# Re: [OPE-L] book review of Kliman's book

From: Dave Zachariah (davez@KTH.SE)
Date: Thu Nov 22 2007 - 13:07:26 EST

```on 2007-11-22 00:49 Philip Dunn wrote:
> Devastating?
>
> p.7
> "So suppose that capital investment was \$100 (including wage payments)
> and the direct labor-time was 10 hours which produced an output that
> sold for \$150, then from this data you can compute that 10 hours of
> labor adds \$50 of “new value” and therefore \$1 must represent 0.2 hours
> of labor."
>
> Oops.
>
> OK, just a slip over including wage payments.
>
> However, the writer gives no indication that he understands the TSSI or
> SSSI or Marx even.
>

For an outsider like myself numerical examples like these do not clarify
much without an explicit formalism. The standard production price vector
is defined by solving

p = (1+r) [  A p + w  ]

What is it precisely that the "TSSI school" suggests? From what I have
understood the TSSI production price vector at time instance t is defined by

p(t) = (1+r)[ A p(t-1) + w ]

If this the case, I think there are some serious problems:

1. What predictions does the theory make regarding real market prices?
2. Is it suggested that p(t-1) are the real market prices from a
previous "period"?
3. If yes, then the predictive power of the theory is weak, because
it takes previous market prices to predict future market prices
4. If no, and thus p(t-1) is the TSSI production price vector from
the previous period, then it appears that p(t) is determined by
the infinite past. If that is taken as historical time, i.e. with
a varying production matrix A, it appears that p(t) is
undetermined. Otherwise, as Ajit Sinha points out, p(t) is merely
step t in an iteration that converges to the standard production
price vector p.
5. If we are truly speaking about historical time, then what is the
sample period of the discrete time model? Is t-1 one year, one
month or one hour ago?

//Dave Z
```

This archive was generated by hypermail 2.1.5 : Fri Nov 30 2007 - 00:00:04 EST