Moon of Alabama Brecht quote
March 26, 2007
Incentive Pay

In a consulting gig today, I had yet another discussion on incentive pay. "Can we motivate these  people stronger by increasing the bonus part of their pay," I was asked. I recommended to abolish incentive payments at all. There are technical arguments against it and more serious philosophical reasons.

Let’s start with the technical problems:
In its simplest form you pay the guy who makes the nuts more if he produces more nuts. You do the same thing with the girl who makes the bolts. The guaranteed outcome will be either too many nuts or too many bolts and more money spent than needed to make an equal and sufficient number of nuts and bolts.

In more complex knowledge worker jobs there are simply no measurable outcomes one could tie  an incentive to. The number of lines of codes written by a programmer says nothing about the quality of the code or its long term maintainability (and costs.)

In one job I was told to better the productivity (value of output/cost of input) by 10% per year to increase my income by 10%, but the variables to do so, like moving the department out of an outrageous expensive office location, were out of my bounds. (The managing board of that media company also did of course not understand the mathematic limit value of "by 10% per year".)

In other situations there are general market interferences. You may well be the best in your trade and in an economic downturn the one who loses the least money. But most incentive pay systems will give you zero for achieving this.

The response to the above calamities is usually to make the system of benchmarking more complex. Lots of consultants make piles of money for presenting such fine tuning. But the inevitable outcome will be either totally confused and unmotivated workers or folks who start gaming a not supervisable system.

In a young company I was with, the head of the marketing department had his salary changed from fixed to variable to reflect the number of new subscribers he could bring in. Before the change he was quite successful in helping the company into a profitable realm. After the change the number of new subscribers per month soared, as did his income. A year later it turned out that most of these new subscribers left as soon as they could. The retention rates were catastrophic because the marketing campaigns were designed to maximize intake of subs, not to keep them once they were on board.

People who believe in incentives assume that money is the main human motivational factor. Rational people, they say, will always work more and work better if this leads to better pay. Essentially they think everybody is corrupt and they cite lots of examples to prove this point.

But that is not true. The fallacy here is to mistake cause and effect.

If one pays people a fair fixed amount for doing their job, they have no reason to game the system. To make them do a better job, some smiles, ice cream surprises on hot days and a bottle of good wine for Christmas will usually have the desired effect.

But as soon as you start to offer incentive pay you urge them to move away from doing a "better  job" to doing a "more profitable job" – more profitable for them that is. This is planting the seeds of corruption where none need to be.

The philosophy behind incentive pay is that of humans as rational beings – rational understood in the very reduced sense of a homo economicus, a useful mathematical model for economic science theorists, but without much resemblance to living creatures.

Adam Curtis makes this point in a much wider sense in his new three part BBC series The Trap: What Happened to Our Dream of Freedom.  He is on to something and I hope to have more on the thoughts he tells in future pieces.

Today the folks I talked to were not immediately convinced, but I did get them to think about the issue. The funny moment I needed came when I refused to make an offer for designing a "better incentive pay system" for them. I said: "such offer would only benefit me and deeply hurt your company."  Somehow, that answer seemed to be unexpected.

Comments

An incentive pay system works best for an economy transitioning from a pure socialist model to a capitalist model where skills are simple and can be readily measured.
Before 1978, Chinese workers were paid a flat salary regardless of their productivity output. Beginning that year, workers in clothing factories were paid according to how many pieces they made. Of course, productivity shot up. China’s entrepreneurs today are most likely those who were willing to make an early gamble to up productivity by starting their own businesses and factories.
The problem in China today is that now even Communist officials are quantitatively measured by how much investment they can attract, how much land they develop, etc. The downside is that they attract manufacturing (high capex/high pollution/high corruption) and development (high corruption/high rural unrest). That is why “socialism with Chinese characteristics” is hyper-capitalism. It’s a Republican’s wet dream!
Once an economy or skill becomes more qualitative (instead of quantitative), measuring and benchmarking just don’t work. Of course this would explain why Republicans in America want to downgrade peoples’ skills; they want something measurable at the expense of quality. And most Americans are dumb enough to eat up the BS; this was demonstrated in 2004 with the Bush re-election.
Large organizations, regardless of whether it is a nominally Marxist capitalist organization like the Chinese Communist Party and large companies like to find easy solutions and models to “manage” large groups of people. They forget that people are complex, and are motivated for different reasons. They hire consultants to find a magic formula when, in fact, there isn’t any. The answer is for executive/senior management to act like human beings and to treat employees as human being
, but more often than not, that is asking too much of them.
They simply don’t know how.

Posted by: Chris Marlowe | Mar 26 2007 19:11 utc | 1

che guevara wrote beautifully on this in his essay ‘socialism & man’ – i’m going to look it up myself after this post of yours, b

Posted by: remembereringgiap | Mar 26 2007 19:54 utc | 2

part of the text of che guevara

Posted by: remembereringgiap | Mar 26 2007 20:01 utc | 3

@Chris Marlowe Large organizations, regardless of whether it is a nominally Marxist capitalist organization like the Chinese Communist Party and large companies like to find easy solutions and models to “manage” large groups of people.
A couple tales from a time when small biz still thrived in US.
1- The owner of a single factory that manufactured tables. It was a union shop. He had one worker who he knew simply produced more than others performing the same function. Said the owner, “he runs most of the time, all day long. I don’t know how he keeps it up.” There was not a question of announcing bonuses, nor any expectation that other workers should perform at such a level. The owner simply sought a way to reward the one employee for his extraordinary productivity, without undermining morale among all workers. So he began inquiring about any special needs or interests that this employee or his family might have.
2- A family-owned and run machine tools factory did so well that it was constantly in need of more capital to expand and meet demand. When an employee excelled, and improved profits, they simply raised that employee’s pay. (Not a bonus, nothing announced in advance, simply recognition of a valuable work history.)
It was the lathe cutter I knew. A bright man, he often could look at the design blueprint developed by the engineers and actually improve the plans, thus reducing the time it took to cut and construct a particular machine.
The owning family tired of constantly raising money, remortgaging the houses of father and sons, and finally accepted a buy-out offer from a large conglomerate, offering ready capital for growth. Despite promises that family management would continue, the corporate accountants were soon in charge, and the new owners no longer appeared on the shop floor, where the old ones had always been.
After a couple of years of corporate alienation, it became a union shop. The corporate owners took a very strict line on compensation levels. The managers on the floor still wanted to reward my friend, the lathe operator, for his innovations, but their options were narrow. Eventually, they offered to let him leave the work floor for a couple hours on most day, no questions asked. He would go to the lockers and read.
b, Don’t know what this means, except that I wonder whether there is a human way to manage any organization that grows beyond a human scale. Numbers seem to gain ascendancy when there is a need to translate all events, production of various sorts into a common medium, to measure and compare. Inevitably, it becomes a false measure and a false comparison, as you demonstrate, betraying the actual, complex activities and the human motive for action.

Posted by: small coke | Mar 26 2007 20:45 utc | 4

hehe

Posted by: slothrop | Mar 26 2007 20:46 utc | 5

Would you want your kidney removed by a guy that has already carried out four operations because he wants to make more money? Would you want to fly an airliner whose pilot has worked extra hours because of incentive pay? Do you like a society where every judgment is quantitative? Do children learn more at school because their math books weigh seven pounds? Are men healthier because they weigh 300 pounds and butt each other’s heads? Are cars more pleasant because they are bulkier? I could go on but you get the idea. A single villa by Palladio is worth all the developments of the world.

Posted by: jlcg | Mar 26 2007 20:53 utc | 6

Barclay’s ceo made £22m this year, his employees, fuck all in comparison. God bless incentive pay.

Posted by: Cloned Poster | Mar 26 2007 21:11 utc | 7

excellent post b, especiall regarding the wisdom of your closing remark.

Posted by: anna missed | Mar 26 2007 22:21 utc | 8

this issue also relates to Adam Curtis’ 2nd part of The Trap, in which the modern world reduces all human activity to numerical values…and those same humans, when given quotas, etc. find ways to appear to comply rather than actually do their jobs — and do a worse job because of this peripheral performance standard.
“teaching the test” rather than teaching children is the saddest example. a sure way to make a kid hate school and associate learning with rote b.s.
btw, google took down The Trap, but someone has put it back up. to find it, go to google video and type in sitruc (curtis backwards) — I don’t wanna link cause I don’t want it to disappear. or it’s available via bit torrent.
curtis really has some interesting things to say about the current ideologies that blind us to their existence.

Posted by: fauxreal | Mar 26 2007 23:24 utc | 9

More often than not, strategic consulting is highly political.
In most instances, the CEO is looking for objective metrics to justify decisions he has already made. If the smart consultant, who is highly paid, can come up with something, the CEO can say to the board “I paid so and so X dollars (the more the better) to come up with this formula to reward productivity.”
If the consultant doesn’t come up with something, it’s all the consultant’s fault and he gets fired. The CEO is safe.
You know, it would be like Cheney in 2002 going to the CIA and offering an analyst 1M to find that Saddam Hussein had WMDs. (Don’t know if he did, but looking back on the current mess, that would have saved us a lot of blood and treasure.)
Of course, if the CEO doesn’t like what he hears, people can end up like Joe Wilson and Valerie Plame.
So, my advice to any consultant would be to ask the CEO, behind closed doors, “What’s your agenda?”
Then fit the consulting to fit what he wants. That’s the way to get repeat business in consulting. When he leaves that company and goes to another firm, follow him. That’s how consulting works.

Posted by: Chris Marlowe | Mar 27 2007 1:26 utc | 10

well, if everybody stopped selling the plebs in Kenya (as well as Kansas) fish and instead sold them fishing rods, who knows what might come of it.
and if we would all come together on the mindlessly obvious understanding that its most always better to trade fishing rods rather than fish, hence more engagement by the global masses in meaningful & tangible production rather that chronic consumption, disconnection & vegetation.
No, I am not going to give you any more fish, but heres some incentive. Its called a fishing rod.

Posted by: jony_b_cool | Mar 27 2007 1:46 utc | 11

Well… if management were any good it would manage, instead of hiring consultants to do what is ostensibly their job.
So if management’s job is not managing, what is it then?
Stealing. As is obvious from the compensation levels of the non-managing management relative to the compensation levels of the peope that do the work.
As Chris Marlowe has pointed out the consultants are just part of the non-management game.
As small coke points out, the further the “owners” of an enterprise are from the management of the enterprise the more likely it is to fail, if it is still in a non-monopoly stituation wher failure is an option, or the more abusive it gets of everyone involved if it is in a monopoly situation, as is more likely the case.
We all as individuals pony up our bucks to be invested in such absentee cash extraction operations, in the hope that somehow we can gain respite from them… through them. That we can become financially independent. As though independence in the “I’ve got mine… to hell with you” sense were possible in any sphere. We spend very many of our waking hours trying to earn the bucks to invest in such schemes…
Really… I hope that those who are not yet caught up in this mess can avoid it, and that more of use who are can really escape.
Poverty is the ultimate luxury. Very few can truly afford it.

Posted by: John Francis Lee | Mar 27 2007 8:49 utc | 12

For high knowledge jobs, another very effective argument is that if people are to work as a team, such as hospital operation room team, software developers, teachers, etc., etc. they must never be induced to compete with each other, as they have to not only collaborate, but interact smoothly and well, with intuitive understanding, and a can-do spirit that often boils down to helping their team-mates and rectifying mistakes.
‘Good performance’ of individuals should be judged qualitatively and only lead to ‘promotion’ or a ‘salary raise’ provided that more responsibilities are involved; the hierarchical structure must make sense, be accepted, and I would even say, conventional. But this must be done very prudently and parsimoniously, otherwise competition will set in anyway. (Who will get to be district manager?)
Another argument in some situations one can use is that strong authority is needed, some ultimate decisions rest with one person or ‘the board’ or whatever; and for everyone to obey the boss, to not undermine his authority, accept and do their best for ‘him’ or the super-ordinate goal, they need to feel ‘equal’ and not in different positions, fractioned by competition (functional slots set apart of course.) Traditional managers take to this.
One can point to some detrimental, even dangerous outcomes of competition that managers have not thought about in detail: a) sabotage, cheating, mobbing, etc. – can be very costly indeed! b) stifling innovation – the competition is set in rigid terms and workers will not exceed their ‘mandate’, – identify your true competitors! c) loss of flexibility d) destroying knowledge networks – and then? e) that ranking people automatically creates ‘poor performers’ or ‘disliked people’, who will then perform poorly, and probably be fired or leave, to then repeat the cycle (that depends on benchmarks and so on..etc.) f) losing valued employees to the competition over a matter of a small bonus – yikes! g) company reputation; employees will always talk out of the ‘shop’, what does that imply for clients, future hires, etc? Etc. Nice scary introduction..get them thinking..

Posted by: Noirette | Mar 27 2007 9:04 utc | 13

In related news I read today that swedish police has had an evaluation of their performance measurements. Apparently the last couple of years they have used the percentage of solved crimes as a tool to compare different districts. What carrots has been offered, I do not know, but it has worked its magic.
Now police spends more time on the crimes that is easiest to solve. Like picking up a local junkie. That is a real win/win because if it turns out to be a crime (possession of narcotics, generally) then the crime is immediately solved, and otherwise no harm done (except to the junkie, but they do not matter).
Anyway after the evalutation they probably will tweak the system, but not realise the fundamental flaws in their central planning.

Posted by: a swedish kind of death | Mar 27 2007 13:48 utc | 14

Great post, b. Summed up a lot of things I think, as do the comments. Thanks all.

Posted by: Argh | Mar 27 2007 13:50 utc | 15

…and what’s good for the private sector goose is good for the public sector gander.
Incentive pay systems in U.S. federal, state and local governments, usually called ‘Pay for Performance,’ undermine the concept of merit-based civil service systems in a very real way while undermining union rights in the last stronghold unions have in the U.S.
Performance ratings and wage disbursements are placed in the hands of managers, with little to no oversight/transparency as to how the ratings are being determined. The few federal demonstration projects that have been done have been predictable for workers: low morale, capable and critical people going without cost-of-living increases, workers feeling jobs are devalued, etc. It destabilizes a work environment that is supposed to be stable and empathize the retention and sharing of institutional knowledge for the service of the public good. It further opens the space for the politicization of public service. Bling for sycophants, crumbs for dedicated civil servants. Even if it isn’t the case, it looks like favoritism, patronage and spoils in place of merit, basically anything but ‘pay for performance.’
The thing is, despite all the right-wing neolib complaints that government workers have a lifetime tenure and complete job security, civil service rules allow for disciplinary proceedings and firing with cause. It just requires competent management to document the reasons, and when there is a union contract, working with employees and give them a chance to contribute, and giving the employee a fair chance to state his/her side of the situation. If management can’t pull that off, how are they not going to completely fuck up pay for performance?
The DHS and DOD are the supposed to be the proving grounds at the federal level, but federal rank and file workers have delayed and pushed back pay for performance along with provisions that refer to fed workers’ collective bargaining rights as “consultation rights.”
And it’s hard to ignore that unions in the U.S. still have some numbers to work with in federal and state/local govts. 36% govt workers are members. Compared to the 7.4% membership in the private sector, the Business Roundtable & the Club for Growth have some work to do on the government side of the ledger.
good article here about the defeat of the DOD pay system and how this could all backfire on the neolib agenda if unions & workers step up.

Posted by: served_cold | Mar 27 2007 14:55 utc | 16

wow. sorry for the long comment. if you skipped over it, the linked article at the end is the worthwhile part. maybe.

Posted by: served_cold | Mar 27 2007 14:58 utc | 17

There are various types of incentives and various types of motivation: there is the long-term motivation of maintaining a roof over one’s head and securing one’s retirement. That is generally a matter of monetary incentives.
But then again, there is the day-to-day motivation to keep doing one’s job well. That is often not just a matter of money, it depends on subjective psychological factors.
These factors include a sense of being part of the company and the team, of being respected as a person and a professional and of being needed.
When compaines start to look on their workforce as just another commodity to be purchased and maintained at the lowest possible expense, people certainly start to notice it and their day-to-day performance suffers.

Posted by: ralphieboy | Mar 27 2007 14:58 utc | 18

yet another in the long series of inferior substitutes.
for the natural human instinct of workman[sic]ship [hat tip Veblen] and joy in the exercise of our faculties, substitute quantitative metrics. having thus demoralised the working person, use fear and competition and greed to remotivate them.
result: more/faster wealth accumulation for the greedheads running the show, and shoddier goods (but in larger quantities) for the masses. and more rapid liquidation of all the resources that feed the accelerated production process.
the obsession with quantification seems to me like an Asperger’s Syndrome behaviour, generalised to an entire culture. it’s also a controlfreak fantasy, going as far back as the OT assertion that G-d knows the number of hairs on your head and not a sparrow falls without Management making note of it. we “understand” math (actually very few of us understand anything but the most elementary math, kid stuff compared to real mathematicians who operate in spaces weirder and more flexible and surreal than most people’s ganja reveries) and so we think that if we can reduce reality to math, we can understand and control reality: other people, nature, ourselves.
the old hammer, nail thing. having a hammer (math) is a good thing, and useful. but if everything starts looking nail-shaped it’s not so good. onions for example are hard to transplant using a hammer, and it’s hard to educate kids, care for the sick, or produce quality work or a decent social atmosphere using nothing but numbers…

Posted by: DeAnander | Mar 27 2007 23:32 utc | 19

De- paraphrasing Curtis – he puts the source of this mindset on game theory extrapolated to all human endeavors as theorized by John Nash –who was a paranoid schizophrenic who had to be involuntarily hospitalized. He thought people were out ot get him, he heard voices telling him this same thing, and all this time did not recognize that he was delusional. Unlike “A Beautiful Mind,” in real life Nash was an s.o.b. to be around and his view of “equilibrium” in human relationships was called “Fuck you buddy.”
when these “great minds” tried out their theory on secretaries they worked with (the “great minds” operated from the mindset that everyone would try to fuck over everyone else…) — these secretaries cooperated and didn’t try to destroy others — and thus defied the “great minds'” assumptions.
so, of course, the theorists decided that the secretaries were unsuitable subjects, not that their theory was bullshit.
sort of like the neocons/bush et al now regarding foreign policy.

Posted by: fauxreal | Mar 28 2007 0:30 utc | 20

askod :
The IRS in the US has done the same thing. All their examiners look for easy targets and let the big tax evaders off completely. The big evaders tie ’em up in court forever. No bonus there.
Do you think this comes under the heading of “unforeseen consequences”?
I don’t. I think it comes under the heading of “plausible deniablity” when the question of why the non taxpaying grillionaires go unaudited.
More evidence of the real motives of the non-managers’ use of consultants.

Posted by: John Francis Lee | Mar 28 2007 5:34 utc | 21

The popularity of this kind of managemet tools among orgnisations’ top management is the consequence of one of the fundamental laws of Systems Theory, namely the law of requisite variety, which says: “The variety of a control subsystem must be equal or superior to the variety of the controlled system”. Variety being a measure of the number of distinct states a system can be in.

Applied to a complex system like a human group, which is fractal (i.e. the level of complexity remains the same at any level of the system), the law of requisite variety means that a small number of persons (for example the top management of an organisation), even highly skilled and informed, cannot master the variety/complexity of the system it has to govern, hence will not be able to tackle a number of situations. The bigger the organisation, the wider the gap.

The only case in which the law of requisite variety is respected is when the control system and the controlled system coincide, i.e. when every element of the controlled system is at the same time part of the control system, which means empowerment of the systems actors at all levels. In a company, the co-operative form is probably the closest to this state.

The reverse corollary is that any controlling group (i.e. the management) tends, consciously or unconsciously, to reduce the complexity of the system it seeks to control. To reduce cognitive dissonance, managers (and their advisers) tend to adopt a simplified representation of the people they manage and of their motivations (hence the “homo oeconomicus”), and to implement this simplistic view in the organisation’s design and management tools (like incentives). Psychological research has confirmed this conclusion: most of managers have actually a reductionist view of their subordinates.

Posted by: Melanchthon | Mar 28 2007 14:05 utc | 22

Melanchthon, thanks for this — I had no idea about systems theory, or that Margaret Mead and husband Gregory Bateson applied ” … systems theory, such as positive and negative feedback, into the social sciences”.
Your hypothesis (“a small number of persons (for example the top management of an organisation), even highly skilled and informed, cannot master the variety/complexity of the system it has to govern,”) is wonderfully succinct.
And the corrollary that a system which has the controlled part equal to the controllers reminds me of just how pleasant it is to work in a group without direct supervision, where the members encourage, direct and discipline themselves.
Something I’ve noticed about human beings vs. other social or “pack” animals — that people can take turns as the “alpha dog” or leader based on the situation; for example the bus driver drives the bus, but yields control to the tour guide as necessary, and both follow the direction of the traffic cop.
This is another part of group dynamics that I find fascinating.

Posted by: jonku | Mar 28 2007 16:16 utc | 23

Something I’ve noticed about human beings vs. other social or “pack” animals — that people can take turns as the “alpha dog” or leader based on the situation
Bookchin points out that this is in fact the model of leadership in most pre-imperial cultures. anthropologists and conquerors often impose their ideas of leadership when asking to talk to the “head man” or “chief” when in fact an indigenous culture has no concept of institutionalised leadership but only fluid and situational leadership. just as early naturalists utterly misunderstood the functioning of the beehive by assigning institutional power and authority to the “queen” bee, when decisions are made not by the almighty queen (who is more of an egg factory than a manager — though she lives long than worker bees there’s no evidence that she “plans” or functions as air traffic controller or any such centralised image of power) but by complex interactions and adjustments of the entire hive as a communal organism.
good points Melanchthon and nicely summarised. [goes away to think]

Posted by: DeAnander | Mar 28 2007 19:30 utc | 24

Do you think this comes under the heading of “unforeseen consequences”?
NO. there are different division of IRS to go after Rich Folk. DaddyBush slashed this. Consequences most deliberate. If they want to curtail activities of the rich, they budget doing so.

Posted by: jj | Mar 28 2007 19:54 utc | 25

The trouble with using quantitative measurements for performance are:
1. They are useful to some degree to weed out inefficiency among more junior people, and those not smart enough to figure out the game;
2. For those who are smart/senior enough to figure out the measurements game; they figure out how to game the system by measuring their own performance more favorably to make them look good.
That is why purely quantitative measurements are such BS, and are simply an excuse for management laziness and incompetence. Robert McNamara believed in the “body count” as a performance measure in the Vietnam War; as a result US troops went out and killed VN civilians to meet their weekly kill quotas.
That explains why the US won the Vietnam war… (sarcasm).

Posted by: Chris Marlowe | Mar 30 2007 19:57 utc | 26

Actually what we’ve got seems uncannily similar to the old Soviet model – people being paid to manufacture goods by tonnage, therefore they make heavier and heavier products until the chandeliers literally pull the ceiling down over your head. Or the old Krokodil cartoon, of the manager of a Soviet nail factory being given an award for exceeding tonnage targets – then we see that his product is one massive nail!
http://www.vdare.com/roberts/soviet_economy.htm
“The Soviet Union wasted its resources because the success indicator for Soviet managers allowed them to be successful even though their output was poorly related to the needs of users. Soviet output basically satisfied no one but the statistician measuring it.”
Sound familiar?

Posted by: badnewswade | Apr 9 2007 12:25 utc | 27