..
Universität Bielefeld/I MW
Working Papers
Institute of Mathematical Economics Arbeiten aus dem
Institut für Mathematische Wirtschaftsforschung
No. 22
A Further Note On Rawls's Theory
By John C. P~rsanyiUniversity of Califürnia, BC1keley, and
University of Bielefeld July 1974
=--- --.---.-..-".--....-----..----..'-- -----
-=-==- ---'--'. ._--..,-
~~:~%:~;~j
--
H. G. Bergenthai
Institut für Mathematische Wirtschaftsforschung an der
Universität Bielefeld Adresse/ Address:
Universitätsstraße
4800 Bielefeld 1
Bundesrepublik Deutschland
Federal Republic of Germany
- ---
By John C. Harsanyi
This is a postscript to my review article "Can the Maximin Principle Serve as a Basis for Yorality? -
ACritique
of John Rawls's Theory" (Working Paper No. CP-351, Center for Research in Management Science, University of California,
~8rkeley, California 94720. Copies are available on request.)
~ince this paper was written in May, 1973, John R2wls has
tried to c'1S\'lersome of my cri ticicisms in a p~perentitled "Svne Reasons for the !taximinCriteric
1" (J..mericanEconomi~~-!.ew, Pc,pers & Proc., ~1ay 1974, pp. 1~1-146).
His de-
fense to the counterexamplesJ have put forward against using the
maxi""d n .[I i 'ci pI e ,"'.sa moral In incirle (in Section 3 of the p e-
eeding paper) is that "the maximin criterion is not meant to apply to s~allsc~le situations, say, to how a doetor should
treat hi spat ic.'; s or a t::niVt rsi ty i ts students.
...
'1:axi minis a maero not a miero principle"
(p. 142).Regretful1y,
I ri\)st
<ay thnt this is a singularly inept defense.
Fjrst of all, though my counterexamples do refer to
~..~ll-', ~te sil~ati0ns, it is v ry c~sy to 2~2pt them to large- sc'ale ~i t c tiens since they hc.~:eint r i Tl' {cally l10thing to do wi th ~Ce le ,'hether s.'all or lc..'_'ge.For I.Vi'!"ple,in stead of ask-
ing
v~het"'it r a~' (~Ol' ';LUJld 8 a life" 2\riilg c"rugin short
supply for t}. ,t L:':].L:ti. it\
or r,t ilt B, "e can o.skwhether, in allocating
SCc lt.:e .~acc 1 ,"I'.'lu'.,;erend other resources,
should society give priority to those patients who could best
\-",,"'cit fFit Jcdical treatment, or should rather jL'e
priority
",) 1',: t ~(~,,] 'sly ~ick pe,ticnts
- -
a policy ploblfffi sl1,Llly~j:tctig (
Cl Lntry
t
cl l'::;l.>,dt's"',:d indivi\.~pals in ;qy n:ajor tine. (Ir, i'!gain, i ~te d of ",)ring v,'hether c r( e
clY g iv (~n
l ~ 1
of ,'"
catiu.J I;
in c p r t < 1:1 , ~ f S t 0 :..;. " l .r? 1
" i t~ L1l ~or 1 ty
~.i..d t'''ld ~~(J','j gifted stu-
i i t i-, ,;1: tO;=j 1 h ~ d
d E"nt s .. : 'a ( ( i.'1 d
.
' 1YL~rJ":"'1d ,,"L::{l:,lyiLt - ~ed
dnor lC-lefits ftom aJditionc31
.;: , C'"
,
, " ~ .' , ,-
..I' -1 ,~('~l~e (.
Y
I
?M rLally C~LO-~ (, ( t i 1, t; t c .
.
, 1 < \; 1I r ,
.11 ( .', (i s, ..J(,1 t.y
c' l '
-
2
-over1ooked the simple fact that the counterexamples I have a~dl1ccd (and the many more counterexarnples one could easily adduce) have nothing whatever to do with scale at all.
In fact, it would ~e a priori r~ther surprising if, at thc most fundamental level, the basic plinciples of morality should tare different :orms for 1arsc-scale and for small-scale situations.
Does Rawls serious1y think t}~t there 1s a certain number x, such that a situation involving ~ore than x people will come under mo- ral principles basically different from a situation involving lC3s than x people ? In any ("a.se,what moral considerations will de- termine this curious r)u;;dary number x itself ? T>1orefundamentally, what are the basic logical reasons that should make large-scale and s~nall-scale situations essentially different from a moral
point of view
? I C2n,:ot :"'-eehow arlyl.ody C2'1 plopose th~ st '1ge ('!OCtline tha t scale is a f'":1da:entalra~ ic ble in ),,(al phi ]:f( lhy, wiUout giving credible ans..:crs to these Cjuestions at the Sdi1etime.
I heve ~_(;j"f>d th.it in !:p'ost fit.uc~tjonsR2Wls's theory
will have ruch the s.fle policy i!lplic-atio.s as ut i1 itarian theory does, but tr t there d:e s_~~ L1pOl.tC'1t c,itpatio!s ~I're this is not the case. ~()rt_(;.~~r,I 1ve tli.d to cl .Y.'lJ'hat, in those situ- at ions vihere two theorif's (~o ~. ve (...uit.e cli ssiIllilar policy impli- cations, Rawls's theory consist.ently yields morally highly unac-_
( "t ~ ( 1 1
~
1'01 icy conclusions whereas utili tarian theory ("onsi st E:üt-'y..l~ 1('s J ~L ]ly +ul1y q.cGe.ntc\b!§.ones (~E..:ctj(J'1S 3 cd 4 of t'1e
prN ~ Hng i p.'r).
11 LO..1 '- ~ s " ~ II da; \ ; ] (' r v i ( 'fl
r
:.,.( 11r ,'1 0 f ~h i 1 0 ~ 0 J ~'.V'-
JV I .:I - 1 'J ' t i t }-<.: i 1
world the Ai '.: 1 1:-'.1.; (i.l. 1 e d ~.~ e t j 1 i I' n t r. j i1 C i 1..1~ ~ '" 11 d
have very similar pY~clira1 (c :.qU(J ces, he adds: "... t.he rnaxirin
!linciple T.'ould lE;.ad
to 1.
(«_. ' le t;,nselupnces
if t~e T c>r'd.i.e ~. eh that t:-tY[thC' je c-on .~..ue1c('s] n..ally CliffE' d". '1 '1
~i (y~c~ (nt ~ith llrow is that I think the world is in f~(t 0
CC'1 c:-t i t..:..t- ( d that t:-tcse two principles do have very difft..;u ,1t c ( l) .-
..
-
1 (
...l.,FA 4 '_
s
i '1". e ; :1" "1t s. (IT}..
j- i 1 3
( « I
J
f
' .;s r, <....,., , .! , 1f J.) t ,. t-, .J .. . 1 j ,,
- 3 -
pp. 751-252.) But we do agree on the main point, viz. on the conditional statement that, if such differences.exist, then they all speak very strongly against the maximLn principle.
In my opinion, if this criticism is valid, then it completely disqualifies Rawls's theory as a serjous competitPr to utilit2rian theory. (Why should anybody choose a theory that often does much würse, and never does any better, than utilita- rian theory does ?) For this reason, I find it rather unfortunate that Rawls's paper does not even try to answer this criticism at all.
To be sure, the maximin principle does have its valuable uses, and we ~ust be grateful to Rawls for calling our attention to it. E\en if it eannot ~rrve as a b?sic-- - principle of j~ral _.-the01Y, it ean be used as a pIineiple of ~~proxiJnate vdlidity in practie~l
~plieationsL ~,ueh as the theory of optimal inevme distribution er of optimal tixation. In ~ueh arplications, its rplative indepen- dence of dctailed interpPJ"sonal utility eOlt~?arions, and of the aetual mathematical furm of ~eople's von Neumann-Morgenstern uti- lity functions for money, is an important advantage, and can be fr'uitfully exploited in economic studies [cf. Arrow, op. cit., p. 259].
Of course, from the point of view of a utilitarian obser- ver, the l~ ~ults of a study of, e~g~, optimal ineome tax rates,
}
- ~('(on tre mc...ri'dnl'H'ineiple,,,;i1lhave only approxirtate valil'iLy.For example, if the study finds that, owing to the dis{ ;entive effeet of very high ~arginal tax rates, the ~1Iginal ineome tax for the highest inc('"e gIOU) ~:.lld }-,e (~ay) 50 I r ee'1t,
then a utili-
t?rianobf.Lc}:td1{ flL1-" ti'1.sLx t; t "d(lrtai'1lyL~
.!iQ.2!lore th.:'l'1:0 LA~r <,t.;1r. """'('(1, re ( n iT,fer t'.;;t, if the!:t rj
had been La ~(,d on the ä\lt"I Je ut i 1 i ty pI L eir1 e instead of the mayi;'1in I i ({ i le, t~Ln t"e T':, (J{ 21 tr:..xlate at the top \'0,1d
pC.lr;:ps 0t Vt l.y r, ueh lower. (Sensitivity analysis may ,-"en e, '",,1e us t0 c!"1..':',ate the aetual p -ereentage poi nts by "hich stl'dÜ s
-
4
-based on the maximin principle are likely to overestimate the optimal tax rates for various incrnne groups.)
It is regrettable that Rawls has ever made the untenable claim that he is proposing a moral theory superior to utilitcric:}
theory. This claim can only obscure the practical merits of ~he maximin principle as an e?sily appJicable postulate of appro,i- mate validity. These practical merits of course do not in any way
provide a reason for abandoning utilitarian moral philosophy.
(Basic philosophical principles must be exactly right, and not merely approximately right.) But they do provide a reason, even for a utilitarian moral philosopher, to use the n~ximin princ~ple as an admissible approximation in many cases. Had Rawls only made this more modest, but much more realistic, claim for the maximin
principle, few ppople
'. 'ould have contradicted hirn.One thing that all of us must hEve Iparned in the last fifty years is the fact that we must never cor~it ourselves
seliously to I.,(}~lprinciplcs or political ido1<'gies that are
bound to lead to morally utterly ~'~J~ policies
fn)!1\time to
time - -however great the advcntages of these principles or
ideologies may be in terms of administrative convenience, ease
of application, and readier understandability.
Nr.
"'ir.
Nr.
Nr.
~~r.
71
Carl Christian von Weizsäcker:The Political
Economy of Stability in Western Countries,
-
The WicksellLectures -
Stockholm, 2nd and 4th May,1972, published
at Almquist and Wiksell, Stockholm 1972.
2 Carl Christian von Weizsäcker:
"(j lf~rn
Capital Theory and the CCilcept of Fxploitation ,
July 1972, l'ublished
in
:~yklos, j'ay 1973.3 Carl rhri tian von PP:i2: 'kl.er:
A ~~W T(chn'cal Progrt.ss
Function (1962), August 1972.
4
Carl Christi an von \'ieizs.'icker:Long Term Global Optimization in Educational Planning.
A Sir1ple F>:0i'1I'.1e,
Septen'ber
1972.5
CarlChristian von Weizsäcker:
'"nt f>r e'1erationell e Einkomrnensvel. teil ung: Finfache
)eispiele 7ür \'~irkui1ljE'n steuerlic,er '1C!ßnah ,..:n und
{"ijr (He cl-"'" -'le St('h-r~trl1k.tL1r, C('io'er 1S72, I.Jb-
li..1-I'd in ,'(,3 l'q)ti.te(~t>r "IJtE.il .j~tl, J'e, G.poTT1baC'h,
9. Fr.ey, B. ('.,;h"n (cds.), 'nibil~J::..n 1974.
6
Carl Christian von Weizsäcker:Yenneth Arrow's ContriLution to Ecenomic ScienC'E's, r"OVl. bt r 1972, lubli~ lJ.:d
in S"erli~h
JOllu''''l of i-'l (,,0-rniGs,
Velo 74, 1972.
r~rl ChI' ,,,t i <.1 ."1'1 <7'-'1
l.y 1}71,
T("lIlal, T"
'i d' ; 1
'tr 1./1.
.
71,
8
Reinhard Selten:A Si p'ple Horiel of Ir1perfect Competi tion, \Alhere 4 are
1', <'1 'Ld 6 bre "1eny, FC'bn'ery 1973, published in Inter-
',1 Je:, 'I cf (' e 'l"':~',l'Y, D" ' '.:>r 1913.
"j. 10
r.
11
r.
~,
vf
Y 1~11.
, 1
t' 1 J 'n
t ,d ,.1,',1 «'"'!h J.g:
;"y i",trr-t'd ,~fjl1... l' 'r HCt:S, July 1973.
,.-, i '1 ; 1 '1
..
,"', .
,
11.')li d i ,1 \.. . .; c
Nr. 1 2
Nr. 13
Nr. 14
Nr. 15
Nr. 1 6
Nr. 17
l~r . 19
~'1' . 21
"r. '2
-
- 2 -
Carl
Christina von Weizsäcker:Grenzen der traditionellen Globalsteuerung, October 1973,
published in Jahrbücher für Nationalökonomie und Statistik.
1974.
Carl
Christian von Weizsäcker:
Reply to Mrs. Robinson, Morishima and Wolfstetter, November 1973.
Carl Christian von Keizsäcker:
Substituion Along the Time Axis (preliminary version)
, February 1973.John
C.
Harsanyi:The Tr~'cing procedure:
A Bayesian Approach to Defining
a Solution for n-Person Noncoo~erative Games, Part I, May 1974.
Jnhn C. I'cl 5 "1yi :
'.1;'e Tl 8t':i.T9 rll(U~,'.r(':A r:>'l '
a Sol t; on [,)r 11 ~,;r'>un :'c ( C~l
'. a y 1 9 7 4 .
n
1.1 rO~Lhto r~li~e r s,
"'tfi ing
p, 1 t 1 T ,
C"oLn ),n ('~, Tl:
Pl:O(~t'1 Li, 1,' <pm, ,d :!c',,:t ~(s ;'i,'.,JlJU 'I:;; Ü-Ltels n('is
in Ta' an: fko~c~ctlische Unter~uchung und Proynose,
~1ay 1974.
Reinhard Selten:
The Crain Store Pra1ox, July 1974.
Cerl C'listicn von T';ei2.~;-:(;k('r:
Politi~~l Li~its of Traditional stabilization Policy, J~re 1974.
'"',(inha~j Selten:
Bargai'Ü'1g tL;:3.er IncOJlplete Infor;atiC'n - :\ ;'\. Ti ca 1 Fx IpIe, July 1974.
r (T"yi:
, 1 1 )< c
~ U 1
o a '~ (~...l
';,']y 1':;/4.
lii..y?
1..i 1 :
,~ r',,' L L ~G tee 1 ,'"' .;1 s 's l' r ln y, lT\...1Y 1 97 4 .
+ . r '....,...)
( . >
! '1
J.