Economic Growth Second Edition
Appendix 2B: Irreversible Investment
Download 0.79 Mb. Pdf ko'rish
|
BarroSalaIMartin2004Chap1-2
2.9
Appendix 2B: Irreversible Investment Suppose that investment is irreversible, so that ˆc ≤ f (ˆk) applies. Reconsider in this case the dynamic paths that start with ˆk < ˆk ∗ at a position such as ˆc 0 in figure 2.1. These paths would eventually hit the production function, ˆc = f (ˆk), after which the constraint from irreversible investment would become binding. Thereafter, the paths would move downward along the production function, so that ˆc = f (ˆk) would apply. Hence, the capital intensity would decline in accordance with ˙ˆk = −(x + n + δ) · ˆk. Therefore, ˆk (and ˆc) would asymptotically approach zero but would not reach zero in finite time. We now argue that such paths cannot be equilibria. When the constraint ˆc ≤ f (ˆk) is binding, so that all output goes to consumption and none to gross investment, the price of capital, denoted by φ, can fall below 1. The rate of return to holders of capital then satisfies (see note 11) r = R/φ − δ + ˙φ/φ (2.89) Profit maximization for competitive firms still implies the condition R = f (ˆk), which can be substituted into the formula for r . Consumer optimization entails, as usual, ˙c/c = (1/θ) · (r − ρ) Therefore, substitution for r from equation (2.89) yields the formula for the growth rate of ˆc: ˙ˆc/ˆc = 1 θφ · [ f (ˆk) + ˙φ − φ · (δ + ρ + θx)] (2.90) Growth Models with Consumer Optimization 135 The condition ˆc = f (ˆk), together with ˙ˆk = −(x + n + δ) · ˆk, implies another condition for the growth rate of ˆc: ˙ˆc/ˆc = −α(ˆk) · (x + n + δ) (2.91) where α(ˆk) ≡ ˆk · f (ˆk)/f (ˆk) is the capital share of income (which is a constant in the case of a Cobb–Douglas production function). Therefore, equations (2.90) and (2.91) imply a condition for ˙ φ: ˙φ = − f (ˆk) + φ · [δ + ρ + θx − α(ˆk) · θ · (x + n + δ)] (2.92) Suppose that the constraint ˆc ≤ f (ˆk) first becomes binding at some date T , where ˆk(T ) < ˆk ∗ applies. At this point, f (ˆk) − δ > ρ + θx. Therefore, when φ = 1 (just at time T ), equation (2.92) implies that ˙ φ < 0. Over time, the rise in R and the fall in φ tend to raise r in accordance with equation (2.81). Nevertheless, households are satisfied with a negative growth rate of ˆc (equation [2.91]) because the rate of capital loss, ˙ φ/φ, rises sufficiently in magnitude to maintain a low rate of return, r . However, equation (2.92) implies, as ˆk decreases and f (ˆk) rises, that ˙φ eventually rises in magnitude toward infinity (regardless of what happens to α[ˆk] in the range between 0 and 1). Therefore, φ would reach zero in finite time and then become negative. This condition violates free disposal with respect to claims on capital. Hence, paths in which the irreversibility constraint, ˆc ≤ f (ˆk), is binding cannot exist in the region where ˆk < ˆk ∗ . The constraint ˆc ≤ f (ˆk) can be binding in the region where ˆk > ˆk ∗ . This possibility was noted and discussed by Arrow and Kurz (1970). Download 0.79 Mb. Do'stlaringiz bilan baham: |
Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2025
ma'muriyatiga murojaat qiling
ma'muriyatiga murojaat qiling