The physician gender pay gap persists after adjustment for specialty, hours, and productivity. A 2016 study in JAMA Internal Medicine found female physicians earned roughly $20,000 less per year than male counterparts after extensive controls. Subsequent analyses confirm similar patterns across specialties and career stages. Most of this literature treats the gap as evidence of discrimination. That framing identifies the problem but offers no mechanism and few remedies.

This essay applies a different lens. Claudia Goldin’s Nobel Prize-winning research showed that the largest modern pay gaps arise not from employers paying women less for identical work, but from labor markets that disproportionately reward what she calls “greedy jobs”: positions demanding long, inflexible hours, continuous availability, and on-demand responsiveness. In her book-length treatment and earlier work, she demonstrated that the key mechanism is non-linear pay. Working 60 hours does not pay 1.5 times what 40 hours pays. It pays two or three times as much, because the market attaches a premium to the marginal hours that signal total availability. The pay gap in these fields is less about gender than about who bears the cost of temporal inflexibility, a cost that falls disproportionately on women after the birth of a first child.

Her conclusion was not “pay women more for the same job.” It was to redesign jobs so that flexibility carries less penalty. This framework has not been applied to medicine in any systematic way. It should be.

Medicine’s greedy job

On the surface, physician work looks standardized. Same residency. Same board exams. Same Current Procedural Terminology (CPT) codes. Same Relative Value Units (RVUs). The system appears neutral. It is not. Physician compensation under productivity-based models has its own version of greedy job economics, and the highest-paid physicians are not simply doing more of the same work. They are operating within structures that reward specific patterns of availability and volume in ways that scale non-linearly.

I am an orthopedic trauma surgeon. I take call. I operate at 2 a.m. I have built a career in one of the highest-intensity, highest-availability specialties in medicine. Nothing in this analysis argues that a surgeon who does more work should not earn more money. That principle is sound. I believe in it. I have benefited from it. The argument is about the shape of the compensation curve.

In a linear system, a surgeon who works 80 percent of a full clinical schedule earns 80 percent of full compensation. That is not how most productivity-based models work. The surgeon working a full schedule with maximum call availability and full operating room (OR) block utilization does not earn 1.25 times what the 80 percent surgeon earns. That surgeon earns 1.5 or 2 times as much, because compensation is tracking a bundle. OR block access accumulates with seniority and availability. Referral networks compound around surgeons who are consistently present. Ancillary revenue from surgery centers and imaging flows to those with the highest throughput. Partnership tracks and leadership roles select for uninterrupted trajectories. Call availability signals commitment, which feeds case volume, which feeds everything else.

The curve is convex. The architecture amplifies small differences in participation into large differences in compensation. Our systems do not reward productivity linearly. They reward a specific pattern of total professional availability that compounds over a career, and that pattern is not distributed evenly by gender.

How the math works

Consider two orthopedic surgeons who finish fellowship in the same year and join the same group. Surgeon A maintains a full operative schedule and takes regular call for the first ten years. Surgeon B reduces to 80 percent for three years after the birth of a child, dropping one OR day per week and reducing call frequency.

In a linear system, Surgeon B would earn roughly 80 percent of Surgeon A’s compensation during those three years, then return to parity. The lifetime earnings gap would be small and temporary. In the actual system, the damage compounds. During the reduced years, Surgeon B loses a full OR block day, which gets reassigned. Referral patterns shift. When Surgeon B returns to full time, the lost block time does not come back automatically; it must be re-earned through seniority and availability. The referral network has already consolidated elsewhere. If the group has a partnership track, the clock may not pause. Three years at 80 percent can produce a 30 to 40 percent lifetime earnings gap. Not because of a discriminatory decision. Because the compensation architecture compounds the interruption at every layer.

Three structural mechanisms

• Temporal flexibility penalties: Women physicians are more likely to seek schedule modifications during childbearing and early parenting years. In a non-linear RVU system, the penalty is amplified and often permanent. Reduced availability during early career years affects partnership eligibility, OR block access, panel size, and referral volume. The gap does not close when hours normalize because the structural advantages lost during the interruption do not come back on their own.
• Specialty stratification: Women are disproportionately represented in lower-RVU specialties; men are overrepresented in high-RVU procedural fields. This is often framed as a preference issue. Goldin’s lens reframes it as a compensation architecture question: why does the payment system assign the relative values it does, and what process produced those ratios?
• Career interruption compounding: The pay gap widens sharply after the birth of a first child. Training overlaps with peak childbearing years. Compensation escalators, whether explicit like partnership tracks or implicit like referral network accumulation, penalize interruption in ways that do not self-correct.

Architected, not organic

In professional sports, pay differences between men’s and women’s leagues track consumer demand. Medicine is different. Physician compensation is set by policy decisions. The Relative Value Scale Update Committee (RUC), a physician panel historically dominated by procedural specialists, recommends relative values for CPT codes to the Centers for Medicare and Medicaid Services (CMS). MedPAC has documented the committee’s outsized influence on the fee schedule. Payers build on those values. Hospitals build compensation models on payer revenue. Each layer encodes assumptions about what work is worth.

We built this system. The RUC process, the RVU structure, call compensation design, productivity bonus formulas: these are human decisions, and they can be changed.

The budget neutrality problem

This argument is sometimes reduced to “pay primary care more and proceduralists less.” That is not the point, and that framing alienates the people whose buy-in any reform requires. The RUC process has produced a relative value scale encoding assumptions about complexity, skill, and time that have not been seriously reexamined in decades. Whether a 15-minute procedure should be valued at several multiples of a 40-minute visit involving diagnostic reasoning, risk stratification, and longitudinal care coordination is a legitimate calibration question.

But there is a hard constraint most policy discussions avoid. Medicare’s physician payment pool is functionally fixed. Relative values are scored against each other. When CMS increased evaluation and management (E/M) values in 2021, the conversion factor dropped and every procedural specialty saw per-unit reimbursement decline. Any meaningful recalibration is a redistribution within a closed system. Physicians see through euphemisms for this immediately.

The structural critique in this essay does not depend on recalibration. The better interventions operate on the shape of compensation within specialties, not between them. These are contract-level and institutional design changes, not federal budget fights.

Structural remedies

Goldin’s work points toward a specific class of solutions: reduce the non-linear penalty for flexibility. Applied to medicine, that means intervening at the level of compensation design.

• Flatten non-linear RVU curves: Calibrate productivity models so that working 80 percent of a full schedule produces closer to 80 percent of full compensation, not 60 percent. A concrete step: require health systems to publish the RVU-to-compensation taper schedule in employed-physician contracts, making the shape of the curve visible and negotiable.
• Increase substitutability: Pay gaps narrow in professions where one professional can substitute for another without loss of value. Medicine resists this through panel ownership, surgeon-specific referral networks, and individual productivity tracking. Team-based models, shared coverage pools, and standardized workflows reduce the penalty for any individual stepping away.
• Decouple income from continuous presence: Outcome-based payment and capitated models reduce the link between physical availability and earnings. A caveat: research on blunt value-based payment suggests poorly designed models carry their own risk. Crude capitation can flatten individual excellence into statistical noise. The goal is to make sure compensation metrics distinguish clinical excellence from mere availability.
• Reduce the compounding penalty for career interruption: Shared call models, time-banked compensation, and re-entry pathways for OR block access would limit the permanent earnings damage from temporary schedule modifications. CMS could pilot a re-entry RVU adjustment for physicians returning from parental or medical leave, smoothing the productivity trajectory rather than penalizing the gap.

An engineering problem

The discrimination narrative, while valid, has produced decades of advocacy with limited structural change. It is hard to design policy around implicit bias. It is far easier to design policy around compensation formulas. Goldin’s framework shifts the question from “Why are women physicians paid less?” to “How did we design a compensation system that converts small differences in time allocation into large differences in lifetime earnings?” Medicine controls nearly every input that produces the gap. We set the relative values. We design the compensation models. We structure the call schedules. If the system produces disparate outcomes, the system can be changed.

The physician gender pay gap is an engineering problem. We have the blueprints.

Michael Suk is an orthopedic trauma surgeon and physician executive.