> 7 bjbjUU 7|7|l
8BN
n:"\^^^^^^$ bbbb\b\bb\\bv{I
|^\\0\??\b
Comment (on R. C. Feenstra and C. R. Knittel, Re-Assessing the U. S. Quality Adjustment to Computer Prices: The Role of Durability and Changing Software, October 2004)
I have read this paper now three times and I am still impressed by its lucidity and rigour. As I see it, the basic problem attacked in this paper is whether a conventional output price index for a durable good, such as computers, can also be used as (partial) input price index. The answer appears to be no, since it is not so much the computer itself that is desired by a firm, but its bundle of characteristics. On the one hand, this bundle of characteristics determines the computers price at the output side of the market, and can be used to compute a hedonically adjusted output price index. On the other hand, this bundle of characteristics interacts with current and future software in the firms production process. In the firms optimization problem the amounts of characteristics thus serve as decision variables.
This means that from the firms viewpoint the relevant input price is the value of the characteristics bundle of the computer (which is equal to its market price) divided by the quantity, somehow measured, of the services provided by the thing (which quantity is a function of the characteristics and the state of the software). This is the price that enters the firms optimization problem (see expression (1) in the paper). Notice, however, that this price is not exogenous to the firm. It can be conceived as a first stage of the overall optimization problem.
However, according to the authors, it is this price upon which the (partial) input price index should be based. What they basically do in this paper is comparing this price index with a hedonically adjusted output price index. The results are interesting, to say the least.
Along the road of implementation, the authors had of course to make quite a number of assumptions. Rather than discussing them one by one, which would lead to all too obvious questions, I will indicate a number of points for further thought.
First, the authors are a little bit ambiguous about the price concept on which to base the, true, input price index. In their notation, the choice is between an index of the form
EMBED Equation.3 (1)
or
EMBED Equation.3 where EMBED Equation.3 . (2)
Although it appears that (2) was used, the theoretical argument rather led to (1). I guess that to a certain extent the choice depends on whether software is seen as a decision variable or as a conditioning parameter.
Second, this naturally leads to the question whether a symmetric treatment of hardware and software would be possible. In the paper, the state of software is treated as exogenous; it is a given and there is no price involved. The authors are, of course, aware of this limitation and mark it in their last section as an area for further research.
Third, in this paper the BLS-type output price index for computers was compared to a true input price index for the UC Davis. The results are, therefore, firm-specific and maybe idiosyncratic. Speaking as a statistician, I would like to see the sample enlarged.
Bert M. Balk
Statistics Netherlands
and
Erasmus University Rotterdam
18 November 2004
PAGE
PAGE 1
!
4
5
6
7
츩읎쇄|0JmHnHu0J
j0JUjCJEHUaJmH sH jLHE
UVmHsHjCJEHUaJmH sH jHE
UVmHsHjCJEHUaJmH sH jHE
UVmHsHjCJUaJmH sH CJaJmH sH 5CJaJmH sH !
./!"
A
B
xy&`#$
. A!"#$%Dd\
c$A??#"`25@aZMڂ;Dq`! @aZMڂ;
@xڥKcAgf_bK8"\!
A,ET(;J9? X{bmiF|?^7&ٙa|A@>u,BL,8?&kyWwX,)k0`U|E{$1XjSdw9V*vl>xV/Ԛ>kިi`,DsD'N3uk]g4:˔XʧD/K/w2bԼx#-y-?ɣi83<8uok{Qw9iv7yqVp.˭(Xd~;'{)6IA)GFHAh[õ9&u
!su^j*TIJjjPNF<+J=A&RriY+VPXF=VML_1[pÄ71b"EѼOh}5{xlo'}PuLCş}_K[OY7.|9<^짢?@ABCDEFGHIK
FMicrosoft Vergelijking 3.0DS EquationEquation.39q=
ht
(xt
)/f(xt
,st
)ht"1
(xt"1
)/f(xt"1
,sEquation Native _1162380968FppIppIOle
CompObj
jt"1
)
FMicrosoft Vergelijking 3.0DS EquationEquation.39q$L
ht
(xt
)/f(xt
,s)ht"1
(xt"1
)/ObjInfo
Equation Native
_1162381132FvIvIOle
f(xt"1
,s)
FMicrosoft Vergelijking 3.0DS EquationEquation.39q@
s=st"1
,stOh+'0CompObjjObjInfoEquation Native \1Table)?SummaryInformation(DocumentSummaryInformation8\CompObjJj (4
P\h
t Comment on Feenstra and Knittelommbblknt blkblkNormal.dotFAlice Nakamuras2icMicrosoft Word 9.0 @F#@o@L/@L/
՜.+,D՜.+,Lhp|
CBS
Comment on Feenstra and KnittelTitle8_AdHocReviewCycleID_EmailSubject
_AuthorEmail_AuthorEmailDisplayName$Your discussion of Feenstra KnittelBBLK@CBS.nlBalk, prof dr B.M.F
i0@0Normal_HmHsHtH<A@<Default Paragraph Font, @,Footer
p#&)@&Page Number
4 6 :::!!%, 5 7 "333333. 7 B <bblkAlice NakamuradD:\Alice,Erwin,2004Conferences\Papers\PostedPapers\Balk,Nov.2004,Comment on Feenstra and Knittel.doc@*W@UnknownG: Times New Roman5Symbol3&: Arial"0F4F4
!20
U+3HComment on Feenstra and KnittelbblkAlice Nakamura
FMicrosoft Word Document
MSWordDocWord.Document.89q