[OPE-L:3461] Re: More on skilled labour

Steve Keen (s.keen@uws.edu.au)
Thu, 17 Oct 1996 14:39:06 -0700 (PDT)

[ show plain text ]

Hi Allin,

I think you might find my answer frustrating, but here goes!

What you have argued does not IMO distinguish, in the manner you intend,
between the case I am putting--that skilled labor can generate surplus
value in addition to that attributable to simple labor--and the case
Sweezy/Meek put--that the additional productivity of skilled labor
simply reproduces the training time involved, with the side effect that
training is not an additional source of surplus value.

In fact, it presents the case for "my" side, because you show that if
training converts a single finger typist into a touch-typist, and this
doubles the typing speed, then an additional source of surplus value has
been created. In my terms, the Sweezy coefficient you use is the
exchange-value of the training, while the assumption of a doubling of
output is the use-value.

What you call the Sweezy coefficient, I would call the Hilferding


Allin Cottrell wrote:
> Steve:
> > > > Let's apply the use-value/exchange-value analysis to this.
> Allin:
> > > [i.e. the analysis of skilled versus simple labour]
> Steve:
> > No, the use-value/exchange-value analysis applies to a lot
> > more than just skilled vs simple labor... [amplification
> of misunderstanding follows]
> Allin:
> I didn't mean to imply that the problem of skilled labour is
> the _only_ field of application of the use-value/
> exchange-value distinction, merely that the skilled-labour
> question was the referent of the "this" in the quotation
> from Steve. But anyway, on to Steve's response to the
> substance of my remarks...
> Allin:
> > > What happens in the course of training/education is either
> > > (a) that the trainee is made more productive at the task
> > > he/she was working at before (in which case the increased
> > > productivity can be quantified in physical terms), or (b)
> > > the trainee is equipped for some more demanding sort of work
> > > that he/she was not capable of before, in which case we
> > > cannot make any before vs. after comparison of physical
> > > productivity. I think the latter is more typical, but in an
> > > attempt to simplify a complex topic, let's consider the
> > > former case.
> Steve:
> > No argument here. Normally training creates an ability that did not
> > exist in the recipient before. But yes, we can proceed with the simpler,
> > if less typical case.
> Allin:
> Fine; back to my original posting...
> > > Without any training, one can do two-finger typing. With
> > > practice, one can get quite accurate at this, but never very
> > > fast. Suppose that as a result of taking a course in
> > > touch-typing, a typist is able to double his/her speed, at a
> > > given degree of accuracy.
> Steve:
> > This is where I think your rendition of the problem a la Sweezy/Meek may
> > go wrong, because you have assumed what must be explained by this
> > analysis.
> Allin:
> I'm assuming nothing that must be explained. I'm merely
> assuming, for the sake of argument, that a particular sort
> of training has the effect of doubling a typist's speed
> (words typed per hour of direct labour input). Surely I'm
> entitled to this hypothesis.
> > > Let me define the "Sweezy coefficient" for a given sort of
> > > skilled labour as the ratio of (a) the sum of direct labour
> > > time plus indirect labour time "transmitted" via the
> > > training process to (b) the direct labour time alone. What
> > > might be an upper bound to the Sweezy coefficient for a
> > > proficient touch-typist? Well, let's suppose it takes two
> > > months of one-on-one tutorials to learn (clearly
> > > over-generous). And suppose that the trainer has a Sweezy
> > > coefficient of 2.0. Then that's six months of simple labour
> > > equivalent. Now suppose the typist's skill is depreciated
> > > over a period of 5 years. Then the typist's Sweezy
> > > coefficient is 1.10 (six months being 10 percent of 5
> > > years).
> Steve:
> > I think we may have to start this discussion over again once you've
> > considered my point above, because I'm having trouble with the
> > "supposes". The supposition that it takes two months of training is
> > fine. What I'm not willing to proceed with are the suppositions of a
> > coefficient for the trainer of 2.0 (I'd want an iterative proof that
> > this is possible, since the trainer became a trainer by being trained by
> > a trainer ... and at the end of the induction there is unskilled labor),
> > or the supposition of depreciation. Depreciation of fixed capital
> > presumes that at the same time as the machine's exchange-value is
> > depreciated, it is losing use-value: certainly, by the time it is
> > scrapped, both its exchange-value has been depreciated to zero, and its
> > use-value has been extinguished. This doesn't apply to a skilled worker:
> > a speed typist will maintain at least that level of skill for, say, 45
> > years, after being trained at 19. So the use-value of this skilled labor
> > is its ability to produce value at a rate twice that of a one-finger
> > typist for 45 years--if indeed training does double typing speed (which
> > is a reasonable observation).
> Allin:
> There's no problem of principle here, just a quibble over
> the numbers chosen. I assumed a Sweezy coefficient of 2.0
> for the trainer. That may be unreasonably high, so let me
> use 1.2 instead. Notice this does _not_ have to equal the
> Sweezy coefficient that is then derived for a skilled
> typist: being able to type fast, and being able to _teach_
> how to type fast are two different skills, with the latter
> presumably being a "higher order" skill. If the ability to
> teach touch-typing is acquired via labour, we know that such
> teaching must have a Sweezy coefficient greater than 1.0 --
> since my numbers are merely illustrative, I don't lose
> generality by positing some specific value such as 1.2 (or
> even 2.0).
> As for depreciation, Steve is in effect suggesting a 45-year
> depreciation horizon for the typing skill rather than my
> illustrative value of 5 years. That's fair enough if "once
> a typist, always a typist". Behind my 5-year figure was the
> thought that a typist may well move on to other work over
> the course of his/her career, so the skill would actually be
> deployed for a much shorter period. Be that as it may, this
> is also just a matter of the numbers, not the principle I
> was trying to get at.
> The combined result of choosing a value of 1.2 for the
> trainer's Sweezy coefficient and something on the order of
> 30 or 40 years for the trainee's subsequent time working as
> a typist would be to reduce the trainee's Sweezy coefficient
> to something like 1.01 (maintaining my other assumptions
> intact). OK, let's go with that. Then the rest of my
> argument can be restated...
> Thus, as a result of the training, a doubling of the
> worker's productivity is associated with a mere 1 percent
> (or so) increase in the worker's Sweezy coefficient. Does
> that indicate a problem with the Sweezy-type analysis?
> First consider the situation from the point of view of the
> rational allocation of labour. Given these numbers it
> clearly makes no sense to leave any typists untrained. The
> process of producing typescript via untrained labour is
> dominated; it does not correspond to socially necessary
> labour-time.
> [I might add here: What we have is a 'corner solution', so
> we don't expect to see any sort of marginal condition met --
> e.g. Sweezy coefficient = ratio of physical productivity of
> trained worker to that of untrained worker.]
> What would happen under capitalism? Suppose typists are
> paid via piece rates. In that case the trained typist will
> make double the hourly wage of the untrained (supposing that
> both sorts of labour-power are in the market). Suppose
> further that the workers pay for their own training, which
> is itself supplied by a capitalist enterprise. If the
> training costs about 4 months' wages it will pay for itself
> in 4 months (via the doubling of the wage), after which
> point the higher wage is a clear gain for the worker. If
> workers could borrow to pay for the training it would
> clearly be in their interest to do so. But then we would
> not expect to find any untrained typists in the labour
> market.
> Now notice something else here -- if we follow Marx in
> assuming that there are forces driving wages to a
> "reproduction cost" minimum (allowing for some
> historical-cultural element), the final result will not in
> fact be a doubling of the typist's wage. Rather, the wage
> will _end up_ increasing only by the amount required to
> finance the training -- and the hourly wage of any remaining
> untrained typists will fall below subsistence. In this way,
> the untrained labour is driven out.
> Under what circumstances would it _remain_ the case that
> trained and untrained typing labour co-exist, with the
> "discrepancy" noted above, namely that typists with a Sweezy
> coefficient of 1.01 have double the productivity of those
> with a coefficient of 1.0? Presumably, it would take the
> imposition of restrictions on training. The logic would be
> like this:
> 1. The skilled typists somehow restrict access to training.
> 2. In order to satisfy their requirements for the production
> of typescript, capitalists have to hire both trained and
> untrained labour-power.
> 3. The untrained workers have to be able to survive; their
> wage must be at least at subsistence.
> 4. So the trained workers make double subsistence,
> comprising a substantial gain even after the cost of
> training is met.
> Ricardo (as everyone knows) pointed out that the existence
> of rent is due not to Nature's bountifulness, but rather to
> Her niggardliness: it is because the best land is
> restricted in supply that its owner receives differential
> rent. The same sort of thing can be seen here. In the
> hypothetical case above, the handsome wage for the skilled
> typists is not due to their productivity as such, but due to
> the fact that they have managed artificially to preserve a
> pool of untrained typists.
> Allin Cottrell.