FOR STUDENTS

Your library closed
at 9pm.

Your case brief
doesn’t care.

A sector landscape in under fifteen minutes, from anywhere. Caspr settles the figures the sources disagree on and gives you the reason each one holds, so “it seemed reasonable” is never your answer.

  • Any hourNo campus, no VPN, no download limit
  • The whyBehind every number a judge will question
  • 15 minAn industry primer, not half a day

No account needed. No subscription.

Your files stay in your account and never train a model. ISO 27001:2022 certified →

THE PROBLEM

The databases are on campus.
You are not.

  • Closed when
    you need it.

    The library databases work on campus, during opening hours, until the download limit hits. The case competition runs overnight, and you are in your apartment.

  • “It seemed reasonable”
    is not an answer.

    Judges and interviewers ask where a figure came from, and then how it was built. A number you cannot explain costs you the room.

  • Scaffolding before
    the contribution.

    The market context chapter and five industry primers come before the part that is yours. Each takes days when built from scratch, in the weeks that matter most.

HOW IT WORKS

From brief to
working context.

In under an hour.

  1. Describe the question

    One line: the market, the case, what the answer must show. Caspr proposes the structure and asks what it needs to know.

    🔒 Anything you upload is encrypted at rest and in transit. It stays in your account. It is never used to train any model. Full security posture →

  2. Caspr reads the market

    25M+ curated sources. For each figure that matters, Caspr goes underneath it, sets the sources against each other and settles the one to use for your case, with the argument.

  3. You write

    Finished and ready to build on: a PDF or an editable PPTX, the figures settled, with a References list for your bibliography. The contribution is yours.

ANALYTICAL AI, NOT GENERATIVE

Every source, credible. Every claim, triangulated. Every report, defensible.

Generative AI is trained to understand how people communicate.
Caspr is proprietary Analytical AI, built from the ground up for business professionals making high-stakes decisions.

Academic Research · Study

US Minimum Wage Employment Effects: Evidence Review

Why do the employment estimates differ, and which should a graduate researcher use?

US literature · the early 1990s to the present · For a graduate researcher choosing an estimate and a design

October 2026

Ten sections, ten findings

Sections 1 · Scope, 2 · Framework, 4 · Panel data, 5 · Border designs, 8 · Syntheses In red, the finding this deck follows next. Select any section to go to it. 2

Published US estimates of the employment effect run from about zero to −2.0

Exhibit 1 · employment elasticity, by study design

Border county pairsDube, Lester and Reich, 2010

≈ 0

Median of journal studiesDube and Zipperer, 72 studies

−0.13

Statewide panelConventional fixed effects, same paper

−0.787

SeattleJardim et al., 2022

−0.2 to −2.0

Four designs, and what looks like four answers to one question

Sections 1 · Scope, 2 · Framework Source: Dube, Lester and Reich; Dube and Zipperer; Jardim et al. Caspr analysis, sections 1, 2.2 and 8. 3

The founding dispute was mostly measurement: survey counts against payroll records

Card and Krueger, survey of 410 stores

+2.4

FTE workers per New Jersey store relative to Pennsylvania, after the 1992 increase

Neumark and Wascher, 230 restaurants

Negative

relative employment effect for New Jersey in payroll records of the same episode

Card and Krueger, payroll reply

No loss

of employment, in BLS ES-202 payroll records

On April 1, 1992, New Jersey raised its minimum wage to $5.05, against $4.25 before; Pennsylvania did not. One law produced positive, negative and near-zero estimates as survey and payroll frames, comparison samples and FTE construction changed. Read it as a measurement and sample dispute, with some residual replication disagreement.

Sections 3 · Foundational cases, 4 · Panel data Source: Card and Krueger; Neumark and Wascher. Caspr analysis, sections 3.1, 3.2, 3.3, 4.2 and 9.3. 8

In Seattle, low-wage hours fell 6–7% while restaurant studies find no job loss

Hourly wages, low-wage jobs

+3%

after the rise to $13 in 2016, in Washington administrative records

Hours, low-wage jobs

−6 to −7%

in the $13 phase: work time contracted

Monthly pay per low-wage job

−$74

the net of higher wages and fewer hours

The University of Washington team measures hours and pay in all jobs below a low-wage threshold; the Berkeley studies measure restaurant employment, in Seattle and six high minimum-wage cities, and find effects close to zero. A difference of estimand, not a contradiction: once bite is large, the answer depends on whether the study counts headcount, hours, earnings or sector.

Section 6 · Seattle Source: Jardim et al.; Reich, Allegretto and Godoy; Allegretto et al. Caspr analysis, sections 6.1, 6.2 and 9.1. 9

Across 138 state increases, jobs missing below the new floor reappear just above it

Cengiz et al., CPS wage bins

138

state minimum wage changes, 1979 to 2016: little net change in low-wage employment

Bite of those changes

37–59%

of the median wage: lower-bite evidence, short of where job losses became sizable

Bunching turns the employment question into a mass balance inside the affected wage range: wage compression is the mechanism, limited net job loss the inference. Restaurant workers, teens and jobs near the floor face different exposure, so they are not samples of one estimand, and these changes should not be pooled with Seattle-style ordinances.

Sections 7 · Bunching, 9 · Divergence Source: Cengiz, Dube, Lindner and Zipperer, 2019. Caspr analysis, sections 7.1, 7.2 and 9.2. 10

Since 1994, designs moved the question from the sign of the effect to its margin

Exhibit 3 · selected studies and syntheses
  1. 1994Card and KruegerNo job loss in New Jersey fast food
  2. 2006Neumark and WascherReview: negative effects common in teen studies
  3. 2010Dube, Lester and ReichBorder counties: restaurant effects near zero
  4. 2019Cengiz et al.138 state changes: little net job loss
  5. 2024Dube and ZippererMedian of 72 journal studies: −0.13

Reviews and synthesesSingle-design studies

Sections 3 · Foundational cases, 4 · Panel data, 5 · Border designs, 7 · Bunching, 8 · Syntheses Source: Card and Krueger; Neumark and Wascher; Dube, Lester and Reich; Cengiz et al.; Dube and Zipperer. Caspr analysis, sections 3, 4.1, 5, 7 and 8.1. 11

Most disputes are method; the genuine one is bite and labor-market structure

DisputeMostlyWhat drives it
Card and Krueger against Neumark and WascherArtifact, with some residual replication disagreementData source, sample frame, FTE construction
State panels against border countiesMethodological artifactCounterfactual trend bias in distant controls
Seattle: University of Washington against BerkeleyArtifact in estimand, partly sectoral scopeHours against headcount; all low-wage jobs against restaurants
Border and bunching evidence against high-bite casesGenuine substantive heterogeneityBite and labor-market structure
Section 9 · Divergence Source: Caspr analysis, section 9.3. 12

For a graduate researcher: fix the estimand first, then the estimate and the design

Your research questionEstimate to useDesign
Conventional state increases−0.13, inside 0 to about −0.2Group-time ATT with local controls
Mechanisms near the wage floorNear-zero net effect, with wage compressionBunching or wage-bin event study
One high-bite cityAn episode-specific effect, not the national anchorAdministrative event study; hours, hiring and earnings first
Monopsony and market structureAverage near the baseline; slope variesInteract treatment with concentration and bite
Section 10 · Design and the anchor Source: Caspr analysis, sections 10, 11.2 and 11.3. 13

Not an assistant. An analyst.

A Study, as it arrives: the figures settled and the case concluded, usually within the hour.

THE STUDENT PROGRAM

Start free.
Earn more as you go.

Verify your university email and the first $150 is on Caspr.

  • START HERE

    Try

    Free. $150 of research on us. Verify your university email. No card.

    • Student prices: Brief $8 · Study $40
    • Ask Caspr about any finding
    • Attach your readings to an analysis
  • YOUR CLASSMATES

    Refer

    Earn up to $200 of research. Spent before your own balance, valid 180 days.

    • + $15 when a classmate joins
    • + $10 when they run an analysis
    • + $10 for 10 Share Card readers
  • AFTER YOU GRADUATE

    Graduate

    $100 toward your first Solo budget. Your analyses move with you.

    • Standard prices from that day
    • Everything you ran stays yours
    • Credit valid 180 days

A university or student union can add its domain, and every student verifies automatically.

How student pricing works 

STUDENT PRICES

Same analysis.
Student prices.

Every analysis below runs at student prices: an Academic Research Brief is $8 and a Study $40. No subscription, and the same analysis every other account gets.

WHY STUDENTS TRUST CASPR

For work that will be questioned.

  • The analysis is done before it arrives.

    Caspr goes underneath every figure that matters, sets the credible sources against each other and tells you which number to use, and why. Nothing is left to find, reconcile or assemble. The sources are there if you want them.

  • Built by the people who set the standard.

    Joy, Caspr’s Founder and CEO, defines what the analysis must do and what good work looks like. Jayant, the Co-Founder and CTO, leads the research that builds the models that do it.

  • Try it on your next assignment first.

    Verify your university email and the first $150 is on Caspr. No card. No subscription.

FREQUENTLY ASKED QUESTIONS

Answered before you have to ask.

Can I use AI research tools for my dissertation?

Most programs that permit it require you to disclose it, and many leave the rule to the department. Check your program’s rule and confirm with your supervisor. The report is marked as generated by Caspr.

How do I cite data from a Caspr report?

Cite the underlying source, exactly as the report shows it. Every table and chart has a source line, claims carry a citation where they are made, and the References list gives the full works. For academic work the report can cite inline, by author and date.

Is AI allowed in case competitions, and will judges accept it?

Rules differ, so read the brief. Investment competitions usually restrict entrants to public information, which is what Caspr works from. Judges look for a number you can explain: Caspr settles which figure fits your case and states how it was built. The file is marked as generated by Caspr.

Can I run an analysis on a student budget?

Yes. An Academic Research Brief is $8 and a Study $40, and the first $150 is on Caspr, with no card.

Are my files private?

Yes. Anything you attach stays in your account, is encrypted at rest and in transit, and is never used to train any model.

Caspr means Business.

GET STARTED

The case brief, before the library opens.
Your first $150. On us.

Verify your university email. No card.

Run the analysis See a sample report →
Run an analysis