Public policy process

In short

The public policy process is the sequence of activities by which a government recognises a problem, decides what to do about it, converts the decision into a law, scheme or rule, delivers it through administrative machinery, and then checks whether it worked. In India this process runs through Parliament and State legislatures, the Union Cabinet and Cabinet Secretariat, line ministries, NITI Aayog, regulators and the courts, and it is where politics, administration and evidence actually meet. It is a core unit of BA Public Administration, Political Science and Governance papers, and the backbone of UPSC/State PSC General Studies Paper II (Governance) and the Public Administration optional. Employers who test it include policy think tanks and research organisations, CSR and development consulting firms, PSU and government management trainee streams, and NGO programme roles.

Explained

What policy is, and what it is not

• Public policy: a purposive course of action (or deliberate inaction) adopted by a public authority to address a perceived public problem. Thomas Dye's classic phrasing: "whatever governments choose to do or not to do."
• Policy is not a single document. It is the whole bundle: the stated goal, the instrument chosen, the money allotted, the rules framed, the officials who deliver it and the way it is enforced.
• Levels in India: constitutional provisions (Fundamental Rights, Directive Principles) → statutes passed by legislatures → delegated legislation (rules, regulations, notifications) → executive policy (Cabinet decisions, schemes, guidelines, office memoranda) → field-level practice.
• Policy versus administration: Woodrow Wilson's politics-administration dichotomy claimed politicians set policy and administrators only execute it. Later writers (Appleby, Simon, Waldo) rejected this — civil servants shape policy by drafting Cabinet Notes, interpreting rules and exercising discretion.
• Types by content:
  • Substantive (what government does: a minimum support price) versus procedural (who decides and how: tendering rules).
  • Distributive (subsidies spread widely), redistributive (transferring from one class to another, e.g. progressive taxation, reservation), regulatory (restricting behaviour, e.g. emission norms) and constituent/self-regulatory (creating institutions, e.g. setting up a regulator) — Theodore Lowi's typology.
• Policy instruments (tools): direct provision, public funding/subsidy, taxation and cesses, regulation and licensing, information and awareness, nudges (behavioural design), market creation (auctions, tradable permits), and contracting/PPP.

The policy cycle: stages and what happens at each

The stages model (Lasswell's seven functions, later simplified by Anderson into five stages) is the standard exam framework. The cycle is analytical, not a real-time sequence — stages overlap and loop.

1. Problem identification and agenda setting. A condition becomes a "problem" only when it is politically recognised. Triggers: a crisis or focusing event, media coverage, judicial direction, data from a survey (NSSO/NSO, NFHS, Census), pressure from interest groups, election manifestos, international commitments (SDGs, Paris Agreement). Kingdon's multiple streams model: a policy window opens when the problem stream, policy stream and politics stream converge.
2. Policy formulation. Framing alternatives and selecting one. In India, the line ministry prepares a concept note, consults stakeholders, obtains the Finance Ministry's view on funds and the Law Ministry's view on legality, and circulates a Cabinet Note for inter-ministerial comments. NITI Aayog (which replaced the Planning Commission on 1 January 2015) advises; the Economic Survey, Finance Commission reports, expert committees and commissions of inquiry feed options in. A pre-legislative consultation policy (2014) requires draft legislation to be placed in the public domain, normally for 30 days — compliance is uneven, a point examiners like.
3. Policy adoption/legitimation. The Cabinet approves; a Bill is introduced, examined by a Department-related Standing Committee or Select Committee, passed by both Houses (Money Bills only by Lok Sabha, with Rajya Sabha recommendations), assented to by the President, and notified in the Gazette. Schemes that need no new law are adopted by Cabinet/Cabinet Committee on Economic Affairs approval plus a budget line.
4. Implementation. Rules are framed under the parent Act, guidelines issued, funds released (Centrally Sponsored Schemes with a Centre-State sharing ratio, usually 60:40, and 90:10 for Himalayan and North-Eastern States), and delivery happens through State departments, district administration, panchayats/urban local bodies under the 73rd and 74th Amendments, and increasingly through digital platforms (DBT, Aadhaar-based payments, PFMS). Pressman and Wildavsky showed implementation is where policies die; Lipsky's street-level bureaucracy shows frontline discretion effectively remakes policy.
5. Evaluation and feedback. Concurrent monitoring (dashboards, MIS), outcome evaluation, impact evaluation, social audit (mandatory for the rural employment guarantee scheme), audit by the CAG (performance audits under Article 149), parliamentary oversight (Public Accounts Committee, Estimates Committee, Committee on Public Undertakings), and evaluation by NITI Aayog's Development Monitoring and Evaluation Office (DMEO). Feedback leads to policy maintenance, succession or termination.

Models of decision-making (the most examined sub-topic)

• Rational comprehensive model: define the goal, list all alternatives, compute all costs and benefits, choose the optimum. Assumes complete information, agreed values and unlimited analytical capacity — unrealistic.
• Herbert Simon — bounded rationality and satisficing: cognitive limits, cost of information and time force decision-makers to accept the first satisfactory option rather than the best. Simon also split decisions into programmed (routine, rule-based) and non-programmed (novel, judgemental).
• Charles Lindblom — incrementalism ("muddling through"): policy moves in small steps from the status quo; only a few alternatives, close to current practice, are compared. Advantages: feasible, reversible, politically acceptable. Limitations: conservative, no innovation, bad for crises and for countries needing structural change.
• Amitai Etzioni — mixed scanning: combines a broad, shallow scan of all options (rational) with a detailed look at a few (incremental); fundamental decisions get the wide scan, routine ones the narrow.
• Yehezkel Dror — normative optimum model: a critique of both; adds meta-policymaking (deciding how to decide), explicit use of judgement, creativity and "extra-rational" elements, plus post-decision feedback. Dror's 18 phases are grouped into meta-policymaking, policymaking and post-policymaking.
• Cohen, March and Olsen — garbage can model: in "organised anarchies" problems, solutions and participants are dumped into a choice opportunity and matched almost by accident; solutions often exist before problems.
• Allison's three models (from the Cuban missile crisis): rational actor, organisational process, governmental/bureaucratic politics — useful for explaining defence and foreign policy decisions.
• Elite, group and institutional theories: policy as the preference of a ruling elite (Pareto, Mosca, Mills), as the equilibrium of group pressure (Truman, Latham), or as a product of institutional rules and path dependence.
• Systems model (Easton): inputs (demands and supports) → the political system as a "black box" → outputs (policies) → feedback.
• Top-down versus bottom-up implementation: top-down (Van Meter and Van Horn, Sabatier and Mazmanian) stresses clear objectives and control; bottom-up (Elmore's backward mapping, Lipsky) stresses frontline realities. Advocacy coalition framework (Sabatier) explains long-run change through competing coalitions sharing belief systems.

Indian institutional machinery and the Centre-State dimension

• Political executive: Union Cabinet and Cabinet Committees (notably CCEA, Cabinet Committee on Security, Appointments Committee of the Cabinet); the Prime Minister's Office; State Cabinets.
• Permanent executive: Cabinet Secretariat (coordination, Committee of Secretaries), line ministries and departments, Department of Expenditure (Expenditure Finance Committee / Public Investment Board clear project proposals above threshold limits, which are revised periodically), Legislative Department and Department of Legal Affairs.
• Advisory and evaluative: NITI Aayog and its Governing Council, DMEO, Finance Commission (Article 280), GST Council (Article 279A), statutory regulators (RBI, SEBI, TRAI, CERC, CCI, IRDAI, FSSAI), and the Law Commission.
• Legislature: Parliament's Standing Committees, questions, debates, Private Members' Bills, and scrutiny of delegated legislation by the Committee on Subordinate Legislation.
• Judiciary: judicial review, public interest litigation and continuing mandamus have produced policy in practice (e.g. directions on CNG in Delhi transport, the midday meal order of 2001, environmental directions). Courts also restrain themselves on "policy choices" where no legal right is breached.
• Federal division: Union, State and Concurrent Lists (Seventh Schedule). Education, forests and labour sit in the Concurrent List, so central policy needs State cooperation. Inter-State Council (Article 263), Zonal Councils and the NITI Aayog platform are the coordination forums; the Punchhi Commission (report 2010) and the Sarkaria Commission (1988) examined Centre-State policy coordination.
• Non-state participants: political parties, business associations (CII, FICCI, ASSOCHAM), farmer and trade unions, think tanks, media, civil society and campaign networks (the RTI and employment-guarantee campaigns are standard Indian examples of civil society driving the agenda).

Evidence, analysis and current Indian practice

• Policy analysis techniques examined: cost-benefit analysis with discounting, cost-effectiveness analysis, multi-criteria/weighted scoring, SWOT and PESTLE scanning, stakeholder analysis (power-interest grid), decision trees, operations research models, and regulatory impact assessment.
• Evaluation designs: process evaluation, before-after comparison, difference-in-differences, randomised controlled trials (used widely in Indian education and health programme evaluation), social audit, beneficiary surveys, and third-party concurrent evaluation.
• Outcome Budget and Output-Outcome Monitoring Framework: every Union ministry publishes output and outcome indicators with its demand for grants; NITI Aayog compiles composite indices (SDG India Index, School Education Quality Index, Health Index, Export Preparedness Index) that work as policy feedback and as inter-State competition ("competitive and cooperative federalism").
• Digital delivery as a policy instrument: Direct Benefit Transfer with Aadhaar-based payments, PFMS for fund flow, JAM trinity (Jan Dhan-Aadhaar-Mobile), the Aadhaar (Targeted Delivery of Financial and Other Subsidies, Benefits and Services) Act, 2016, and the Digital Personal Data Protection Act, 2023 (whose rules were notified in 2025, with phased commencement) as the current data-governance frame.
• Recent legislative style: large-scale codification and consolidation — the four Labour Codes (Code on Wages, Industrial Relations Code, Social Security Code, Occupational Safety Health and Working Conditions Code) came into force on 21 November 2025, replacing the Factories Act, Minimum Wages Act, Industrial Disputes Act and BOCW Act; the three new criminal laws (Bharatiya Nyaya Sanhita, Bharatiya Nagarik Suraksha Sanhita, Bharatiya Sakshya Adhiniyam) came into force on 1 July 2024; the Income-tax Act, 2025 replaced the Income-tax Act, 1961 from 1 April 2026 with a complete renumbering of sections and forms; GST has had 5%, 18% and 40% slabs since 22 September 2025. These are the textbook illustrations of policy succession rather than termination.
• Reform commissions on process: the Second Administrative Reforms Commission (2005-09, 15 reports) recommended citizen-centric administration, RTI-based transparency, a Public Services Bill, and evidence-based evaluation; the First ARC (1966-70) under Morarji Desai shaped the Secretariat and Lokpal debate.
• Limitations of the cycle model (examiners ask for critique): it suggests a tidy linear order that does not exist; it is silent on power and interests; it treats stages as separate when formulation and implementation blur; and it under-states the role of courts, markets and international agreements.

Key points

• Public policy — purposive action or deliberate inaction by a public authority on a public problem (Dye).
• Policy cycle — agenda setting, formulation, adoption, implementation, evaluation, with feedback; analytical device, not a real sequence.
• Agenda setting — a condition becomes a problem only when politically recognised; Kingdon's three streams and policy window explain timing.
• Lowi's typology — distributive, redistributive, regulatory and constituent policies, each with its own politics.
• Rational model — complete information and full comparison of alternatives; normatively neat, practically impossible.
• Bounded rationality / satisficing — Simon: limited information and capacity force a "good enough" choice.
• Incrementalism — Lindblom: small steps from the status quo; feasible but conservative.
• Mixed scanning — Etzioni: broad scan for fundamental decisions, detailed analysis for routine ones.
• Normative optimum — Dror: adds meta-policymaking, creativity and extra-rational judgement.
• Garbage can — Cohen, March and Olsen: fluid matching of problems, solutions and participants in organised anarchies.
• Street-level bureaucracy — Lipsky: frontline discretion effectively makes policy at the point of delivery.
• Implementation gap — Pressman and Wildavsky: the distance between intent and delivered outcome; the commonest Indian failure point.
• Cabinet Note — the inter-ministerial instrument through which a proposal reaches Cabinet, with Finance and Law Ministry views.
• Pre-legislative consultation policy, 2014 — draft laws and rules to be published, normally 30 days, for public comment.
• NITI Aayog and DMEO — advisory think tank (from 1 January 2015) and its evaluation arm; indices as feedback instruments.
• Oversight institutions — CAG performance audit (Article 149), Standing Committees, PAC, Estimates Committee, social audit.
• Centrally Sponsored Schemes — cost shared with States, typically 60:40 and 90:10 for Himalayan/North-Eastern States.
• Policy succession — most "new" policy replaces existing policy (Labour Codes 2025, Income-tax Act 2025) rather than ending it.

Formulas and examples

Example 1 — Cost-benefit appraisal of a rural road (NPV and BCR). A State proposes a rural road costing ₹5 crore now, with net benefits (time and crop-loss savings) of ₹1.5 crore a year for 5 years. Social discount rate r = 10%.

• Discount factors: 1/1.1 = 0.9091; 0.8264; 0.7513; 0.6830; 0.6209. Sum = 3.7907.
• Present value of benefits = 1.5 × 3.7907 = ₹5.686 crore.
• NPV = 5.686 − 5.00 = +₹0.686 crore → accept.
• BCR = 5.686 / 5.00 = 1.14.
• If the discount rate were 15%: factors 0.8696 + 0.7561 + 0.6575 + 0.5718 + 0.4972 = 3.3522; PV = 1.5 × 3.3522 = ₹5.028 crore; NPV = +₹0.028 crore, BCR = 1.01 — the project is barely viable, showing how sensitive appraisal is to the discount rate.

Example 2 — Choosing between two instruments (cost-effectiveness). A district must raise full immunisation coverage. Option A: mobile vaccination vans, cost ₹80 lakh, expected 20,000 additional children immunised. Option B: incentive payment to mothers, cost ₹50 lakh, expected 10,000 additional children.

• CER (A) = 80,00,000 / 20,000 = ₹400 per child.
• CER (B) = 50,00,000 / 10,000 = ₹500 per child.
• A is more cost-effective per child; but B costs ₹30 lakh less in total. If the budget ceiling is ₹50 lakh, B is the only feasible option — the standard exam point that cost-effectiveness ranking does not survive a binding budget constraint.

Example 3 — Weighted multi-criteria choice of a scheme design. Criteria and weights: targeting accuracy 0.35, speed of rollout 0.25, administrative cost 0.20, political acceptability 0.20. Scores out of 10:

• Universal: 0.35×5 + 0.25×9 + 0.20×7 + 0.20×9 = 1.75 + 2.25 + 1.40 + 1.80 = 7.20.
• Means-tested: 3.15 + 1.50 + 1.60 + 1.00 = 7.25.
• Subsidised price: 2.10 + 1.75 + 1.00 + 1.60 = 6.45.
• The top two are within 0.05 — the honest conclusion is that the ranking is not robust, and the weights (a political judgement) decide the answer.

Example 4 — Impact evaluation by difference-in-differences. A State introduces a school-meal upgrade in treatment blocks. Attendance (%): treatment 62 → 74; control 60 → 64.

• Treatment change = 74 − 62 = +12 points. Control change = 64 − 60 = +4 points.
• Impact = 12 − 4 = +8 percentage points attributable to the policy; the remaining +4 is the common trend (season, general awareness).
• A before-after comparison alone would have overstated the effect as +12 points — a 50% overstatement.

Example 5 — Centre-State cost sharing of a Centrally Sponsored Scheme. A scheme has a total outlay of ₹1,000 crore for the year, shared 60:40 with general-category States and 90:10 with Himalayan and North-Eastern States. Suppose ₹700 crore is spent in general States and ₹300 crore in Himalayan/NE States.

• General States: Centre 0.60 × 700 = ₹420 crore; States ₹280 crore.
• Himalayan/NE: Centre 0.90 × 300 = ₹270 crore; States ₹30 crore.
• Total Centre share = 420 + 270 = ₹690 crore (69%); total State share = ₹310 crore (31%).

Where it is used

• Think tanks and research organisations — NITI Aayog consultants, PRS Legislative Research, Centre for Policy Research, Takshashila, NCAER, IDFC Institute: policy briefs, legislative analysis, evaluation studies.
• Government and PSU careers — civil services (policy desks in ministries), State administrative services, Parliament and legislature secretariats, regulators (RBI, SEBI, TRAI, CCI) in research and policy cells.
• Development and CSR sector — programme design and monitoring in NGOs, foundations and CSR teams under the Companies Act CSR provisions; impact assessment mandated for large CSR projects.
• Consulting and advisory — government advisory practices of large consulting firms and development consultancies working on urban missions, health, education and energy transition projects.
• Corporate public policy and regulatory affairs — government relations roles in telecom, pharmaceuticals, fintech, e-commerce and energy companies, which track draft rules during consultation and file comments.
• Real Indian examples to cite — the Right to Information campaign driving a statute; employment guarantee with social audit built in; GST designed through the GST Council as a federal bargain; the shift from the Planning Commission to NITI Aayog as a change in the formulation machinery.

Interview questions

• Q: What is public policy in one line? A: A purposive course of action or inaction adopted by a public authority to deal with a public problem, expressed through laws, rules, schemes and administrative practice.
• Q: Name the stages of the policy cycle. A: Problem identification and agenda setting, formulation, adoption/legitimation, implementation, and evaluation with feedback — overlapping rather than strictly sequential.
• Q: Why is the rational comprehensive model criticised? A: It assumes complete information, agreed and ranked values and unlimited analytical capacity, none of which exist; Simon replaced it with bounded rationality and satisficing.
• Q: How does incrementalism differ from mixed scanning? A: Incrementalism only considers small changes from the status quo; mixed scanning adds a broad, shallow scan of all options so that fundamental decisions are not missed.
• Q: Where do Indian policies most often fail, and why? A: At implementation — vague guidelines, weak frontline capacity, delayed fund flow, poor inter-departmental coordination and unmonitored discretion, the gap Pressman and Wildavsky described.
• Q: What role does a Cabinet Note play? A: It is the formal proposal circulated by a ministry for inter-ministerial comments, including Finance and Law Ministry views, on which the Cabinet takes its decision.
• Q: Is pre-legislative consultation legally binding in India? A: The 2014 policy requires draft legislation and major rules to be published for about 30 days for comment, but it is an executive policy, not a statute, and compliance has been inconsistent.
• Q: How does the judiciary influence policy? A: Through judicial review, public interest litigation and continuing mandamus it has effectively created policy in areas like environment, food and education, while formally declining to substitute its own view on pure policy choices.
• Q: Difference between monitoring and evaluation? A: Monitoring tracks inputs, activities and outputs continuously against targets; evaluation asks periodically whether outcomes and impacts were achieved and whether they are attributable to the policy.
• Q: How would you evaluate a scheme if you could not run an RCT? A: Use a quasi-experimental design — difference-in-differences with a comparable control group, propensity-score matching or a regression discontinuity at an eligibility cut-off — supported by process evaluation and beneficiary surveys.
• Q: What is policy termination, and why is it rare? A: Formally ending a policy; it is rare because beneficiaries, staff and political sponsors resist, so governments prefer policy succession — replacing one policy with another, as the Labour Codes replaced four old labour statutes in November 2025.
• Q: Does federalism help or hinder policy in India? A: Both — States adapt policy to local conditions and compete on outcomes, but Concurrent List subjects and cost-sharing conditions create friction, which the GST Council, Inter-State Council and NITI Aayog forums try to manage.

Common mistakes

• Treating the policy cycle as a real chronological sequence and never mentioning its critique — examiners expect the limitation that stages overlap and power is ignored.
• Mixing up the theorists: attributing satisficing to Lindblom, incrementalism to Simon, or mixed scanning to Dror instead of Etzioni.
• Describing only formulation and adoption, and skipping implementation and evaluation, where most marks and real failures lie.
• Writing a policy answer with no Indian institution or statute in it — a good answer names the Cabinet, the ministry, NITI Aayog, the CAG, a Standing Committee or a specific Act.
• Confusing outputs with outcomes (toilets built versus open-defecation-free status; classrooms built versus learning levels).
• In cost-benefit sums, forgetting to discount future benefits, or treating a single discount rate as settled rather than testing sensitivity.
• Claiming policy is purely an administrative or purely a political act — the politics-administration dichotomy has long been rejected.

Related topics

E-governance · Right to Information Act · Administrative ethics and accountability · Centre-State relations and federalism · Development administration and planning in India · Public financial management and budgeting