Every fresher who has ever signed a joining letter has run their eyes over that one clause tucked between the salary structure and the leave policy: “The employee shall serve the Company for a minimum period of ___ years, failing which a sum of ₹___ shall become payable as liquidated damages.”

For decades, this clause lived in a grey zone. Was it a legitimate business safeguard, or was it, as one Kerala High Court judgment memorably framed the debate, a modern-day echo of indentured labour dressed up in corporate stationery? In May 2025, the Supreme Court of India finally settled the argument in Vijaya Bank & Anr. v. Prashant B. Narnaware (2025 INSC 691), giving employers a clearer green light — but with conditions attached that employees would do well to know.

This article traces the law from its 1872 origins to its 2025 clarification, and lays out precisely when a bond will hold up in court and when it will collapse like a badly drafted contract.

What, Exactly, Is an Employment Bond?

An employment bond (also called a service bond or training bond) is a contractual promise by an employee to serve an employer for a fixed minimum period, commonly one to three years, usually after receiving specialised training, a sponsored certification, or a costly recruitment process. If the employee exits early, the bond obliges them to pay a pre-agreed sum. Such clauses are especially common in IT and ITES firms after structured onboarding programmes, in banks and public-sector undertakings, and in manufacturing units that invest heavily in technical training.

At first glance, this looks like a straightforward promise. Indian contract law, however, treats any clause that smells of restraining someone’s livelihood with deep suspicion — and that suspicion is codified in a single, powerful sentence.

The Statutory Battlefield: Sections 23, 27, 73 and 74

Four provisions of the Indian Contract Act, 1872 do all the heavy lifting in every employment bond dispute:

• Section 23 voids any agreement whose object or consideration is unlawful or opposed to public policy.
• Section 27 declares that an agreement restraining a person from exercising a lawful profession, trade, or business is void to that extent, the only express exception being agreements for the sale of goodwill of a business.
• Section 73 allows compensation for loss actually suffered from a breach, excluding remote or indirect damage.
• Section 74 governs situations where the contract itself names a sum payable on breach, permitting the court to award reasonable compensation not exceeding that stipulated amount, regardless of whether actual loss is strictly proven.

Section 27 is the real battleground. Employees challenging a bond almost always argue that being forced to pay a penalty for leaving effectively restrains their freedom to work elsewhere. Employers counter that the bond merely secures a minimum tenure, it does not stop anyone from resigning; it only prices the consequence of doing so.

The Case Law Trail

Niranjan Shankar Golikari v. Century Spinning (1967 AIR 1098)

This foundational ruling drew the line that still governs the field today: restrictive covenants operating during the subsistence of an employment contract are valid, because during that period the employee owes exclusive service to the employer. Restraints that continue to bite after the employment ends, however, generally fall foul of Section 27.

Superintendence Company v. Krhaisn Murgai (1980 AIR 1717)

This case reinforced the same in-service versus post-service distinction, confirming that a negative covenant preventing an employee from taking up other work while still employed does not offend Section 27.

Vijaya Bank v. Prashant B. Narnaware (2025 INSC 691)

The facts read like a case study built for a law school seminar. A Senior Manager at Vijaya Bank had agreed, as part of a recruitment drive, to serve a minimum of three years or pay ₹2 lakh in liquidated damages on premature resignation. He resigned early to join IDBI Bank, paid the amount under protest, and then challenged the clause, first before the Karnataka High Court, which ruled in his favour, and finally before the Supreme Court.

The Court, led by Justices P.S. Narasimha and Joymalya Bagchi, reversed the High Court and upheld the bond. Its reasoning rested on three pillars:

1. Timing matters. The restriction operated only during the currency of the employment contract and ceased the moment employment ended, it did not stop Narnaware from joining IDBI Bank at all, it simply attached a price to leaving early.
2. Purpose matters. The clause was intended to secure a minimum service period and curb attrition, not to bar future employability, a legitimate business interest, particularly for a public-sector bank that invests in recruitment and training.
3. Proportionality matters. ₹2 lakh, measured against recruitment costs and the employee’s own pay scale, was neither unconscionable nor oppressive, it was a reasonable, not punitive, figure.

The upshot: minimum-service clauses with a liquidated-damages backstop do not automatically violate Section 27 or public policy under Section 23, provided they operate only within the employment period and are reasonably proportionate.

So When Is a Bond Actually Enforceable?

Stitching the case law together, a bond survives judicial scrutiny when it clears roughly four hurdles:

1. It operates only during employment. A clause that follows the employee out the door barring them from joining a competitor, or from working in the same industry for years afterward is far more vulnerable, since post-termination restraints sit squarely within Section 27’s prohibition.
2. It doesn’t functionally trap the employee. Courts distinguish between a bond that makes leaving costly and one that makes leaving impossible. The former is a commercial safeguard; the latter edges toward involuntary servitude and public-policy objections under Section 23.
3. The sum is a genuine pre-estimate of loss, not a penalty in disguise. This is where Section 74 does its quiet, decisive work.
4. The restriction is proportionate to a legitimate interest. Recruitment cost, training expenditure, notice-period shortfall, or the value of a sponsored certification are all legitimate interests. A blanket, disproportionately large sum imposed on a fresher who attended a two-week orientation is not.

Notably, no Indian court will compel an employee to actually continue working — specific performance of a personal service contract is barred by Section 14 of the Specific Relief Act. An employee can always resign; the only question the law asks is what they must pay for doing so.

Recovering Compensation: Where Sections 73 and 74 Diverge

This is the part employers frequently get wrong, and employees frequently misunderstand.

If the contract is silent on the amount (no specific sum was fixed) the employer must sue under Section 73 and prove actual loss: real training costs, documented recruitment expenses, or demonstrable productivity loss. Vague assertions of “inconvenience” will not survive judicial scrutiny.

If the contract names a specific sum, i.e. the classic “₹2 lakh if you leave before three years” clause, then Section 74 applies. Here, the employer need not prove the exact loss down to the rupee, but the amount must still be a reasonable compensation, not exceeding what was stipulated, and courts retain full discretion to scale it down if the figure looks like a punitive penalty rather than a genuine pre-estimate of damage. The Supreme Court’s own approach in ONGC v. SAW Pipes (AIR 2003 SC 2629) is instructive here: even a clear, unambiguous liquidated-damages clause will be enforced at its stated value only if that value is not, on its face, unreasonable or oppressive.

A useful real-world illustration comes from a Kerala High Court dispute involving a training bond for prospective employees. A trainee who resigned five months into a structured programme argued that she had never been a “permanent employee” and so owed nothing. The Court disagreed and said that the employer had invested real time, money, and opportunity cost in training her, and that investment itself constituted the loss to be compensated, even if the exact figure required judicial determination.

Conclusion: A Price Tag, Not a Padlock

The doctrinal thread running from Niranjan Shankar Golikari in 1967 to Vijaya Bank v. Narnaware in 2025 is remarkably consistent: Indian law will let an employer put a price on early exit, but it will never let an employer install a padlock on the door. A bond that says “leaving costs you this much, and here’s why” tends to survive; a bond that says “you may not leave, or may never work in this field again” tends to fall.

For employers, the lesson from the 2025 ruling is to draft with restraint — tie the sum to demonstrable cost, keep the restriction confined to the employment period, and resist the temptation to inflate the figure into a deterrent. For employees, the lesson is equally clear — a bond is not an unbreakable chain, but it is not toothless paper either; the exit door remains open, provided you’re prepared to pay a reasonable, provable price for walking through it.