( 1996 ). In thisWeld study of home telecommunications sales representatives, at the
two-month mark on the job, new employees transitioned from a relativelyWxed
pay scheme to a system where 100 percent of pay depended upon sales perform-
ance. During theWrst two months of employment the correlation between sales
performance and turnover was very modest, but under the commission-based
system, this correlation increased dramatically.
Another supportive example is Trevor et al.’s ( 1997 ) study of managerial and
professional employees of a petroleum company. Of interest here, turnover rates
among top performers reduced substantially if they experienced signiWcant salary
growth. According to Trevor et al.:
The greatest diVerences in retention under conditions of low and high salary growth were
by far in the top two performance categories. We contend that the fate of these few
employees is disproportionately important to the organization.... Tomorrow’s stars and
perhaps even franchise players may be among today’s few top performers; their retention, at
least in part, appears to depend on paying them according to their performance. ( 1997 : 57 )
Sturman et al. ( 2003 ) analyzed the economic beneWt of ‘winning the talent war’—
retaining higher performers through the use of individual pay-for-performance
plans. Even under the most conservative assumptions regarding the value associ-
ated with employee performance variability, Sturman et al. demonstrate that
performance contingent pay yields substantialWnancial beneWts over the non-
contingent pay strategy.
Group–based pay plans have interesting implications for voluntary turnover.
While some evidence suggests reductions in voluntary turnover (e.g. Azfar and
Danninger 2001 ; Wilson and Peel 1991 ), group-based pay plans are also imbued
with the ‘free rider’ problem, where incentive eVects are diluted by peers’ attempts
to free-ride oVothers’ contribution. The free-rider problem should be particularly
disturbing to higher-performing workers (Hansen 1997 ; Weiss 1987 ), leading to
greater dissatisfaction and turnover among this group (see Park et al. 1994 ).
Further, if group incentive plans are particularly dissatisfying for employees who
perceive themselves as ‘high performers,’ then this problem may be magniWed by
the ‘Lake Wobegone eVect:’ the statistical anomaly that almost all employees
perceive their performance to be above average (Meyer 1975 ). Moreover, increases
in group size may exacerbate this problem. Described as the 1 /nproblem (where
n¼group size), direct returns to individuals in group settings are diluted by a
factor of 1 /n(Hansen 1997 ; Kruse 1993 ; Weitzman and Kruse 1990 ). Guthrie ( 2000 )
reported results consistent with this argument, showing that organizational turn-
over rates under group incentive plans increased as a function of group size.
Pay form/payment system eVects: summary. A review of the literature suggests
strongly that pay-for-performance plans, both at the individual and group levels,
can have both statistically and practically signiWcant eVects. Research also suggests
that eVects of pay-for-performance are due to both incentive and sorting eVects.
Research examining pay-for-performance issues needs to especially be aware of
350 james p. guthrie