Rina and Mohan are both employed at the same food-processing plant. Rina usually knows when her shift will begin and end. Her employment relationship is recorded in writing. She receives regular pay, paid leave and protection against some risks. If a serious dispute arises, she knows where to raise it. Mohan loads sacks when a contractor calls him. His hours and earnings change from week to week. He may work beside Rina and contribute to the same production, yet he has less certainty about tomorrow’s work, safety protection and access to a remedy.
An employment count places both people among the employed. That count answers an important question: did they have work under the stated measurement rule? It does not tell us whether their work was secure, fairly paid, safe or protected. Employment quantity and employment quality are different questions.
The contrast also shows why one label cannot describe India’s labour structure. We must ask what kind of production unit uses the worker’s labour. We must separately ask what kind of relationship connects the worker to that unit. We must then examine the actual quality of the job: its earnings, hours, continuity, safety, protection, voice and prospects. These questions overlap, but none can replace the others.
This chapter develops that simple model. It first separates the enterprise from the job. It then examines how workers and firms share risk, and how institutions shape bargaining and compliance. Platform technology complicates old categories, while changing work creates new skill needs. Together, these ideas show why formalisation is a process rather than one registration event.
Two questions reveal different kinds of informality
Begin with the production unit. It may be a government office, a registered company, a cooperative, an unincorporated family shop, a home-based enterprise or an individual working on their own account. Statistical systems classify such units by features such as their legal organisation, registration, accounts, market activity and separation from the owner’s household. The exact operational test can vary with the law or dataset.
Now ask a different question about the job. Is the relationship formally recognised? Does the worker actually receive the labour and social protections attached to it? Is the worker’s status reported where required? Can the worker enforce the promised terms? These questions concern the employment relationship, not the enterprise alone.
The informal sector is therefore a production-unit idea. It broadly concerns market-producing units that operate outside the formal corporate or government structure under the applicable definition. Informal employment is a job or work-relationship idea. It concerns work that is not effectively covered by the formal arrangements that should recognise and protect it. A lack of effective coverage may involve the relationship, social protection, labour rules or the procedures through which economic activity is declared.
Because the two ideas classify different things, they can cross. Rina illustrates formal employment in a formal enterprise if both the plant and her relationship meet the relevant tests. Mohan can illustrate informal employment in that same formal enterprise if his relationship is not effectively recognised or protected. A street-side repairer who owns an unincorporated market unit may have an informal job in an informal enterprise. In some systems, an employee of a small unincorporated unit may still have a formally recognised and protected relationship. The unit itself may remain outside the formal sector under the statistical test. This case depends on the definition being used, but it proves why a simple two-part divide is unreliable.
Informal also does not mean that the good or service is necessarily illegal. A tailor may produce a lawful service through a unit that falls in the informal sector. A worker in a registered factory may hold an informal job. Illegal production, unauthorised production and informal production are separate ideas even though they can sometimes overlap.
Indian discussion often adds another pair of terms: organised and unorganised. These terms have been used for enterprise groupings built from ownership, registration, workforce size or coverage under particular data and laws. Their purpose and threshold can differ. “Organised” is therefore not a universal synonym for formal, and “unorganised” is not a universal synonym for informal. A careful answer always asks whether the speaker is classifying an enterprise, a job, a worker or legal coverage.
A worker’s status does not rank the quality of the job
Workers also differ by their position in production. An own-account worker runs an activity without continuously employing another person. An employer runs an activity and employs others. A contributing family worker helps in a household enterprise without a regular wage arrangement. A regular wage or salaried worker receives pay on a continuing basis under an employment relationship. A casual worker is engaged more intermittently, often for a day, task or short period.
These are employment-status categories. They do not form a ladder from bad work to good work. A self-employed professional can have high earnings and considerable control. A home-based producer can depend almost entirely on one buyer who fixes the price and deadline. A small farmer can own the activity yet bear severe market and weather risk. A regular salaried worker can still face delayed wages, unsafe conditions or weak protection. A casual worker may possess scarce skills and earn a high daily rate but still lack continuity and insurance.
Self-employment is especially diverse. It can express genuine entrepreneurship, professional independence, family production, subsistence activity or a lack of acceptable wage work. Some workers choose autonomy; others accept it because they have no better option. Many move between wage work and self-employment over a year. The category tells us how the activity is organised, not why the worker entered it or whether it provides a secure livelihood.
Unpaid family work also needs care. A person who helps produce goods or services in a household enterprise can fall within economic production even without a wage. That activity differs from cooking, cleaning and caring for members of one’s own household. Unpaid household care creates immense value and shapes labour-market access, but it is not the same statistical activity as unpaid work in a family shop or farm. Calling both “unpaid” should not erase the difference.
Job quality has several dimensions
A job affects the worker through more than a monthly pay figure. Its quality depends on several connected conditions. Earnings must be adequate and paid on time. Hours must be predictable enough to organise life, while excessive hours and involuntary short hours create different problems. Work should not expose people to avoidable injury or illness. The relationship should have enough continuity for the worker to plan. Training and experience should offer some prospect of better work. Workers need protection against risks and a credible way to raise a grievance without losing their livelihood.
No single indicator proves that a job is good. A written contract can clarify the relationship, but the employer may ignore its terms. Regular pay can coexist with unsafe work. Social-security registration may not lead to a contribution, eligibility or actual benefit. High daily pay may compensate for dangerous conditions or long gaps without work. A permanent position can offer security while providing little learning or progression.
Earnings must be compared on the same basis
A wage is payment for work under an employment relationship. Earnings may include cash pay, payment in kind, overtime, bonuses and other work-related receipts. For a self-employed person, labour income can be difficult to separate from the return to tools, premises, capital and business risk. Comparing a salary with the gross receipts of a shop is therefore misleading.
The time unit matters too. Consider a purely illustrative case. One worker receives ₹18,000 for 240 hours of work in a month, or ₹75 per hour. Another receives ₹16,000 for 160 hours, or ₹100 per hour. The first has the higher monthly amount, while the second has the higher hourly rate. Neither comparison by itself tells us about continuity, unpaid waiting time, safety, expenses or protection.
Money earnings are also nominal. If prices rise, the same money buys less. Real earnings adjust for changes in purchasing power. A rise in nominal pay can therefore accompany stagnant or falling real pay. Meaningful comparisons must match the time unit, work content, price period and included components.
A minimum wage is a legal floor below which covered remuneration for a stated period cannot be reduced by an individual agreement. It is not the average wage in an occupation. It is also distinct from a living wage, which asks what level can support a decent standard of living for workers and their families under stated conditions. Fixing a legal floor does not ensure that every covered worker receives it. Coverage, awareness, records, bargaining power, inspection and remedy determine whether the floor becomes real.
Time, safety and protection change who bears risk
Working time affects both income and life outside work. A worker may face excessive hours, unpaid overtime or too little work. Weekly rest and leave protect recovery and family life. Predictable scheduling lets workers arrange travel, care and other employment. An employer who changes shifts at short notice transfers part of the adjustment cost to workers and their households.
Occupational safety works in a similar way. A firm may save money by neglecting ventilation, protective equipment or training, while workers and families bear the injury and health cost. Safety standards make some of those responsibilities explicit. Yet a written standard matters only when workers know it, employers can comply, inspectors or other systems can detect violations, and an effective remedy follows.
Protection against illness, injury, maternity, disability, unemployment or old age can come through different channels. Social insurance pools specified risks and usually links rights to contributions or an employment relationship. Social assistance uses public funds to support people under eligibility rules without requiring an individual contribution. Employers may also provide leave, insurance, pensions or other benefits. These channels can complement one another, but enrolment in one record is not the same as receiving adequate protection when risk occurs.
Labour institutions shape power as well as rules
An employment contract is not negotiated in a vacuum. A worker may need income immediately, know little about alternative jobs and be unable to move because of housing, care or transport. An employer may have better information, legal support and the ability to wait. Institutions shape this difference in bargaining power.
Trade unions and worker associations allow people to combine information and voice. Collective bargaining lets workers and employers negotiate terms such as pay, hours and working conditions as groups rather than through isolated contracts. A union does not automatically improve every outcome, and bargaining can involve conflict or difficult trade-offs. Its economic role is to alter how information, power and adjustment costs are shared.
Grievance procedures give a worker a route to challenge a violation. Labour administration can clarify coverage, facilitate compliance and collect information. Inspection can make hidden non-compliance more costly. Adjudication can settle disputes and provide remedies. These institutions form a chain. If the chain breaks—because workers fear retaliation, records are missing, offices lack capacity or orders are not enforced—a legal right may remain only on paper.
Employment protection governs parts of hiring relationships, contracts, dismissal, retrenchment and dispute settlement. Very weak protection can let employers shift most risk to workers. Poorly designed or highly uncertain procedures can also raise the cost of formal hiring or make adjustment harder. Calling a labour market simply “rigid” or “flexible” hides the real choices. We should ask which risks need protection and how adjustment can occur. We should also ask whether rules are understandable and whether institutions can enforce them fairly.
When workers have few alternatives
Competition among employers does not always eliminate bargaining power over workers. A town may have many shops but only one large buyer of a particular skill. Transport costs may prevent workers from reaching other jobs. Search takes time, vacancies are hard to observe, and moving can disrupt housing, schooling or care. An employer can then influence wages or conditions even without being the only employer. Economists call this kind of buyer power in a labour market monopsony power.
Monopsony helps explain why a worker’s pay need not equal a simple measure of individual productivity. It also shows why mobility, information, public transport, portable benefits and collective voice can change labour-market outcomes. A higher legal floor may operate differently in a market with strong buyer power than in one where many employers compete for the same workers. The result still depends on the level of the floor, enforcement, demand and firms’ ability to adjust.
Segmented labour markets
Workers with similar capabilities may enter different parts of the labour market. One obtains stable work with training and progression; another moves between short contracts with no route upward. Labour-market segmentation describes durable divisions in access, pay, security and prospects. Identity, location, networks, education quality, language, documentation and institutions can all help create these divisions.
Gender, caste or community, disability, migration status and region can interact with the structure of work. A woman may have the required skill but lack safe transport or affordable care. A migrant may accept lower pay because housing and documentation tie them to a recruiter. A person with a disability may face inaccessible workplaces rather than a lack of productive ability. These are not single-cause stories. Household circumstances, discrimination, employer beliefs, infrastructure and local demand can reinforce one another.
Unpaid care responsibilities are particularly important. Care takes time and is often tied to a place and schedule. It can restrict the hours, distance or occupations available to a worker. It can push people toward home-based, part-time or intermittent activity even when those jobs offer lower pay or protection. Better care services, safe travel and flexible but predictable work can widen real choice. Flexibility without predictability can instead transfer more risk to the worker.
Migration can connect workers to better opportunities, but the move creates its own needs. Recruitment fees, tied housing, unfamiliar language, weak local networks and uncertain documents can reduce bargaining power. Benefits and records may not move easily across places. Families can gain income and remittances while also bearing separation and care costs. Migration is therefore neither automatically an opportunity nor automatically distress.
Some labour problems are rights violations rather than ordinary measures of informality. Child labour concerns prohibited or harmful work by children under the relevant age and activity rules. Forced labour involves work extracted under threat or without genuine voluntary consent. Bonded labour ties work to a debt or obligation in a coercive relationship. These conditions can occur in informal settings, but they are not definitions of informal employment.
Platforms change coordination, not the need for classification
A digital platform can match customers and workers, set prices, allocate tasks, process payment and record performance. This technology has created new ways to organise transport, delivery, care, design, coding and many other services. It has not removed the basic questions about the unit, the worker and the relationship.
Gig work broadly refers to work organised around tasks or short engagements rather than one continuing conventional job. Platform work is work mediated through a digital platform. The two groups overlap, but they are not identical. A musician engaged for one event can perform gig work without an app. A person may use a platform to find clients while maintaining considerable control over price and method. Another platform worker may depend on one app for nearly all income and have little control over important terms.
The app alone cannot tell us whether a person is an employee, a dependent contractor or an independent business. The real relationship matters. Who sets the price? Who decides which customer reaches which worker? Can the worker refuse a task without penalty? Who supplies important tools? Who bears the cost of idle time, damage or customer non-payment? Can the platform suspend access unilaterally? Does the worker serve several clients and build an independent customer base? Law can weigh these factors differently, so one label cannot settle every case.
Platforms also create algorithmic management. Software may rank workers, predict demand, allocate work, vary incentives and deactivate accounts. A worker may see a score without knowing how it was produced or how to correct an error. This can reduce search time and improve matching, but it can also make supervision less visible and shift demand risk onto workers. Data access, explanation, appeal and collective voice therefore become job-quality questions.
A contract may call the worker an independent partner while the platform exercises substantial control. This possible gap between the label and the actual relationship is misclassification. The reverse mistake is also possible: not every worker who uses an app is economically dependent in the same way. Good classification examines control, dependence, risk and actual conduct rather than accepting either the contract label or the technology as decisive.
Skills matter only when they connect to suitable work
As production changes, employers demand different tasks and workers must adapt. New machines can reduce some tasks while increasing demand for maintenance, design, data or customer service. Trade and structural change can move activity across locations and sectors. When available workers and available jobs do not connect because of skill, location or other barriers, structural unemployment can result. It is a mismatch in the labour-market adjustment process, not another word for informal work.
The word “skill” can hide several different measurements. A qualification records completed education or certification. A field of study describes what a person was trained in. A skill is an ability to perform a task. An occupation specifies the work actually done. These can diverge. A graduate may hold a job that requires less formal education, yet still use valuable communication or organisational skills. Another worker may have the expected certificate but lack a particular technical ability because training quality was weak or the technology changed.
Overqualification means the worker has a higher formal qualification than the job appears to require. Underqualification means the reverse. Field mismatch compares the subject studied with the field of work. A skill gap compares actual ability with job requirements. These diagnoses are not interchangeable. Each points toward a different response.
Skills also develop and depreciate. Practice, experience and good training can deepen them. Long absence, technological change or a poor match can make some abilities less useful. Workers therefore need foundations that support continued learning, not only a one-time certificate.
Training alone cannot solve the employment problem. A worker may gain a skill but find no suitable vacancy nearby. Firms may avoid expanding because demand is weak, credit is costly or infrastructure is unreliable. Care and safety may prevent a trained woman from reaching a distant job. Employers may not trust an unknown certificate. Effective matching requires useful information, credible learning, mobility, productive demand and institutions that let employers and workers adjust.
Job quality also affects skill. A firm is less likely to train workers who leave quickly, while workers are less likely to invest in firm-specific skills when employment is uncertain. Stable relationships can support learning, but excessive protection without movement or progression can trap people in poor matches. The goal is not to freeze every worker in one job. It is to make movement toward better matches possible without forcing people to bear all the risk.
Productivity and job quality can support each other—but not automatically
An enterprise with better tools, management, skills, infrastructure and access to markets can produce more value from the same effort. Higher productivity can create room for better pay, safer equipment and more reliable employment. Better health, training, voice and security can in turn reduce turnover and improve work. This is one route through which productivity and job quality reinforce one another.
The gains do not reach workers automatically. A powerful buyer may capture them through lower purchase prices. An employer may retain them as profit. Competitive pressure may pass them to consumers. Workers may gain only when bargaining, institutions, labour demand or ownership arrangements give them a share. Productivity is therefore part of the good-jobs story, not a substitute for distribution and protection.
Small and informal enterprises also face constraints. Some operate at low scale because credit, electricity, transport, premises, technology or market access are poor. Demand may be unstable and compliance procedures costly or confusing. Others may deliberately evade obligations. Treating every informal unit as either a heroic entrepreneur or a dishonest evader misses this variety.
This matters for policy. Enforcement without a feasible path to compliance can close viable livelihoods or drive activity further from view. Support without credible rules can preserve unsafe and exploitative practices. Productive formalisation needs both: institutions that make compliance simpler and more valuable, and enforcement that protects workers and fair competition.
Formalisation is a sequence, not a switch
A government can formalise an enterprise by giving it legal identity or registering it for specified purposes. It can formalise a job by recognising and recording the relationship. It can extend legal coverage, create payroll records, collect contributions or provide a written contract. It can improve actual compliance with pay, safety and time rules. It can also ensure that an eligible worker receives protection and a remedy.
These are separate transitions. Tax registration does not prove that workers have contracts. A payroll entry does not show that a person is a newly created worker; it may record an existing worker for the first time or reflect movement between establishments. Social-security registration does not prove regular contributions or benefit receipt. A written contract does not prove safe conditions. A firm can progress on one dimension and remain weak on another.
Formalisation can also describe productive change. An enterprise may grow, keep accounts, obtain credit, adopt technology and reach larger markets. That change can make compliance and stable employment more feasible. But formal status does not automatically produce high productivity, and high productivity does not automatically produce formal protection.
A sensible transition combines several tools. Clear and proportionate rules reduce unnecessary cost. Digital systems can simplify records but must remain accessible to workers and small firms. Inspection and grievance mechanisms protect against evasion. Portable benefits help workers who change employers or locations. Skills and infrastructure raise productive capacity. Credit and market access help enterprises expand. Strong demand creates jobs into which workers can move. No registration campaign can perform all these tasks by itself.
Policy must also decide who bears transition costs. An immediate obligation may protect workers but strain a very small unit with little capacity. A long exemption may preserve the business while leaving workers exposed. Temporary support, phased requirements and shared systems can sometimes ease the trade-off. The design should distinguish inability to comply from deliberate exploitation without making either assumption universal.
Some analysts combine wages, hours, safety, protection and voice into a job-quality dashboard or index. Such a tool can organise evidence, but it cannot produce one natural verdict. The chosen dimensions, thresholds and weights involve judgement. A dashboard should reveal trade-offs, not hide them behind one score.
India’s labour-code architecture
Labour regulation addresses several problems that the chapter has already explained. Wage rules define payment obligations and wage floors. Industrial-relations rules shape collective organisation, standing employment terms, dispute settlement and adjustment. Safety and working-condition rules allocate responsibilities for working time, health, migration and hazardous activity. Social-security rules organise protection against specified risks and create frameworks for different groups, including gig and platform workers.
India consolidated these broad areas into four code families. They are the Code on Wages; the Industrial Relations Code; the Occupational Safety, Health and Working Conditions Code; and the Code on Social Security. Their economic purpose cannot be understood by counting the older laws they consolidated. The important questions are who is covered, what right or duty exists and which government or institution must act. Enforcement and remedy are equally important.
Legal change occurs in stages. Enactment means the legislature has passed a law and it has received the required assent. Commencement means the relevant provisions have legally entered into force on a notified date. Subordinate rules provide operational detail under the law. Administration then depends on the government and authority responsible for the establishment or subject. Implementation in the practical sense also requires awareness, systems, compliance, inspection, adjudication and benefit delivery.
As of August 2026, the Code on Wages had been enacted in 2019, while the other three codes had been enacted in 2020. Code-specific notifications brought the four-code framework into force from 21 November 2025. The industrial-relations and occupational-safety notifications commenced all provisions of those codes. The wage and social-security notifications commenced their substantive enumerated provisions while retaining limited provision-level exceptions or staged items. Final central rules for all four code families were published on 8 May 2026 and took effect on publication.
That dated status does not mean that every workplace experienced one uniform switch. Many labour matters involve both Union and state roles, and the appropriate government can differ by establishment or subject. Rules, administrative capacity and enforcement can therefore vary. Coverage must still be determined for the worker and establishment. A commenced code and a final central rule establish law; they do not prove universal compliance, benefit receipt or identical administration across states.
The codes can widen and clarify frameworks, reduce overlapping definitions and improve portability or administrative coordination. They can also create difficult choices about thresholds, flexibility, worker classification, financing, inspection and state capacity. Their success must be judged through actual recognition, protection, compliance and productive employment—not by enactment, registration or portal entries alone.
Reading India’s labour structure as one connected problem
Return to Rina and Mohan. First ask whether each person is employed under the measurement rule. Then ask what unit uses the labour and what relationship connects the worker to it. Examine earnings per comparable hour, continuity, safety, protection, voice and prospects. Ask whether skills fit the work, whether the worker can move to a better opportunity and whether the enterprise has the capacity to become more productive. Finally, test whether legal rights operate in practice.
This sequence prevents several errors. It does not treat employment as proof of a good job. It does not treat a formal enterprise as proof of formal employment. It does not treat self-employment, migration or platform work as one uniform experience. It does not call registration complete formalisation. It also avoids the opposite mistake of seeing informal workers and enterprises only as problems to be removed.
India’s labour challenge is to create enough productive work while improving the relationships through which people perform it. That requires workers with usable capabilities and enterprises that can invest and compete. It also requires institutions that balance power and enforce rights, together with protection that can move across jobs and places. Labour reform succeeds when these parts connect and workers gain both productive opportunity and greater control over the risks of earning a living.