The Puzzle and Persistence of Biglaw Clustering

Elite U.S.-based global law firms (Biglaw) concentrate their offices in the costliest districts of superstar cities, especially two neighborhoods in Manhattan. This pattern has persisted despite both the dispersal of Biglaw clients across less-dense, lower-cost U.S. locations and the development of telework capacity. This suggests a puzzle: Law is among the occupations with the highest theoretical propensity for remote work, yet prior to the pandemic, Biglaw required in-person work in the priciest places – meaning it paid (and continues to pay) a premium on both of its biggest expenses, wages and real estate. Yet Biglaw is famously profit-driven, suggesting that this double premium is necessary in some way. How might this equilibrium be explained, and what might lead it to change?

In a recent article, I study the strategy of spatial clustering among top firms and explore the potential for remote work – specifically, the mass telework experiment begun in 2020 – to upend it.

Biglaw has opted for a tightly clustered spatial distribution over several logical alternatives, including headquartering in lower-cost cities, dispersing offices more generally, and relying on more remote work. None of these alternative strategies were novel or mysterious prior to 2020, yet none were pursued. In fact, the data arguably show the true extent of Biglaw concentration to be even greater than many realized. Yet far from reducing productivity, the abrupt shift to remote work starting in 2020 was accompanied by record profits. This raises several questions, including what held Biglaw back from pursuing remote work or a diffuse office footprint previously and whether the pandemic shift towards population dispersal in general and telework in particular will have long-lasting impacts on the geographic human capital strategies of Biglaw.

Novel Data on Biglaw Clustering

First, the data. In the United States, the top firms in Biglaw are disproportionately headquartered in the largest cities. The top 25 firms in the AmLaw 100 by revenue per lawyer (a common proxy for profitability) are all based in one of six cities. If a central planner were to distribute the top 25 Biglaw headquarters proportionately by population, then the Los Angeles Metropolitan Statistical Area (MSA), at 4 percent of national population, would be predicted to have exactly one such headquarters. Instead, L.A. has four, yielding a representation score (ratio of actual-to-predicted headquarters) of 4:1, or, simply, four. Although the New York MSA – at a bit under 6 percent of national population – is the nation’s largest, if New York’s Biglaw headquarters  share matched its predicted share based on population it would have only one or two such headquarters. Instead, it has 15 of the top 25, a representation score exceeding 10. Indeed, every MSA that is home to even a single top-25 Biglaw headquarters has a representation score greater than one. But even these examples of overrepresentation mask the extent to which Biglaw is concentrated.

The experience of L.A. Biglaw illustrates why. Los Angeles is the second-largest city in the United States, and the L.A. MSA is the second largest. Yet for each of the four Biglaw outfits headquartered in Los Angeles, the firm’s largest office is actually located a continent away, in New York. Thus, in addition to the 15 top-25 Biglaw firms that are based in New York, all four top-25 Biglaw firms that are nominally based in Los Angeles operate their largest office in New York, meaning 19 of the top-25 Biglaw firms have their largest office in New York. This gives New York an adjusted representation score of 12.99 – meaning, New York contains about 13 times as many top-25 Biglaw headquarters or largest offices as would be predicted by its metro-area population share alone. Table 1 presents these and related data in this form for the first time.

Table 1

HQ City of Top 25 AmLaw 100 Firms by RPL HQ of Top 25 (Total N of Largest Offices Regardless of HQ Location) Share of Top 25 (Share of Total Largest Offices Regardless of HQ Location) Metropolitan Statistical Area & 2019 Population (U.S. total: 328,239,523) Metropolitan Statistical Area 2019 Share of U.S. Population Representation Score (Adjusted Representation Score) Indicating Share of Total HQs (Largest Offices)  Divided by MSA Share, with 1.0 as Perfectly Representative
New York 15 (19) 60% (76%) New York- Newark-Jersey City, NY-NJ-PA: 19,216,182 5.85% 10.26 (12.99)
Los Angeles 4 (0) 16% (0%) Los Angeles- Long Beach- Anaheim, CA: 13,214,799 4.03% 4 (0)
Boston 2 (2) 8% (8%) Boston- Cambridge- Newton, MA-NH: 4,873,019 1.48% 5.41 (5.41)
Silicon Valley 2 (2) 8% (8%) San Jose- Sunnyvale- Santa Clara, CA: 1,990,660 0.61% 13.11 (13.11)
Chicago 1 (1) 4% (4%) Chicago- Naperville- Elgin, IL-IN- WI: 9,458,539 2.88% 1.39 (1.39)
Washington, D.C. 1 (1) 4% (4%) Washington- Arlington- Alexandria, DC-VA-MD- WV: 6,280,487 1.91% 2.09 (2.09)


Clustering of Biglaw is most extreme in New York City. If the number of Biglaw headquarters in the Big Apple were proportional to the share of the U.S. population living in that metro, the expected number of headquarters would be only one or two. Instead, of the 25 top Biglaw firms, New York City is home to 15 of these firms’ headquarters and 4 other of these firms’ largest offices. While orthodox economic theory would presume that this choice is efficient for the firms in question, the experience of the past few years adds nuance to that story.

“Explore” vs. “Exploit” Strategies

Ordinarily, firms face tradeoffs between the exploitation of proven strategies and the exploration of novel and uncertain ones. When it comes to location decisions, Biglaw has traditionally favored exploitation, i.e., in-person work in the commercial core of global cities. That strategy boasts many advantages, but it is not obviously required by the external factors that are traditionally associated with production location decisions.

Under an exploitation model, firms focus on refining their existing, successful strategies. In contrast, when firms employ an exploration model, they focus on discovery and experimentation. Though exploration can lead to success, it requires firms to spend money and time on actions they may ultimately deem unsuccessful. To avoid the uncertainty associated with exploration, Biglaw leadership has chosen to exploit known successes by continuing to pay a double premium for labor and real estate in major cities.

Biglaw is not the only industry that traditionally prefers exploitation strategies. Pre-pandemic, clustered in-person work was said to be preferred in knowledge industries because of agglomeration effects: its advantages in facilitating informal and non-verbal communication, monitoring, intrafirm knowledge spillovers, a sense of shared purpose, and organizational nimbleness. Like many industries, Biglaw was willing to pay a double premium for these benefits pre-covid.

But the explore vs. exploit question contains a sequential dimension: Exploration naturally precedes exploitation, and until recently there was little impetus for further exploration. Biglaw has now spent two and a half years exploring full and partial telework.

Lower Levels of Telework Pre-Pandemic Probably Reduced Biglaw Profits

Before the pandemic, periodic telework was a complement to the Biglaw lawyer’s routine of in-office work. From this initial phase, remote work arrangements evolved ad hoc. Due to their bespoke nature, the negotiation of these arrangements created more frictions relative to a universal policy or mandate and positioned telework as an individual perk or accommodation rather than an entity-level cost-saving strategy to maximize profits.

The pandemic changed telework from a complementary arrangement to a substitutional one. In 2020, telework became the dominant strategy for most top Biglaw firms, and at many this continued through 2021 and some of 2022. Notably, during this period, firms saw gains in net income. Association is not causation, but widespread concerns around decreased team efficiencies stemming from telework do not appear to have shown up in profits. Arguably, the experience of the past two and a half years suggests that the comparatively low share of Biglaw work that was done remotely pre-pandemic may have negatively impacted firm profits.

The Pandemic Bright Spot for Biglaw: An Elimination of the “Explore/Exploit” Gap in Telework

Beyond stickiness and path dependency, two types of costs likely reduced the level of telework prior to the pandemic: (1) agency costs (especially adverse selection) and (2) the transaction costs of coordination and switching. This plausibly led to a situation where remote work was underprovided in Biglaw, as in other domains.

The pandemic and the mandatory shuttering of Biglaw offices in early 2020 radically altered both the agency costs and transaction costs of remote work. Then, Biglaw offices reopened, but few mandated in-person presence, and those that did do not appear (as of summer 2022) to have enforced it rigidly.

The agency costs theory of under-provision holds that firms in the pre-pandemic world feared that offering telework to employees would negatively select for less productive employees. These employees would be enticed by the possibility that an employer would have a harder time monitoring remote productivity. Pre-pandemic telework was also limited by the transaction costs incurred by firms when dealing with a mix of remote and in-person employees. An apt example is the increased work involved in coordinating a meeting when some employees are remote and others are in person. Firms have now been coordinating such meetings for over two years.

Since 2020, top firms in Biglaw have adapted successfully to a telework-first strategy. This adaptation has turned telework from a potentially costly exploration strategy to another exploitation strategy. Thus, firms in the post pandemic world enjoy a choice between two tested, successful strategies to exploit.

In Biglaw, two possible futures for telework stand out. First, top firms may simply revert to 2019 telework norms (or close), which position telework as a complement or accommodation for the benefit primarily of the employee. Alternatively, they may build on the experience of high profits during the telework-first years and evaluate how remote work can shrink real estate or salary costs. This may become easier to contemplate as firms’ commercial leases, which are often 10 or 20 years in duration, come up for renewal.

The path Biglaw takes should be of broad general interest.  It will affect not only the degree of elite law firm clustering, but real estate values, municipal finances, and other dimensions of economic and social activity.

This post comes to us from Professor Gregory H. Shill at the University of Iowa College of Law. It is based on his recent article, “The Puzzle and Persistence of Biglaw Clustering,” available here.