Public Sector Consulting Case Interview: Government Guide

Updated Nov 11, 2026By Esteban Ronsin · Founder · former Deloitte and Accenture Strategy consultant

A public sector consulting case interview is a 30-to-45 minute case in which you advise a government agency, and public value and political feasibility take the place of profit as the test of a good answer. Booz Allen Hamilton, Deloitte, McKinsey, and BCG build these cases around policy impact, service delivery, and social value rather than EBITDA and shareholder return. This guide covers the frameworks, the case types, a fully worked government case with real math, a walkthrough of McKinsey's published Diconsa practice case, the social ROI calculation, and what each major firm tests.

Preparing for a government consulting interview?

Road to Offer lets you practice structured cases and get AI feedback on your framework, stakeholder mapping, and quantitative reasoning.

Try your first case free
PESTLE visual for public sector cases showing policy, economics, stakeholders, technology, legal constraints, and environment

What Is Public Sector Consulting and Who Hires For It?

Public sector consulting covers work for government agencies, multilateral institutions (World Bank, IMF, UN), public utilities, and state-owned enterprises. The work spans strategy, operations, technology implementation, program evaluation, and policy design.

The market is large and durable, though its exact size depends on definition. Mordor Intelligence sizes global public sector consulting and advisory at roughly $31 billion in 2025, while Market Research Future, using a broader scope, puts it near $76 billion. Both agree on the shape: low-single-digit growth and steady, less-cyclical demand than commercial strategy.

The firms that hire heavily split into three groups:

  • Pure-play government firms. Booz Allen Hamilton is the largest, reporting $11.98 billion in revenue for fiscal year 2025 across Defense ($5.9B), Civil ($4.2B), and Intelligence ($1.9B), with a record $37 billion backlog. Guidehouse and Accenture Federal Services also concentrate here.
  • MBB public-sector groups. McKinsey, BCG, and Bain serve governments through dedicated groups alongside their commercial work, emphasizing transformational strategy.
  • Big Four and broad-line firms. Deloitte Government and Public Services, EY, and KPMG run large public sector advisory teams focused on operations, technology, and compliance.

Engagements often run longer than commercial strategy projects, client relationships are more durable, and federal or defense roles can require security clearance eligibility, so mission alignment and stability tend to weigh more heavily.

Public sector practice cases at a glance

How a public sector case works: you optimize for service outcomes and public value rather than profit, and weigh stakeholders and budget limits.

Start free now: National Health System is a service improvement case: the Sydney Department of Health has found that people over 60 are dissatisfied with the time and quality of care. It has 8 questions, 1 with math, 1 exhibit, and fewer than 10 completed runs so far, so no median time yet. An AI interviewer runs it out loud, shows each exhibit when you reach it and scores the debrief. Your first full case is free when you create an account.

Case (style, industry)Quant difficultyQualitative difficultyThe math you runMedian time
National Health System (Independent, health services). 8 (1 with math), 1 exhibit.
easymediumOlder share, visits, specialists neededFewer than 10 runs
Quahog Public Schools (McKinsey, public sector). 6 (1 with math), 2 exhibits.
mediumhardBook shortage addressed and remainingFewer than 10 runs
Voter Turnout in Philadelphia (McKinsey, public sector). 8 (2 with math), 2 exhibits.
easyhardRegistration and turnout percent change, exhibit readFewer than 10 runs
Colombian Hippos (McKinsey, public sector). 7 (3 with math), 1 exhibit.
mediumeasySanctuary five year net, Duke lease netFewer than 10 runs

All four are adapted from consulting club casebooks (INSEAD, Chicago Booth, Wharton and Duke Fuqua). More are in the case library, 92 cases in all. Every debrief scores seven categories: structure (25% of the score), quantitative analysis (20%), business judgment (15%), and problem understanding, exhibit reading, communication and synthesis (10% each), with feedback on every question. Median times come from completed runs between June 25 and September 22, 2026, and count runs, not people.

How Government Cases Differ From Private Sector Cases

The format of a government case is identical to a private sector case: you receive a scenario, structure your approach, analyze data, and deliver a recommendation. The content is where the differences appear.

DimensionPrivate SectorPublic Sector
Objective functionMaximize profit / shareholder valueMaximize public value / citizen outcomes
Success metricsEBITDA, revenue growth, ROI, market shareCost per person served, policy adoption rate, social ROI
Primary stakeholdersCEO, board, investorsMinister, legislature, citizens, oversight bodies
Decision constraintsMarket forces, competitive dynamicsPolitical cycles, regulatory approvals, budget appropriations
Timeline pressureQuarterly earnings cyclesAnnual or multi-year budget cycles
Implementation complexityOrganizational changePolitical economy, legislative process, public trust

The most important shift is in how you define the right answer. In a private sector case, the recommendation with the highest NPV is directionally correct. In a public sector case, the highest-impact intervention may be politically unfeasible or constitutionally constrained. A strong answer addresses feasibility across three dimensions at once: effectiveness (will it work?), authorization (can it be approved?), and capability (can it be implemented?). That is the basis of Moore's Public Value Triangle, covered below.

How Do You Walk Through a Public Sector Case Live?

The mechanics mirror a standard case, with two public-sector adaptations. Use this sequence:

  1. Clarify the objective in public-value terms. Ask what success looks like for citizens, not just for the budget. "Are we optimizing for coverage, equity, or cost per person served?"
  2. Structure with your own issue tree first. Build a MECE tree for the specific problem. Do not open with a named framework. PESTEL and Moore's Triangle are lenses you apply onto the tree, not the tree itself.
  3. Run a 90-second PESTEL scan to flag which external factors are hard constraints versus background context.
  4. Drive the analysis, doing the quantitative work where the data points. Government cases contain real math (see the worked example below).
  5. Pressure-test against Moore's Triangle before you recommend: is it valuable, authorized, and executable?
  6. Synthesize with a named stakeholder. Lead with the recommendation, then name who must authorize it and how you would manage the most likely opponent.

If you want a deeper reference on the universal mechanics, see the case interview prep guide and the case interview frameworks complete guide.

The PESTEL Framework for Government Cases

The profitability framework and market entry framework are starting points for private sector analysis. For government cases, PESTEL (Political, Economic, Social, Technological, Environmental, Legal) is the primary environmental scan.

PESTEL is not a replacement for an issue tree. It is a structured read of the external environment that precedes your analytical framework. In a government case the political and legal letters often decide what is possible at all, so read them first.

PESTEL Framework for Government Cases

  1. 01

    Political. Political will, election cycles, ministerial priorities, coalition dynamics, opposition risks

  2. 02

    Economic. Budget constraints, fiscal position, macroeconomic context, funding sources (domestic vs. donor)

  3. 03

    Social. Demographics, public opinion, cultural context, equity and inclusion considerations

  4. 04

    Technological. Digital infrastructure, data availability, implementation readiness, cybersecurity

  5. 05

    Environmental. Climate targets, resource constraints, environmental regulations, sustainability mandates

  6. 06

    Legal. Constitutional constraints, procurement law, data privacy regulations, international obligations

How to use PESTEL in a case: once your tree is drawn, scan the six letters and sort what you find into three piles. The first holds what will not move during the engagement, such as a statute or an election date. The second holds what you can plan around, such as budget timing or public sentiment. The third holds what you can exploit, such as a new shared data platform or a climate target that unlocks funding. Two or three items usually carry the answer, so name those, mention the rest in a line, and spend your time on the pile that blocks the recommendation.

Moore's Public Value Triangle: The Three-Part Test Every Policy Must Pass

Mark Moore's Public Value Strategic Triangle (Harvard Kennedy School, 1995, updated 2013) is the foundational test for whether a government initiative is viable. It checks three dimensions at once:

  1. Public Value (societal benefit). Does the initiative deliver outcomes citizens and society genuinely value? Not what politicians say citizens want, but what evidence and direct engagement show citizens need.
  2. Authorization (political and legal legitimacy). Has the initiative secured enough political support and legal authority to proceed? Can it survive a change of government? Is it constitutionally permitted?
  3. Operational Capacity (implementation feasibility). Does the agency have the people, budget, technology, and organizational capability to execute? A brilliant strategy with 20% implementation capacity is not a viable recommendation.

A recommendation is only as strong as its weakest vertex. Say out loud which vertex is most at risk and what you would do about it: pilot first, phase the funding, or build the missing capability before launch.

McKinsey Diconsa Practice Case: What It Asks and How to Structure It

The McKinsey Diconsa case asks whether Diconsa, a chain of about 22,000 state-owned stores serving rural Mexico, could deliver basic financial services, starting with paying out government benefits. It is a practice case published on McKinsey's careers site, written for interview prep, not a record of a real engagement (the official case).

The setup. The client is the Bill & Melinda Gates Foundation, which wants to reduce extreme poverty and has asked McKinsey to design a basic financial-services offering for remote Mexican communities. Many rural residents have no bank account and collect benefits in cash at a limited number of state-owned bank branches, often far away, with crime a risk on the journey. The 22,000 Diconsa stores already sell food and essentials, and a warehouse and truck network keeps them stocked. The study would begin with benefit payments and could later add savings, bill payments, insurance and loans. The published case runs four questions: what to investigate, a savings calculation, a survey-and-store-data exhibit, and how to sell the services to a low-education, high-fraud population.

Question 1, how I would structure the investigation. Group the work into four branches, then test each against the three vertices of Moore's triangle:

  • Resident need and trust: do recipients want to collect benefits at a store, and would they trust it with cash and later with savings?
  • Store and network capability: cash handling, security, staff training, connectivity, and whether the warehouse and truck network can move cash safely.
  • Economics and partners: what the state-owned bank pays or saves per transaction, what a store earns, and who carries the float.
  • Rules and authorization: which regulator and which agency must approve, and what the bank's own concerns are.

Because the owner of the stores is the government, authorization looks easier than capability, but treat that as a hypothesis to test, not a fact. Resident need and trust maps to public value, store capability to operational capacity, rules and approvals to authorization, and the economics branch tests whether the offer pays for itself under all three.

Question 2, the savings math. The case gives you these inputs: a family spends 50 pesos a month on transport and food to collect benefits; a store would cut that by 30% for the 60% of rural families who live furthest from a branch and by 20% for the 40% who live nearer; 20% of Mexico's 140 million people are rural; half of those receive benefits; a family has four members. Two routes reach the same answer.

StepCalculationResult
Rural people140 million × 20%28 million
Rural people on benefits28 million × 50%14 million
Families on benefits14 million ÷ 43.5 million
Yearly cost per family today50 pesos × 12600 pesos
Blended saving rate60% × 30% + 40% × 20% = 18% + 8%26%
Saving per family per year600 × 26%156 pesos
Total yearly saving3.5 million × 156 pesos546 million pesos

Cross-check by splitting the families instead: today's total is 3.5 million × 600 = 2.1 billion pesos; 2.1 billion × (0.6 × 0.3 + 0.4 × 0.2) = 2.1 billion × 0.26 = 546 million pesos. Then add the so-what, which is my own read and not part of the published case: 156 pesos a year is about 13 pesos a month per family, so travel savings alone are a thin reason to build the service. A strong answer says the case rests on safety, time saved, and the later products such as savings accounts.

Questions 3 and 4. Question 3 shows survey scores and store data for three regions and asks what you notice. Look for the region that differs, then ask whether the store data could explain the gap. Question 4 asks how the bank could sell to people with little banking experience in areas with fraud. Group your answer around the two problems named, unfamiliarity and fear of fraud, and answer each with an action such as in-store demonstrations and a simple first product.

For more public-sector reps, use the four cases in the practice table above.

A Fully Worked Government Case: Clearing a Permit Backlog

This example uses an invented city so you can see structure, math and a decision working together.

Prompt. A mid-sized city's building-permit office is slow. Applicants wait about three months for a decision and the council wants the wait under 30 days within 15 months. The office has 10 reviewers. It receives 600 applications a month, and 40% of them arrive incomplete and need a second review. Each reviewer completes 70 reviews a month, and 2,100 reviews are waiting. A fully loaded reviewer costs $90,000 a year. What should the city do?

Structure. Split the problem into three branches: demand (how much review work arrives), capacity (how much the office can clear), and the backlog already waiting. Then test any fix against the public-value triangle.

The quantitative leg.

StepCalculationResult
Reviews demanded per month600 + (600 × 40%)840
Reviews cleared per month10 × 70700
Monthly shortfall840 − 700140
Wait today2,100 ÷ 700 × 30 daysabout 90 days

Each month the backlog grows by 140 reviews. A backlog of 700 at 840 reviews a month is a 25-day wait, so the city must also remove roughly 1,400 of the 2,100 waiting reviews.

Option A, hire only. Four more reviewers give 14 × 70 = 980 reviews a month against demand of 840, a surplus of 140, so clearing 1,400 takes 10 months. The hires start work right away, so the backlog shrinks from month one and the goal is met at about month 10. Cost: 4 × $90,000 = $360,000 every year.

Option B, checklist plus two hires. An online checklist blocks submission until required documents are attached and cuts incomplete applications from 40% to 15%. Demand falls to 600 + (600 × 15%) = 690. The checklist takes about three months to build, and until it launches demand stays at 840. Two hires lift capacity to 12 × 70 = 840, so during the build the backlog holds at 2,100 and the wait at about 75 days. After launch, demand is 690 against capacity of 840, a surplus of 150 a month, so clearing 1,400 takes 1,400 ÷ 150, about 9.3 months, and the goal is met at about month 12.3, when the backlog is near 700 and the wait is 700 ÷ 840 × 30 = 25 days. The backlog keeps shrinking after that if the city keeps both hires, so 25 days is simply where this plan stops. Cost: a $250,000 one-time build, then 2 × $90,000 + $40,000 upkeep = $220,000 a year.

Each piece on its own. Two hires alone give 840 of capacity against 840 of demand, so the backlog never shrinks and the wait stays near 75 days. The checklist alone leaves 700 of capacity against 690 of demand, a surplus of 10 a month, which would take over 10 years to clear the backlog. Neither piece reaches the target by itself.

Break-even. Year-one cost is $360,000 for A and $250,000 + $220,000 = $470,000 for B. B costs $110,000 more in year one and $140,000 less every year after ($360,000 − $220,000), so cumulative cost crosses over at about 1.8 years, roughly month 21. After that B is ahead.

Decision. Recommend Option B: it meets the 15-month goal at about month 12 (Option A is faster, at about month 10), its yearly cost is lower from year two, and it repays its extra year-one cost at about month 21. If the council holds the original one-year goal, Option A is the only one that meets it. Name the risks as well: the council must approve $250,000 of capital (authorization), the city's IT team must deliver the checklist in about three months (capacity), and a slipped build pushes the date out month for month, so the two reviewers should be hired first.

Synthesis. "I recommend the checklist plus two hires. It brings the wait from about 90 days to about 25 by around month 12, inside your 15-month goal, and its yearly cost is lower than hiring alone from year two on. The council needs to approve the $250,000 build, and I would start the hiring immediately so the backlog stops growing while the checklist is built." For more on delivering this final step, see case interview synthesis, and for the underlying arithmetic technique see case interview math practice.

Structure a McKinsey public sector operations case. Fifty escaped hippos in Colombia could reach 250 and threaten people and crops. Answer this McKinsey operations case step in the article. Browse more practice cases in the case interview library

Six Common Government Case Types

Government cases cluster around six recurring problem types. Each is a slight adaptation of a private sector equivalent.

Case TypeDescriptionKey Approach
Service delivery efficiencyServe 20% more citizens on the same budgetValue chain analysis + cost per unit benchmarking
Policy designShould the government implement Policy Y, and in what form?PESTEL + Moore's Triangle + stakeholder analysis
Program evaluationIs Program Z achieving its intended outcomes?Logic model + impact measurement + counterfactual
Digital transformationHow should the agency modernize its IT?Current-state map + risk assessment + roadmap
Budget optimizationWhere should the ministry reallocate $500M for impact?Portfolio analysis + Social ROI across programs
Public-private partnershipShould the government partner with a private provider?Feasibility + risk allocation + contract structure

For most of these, the analytical skeleton mirrors a private sector case. The difference is in the inputs (social outcomes instead of financial returns) and the constraints (political and legal rather than competitive). A digital transformation case inside a federal agency, for instance, uses the same current-state-to-roadmap structure but adds procurement law and legacy-system risk as binding constraints.

Social ROI: Quantifying Impact Without a P&L

The Social Return on Investment (SROI) methodology quantifies social outcomes in financial terms. It is the public sector counterpart to NPV analysis and pairs naturally with cost-benefit analysis when you are asked to value a government program.

Social ROI = (Total Social Value Created) ÷ (Total Investment)

A ratio of 4.2 means every $1 invested generates $4.20 in social value.

How to calculate it in a case: identify two or three measurable outcomes (graduates, readmissions prevented, citizens lifted above the poverty line), quantify the units, monetize each with government cost data or productivity estimates, sum the social value over the relevant horizon, total the investment (operating plus capital plus opportunity cost of public funds), and divide.

Worked example: early childhood education. A state is evaluating a $200M early-childhood expansion (10,000 additional children per year, 5-year program).

  • Outcome 1, graduation. Early childhood education raises high school graduation by roughly 8 percentage points. 10,000 × 8% = 800 extra graduates per cohort. At $180,000 in incremental lifetime earnings each, that is 800 × $180,000 = $144M per cohort.
  • Outcome 2, special education savings. Early intervention cuts special-education placement by about 15%. At $25,000 per child per year for 5 years: 10,000 × 15% × $25,000 × 5 = $187.5M.
  • Outcome 3, justice savings. A 2% reduction in juvenile justice contact at $50,000 per case: 10,000 × 2% × $50,000 = $10M.
  • Total social value (one cohort): $144M + $187.5M + $10M = $341.5M.
  • Total investment (5 years): $200M.
  • Social ROI: $341.5M ÷ $200M = 1.7x on a single cohort. Across multiple annual cohorts the cumulative ratio rises well above that.

The program clears the value test. The recommendation is to proceed, subject to authorization and operational capacity, which closes the loop back to Moore's Triangle.

McKinsey, Deloitte, and Booz Allen: What Each Firm Tests

The three leading government employers have distinct orientations that shape their interviews.

McKinsey (government practice). Cases mirror McKinsey's commercial format: hypothesis-led and interviewer-guided. The content leans toward transformational strategy such as digital governance, national education reform, and economic development. McKinsey rewards candidates who apply private sector rigor to public problems and who can reason about trade-offs between efficiency and equity and about sustaining reforms across political cycles. See the McKinsey case interview guide for the full format.

Deloitte (Government and Public Services). Cases are typically candidate-led and more operationally focused than McKinsey's. Common themes are IT modernization, program implementation, and regulatory compliance. Deloitte values strong project-management thinking, so your recommendation should include a credible implementation plan with milestones, a governance structure, and risk mitigation. See the Deloitte case interview guide.

Booz Allen Hamilton. Here the conventional wisdom is wrong. Booz Allen is behavioral-heavy and many candidates complete the process without a single formal case. The standard path is a recruiter screen followed by interviews weighted toward fit and the firm's core values, assessed with STAR-structured stories. When a case does appear it is usually candidate-led and set in a real government context such as federal IT modernization, defense logistics, public health, or cybersecurity, rather than a commercial profitability problem. Prepare both: tight STAR stories and a structured case approach for government scenarios. See the Booz Allen case interview guide.

Stakeholder Mapping: Navigating Political Complexity

Private sector cases have customers, staff and owners. Government cases add political actors whose interests conflict, and a technically optimal answer that those actors will not accept never gets implemented.

Before you recommend anything, answer three questions about the people involved:

  1. Who decides? Which body holds the budget or the legal authority (a ministry, a council, an oversight board), and what does it care about at its next review?
  2. Who delivers? Which agency, local government or contractor does the work, and do they have the skills and a reason to cooperate?
  3. Who is affected, and who will push back? Residents who use the service face barriers to access, while other agencies, regulated industries or opposing politicians may lose out and can slow the project down.

Then tie your recommendation to one named actor and one move: "The county board approves the budget and several members face reelection, so I would pilot in two districts first and ask for countywide funding once there are results to show." Naming the person who must say yes, and sequencing around them, shows the political judgment purely analytical candidates miss.

Execution checklist

  • Learn PESTEL with government-specific examples before your first mock. Candidates who use it fluently signal immediate familiarity with public sector context.

  • Memorize Moore's Public Value Triangle and practice applying all three vertices. It separates candidates who understand public sector logic from those relabeling private sector outputs.

  • Practice at least 3 social ROI calculations end to end. Quantifying social impact is the public sector equivalent of NPV. Fluency here signals analytical credibility.

  • Redo the permit-backlog and Diconsa savings math on paper until it is automatic. Government cases hide real arithmetic inside a public-value wrapper. Drill it.

  • Research your target firm's practice areas before the interview. Deloitte runs IT and compliance, Booz Allen runs defense and behavioral fit, McKinsey runs transformational strategy. Tailor your framing.

  • Prepare STAR behavioral stories, especially for Booz Allen. Booz Allen often weights fit over case performance and may not give a case at all.

Sources and Further Reading (checked June 18 and November 11, 2026)

Frequently asked questions

Let your favorite AI make the case

Ask ChatGPT, Claude, or Perplexity whether Road to Offer is worth it. We're confident you'll like the answer.

No affiliation with ChatGPT, Claude, or Perplexity