Must I Retire Now? Must You?

I will leave the philosophical assumptions unquestioned.  What about the economic assumptions?

1) What happens if everyone follows the philosopher’s advice and
retires so long as they are below the median of the unemployed?  Is
there a stable equilibrium?  Yes, in equilibrium every worker with a
job must be better than the average worker without a job.   This
certainly seems possible although it is hard to see how it is optimal –
can no change in wages or job assignments make it beneficial to hire
more workers?  The fixity of jobs assumption is very strong.

2) More generally, if workers are
paid their marginal product and are appropriately assigned (e.g. better doctors work on harder cases) then no worker need retire.  With appropriate assignment, when a below-median doctor does retire he would not be replaced by an above-median doctor.  Instead, the new better doctor would be slotted in for
more difficult work, everyone else would move down slightly and the
retiring doctor would be replaced by one only marginally better. 

3) What happens in general equilibrium?  With flexible markets everyone gets a job so the worker who retires because he is below median is replaced by a worker from another industry.   It’s no longer obvious that this is optimal.

Most generallly, comparative advantage tells us that markets find a place for even the lowest-quality workers.  For the argument to apply we need a relatively fixed number of jobs, relatively fixed wages and a large reserve army to draw from.  Supreme Court justices come to mind.


Another company is, does a company actually benefit from hiring better quality workers?

In an inverse of this argument, I once worked at a hedge fund with a policy that any new hire had to be better than the median existing worker, thus moving the talent level up. Result: major headlines when the company nearly went bust, had to layoff most of the staff. It wasn't LongTermCapital but tried to hire a similar level of superstars.

Likewise, this is reminiscient of the McKinsey-taught policy that companies should rank their workers 1-5 and try to get rid of all the 5's. Enron is the most famous company known to have used this method. HP also started using it during the Carly Fiorina era of value destruction/declining stock prices.

There are least 3 possible conclusions implied by these events:
1. You are better off with less talented but less overconfident people.
2. You are better off with more talented people, but you don't want to rank them or tell them they are more talented, b/c (a) intra-firm competition will be too extreme and demoralizing and (b) the better employees are more likely to leave than the worse ones
3. The definition of "talent" should not be an obvious one like smarts, work ethic, etc., but a less obvious one like "safe, reliable, a company man but not a superstar"

oops, I meant "another question to ask is". I guess I'd better retire from the comments!

One big assumption here is that older workers and younger workers are substitutes in production. Actually, they are more likely to be compliments. The older workers have built up a lifetime of human capital that is essentially irreplaceable. In some cases that human capital might become obsolete, due to technological change or some other change. But in most cases, it is a great lose when an older worker retires because he or she has built a treasure trove of specific knowledge and contacts that you cannot get elsewhere.

That being said, I believe that if we were to retire half of those NCOs of paygrade E-5 or above in the US Air Force, and probably other branches of the military, they would not be missed a bit.

So if the idea is to hire someone better than the person being retired, what makes the employer so sure they're not going to screw up a second time?

Eddie has it exactly right. A coherent argument can be made for Professor Smilansky's claim with respect to government employees and employees in professions with legally fixed populations (the usual result of successful professional licensing associations) only. This argument may apply reasonably well to people in certain corporate heirarchies, especially in highly regulated or inefficient industries as well. Even in these cases, semi-competent employees will almost certainly do more good by keeping their job and giving the surplus over the best paying alternative job to some charity.

Even in the professor's world the sun doesn't quite revolve around the earth. Given a shortage of workers in a sector compensation rises - even if that sector is a government one.

Compensation may amount to allowances on top of wages, but the net effect is the same, and even in a unionised sector wages are not infinitely non-elastic.

I think the opposite happens. Aren't the most talented workers more likely to become rich and retire early, while the less talented ones will have to work more to save enough for retirement?

I'm a forty-something in early semi-retirement. For what it's worth, in the computer technology sector it's gotta be some combination of luck and talent. My options ended up OK, someone else with equal talent ended up under water. When I retired, I was probably replaced by someone of similar talent who was not as lucky. (There is also the factor that, in my experience, most technologists will hold their options until they go underwater. They're too afraid of taxes.)

"Assume ... that the number of jobs is relatively fixed ... ."

This is the same assumption that the French Government made in setting up their wildly successful no more than 35 hours a week plan. It has been a staple of liberal economic nostrums since Hector was a pup, and it hasn't worked yet. why do they continue? Why read anything that rests on that assumption?

Let's see:

a) It is virtually impossible to fairly measure one's ability relative to one's peers, because of the multitude of factors involved, and the multitude of future scenarios which may exist
b) Even if you could do (a), there is usually a significant cost in bringing some new person on, relative to the cost of retaining you. That cost should be factored into your value to the public good - essentially the closer you are to retirement, the less worthwhile it would be to replace you because of the switching costs.
c) B is balanced by the costs to the public of your pension/payouts should you retire.
d) One should not work for the public good. One should work for one's own benefit, and any benefit to the public will be coincidental. (Objectivist-style argument)
e) If you could measure yourself relative to your peers, and if you could properly calculate the benefit of your effort for the public until retirement, and you could properly calculate the switching costs of leaving, and you could properly calculate the net public benefit for someone to come on to replace you, and you were motivated by a desire to maximize the public good rather than maximize your personal good, then you could perform the analysis described in the initial post.
f)Note that if you decide to 'retire', in order to maximize the public good it is essential that you forego any pension/payments/benefits from the firm. Because those are public costs with no good whatsoever.
g) In fact, if maximizing public good was your objective, you would take the minimal salary possible.
h) Which would mean that were you to 'retire', you would have no savings and no means of supporting yourself. Which would demand that you either find another job immediately, or commit suicide by jumping into a furnace, to minimize the public cost of your unemployment/funeral/etc

Gives new meaning to 'Work until you die'

It also strikes me that if you could estimate all those costs and benefits, deliberately reducing your own salary would increase your net benefit, making it harder to justify bringing someone on to replace you. As you got closer to retirement, if you were satisfied with the level of public benefit you brought, you could gradually increase your salary as the switching cost became signficant

The Myth of Measurement

The inputs for this microeconomic utility calculus must come from an allocation of estimated total product among factors of production. There is no objectively correct method of allocating such costs. Rather cost accounting methods are driven by the decision processes they are intended to inform. The decision between methods is always subjective and always subject to bias. A simple work-around is to assume that compensation approximates marginal product. But such a price signal can only be taken at face value in an enterprise fully subjected to market forces (as mentioned in other comments above).

Its not that measurement is impossible, only that it is always an estimate and usually a very rough estimate. Synergies of two factors are not easily bifurcated. Synergies of multiple factors are nigh impossible to specifically identify with particular factors. Outside this margin of measurement uncertainty market forces are likely to obviate the need for any psycologically tortured self-selection. Positions shielded from market discipline may yield only the smallest indirect product regardless of the talent of the person who occupies them.

Most people of working age don't have the option to "retire immediately"...if they retired from career "A", they would need to find another job in some other career field. So, assessment of effect on society as a whole would depend on their *comparative* skills in career "A" versus all the other possible careers. (Obvious exception where public safety is concerned)

Seems like an exceptionally shallow argument.

I wanna write something clever but since there's an even chance I'm no better than average, I'm not going to bother. In fact, I really never should have gone to work in the first place.

This argument assumes complete substitution with equal competence among all jobs. Someone who is below average at a particular job may be above average at another position. It is part of management's task to figure out where someone best contributes to the firm/organization.

Another thing to consider is the substitution cost. Finding and bringing in some one new generally costs about 75% of annual personnel costs. Assume the new person evenutally gets up to the 60 percentile in performance. To outperform financially the 50 percentile worker will take 7.5 years. The 40 percentile person 3.25 years. So, the current person would have to be performing below the 40 percentile to make the risk worth it. Because if you end up with another 40%er, then you have the 75% of personnel costs to expend again on an immediate basis.

Lets all spend some time in the business world shall we.

there is usually a significant cost in bringing some new person on, relative to the cost of retaining you

Comments for this post are closed