Access to justice in a world without lawyers: evidence from Texas bodily injury claims.

• Author: Silver, Charles
• Position: 2008 ABA Section of Litigation Access to Justice Symposium

Introduction I. Data Description II. Results A. Descriptive Statistics B. Multivariate Analysis 1. Changes in the Rate of Self-Representation over Time 2. Claim-level Factors a. Claim Size b. Claim Complexity c. Cost of Legal Services d. Regression Analysis III. Discussion A. Implications of Our Findings B. Time Trends C. Effect of 1995 Tort Reforms D. The Proof is in the Pudding E. Experience in Other States F. Wealth and Income and the Decision to Hire Counsel G. Limitations of Our Study Conclusion INTRODUCTION

Price controls are unpopular, except when they are imposed on plaintiffs' lawyers. Sixteen states have caps on contingency fees; (1) none cap defense-side fees. (2) The federal government limits the fees plaintiffs can pay in Federal Tort Claims Act ("FTCA") cases, (3) Social Security disability cases, (4) and claims for veterans' benefits. (5) In 2003, 2004, and 2005, the House of Representatives passed a bill capping contingency fees in medical malpractice cases--only to see it die in the Senate each time. (6) Judges have followed suit; in recent aggregate products liability proceedings, three federal judges capped the plaintiffs' lawyers' fees after noting a "trend in the states to limit contingent fees to 33-1/3% or less of net recovery." (7)

Claimants are the supposed beneficiaries of these restrictions, which are said to protect them from excessive legal fees. (8) The assertion is facially plausible: lawyers' fees are the biggest component of litigation costs, so policies that reduce fees may help claimants. Yet, the pressure to cap contingent fees comes from tort reform groups, representing drug manufacturers, medical providers, liability insurers, and other repeat players on the defense side in litigation. Like all interest groups, tort reform groups advocate policies that help their supporters. (9) Because tort reform groups have defendants' interests at heart, it is safe to assume that these groups expect caps on contingency fees to help defendants--most likely because price controls tend to cause the supply and quality of services to decline. (10) When lawsuits become less profitable for lawyers, plaintiffs' attorneys will predictably become more selective and may even move into other uncapped lines of work, or retire. Because expected marginal earnings also decline, lawyers will predictably invest less in the cases they do accept-that is, they will litigate less intensively.

Will plaintiffs' ability to recover for their injuries also decline? Although most academics believe "meaningful access" to the tort system requires legal counsel, (11) claimants can always represent themselves. (12) Many tort claims settle without formal litigation, and nearly all settle without trials. (13) Some studies contend that claimants who represent themselves don't do that much worse than claimants represented by attorneys once attorneys' fees are subtracted. (14) If self-representation is a viable option, caps on contingency fees may have less of an impact on access than one might otherwise expect. (15) Conversely, if tort recoveries flow only or mostly to plaintiffs who either actually hire lawyers or can credibly threaten to do so, caps on contingency fees may make it difficult or impossible for many victims to obtain justice.

We explore this issue by looking at patterns of representation among claimants who received payments for bodily injuries ("BI") in Texas during 1988-2005. Our database includes every closed claim with a payout greater than or equal to $10,000 (nominal) from five commercial lines of insurance. Approximately 7% of successful BI claimants represented themselves. Paid claims brought by persons who represented themselves had smaller payouts and were almost always resolved without litigation. Self-representation was roughly four times as common (12% versus 3.3%) in small claims (less than $10,000) as compared to large claims (greater than $25,000) (both amounts in 2008 dollars). The filing of a lawsuit marked a fundamental divide: almost 16% of claims resolved without filing a lawsuit involved self-represented claimants, compared with a mere 0.2% of claims resolved after a lawsuit was filed.

This Article proceeds as follows: Part II describes our data, Part III presents our results, Part IV discusses the implications of our findings, and Part V concludes. (16)

1. DATA DESCRIPTION

   We have published a series of articles on tort litigation (17) using the Texas Closed Claim Database ("TCCD"), which is maintained by the Texas Department of Insurance ("TDI"). The TCCD, which we describe more fully in other works, contains individual reports of all Texas BI claims covered by five lines of commercial insurance (auto, general commercial, multi-peril, medical professional liability, and other professional liability) that closed with payments exceeding $10,000 (nominal) from 1988 to 2005. (18)

   Because the TCCD uses a fixed nominal dollar reporting threshold, the dataset is subject to bracket creep. Over time, the lower end of the bracket creeps downward, encompassing smaller and smaller claims. Thus, a claim worth $10,000 (nominal) in 2005 would have been worth only $6,057 in 1988. More generally, a claim worth $10,000 in 2005 would have been worth less than $10,000 in any year prior to 2005, and thus would not have been reported to TDI. Bracket creep has the potential to make time trends in the TCCD misleading.

   In other publications, we address the problem of bracket creep by adjusting payments to 1988 dollars and excluding claims with real payments below $10,000. (19) This approach has an important drawback here. Claims with real payments below $10,000 are common, and the frequency of self-representation is higher in claims with smaller payouts. Indeed, the highest rate of self-representation occurs in the bracket creep claims. To preserve information about self-representation, we therefore include bracket creep claims in our summary tables. When studying time trends, however, we adjust for bracket creep to avoid inflation-generated distortions.

   When submitting closed claim information to the TDI, insurers use "short forms" for claims with nominal payouts greater than $10,000 but less than $25,000, and "long forms" for claims with nominal payouts of $25,000 or more. (20) Although both forms indicate whether the claimant hired an attorney, long forms contain much more information than short forms. (21) Many fields of potential interest--age, employment status, injury type, and injury cause--appear only on the long form, (22) and neither form has all the information needed to fully study the choice between self-representation and representation by counsel. (23)

   The TCCD contains only insured claims. (24) Rates of self-representation may differ in uninsured cases, which lawyers are likely to find less appealing. The TCCD also contains only claims covered by commercial or professional lines of insurance. (25) Cases involving personal lines of insurance, for example, personal auto, homeowners, or watercraft policies, may have a different pattern of self-representation as well. Unfortunately, we lack a source of data with which to explore these issues.

2. RESULTS

   1. Descriptive Statistics

      Table 1 presents descriptive statistics on all claims in the database. Commercial auto claims are the most common, the least likely to involve death as the injury alleged, and have the smallest mean and median payouts. Medical malpractice claims are much less numerous than commercial auto claims, but have the highest frequency of death, as well as the highest mean and median payouts.

      Table 2 provides further descriptive statistics on whether the plaintiff was represented by a lawyer and the resolution stage (pre-suit versus post-suit) of each claim. Table 2 indicates that across all lines of coverage 6.9% of claimants represented themselves, ranging from 2.5% (medical professional) to 7.9% (commercial auto). In cases where no suit was filed, the rate of self-representation was far higher, ranging from 13.4% (commercial auto) to 24% (multi-peril). In claims where a lawsuit was filed, however, the rate of self-representation was essentially zero across all lines of coverage. For unrepresented claimants, there is an unbridgeable chasm between pre-suit and post-suit resolution. The probability of suit varied widely, however, with commercial auto accounting for the lowest rate (42%) and medical professional having the highest (89.5%). (27)

      Table 3 provides descriptive statistics on the relationship between self-representation and payout. As it shows, across all lines of coverage the frequency of self-representation declines steadily as the payout increases. Coverage line still matters, however. Medical malpractice claimants retained lawyers more often than auto claimants at all payment levels except the very smallest.

   2. Multivariate Analysis

      1. Changes in the Rate of Self-Representation over Time

         We used multivariate analysis to study the factors that influence self-representation and to determine whether the frequency of self-representation changed over time. Because the dependent variable, "lawyer/no lawyer," has only two possible values (1 when a claimant had counsel; 0 otherwise), we used logit regressions. (30) To avoid reporting time trends influenced by inflation, we excluded all "bracket creep" claims with real payments below $10,000 (1988$). For independent variables, we used (year-1988) and a dummy for whether the claim was filed after Texas adopted tort reform in 1995. (31) We report the results for each line of coverage and for the entire dataset. The results are shown in Table 4.

         In Table 4, a confidence interval with only positive (negative) values indicates a statistically significant higher (lower) probability of representation by counsel as the value of an independent variable increases. Conversely, a confidence interval that spans zero indicates that there is no statistically significant...