THE HOUSES OF THE STREET (1960–2025) — POWER OF THE GREAT INVESTMENT BANKS, IN MARKET SHARE

Why they got so powerful. Three structural advantages, compounding for four decades. (1) They stayed private the longest: through the 1980s every rival sold itself to a deep-pocketed parent — Salomon to Phibro, Lehman to American Express, Dean Witter to Sears, Kidder to GE — and every parent's capital dissolved the partners' discipline, because it was no longer their money at risk. Goldman's partners kept their entire net worth in the firm until the 1999 IPO, the last major house to convert. When your own fortune is the trading book, you read the risk reports differently. (2) The culture machine: Gus Levy's "long-term greedy" — take the loss today, keep the client forever — plus the most selective recruiting on the Street and a partnership tournament that made the firm, not any product, the object of loyalty. Rivals were serially captured by their hot desk: Salomon became its bond floor, Drexel became Milken's junk machine, Bear and Lehman became their mortgage books. Goldman never let one desk become the firm. (3) They side-stepped the fatal trade: in December 2006, when the mortgage desk started losing money ten days straight, Goldman's leadership ordered the whole book "closer to home" — and the firm entered 2008 hedged short the very securities that killed Bear, Lehman and Merrill. Ruthless, controversial, litigated — and the reason there's a band left to argue about.

Why they stayed on top while the Street burned. Strategy: they never bet the house on a single product war. Advisory when advisory paid, trading when trading paid, principal when principal paid — the revenue mix rotated while the franchise compounded. Every rival's obituary names its one fatal concentration; Goldman's near-death moments (Penn Central's paper in 1970, 1994's rate shock, 2008 itself) were survived precisely because no single position was the firm.

Why the aura cracked anyway. Victory brought the scrutiny the partnership had always avoided: the "vampire squid," the Abacus settlement (2010), 1MDB (a $2.9B penalty and a guilty-pleading subsidiary), and a billion-dollar wander into consumer lending (Marcus) that proved even Goldman can be tourists somewhere. The house that survived every war on the Street was nearly wounded worst in peacetime, by itself.

What happened over six months in 2008 looks like chaos until you see it as a single forced land reform. It was a set of forced swaps, not rescues freely chosen — and, like Hideyoshi at the Kantō, the sovereign decided who got what. In March, Bear Stearns — 85 years old, never a losing quarter until its final one — was handed to JPMorgan for $2 a share (raised to $10 to quell the shareholders' mob), with the Fed financing $29B of the corpse's worst assets. The precedent set, the Street assumed everyone would be caught.

Then came the weekend of September 13–14, when the sovereign chose differently: (1) Lehman — which had survived the Civil War, the Depression, and its own 1984 civil war — was allowed to fail outright, the largest bankruptcy in world history (~$613B), because Washington needed one head on a spike; (2) Merrill Lynch, reading the lesson in real time, married Bank of America in 48 hours rather than be Monday's corpse; (3) within the week, Goldman Sachs and Morgan Stanley — the last two independent houses — surrendered the broker-dealer model itself, converting to bank holding companies: trading the freedom of 30:1 leverage for the shelter of the Fed's discount window. The five great houses of 2007: two dead, one captured, two converted. The independent investment bank, as a form, was abolished in eleven days.

The regranting decided the next twenty years. JPMorgan got Bear (and then WaMu) at fire-sale prices and emerged the largest bank on Earth. Barclays bought Lehman's U.S. business out of bankruptcy for ≈$1.75B — the actual trade of the decade — while Nomura took Asia and Europe for, famously, two dollars plus retention. Mitsubishi's $9B wire (allegedly a physical check, delivered on a holiday weekend) kept Morgan Stanley alive; Buffett's $5B blessed Goldman. And the deepest irony: the universal banks whose model the houses had spent 70 years mocking — the boring, deposit-funded giants Glass-Steagall had caged until 1999 — ended up owning the Street. The exile from independence became the throne: the survivors, fused to fortress balance sheets, grew larger than the free houses had ever been.

How the share estimates were made (Fermi method): Anchors are league tables, reported segment revenues (post-IPO), and partnership capital rankings (pre-IPO, when the houses were private and told no one anything). Documented checkpoints keep the bands honest: Merrill's retail machine made it the revenue giant of the Street from the 1960s onward; Salomon was crowned "the King of Wall Street" in 1985 on its bond monopoly; Drexel was briefly the most profitable firm on the Street (1986) at the peak of junk; the five independents' 2007 leverage and revenues are in their final filings; and 2025 disclosures show JPMorgan first in global investment-banking fees with 8.4% share and first in Markets revenue, Goldman first in M&A advisory revenue for the 23rd straight year, and Morgan Stanley at $33.1B of Institutional Securities revenue. Pre-1975 figures are koku-equivalents at best — partnership revenue was a secret — so early bands are inferred from capital, headcount and syndicate standing. Error bars are roughly ±1–2 points.

Two things that look wrong but aren't. (1) Revenue share is not prestige. Merrill's giant band in the early decades is retail commissions from the thundering herd; Morgan Stanley's slim band ruled the Street anyway, because under the syndicate hierarchy it was first on every tombstone — precedence, not revenue, was the currency of the club, which is exactly why Mayday (1975), which repriced revenue, destroyed the club. (2) The biggest bankers of the early era aren't on the chart at all. Glass-Steagall (1933) caged the commercial banks — Morgan Guaranty, Bankers Trust, Citibank — outside the securities business entirely, like captive IBM outside the merchant chip market. The JPMorgan band appears only in the 1990s as the wall eroded, and the wall's formal demolition (1999) was itself forced by a merger already consummated in defiance of it. When the caged banks finally invaded, they ended up owning the Street — the chart's last era is their revenge for the first.