@article{99a0d0487f93479684f7714f64d67326,
title = "Infrastructure, growth and the two dimensions of industrial policy",
abstract = "When sustained growth depends on establishing an indivisible infrastructure for directly productive activities (DPAs) both discrete and marginal departures from optimal growth can undermine the dynamic efficiency of the market. Producers{\textquoteright} anticipations of paying monopoly fees for infrastructure services dampen their incentive to invest in DPAs and may prevent the economy from reaching a minimal level of activity that would justify investment in a large indivisible infrastructure; and potential investors in infrastructure may be intimidated by the prospect of expropriatory regulation. Thus a credible prior commitment to effective but fair regulation is necessary for achieving optimal growth. But it may not be sufficient. Low-level expectations before the infrastructure is established can be self-fulfilling, indicating a role for coordinative industrial policy. Investment subsidies, even in conjunction with regulation, cannot induce an efficient equilibrium.",
author = "Moshe Justman",
note = "Funding Information: Proof First note that r'(yo, s) is non-decreasing in s. To see this assume that r'(yo, s) > r'(yo, s+ 1). Then r'(yo, s) #:r'(yo, s+ I) implies that min {r'(yo, s), r'(yo, s+ 1)}~s+ 1 and max {r'(yo, s), r'(yo, s+ I)} ~s, and therefore that r'(yo, s) > sand r'(yo, s + 1) < s + 1. Thus households postpone use of the infrastructure until s + 1 when it is available at time s. As this course of action is also feasible when the infrastructure is only available at times s+ 1 we must have, from the uniqueness of the optimal path, y'(Yo, s) = Y(Yo, s+ I) (as the stationarity of access fees implies that households cannot be made better off by postponing structural change.) But then r'(yo, s)= r'(yo, s+ I). Hence r'(yo, s) is non-decreasing in s. Moreover, the monotonicity of the growth path (cf. Lemma 4) implies that YO~YI(YO,s) and that r{"} is nondecreasing in Yo. Hence: r'(yo, s) = r'(JHyo,s),s-l)+1~r'(yo,s-l)+1. II Acknowledgements. Research for this paper was supported by the Jerusalem Institute for Israel Studies, and generous administrative assistance was provided by the Graduate School of Industrial Administration, Carnegie-Mellon University and the Monaster Center for Economic Research, Ben-Gurion University. It is an outgrowth of a long and continuing collaboration with Morris Teubal on technological infrastructure which provided much of the motivation for this paper and some of its conceptual underpinnings. I am indebted to him for many comments, reflections, and specific suggestions. I have also benefited from stimulating discussions with colleagues in the Industrial Development Policy Group at the Jerusalem Institute; from the comments of seminar participants at Carnegie-Mellon University, the University of Pittsburgh, Case Western Reserve University, and Ben-Gurion University; from the helpful suggestions of Ezra Einy and Mark Gradstein; and from the thoughtful and extensive comments of two anonymous referees, and the editor, Mathias Dewatripont.",
year = "1995",
month = jan,
day = "1",
doi = "10.2307/2297845",
language = "English",
volume = "62",
pages = "131--157",
journal = "Review of Economic Studies",
issn = "0034-6527",
publisher = "Oxford University Press",
number = "1",
}