United States Court of Appeals
For the Eighth Circuit
___________________________
No. 18-3689
___________________________
Roderick Ford,
Plaintiff Appellee,
v.
TD Ameritrade Holding Corporation; TD Ameritrade, Inc.; Frederic J. Tomczyk,
Defendants Appellants.
------------------------------
Securities Industry and Financial Markets Association; Chamber of Commerce of
the United States of America,
Amici on Behalf of Appellants,
Better Markets, Inc.,
Amicus on Behalf of Appellee.
____________
Appeal from United States District Court
for the District of Nebraska - Omaha
____________
Submitted: September 23, 2020
Filed: April 23, 2021
____________
Before COLLOTON, GRUENDER, and GRASZ, Circuit Judges.
____________
COLLOTON, Circuit Judge.
A customer of TD Ameritrade, Inc., sued the company and two other
defendants for securities fraud in the District of New Jersey. He purported to sue on
behalf of himself and all similarly-situated customers of TD Ameritrade. The district
court in New Jersey later appointed Roderick Ford as lead plaintiff, and the court then
transferred the action to the District of Nebraska. The district court in Nebraska
certified a class under Federal Rule of Civil Procedure 23(b)(3), and the defendants
appeal that order. We conclude that the proposed class does not satisfy the
requirements of Rule 23, and we therefore reverse.
I.
TD Ameritrade offers brokerage services to retail investors. The company is
the nation’s third largest discount brokerage, serving over six million clients. TD
Ameritrade customers can trade stocks by submitting orders through the company’s
online platform. The company itself does not execute customer orders, but instead
routes orders to trading venues (such as a stock exchange) for fulfillment. The
company generally transmits orders using a computerized routing system.
Ford was appointed in 2014 as lead plaintiff for a group of investors who
purchased and sold securities through TD Ameritrade between 2011 and 2014. He
alleges that TD Ameritrade’s order routing practices violate the company’s “duty of
best execution” by systematically sending customer orders to trading venues that pay
the company the most money, rather than to venues that provide the best outcome for
customers. The duty of best execution requires that brokers “use reasonable efforts
to maximize the economic benefit to the client in each transaction.” Newton v.
-2-
Merrill Lynch, Pierce, Fenner & Smith, Inc., 259 F.3d 154, 173 (3d Cir. 2001)
(internal quotation omitted).
Ford maintains that TD Ameritrade caused customers to suffer economic loss
by leaving orders unfilled, filling orders at a sub-optimal price, and filling orders in
a manner that adversely affected performance after execution. The complaint asserts
that TD Ameritrade, its parent company, and its chief executive officer, Frederic J.
Tomczyk, violated § 10(b) of the Securities Exchange Act of 1934, 15 U.S.C.
§ 78j(b), and the Securities and Exchange Commission’s Rule 10b-5. 17 C.F.R.
§ 240.10b-5. The complaint also asserts that Tomcyzk is jointly and severally liable
as a “controlling person” of the company under § 20(a) of the Act. 15 U.S.C.
§ 78t(a).
Ford moved for class certification in 2017. A magistrate judge concluded that
the proposed class did not satisfy the requirements of Rule 23(b)(3) and
recommended denying certification. The judge reasoned that determining whether
each TD Ameritrade customer suffered economic loss as a result of the company’s
order routing practices would entail an order-by-order inquiry, and that common
issues thus did not predominate over individual questions.
On review of the recommendation, however, the district court determined that
Ford’s expert had developed an algorithm that could solve the predominance problem
by making automatic determinations of economic loss for each customer. The court
certified a class consisting of “[a]ll clients of TD Ameritrade between September 15,
2011 and September 15, 2014 who placed orders that did not receive best execution,
in connection with which TD Ameritrade received either liquidity rebates or payment
for order flow, and who were thereby damaged.”
-3-
This court granted the defendants permission to appeal the class certification
order. See Fed. R. Civ. P. 23(f). We review the order for abuse of discretion. IBEW
Local 98 Pension Fund v. Best Buy Co., 818 F.3d 775, 779 (8th Cir. 2016).
II.
A.
To justify certification of a class, plaintiffs must meet all of the requirements
of Federal Rule of Civil Procedure 23(a) and satisfy one of the three subsections of
Rule 23(b). The district court certified a class based on Rule 23(b)(3), which requires
that “questions of law or fact common to class members predominate over any
questions affecting only individual members, and that a class action is superior to
other available methods for fairly and efficiently adjudicating the controversy.”
“An individual question is one where ‘members of a proposed class will need
to present evidence that varies from member to member,’ while a common question
is one where ‘the same evidence will suffice for each member to make a prima facie
showing [or] the issue is susceptible to generalized, class-wide proof.’” Tyson Foods,
Inc. v. Bouaphakeo, 577 U.S. 442, 453 (2016) (quoting 2 William B. Rubenstein,
Newberg on Class Actions § 4:50, at 196-97 (5th ed. 2012)). If the plaintiffs’ method
of proving their claim would “include individualized inquiries that cannot be
addressed in a manner consistent with Rule 23, then the class cannot be certified.”
Harris v. Union Pac. R.R. Co., 953 F.3d 1030, 1035 (8th Cir. 2020) (internal
quotation omitted).
Ford alleges that TD Ameritrade violated § 10(b) of the Securities Exchange
Act and Rule 10b-5. We do not address the merits at this stage, but we do consider
the nature of the underlying claim to determine its suitability for class certification.
See Harris, 953 F.3d at 1033. Section 10(b) forbids the use, in connection with the
-4-
purchase or sale of a security, of “any manipulative or deceptive device or
contrivance in contravention of” regulations promulgated by the SEC for the
protection of investors. 15 U.S.C. § 78j(b). The SEC promulgated Rule 10b-5 to
enforce § 10(b). Rule 10b-5 prohibits making an untrue statement of material fact or
omitting to state a material fact in connection with the purchase or sale of a security.
17 C.F.R. § 240.10b-5(b). It also forbids engaging in “any act, practice, or course of
business which operates or would operate as a fraud or deceit upon any person, in
connection with the purchase or sale of any security.” Id. § 240.10b-5(c).
The Supreme Court has “long recognized an implied private cause of action to
enforce [§ 10(b)] and its implementing regulation.” Halliburton Co. v. Erica P. John
Fund, Inc., 573 U.S. 258, 267 (2014). To recover damages for violations of § 10(b)
and Rule 10b-5, a plaintiff must prove “(1) a material misrepresentation or omission
by the defendant; (2) scienter; (3) a connection between the misrepresentation or
omission and the purchase or sale of a security; (4) reliance upon the
misrepresentation or omission; (5) economic loss; and (6) loss causation.” Amgen
Inc. v. Conn. Ret. Plans & Tr. Funds, 568 U.S. 455, 460-61 (2013) (internal quotation
omitted).
B.
This case involves a dispute about a broker’s compliance with its duty of best
execution. Best execution cases differ from typical securities fraud cases under Rule
10b-5, where the alleged fraud directly affects the price of a security. See Newton,
259 F.3d at 173, 179-80. When a broker’s fraud directly affects the price of a
security, the customer trading in that security in reliance on the broker’s
representation can easily demonstrate that, but for the broker’s fraud, the customer’s
trade would have executed at a more favorable price. See id. at 180.
-5-
Here, by contrast, the economic loss allegedly caused by TD Ameritrade’s
order routing practices is “the difference between the price at which [customers’]
trades were executed and the ‘better’ price allegedly available from an alternative
trading source.” Id. at 178. To justify class certification, Ford must show that he can
establish this type of economic loss for a class of plaintiffs in a manner consistent
with the predominance requirement of Rule 23.
Ford’s expert proposes to analyze “hundreds of millions of data points”
through an algorithm. The algorithm would assess execution quality by using class
trading history data provided by TD Ameritrade and data about the state of the market
at the time of each trade. The expert shared the “specification” and “logic” of his
proposed algorithm with the parties and the court, but he has not disclosed the
software code he proposes to use for automating the analysis.
In Newton, the Third Circuit affirmed the denial of class certification in a best
execution case. 259 F.3d at 162. The court observed that “[w]hether a class member
suffered economic loss from a given securities transaction would require proof of the
circumstances surrounding each trade, the available alternative prices, and the state
of mind of each investor at the time the trade was requested.” Id. at 187. The alleged
injuries arose out of the execution of “hundreds of millions of trades.” Id. at 190.
Because “[d]etermining which class members were economically harmed would
require an individual analysis into each trade and its alternatives,” the Third Circuit
concluded that individual questions were “overpowering.” Id. at 189. We find
Newton’s reasoning persuasive despite Ford’s attempts to distinguish it.
Ford contends that Newton is inapposite because it involved different
technology. In Newton, the plaintiffs argued that their expert could “devise a formula
for calculating injury and damages.” Id. at 191. Ford maintains that his expert has
already developed an advanced algorithm that can calculate injury and damages on
a class-wide basis, and that the inquiry can be completed upon discovery of class-
-6-
wide trading history. As such, Ford argues that we need not share Newton’s
reluctance to “rely on a formulaic nostrum given the consequences if it fails to meet
expectations.” Id. Even with the proposed algorithm, however, we conclude that
determining economic loss in this case entails individualized inquiry inconsistent
with the predominance requirement of Rule 23.
To succeed on the merits of his claim, Ford must show that TD Ameritrade’s
order routing practices caused its customers to suffer economic loss. See Amgen, 568
U.S. at 460-61. This requirement derives from the “standard rule of tort law that the
plaintiff must allege and prove that, but for the defendant’s wrongdoing, the plaintiff
would not have incurred the harm of which he complains.” Newton, 259 F.3d at 177
(internal quotation omitted).
Ford’s expert proposes to “establish that a ‘better’ price was obtainable for
each executed trade,” id. at 178, by comparing the trade’s actual price with the
National Best Bid and Offer (NBBO) price. The NBBO represents the highest price
a buyer was willing to pay, and the lowest price a seller was willing to accept, for a
particular stock at a given time. But sometimes a trade fails to execute at the NBBO
price through no fault of the broker. For example, volatile or otherwise unusual
market conditions can prevent a trade from executing at that price. The parties’
experts agree that certain transactions must be excluded from the algorithm’s analysis
to account for instances where TD Ameritrade could not have prevented execution
at a price inferior to the NBBO.
The parties’ experts disagree, however, about which transactions should be
excluded. Ford argues that this disagreement is unresolved only because TD
Ameritrade successfully moved to limit discovery of class-wide trading data. Once
discovery is complete, he contends, he will be able to identify all necessary
exclusions.
-7-
The process will not be that simple. Ford’s expert explains that third-party
companies provide historical stock market information that identifies periods when
stocks were traded during unusual market conditions, and argues that his algorithm
can filter out unusual market conditions using these data. But TD Ameritrade’s
expert maintains that not all relevant unusual market conditions are recorded in these
market data, and that others must be identified on a case-by-case basis. As one of
Ford’s experts acknowledged, there is no definitive list of unusual market conditions
that account for transactions that depart from the best available price. As a result, the
algorithm’s use of published market data will not identify all legitimate exclusions,
and the experts will have to bring their own judgment to bear to identify further
exclusions on a trade-by-trade basis. A trier of fact will then have to resolve these
disputes through individualized determinations about the appropriateness of
particular exclusions.
Nor do advances in technology render “the state of mind of each investor at the
time [a] trade was requested” irrelevant to the economic loss determination. See
Newton, 259 F.3d at 187. Consider a trader who places two orders to buy shares of
a stock, one that he cancels before it is executed, and a second that is identical to the
first, but executed at a better price than would have applied to the first order. Even
if he canceled the first trade because of a delay in execution caused by TD
Ameritrade’s order routing practices, whether the cancellation caused economic loss
depends on the trader’s strategy. If he intended the second order to replace the
canceled one, then he is better off than if his first order had been executed. But if he
would have placed the second order even if the first order had been executed, then he
might be worse off, because he will have fewer shares available to sell for a profit if
the price of the stock later goes up. Ford’s algorithm cannot account for each
customer’s trading strategy.
We conclude that despite advances in technology, individual evidence and
inquiry is still required to determine economic loss for each class member. See id. at
-8-
187-88. Advanced computing power can expedite that determination, but it cannot
change its underlying nature by converting individual evidence into common
evidence. In this case, the prevalence of these individualized inquiries precludes
class certification under Rule 23(b)(3).
Another concern with predominance is the nature of the trading conduct at
issue. In Newton, the plaintiffs challenged the execution of orders at the NBBO price
on the ground that their broker failed to investigate whether the orders could have
been executed at more favorable prices on alternative trading systems. See 259 F.3d
at 169-70. In affirming the denial of certification, the court reasoned that “the NBBO
listed price may or may not have provided a class member with the best price,”
depending on the facts of each trade. Id. at 180. The court thus declined to presume
economic loss across the class, and it was unpersuaded that the plaintiff’s proposed
formula could determine economic loss for each class member without an array of
individual inquiries. Id. at 180-81, 187-88.
Ford attempts to distinguish Newton on the ground that TD Ameritrade
executed orders at prices inferior to the NBBO price when the orders could have been
executed at the NBBO price. He argues that this practice is inconsistent with the duty
of best execution. His expert’s analysis of a limited set of Ford’s own trades,
however, revealed that a substantial majority were executed at a price better than or
equal to the NBBO price. The economic loss analysis for these trades does not differ
from Newton. And insofar as Ford suggests that execution of other trades at prices
inferior to the NBBO price necessarily results in economic loss, we disagree. There
are circumstances in which a trade legitimately might execute at a price inferior to the
NBBO price, such as when the order size exceeds the number of shares available at
the NBBO price at the time of the order. As a result, the price that a class member
received on trades executed at prices inferior to the NBBO price “may or may not
have provided a class member with the best price,” depending on the facts of each
-9-
trade. Id. at 180. Assessing the relevant facts of each trade requires individualized
inquiries.
Ford argues that a district court in New York approved use of a similar
algorithm in a best-execution class action after Newton. In In re NYSE Specialists
Sec. Litig., 260 F.R.D. 55 (S.D.N.Y. 2009), the court certified a class in a securities
fraud case and approved the use of an algorithm to prove economic loss. Id. at 80.
The type of securities fraud at issue involved specialists using their knowledge of
impending customer orders to trade ahead of those orders for the benefit of their own
accounts. Id. at 64. To measure the resulting economic harm to the customers, the
finder of fact was required to match customers’ orders with specific trades made by
the specialists. See id. at 66-67. The next step was to compare the outcome that
customers received with the outcome they would have received but for the specialists’
alleged misconduct. See id. at 80.
That an algorithm could perform the limited matching function in NYSE
Specialists and satisfy Rule 23 does not establish that an algorithm can solve the
predominance problem in this case. Indeed, one of Ford’s experts acknowledged that
the algorithm in NYSE Specialists did not have to identify reasonably available prices
for executed trades across all market centers or take into account all of the exclusions
that must be considered here. Measuring economic loss in this case is a more
complex task, and the individual questions preclude a conclusion that common issues
predominate.
The duty of best execution requires that brokers “use reasonable efforts to
maximize the economic benefit to the client in each transaction.” Newton, 259 F.3d
at 173 (internal quotation omitted). The duty regulates a broker’s process of routing
orders for execution, but does not guarantee a specific outcome. As Ford’s expert
acknowledged, compliance with the duty of best execution does not guarantee that
the customer will get the best deal possible. Nor does a violation of the duty of best
-10-
execution necessarily cause a customer economic loss. As in Newton, “[b]ecause
economic loss cannot be presumed, ascertaining which class members have sustained
injury means individual issues predominate over common ones.” Id. at 190. The
district court therefore abused its discretion in certifying a class under Rule 23(b)(3).
C.
There is an independent problem with the class as defined by the district court:
it is an impermissible “fail-safe class.” The class consists of “[a]ll clients of TD
Ameritrade . . . who placed orders that did not receive best execution, in connection
with which TD Ameritrade received either liquidity rebates or payment for order
flow, and who were thereby damaged.” This definition incorporates two contested
elements of liability—failure to seek best execution and economic loss. By defining
the class to include only those customers who were harmed by TD Ameritrade’s
alleged failure to seek best execution, the district court certified a class in which
membership depends upon having a valid claim on the merits.
Such a class is impermissible because it allows putative class members to seek
a remedy but not be bound by an adverse judgment. Orduno v. Pietrzak, 932 F.3d
710, 716 (8th Cir. 2019). Fail-safe classes are also unmanageable, see Fed. R. Civ.
P. 23(b)(3)(D), “because the court cannot know to whom notice should be sent.”
Orduno, 932 F.3d at 717. If a fail-safe class is certified as a means of avoiding a
predominance problem under Rule 23(b)(3), “its independent shortcomings are an
alternative basis” to reverse class certification. Id.
* * *
For the foregoing reasons, we reverse the district court’s order certifying a
class and remand for further proceedings.
______________________________
-11-