An exploration of "noneconomic" damages in civil jury awards.

• Author: Kritzer, Herbert M.
• Position: III. Results B. Bureau of Justice Statistics Data through Summary and Conclusion, with appendices, footnotes, and tables, p. 999-1027 - The Civil Jury as a Political Institution

2. Bureau of Justice Statistics Data

   For the BJS dataset, we again looked at all cases together and then split the cases into three categories: auto accidents, medical malpractice, and other personal injury. Figure 2 shows the plots for all cases and for the three separate categories. Figure 2 shows strong linear relationships for all of the cases taken together and for both auto and other personal injuries; however, the fit for the medical malpractice cases is less clear.

   Table 6 summarizes the regression results for the broken lines shown in Figure 2. (139) Note the last column in the table, which displays the percentage of plaintiff's verdicts for which nonzero values were reported for both economic and noneconomic damages, and hence, are included in the analysis. Both auto accident cases and the other personal injury cases produced good fits with positive regression coefficients, indicating that noneconomic damages increased in a linear fashion as economic damages increased. The exception is the small subset of medical malpractice cases, which also demonstrated a low [r.sup.2] in the Cook County data discussed above. One difference is that these data include states that have imposed caps on noneconomic damages in medical malpractice cases, (140) and it may be that the weak relationship in such cases reflects in part the presence of caps in some states. Hence, the table also shows separate regressions for those cases in which a cap did and did not apply; the number of cases is quite small, but it is clear that there is little difference between the two subsets of cases. (141)

   Table 7 reports the statistics concerning the noneconomic to economic damage ratios for all of the cases and for the three subsets of cases broken down by the amount of economic damages. Here, we use the same categories used by Eisenberg and his coauthors in their analysis of the punitive to compensatory damage ratios in the BJS dataset. (142) The best summary figures to look at are the medians because a small number of extreme values can substantially inflate the means. Overall, there was a fairly consistent pattern in the median ratios of the noneconomic to economic damages: the medians declined as the amount of economic damages increased. The one exception is the highest economic damage category for auto accident cases; however, it should be noted that there are only eight observations in this category.

   Looking at the cases overall, the median was just over 1 (1.19), indicating that in the median case the amount of noneconomic damages was about 20% more than the economic damages. This is consistent with the overall regression shown in Table 6, which showed the overall regression coefficient as close to 1 (.904). One noteworthy difference between what Table 7 shows and what the ratio between punitive and compensatory damages in Table 1 shows is that the standard deviation in the noneconomic to economic damage ratio tended to stay high--up to $100,000 in compensatory damages--while the standard deviation in the punitive to compensatory ratio dropped when the compensatory damages reached $10,000. However, this result is likely generated by a very small number of cases. We say this because an alternate measure of variation, the Interquartile Range (IQR) which is the difference between the first and third quartiles (both of which are shown in Table 3), does drop sharply by the time economic damages reach $10,000 when we look at all cases together or at auto accident cases; the IQR does not drop until economic damages reach $100,000 for medical malpractice and other personal injury cases.

3. RAND Jury Verdict Study Data

   The number of cases and the coding detail for the RAND data allowed us to split those data into five categories: auto (including common carrier), premises liability, medical malpractice, product liability, (143) and other personal injuries. Figure 3 shows the relationship between noneconomic and economic damages (both logged); in this figure we omit the points because they obscure the lines. As with the previous figures, the solid line is fit using the LOWESS procedure and the broken line is the simple regression line. A clear pattern of linear increase appears for all of the case subsets, although the LOWESS lines suggest some deviations for premises liability cases and products liability cases. Table 8 shows the regression results for the various subsets. The strongest relationship, both in terms of the [r.sup.2] and the slope, is for auto accident cases, although it is also the type of case for which the proportion of cases with information on both economic and noneconomic damages was the lowest.

   Interestingly, both the [r.sup.2] and the slope for auto cases in the RAND data are very similar to what we reported above for the Cook County and the BJS datasets. As with the prior two datasets, the [r.sup.2]s and slopes drop off for the other types of cases. We also grouped the cases other than auto and medical malpractice, and reran the regression to provide a comparison to the "Other Personal Injur/' category in the BJS data. The slope and [r.sup.2]s were .570 and .4039, respectively, which are only slightly lower than the comparable figures from our analysis of the BJS data.

   One question about the results for the medical malpractice cases is what difference California's limits on noneconomic damages in medical malpractice cases make. (144) Table 8 shows separate results for the California medical malpractice cases and medical malpractice from other states; clearly, the relationship between economic and noneconomic damages in medical malpractice cases was weaker in California than in the other states in the study. (145)

   Table 9 shows the noneconomic to economic damage ratios for all case types except for medical malpractice, which appear in Table 10. Table 9 shows that the median and mean ratios tended to decline sharply as the amount of economic damages increased; similarly, the amount of variation in the ratios, as measured either by the standard deviation or the interquartile range, tended to decrease. This was also true in medical malpractice cases, as Table 10 shows. Table 10 also provides further evidence on the impact of the cap on noneconomic damages in California. (146) California's ratios were similar to those of other states in the RAND data until economic damages reach the $100,000 to $999,999 category, at which point they dropped sharply compared to the other states. Another way to see the drop-off in noneconomic damage payments in California is shown in Figure 4, which plots the logarithms of economic and noneconomic damages in California as filled circles and the other states as open squares. The fitted line shown is across all medical malpractice cases. We added a vertical line to divide the figure between cases with less than $100,000 in economic damages and cases with economic damages of $100,000 or more. There is little or no difference in the scatter between California and the other states when the economic damages were less than $100,000. After $100,000, a small proportion of the California cases fall above the line, while for the other states there is a fairly even scatter above and below the line. The figure also has a horizontal line at the $250,000 cap on noneconomic damages; we presume that for these California cases, jury awards would have been reduced to no more than $250,000 by the judge.

   SUMMARY AND CONCLUSION

   We modeled most of our analyses above on Eisenberg's and his colleagues' work regarding the relationship between punitive and compensatory damages. (147) Using three primary data sources plus three supplemental sources discussed in the Appendix, we looked at how well noneconomic damages could be predicted by economic damages and at how the ratio of noneconomic damages to economic damages changed as the magnitude of the economic damages awarded by juries increased. (148) One important caveat regarding our analysis is that we have, with some exceptions, focused our analysis on cases in which the jury awarded explicit, nonzero amounts for both economic and noneconomic damages.

   Using the Cook County data, our study is the first to provide detailed breakdowns of damage awards both for economic damages and noneconomic damages. Although medical expenses and lost income make up a large proportion of economic damages, pain and suffering is the most important type of noneconomic damages. However, readers should note that noneconomic damages also take the form of disability, disfigurement, emotional distress, loss of consortium, loss of normal life, and loss of society.

   In our analysis, we found a mixture of consistent and inconsistent patterns across our various datasets. One fairly consistent pattern was the tendency for the ratio of noneconomic to economic damages to decline as the amount of economic damages increased. Moreover, the variability of the ratio also tended to decline as the amount of economic damages increased. We found less consistency in our simple regression models where we predicted the logarithm of noneconomic damages from the logarithm of economic damages. In all of those models the slopes of the fitted line were positive, but the slopes and the measures of fit ([r.sup.2]) varied from one dataset to another and among types of cases within those datasets with multiple case types. Also, when we had the same type of case across datasets, we found variation in the fit and slope. The latter was most striking for medical malpractice cases where we found a very weak relationship within the Bureau of Justice Statistics and National Center for State Courts 2005 verdict study, and that weak relationship held up even when we added controls for whether a state had a cap on noneconomic damages in medical malpractice cases.

   With two of the datasets we were able to extend our regression models with regard to medical malpractice cases. Using the RAND jury study from 1995-1999...