoon, just before the
market�s close. Traders were working with an imaginary portfolio of a
million dollars. Each day before 4pm one could, for example, enter
trades buying 200,000 shares of each of 50 stocks including the dozen
or two stocks releasing earnings just after that day�s close. If
nothing further were done, only the first of the 50 trades submitted
would be completed, investing essentially all of the contestant�s
portfolio in one stock (with perhaps a single share of a later
low-priced stock being purchased as well with the few dollars
remaining after the first trade was entered). Before the 4pm market
close, the contestant could alter the priority of her 50 orders, which
would change which stock trade would consume 99.9% of that day�s
investment.
But according to allegations by blog commenters, if a contestant kept
her computer CNBC pending trades window open at the close, she could
wait for earnings to be released after 4pm and watch the movements of
those stocks in after-market trading until perhaps 4:30pm. (Scroll
down to comments on a non-CNBC blog [4]here at 5/25/2007 10:09 AM,
5/25/2007 3:16 PM, 5/25/2007 3:38 PM, 5/25/2007 3:59 PM, 5/25/2007
4:31 PM, and 5/27/2007 6:14 AM.) The stock that jumped the most in
after-market trading until 4:30pm could be moved to top of the
priority list, and CNBC would then process that stock trade as if it
had been made BEFORE the 4pm close, rather than AFTER the earnings
were released a few minutes after 4pm.
If these allegations are true--and for now CNBC is not saying anything
informative--CNBC has a mess on its hands, since presumably the same
tactic could have been performed during the rest of contest as well.
At least as to the 20 finalists, one hopes that CNBC has adequate
computer time stamps to show when the priority of a set of orders was
changed by a contestant. If not, they would be forced to rely on the
statistical probabilities of one person (or several people) picking
strings of winners.
([5]show the rest of the post)
Although some blog commenters seem to view the contest rules
(available from a link on the right side of [6]this page) as
ambiguous, I think that the trading strategy described is adequately
prohibited (though, of course, in hindsight language could always have
been clearer). The rules provide:
Each Participant can make a maximum of fifty (50) trades per �Day�,
based on the time the trade is entered by the Participant on the
Site, not the time the trade is executed. A �Day� is defined as
4:00 p.m. to 3:59:59 p.m. on the next trading day.
A trade is defined as a single purchase of one or more shares of a
single stock for �cash� or a sale of one or more shares of a single
stock for �cash.� . . . Participants will have the ability to
prioritize trades within a given day of trading; provided that, all
sales will be processed before all purchases. If a Participant does
not elect to grant a ranking of priority, trades will default to
prioritization based upon the time the trade is entered with
earlier trades taking priority over any subsequently entered
trades.
If a Participant submits a trade that either exceeds the daily
maximum number of trades or is otherwise inappropriate, such trade
will not be processed. . . . Stocks will be purchased in the order
of priority until funds in the Participant�s account are depleted.
Pending trades can be cancelled up until 3:59:59 p.m. on the same
day of the trade, however, all trades are final after 3:59:59 p.m.
As I read it, one can prioritize trades only within a trading day,
which is defined as ending before 4pm. Further, �all trades are final
after 3:59:59 p.m.,� which I interpret as meaning that after 4pm one
can�t instruct CNBC not to execute a trade investing 99.9% of one�s
portfolio in one stock and instead to execute a different trade
investing 99.9% of one�s portfolio in a different stock.
The complete failure of the [7]corporate blog to explore this issue�or
even to explain the nature of the problem�is disappointing. A
corporate blogger would have access to officials who could detail the
allegations, who might be filing charges, when the charges first
arose, and to what extent this strategy might have affected who made
the finals in the first place. Personal Note: VC readers may remember
that I [8]blogged about early problems with the CNBC stock trading
challenge. With a $1,000,000 annuity as the prize, the contest
proceeded in two stages: a 10-week contest followed by a 2-week finals
for the 10 weekly winners and 10 others with the best overall
portfolio results.
By the way, although CNBC erased my portfolio so I can�t give you my
exact final results for the 10-week contest, I was in the top 1% of
the more than a million portfolios (with as I recall over a 45%
return). Unlike the weekly and 10-week winners, who did much better
than I did by picking the best (or among the best) performing of those
companies releasing earnings results on several days, I traded
ordinary stocks based on daily predictions generated by some
statistical models I�ve developed.
([9]hide)
References
1. http://contests.cnbc.com/milliondollar/ContestClosed.jsp
2. http://contests.cnbc.com/milliondollar/ContestClosed.jsp
3. http://www.cnbc.com/id/17506588/site/14081545/
4.
https://www.blogger.com/comment.g?blogID=26194910&postID=4296240513484372156
5. file://localhost/var/www/powerblogs/volokh/posts/1180718507.html
6. http://contests.cnbc.com/milliondollar/ContestClosed.jsp
7. http://www.cnbc.com/id/17506588/site/14081545/
8. http://volokh.com/archives/archive_2007_03_11-2007_03_17.shtml#1173763490
9. file://localhost/var/www/powerblogs/volokh/posts/1180718507.html
_______________________________________________
Volokh mailing list
[email protected]
http://lists.powerblogs.com/cgi-bin/mailman/listinfo/volokh
