game is identical in equilibrium:
Ui(e↑(AU), e-i(AU)) = uj(e*(AU),e-j(AU)) for all i = j. (19)
The equilibrium utility for the two-player contest is derived in Eq. (7). For the
n -player contest with linear CSF it is
(∖ 2
i—α P v(m~ι)βi i
ɑi Xj ∈m αP I
i
• for all non-active contestants i / M: u*(e*(P),e-i(P)) = 0.
Condition (19) immediately implies that the set of non-active contestants must be
empty, i.e. all contestants will be active under AU. Closer inspection of the expression
for equilibrium utility for the two- and n-player contest also reveals that condition
(19) is satisfied if αiAU = βir for all i ∈ N which coincides directly with policy AA.
Hence, the different interpretations of the normative objective of affirmative action do
not yield different policies for the class of contest games considered here.
7 Concluding Remarks
The implementation of affirmative action policies is a highly controversial topic in
public policy discussion. One of the frequent critical remarks is focused on the poten-
tial disincentives for effort contribution that could be generated by affirmative action
policies. It is argued that there might exist discouraging effects on targeted and non-
targeted groups that could finally imply a reduction of effort levels.
This claim is analyzed for a stylized contest game where contestants could be hetero-
geneous because of past discrimination. If this is the case then, from a normative
perspective, the contest rule should be biased in favor of discriminated contestants to
induce a level playing field. This affirmative action bias is implemented through the
specification of different individual effort weights that are tailored to the individual
discrimination parameter of each contestant in such a way that the normative objective
is satisfied. At the same time the biased contest rule leads to a change in the incentive
structure of the game that affects the optimal effort choice by each contestant. Hence,
the consequences of the implementation of affirmative action can be analyzed with
respect to the equilibrium effort that this policy induces. Using this approach it can
be shown that for the two-player contest game a trade-off between affirmative action
22