Permanent work should be done by permanent people

Walk along a gear line in many Indian auto-component plants and you will often find two operators at neighbouring machines doing the same work in the same way. One is a permanent employee, a member of the recognised union, covered by its long-term settlement. The other was engaged through a contractor, earns perhaps half as much, may have worked on that line for years, and could be gone next month. The arrangement is so common that most managers no longer see it.

I did not see it clearly myself until August 2023, when a contract worker at our Chakan I plant lost two fingers while loading a hobbing machine. I want to explain why that accident changed my view, and why I now believe that core production work in a precision plant should be done by permanent employees.

Two operators, one machine, two rates of pay

The use of contract labour in Indian manufacturing has a history that makes it easy to defend. Demand in the automotive industry is cyclical, and permanent employment in large plants is hard to reduce when volumes fall. Contract workers give a plant a buffer. In periods of rapid growth they also let a company add shifts quickly. For work outside core production, such as housekeeping, canteens, material handling and security, contracting is often sensible, and the contractor may do the work better than the plant would.

The problem is drift. Over the years, the buffer becomes the base. At Chakan I in 2023, contract workers made up 45 per cent of the shop floor: 1,180 people beside 1,450 permanent employees. Many of them operated machines on core lines. Some had done so for longer than the permanent operators next to them. They were trained by the contractor's supervisors, not by us, and they sat partly outside our safety system. The worker who was injured had been on the line for three years.

Our recognised union, the Kavirath Kamgar Sanghatana, began a go-slow the day after the accident and demanded that every contract worker in production be made permanent. Output fell by about 30 per cent for eleven days. Our largest customer's assembly line came within two shifts of stopping, and we air-freighted parts from our Hosur plant at a cost of ₹3 crore. The plant head and our head of human resources proposed moving two product lines to Hosur and holding firm. It was a respectable position, and many companies in our industry would have taken it.

What the contract saves and what it costs

I did not take it because, when we looked at the evidence, the grievance was largely justified. An independent audit of every contractor-supervised area at Chakan I found weaker machine guarding and thinner training than in areas we supervised ourselves. Our own injury data had shown the difference for at least two years, but we reported injuries by plant, not by type of employment, so nobody had noticed.

The quality figures told a similar story. Lines with a higher share of contract operators had lower first-pass yield, largely because of turnover: a contract operator who leaves after eight months takes the skill with him, and the next one starts again. In a gear plant, where the difference between an acceptable gear and a noisy one may depend on how carefully an operator sets up a fixture, that turnover is a quality risk. For electric-vehicle gears, which we were beginning to make in volume, it was a serious one.

The visible saving from contract labour is the wage difference, and it is large. The less visible costs are injuries, rework, scrap, customer complaints, recruitment and the slow erosion of trust in a workforce divided into two classes. Those costs are spread across budgets that nobody compares with the wage saving.

A settlement that pays for skill

We proposed a framework that the union accepted within a week. Of the plant's production roles, 420 were classed as core, meaning roles that need trained operation of a machine. These would become permanent over three years. They would be filled from contract workers already doing them, through a practical skills test with a fixed pass standard and a second attempt after retraining. Loading, cleaning and logistics roles would remain contracted, but under our own safety system, with our supervisors responsible for training. The union ended the go-slow. In March 2024 it signed a four-year settlement in which part of each wage increase depends on first-pass yield, attendance and safety.

The cost is real. At full run-rate the additional wages come to about ₹34 crore a year, and the go-slow cost us about ₹19 crore of output. By March 2026, 386 of the 420 roles had been filled. Of those who took the test, 112 failed at the first attempt and 71 of them passed after retraining. Some workers who had operated machines for years did not meet the standard, which was painful for them and for us.

The results so far justify the change. Reportable injuries at Chakan I fell by 58 per cent. First-pass yield rose from 93.1 to 96.4 per cent. The plant has lost no production days since the settlement, and the operators we train now stay. We are applying the same framework, more gradually, at our other plants.

I should acknowledge the arguments against. Permanent employment reduces flexibility, and a downturn will test us; we have planned for it with overtime limits and a smaller contracted buffer in non-core roles. Some competitors will keep lower wage costs, and a customer comparing quotations will not ask how our operators are employed. And a skills test can be designed to exclude as easily as to include, which is why we agreed its standard with the union in advance.

For directors of Indian manufacturers, I would suggest three questions. What share of the people operating machines on core lines are contract workers, and how long have they been there? Are injuries and quality reported by type of employment as well as by plant? And when the company calculates its saving from contract labour, does it also count what that labour costs in turnover and rework? A board that asks those questions will find out whether its flexibility is real, or whether it has quietly built a permanent workforce that it has chosen not to employ.