866 Macroeconometric Modeling for Policy
we obtain the solution for the bargained nominal wage:
Wb=Q
( 1 +T 1 )
F(Pq( 1 +T 1 ),Z,,U). (17.19)Letting lower-case letters denote logs of variables, a log-linearization of (17.19)
gives:
wb=mw+qt+(
1 −δ 12)(
p−q)
+δ 13 z−δ 15 u−δ 16 T1. (17.20)
0 ≤δ 12 ≤1, 0<δ 13 ≤1,δ 15 ≥0, 0≤δ 16 ≤1.The elasticity of the wedge variable (p−q)is( 1 −δ 12 )in (17.20). In econometric
models of wage-setting in manufacturing, the hypothesis ofδ 12 =1 is typically not
rejected, meaning that the wedge variable drops out and the bargained nominal
wage is linked one-to-one with the producer priceq(see, e.g., Nymoen and Rød-
seth, 2003). However, at the aggregate level, a positive coefficient of the wedge is
typically reported. This may be due to measurement problems: since gross domestic
product (GDP) is an income variable, the price deflatorqis not a good index of “pro-
ducer prices.” That said, the estimated importance of the wedge may also reflect
that the economy-wide average wage is influenced by the service sector, where
wage claims are linked to cost of living considerations, implying that( 1 −δ 12 )is
different from zero.
Irrespective of the split betweenqandp, productivityzis found to be a main
determinant of the secular growth in wages in bargaining-based systems, so we
expect the elasticityδ 13 to be close to one. The impact of the rate of unemployment
on the bargained wage is given by the elasticity−δ 15 ≤0. Blanchflower and Oswald
(1994) provide evidence for the existence of the empirical law that the value of
−δ 15 is 0.1, which is the slope coefficient of theirwage curve. Other authors instead
emphasize that the slope of the wage-curve is likely to depend on the level of
aggregation and on institutional factors. For example, one influential view holds
that economies with a high level of coordination and centralization are expected to
be characterized by a higher responsiveness to unemployment (a higher−δ 15 ) than
uncoordinated systems that give little incentive to solidarity in wage-bargaining
(Layard, Nickell and Jackman, 2005, Ch. 8). Finally, from the definition of the
wedge, one could setδ 16 =δ 12 , but we keepδ 16 as a separate coefficient to allow
for separate effects of the payroll tax on wages.
Equation (17.20) is a general proposition about the bargaining outcome and its
determinants, and can serve as a starting point for describing wage formation in any
sector or level of aggregation of the economy. In the following we regard equation
(17.20) as a model of the average wage in the total economy and, as explained
above, we therefore expect
(
1 −δ 12)
0, meaning that there is a wedge effect in
the long-run wage equation.
Equation (17.15) already represents a price-setting rule based upon so-called
normal cost pricing. Upon linearization we have:
qf=mq+(w+T 1 −z), (17.21)