ASi" kUCTURE FlOR DEVELOPMENT

- :. : : . - I-
~~*r~
4.~ accounts and ach-ieve a better overall quality of ser-
.>gre-A1_;The4iagoftio, ofcowm~iM
vice. An additional gain fr-om these contracts is that
gp 4 r. & go s governument force account work groups have ber*wterecoveragn
come more en compelled to compete
: ~: with private contractors. A survey of contractors
sugest, lowve, tatmaintenance contracts o
- ... longer duration and wider scope are needed to jus-
F-di;ii9i>'e.>a-r.&-',;4.rr ti- capitalization -9=fy and acquisitiono of specialized
* FiguiE l (1984.>
. of a public entity depends on the actions of governr-
~ g - >- w ; ment, managers, and workers, the best contract is
.;-,. 7 Uu8~ tt.................... the one that most effectively alters incentives to
but a:perf.mancagreement in989 was-. whidnever of these three performs least well
-=needed to imp . eBnanczl outcomes If the problem is with the government, the perr>
ee - *--- L::->----: formance agreement may be the preferred instru-
190
* :
-; ! ment -; -because 0 | performance agreements are reapro-
. . For -: example, ..... = a 1989 :-;.caL performance agreement by
Togo's water utility illustrates how managers can
* . use m such ; - .: a -. ent toget the goverment to endorse
needed tariff inaeases. The performance
agreement was a complement to conmmercialization
Average @ ,. : X . in 1984. The utility's managers wanted an explicit
.perfonance agreement to commit the govenment
to tariff increases. Although commercialization unproved
performance with respect to noncommercial
*St< ; > s- I- ; 5 , goals-a 73.5 percent increase in the number of con-
* ,iSg~? B8687' 88< 89 o "-91 nections in just five years-it did not help financial
-Sr, ., ~performance because the govenmnent did not aurfl[
fjjjnjal c~ifl~fl.ntflbofl (198449)thorize - .
needed tariff increases. By 1989 the cost re-
: . - -covery ratio was 7 percent lower. than in 1984 (Fig-
* { lPerio& wth6e agreement (1989u91)
Lttrt '"
re 2Z1). The performance agreement was needed so
~~~~~~~~~that the government and the utility could agree on
~ AiilA~$J 4 k&Jt% > the steps to achieve finncial autonomy. Within a
year, the cost recovery ratio was 16 percent higher
than its 1984 value. However, if the problem is one
of weak commitment by the government, no reme-
How EFFECnvE s coNTRAcnmc our? Contracting dial instrument short of privatization is likely to be
out tends to be more cost-effective than using public very effective.
employees to handle maintenance (known as force If the problem is with management, the choice of
account). Brazil's switch to road maintenance by contract depends on whether abilities or incentives
contract reduced costs by some 25 percent for equiv- are in question. Performance agreements with in-
* alent service quality In Colombia rural microenter- cumbent public managers assume that their capabilprises
charge about half the rates implicit in force ities are adequate. Thus, in the case of an organiza-
46