O'Connell Early Settlement Offers: Toward Realistic Numbers and Two‐Sided Offers

• Date: 01 June 2010
• DOI: http://doi.org/10.1111/j.1740-1461.2010.01182.x
• Published date: 01 June 2010

O’Connell Early Settlement Offers: Toward

Realistic Numbers and Two-Sided Offersjels_1182379..401

Bernard Black, David A. Hyman, and Charles Silver*

In a prior article in this journal, we estimated the effect of an “O’Connell” early settlement

offer program on payouts in medical malpractice litigation. Using Texas data and a base set

of assumptions, we predicted that early offers would result in a 16 percent (20 percent)

decline in payouts in currently tried (settled) cases. The overall decline came almost entirely

from a sharp drop in payouts in cases with small economic damages. We compared our results

with the estimate by Hersch, O’Connell, and Viscusi (2007) (HOV) of a 70 percent reduction

in payouts, reconciled the two estimates, and explained why HOV’s estimate reflected the

compound effects of a series of unreasonable assumptions. In a reply in this journal, Hersch,

O’Connell, and Viscusi (2010) (HOV-2) complain that we misunderstand both the early offer

proposal and their analysis. Remarkably, they do not dispute our estimates, given our

assumptions. In this rebuttal, we defend our assumptions and provide an alternate analysis of

settled cases based on insurer allocations, which also produces an estimated 20 percent

payout decline. We also develop further our proposal for two-sided early offers. This proposal

would reduce the predicted payout decline by about 2 percent in both tried and settled cases.

We explain in greater detail the problems with HOV’s analysis. If we correct an error they

made in understanding the Texas data set, and leave their other assumptions unchanged,

their payout decline estimate drops to 30 percent, not far from ours.

I. Introduction

In a prior article in this journal (hereinafter, BHS),1we analyzed an early settlement

proposal that Professor Jeffrey O’Connell has been promoting since 1982.2Under his

proposal, defendants can offer to settle for full economic damages plus attorney fees,

possibly with a minimum damages offer in some cases. Plaintiffs who refuse the offer

*Address correspondence to Bernard Black; email: bblack@northwestern.edu; from the fall of 2010, Black will be

Nicholas J. Chabraja Professor at Northwestern University, Law School and Kellogg School of Management. Black is

currently Hayden W. Head Regents Chair for Faculty Excellence, University of Texas School of Laws, and Professor

of Finance, University of Texas, Red McCombs School of Business. Hyman is Richard and Marie Corman Professor

of Law and Professor of Medicine, University of Illinois; Silver is McDonald Endowed Chair in Civil Procedure,

University of Texas Law School.

We thank Hyun Kim for superb research assistance and Vicky Knox of the Texas Department of Insurance for her

answers to our many questions about the TDI database.

1Black et al. (2009). <http://www3.interscience.wiley.com/cgi-bin/fulltext/123214961/PDFSTART>.

2O’Connell (1982).

Journal of Empirical Legal Studies

Volume 7, Issue 2, 379–401, June 2010

379

face a large stick in the form of a heightened burden of proof—they must prove gross

negligence beyond a reasonable doubt, making the offer one “that can’t be refused.”3Our

article responded to an empirical study by O’Connell and co-authors Joni Hersch and W.

Kip Viscusi (hereinafter, HOV), where they estimated that an early offer program would

reduce payouts and defense costs by roughly 70 percent.4

In BHS, we estimated that, under a base set of assumptions and using basically the same

data set as HOV, an early offer program would produce a 16 percent (20 percent) overall

decline in payouts in currently tried (settled) cases. We also provided sensitivity analyses

showing how this estimate would vary depending on program details and on changes in our

assumptions. Under no plausible set of assumptions was the estimated payout decline even

close to HOV’s estimate. We showed: (1) that the overall payout decline would come mostly

from large drops in payouts in cases with “small” economic damages of $0–$100,000; and (2)

elderly and deceased plaintiffs would experience large payout declines, with much more

limited effects on babies, children, and employed adults in nondeath cases.5

Why do our results differ so greatly from HOV’s? In BHS, we concluded that HOV’s

payout decline estimates reflect the compound effects of a series of unreasonable assump-

tions, including: (1) only one-third of paid damages are economic; (2) no adjustment of

payouts for the time value of money; (3) current payouts include full payment of economic

damages; (4) a fee of 10 percent of economic damages (9 percent of gross payout) is a

market-clearing price for the services of plaintiffs’ attorneys; (5) plaintiffs incur no out-of-

pocket litigation costs; and (6) no minimum damages offer in many cases.

In a reply in this journal, Hersch, O’Connell, and Viscusi (hereinafter, HOV-2)

challenge some of our assumptions and complain that we misunderstood their proposal

and “misrepresent[] fundamental aspects of [their] analysis.”6Remarkably, they do not

dispute our estimates given our assumptions. The central question is, therefore: Whose

assumptions are more reasonable—ours or theirs? In this rebuttal, we defend our assump-

tions and analysis, provide additional sensitivity tests, and develop an alternate analysis of

early offers in settled cases, based on insurer allocations of damages. Our alternate analysis,

which also produces an estimated 20 percent payout decline in settled cases, responds to

HOV’s concern with whether one can reliably extrapolate from tried cases to settled cases.

Extrapolating from tried cases and relying on insurer allocations have different strengths

and weaknesses; we consider both approaches to be reasonable. As it turns out, the two

approaches produce similar results.

We also develop further our proposal for two-sided early offers, which HOV dismiss,

and show that it does not have the flaws they attribute to it. Two-sided offers would better

3Hence the title of Professor O’Connell’s initial article, which begins “Offers That Can’t Be Refused.” Id.

4Hersch et al. (2007); see also the similar estimates in O’Connell and Born (2008) and O’Connell and Robinette

(2008).

5All amounts in BHS and this rebuttal are in 1988 dollars; to convert to 2008 dollars, multiply by 1.82.

6Herschet al.(2010:164) (abstract). <http://www3.interscience.wiley.com/cgi-bin/fulltext/123294628/PDFSTART>.

380 Black et al.