# Labour Turnover Rate Formula for Garment Factories: Monthly, Annual and Retention Rate

> Labour turnover rate formula for garment factories: monthly and annual rates, retention rate, a worked example, a free calculator and what a leaver costs.

**Source:** [https://scanerp.pro/blog/labour-turnover-rate-garment-factory.html](https://scanerp.pro/blog/labour-turnover-rate-garment-factory.html)

---

# Labour Turnover Rate Formula for Garment Factories: Monthly, Annual and Retention Rate

· November 18, 2026 · 9 min read Workforce

**Quick answer.** The labour turnover rate is the share of your average workforce that left during a period.

**Turnover % = separations in the period ÷ average headcount × 100**

Average headcount = (headcount at start + headcount at end) ÷ 2. Illustrative example: headcount **240** on the 1st, **250** at month end, **18** leavers. Average = 245. Monthly turnover = 18 ÷ 245 × 100 = **7.35%**.

Want this number from your own floor, every month? [Ask on WhatsApp →](https://wa.me/9779704778917?text=Hi%2C%20I%20used%20your%20labour%20turnover%20calculator.%20I%20want%20to%20see%20how%20Scan%20ERP%20tracks%20attendance%20and%20operators%20automatically.)

"People keep leaving" is a complaint, not a measure, and a factory cannot manage a complaint. Managing labour turnover in the garment industry starts with one number, calculated the same way every month, that shows whether the problem is growing, which lines it hits, and whether the people leaving are long-serving operators or last month's new joiners. This guide is about that calculation: the labour turnover rate formula, monthly and annual versions, voluntary and involuntary leavers, the retention rate, and how to put a cost on each leaver. For why sewing operators leave and what the research says reduces it, see our separate guide to [garment worker retention](https://scanerp.pro/blog/garment-factory-worker-retention.html).

## How do you calculate labour turnover rate?

**Labour turnover rate % = separations in the period ÷ average headcount × 100, where average headcount = (headcount at start + headcount at end) ÷ 2.** Illustrative example: 18 workers leave in a month and headcount moves from 240 to 250. Average headcount = 245, so monthly turnover = 18 ÷ 245 × 100 = 7.35%.

Turnover % = (Separations ÷ Average headcount) × 100

Where:

Average headcount = (Headcount at start + Headcount at end) ÷ 2

tutor2u calls this the separation rate: the number of employees who leave over a period, divided by the average number of employees in that period ([tutor2u, Labour turnover](https://tutor2u.net/business/topics/labour-turnover)). The US Bureau of Labor Statistics uses the same structure for its national statistics: the separations rate "is computed by dividing the number of separations by employment and multiplying that quotient by 100" ([BLS, JOLTS technical note](https://www.bls.gov/news.release/archives/jolts_03062024.htm)).

A **separation** is anyone who comes off the payroll for any reason. BLS defines separations as "all employees separated from the payroll during the month" and splits them into quits (voluntary), layoffs and discharges (involuntary), and other separations such as retirements, transfers, deaths and disability ([BLS, JOLTS definitions](https://www.bls.gov/jlt/jltdef.htm)).

### Worked example, step by step

**Illustrative example (one month, one factory):**

Headcount on the 1st: **240** · Headcount at month end: **250**

Leavers in the month: **18** (15 resigned or stopped coming, 3 dismissed) · Joiners: **28**

**Step 1:** Average headcount = (240 + 250) ÷ 2 = **245**

**Step 2:** Turnover = 18 ÷ 245 × 100 = **7.35%**

**Step 3 (check the roll):** 240 − 18 + 28 = 250, which matches the month-end headcount.

Monthly labour turnover = 7.35%

The roll check in step 3 is worth doing every month. If start − leavers + joiners does not equal the end headcount, someone has been missed, often a worker who stopped coming but was never marked as a leaver.

## How do you convert a monthly turnover rate to an annual rate?

**Add the separations of all 12 months, divide by the average headcount over the year and multiply by 100.** That is how the US Bureau of Labor Statistics computed annual separation rates before 2023. Multiplying one month's rate by 12 is only a rough projection, because a single month is rarely typical.

In BLS's words, annual levels "are the sum of the 12 published monthly levels", and annual rates were computed "by dividing the annual level by the Current Employment Statistics (CES) annual average employment level, and multiplying that quotient by 100" ([BLS, JOLTS release, March 2019](https://www.bls.gov/news.release/archives/jolts_03152019.htm)). Since its January 2023 data, BLS publishes "annual average rates" instead, dividing the sum of 12 monthly separations by the sum of 12 monthly employment levels. That gives an average *monthly* rate, not a full-year turnover rate, so check which one you are reading before comparing.

**Illustrative example:** if the 12 monthly separation counts add up to 196 and the average headcount across the year is 245, annual turnover = 196 ÷ 245 × 100 = **80.0%**. Projecting from the single month above (7.35% × 12) would have given 88.16%. The gap is the reason not to annualise from one month: festival months, peak-season months and slack months all look different.

A related shortcut is to add the 12 monthly rates. Tracey Smith of Numerical Insights notes that with the start-and-end average, "adding 12 monthly values doesn't equal the same number as the annual calculated value"; the article describes a third method, which the author matches to SHRM and CIPD definitions, under which the monthly values do add up to the annual figure ([Numerical Insights](https://www.numericalinsights.com/blog/how-to-choose-a-turnover-definition-and-correctly-measure-retention)). We recommend the BLS approach: keep monthly separation counts and monthly headcounts, and compute the annual rate from the totals.

## Voluntary, involuntary and replacement rates

One turnover number hides two very different problems. Split it using the same denominator:

| Measure (illustrative month) | Count | Calculation | Rate |
|---|---|---|---|
| **Voluntary turnover** (resigned, stopped coming) | 15 | 15 ÷ 245 × 100 | **6.12%** |
| **Involuntary turnover** (dismissed, laid off) | 3 | 3 ÷ 245 × 100 | **1.22%** |
| **Total turnover** (separation method) | 18 | 18 ÷ 245 × 100 | **7.35%** |
| **Replacement rate** (joiners who filled a vacancy; tutor2u counts all new hires, we count only replacement hires so expansion hires do not inflate it) | 20 | 20 ÷ 245 × 100 | **8.16%** |

Voluntary and involuntary rates add up to the total: 6.122% + 1.224% = 7.346%, shown rounded as 7.35%. The **replacement rate** divides new hires by the same average headcount ([tutor2u](https://tutor2u.net/business/topics/labour-turnover)). Count only joiners who filled a vacancy; in the example, 8 of the 28 joiners staffed a new line and are left out. A replacement rate below the separation rate means not every vacancy was refilled; above it, the factory is refilling past vacancies. Whether headcount actually shrinks also depends on expansion hires, which this rate leaves out.

Workers who simply stop coming are a hard case to record. Decide in writing after how many days of unexplained absence a worker is counted as a voluntary leaver, and use that rule every month, or the voluntary rate will move with whoever prepares the report. Absence that ends with the worker returning belongs in your [absenteeism rate](https://scanerp.pro/blog/absenteeism-rate-garment-factory.html), not in turnover.

## What is the labour retention rate formula?

**Retention rate % = workers on the roll at the start of the period who are still employed at the end ÷ headcount at the start × 100.** Joiners are left out. In the illustrative month, 13 of the 240 starting workers left, so retention = 227 ÷ 240 × 100 = 94.58%, not 100 − 7.35.

Retention % = (Starting workers still employed at end ÷ Headcount at start) × 100

Numerical Insights puts the point directly: "retention is not 100% - [Turnover Rate]", because retention follows only the group that was there at the start, while turnover counts every exit, including people hired during the period ([Numerical Insights](https://www.numericalinsights.com/blog/how-to-choose-a-turnover-definition-and-correctly-measure-retention)). In the example, 5 of the 18 leavers had joined that same month. Only 13 came from the starting roll of 240, which is why retention (94.58%) is higher than 100 − turnover (92.65%).

### Measure new-joiner turnover separately

Those 5 early leavers are the most expensive kind. We recommend a separate early-leaver rate: joiners who left within 90 days ÷ joiners in the intake × 100. At month end, 5 of 28 joiners have already left: 5 ÷ 28 × 100 = 17.86% so far. The intake’s 90-day window is still open, so recheck at day 90; the final rate can only be the same or higher. The reason to track it is in the research our retention guide summarises: at a large garment firm in Ethiopia, "more than 40% of workers" had left after 12 weeks, while workers "are hardly productive in the first 8 weeks" because they are being trained ([Abebe, Caria, Dercon and Hensel, IGC working paper, 2019](https://www.theigc.org/sites/default/files/2021/05/Abebe-et-al-2019-Working-paper-2.pdf)); at a menswear factory in Bangalore, 36% of first-time workers left within three months ([Stanford King Center, 2019](https://kingcenter.stanford.edu/news/socialization-training-key-factor-retaining-first-time-women-workers)). Those are findings from particular factories, not benchmarks, but they show why an annual average can hide where the losses happen.

## What a leaver costs a sewing line

We found no reliable published replacement cost per sewing operator for South Asian factories (our [retention guide](https://scanerp.pro/blog/garment-factory-worker-retention.html) explains why the figures in circulation do not apply), so build your own from four parts:

1. **The gap.** Until the post is filled, the operation she ran has less capacity. A sewing line runs at the pace of its slowest operation, so one empty seat at a bottleneck can cost more than one person's output. The mechanism is explained in our guide to [line balancing in the garment industry](https://scanerp.pro/blog/line-balancing-garment-industry.html).
2. **Hiring.** Time spent recruiting, testing and processing joining papers.
3. **Training.** Trainer time, and the machine and minutes the trainee uses while learning.
4. **The learning curve.** The difference between the minutes you pay a new operator and the standard minutes she produces while she gets up to speed.

Learning-curve cost = (Paid minutes × target efficiency − Standard minutes produced) × all-in CPM

Take both minute counts from the new operator's own output records until she reaches your target efficiency.

Multiplying paid minutes by your target efficiency means you only charge the new operator for the gap between her output and a trained operator’s, not for the normal loss every operator has. The all-in CPM is the same one used for quoting; our [garment costing guide](https://scanerp.pro/blog/garment-costing-cmt-calculation-guide.html) shows how the all-in cost per minute is built. Multiply the result by the number of early leavers and you have a figure management can compare against the cost of better induction.

## Common mistakes in labour turnover calculation

- **Dividing by month-end headcount.** In a month with heavy hiring, this makes turnover look lower. Use the average.
- **Annualising from one month.** 7.35% × 12 is a projection, not a measurement.
- **Reporting retention as 100 − turnover.** They measure different groups.
- **Leaving out workers who just stopped coming.** Without a written rule, they never appear as leavers.
- **Counting expansion hires as replacements.** It inflates the replacement rate.
- **Mixing contract or piece-rate workers in one month and leaving them out the next.** State which groups are on the roll and keep it fixed.
- **One factory-wide number.** Compute it per line and per section too; turnover concentrated on one line may point at that line's supervisor or work rather than at the labour market.

## Where the inputs come from

The formula needs four clean counts: headcount at start, headcount at end, leavers (with reason and joining date) and joiners. A paper register makes all four arguable. With [ZKTeco biometric attendance](https://scanerp.pro/blog/biometric-attendance-garment-factory.html) connected to Scan ERP, punches arrive automatically, so a worker who has stopped coming shows up as absent in the daily attendance report. Because production is tracked by QR bundle scans, per-operator output from the same scans shows how a new operator's efficiency moves week by week. Turnover sits alongside the other measures in our list of [garment factory KPIs for production managers](https://scanerp.pro/blog/garment-factory-kpi-metrics-production-manager.html).

## Sources

- US Bureau of Labor Statistics, [JOLTS definitions](https://www.bls.gov/jlt/jltdef.htm): separations, quits, layoffs and discharges, other separations.
- US Bureau of Labor Statistics, JOLTS news releases: [March 2019](https://www.bls.gov/news.release/archives/jolts_03152019.htm) (annual levels and annual rates, earlier method) and [March 2024](https://www.bls.gov/news.release/archives/jolts_03062024.htm) (separations rate definition and annual average rates).
- Tracey Smith, ["How to Choose a Turnover Definition and Correctly Measure Retention"](https://www.numericalinsights.com/blog/how-to-choose-a-turnover-definition-and-correctly-measure-retention), Numerical Insights.
- tutor2u, [Labour turnover](https://tutor2u.net/business/topics/labour-turnover): separation and replacement rates.
- Abebe, Caria, Dercon and Hensel, [IGC working paper](https://www.theigc.org/sites/default/files/2021/05/Abebe-et-al-2019-Working-paper-2.pdf), 2019; Stanford King Center, [socialization training and first-time women workers](https://kingcenter.stanford.edu/news/socialization-training-key-factor-retaining-first-time-women-workers), 2019 (both as cited in our retention guide).

### Know Who Stays, From Punches and Scans

Scan ERP reads attendance from biometric punches and per-operator output from QR bundle scans on your own floor, so absences and a new operator's weekly efficiency come from your own floor, not estimates. Ask the founder on WhatsApp and see it on a real line.

[Ask on WhatsApp](https://wa.me/9779704778917?text=Hi%2C%20I%20used%20your%20labour%20turnover%20calculator.%20I%20want%20to%20see%20how%20Scan%20ERP%20tracks%20attendance%20and%20operators%20automatically.) [Request a Free Demo](https://scanerp.pro/#contact)
