Paying by the piece is one of the oldest arrangements in the history of work. Weavers were paid by the yard, spinners by the hank, harvesters by the bushel, and seamstresses by the finished garment. It is usually discussed as though it were a phase that industrial economies passed through and left behind. It is not a phase. It is a function of one variable, and that variable is how cheaply output can be counted.
A piece rate requires a countable piece
Piecework is not a management philosophy. It is what happens when someone can count the output cheaply enough to pay for it directly. This sounds obvious and has a consequence that is not: the history of piecework is not really a history of employment practices. It is a history of measurement.
A bushel of grain is countable without instruments. A yard of cloth needs only a rod. A finished garment counts itself. These were paid by the piece for centuries because the counting was free, obvious, and hard to dispute. Work that could not be counted this way was paid by time, not because anyone preferred time wages but because there was no alternative unit.
This explains a pattern that otherwise looks like ideological drift. Piece rates did not retreat during the twentieth century because employers became more generous or workers more powerful in some general sense. They retreated in specific settings where output stopped being individually attributable, which happened as production became more interdependent. On an assembly line, one person's output is determined by the line speed and by everyone upstream. Counting an individual's pieces stops meaning anything, so time wages take over. The unit of measurement dictated the pay basis, not the other way around.
- Ancient to medievalCounted by natureBushels, yards, hanks, finished garments. The unit is obvious and the counting is free, so payment follows output without anyone deciding it should.
- 16th to 18th c.The putting-out rateMerchants supply material to households and pay by the finished piece. The worker absorbs the cost of a bad batch, a slow week, and their own equipment.
- 19th c.The sweated tradesPiece rates in tailoring, chainmaking, and box making fall to levels that cannot sustain the worker, because rates are set unilaterally against desperate supply.
- 1909The Trade Boards ActBritain establishes tripartite boards to set minimum rates in named low-paying trades. Churchill argued it a national evil that anyone should receive less than a living wage for their utmost exertions.
- Early 20th c.The stopwatchScientific management makes new things countable by breaking tasks into timed components. Measurement expands and piece rates follow it into new work.
- Mid 20th c.Retreat on the lineInterdependent production makes individual output unattributable. Time wages dominate where counting a person's pieces stops meaning anything.
- 21st c.Counting becomes freeDeliveries, tickets, calls, rides, and tasks are all logged automatically. The precondition for piece rates returns across work that never had them.
What the rate actually transfers
A piece rate moves the risk of a slow day from the person who organizes the work to the person who does it, including every cause of slowness they do not control. This is the substance of the arrangement and it is routinely obscured by arguing about incentives.
Consider what determines how many pieces someone completes. Their own speed and skill, certainly. Also the quality of the material supplied, the condition of the equipment, whether the work arrives steadily or in bursts, whether the weather holds, whether the route is congested, whether demand appeared that day at all. A time wage places all of those on the employer, who has both the information and the authority to manage them. A piece rate places them on the worker, who typically has neither.
This is why the sweated trades were sweated. The nineteenth century found that where labour was abundant and desperate, and rates were set unilaterally by the party buying, piece rates could fall below subsistence while the work itself remained as demanding as ever. Workers responded by working longer, which increased supply and pushed rates down further. The mechanism is self-reinforcing and does not correct on its own, which is precisely why Parliament eventually intervened with the Trade Boards Act of 1909 to set minimum rates in the worst affected trades.
The lesson from that intervention is specific and often misread. Parliament did not abolish piecework. It set a floor under the rate. The problem was never that output was being counted. It was that one party set the price of a piece and the other could only choose how many pieces to attempt.
Time wage
- Slow day costsThe employer
- Bad material costsThe employer
- Idle time costsThe employer
- Rate set byNegotiation
- Worker optimizesHours present
- QualitySupervised
Piece rate
- Slow day costsThe worker
- Bad material costsThe worker
- Idle time costsThe worker
- Rate set byThe buyer
- Worker optimizesCounted output
- QualityWhatever passes
Measurement changes what the work becomes
People optimize the counted thing, and the counted thing is never quite the valuable thing. This is the second-order effect, and it is more consequential than the first because it changes the work itself rather than merely its price.
The gap is always present and its size varies. Paying a weaver by the yard rewards yards, which is close to what a bolt of cloth is for, so the distortion is small and quality inspection closes most of the rest. Paying by calls handled rewards ending calls, which is a different thing from resolving problems. Paying by tickets closed rewards closure, which can be achieved by solving an issue or by declaring it solved. Paying by deliveries completed rewards speed in a job where the uncounted variable is whether anyone got hurt.
Every piece-rate system therefore needs a second apparatus to police the difference between the counted thing and the wanted thing: inspection, quality penalties, customer satisfaction scores, safety rules. That apparatus is a cost, and it exists purely because the measure is a proxy. The cheaper measurement gets, the more tempting it becomes to pay on a proxy that is easy to count rather than on an outcome that is hard to define, and the larger the policing apparatus has to grow to compensate.
The final stage is the cost of having paid on a proxy. It is rarely counted against the savings that motivated the switch.
Why it is returning now
The precondition has been met across an enormous range of work that previously could not support it. Deliveries are logged with timestamps. Support interactions are recorded and categorized. Rides, tasks, keystrokes, code contributions, and response times are all captured automatically as a byproduct of the work passing through a system that records it.
Measurement is now nearly free, and it arrives without anyone deciding to measure. That is the genuinely new condition. Historically, counting output required someone to stand there and count, which meant measurement was a deliberate investment weighed against its benefit. Now the count is a side effect of the work being routed through software, and the marginal cost of paying on it is close to zero.
The consequences follow the historical pattern exactly. Where counting is possible, pay drifts toward the count. Where pay attaches to the count, risk shifts to the worker. Where risk shifts and rates are set unilaterally by the buying side, rates fall until something stops them. Nothing about this requires new intentions on anyone's part. It required only that measurement get cheap, and it did.
- 1DecisiveWho sets the rateThe sweated trades and a well-paid piece-rate machinist differ on this and little else. Unilateral rate-setting against desperate supply is the failure mode.
- 2Who absorbs the varianceBad material, idle time, a route with no demand. If the worker carries causes they cannot control, the rate is a risk transfer wearing the language of fairness.
- 3How far the count sits from the valueYards of cloth is close. Calls handled is far. The gap determines how much policing apparatus is needed and how badly the work is distorted.
- 4Whether a floor existsThe Trade Boards answer. A floor does not abolish the count, it removes the self-reinforcing spiral where working longer drives rates lower.
- 5Least discussedWhether the count is visible to the countedA worker who cannot see how the measure is computed cannot contest it, which returns the whole arrangement to a single party's discretion.
The question worth asking
The argument against piecework has usually been made as an argument against counting, and that argument is weak. Counting output is not inherently exploitative. A skilled worker on a fair piece rate can earn more than a time wage would pay them, retain real autonomy over their pace, and prefer the arrangement. Many historically did.
The argument that survives is narrower. Piece rates concentrate risk on the party least able to absorb it, they attach pay to a proxy that diverges from value, and they collapse toward subsistence whenever one side sets the rate and the supply of willing workers is abundant. None of those are consequences of measurement. They are consequences of who holds the pen.
So the useful question about any output-based arrangement is not whether the work is being counted. It is who sets the rate, who absorbs a bad day, how far the counted thing sits from the valuable thing, whether there is a floor, and whether the person being measured can see the measurement. Those five questions would have identified the sweated trades in 1890 and they identify the same structure today, under whatever name it currently travels.
Piecework never left. It was waiting for the counting to get cheap again.
Key takeaways
- Piece rates require countable output, so the history of piecework is a history of measurement rather than of management fashion.
- Piece rates retreated on assembly lines because interdependent production made individual output unattributable, not because of any general shift in values.
- The arrangement transfers the risk of a slow day to the worker, including every cause of slowness they do not control.
- The sweated trades collapsed because working longer increased supply and drove rates lower, a spiral that does not self-correct.
- The Trade Boards Act set a floor under the rate rather than abolishing the count, which correctly identified where the problem was.
- People optimize the counted thing, so every piece-rate system needs a policing apparatus to cover the gap between the measure and the value.
- Measurement is now a free byproduct of work passing through software, which restores the precondition across jobs that never had it.
This report was produced by Kariaa Research. All data, resources, and analysis are proprietary. For questions, contact research@kariaa.com.