THE UNCERTAINTY OF LEGAL RIGHTS by Steven Scott Stephens. New York: Routledge, 2001. 114pp. Cloth $50.00.
ISBN: 0-8153-3764-7.
Reviewed by James R. Rogers, Department of Political Science, Texas A&M University.
This slim volume brings a novel data set to its attempt to quantify the uncertainty that underlies litigation.
The book addresses three interrelated hypotheses. The first, and substantively the narrowest, considers how well
capital markets are able to predict the trial outcomes of civil suits between two firms. Drawing on these results,
Stephens then addresses two additional hypotheses. First, Stephens broadly applies his results to draw conclusions
regarding the overall predictability, or lack thereof, of legal proceedings--hence the title, THE UNCERTAINTY OF
LEGAL RIGHTS. Secondly, Stephens applies his results empirically to validate one of two rival theories of why litigants
fail to settle their lawsuits out of court. We consider each of these hypotheses in turn.
The measured impact of lawsuits on corporate finance provides the empirical basis used to test the book’s broader
hypotheses, so we must start here even though the substantive interest of this hypothesis–at least for the average
LPBR reader–may be less than the others.
Stephens builds on a paper by Sanjai Bhagat, James Brickley, and Jeffrey Coles (1994). They used an event study
approach to determine the impact of one company filing a lawsuit against another company on the expected future
earnings of those firms as represented by their stock prices. The theory behind the approach is this. The efficient
markets hypothesis provides that the price of assets in capital markets reflect all of the information that agents
hold regarding the value of that asset. Put another way, the price of a financial asset aggregates all available
information, whether that information is public or private. The intuition for the efficient markets hypothesis
is that the market itself provides an incentive for participants to reveal any private information they have.
Assume that an individual held private information not reflected in the price of an asset, and assume that the
information is positive news about the firm’s future value. The firm would be undervalued relative to the value
suggested by the private information held by that one individual. An arbitrage possibility would therefore exist
for that individual, and he would have the incentive to purchase as much of the undervalued stock as he could.
However, the very act of purchasing the stock would bid up its price to
the point that the individual would be indifferent between buying one more share or not. But the stock’s new higher
price now PUBLICLY reflects the higher expected earnings of the asset. Hence, the price mechanism provided the
incentive to the individual to reveal private information to all.
One implication of the theory is that the price of a firm’s stock changes only with the revelation of new information
and unexpected events. (The impact of known events on the earnings potential of a firm
Page 294 begins here
would be reflected in the current price of the stock.). However, the empirical world is not so tidy when it comes
to identifying the actual impact of new information on stock prices. Stock prices change all the time, and much
of that change reflects only random noise. The empirical trick used to separate random noise in price changes
from the effect of identifiable events is, first, to estimate the normal magnitude of the random noise. Once that
is identified, then price changes greater than that magnitude can be ascribed (at stipulated levels of confidence)
to the impact of new information reaching the market.
Bhagat, Brickley and Coles used this approach to measure and explain the wealth impact of one company filing suit
against another company. In doing so, however, they did not control for whom ultimately won the case. Stephens’
contribution to this line of research in the finance literature is to separate the ultimate winners from the ultimate
losers, thus developing a rough proxy for lawsuit quality in estimating the impact of filing suit on both corporate
plaintiffs and corporate defendants. Stephens finds, however, that controlling for the ultimate winners changes
the results only very slightly from those found in Bhagat, Brickley and Coles (pp. 72-75).
If Stephens’ study ended here, there would not be much of interest either to finance literature itself or to the
study of judicial politics more broadly. However, perhaps in an academic illustration of “making lemonade from
lemons,” Stephens uses his data and results in an attempt to illuminate questions broader than the wealth impact
of suit filings.
To understand how Stephens applies his results to additional questions, we need to return to the theory underlying
Stephens’ work. Recall that the efficient markets hypothesis implies that only new information pertinent to the
expected future stream of corporate earnings will affect the price of a company’s stock. The empirical
implication also works the OTHER way: when the price of a stock changes, we can deduce that new information pertinent
to expected corporate earnings has been revealed to the market.
In the context of trial outcomes, this means that Stephens can examine variation in the price of stock on the day
that the outcome of a trial was announced (as well as variation in the next several days) and deduce whether the
market had anticipated the outcome of the trial. Given that the efficient markets hypothesis states that markets
react only to new information, and given that evidence relevant to the trial outcome is revealed during the trial
itself BEFORE the announcement of
the outcome, then, if trial outcomes are predictable, there should be no significant change in the price of the
firm’s stock when the trial outcome is announced.
Stephens finds that markets do not fully anticipate the trial outcome. As Stephens writes, “The most notable finding
of this research is the significant differences between abnormal returns to winners and losers at the time trial
level resolutions are announced. This strongly suggests that the markets did not fully anticipate the results
of the suits, even after all the privately held information that could have affected the result was made public”
(p. 87). Stephens finds that the price of stock increases on average 2.475 percent for the winners of lawsuits
and declines 2.200 percent for the losers of lawsuits.
Because it touches on the transparency and predictability of the legal
Page 295 begins here
process, Stephens claims that this result should bother us at least a little bit. There are, he argues, at least
three sources of ex ante uncertainty related to a trial. The first source of uncertainty is the evidence that
the parties will introduce at the trial. The second source of uncertainty is the factual conclusions that the
fact-finder
will draw from the evidence. The final source of uncertainty is the legal judgments that the judge will make.
Stephens’ results, he claims, demonstrate that “even after the first [source of uncertainty] is eliminated, the
remaining two sources can still account for a considerable degree of uncertainty” (p. 91). One attribute of a
well-working legal system for Stephens is that “legal rights and obligations are clearly defined” (p. xi), meaning
not only that there are clearly defined in the abstract, but that there is predictability regarding their enforcement.
His results, he claims, thus raise concerns about the transparency of legal proceedings.
I am, however, skeptical of this broader claim. To be sure, if all disputes went to trial I would agree that Stephens’s
worries were well merited. But Stephens did not study the set of all disputes, he studied the set of disputes
that went to and were resolved in a trial. Presumably, the disputes in which an ultimate winner is relatively
clear do not go to trial (or at least are often resolved prior to going to trial). If these disputes DID go to
trial, their resolution would likely be as clear to the judge and jury as they were to the parties themselves.
It is the clear expectation of a given outcome, after all, that provides one inducement for the parties to settle
prior to a trial: why assume the cost of a trial if the outcome is known beforehand and those costs can be avoided
by settling the case without a trial.
Presumably, then, it is the unclear cases–cases without clear legal precedent or cases with ambiguous and competing
factual claims–that are the ones that are most often litigated. Given the conditions under which actual trials
are likely to occur, it doesn’t strike me as particularly surprising or concerning that trial outcomes are as much
a surprise to onlookers as they are to the participants themselves. This possibility is particularly notable given
that the settlement literature is least developed in considering negotiations that continue after a trial has begun
(Dougherty 2000, 102).
Finally, Stephens also applies his results to test the implications of the “divergent expectations” and the “asymmetric
information” theories of settlement by drawing on empirical implications of the respective theories developed in
a 1998 NBER working paper by Joel Waldfogel. Waldfogel argues that the divergent expectations theory predicts
plaintiff win rates of around 50 percent, while the asymmetric information theory predicts that the better informed
party wins a disproportionate amount of the time (having the better assessment of the case outcome to begin with).
Stephens argues that “the rough equivalence of the magnitudes of the reaction in either direction (+2.475% for
winners, and –2.200% for losers) is consistent with markets expecting, on average, a 50% probability of either
side prevailing. Of course, as Stephens himself recognizes, “If it is equally likely that a plaintiff or defendant
will be better informed than the opponent in a given case, then even if the [asymmetric information] theory is
correct, the plaintiff win rate will
still approach 50%” (p. 88).
Page 296 begins here
Stephens’ evidence for the divergent expectation’s theory and against the asymmetric information theory thus boils
down to the claim that there cannot be any asymmetric information left by the end of a trial (because it has all
been revealed during the trial itself) so, therefore, the asymmetric information theory should predict that the
parties should always settle between themselves prior to the announcement of a decision. However, this also is
something of a weak reed given that, presumably, avoidance of court and transaction costs is a major inducement
to settle prior to going to trial. By the end of a trial, but before the judgment, the costs have been almost
fully assumed by the respective parties. Hence, there cannot be much to gain by settling at that point beyond
attempting to insure against the
residual uncertainty of judge and jury. So I am unpersuaded that Stephens’ study helps very much to distinguish
between the plausibility of the divergent expectations theory and the asymmetric information theory of litigation.
Overall, Stephens does provide evidence that the announcement of a case decision affects capital markets. The
magnitude of that effect is of some intrinsic
interest in itself, even if it does little more than confirm results of earlier studies. However, where Stephens
attempts to draw broader implications from his results–identifying random elements inherent in the legal process
and empirically testing the distinction between the divergent expectations and asymmetric information theories
of settlement–his claims did not ultimately persuade this reviewer.
REFERENCES:
Bhagat, Sanjai, James A. Brickley, and Jeffrey L. Coles. 1994. “The costs of inefficient bargaining and financial
distress.” JOURNAL OF FINANCIAL ECONOMICS 35: 221-47.
Doughety, Andrew F. 2000. “Settlement.” In ENCYCLOPEDIA OF LAW AND ECONOMICS, Vol V., eds. B. Bouchaert and G.
DeGeest. Cheltenham.
Edward Elgar Publishing Co.
Waldfogel, Joel. 1998. “Reconciling Asymmetric Information and Divergent Expectation Hypotheses.” NBER Working
Paper No. W6409.
**************************************************************************
Copyright 2002 by the author, James R. Rogers.