/LMAR=0/XLINE=3/FONT#0=BASL30/FONT#1=BASI30/FONT#2=BASB30/FONT#3=STA200/FONT#4=NGB25
␈↓ ↓H␈↓␈↓βS␈↓∧ Artificial Intelligence Laboratory, STANFORD UNIVERSITY, Stanford, California 94305

␈↓ ↓H␈↓∧Telephone 415 497-4430␈↓ 
≤May 13, 1977 




␈↓ ↓H␈↓Professor Wassily Leontief
␈↓ ↓H␈↓Department of Economics
␈↓ ↓H␈↓New York University
␈↓ ↓H␈↓New York, N.Y.

␈↓ ↓H␈↓Dear Professor Leontief:

␈↓ ↓H␈↓        This␈α↔is␈α⊗a␈α↔question␈α⊗about␈α↔the␈α↔possible␈α⊗applicability␈α↔of␈α⊗input-output␈α↔analysis␈α↔to␈α⊗the
␈↓ ↓H␈↓management of a discrete in∨ationary shock like the sudden rise in oil prices.

␈↓ ↓H␈↓        When␈αthe␈α
price␈αwent␈α
up,␈αevery␈α
individual␈αand␈α
business␈αknew␈α
that␈αit␈α
would␈αbe␈α
a≥ected␈α-␈α
not
␈↓ ↓H␈↓merely␈αdirectly␈α
but␈αindirectly␈α
as␈αwell.␈α
 A␈αbusiness␈α
could␈αcalculate␈α
the␈αdirect␈α
e≥ect␈αof␈α
the␈αincrease␈α
in
␈↓ ↓H␈↓oil␈α∂price␈α∞on␈α∂its␈α∞own␈α∂costs␈α∞in␈α∂so␈α∞far␈α∂as␈α∂it␈α∞used␈α∂oil,␈α∞but␈α∂it␈α∞could␈α∂not␈α∞compute␈α∂the␈α∂indirect␈α∞e≥ects
␈↓ ↓H␈↓through␈α
its␈α∞other␈α
suppliers.␈α∞ Therefore,␈α
its␈α
price␈α∞increases␈α
were␈α∞guesses,␈α
often␈α∞overestimates,␈α
and
␈↓ ↓H␈↓were␈α∃subject␈α∃to␈α∃later␈α∃correction␈α∃-␈α∃usually␈α∃increases.␈α∃ It␈α∃seems␈α∃reasonable␈α∃that␈α∃this␈α∃lack␈α∀of
␈↓ ↓H␈↓information made the total in∨ation greater than would otherwise have been the case.

␈↓ ↓H␈↓        In␈αprinciple,␈αgiven␈αa␈αproduct-by-product␈αinput-output␈αmatrix,␈αthe␈α≡nal␈αe≥ects␈αof␈αa␈αprice␈αrise
␈↓ ↓H␈↓in␈α∞a␈α∞single␈α∞quantity␈α∞could␈α∞be␈α∞computed.␈α∞ The␈α∞idea␈α∞is␈α∞that␈α∞a␈α∞collection␈α∞of␈α∞compatible␈α∞price␈α∞rises
␈↓ ↓H␈↓might␈α∪have␈α∀been␈α∪announced,␈α∀such␈α∪that␈α∀if␈α∪they␈α∀were␈α∪all␈α∀made,␈α∪each␈α∀industry␈α∪and␈α∀class␈α∪of
␈↓ ↓H␈↓individuals␈α∂would␈α∂have␈α∂the␈α∂income␈α∂corresponding␈α⊂to␈α∂that␈α∂of␈α∂a␈α∂number␈α∂T␈α∂of␈α⊂years␈α∂previously,
␈↓ ↓H␈↓according␈αto␈αhow␈αmany␈αyears␈αof␈αeconomic␈αprogress␈αare␈αlost␈αby␈αthe␈αincrease.␈α An␈αallocation␈αof␈αthe
␈↓ ↓H␈↓immediate shock to wealth might be made by delaying some of the increases.

␈↓ ↓H␈↓        My␈α∃question␈α∃is␈α∃the␈α⊗following:␈α∃␈↓↓Is␈α∃the␈α∃input-output␈α⊗data␈α∃base␈α∃too␈α∃aggregated␈α⊗for␈α∃this
␈↓ ↓H␈↓↓application␈↓?␈α∂ Thus␈α∂if␈α∂airplanes␈α∂depending␈α∂heavily␈α∂on␈α∂fuel␈α∂and␈α∂electri≡ed␈α∂railroads␈α∂in␈α∂a␈α∞region
␈↓ ↓H␈↓with␈αwater␈αor␈αnuclear␈αpower␈αwere␈αboth␈αcalled␈αtransportation,␈αthe␈αallocated␈αincrease␈αin␈αfares␈αwould
␈↓ ↓H␈↓be␈αtoo␈αmuch␈α
for␈αthe␈αelectri≡ed␈α
railroad␈αand␈αinsu≠cient␈α
for␈αthe␈αairline.␈α
 Some␈αerrors␈αof␈α
this␈αkind
␈↓ ↓H␈↓are inevitable; the question is whether the data are adequate for a ≡rst approximation.

␈↓ ↓H␈↓        The␈α∂question␈α∞is␈α∂neutral␈α∞as␈α∂to␈α∂whether␈α∞a␈α∂government␈α∞could␈α∂properly␈α∞impose␈α∂such␈α∂a␈α∞price
␈↓ ↓H␈↓increase␈α
schedule␈α
or␈α
whether␈α
it␈α
might␈α
have␈α
merely␈α
been␈α
published␈α
as␈α
a␈α
guideline␈α
to␈α
what␈α
might␈α
be
␈↓ ↓H␈↓expected to occur.


␈↓ ↓H␈↓Sincerely,



␈↓ ↓H␈↓John McCarthy
␈↓ ↓H␈↓Director
␈↓ ↓H␈↓Professor of Computer Science