Career questions invite a single number.
Not the proportion of lawyers who are partners today. Not the percentage of partners who belong to a particular group. Not a survey of whether associates want partnership. The question behind the headline is narrower:
Of the lawyers who begin on a partner track, what proportion ultimately become equity partners?
The honest answer has to begin with the denominator, the cohort, the firm structure, and the time horizon. Without those coordinates, a precise-looking percentage describes a career system more confidently than it has been measured.
This article is about what the available evidence can establish and what an individual planning file still has to learn.
Stock is not flow
The 2025 Thomson Reuters Institute and Georgetown Law report compares the average composition of 183 participating US law firms in two periods. Equity partners fell from 31.2% of lawyers in 2005–2009 to 27.8% in 2020–2024. Non-equity partners rose from 14.3% to 19.1%, while associates fell from 44.5% to 40.2%. [R005-C01 · DUAL-VERIFIED]
Those figures describe the workforce inside reporting firms.
They do not say that 27.8% of a first-year cohort will become equity partners. Partners remain in the stock for multiple years. Associates enter and leave. Firms merge, shrink, grow, and change titles. Non-equity partnership can be a destination, a waiting room, a different economic role, or a firm-specific label.
Using workforce composition as promotion odds is like using the proportion of directors on a board to estimate the chance that today's management cohort will become directors. The quantities are related to the same system and still answer different questions.
R005-M01 · MECHANISM — A level's share of the current workforce is not the probability that an entering cohort will reach that level.
The denominator keeps moving
“Starting associates” sounds clear until we define it.
Does the cohort include only first-year associates hired from law school? Lateral associates? Lawyers who join a firm with prior experience? People who leave and later return? Lawyers who become partners at another firm? Lawyers who choose an in-house, government, public-interest, or entrepreneurial route?
“Equity partner” also changes by firm. Some firms operate one tier. Others have equity and non-equity tiers, salaried partners, fixed-share partners, income partners, or counsel roles with different economics and voting rights.
A universal percentage can conceal these structural differences.
The correct analysis is firm- and cohort-specific. Follow a defined entry cohort over enough time. Record every exit and destination. Distinguish partnership at the original firm from partnership elsewhere. Distinguish equity status from title. Preserve the time at risk.
That dataset would not produce one answer for “law.” It would produce distributions across firms, practices, offices, years, and people.
Representation data is not transition probability
Representation snapshots are not promotion rates.
A single partner-track probability can erase unequal attrition, opportunity allocation, sponsorship, client access, evaluation, and title structures.
The article should not convert a representation gap into a causal story the data does not establish. It should also not offer “visibility” as a cure for structural inequality.
Public authority can make work more inspectable. It cannot repair a biased allocation system by itself.
R005-M02 · MECHANISM — A career estimate that ignores subgroup and firm structure can look neutral while describing nobody accurately.
What a planning model would require
A planning model might begin with a hypothetical cohort of 100 associates.
It would then apply year-by-year continuation rates, distinguish voluntary and involuntary departures, allow lateral moves, define the partnership decision year, separate non-equity from equity promotion, and decide whether partnership at another firm counts.
Different assumptions can produce materially different results.
That sensitivity is not a defect to hide. It is the main result.
A published planning figure would have to display:
- the starting cohort definition;
- every annual continuation input;
- the partnership horizon;
- the treatment of lateral movement;
- the treatment of non-equity partnership;
- the central result;
- a sensitivity band;
- the fact that it is modeled, not observed.
A solitary odds badge would hide the main result: sensitivity to assumptions.
What the imperfect data does show
The legal-market structure is increasingly tiered in the reviewed Thomson Reuters comparison: a smaller equity share and a larger non-equity share in the recent period. [R005-C04 · SYNTHESIS · DUAL-VERIFIED]
That comparison does not establish a universal individual route. Revenue, client access, sponsorship, evaluation, and title rules still need firm-specific evidence. The structural result is useful because it changes the question from “What are the odds?” to “Which destination, cohort, and decision process are we actually measuring?”
Replace “odds” with evidence gates
A probability can create fatalism. Evidence gates create work.
Gate one is technical credibility. Can the lawyer perform at the expected level in the relevant practice? This is necessary and often insufficient.
Gate two is trusted execution. Have senior lawyers and clients seen the person handle increasingly consequential work?
Gate three is internal sponsorship. Is someone with power prepared to spend political capital on the person's advancement?
Gate four is economic relevance. Does the candidate retain, deepen, originate, or help the firm win valuable client relationships under the firm's actual model?
Gate five is leadership and risk. Will peers trust the person with people, standards, cross-practice work, and the firm's reputation?
Gate six is evidence at decision time. Can the committee see a coherent record of these contributions, or must someone reconstruct it from memory?
The gates are not universal scoring criteria. Firms weight them differently. Their value is diagnostic: they convert a distant title into evidence that can be gathered, reviewed, and discussed.
R005-M03 · MECHANISM — A career path becomes more actionable when the destination is decomposed into evidence-bearing decisions rather than expressed as one probability.
Public authority has a narrow role
A public record can support some of the evidence gates.
It can demonstrate a point of view on client-relevant problems. It can show the ability to explain complex issues clearly. It can document the development of a practice area. It can make a lawyer easier to evaluate for speaking, writing, recruitment, or client conversation.
It cannot prove confidential client work. It cannot substitute for internal sponsorship. It cannot guarantee origination. It cannot force an equity decision.
The best public work is adjacent to the actual practice, not a performance of expertise detached from it.
A useful article might explain a regulatory trade-off without disclosing a client. A chart might clarify a market structure using released data. A field note might document how the lawyer frames a recurring decision, clearly labelled as professional judgment rather than universal advice.
Visibility should make real work more legible. It should not create a second fictional career.
Ask for the firm's denominator
Before using any public partner-track statistic, ask the questions the statistic avoids.
How many first-year associates entered this firm or peer group in the relevant years?
How many remained at years three, five, seven, and the partnership decision?
How many became non-equity partners, equity partners, counsel, lateral partners elsewhere, or leaders outside private practice?
How does the result differ by practice, office, hiring route, gender, race/ethnicity, and flexible-working pattern where lawful and responsibly measured?
What changed in the firm's structure during the period?
If the firm cannot answer, the gap itself is management information. It means the career system is being narrated more precisely than it is measured.
The planning file can still move forward. Record which definition the firm uses, which gates apply to the candidate's practice, which evidence is observable, which opportunities depend on allocation by others, and when the next real decision will be reviewed. That produces a better conversation than importing a universal percentage from another population.
Honest limits
A firm-composition table is not an individual forecast. The Thomson Reuters data describes firm composition. The NALP data describes representation. The Briscoe and von Nordenflycht study uses records from one large firm and examines revenue-credit patterns.
A later modeled range can support planning only if its assumptions and sensitivity are visible. It cannot predict one person's outcome. Firm economics, practice demand, sponsorship, performance, client access, geography, identity, life choices, lateral moves, and changing partnership structures all matter.
Crafted Virtue cannot promise partnership or claim that public writing causes it.
It can help build a governed record of relevant judgment and measure the evidence that exists today.
That is less satisfying than a precise odds badge.
It is more honest—and more useful.