Economic Growth Second Edition
Download 0.79 Mb. Pdf ko'rish
|
BarroSalaIMartin2004Chap1-2
2.6
Transitional Dynamics 2.6.1 The Phase Diagram The Ramsey model, like the Solow–Swan model, is most interesting for its predictions about the behavior of growth rates and other variables along the transition path from an initial factor ratio, ˆk (0), to the steady-state ratio, ˆk ∗ . Equations (2.24), (2.25), and (2.26) determine the path of ˆk and ˆc for a given value of ˆk (0). The phase diagram in figure 2.1 shows the nature of the dynamics. 19 We first display the ˙ˆc = 0 locus. Since ˙ˆc = ˆc · (1/θ) · [ f (ˆk) − δ − ρ − θx], there are two ways for ˙ˆc to be zero: ˆc = 0, which corresponds to the horizontal axis in figure 2.1, and f (ˆk) = δ + ρ + θx, which is a vertical line at ˆk ∗ , the capital-labor ratio defined in equation (2.29). We note that ˆc is rising for ˆk < ˆk ∗ (so the arrows point upward in this region) and falling for ˆk > ˆk ∗ (where the arrows point downward). Recall that the solid curve in figure 2.1 shows combinations of ˆk and ˆc that satisfy ˙ˆk = 0 in equation (2.24). This equation also implies that ˆk is falling for values of ˆc above the solid curve (so the arrows point leftward in this region) and rising for values of ˆc below the curve (where the arrows point rightward). Since the ˙ˆc = 0 and the ˙ˆk = 0 loci cross three times, there are three steady states: the first one is the origin ( ˆc = ˆk = 0), the second steady state corresponds to ˆk ∗ and ˆc ∗ , and 19. See the appendix on mathematics for a discussion of phase diagrams. Growth Models with Consumer Optimization 103 the third one involves a positive capital stock, ˆk ∗∗ > 0, but zero consumption. We neglect the solution at the origin because it is uninteresting. The second steady state is saddle-path stable. Note, in particular, that the pattern of arrows in figure 2.1 is such that the economy can converge to this steady state if it starts in two of the four quadrants in which the two schedules divide the space. The saddle-path property can also be verified by linearizing the system of dynamic equations around the steady state and noting that the determinant of the characteristic matrix is negative (see appendix 2A, section 2.8, for details). This sign for the determinant implies that the two eigenvalues have opposite signs, an indication that the system is locally saddle-path stable. The dynamic equilibrium follows the stable saddle path shown by the solid locus with arrows. Suppose, for example, that the initial factor ratio satisfies ˆk (0) < ˆk ∗ , as shown in figure 2.1. If the initial consumption ratio is ˆc (0), as shown, the economy follows the stable path toward the steady-state pair, (ˆk ∗ , ˆc ∗ ). This path satisfies all the first-order conditions, including the transversality condition, as shown in the previous section. The two other possibilities are that the initial consumption ratio exceeds or falls short of ˆc (0). If the ratio exceeds ˆc(0), the initial saving rate is too low for the economy to remain on the stable path. The trajectory eventually crosses the ˙ˆk = 0 locus. After that crossing, ˆc continues to rise, ˆk starts to decline, and the path hits the vertical axis in finite time, at which point ˆk = 0. 20 The condition f (0) = 0 implies ˆy = 0; therefore, ˆc must jump down- ward to 0 at this point. Because this jump violates the first-order condition that underlies equation (2.25), these paths—in which the initial consumption ratio exceeds ˆc (0)—are not equilibria. 21 The final possibility is that the initial consumption ratio is below ˆc (0). In this case, the initial saving rate is too high to remain on the saddle path, and the economy eventually crosses the ˙ˆc = 0 locus. After that crossing, ˆc declines and ˆk continues to rise. The economy converges to the point at which the ˙ˆk = 0 schedule intersects the horizontal axis, a point which we labeled ˆk ∗∗ . Note, in particular, that ˆk rises above the golden-rule value, ˆk gold , and asymptotically approaches a higher value of ˆk. Therefore, f (ˆk) − δ falls below x + n asymptotically, and the path violates the transversality condition given in equation (2.26). This violation of the transversality condition means that households are oversaving: utility 20. We can verify from equation (2.24) that ˙ˆk becomes more and more negative in this region. Therefore, ˆk must reach 0 in finite time. 21. This analysis applies if investment is reversible. If investment is irreversible, the constraint ˆc ≤ f (ˆk) becomes binding before the trajectory hits the vertical axis. That is, the paths that start from points such as ˆc 0 in figure 2.1 would eventually hit the production function, ˆc = f (ˆk), which lies above the locus for ˙ˆk = 0. Thereafter, the path would follow the production function downward toward the origin. Appendix 2B (section 2.9) shows that such paths are not equilibria. 104 Chapter 2 would increase if consumption were raised at earlier dates. Accordingly, paths in which the initial consumption ratio is below ˆc (0) are not equilibria. This result leaves the stable saddle path leading to the positive steady state, ˆk ∗ , as the only possibility. 22 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