back start next
[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [ 99 ] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] [174] [175] [176] [177] [178] [179] [180] [181] [182] [183]
99 -The term is due to Irving Fisher (1933). For a modern treatment, see Bernanke and Gertler (1989). Gertler (1988) surveys work in this area. Section 8.7 develops a model of investment and the effects of changes in entrepreneurs wealth when ftnancial markets are imperfect. If a small decline in borrowers wealth causes a discontinuous drop in their abilm to borrow, the increase in costs is no longer small (see, for example, Mankiw, 1986a, and Bernanke and Gertler, 1990). But it is not clear why a small fall in borrowers wealth would induce lenders to stop lending if at the same time labor costs (for example) had dropped sharply. In addition, it is not clear why small redistributions would have large effects on the number of entrepreneurs who can borrow. In such settings, redistributions matter: transferring wealth from entrepreneurs to lenders makes the entrepreneurs more dependent on external finance, and thus reduces investment. Thus if debt contracts are not indexed, nominal disturbances are likely to have real effects. Indeed, price and wage flexibility can increase the distributional effects of nominal shocks, and thus potentially increase their real effects. This channel for real effects of nominal shocks is known as debt-deflation? This view of the nature of nominal imperfections must confront the same issues that face theories based on frictions in nominal price adjustment. For example, when a decline in the money stock redistributes wealth from firms to lenders because of nonindexation of debt contracts, firms marginal cost curves shift up. For reasonable cases, this upward shift is not large. If marginal cost faUs greatly when aggregate output faUs (because real wages decline sharply, for example) and marginal revenue does not. the modest increase in costs caused by the fall in the money stock leads to only a small decline in aggregate -* If marginal cost changes little and marginal revenue is very responsive to aggregate output, on the other hand, the small change in costs leads to large changes in output. Thus the same kinds of forces needed to cause small barriers to price adjustment to lead to large fluctuations in aggregate output are also needed for smaU costs to indexing debt contracts to have this effect. This discussion suggests an alternative interpretation of the Lucas model. Recall that Lucass model is based on the assumptions of imperfect information about the aggregate price level and considerable intertemporal substitution in labor supply, and that neither of these assumptions appear to be good first approximations. The discussion here, however, suggests that Lucass central results do not rest on these assumptions. Suppose th? price-setters choose not to acquire current information about the pric level, and that the behavior of the economy is therefore described by the Lucas model. In such a situation, price-setters incentive to obtain information about the price level, and to adjust their pricing and output decisions accordingly, is determined by the same considerations that determine their incentive to adjust their nominal prices in menu-cost models. As we ha\ e seen, there are many possible mechanisms other than intertemporal sub-stttution that can cause this incentive to be small. Thus neither unavail-
Another recent line of work investigates the consequences of the fact that at any given time, not all agents are adjusting their holdings ot high-powered money. Thus when the monetary authority changes the quantity of high-powered money, it cannot achieve a proportionate change in everyones holdmgs. As a result, a change in the money stock generally affects real money balances even if all prices and wages are perfectly flexible. Under appropriate conditions (such as an impact of real balances on consumption), this change in real balances affects the real interest rate. And if the real interest rate affects aggregate supply, the result is that aggregate output changes. See Grossman and Weiss (1983); Rotemberg (1984); Lucas (1990b); Fuerst (1992); and Christiano and Eichenbaum (1992b). ability of information about the price level nor intertemporal substitution is essential to the mechanism identified by Lucas. The friction in nominal adjustment may therefore be a small inconvenience or cost of obtaining information about the price level (or of adjusting ones pricing decisions in light of that information). Whether this friction is important in practice remains an open question. 6.13 Empirical Applications The Average Inflation Rate and the Output-Inflation Tradeoff Ball, Mankiw, and D. Romer (1988) point out that if the real effects of aggregate demand movements arise from frictions in price adjustment, then the average rate of inflation is likely to influence the size of those effects. Their argument is straightforward. The higher average inflation is, the more often firms must adjust their prices to keep up with the price level. This implies that when there is an aggregate demand disturbance, firms can pass it into prices more quickly. Thus its real effects are smaller. Ball, Mankiw, and Romers basic test of this prediction is analogous to Lucass test of his prediction that the variance of aggregate demand should influence the real effects of demand shocks. Following Lucas, they first estimate the real impact of aggregate demand shifts (denoted ,) in a large number of countries using the specification in equation (6.34). They then ask how those estimated impacts are related to average inflation. Figure 6.5 shows a scatterplot of the estimated t,s against average inflation for the 43 countries considered by BaU, Mankiw, and Romer. The figure suggests a negative relationship. The corresponding regression (with a quadratic term included to account for the nonlinearity apparent in the figure) is = 0.600 - 4.835, + 7.118? (0.079) (1.074) (2.088) (g.gi) = 0.388, s.e.e. = 0.215,
-0.3 0.2 0.4 Mean inflation FIGURE 6.5 The output-inflation tradeoff and average inflation (from Ball, Mankiw, and Romer, 1988) where , is average inflation m country i and the numbers in parentheses are standard errors. The point estimates imply that / = 4.835 -2(7.118)¥, which is negative for ¥ < 4.835/[2(7.118)] = 34%. Thus there is a statistically significant negative relationship between average inflation and the estimated real impact of aggregate demand movements. Recall that the Lucas model predicts that the variance of aggregate demand shocks affects , and that the data appear consistent with this prediction. Moreover, countries with higher average inflation generally have more variable aggregate demand. Thus it is possible that the results in (6.91) arise not because directly affects , but because it is correlated with the standard deviation of nominal GNP growth ( ), which does dhectly affect . Alternatively, it is possible that the earlier results, which appeared supportive of the Lucas model, in fact arise from the fact that and are correlated. The appropriate way to test between these two views is to include both variables in the regression. Again quadratic terms are included to allow for nonlinearities. The results are t, = 0 .589 - 5.729, + 8.4062 i241(Tx - 2.380a-2 (0.086) (1.973) (3.849) (2.467) (7.062) (6.92) = 0.359, s.e.e. = 0.219. The coefficients on the average inflation variables are essentially the same as in the previous regression, and they remain statistically significant. The
[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [ 99 ] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] [174] [175] [176] [177] [178] [179] [180] [181] [182] [183]
|