Indian manufacturing wages and cost structure: what sits behind a piece price

Why a wage comparison misleads

A comparison of hourly wages between countries is easy to find and easy to misuse. It compares one input, at an average, in a sector that may not match the part. A precision machined component and a hand welded frame sit in the same national statistic and have very different labour content.

The wage also ignores what the employer pays on top of it. Provident fund, insurance, gratuity and bonus are statutory or customary additions. Then there is the cost of supervision, quality, maintenance and the rework that a less stable process generates. A low wage with high rework can produce a higher cost per good part than a higher wage with a stable process.

The cost structure of an engineered part

For a metal part, material is usually the largest line. A copper busbar or a brass terminal carries the metal price in the quote, and the metal price moves with the exchange and the commodity market. The conversion cost, which covers labour, machine time, power and overhead, is the part the supplier actually controls.

That split explains why a wage increase does not always move a quote much. For illustration, if labour is a tenth of the cost, a ten per cent wage rise is a one per cent change in the total. If the part is hand assembled from many small components, labour may be a third of the cost and the same rise is material.

Ask the supplier to break the price into material, conversion, finish, packing and freight. A supplier who will not separate material from conversion is usually indexing the whole price to a commodity, which is a different commercial arrangement.

• Material, including yield loss and scrap credit
• Direct labour at the process steps
• Machine time and depreciation
• Power and utilities
• Tooling, consumables and maintenance
• Finish and outside processing
• Overhead, quality and administration
• Packing, inland freight and margin

Wages by skill band

Indian manufacturing wages are not one number. A helper, a machine operator, a setter, a toolmaker and a graduate engineer sit on different bands, and the gap between the operator and the toolmaker is wide. A plant that runs three shifts needs setters on every shift, and setters are the constraint in many small firms.

Minimum wages are set by state and by scheduled employment, and they are revised. A supplier in Maharashtra and a supplier in Tamil Nadu may have different statutory floors for a similar job.

The statutory additions to a wage

An Indian employer contributes to the Employees Provident Fund for eligible employees, commonly twelve per cent of basic pay, with a small additional charge for insurance and administration. The Employees State Insurance scheme applies to eligible employees below the wage ceiling, with an employer contribution of just over three per cent of wages.

Gratuity accrues for employees who complete the qualifying period, and the annual accrual is usually provided at around four to five per cent of basic pay. Bonus is payable to eligible employees within the statutory wage limit, at a minimum of eight and a third per cent of eligible salary, and it can be higher by agreement.

These are not optional extras that a buyer can ignore. They are part of the employment cost and appear in a proper should cost. Rates and thresholds change, and the labour codes have been reshaping the framework, so confirm the current position with the supplier and with a local adviser rather than relying on a figure from an old model.

Productivity is the multiplier

A low wage with a long cycle time can cost more than a higher wage with a short one. The comparison that matters is cost per good part, and that includes the cycle time, the setup time, the scrap rate and the rework. A plant that runs a stable process with quick changeovers will beat a cheaper plant that spends a shift setting up.

Automation changes the labour share. A CNC cell with a robot loader has fewer operators per part and a higher depreciation line. The quote may still be competitive, but the reason is machine time and yield rather than cheap hands. That is the kind of supplier that can hold a tolerance over a long programme.

The overheads that a wage table hides

Power is a real line in India. Tariffs vary by state and by category, and industrial users may pay a demand charge as well as a unit rate. A plating line or a furnace is power intensive, so the tariff can decide whether a process is viable at a site.

Building a should cost from the process

Start with the drawing and the routing. Estimate the material weight including yield, then the cycle time for each operation, then the labour minutes, then the power and consumables. Add overhead as a percentage that you can defend, and add a margin that reflects the volume and the risk.

The output is a range, not a number. A should cost that lands close to the quote is useful because it tells you which lines to discuss. A should cost that is far below the quote usually means an assumption is wrong, often the yield or the number of operations.

Use the model in the negotiation as a set of questions rather than an accusation. Ask why the material yield is higher than your estimate. A supplier who can explain the gap is a supplier who understands the cost.

1. 01

   Weight

   Calculate blank or bar weight and the material grade.

2. 02

   Routing

   List each operation and the machine it runs on.

3. 03

   Time

   Estimate cycle and setup time, then convert to labour minutes.

4. 04

   Consumables

   Add tooling, chemistry, power and packaging.

5. 05

   Overhead

   Apply a defensible rate and check the headcount.

6. 06

   Compare

   Set the quote against the range and discuss the gap.

Reviewing the model as the inputs move

Wages rise, the currency moves and metal prices change. A cost model dated two years ago is a historical document. For a long programme, agree which inputs are indexed and which are fixed. Metal is usually indexed to a published reference, currency to a stated source, and labour to an annual review.

Keep the index simple enough that both sides can calculate it without a dispute. A complicated formula with three sources and a lag is hard to audit, and it usually ends up being renegotiated anyway. A single reference per input, with a stated review date, works better.