928 Monetary Policy, Beliefs, Unemployment and Inflation
the labor market, in contrast with more statistical (or time series) models of unem-
ployment, such as Ball and Mankiw (2002) for example. Using such behavioral
models, when estimated as reduced-form unemployment equations, many labor
economists have concluded that labor supply factors play a crucial role in deter-
mining long-term unemployment. For a recent affirmation, see Layard, Nickell and
Jackman (2005) and the review of this book by Blanchard (2007). But the empirical
claims that effects of labor supply-side or wage-pressure variables, such as unioniza-
tion, income out of work and the rigor with which rules on this are applied, have
been so important is challenged by other economists.^17 Among others, these critics
include Madsen (1998), Oswald (1997), Henry and Nixon (2000) and Blanchflower
(2007). Empirical reasons for not accepting the claims of the supply-side propo-
nents are discussed in section 18.4.2. The alternative model suggested there is
that, based on econometric evidence, the labor supply variables that have charac-
terized much of UK research do not satisfactorily account for the large changes in
unemployment since the early 1970s, so the alternative emphasizes other possible
determinants of unemployment, based on effects from the external economy.
To make the present study comparable with existing research, the model of
wage and price determination used here starts from the model described in Layard,
Nickell and Jackman (2005) and Nickell (1998), the basic equations of which are a
price equation based on markup pricing and a wage equation derived from union–
firm bargaining, which are solved to give an unemployment equation.^18 Thus the
price and wage equations (ignoring time subscripts) are:
p−w=zp−β 2 (p−pe) (18.15)w=γ 2 pe+( 1 −γ 2 )p−γ 1 u−γ 11 u+zw, (18.16)wherepeis expected prices. Solving (18.15) and (18.16) foru, ignoring the price
surprise terms, gives a lagged equation in unemployment and the exogenous wage
and price variables(zp,zw), that is,
u=
γ 11
γ 1
u− 1 +zp+zw
γ 1. (18.17)
This shows that in the long run (where unemployment is not changing, sou=
0), unemployment depends onzwandzp, the “push” or driving variables in the
underlying wage and price equations respectively. Details on the components of
each of these “push” variables are discussed next.
From this point on, the approach is to review the empirical role of each of a
large set of wage and price “push” variables, taking those used in recent research
on both the wage and the price equation as the starting point. Recent examples of
wage “push” variables (where each is an element in thezwterm in (18.17) above)
have included the terms of trade (TT), a measure of skill shortage (Skill), the tax
and price wedge (T), the replacement ratio (RR), a measure of union power (UP), an
index of industrial turbulence (IT), and the real interest rate (r) (see, for example,
Nickell, 1998, for models of the UK). From the pricing side, recent open-economy