That story was interesting, but it is properly worth pointing out that the productivity was fruit picking, a line of work were the results are easy to measure and compare, the requirements of skill are modest and the difference between the highest and lowest performer small. Care should therefore be taken before this is applied to other areas where these assumptions no longer holds, such as software development.
"a line of work were the results are easy to measure and compare, the requirements of skill are modest and the difference between the highest and lowest performer small."
Definitely. And it should also be noted that the work in the story is seasonal. I'm not sure that the incessant competition in this performance-pay model would be entirely conducive to retaining staff long-term. However, this may not be a big problem with unskilled industries where staff turnover is high anyway.
Yes. The seasonal thing is central. Important assumptions here are basically:
A - Workers have no incentive not to screw the employer.
B - Employing/Retaining staff's not an issue. This is also part of the seasonal work. It's not worth looking for a different job when it's only a couple of weeks.
If this were a more perfect market (infinite time, perfectly informed participants ...), the decreasing marginal rates (to avoid being over generous), would disappear. You'd be better off pinching your neighbour's picker that picks twice 2x fruit with 2 X minimum wage then keeping you 0.9x picker at 1 X minimum wage.
True. That is what makes it a great read (I thought it was) & a great experiment. You can easily understand the pieces.
But hopefully, some rules & guidelines might emerge from this type of experimentation. These would (if they are good rules & guidelines) be applicable more widely.
It'd be difficult to work out the laws of mechanics by examining an engine. But playing with a lead ball in a lab can eventually help you build a better engine.
Also be careful about side effects. In my town the city tried a neat rule for garbage men: finish your route and you can go home, but we'll pay you the full-time wage. This had two results:
1) garbage routes were completed more quickly
2) on-job injuries (pulled backs and hernias) went way up
Joel would say that, since "incentive pay" comes out of his profit!
The prospect of more money wouldn't convince me to do a job I actively disliked but at the end of the day I'm a professional. I do what I do for money. The company isn't shy about boosting its revenue and making profits, why should I be?
Joel would be one of the least concerned about how incentive pay affects his profit - he spends way more than average per developer, but in the areas where he thinks it has a positive effect, like above average salary, good desks/chair, etc.
You may dislike Joel, but but that's a stupid argument.
I like Joel, and have subscribed to his blog for years, and what he say often seems to make a lot of sense. However, I know a lot of people in this industry, and I don't know a single one who uses a Fog Creek product. That, to me, speaks volumes.
That works because the managers can do the job of their employees. Most tech fields to which I've been exposed, managers might have once been able to do the job of employees, but are either out of touch with the technology of today or their skills have ossified.
Not to mention the managers that could never do the job in the first place...
I'd hate to be a worker there - you go from an easy job to a really hard one - without really getting more money. I wonder if they took that into account.
If I knew about this, I'd find a different farm to work on.
Also adding teams added a lot of stress to the workers, again that's not someplace I'd like to work.
I think they should have rolled back the teams idea and stopped.
Well if everyone followed your path, they would find themselves making minimum wage. You could chose to work at the minimum wage farm, or the productive farm where you would be making more than minimum wage. You would also find that the minimum wage farm would be likely to go out of business.
... or the productive farm where you would be making more than minimum wage.
There is nothing to say that you will get paid more working in the "productive" farm, especially when "high productivity" is defined as "you doing more work for less money with the same tools".
The farmer has all the data, he sets your pay, and he wants you to work for minimum wage. The trick used to boost the amount of work done was to reduce the influence that the workers had on their pay, not to increase their pay.
Doing some quick calculations using the numbers in the paper linked to from the article, (and assuming that the farmer was successful in setting his pay structure so that those working the least were paid a minimum wage of £5/hr), those working the hardest worked harder (from 8.05 kg/hr to 12.16 kg/hr) but got paid less (from £22.45/hr to £12.28/hr).
Assuming you don't care whether the farm goes out of business or not (you're only working there for one summer) which one would you like to work in again?
No, the implication was that the whole point of doing this was to get as much output as possible while paying minimum wage.
That was the whole premise - the whole thing about tuning the piece rate - it was to make sure each person earned minimum wage, and no more, but also not less, because otherwise they are paying extra (to make up the difference), without getting more fruit.
I'm getting down mods, and you're getting ups. But I read it _again_ and I don't see anywhere at all where it says that!
Exactly the opposite: "their wages didn't fall below the legal hourly minimum. Farmer Smith tried to adjust the piece rate each day so that it was always adequate but never generous"
Where are you seeing that he is paying more? All I see is him paying the same, but getting more output via competition.
And again:
"(If the managers made a mistake in their estimate, and the pickers didn't earn minimum wage, Farmer Smith would make up the shortfall"
I went over it again and had a look at the linked paper.
The farmer originally had the minimum wage problem. The piece rate needed to keep everyone over minimum wage regardless of skill level or field quality. So
(Piece rate) = (Minimum Wage) / (the lowest performer's fruit quantity).
You can scale incentive on additional output, but it has to go up from there.
In this environment, he could probably work out a pay scale that works & it is to his advantage if worker earn more then minimum wage (because he probably pays less for marginal output).
Problem was that another productivity factor is the field. So, if he sets the above rate based on the worst field he is over paying workers on the best field. This is where 'generous' comes in. It means they were getting paid for being on a good field not working faster. His solution was scaling based on group productivity.
In a perfectly individualistic market, this is a good system . But the workers colluded to work less for the same pay (as opposed to getting paid more).
The economist 'fixed' that by removing the group incentive (they call it an externality) to pick less.
I think the most accurate description of Farmer Smith's intention would be to maximize the amount of productivity he received per dollar paid. production/cost = x Solve for x, maximizing x. His objective was to maximize x, not minimize cost.
Exactly the opposite: "their wages didn't fall below the legal hourly minimum. Farmer Smith tried to adjust the piece rate each day so that it was always adequate but never generous"
That was before the economists tried the new payment methods.
The entire paragraph, for reference:
The owner had been paying a piece rate - a rate per kilogram of fruit - but also needed to ensure that whether pickers spent the day on a bountiful field or a sparse one, their wages didn't fall below the legal hourly minimum. Farmer Smith tried to adjust the piece rate each day so that it was always adequate but never generous: The more the work force picked, the lower the piece rate. But his workers were outwitting him by keeping an eye on each other, making sure nobody picked too quickly, and thus collectively slowing down and cranking up the piece rate.
To summarize, the farmer was paying a modified market rate for the productivity: using minimum wage as the mean to determine the rate to be paid out. The problem was that the workers were the only supply and they were collaborating to lower the output (even in good fields) so that they would work less for the same pay.
A good analogy for this would be OPEC deciding to lower oil production. The oil supply is there, but by exporting less they do less work and increase the market value by limiting supply.
After the new methods of payment were introduced, the more productive workers could earn more by being more productive.
After the new methods of payment were introduced, the more productive workers could earn more by being more productive.
The change was to do the calculation of how much to pay per kg before the day's work, rather than after. The most productive were always paid the most, and the least productive were still paid as close to minimum wage as possible.
With the rate set at the start of the day, the people working in a particular field had no control over that day's pay per kg. As noted in footnote 5, the workers also had little control over the price for future days, because it was harder for them to correlate their productivity with the pay rate:
First, ... it is difficult for workers to disentangle changes in the piece rate due to changing conditions and those due to management learning about workers' true ability ... Second, the effect of a worker's current performance on the unit wage she faces in the future is weak as the unit wage is field-day specific and workers are reallocated to different fields in different days
So, instead of using the day's average, the farmer sets the rate based on an historical average, or a direct measurement of how hard a particular field is to work. In either case, it disconnects how hard someone works, and how much the they get paid that day. This makes collusion between workers harder, as they can no-longer game the system by forcing those directly working with them to work less. They would need co-operation between everyone working on the farm - a level of control that only the farm owner has.
Looking at Table 1, you'll see that productivity rises from 4.92 kg/hr to 8.06 kg/hr. The difference between the hardest workers and the weakest workers can be approximated using the standard deviation, which went from 3.13 to 4.10. Putting that into a little table:
weakest workers hardest workers
kg/hr before 1.79 8.05
kg/hr after 4.93 12.16
Now, we know that the farmer will set the pay rate so that the weakest workers receive minimum wage. For simplicity, we can call the minimum wage £5 an hour. In the "before" case, he must pay £5 for 1.79 kg, so the pay rate = 5/1.79 = £2.79/kg. And "after", he must pay only 5/4.93 = £1.01/kg. This gives the following table:
weakest workers hardest workers
before £5.00/hr £22.45/hr
after £5.00/hr £12.28/hr
Now you can clearly see ars's worry. The hardest workers are picking 50% more, but are being paid almost 50% less, while the weakest workers pick 175% more, but are paid the same amount.
No wonder the farmer was pleased, but it is hard to see how the workers will be happy with the "work harder and then you will get paid more" line.
It's interesting that the deviation went down when the pay scale was 'fixed.' Makes you wonder if it was the same workers that were productive or not. Maybe it was a good opportunity to make a good wage if you were willing to take social flak.
I'd like to point out that this is coincidental. The standard deviation might have gone up to the benefit of the (hardest) workers. The farmer would probably still been better off too.
Another interesting thing that you can clearly see in your numbers what a blow to the farmer (and advantage to the workers)it is to have weak workers. They set the pay rate.
The next con could be to collectively hire bad workers.
Oh, the standard deviation did go up, just not enough to account for the increased average. The deviation went from 3.13 to 4.10, but the number which matters for the hardest working is ratio between them and the least working, which went down from 9:2 to 5:2.
With manual labour there is a limit on the amount of fruit that one person can pick per hour, hour after hour. That caps how much the most dedicated person can work, they simply can't work any harder without inventing better tools to do the job. (Which is where us technologists come in :-)
Essentially, it is harder to be 5x more productive than someone motivated than someone unmotivated, so when everyone is motivated and the pay is kept to the minimum, it is not a coincidence that the hardest workers will get paid less.
Welcome to the free labour market, I guess. PG has an essay recommending to work where your (presumably) better than average productivity can be easily measured, so that you can be remunerated fairly for it. This seems to be a corollary: don't work somewhere where it is hard to work more productively than everyone else, or you'll get shafted if they start working just a little harder.
Minimum wage is the regulation that's causing all the trouble here.
Without minimum wage, Many of Farmer Smith's problems would have gone away. He could just pay the lowest rate at which he could still get enough workers to pick his fields. He still might have to adjust the rate by field, bit once that's done he wouldn't have to worry about motivating workers.
Remember though that their original rates were inflated due to the collusion between the workers. They are taking a pay cut of course, but that's because they can't control the supply as easily.
