Endogenous institutions and endogenous technology

Combining the main mechanisms analyzed in previous sections yields a model where the distribution of human capital, the technologies used by firms and the policy imple­mented by the state are all endogenous - as they are in reality.

This structure makes clear the presence of important multiplier effects: a transitory shock affecting inequality (e.g., more idiosyncratic uncertainty v²) or the political sys­tem (a higher λ) will be amplified through technology decisions, the policy choice, and the intergenerational transmission mechanism, and may thus have considerable long­term consequences.⁴⁴ Most importantly, in accounting for changes in inequality one can no longer treat technological and institutional factors as separate, competing explana­tions: both are jointly determined, and complementary. The model thus shows how, in the words of Freeman (1995), one needs to think of “the Welfare State as a system'”.

To demonstrate these points I shall assume from here on a piecewise-linear techno­logical frontier. Flexibility is free up to c_l, then has a marginal cost of M > 0, up to a maximum level c_h > cl,

Furthermore, since R(τ, σ) is decreasing in τ, we have the proposition.

Proposition 9. More skill-biased technologies appear first in, and less skill biased technologies disappear first from, countries that have less redistributive fiscal, educa­tional or labor market institutions. For any M > 0:

(1) If σ∏ is a steady-state equilibrium technology under a constant redistributive policy τ, this remains true under any less progressive policy τ' < τ.

(2) If σ^ is a steady-state equilibrium technology under a constant redistributive policy τ', this remains true under any more progressive policy τ > τ'.

These results are illustrated in Figures 5 and 6 for two cases where: (i) R( r_, σ∏) < R(τ, σ_b), implying that for each M there is a unique technology compatible in the long-run with each policy τ ∈ {τ, τ};⁴⁶ (ii) R( r_, a_L) < R(τ,oh), implying that for either policy τ ∈ {τ, τ} there is a range of M’s where multiple technologies are sus­tainable. The message is essentially the same in both cases, showing how a world-wide shift in the set of feasible technologies can result in different evolutions of both pro­duction processes and the skill premium across countries. in particular, the model can help explain why skill-biased technical change and reorganization occurred first, and to a greater extent, in the United States compared to Europe - and within Europe, more so in England than on the Continent.⁴⁷

Indeed, consider two countries, C₁ and C₂, that are initially identical in all respects, including both using the technology σc, except that one is in a laissez-faire equilibrium, τ = τ, and the other in a welfare state, τ = τ. Suppose now that the technological frontier gradually flattens (M declines), meaning that flexibility becomes cheaper to achieve. As shown on Figures 5 and 6, the more skill-biased technology σ∏ becomes (all or part of) another feasible equilibrium in C₁ before it does in C₂; similarly, σc first ceases to be viable (by itself or as part of a mixed equilibrium) in the laissez-faire country, while it is still sustainable in the more redistributive one.

Going further, there are in fact reciprocal interactions between the economy’s tech­nology response and policy response to shocks. Proposition 9 and Figures 5 and 6 show through which the wage-compression policies of continental European countries may have caused techno­logical progress there to be less skill-biased than in the United States.

Figure 5. The response of technology and policy to a decline in the cost of flexibility [case (i)]. Under each range of M appears the unique (σ, τ) suchthat (σ,∆ = D(τ; 1 — 1∕σ)) isa stable steady state given τ and M. The subset reached via solid lines corresponds to the stable steady states in (σ, ∆, τ) jointly, when policy is endogenous as well.

Figure 6. The response of technology and policy to a decline in the cost of flexibility [case (ii)]. Under each range of M appear the values of (σ, τ) such that (σ,∆ = D(τ; 1 — 1∕σ)) is a stable steady state given τ and M. The subset reached via solid lines corresponds to the stable steady states in (σ, ∆, τ) jointly, when policy is endogenous as well.

that feasible new technologies are not implemented unless institutions are sufficiently inegalitarian. But, conversely, the occurrence of technical change alters these same in­stitutions, as seen in Proposition 7.

where λ and λ were defined in (22) and (23). These inequalities imply that: (i) under the technology a_L, both social contracts τ and τ are political steady states; (ii) under σ∏, τ is a political steady state, while τ is one if and only if we also have λ < λ (γ∏; B).

When this last inequality holds, the set of stable politico-economico-technological steady states (with endogenous τ, ∆ and σ) is the same as described on Figures 5 and 6. When λ > λ(yh ; B), however, the more redistributive social contract τ is not politi­cally sustainable under the amount of inequality that results, in the long run, from the technology σ∏. Therefore one must remove from the set of steady states on each figure the “branches” corresponding to this outcome; these are indicated by the dashed lines. The remaining solid lines then indicate that only certain politico-technological configu­rations can be observed in the long run: (a) for low values of M, e.g. for M < R(τ,a_L) on the first figure, the only feasible social contract is τ, together with the technology σ∏;

(b) on the secondfigure, for M ∈ (R(τ,σL),R(τ~, o_l)) only the egalitarian social con­tract and the egalitarian technology, or the inegalitarian social contract and inegalitarian technology, are mutually compatible.

5.