Why We Work Analysts to Death
How the eighty-hour workweek reveals the hidden structure of human capital markets
Institutions are heartless, but they are not impractical. And even hazing becomes impractical very quickly.
When I worked in BigLaw, I knew the long hours were part of the bargain. But something still didn’t quite add up. As the folk theory goes, the long hours are evidence that banking, consulting, and law firms — which rigorously optimize every facet of their business — have somehow failed to update their talent development practices since the 1960s.
That implication is hard to square with how these firms actually behave.
Three Myths about the First-Year Grind
There are three standard explanations for long junior hours in professional services, and none of them explain collective behavior among firms.
1. Extraction: “Firms work juniors long hours because it’s profitable.”
The most serious folk explanation holds that junior labor is cheap relative to the revenue it supports, and long hours pad the bonuses of those higher up the food chain. This resonates because it fits basic intuitions about capitalism.
But the underlying premise is shaky. “More hours worked” does not translate proportionally into “more money made.” Consulting and banking don’t usually bill by the hour, and law firm engagements are increasingly fee-capped.
The question is not whether another hour can generate revenue. It’s whether it generates more value than it destroys in industries where the cost of mistakes is extremely high. John Pencavel, writing for SIEPR, concluded that weekly output is largely insensitive to increases in work time beyond sixty hours, while Dawson and Reid, writing in Nature, found that twenty-four hours of wakefulness degrades cognitive performance to the equivalent of a 0.10% blood alcohol concentration. In first-year analyst terms, that’s about five shots of Fireball.
So pure extraction cannot be the sole mechanism. It can explain the first sixty hours, but not the next twenty.
2. Training: “The hours are an investment in developing talent.”
This HR-approved framing posits that juniors learn by doing and long hours accelerate the learning curve. This reasoning makes a bit of sense, even among some juniors, because the early-career learning curve is steep, repetition builds fluency, and certain skills only develop under pressure.
But again, nothing exhausts like exhaustion, and the learning curve flattens quickly after 11:00pm. No serious person believes late night work is pedagogically rich.
3. Inertia (d/b/a Tradition): “That’s just how it’s always been.”
There’s a partly anthropological, partly psychoanalytic account that treats long hours as the founding myth of corporate civilization, so your boss will haze you as he and his fore-bosses of old were hazed.
Still, tradition is a weak constraint in industries where junior talent is mobile, client relationships are fragile, and reputation is repriced daily. Culture can transmit a practice, but at some point, even the Romans thought that gladiators were too expensive.
Equilibrium
“Maybe none of these explanations works in isolation, but what if they’re all operating at once?”
I mean, sure. Extraction, training, and tradition each account for some additional junior hours somewhere, and most of us have worked under someone who seemed genuinely animated by one of these three.
But each of the folk explanations depends on the particular culture and economics of a given firm. On the other hand, long junior hours have persisted within firms through leadership changes, within industries through technological revolutions, and across geographies through shifting global fashions. They have even survived headline-grabbing lawsuits.
This points instead to a stable equilibrium created by structural features of professional services markets. The eighty-hour workweek reflects a fairly consistent quantity of labor sold within a relatively narrow band of wages. The status quo persists because firms continue to demand these hours and ambitious grads continue to supply them.
Demand
The demand side of the labor market in professional services is shaped by management’s ability to sell work performed by juniors to firm clients. Here, the structural constraints are coordination problems and adverse selection, but the root cause traces back to the client paying the firm’s bills.
What Rolls Downhill
In professional services, work product is complex. And at marquee prices, it has no excuse to not be perfect.
Because the work must be perfect, the cost of a mistake is convexly negative. Experts are assembled around bet-the-company events, so a material misstatement on investor-facing docs or a missed filing deadline is very expensive. Market share among banks, consultancies, and law firms is also disproportionately allocated by reputation.
Because the work is complex, it’s the joint product of many contributors, with several layers of review. A slide deck must be created by an analyst before it can be reviewed by an associate, pending sign off by a VP or MD.
These constraints incentivize firms to internalize redundancy where it is least expensive, that is, as low as possible on the totem pole. It’s not a partner who’s staying up till 2:00am to re-reread an outgoing memo. It’s the person taking home a tenth of her pay.
So clients demand perfection and responsiveness that match the sticker price. Within the firm, this causes grunt work to cascade downhill. These coordination problems will be smoothed as AI reduces the turnaround time for doc gen and review, but it won’t eliminate them entirely.
The Market for Lemons
Firms and the clients that engage them also face a two-tiered adverse selection problem.
At the lower tier, employers have no actual way of knowing how good you are on the day you collect your badge. When it’s all said and done, the firm got to know the winning candidate by reading a cover letter that may or may not have been written by Claude, and by having six different people talk to him for a half hour each. And he has every incentive to overstate how well he handles pressure, how much he covets my corner office, and how many people actually showed up to his frat’s charity bake sale.
And now some twenty-something I barely know shows up, and I have to pay him $3,000 a week plus insurance and payroll tax until I figure out if he’s any good. This incentivizes me to get that information as quickly as possible. Voluntelling him to fill his plate with enough deals to round out an eighty-hour workweek increases the surface area for performance review per unit paycheck. It’s also a happy coincidence that the traits I care most about — endurance, grace under pressure, prioritization, efficiency, coachability in extremis — only surface when a person is marched into the hinterlands of their bandwidth.
But why do I really care if my fellow partners/MDs largely absorb the dollar cost of the new hire’s underperformance or churn? Because my ability to collect fees from my clients now depends on them believing that the work I’m billing for (which they know is farmed to juniors) is actually worth it.
This is where the upper tier of the adverse selection problem comes in. In practice, a client’s main interaction with a bank, consultancy, or law firm comes through the partner/MD in charge, punctuated by update calls, invoices, and the occasional snappy presentation.
In law, where most everyone except Wachtell gives itemized invoices, the client sees that juniors spent 85% of the time to generate 65% of the bill, and asks, “Are the tasks these people performed worth it, and done well?”
For banking and consulting, the trust problem is even more salient. If a CFO approves $500,000 to $2 million for an MBB consulting assignment, or a percentage of deal size for M&A advisory, she will rightly ask, “I know the MD I play tennis with wasn’t doing all of that by herself...Who else was helping, and how do I justify to the board how much we paid?”
So a partner/MD’s ability to minimize write-downs depends on how much her clients trust her juniors. Employers outsource the burden of filtering good candidates to colleges, who screen applicants. Clients, in turn, outsource the task of filtering good junior bankers, consultants, or lawyers to the firms that hire and train them. If a firm can credibly say, “our juniors are swift as a coursing river,” they signal that their selection process is sound and hence the invoices deserve to be paid.
This incentivizes firms to solve the lower-tier adverse selection problem — namely, to ramp the juniors up to an eighty-hour workweek as quickly as possible — and to tacitly relish their reputation as a hazer.
The eighty-hour workweek then is less a labor practice than a screening technology. And somewhere along the way, it also became a Dickensian form of virtue signaling: you can tell the factory’s working, m’lord, by how skinny the waifs are inside.
Supply
At equilibrium, supply in a known amount meets demand at a given price, which means the workers themselves need to participate. Across professional services, the equilibrium price of junior labor is well known. For lawyers, it’s the Cravath scale; for bankers, it’s around $110K base plus bonus at bulge brackets; and for consulting, it’s around $110K base plus performance and signing bonus at MBB. These are the clearing prices at which recent college and law school graduates are prepared to sell up to eighty hours per week of labor.
Assuming full busyness and top-tier bonus pay, this ballparks hourly wage at around $50, which puts a star first-year at Goldman roughly on par with the top decile of Uber drivers. One could argue that the marginal cost of the sixtieth hour is much higher than the fortieth, so the disutilities (econ-babble for “this sucks”) at a banking job actually outweigh those in professional driving.
Bright recruits know that the wages don’t justify the pain on a one-year time horizon. The real question is how to calculate the expected value of deferred compensation.
Pie-Eating Contests
If we’re honest, most new hires don’t expect to retire as rain-makers from the same firm where they started. But tournament logic still shapes the internal social order of firms: who gets the best work, the strongest mentors, the highest evaluations, and ultimately the best exits. The day-to-day experience of a junior employee is still organized around the proverbial pie-eating contest, where the prize is more pie.
Even if you don’t know or don’t expect to become a partner/MD one day, your dominant strategy is to behave like one. Embodying future-partner/MD material takes immense work, but it paradoxically reduces friction in office life on multiple fronts. Coworkers like to work with dependable teammates. Bosses give interesting work to rising stars. It’s a lot easier to land your next job if you’re crushing at your current one. And who knows, you might even just get promoted.
So the up-or-out structure of professional services firms creates a tournament where your best option is striving to win. Not because you actually hope to win, but because you fully expect to lose.
College II: This Time It’s Personnel
Taking a step back, human capital cannot be sold. It can only be rented.
The labor market is the human capital rental market, where employers rent temporary access to a productive asset whose quality they cannot observe directly. That opacity gives rise to institutions analogous to those in the capital markets, where the information asymmetry between corporate insiders and outside investors is also extremely high.
Universities exist as informational intermediaries that decrease transaction costs and increase liquidity in the labor market. At various turns, universities play familiar roles:
Underwriters (we diligenced this graduate for four years when he lived in our dorms and took our exams)
Rating agencies (we are willing to stake the Harvard brand on this graduate)
Placement agents (on-campus recruiting)
Credible universities reduce search costs for employers and increase trust in entry-level candidates. This results in a labor market where untested graduates find jobs more quickly and in greater volume than they otherwise would.
Few would dispute that a Harvard degree increases your personal future cash flows over baseline and has a positive net present value. This is because a Harvard degree does more than help land your first job. Your Harvard degree is a portable certification that will always sit on the top line of your resume, reducing friction with future employers and helping you win every subsequent job. If we decompose the NPV of a Harvard degree into its constituent value drivers, one of them reflects lower transaction costs each time you bring your labor back to market.
The same dynamics play out in the market for experienced hires, with name-brand employers playing a similar role to prestigious universities. PE and hedge funds outsource the rating and underwriting of mid-career banking talent to bulge brackets. BigLaw or MBB experience certifies a candidate to in-house legal or strategy teams respectively. Like having a Harvard degree, the NPV of being ex-Goldman, ex-McKinsey, or ex-Cravath is large and positive, reflecting reduced friction at every subsequent rental event in the labor market.
So elite employers do not merely consume credentials. They produce them.
This is the deferred compensation that motivates college grads to apply for top firms and new analysts to accept hourly pay on par with an Uber driver. I’d go so far as to say that the portable certification of being ex-BigFirm is not the consolation prize for failing to make partner. It is the prize on probability-weighted terms.
So there’s a reason why landing that first big job feels like winning the lottery. It’s a human capital arbitrage opportunity of the sort I’ve explored before. You play the pie-eating game for a couple years. You get rewarded with a portable, positive-NPV certification, even if you lose. And unlike college, you don’t pay tuition. You get paid.
But like most human capital investments, this trade is gated by a lock-up period of discipline and execution. No partial credit. No early exit. “Ex-Goldman” doesn’t open that many doors if you got fired after three months.
Credential Factories
Elite firms have persisted not because they employ the brightest people, but because they produce trusted information about them. The eighty-hour workweek merely reveals that both the supply side (employees) and the demand side (the firms) have a shared interest in preserving the value of credentials.
The analyst wants “ex-McKinsey” to remain valuable because she expects to carry it into every future labor-market transaction. The MD wants “McKinsey analyst” to remain valuable so that her clients implicitly trust the people staffed on their engagements. Supply and demand converge on the same objective: preserving the firm’s reputation for producing credible information about human beings.
So the eighty-hour workweek is not per se an economic necessity. It’s an expedient manifestation of the structural position that elite institutions play in human capital markets: as informational intermediaries that discover and certify qualities about human beings that the labor market cannot observe for itself.
This also reveals where AI poses its deepest challenge. The obvious story is that AI automates junior work. The more interesting story is that AI changes the production function of elite credentials.
At the start I said that institutions are heartless, but they are not impractical. The same can be said for the human capital markets. If AI changes what can be learned about people at work, universities and elite employers won’t cling to old rituals out of nostalgia. They’ll redesign themselves to manufacture whatever information the market still cannot cheaply obtain. Whether that still looks like the eighty-hour workweek is a different question entirely.
Works Cited
“Centerview Settles Lawsuit over Analyst’s Need for 8 Hours’ Sleep.” Financial Times, 22 Feb. 2026, https://www.ft.com/content/a5271651-825e-4007-a029-9e2c16b5d770. Accessed 8 July 2026.
Dawson, D., and K. Reid. “Fatigue, Alcohol and Performance Impairment.” Nature, vol. 388, no. 6639, 1997, p. 235. https://doi.org/10.1038/40775.
Pencavel, John. The Productivity of Working Hours. SIEPR Discussion Paper No. 13-006, Stanford Institute for Economic Policy Research, Oct. 2013, https://siepr.stanford.edu/publications/discussion-paper/productivity-working-hours.
Claude Opus 4.8 (Anthropic) was used for spell-checking, proofreading, fact-checking, and citation formatting.
Cover photo licensed from Vecteezy.
Edited 9 July 2026 to make minor formatting fixes.
All ideas, arguments, source selection, and writing are my own.