1 Galanter, & Palay, , Tournament of Lawyers 15. All page references in text are to this volume.Google Scholar

2 The term “law factories” was used as early as 1906, when the largest firms had fewer than 20 lawyers. C. Chamberlayne, “The Soul of the Profession,” 18 Green Bag 397 (1906), quoted in Robert Nelson, Partners with Power 170 (Berkeley: University of California Press, 1988) (“Nelson, Partners”). Berle's comment is from his article “Modern Legal Profession” in 5 Encyclopedia of the Social Sciences 343 (1933), quoted in Galanter & Palay at 17.Google Scholar

By 1933, when Berle was writing, the largest New York firm had 70 attorneys; the largest Chicago firm in 1935 had 43 attorneys. Galanter & Palay at 17; Nelson, Partners 41. It appears that a firm of 25 lawyers would have easily ranked among the nation's “top 50.”.Google Scholar

The term “boutique” refers to a law firm that is highly regarded but is nonetheless conspicuously small and specialized.Google Scholar

3 “The Am Law 100,” 13 Am Lawyer (spec. supp.) (July-Aug. 1991) (“The Am Law 100”).Google Scholar

4 To give one example: the receipts of the nation's 50 largest firms increased, in real dollars, an average of 10% per year from 1972 to 1987—more than double the rate of growth in the legal services field generally. Sander, R. & Williams, E., “Why Are There So Many Lawyers? Perspectives on a Turbulent Market,” 14 Law & Soc. Inquiry 431, 437 (1989). Unless otherwise specified, all dollar figures in this review will be adjusted for inflation and expressed in “real” 1990 dollar values.CrossRefGoogle Scholar

5 “The Am Law 100” (cited in note 3).Google Scholar

6 Id. at 8.Google Scholar

7 The emergence of very large law firms has raised another significant concern—namely, that expansion is transforming the practice of law from a “profession” to a “business.” A succinct exposition of this argument is in Deborah Rhode, “Ethical Perspectives on Legal Practice,” 37 Sun. L Rev. 589, 631–38 (1985). These issues are also central to Nelson's fascinating exploration of the big firm (Partners, cited in note 2). While we do not pursue those issues here, they give additional interest to the question of what drives law firm growth.Google Scholar

8 A law partner interviewed by Robert Nelson in the late 1970s had a similar viewpoint: “I don't think we will ever be larger than say about [75] to [80] lawyers, because I think this firm will lose its esprit de corps if it gets up to that point. The administration probably will become unmanageable.” Nelson, Partners 112. It would be interesting to know what those partners think today.Google Scholar

9 It is easily shown that large firms are generally more profitable than small firms (in the sense of generating more income per partner), but this may be (and is generally assumed to be) because the individual members of large firms are more talented and skilled (or at least more admired) than the individual members of small firms. One would need to control for lawyer quality to make a meaningful comparison of scale economies.Google Scholar

10 Nelson, Partners 63. This seems consistent with our casual observations and logical as a theoretical matter. If economies of scale militated for 1,000-lawyer firms, one would think that such firms would have emerged long ago, through more aggressive patterns of merger, as happened in the accounting field. The fact that expansion has occurred almost entirely through internal growth of firms suggests that any economies which exist are not compelling.Google Scholar

11 Gilson, Ronald & Mnookin, Robert, “Sharing among the Human Capitalists: An Inquiry into the Corporate Law Firm and How Partners Split Profits,” 37 Stan. L. Rev. 313, 317–18, 356 n.70 (1985). As an example of realizing economies through sharing, Gilson and Mnookin give the common example of two independent practitioners sharing office space and a secretary. There are more elaborate examples: in the legal districts of several large cities, some office buildings maintain substantial law libraries that are jointly sponsored by a dozen or more independent offices in the building.CrossRefGoogle Scholar

12 Id at 317; Nelson, , Partners 63 (cited in note 2).Google Scholar

13 Such norms are well documented in Erwin Smigel's account of Wall Street law firms in the late 1950s. In explaining why one of the most famous of these firms did not have a written partnership agreement, Smigel noted, “This firm and many other large old … [law firms] regard themselves as special institutions. It is not unusual for them to liken themselves to the British government. They claim that an equivalent to common law grows up in the organization and is passed along and accepted by the new members. It is sort of a continuing gentlemen's agreement.” Erwin Smigel, The Wall Street Lawyer 211 (1964) (“Smigel, Wall Street”).Google Scholar

14 Accordingly, we hope that our conclusion, infra, that the portfolio theory is, at best, a partial explanation of law firm growth is taken as an exploration of an idea, and not as a dismissal of the many other issues explored in Gilson & Mnookin, 37 Stan. L. Rev.Google Scholar

15 Id. at 356 n.70.Google Scholar

16 This follows from the widely held view that professionalism in large firms has declined in recent years, with a resulting decline in the stigma attached to partner defections. Recall that in Gilson and Mnookin's view, the norms of professionalism inhibit grabbing, leaving, and shirking. Thus, a decline in professionalism might be expected to stall or reverse the growth of firms. For sources related to this point. see note 7.Google Scholar

17 B. Curran, K. Rosich, C. Carson, & M. Puccetti, The Lawyer Statistical Report: A Statistical Profile of the U.S. Legal Profession in the 1980s 13 (Chicago: American Bar Foundation, 1985). These figures are as of 1980.Google Scholar

18 Nelson uses this term in a narrower context: to describe “the advantages of joint production of different services.” An example of this occurs when “the demand for particular fields fluctuates, … [large] firms can move productive resources between fields, thus maximizing their utilization.” Nelson, Partners 62. The term “economies of scope” is equally applicable, however, to the more subtle mechanism Nelson goes on to develop, and which we summarize in the text.Google Scholar

19 Nelson points out a number of empirical problems with the portfolio theory. Most fundamentally, he found that “In my interviews with firm leaders concerning mergers and branching, risk avoidance was never mentioned as a motivating factor.”Id. at 64–66.Google Scholar

20 Economies of scope could drive growth in an era when new specialties are opening up. In the 1970s, most notably, many firms added practices in environmental law, ERISA (pension regulation), and health and safety regulation. We are skeptical, however, that these new areas accounted for more than a small fraction of growth in most firms during the 1970s, and we suspect they provide even less help in explaining rapid growth before and since that decade.Google Scholar

21 Nelson, Partners 68. In a similar vein, Nelson observes that economies of scale and scope “do not clearly militate for … [continued growth]. The flexibility of computerized research and word-processing systems suggests that law firms can achieve economies of scale at a modest size, certainly at a more modest size than current levels. … Rarely do the constituent work groups inside firms exceed thirty lawyers; most work is performed by much smaller work teams, generally two to six lawyers.”Id at 63.Google Scholar

22 Although we skip over this historical section (chaps. 2–4 of the book) quickly, we recommend it highly as a superbly written synthesis of a wide range of historical materials combined with clear theoretical perspectives.Google Scholar

23 In the argot of economists, the “up-or-out” system (as characterized by Galanter and Palay) is a deferred compensation scheme. Deferred compensation schemes have a long history in the economics literature. Economists predict their use in situations where the behavior of economic agents cannot be perfectly monitored. The notion that a promotion system could incorporate deferred compensation, and thus create an employee “tournament,” has been developed more recently. The formal theoretical development of Galanter and Palay's tournament is due to James Malcolmson. See Malcolmson, , “Work Incentives, Hierarchy, and Internal Labor Markets,” 92 J. Pol. Econ 486 (1984).CrossRefGoogle Scholar

24 It should be noted that there is nothing sacred about a fixed promotion rate in cementing the commitment of the firm. It is likely that there are a number of schemes which could assure a similar degree of commitment. For example, Gilson and Mnookin have shown that the “out” part of the up-or-out system could substitute for a fixed promotion rate as a means of persuading associates that the firm will not take advantage of them. Gilson & Mnookin, “Coming of Age in a Corporate Law Firm: The Economics of Associate Career Patterns,” 41 Stan. L Rev. 567, 578–81 (1989). The commitment scheme actually chosen should be the one which minimizes the cost of such a scheme. For more discussion of the costs of a tournament scheme, see id at 580 n. 38.CrossRefGoogle Scholar

25 Galanter and Palay do not discuss very much how a firm picks a particular promotion rate, but there is a clear pattern of promotion rates across firms. At one end of the spectrum are firms that typically promote a small fraction of associates to partnership, maintain a high associate/partner ratio, and have very high partner salaries. Davis, Polk & Wardell of New York, for example, has a 3:1 associate/partner ratio, a promotion rate well under 10%, and generated roughly $1 million in profits per partner in 1990. Mayer, Brown & Platt of Chicago, in contrast, has an associate/partner ratio of 1.2:1, a promotion rate of over 30%, and profits per partner of $305,000 in 1990. The tournament theory can account for these patterns. If a key purpose of partner promotion is to create a credible reward for associates, a small chance of a very big award may be as effective as a better chance of a smaller (but still rather impressive) award. Ceteris paribus, the tournament theory would predict an inverse relationship between the probability of an individual associate being promoted to partnership and the lifetime earnings of partners.Google Scholar

However, a simpler human capital theory could not only account for these patterns but could also explain why partner incomes diverge so much in the first place. If a partner's income is related to his talent (e.g., his ability to win cases, acquire reputation, and coordinate associates), then firms with high-earning partners also have, on average, more talented partners. If the partners are exceptionally talented, then the firm must be highly selective in choosing new partners; thus, fewer associates will qualify for partnership.Google Scholar

A third, more commonsense possibility is that the firms that have acquired such national reputations that they can command the huge fees that support high partner incomes can also secure loyal associates who know they have little or no prospect of partnership but expect to acquire valuable skills, contacts, and references through the firm. The ability to attract large numbers of such associates allows the firm to maintain a high associate/partner ratio and this ratio generates extra profits for the firm.Google Scholar

We suspect that all three influences are present and interact with one another.Google Scholar

26 We wonder how crucial this rate is in firms with very low promotion rates, including many of the most prestigious firms in the country (see note 23). Where a firm typically promotes fewer than 10% of its associates to partner, a much more important method of securing associate loyalty might be the firm's placement system—its assurance that the great majority of associates passed over for partnership will nonetheless be successfully placed with corporate clients or lower-tier law firms. This sort of arrangement would free a firm of the fixed growth hypothesized by Galanter and Palay (see discussion infra).Google Scholar

27 Galanter and Palay use the term “exponential growth” to describe growth at a constant rate. Technically, this is an overbroad term. An exponential function is one of the form xn′, where n is any independent variable (n can itself be a function). A geometric progression is a special case exponential function, where the variable n takes on the values of an arithmetic progression (e.g., the integers). Geometric progressions crop up in any number of phenomena; they have the common characteristic that the ratio of each term to its predecessor is always the same. Thus, 1, 2, 4, 8, 16, 32 is a geometric progression of xn′ where x is “2” and n takes on the values of the integers from 0 through 5. We will therefore use the term “geometric growth” in place of “exponential growth.”.Google Scholar

28 These figures ignore retirements, of course; but since this pattern of hiring quickly generates a pyramidal age structure, with young partners greatly outnumbering senior partners, factoring in retirement does not change the geometric nature of growth—it simply lowers the rate a bit.Google Scholar

29 Noneconomists may be troubled by the theory's simplifying assumptions that high partner earnings simply reflect a reasonable return on human capital and that all parties are rational income maximizers. We sidestep these paradigmatic issues here and evaluate the theory on its own terms.Google Scholar

30 For example, the authors note (at 137) that since the 1960s, “the big firms have continued to flourish. They have become larger, more numerous, more prosperous.”.Google Scholar

31 See F. Scherer, Industrial Market Structure and Economic Performance 260, 384, 427–28, 445–47 (2d ed. Boston: Houghton-Mifflin, 1980), for examples of the effect of “first-mover” advantages on market structure and firm size as they apply to other markets.Google Scholar

32 It also seems to us that, in many firms, decisions about how many associates to hire each year are viewed as short-term decisions, while the decision to promote an associate to partnership is seen as the quintessential long-term decision. The implicit divorce between these decisions seems at odds with the tournament model.Google Scholar

33 Our view is consistent with those studies that have closely scrutinized actual firm operations. in Nelson's case studies of four Chicago law firms, he never indicates that the partners he interviewed and observed believed that the partnership had any commitment to promoting a specified proportion of associates, or that anything like the Galanter-Palay tournament dictated the firm's growth. Erwin Smigel's portrait of Wall Street firms in the late 1950s jars with the tournament story in several respects: Promotions to partnership were infrequent, unpredictable, and came at varying points in associate careers. The ability of these firms to place associates with clients struck us as a more important source of firm leverage than the chance ac partnership. See Smigel, Wall Street chap. 4 (cited in note 13).Google Scholar

We should note that our own “sampling” of law firm partners is highly skewed to firms in Chicago and the West Coast: e.g., Kirkland & Ellis; Schiff, Hardin & Waite; McDermott, Will & Emery; Seyfarth, Shaw, Fairweather & Geraldson; O'Melveny & Myers; Morrison & Foerster, and Irell & Manella.Google Scholar

34 Nelson, , Partners 121 (cited in note 2).Google Scholar

35 Galanter and Palay created two samples of big firms. Group I consisted of 50 of the 56 largest firms as reported in the 1986 National Law Journal survey of big firms (6 firms were passed over because they are not classic “big firms”—e.g., Hyatt Legal Services—or because of problems in data availability). Group II consisted of “fifty smaller but still large firms that were roughly the 200th to 250th in the National Law Journal list in 1988.” Galanter & Palay, app. A.Google Scholar

36 For many firms there is a kink in the curves around 1970, which Galanter and Palay take to reflect an acceleration of growth due to surging demand, but they show that the kink is, in many cases, simply a change in the rate of geometric growth.Google Scholar

37 Galanter and Palay discuss the fact that one could set the associate/partner ratio and the promotion rate low enough to create zero growth (e.g., if there is one associate for every partner, a seven-year apprenticeship, and a 20% promotion rate) or an actual decline in size. But they suggest (at 109) overly restrictive conditions for such a pattern:.Google Scholar

In order to maintain the tournament structure necessary to motivate associates, the ZGPP [zero-growth] firm would need to require regular and consistent retirements. … Of course, maintaining the necessary age distribution might conflict with conducting the tournament purely on the basis of productivity. And if the zero growth promotion policy resulted in a promotion rate lower than that of the rest of the market, the firm would need to compensate by offering entering attorneys either higher money wages while they remained associates or greater rewards if they became partners.Google Scholar

The first critique is accurate only in theory; in practice, even the rapidly growing firms have marked fluctuations in partner promotion rates (see figs. 1 and 2), and these could offset irregularities in the age distribution of partners. The second critique is valid but applies with equal force to firms that grow geometrically but less rapidly than the fastest growing firms; the low-growth firms will tend to have lower promotion rates and may need to offset these with higher salaries. For related points, see supra note 24 and text following note 31, and the Conclusion.Google Scholar

38 The most vivid example that comes to mind in the organizational context was provided by the French economist Gibrat, who demonstrated that logarithms of firm sizes within or even across industries rend to approximate a normal distribution. A key assumption in deriving this result is that changes in firm size from period to period are proportionate rather than arithmetic. See Kalecki, , “On the Gibrat Distribution,” 13 Econometrica 161 (April 1945).CrossRefGoogle Scholar

39 It would be easier to understand Galanter and Palay's emphasis on this issue if arithmetic growth were in fact, as they imply, a widely held misconception among lawyers and legal scholars. But the book offers no citation to scholars who subscribe to linear growth.Google Scholar

40 Galanter and Palay note that “[m]any observers subscribe to a ‘shock theory’ of recent firm growth.” Although they provide no references, Nelson's analysis of legal demand growth is roughly consistent with this view. See Nelson, Partners 7–8 (cited in note 2), and note 18 and accompanying text.Google Scholar

41 In 12 out of 74 firms for which Galanter & Palay have data, the growth rate declined after 1970; in 7 cases, the declines were statistically significant. (See second-period coefficients for KEF models, listed in Galanter & Palay, app. B.).Google Scholar

42 Winston &a Strawn's average annual growth rates have been 0% for 1935–50, 1.7% for 1950–65, 5.5% for 1965–79, and 9% for 1979–86. These figures were computed by the authors from Galanter and Palay's dataset on large firm histories of growth.Google Scholar

43 A more extreme example of irregular growth is the law firm of Taft, Stettinius & Hollister (also computed from Galanter and Palay's dataset). The firm's average annual growth rates during each decade since 1926 have been: 1926–36 9.0%/year. 1936–46 1.6%/year. 1946–56 2.5%/year. 1956–66 8.3%/year. 1966–76 4.8%/year. 1976–86 6.1%/year.Google Scholar

44 Average calculated by the authors, based on the 91 firms for which KEF results are reported in Galanter & Palay, app. B.Google Scholar

45 Consider the gross national product (GNP) of the United States. Certainly in these troubled times we are keenly aware of the inconsistency and unpredictability of economic growth. But if we regress real (inflation-adjusted) American GNP on a 1939–87 trend line, we get an impressive.977 R2—a result that occurs despite the ups, downs, and sharply changing growth rates of the American economy over the past half-century.Google Scholar

46 Our calculations are based on the 75 firms for which data is given.Google Scholar

47 Martindale-Hubbell, the source used by Galanter and Palay in their census of firm sizes over time, contains enough information (e.g., names, ages, and dates of admission to the bar for associates and partners) to make feasible a more refined analysis, at least for the last 45 or 50 years. By counting the cohort of new associates each year and determining how many of those associates eventually become partners, one could construct a true “promotion rate” over time for each firm. Of course, to do this for 100 firms over a 50-year period would be an enormous task; given the theoretical emphasis of the book, it is understandable that Galanter and Palay chose to proxy the promotion rate. However, a rigorous resting of the tournament theory of growth will require the more refined measure.Google Scholar

48 Note that the promotion rate could “drift” and yet not show any long-term trend upward or downward.Google Scholar

49 Figs. 1 and 2 show, for two law firms, the ratio of new partners to existing associates on an annual basis and as a five-year floating average. Baker & Botts (fig. 1) shows sharp fluctuations, a significant upward trend over time, and clear cycling in the five-year trend. Robinson & Cole (fig. 2) illustrates sharp annual fluctuations, but a seemingly random long-term trend promotion rate that is small but positive.Google Scholar

50 Although law firm earnings were traditionally closely held secrets, the data for analyses of this sort do exist. See, e.g., Samuelson, S. S. & Jaffe, L. J., “A Statistical Analysis of Law Firm Profitability,” 70 B.U.L. Rev 185 (1990), which drew on data Price Waterhouse has been gathering for some 30 years on the profits and partner earnings of a substantial number of large law firms.Google Scholar

51 For an expanded discussion of this point, see Sander, &. Williams, , 14 Law & Soc. Inquiry at 467–73 (cited in note 4); Galanter, Marc, “The Day after the Litigation Explosion,” 46 Md. L. Rev 3 (1986).Google Scholar