BANC A NAZIONALE DEL LAVORO
QUARTERLY REVIEW
EDITORIAL BOARD
Francesco Bignardi
Guido Carli
Cesare Cosciani
Antigono Donati
Lucio Izzo
Alberto Mucci
Nerio Nesi, Chairman
Francesco Parrillo
Romano Prodi
Pasquale Saraceno
Luigi Ceriani, Editor
A. Roncaglia, Asst. Ed.
VIA V. VENETO, 119
I - 00187 ROMA
December 13, 1984
Dear Josef,
I found your paper on Sraffa very interesting and very provoca
tive - as usual. I feel largely, but not completely, in agreement with
you. Namely, I strongly agree with the idea that Sraffa's analysis must
be considered a "reference system", while his analytical tools - R, the
basic-non basic dichotomy - can be of use in "applied" analysis. On the
other hand, as you can imagine from my paper on Sraffa in the BNL Quarterly,
I feel at unease with some aspects of your presentation of the aims and
scope of Sraffa's analysis (e.g. the first 6 and last 3 lines of the paper),
which in any case do not seem in a necessary connection with the rest
of the paper.
I have some other comments on a few points of detail: 1) Page
1, first few lines: For Sraffa (I960, Preface) the standard commodity
is only "a particular point", not the main aim of his research. Thus the
duality between "standard proportions" (vector d^) and prices (in Section
2) is a derived aspect of Sraffa's analysis; 2) Page 2, last few lines:
it would be more precise to write "This applies only to commodities which
directly or indirectly enter as input into the production of all other
commodities... Other goods are called non-basics". (As you note later,
some non-basic might be used as input in its own production or in the
production of other non-basics); Page 3, line 7, more precise: "...produc
ing one unit of a particular commodity"; 4) Page 3: System (3) is dual
to system (2). In (1), s^ f s^ ;.while in (3) = R. ; 5) Page 6, last
7 lines. I do not understand thh underlying reasoning. At first sight,
(8) = (9) = 1 implies overdetermination of the system. On lines 1-2 of
page 7 : any unit of measure is, in itself, independent of distribution;
but the relative price of any other commodity in terms of any given unit
of measure will change, in general, when distribution changes. If we choose
the standard commodity as unit of measure, the standard net product will
not change in value, when distribution changes, because it is simply
a given quantity of the standard commodity; but UeriwQ T ackward cases the
actual net product will; 6) Page 7, last line: one might add: "(unless
a change in their price, due to changes in their conditions of production,
affects the wage rate)"; 7) Page 9, line 6: it is unclear why matrix A
should become decomposable; 8) Page 9, line 9 from the bottom: I would
object to the statement that the system analysed has neither past nor
4
Urn*. 10 turr
v t
gr- y ,7 r Au
i t f, ™ /
At -fi*
0 ft
)" *
fA/ 4