Should Med, Law, and MBA Students Save Money? The Case for Lifecycle Consumption Smoothing

The most common rule in personal finance is “save 15% of your income from your first paycheck and never stop.” That advice works well for most steady-income earners. It works badly for medical, law, and MBA students whose income today is near zero but whose expected income after graduation is high. Forcing a 15% savings rate on a third-year medical student earns nothing in compounding while costing them sleep, health, family stability, and sometimes the degree itself.

Economists have a name for the alternative: lifecycle consumption smoothing. The idea is that the right financial target is not a constant savings rate but a reasonably stable standard of living over time. In a world where income is low when young, higher in midlife, and zero in retirement, the optimal plan can involve saving little or even borrowing early, then saving aggressively later when income arrives. The catch, as Yale’s James Choi documents, is that this elegant theoretical answer often clashes with the realities of willpower, uncertainty, and the difficulty of suddenly becoming a super-saver later in life. What follows lays out the academic case, applies it to medical, law, and MBA students with real outcome data, names the risks, and proposes a middle ground: “smooth the floor, not the fantasy.”

Where popular advice and academic finance disagree

Choi’s 2022 paper in the Journal of Economic Perspectives surveys the fifty most popular personal finance books and compares each prescription against normative academic models. The bottom-line finding: popular advice and academic models frequently disagree, and both have a point. Popular advice tends to recommend constant high savings rates, simple rules, and broad debt avoidance. Academic models often recommend variable savings rates that respond to lifetime resources, with low or zero saving when income is low and high saving when income is high.¹

Choi’s charitable read of popular advice is that it accounts for two things academic models often ignore: limited willpower and limited computation. Telling a 25-year-old to optimally smooth marginal utility across her lifetime is a math problem most people cannot solve and a behavior most people cannot maintain. Telling her to save 15% from every paycheck is wrong in some utility-maximizing sense but right enough often enough to be useful.¹

The lifecycle framework itself comes from Modigliani and Brumberg (1954) and Friedman (1957). Both formalized the same intuition: a rational household sets consumption based on lifetime resources, not today’s paycheck. Income is typically low when young, peaks in midlife, and falls to zero in retirement; consumption (the right target) should be relatively flat across that arc. The savings rate is just the residual: today’s income minus today’s consumption.

Why uncertainty changes the textbook answer

Christopher Carroll’s buffer-stock saving model is the most important correction to the textbook lifecycle answer. When labor income is risky, households rationally maintain a liquid cushion and avoid borrowing aggressively against future income, because bad income outcomes are possible. The result is a consumption path that looks much more sensitive to current income than the pure lifecycle model predicts. Households behave less like infinitely-foresighted optimizers and more like buffer-stock savers protecting themselves against shocks.²

Gourinchas and Parker estimated this empirically and found that young households often act like buffer-stock consumers early in their working lives, then shift toward retirement accumulation around midlife. The shift is not because households become wiser; it is because the relative weight of retirement risk grows as retirement approaches.³

For a professional student, the buffer-stock framing is the right starting point. Future income is high in expectation but uncertain in realization. Borrowing more today only makes sense if the future income is unusually credible, the debt terms are survivable, and the money funds completion, stability, health, or family logistics rather than lifestyle inflation.

The behavioral bridge: Save More Tomorrow

Thaler and Benartzi’s “Save More Tomorrow” program offers a practical bridge between lifecycle theory and real self-control. Rather than demanding maximum saving immediately, participants pre-commit to escalating their savings rate when future raises arrive. In one early implementation, average saving rates rose from 3.5% to 13.6% over 40 months. The mechanism works because it harnesses two well-documented biases (loss aversion and present bias) in service of long-run wealth, instead of fighting them.⁴

For professional students, this is the operational model. Borrow humanely during school. Pre-commit that the first attending paycheck, the first BigLaw bonus, or the first post-MBA raise will flow into debt repayment, retirement contributions, and an emergency fund before lifestyle expands. The pre-commitment matters because the moment income arrives is the moment the temptation to upgrade lifestyle peaks.

Why this matters for professional students specifically

Professional students are the cleanest real-world example of lifecycle consumption smoothing. Their current income is near zero or negative. Their expected post-graduation income is high enough that extreme deprivation today is not actually optimal. The catch is that “expected income” is not the same as “guaranteed income.” The next three sections work through medicine, law, and business school individually because the income distributions, completion risks, and debt structures are materially different.

Medical school: the strongest case for smoothing

Medicine is the textbook example. Training is long and expensive, but post-training income is high and relatively stable. AAMC’s most recent fact card reports that 70% of the class of 2025 graduated with education debt, with a median education debt of $215,000 among indebted graduates. Among graduates with debt, 85% owed $100,000 or more, 59% owed $200,000 or more, and 28% owed $300,000 or more. Sixty-five percent of graduates were planning to enter a loan-forgiveness or loan-repayment program.⁵

Physician earnings are unusually high relative to most occupations. The Bureau of Labor Statistics reports that physicians and surgeons had a median annual wage greater than or equal to $239,200 in May 2024. (BLS does not publish a median above $239,200; the actual median is higher.) The path to that income requires four years of medical school plus three to nine years of residency, with possible fellowship afterward.⁶

For most medical students, modest borrowing for safe housing, nutritious food, board exam preparation, transportation, and family stability is rational. The expected income gap is wide enough and the completion path is well-defined enough that lifecycle smoothing meaningfully improves lifetime well-being. The caveats: residency income is much lower than attending income, interest can compound for a decade or more, specialty choice changes the income trajectory by a factor of two or three, burnout and attrition are real, and PSLF eligibility depends on employer choice. The smoothing argument is strongest for a med student headed for an in-demand specialty at a non-profit hospital with a credible PSLF path and weakest for one targeting a long fellowship in a low-income subspecialty in private practice.

Law school: bimodal and dangerous

Law school is much more bimodal than medicine. NALP’s Class of 2024 data, summarized by LawHub, reports that 81.4% of 2024 ABA graduates obtained long-term, full-time legal jobs, while 9.7% were employed in part-time or short-term jobs, pursuing additional degrees, or unemployed and seeking.⁷

The bimodal salary distribution is the harder problem. NALP reports a record overall median salary of $95,000 for the Class of 2024, with a six-year stretch of consecutive increases. The full distribution shows two peaks: a left-side cluster of salaries from $55,000 to $100,000 that accounted for 53% of reported salaries, and a sharp right-side peak at large-firm starting salaries (4.4% of reported salaries at $215,000 and 18.7% at $225,000). Most law graduates are in the left peak. The Bureau of Labor Statistics reports a median lawyer wage of $151,160 in May 2024, with the bottom decile under $72,780 and the top decile over $239,200, hiding the early-career bimodality that matters most for debt decisions.⁸⁹

This means lifecycle smoothing is rational for some law students and dangerous for others. A student at a highly ranked school with strong BigLaw placement, a large scholarship, or a credible public-service path with PSLF is in a fundamentally different position from a student paying full freight at a lower-ranked program with weak placement statistics. Law school borrowing deserves to be underwritten like a high-variance investment, not like a guaranteed annuity.

MBA: the most program-specific

MBA economics are the most heterogeneous of the three. Altonji and Zhu’s 2025 NBER working paper estimates causal earnings returns for 121 graduate degrees using Texas administrative records. Across the 18 most popular programs, MD delivered an average earnings return of 110%, JD 59%, and MBA 16%. Earnings effects depend on undergraduate major (students from lower-paying majors benefit more from an MBA or JD), and school-specific returns are higher for higher-ranked JD and MBA programs.¹⁰

That dispersion is the central planning fact. A sponsored part-time MBA at a strong regional program for a working professional is a very different financial product from a full-price two-year MBA at a lower-placement school for a career-changer. Treating “an MBA” as a single asset class is the most common analytical mistake in this category. Lifecycle smoothing can be defensible at a top full-time MBA with a credible BigConsulting, BigBank, or BigTech path, especially for someone leaving a lower-paying field. It is much harder to defend at a full-freight program where the cost-adjusted return is in single digits.

The risks worth taking seriously

Four risks deserve explicit attention, because each one breaks the smoothing argument differently.

1. Future income is uncertain. Carroll’s buffer-stock logic is the right discipline. Even when expected income is high, bad outcomes are possible: non-completion, licensing failure, residency match failure, recession, immigration constraints, health shocks, family shocks, burnout. The plan should survive the bottom decile of outcomes, not just the median.

2. Debt terms matter enormously. Federal student loans, income-driven repayment plans, PSLF eligibility, private loans, and credit cards are not interchangeable. AAMC reports federal Direct Unsubsidized loans for graduate and professional borrowers at 7.94% for AY 2025-26 and Direct PLUS loans at 8.94%.⁵ Private loans run higher and lack federal protections; credit-card debt is far worse. The smoothing argument that works for federal student debt at 8% does not work for credit-card debt at 22%. See also debt avalanche vs. snowball for the post-graduation payoff math.

3. Federal policy is changing. The One Big Beautiful Bill Act and related Department of Education implementation introduce new federal student-loan rules effective July 1, 2026. Graduate students face an annual cap of $20,500 and a lifetime cap of $100,000. Professional students face an annual cap of $50,000 and a lifetime cap of $200,000. A universal lifetime borrowing limit of $257,500 applies across all federal loans. Grad PLUS loans are discontinued for new borrowers (current borrowers get up to a three-year teach-out to complete their programs).¹¹ Anyone planning a debt-funded path that exceeds these limits needs to confirm what private lenders charge and whether the math still works.

4. Behavioral risk often dominates the spreadsheet. Choi’s point cuts both ways. If a student smooths consumption by borrowing more during school but never actually ramps up savings later, the consumption was not smoothed; it was front-loaded and the pain was pushed into the future. The Save More Tomorrow evidence shows the ramp can work, but it has to be planned and automated. A vague intention to “save aggressively when I start earning” is not a plan.⁴

Run the math on your own numbers

The interactive calculator below quantifies the buffer-stock argument for your specific case. Set your years of low income, your monthly living expenses during school, your total tuition, your expected attending income, and your loan rate. The output shows your debt at graduation, your monthly debt service on a 10-year standard amortization, and your debt-service share of after-tax attending income under both expected and P10 income scenarios.

The middle ground: smooth the floor, not the fantasy

The right framing is not “borrow as much as possible because future income will inflate the debt away.” It is also not “save 15% from year one, no matter what.” The middle ground is to smooth a humane consumption floor while preserving optionality and pre-committing to ramp savings later. Five practical rules.

1. Borrow first for completion probability. Tuition, required fees, exam costs, safe housing, food, transportation, childcare, and basic medical needs are legitimate uses of student debt. Each one increases the probability of finishing the degree, which is the precondition for the entire smoothing argument working.
2. Avoid high-interest consumer debt. A federal student loan at 8% is a different instrument from a credit card at 22%. The latter is almost never a good smoothing tool.
3. Stress-test the downside. Run the plan under delayed graduation, lower-paying specialty match, no BigLaw offer, no PSLF eligibility, and a bad first-year market. If the plan still works, the smoothing argument has held. If it does not, the plan was leaning on the median outcome, not the downside.
4. Cap lifestyle debt. Borrowing for “I need to survive school” is different from borrowing for “I want to live like my future attending or partner self.” Premium apartments, frequent restaurant meals, expensive cars, and recurring lifestyle commitments that cannot easily be reversed do not belong in the smoothing argument.
5. Pre-commit to ramping savings later. The first attending paycheck, the first BigLaw raise, or the first post-MBA bonus should automatically flow into debt repayment, an emergency fund, retirement contributions, and taxable investing before lifestyle expands. Save More Tomorrow shows this works when it is automated; it usually fails when it is left to willpower.⁴

Frequently asked questions

Should I save anything during medical school?

Probably not in retirement accounts. With near-zero income, you have no marginal tax rate to defer against, and any cash you save comes from extra borrowing at 7-9% loan rates. The exception is unmatched employer contributions if your spouse’s plan or a side job offers them. Building a small liquid emergency fund (a few thousand dollars) is reasonable; building a Roth IRA from borrowed money is usually not.

Does this argument apply to PhD students or PA/NP students?

Partly. PhD students are typically funded with stipends and tuition waivers, so the borrowing question is smaller; the income ramp is usually flatter (academic salaries grow slowly), so the smoothing case is weaker. Physician assistants and nurse practitioners have shorter, less-debt-heavy training paths and more uniform salary distributions, which usually makes the standard "save consistently" advice work fine.

What about IDR plans and PSLF?

Income-driven repayment plans cap monthly payments as a percentage of discretionary income, and Public Service Loan Forgiveness can forgive remaining federal balances after 120 qualifying payments in qualifying public-service employment. Both can materially change the smoothing math by reducing post-graduation debt service. The catch: the rules have changed repeatedly, eligibility depends on consistent qualifying employment, and policy risk is non-trivial. A plan that requires PSLF to work needs a backup plan.

I'm at a lower-ranked law school with weak BigLaw placement. Should I borrow less?

Probably yes. The Altonji-Zhu finding that returns are higher for higher-ranked JD programs is the key fact.¹⁰ Combined with the NALP bimodal-salary distribution, a lower-ranked full-price JD is a high-variance bet that the buffer-stock model treats unfavorably. Sources of capital with strong outcome contingencies (large scholarships, employer sponsorship, deferred admission with one or two years of paid work) reduce the downside.

Doesn't the new $200k aggregate cap on professional student loans solve the overborrowing problem?

It caps federal exposure but pushes borrowing into private markets for students whose schools cost more than the cap allows. Private lenders price for risk; rates are often higher and protections (IDR, PSLF, deferment, forgiveness) are absent. The new caps may restrain federal balances while shifting the same dollar of borrowing onto worse terms.¹¹

How is this different from `lifecycle-asset-allocation`?

That guide covers lifecycle asset allocation (how much equity to hold at each age, given that human capital is bond-like and large when young). This guide covers lifecycle consumption (how to time spending and saving across a career given that income is not flat). Different lifecycle questions, both grounded in the same underlying framework.

Related guides

• Lifecycle Asset Allocation covers the asset-allocation side of lifecycle thinking: when human capital is bond-like and large, it changes the optimal equity share. Different lifecycle question, complementary framework.
• FIRE Calculator assumes a constant savings rate. This guide explains when that assumption is wrong and what to do instead.
• The Complete Guide to Financial Independence covers the FIRE framework that this guide’s buffer-stock correction sits on top of.
• The Order of Investing Operations explains the default savings waterfall. This guide explains why professional students start the waterfall differently.
• Human Capital Risk for Tech Workers covers the equity-like, sector-correlated version of the human-capital framing. Trainee human capital is bond-like but deferred; the implications differ.
• Roth vs. Traditional uses the same current-vs-future logic at the tax-bracket level. Low income today and high income tomorrow argues for Roth, just as it argues for borrowing rather than saving.

Sources

1. Choi, J. J. (2022). “Popular Personal Financial Advice versus the Professors.” Journal of Economic Perspectives 36(4), 167-192. Yale School of Management. Surveys the 50 most popular personal finance books and compares each prescription to normative academic models.
2. Carroll, C. D. (1997). “Buffer-Stock Saving and the Life Cycle/Permanent Income Hypothesis.” Quarterly Journal of Economics. Buffer-stock consumption model under labor-income uncertainty.
3. Gourinchas, P-O. & Parker, J. A. (2002). “Consumption Over the Life Cycle.” Econometrica 70(1), 47-89. Empirical estimates of the age-related shift from buffer-stock to retirement accumulation.
4. Thaler, R. H. & Benartzi, S. (2004). “Save More Tomorrow: Using Behavioral Economics to Increase Employee Saving.” Journal of Political Economy 112(S1). Average savings rate rose from 3.5% to 13.6% over 40 months under the pre-commitment design.
5. Association of American Medical Colleges. Medical Student Education: Debt, Costs, and Loan Repayment Fact Card for the Class of 2025. 70% of graduates with education debt; $215,000 median among indebted graduates; 65% planning loan-forgiveness or repayment program; federal Direct Unsubsidized 7.94% / Direct PLUS 8.94% for AY 2025-26.
6. U.S. Bureau of Labor Statistics. Occupational Outlook Handbook: Physicians and Surgeons. Median wage greater than or equal to $239,200 (May 2024); training described as four years of medical school plus three to nine years of internship and residency.
7. LawHub. 2024 Graduate Job Outcomes, Aggregated and by School. 81.4% of 2024 ABA graduates obtained long-term, full-time legal jobs; 9.7% in part-time / short-term jobs, pursuing additional degrees, or unemployed and seeking.
8. National Association for Law Placement. Class of 2024 Selected Findings. Record overall median salary of $95,000; bimodal distribution with 53% of reported salaries between $55,000 and $100,000; right-side peaks at $215,000 (4.4%) and $225,000 (18.7%).
9. U.S. Bureau of Labor Statistics. Occupational Outlook Handbook: Lawyers. Median annual wage $151,160 (May 2024); bottom decile under $72,780; top decile over $239,200.
10. Altonji, J. G. & Zhu, Z. (2025). “Returns to Specific Graduate Degrees: Estimates Using Texas Administrative Records.” NBER Working Paper 33530. Across the 18 most popular programs: MD ~110%, JD ~59%, MBA ~16% earnings returns; school-specific returns higher for higher-ranked JD and MBA programs.
11. U.S. Department of Education and One Big Beautiful Bill Act implementation. Federal student loan changes effective July 1, 2026: graduate annual cap $20,500 / lifetime $100,000; professional annual cap $50,000 / lifetime $200,000; universal lifetime cap $257,500; Grad PLUS discontinued for new borrowers with up to a three-year teach-out for current borrowers.
12. Yale Insights. “Personal Finance: Popular Authors vs. Economists”. Yale School of Management interview/summary of Choi’s paper.