UNITED STATES OF AMERICA v. H&R BLOCK, INC. et al

Filing 111

REDACTED DOCUMENT- Declaration of Dr. Warren-Boulton to 76 Notice (Other) by UNITED STATES OF AMERICA. (Buterman, Lawrence)

UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF COLUMBIA

UNITED STATES OF AMERICA,

Plaintiff,

v.

Civil Action No. 11-00948 (BAH)
Judge Beryl A. Howell

H&R BLOCK, INC.;
2SS HOLDINGS, INC.; and
TA IXL.P.,
Defendants.

REDACTED VERSION
FOR PUBLIC FILING

DECLARATION OF FREDERICK R. WARREN-BOULTON
SUMMARIZING EXPECTED DIRECT TESTIMONY
INTRODUCTION

I. I have been a practicing economist for more than 40 years and much of my professional
experience has involved the analysis of proposed mergers and the study of the appropriate
methods for such analysis. I previously served as the DOJ Antitrust Division's chief economist
and in that capacity I supervised approximately 40 economists in the review of numerous
mergers and other competition issues.
2.

I was retained by the Antitrust Division to provide my opinion concerning the likely

competitive effects of the transaction at issue in this case. I have concluded that the relevant
market to analyze this transaction is the market for digital do-it yourself tax preparation products
(digital DIY) and that the transaction is likely to result in significant harm to consumers arising
from higher prices and reduced non-price competition. My conclusions are based on work that I
did, or that was undertaken at my direction, including: (a) a review of documents and testimony
conclusively indicating that the parties view each other as direct competitors and undertake
significant competitive responses to each other's actions; (b) an analysis of data regarding

pricing and switching (including data produced by the parties and the IRS) to determine
diversion ratios; (c) the application of the hypothetical monopolist test to define the relevant
market; and (d) the use of a merger simulation model to predict the unilateral price increases
expected from the merger. These are standard methods applied by economists in merger analysis.
3. I have also considered the report and opinions ofthe Defendants' expert, Dr. Christine
Meyer, and concluded that nothing in her report or opinions provides a basis to revise my own
opinions. To the contrary, Dr Meyer's critical conclusions appear to rest largely not on the
application of standard economic models or analysis, but on the misuse or misunderstanding of
surveys and other work conducted by the parties that itself appears to be flawed.
4. My analysis is confirmatory of what common sense tells us about this transaction. Using
digital DIY products to do your taxes is very different than using pen-and-paper or paying
someone else to do your return. That is why a robust market for digital DIY products has
emerged. That market is now dominated by three competitors, Intuit, H&R Block (HRB) and
Second Story Software (2SS), who collectively have a nearly 90% share. For many years, 2SS
has been a maverick competitor, pricing substantially below HRB and Intuit and introducing
innovations such as "free" products that have forced competitive responses from HRB and Intuit.
One such significant 2SS innovation allows all taxpayers to prepare and e-file their federal tax
returns for free. 2SS does this with the expectation that many consumers will purchase add-on
features, such as a corresponding state tax return, and some will, in successive years, switch to a
"paid" TaxACT product.
5. H&R Block, which makes HRB at Home, and Intuit, which makes TurboTax, were forced to
respond to 2SS' s business model by redesigning their product lineups to include their own free
federal tax return products, though with fewer tax forms than TaxACT. HRB and Intuit seek to

2

generate revenue from customers who start with their free products by selling similar add-ons
and converting customers to paid products. However, for HRB and Intuit, these free federal
products "cannibalize" their paid products to a much greater degree than for 2SS. In other words,
a significant number of customers who would have purchased HRB and Intuit paid products have
moved to free federal products and only upgrade as needed. Moreover, new customers coming
into the market who would have selected paid products also often start (and sometimes finish)
their returns with free products. 2SS does not have significant cannibalization concerns because,
among other reasons, while its menu of products offers similar functionality to HRB' sand
Intuit's, 2SS earns a smaller dollar margin on its highest priced products.
6. Since 2SS made its "Free-For-All" offer to all taxpayers, free federal tax preparation has
become the most popular product in the digital DIY market. At the same time, the average
inflation-adjusted price paid by digital DIY customers has declined 10%.
7. Defendants argue that, because the "free" marketing message is an important customer
acquisition tool, they have little or no incentive post-merger to eliminate free federal products or
otherwise raise prices or reduce quality. The question, however, is not whether HRB or Intuit
will stop offering and marketing a "free" product (i.e., free federal), but how the merger will alter
the marketing and pricing of2SS and HRB's products including other components of the bundle
purchased by customers who buy the "free" (federal) product. After the acquisition, HRB will
have an incentive to charge higher prices for products across their product lineup, and, HRB will
have an incentive to shift customers to its higher priced paid products by reducing the value
proposition provided by 2SS.

3

8. There is a substantial likelihood of consumer harm when a merger eliminates a maverick
competitor, leaving the market in the hands of an effective duopoly of competitors who had been
charging higher prices than the maverick.
RELEVANT PRODUCT MARKET
9. The relevant product market for this case is digital DIY tax preparation products, which
includes products from HRB, Intuit, and 2SS, as well as products from several fringe firms. The
hypothetical monopolist test set out in the Horizontal Merger Guidelines, and widely used by
economists, substantiates this market definition. In my opinion, a monopolist of digital DIY tax
preparation products would increase prices significantly, notwitstanding switching by some
customers to other forms oftax preparation. Under the test, the hypothetical monopolist need not
control every substitute product, only enough of them that it would profitably impose a
significant increase in price on at least one product sold by one of the defendants.
10. In brief, my conclusion as to market definition is supported by substantial evidence showing
that:
10.1. Defendants viewed each other as significant competitors.
10.2. Intuit and HRB have reacted to competitive initiatives by 2SS.
10.3. Consumers perceive TaxACT as similar to

HRB.

My analysis reveals that on key consumer drivers
identified by HRB, consumers seem to view TaxACT and HRB as close competitors.
10.4. Neither pen-and-paper nor assisted would prevent a hypothetical monopolist of digital
DIY from raising price substantially, a conclusion supported by testimony to this effect by

4

former and current executives ofHRB, Mark Ernst, former Chairman and CEO ofHRB, and
HRB's Jason Houseworth.
10.5. Defendants market to the same customers, as, Jason Houseworth, HRB's Head of Digital
Tax Solutions, admitted as much in his deposition. Targeted marketing efforts by 2SS have
impacted the sales ofHRB and Intuit.
10.6. While defendants point to a survey conducted for litigation and to a price simulator to
indicate a broader relevant market, the survey and price simulator contradict their documents and
the testimony of their executives, and I find neither credible. Professor Ravi Dhar has evaluated
the April 2011 survey and concluded it was "severely flawed" and "fails to meet the basic
premises of good survey design." I find his reasoning persuasive. The price simulator produces
an estimate of diversion from HRB products to TaxACT, other digital DIY products and other
methods oftax preparation, but the pricing simulator results show significant violations of basic
principles of economic theory. Even if this price simulator_were found reliable, and I do not
believe it is, the pricing simulator output relied on by Dr. Meyer cannot be used to inform the
diversion ratio most important to this case: the diversion from TaxACT to HRB products. With
respect to the diversion ratio from HRB to TaxACT, Dr. Meyer's analysis relies entirely on a
single comparison of two groups of pricing scenarios, but I find that comparisons using
alternative scenarios give wildly different results. Defendants' documents also contain other
surveys that suggest substantially different results.
HARM TO COMPETITION
11. The proposed acquisition will combine the second and third largest digital DIY firms,
resulting in two firms together controlling 90% of the market.

5

12. After analyzing both the potential unilateral and coordinated effects, I have concluded that
the proposed acquisition likely will cause anti competitive harm. Unilateral effects occur as a
result of the combined firm acting without cooperation from its competitors; coordinated effects
occur as a result of the combined firm coordinating its actions with rivals, here, Intuit.
Coordinated effects do not require explicit or even implicit agreement among firms. Dr. Meyer
argues that unilateral and coordinated effects would be inconsistent with HRB' s claimed strategy
to continue to offer TaxACT as a free product. I do not predict that this merger will necessarily
cause the merged firm to cease offering free products. I conclude that this merger gives HRB the
incentive and the ability to compete less aggressively than TaxACT would have with respect to
price or quality.
COMPETITIVE HARM FROM UNILATERAL EFFECTS
13. The likelihood and magnitude of unilateral price effects depends to a large extent on the
degree of direct competition between defendants. When considering a price increase, defendants
take into account what they would lose to their competitors as a result and compare it to what
they would gain from those customers who will pay the higher price. Today, the loss for each
defendant would include business the other would capture. If the proposed acquisition takes
place, this would change. A price increase on a 2SS product would lead some customers to use
an HRB product, and HRB would expect to generate significant incremental profit per such
customer, especially since the merged firm will earn more per customer on the HRB product than
on the 2SS product.
14. A valuable measure of the degree of direct competition between firms is the diversion ratio,
which measures the portion of lost customers that would go to a specific firm as a result of a
price increase. Substantial evidence supports the proposition that significant diversion exists

6

between the defendants. Documents show that defendants consider each other competitors: they
track each other's market share, prices, advertising, web traffic, and product quality; and they
respond to each other in ways that benefit consumers. Empirical evidence also shows that
defendants take business from each other by competing harder, (e.g., through increased
marketing). Documents demonstrate that defendants consider one another's pricing when they
make pricing moves.
15. Merger simulations are a well-accepted tool for putting some structure on this analysis. They
provide a way to examine the interaction of diversion ratios, margins, price, and quantities. The
merger simulations I have run indicate that in the absence of efficiencies TaxACT prices would
increase between 8-15% while HRB prices would increase between 2-4%, and cause harm to
consumers in the many millions of dollars annually.
16. Dr. Meyer's criticisms of the merger simulations that I have run are not compelling for the
following reasons:
16.1. By their nature, economic models make simplifYing assumptions. The question is not
whether the model captures every detail of the real world, but whether it reflects the essence of
competition. Dr. Meyer correctly identifies ways in which my merger simulation is simpler than
the real world but fails to demonstrate that the simplifYing assumptions undermine the validity of
the results.
16.2. Switching data reflect who filed in one year using one method and then switched the next
year to another method. Switching rates and diversion rates measure different things (the former
measures switching regardless of cause, while the latter seeks to measure it as a result of a price
increase), but that does not mean that the former cannot serve as a proxy for the latter. In this
case, evidence that switching to assisted tax preparation overstates diversion indicates that any

7

bias in the merger simulation when rnn using switching rates as a proxy for diversion rates may
favor defendants.

HRB

documents imply that more than half of switching from

digital DIY to assisted preparation occurs for reasons nnrelated to price, such as a change in tax
complexity. Data provided by the IRS suggests that taxpayers that switched from a digital DIY
product to assisted preparation are about twice as likely to have an observable increase in tax
complexity as taxpayers that stayed within the digital DIY market.
16.3. Moreover, I did not derive my estimated diversion ratios only from switching rates. I
derived similar diversion ratios based on a methodology that HRB has used in the course of its
business to analyze the effects of customer switching, which relied on market share as the basis
for diversion within the digital DIY market. Diversion proportional to market share assumes that
market shares, which reflect consumers' first choices, also reflect consumers' second choices.
Finally, diversion rates much lower than those estimated from market shares and IRS data would
be inconsistent with the wealth of documentary evidence that illustrates that HRB and 2SS view
each other as significant competitors.
COMPETITIVE HARM FROM COORDINATED EFFECTS
17. The evidence shows that HRB and Intuit share similar incentives to structure their product
lineups and prices in a way that preserves the profitability of their high priced products. 2SS's
incentives are different, as made clear from their disruptive conduct over the years, because they
do not share HRB's and Intuit's concern about cannibalizing higher priced products. 2SS's
different incentives make it the market "maverick" and have forced HRB and Intuit to restructure
their digital DIY product lineups in a way that makes it more difficult for them to sell their
higher margin products. As the Merger Guidelines explain, "An acquisition eliminating a

8

maverick firm in a market vulnerable to coordinated conduct is likely to cause adverse
coordinated effects."
18. This merger would eliminate this stark contrast in incentives among 2SS, HRB, and Intuit,
thus increasing the likelihood that HRB and Intuit could reach a common understanding about
the best way to structure their product lineups to better "up-sell" customers. Customers would
lose the competition-enhancing benefits that come from having such an effective maverick in
this market.
ENTRY, EXPANSION, AND REPOSITIONING
19. Obstacles faced by new entrants and small rivals make it unlikely that they would deter or
counteract a unilateral or coordinated anti competitive effect. Trust and reputation are important
drivers for consumer demand in this industry, and customers need convincing before large
numbers can be expected to rely on a product - no matter how much excess production capacity
a vendor possesses. While Dr. Meyer points to the FFA as a way to enter and expand, she fails
to point out that the FFA's role as a sales channel has decreased substantially over the past
several years.
EFFICIENCIES
20. Defendants' claimed efficiencies do not appear, by and large, to be verifiable or mergerspecific, for reasons set forth in the expert report of Dr. Mark Zmijewski. Two additional points
merit mention here: The claimed fixed cost savings from efficiencies do not appear to fit within
any accepted theory under which fixed cost savings would counter the threat of anticompetive
harm from an acquisition. Second, even if! accepted all variable efficiencies that Dr. Meyer
credits, which I do not, I would still find significant unilateral harm likely. Incorporating her

9

efficiencies into the merger simulation still results in likely harm to consumers of millions of
dollars annually.

Frederick R. Warren-Boulton
September 1,2011

10