4.3 Dynamic panel results
4.3.3 The country group-specific analysis
4.3.3 The country group-specific analysis
The above results are estimated with a full sample. However, the inflationary effect of fiscal deficits could vary across different country groups. According to previous empirical studies, the fiscal deficits might play a stronger role for inflation in de-veloping countries, but play a weaker or even none role for inflation in developed economies (Giannaros and Kolluri, 1986; Protopapadakis and Siegel, 1987; Barhart and Darrat, 1988; De Haan and Zelhorst, 1990; Cat˜ao and Terrones, 2005; Kwon et al., 2009).
Therefore, I break down the full sample into two country groups of high-income vs. middle- and low-income countries and OECD vs. non-OECD countries to esti-mate equation (16),17 where the high-income and OECD country groups represent the countries of a higher development level.18 This breakdown is based on the classi-fication of the World Bank list of economies (July 2009). In addition, the classifica-tion of “OECD” in this case are the OECD members which are in the high-income group. Therefore, the high-income group contains the OECD group.19
First, the results of high-income countries are shown in Table 14 and the upper part of Figure 6. The results of middle- and low-income countries is shown in Table 15 and the lower part of Figure 6. As we can see, from quantile 0.1 to 0.9, the summations of the coefficients of current and lagged deficit-to-money ratios of high-income countries are 0.0094, 0.0188, 0.0232, 0.0224, 0.0272, 0.0349, 0.0441, 0.0581
17Exogenous controlled variables (the growth rate of real GDP per capita, oil price inflation, openness and the exchange rate regime) are not considered here.
18Data of inflation and deficit-to-money ratio are both stationary in high-income, middle- and low-income, OECD and non-OECD countries. The t-statistics of the Levin-Lin-Chu test (one lag) for inflation in high-income, middle- and low-income, OECD and non-OECD countries are -16.51, -24.19, -12.27 and -25.95 respectively. The t-statistics of the Levin-Lin-Chu test (one lag) for the deficit-to-money ratio in high-income, middle- and low-income, OECD and non-OECD countries are -14.77, -21.55, -11.90 and -23.25 respectively.
19The classification “OECD” is broadly consistent with the “advanced economies” of Cat˜ao and Terrones (2005). Cyprus is in their “advanced economies” but not in my “OECD” classification, and Hungary and Korea are not included in Cat˜ao and Terrones’ evaluation but are included in mine.
‧
and 0.0712. The summation of the coefficients becomes larger when the quantile goes higher as above, but the coefficient is only significant at the 5% level at quantiles 0.8 and 0.9, and significant at the 10% level at quantiles 0.2, 0.3 and 0.7. On the other hand, the summations of the coefficients of current and lagged deficit-to-money ratios of middle- and low-income countries are 0.0210, 0.0235, 0.0301, 0.0291, 0.0328, 0.0369, 0.0434, 0.0733 and 0.1362 from quantiles 0.1 to 0.9. We can see that the higher the quantile, the larger the summation of the coefficients. They are insignificant at quantile 0.1 and significant at the 5% level at the other quantiles except for quantile 0.7, where the estimate is significant at the 10% level.
Second, the results of the OECD countries are shown in Table 16 and the upper part of Figure 7, and the results of non-OECD countries are shown in Table 17 and the lower part of Figure 7. Accordingly, the summations of the coefficients of OECD countries are 0.0142, 0.0217, 0.0215, 0.0224, 0.0288, 0.0365, 0.0412, 0.0378 and 0.0436 from quantile 0.1 to 0.9. They are significant at the 1% level at quan-tiles 0.2, 0.3 and 0.6–0.8, significant at the 5% level at quanquan-tiles 0.1, 0.4 and 0.5, and significant at the 10% level at quantile 0.9. The coefficient is still larger as the quantile goes higher. On the other hand, the summations of the coefficients of non-OECD countries are 0.0181, 0.0246, 0.0286, 0.0302, 0.0339, 0.0386, 0.0480, 0.0748 and 0.1148 from quantiles 0.1 to 0.9. The coefficient becomes larger as the quantile goes higher, and they are significant at the 5% level at quantiles 0.2–0.7 and significant at the 1% level at quantiles 0.8 and 0.9. Compared the summations of the coefficients of the current and lagged deficit-to-money ratios, we can see that fiscal deficits impact middle- and low-income or non-OECD countries more strongly than high-income or OECD countries, especially at a high inflation level.
In conclusion, we can find that in whatever country group, the higher the quan-tile, the larger the summation of the coefficients. However, in higher development countries such as in high-income and OECD countries, fiscal deficits play a weaker role in inflation. Contrarily, the impact of fiscal deficits on inflation is generally
‧
greater in developing countries, which becomes more apparent when inflation is at a high level. In addition, in developing countries, fiscal deficits play no role only when inflation is at a low level (quantile 0.1).
Compared with the aforementioned previous research whose estimates of OECD or advanced countries are small and insignificant, the estimates of OECD countries are significant except for quantile 0.9 in my study. Nevertheless, the estimates of OECD countries are all smaller than the estimates of non-OECD countries, and both the D-GMM L and D-GMM D estimators are negative and insignificant. Hence, the inflationary effect of fiscal deficits is not robust and may not exist in OECD countries.
There are some explanations for various inflationary effects of fiscal deficits. In developing countries, a smaller taxable capacity, political instability, a less inde-pendent central bank, and limited access to domestic and external debt financing could lower the relative costs of seigniorage and inflation tax, so the governments tend to rely on monetary accommodation. On the other hand, in higher develop-ment countries, financial depths are deeper, so public bonds could be absorbed and the governments would depend less on monetization (De Haan and Zelhorst, 1990;
Cukierman et al. 1992; Alesina and Summers 1993; Aisen and Veiga, 2008).
4.3.4 The subsample period analysis and central bank independence
For examining whether the parameters are robust or not, I ran regressions for three sub-sample periods: 1970–2006, 1980–2006 and 1990–2006, which contained 96, 101 and 98 countries respectively.20 These three sub-sample periods are divided arbitrarily, because the main results above are estimated over the period from 1960–
2006. Equation (16) is estimated as follows.21 As the observations of the main
20Recall that the datasets of the main results over 1960–2006 contains 91 countries.
21As the country group-specific analysis, exogenous controlled variables are not considered here.
Data of inflation and the deficit-to-money ratio are both stationary over 1970–2006, 1980–2006 and 1990–2006. The t-statistics of the Levin-Lin-Chu test (one lag) for inflation over 1970–2006, 1980–2006 and 1990–2006 are -27.61, -25.17 and -293.28; and the t-statistics of the Levin-Lin-Chu test (one lag) for the deficit-to-money ratio is over 1970–2006, 1980–2006 and 1990–2006 are -25.32,
‧
results, the observations during 1970–1973, 1980–1983 and 1990–1993 are dropped.
First, the results covering 1970–2006 are shown in Table 18 and the upper part of Figure 8. From quantile 0.1 to 0.9, the summations of the coefficients of current and lagged deficit-to-money ratios are 0.0039, 0.0119, 0.0178, 0.0222, 0.0266, 0.0315, 0.0440, 0.0634 and 0.1072 respectively, and insignificant at quantile 0.1 and 0.2, and significant at a 10% level at quantile 0.3, significant at a 5% level above quantile 0.4, and significant at a 1% level at quantile 0.8 and 0.9. The estimates of D-GMM L and D-GMM D are 0.0453 and 0.0704, and both are significant at a 1% level.
Second, the results covering 1980–2006 are shown in Table 19 and the lower part of Figure 8. From quantile 0.1 to 0.9, the summations of the coefficients of current and lagged deficit-to-money ratios are -0.0014, 0.0051, 0.0098, 0.0140, 0.0182, 0.0219, 0.0291, 0.0448 and 0.0961 respectively, and insignificant from quantile 0.1 to 0.4.
They are significant at a 10% level at quantiles 0.5 and 0.8, and significant at a 5%
level above quantile 0.6, 0.7 and 0.9. The estimates of D-GMM L and D-GMM D are 0.0414 and 0.0958, and both are significant at a 1% level.
Third, the results covering 1990–2006 are shown in Table 20 and the upper part of Figure 9. From quantile 0.1 to 0.9, the summations of the coefficients of the current and lagged deficittomoney ratio are 0.0234, 0.0198, 0.0139, 0.0118, -0.0106, -0.0071, -0.0037, -0.0016 and 0.0190, and negatively significant at a 5% level from quantile 0.1 to 0.4, which is significant at a 10% level at quantile 0.5 and 0.8 and is insignificant at quantile 0.4, 0.5 and 0.9. The estimates of D-GMM L and D-GMM D are 0.0253 and 0.0330 respectively. The former is significant at a 5%
level and the latter is significant at a 1% level. However, the estimates of quantile regression are insignificant, so the results are not robust.
Compared with the main results covering 1960–2006 (see Table 5 and Figure 2), we know that the results covering 1970–2006 are broadly unchanged. The results covering 1980–2006 shifts slightly downward, but they are still similar with the main
-25.74 and -25.05 respectively.
‧
results. The higher the quantile, the larger and more significant are the coefficients.
However, the results covering 1990–2006 appreciably shift downward, and the sign of the coefficients at the low quantile becomes significantly negative, which is opposite to the theoretical prediction. In addition, although the signs of the coefficients D-GMM L and D-GMM D are both positive, the results are not robust.
Therefore, the outcomes for various sub-sample periods confirm the results that fiscal deficits would be more inflationary as inflation is higher. Yet, after 1990, the inflationary effect of fiscal deficits is not detected, which means that the “fiscal dominance” hypothesis does not hold after the 90s.
A reasonable conjecture is that central banks around the world became more independent after the 90s (Crowe and Meade, 2007; Cukierman, 2008),22 because a less independent central bank might be more possibly forced to finance budget deficits by money creation (fiscal dominance). Hence, the deficit-inflation relation-ship might emerge after controlling the effect of central bank independence.
For capturing the effect of central bank independence, I take the central bank in-dependence index as controlled variable to estimate the deficit-inflation relationship after 90s again. Ranging from 0 to 1, the smaller the dummy the less indepen-dent the central bank. However, the central bank independence index is only up to the year 2000 and not available to a full sample during 1990–2006.23 Due to the limitation of data availability, I investigated the deficit-inflation relationship during 1990–2000 and the number of countries drops to 57 (see Appendix E).24
After controlling the effect of central bank independence, the results over 1990–
2000 are reported in Table 21 and plotted in the lower part of Figure 9. From quantile
22In my horizons, although some studies have investigated the deficit-inflation relationship for various sub-sample periods, no studies have estimated data from the 1990s to the 2000s.
23The central bank independence index is developed by Cukierman et al. (1992), but the data they provided is only up to the year 1988. Based on the measurement of Cukierman et al. (1992), Polillo and Guill´en (2005) updated the data to the year 2000. The index I used here is calculated by Polillo and Guill´en (2005).
24All variables are stationary over 1990–2000, and the t-statistic of the Levin-Lin-Chu test (one lag) for inflation and deficit-to-money ratio are -95.88 and -12.24 respectively.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
0.1 to 0.9, the summations of the coefficients of the current and lagged deficit-to-money ratio are -0.0210, -0.0169, -0.0059, -0.0023, -0.0010, -0.0006, 0.0023, 0.0075 and 0.0193 respectively, and all of them are insignificant. The estimates of D-GMM L and D-GMM D are -0.0547 and -0.0574 respectively, and the former is insignificant and the latter is only significant at a the 10% level.
Accordingly, different from the prediction, the deficit-inflation relationship is still not detected after controlling the effect of central bank independence. Therefore, rather than central bank independence, there are other reasons for why “fiscal dom-inance” does not hold after 90s, such as a better coordination between the fiscal and monetary policy.
‧
Sustained fiscal deficits might cause inflation by means of money creation, and the economy in a higher inflation level would be more strongly impacted by deficits.
In this study, I investigated the deficit-inflation relationship with a panel dataset which covers 91 countries over 1960–2006. Empirically, I followed the theoretical model of Cat˜ao and Terrones (2005) to scale fiscal deficits by narrow money and considered fiscal deficits distributed-lag. Furthermore, a new econometric method, dynamic panel quantile regression of Lin (2010) was utilized for eliminating biases raised by the dynamic term and estimating the impacts of fiscal deficits at various inflation levels with a panel dataset.
Empirical results of the dynamic panel quantile regression show that sustained fiscal deficits will be more inflationary the higher the inflation, and play a weak or no role in inflation when inflation is at a low level. The estimated relationship remains while taking one or more lagged deference dependant variables as instruments. Scal-ing deficits by GDP and controllScal-ing money growth, the results are similar except that the estimates become significant at a low quantile (0.1). Accordingly, fiscal consolidation would be more effective in price stabilization as inflation rises higher.
Also, the results remain robust while controlling the growth of GDP per capita, oil price inflation, openness and the exchange rate regime, so the deficit-inflation rela-tionship can be stable. Compared with higher developed countries (represented by high-income and OECD countries), the impact of fiscal deficits on inflation is gen-erally greater among developing countries (represented by middle- and low-income and non-OECD countries), which becomes more apparent when inflation is at a high level. Finally, the deficit-inflation relationship does not notably change during 1960–2006, 1970–2006 and 1980–2006. However, it is not detected over 1990–2006.
Dynamic panel quantile regression can help us observe the impacts of fiscal deficits at various inflation levels and allows for intrinsic dynamic adjustment. It is an outstanding econometric method for investigating the deficit-inflation relation-ship, and therefore the relationship becomes clearer.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table4:Cross-sectionalresultsover1960–2006 dependentvariable:averageinflation quantile OLS0.10.20.30.40.50.60.70.80.9 (A) intercept0.8038∗∗∗ 0.5273∗∗∗ 0.5497∗∗∗ 0.6352∗∗∗ 0.6687∗∗∗ 0.7211∗∗∗ 0.8067∗∗∗ 0.9077∗∗∗ 1.0556∗∗∗ 1.1622∗∗∗ (0.0671)(0.0418)(0.0394)(0.0529)(0.0487)(0.0612)(0.0704)(0.0982)(0.1183)(0.1990) averagedeficit/money0.8687∗∗∗ 0.5287∗∗∗ 0.5869∗∗∗ 0.5803∗∗∗ 0.7813∗∗∗ 0.7457∗∗∗ 0.6716∗∗ 0.8603∗∗ 0.81441.7967∗ (0.2136)(0.1465)(0.1459)(0.2116)(0.2053)(0.2365)(0.3297)(0.4118)(0.5407)(0.9735) (B) intercept0.8521∗∗∗0.4736∗∗∗0.5655∗∗∗0.6486∗∗∗0.7006∗∗∗0.8317∗∗∗0.9160∗∗∗0.9686∗∗∗1.1076∗∗∗1.2607∗∗∗ (0.0703)(0.0478)(0.0417)(0.0411)(0.0641)(0.0903)(0.0930)(0.1098)(0.1408)(0.2240) averagedeficit/GDP4.7201∗∗∗ 4.2520∗∗∗ 3.7353∗∗∗ 3.2610∗∗∗ 2.9507∗∗ 2.18951.70033.30684.57728.1045 (1.5965)(1.1891)(0.8453)(0.8604)(1.2634)(2.0928)(2.8107)(3.4130)(4.3226)(6.2514) (C) intercept-0.4520∗∗∗ -0.1403-0.5977∗∗ -0.5678∗∗∗ -0.5579∗∗∗ -0.5066∗∗∗ -0.5113∗∗∗ -0.4849∗∗∗ -0.5360∗∗∗ -0.4291∗∗ (0.0801)(0.3432)(0.2499)(0.1226)(0.0839)(0.0692)(0.0736)(0.0862)(0.1301)(0.1675) averagedeficit/GDP0.60312.59151.13920.1848-0.13210.11230.51900.62640.5740-0.1386 (0.7813)(2.0766)(1.3133)(1.0395)(0.9156)(0.7143)(0.5302)(0.5106)(0.5944)(0.9388) averagemoneygrowth1.1724∗∗∗ 0.6631∗ 1.2118∗∗∗ 1.2456∗∗∗ 1.2775∗∗∗ 1.2526∗∗∗ 1.2569∗∗∗ 1.2467∗∗∗ 1.3152∗∗∗ 1.3279∗∗∗ (0.0656)(0.3396)(0.2516)(0.1163)(0.0688)(0.0517)(0.0606)(0.0714)(0.1032)(0.1233) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Figure 1: Cross-sectional results over 1960–2006
The upper is the estimates of average deficit/money. The middle is the estimates of average deficit/GDP. The lower is the estimates of average deficit/GDP and average money growth is controlled. The black solid line: the coefficients of quantile regres-sion. The black dotted line: the 95% confidence interval of quantile regresregres-sion. The grey solid line: the coefficient of OLS.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table5:Dynamicpanelresultsover1960–2006I(A) dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.0038∗ 0.0210∗∗∗ 0.49000.49700.50030.50240.50570.51100.51420.52320.5369 (0.0022)(0.0021)(3.1146)(3.1148)(3.1147)(3.1147)(3.1148)(3.1147)(3.1146)(3.1147)(3.1147) inflationt−10.7294∗∗∗ 0.2176∗∗∗ 0.22660.3102∗∗ 0.3299∗∗ 0.3677∗∗∗ 0.3735∗∗∗ 0.3697∗∗∗ 0.4447∗∗∗ 0.4531∗∗∗ 0.5687∗∗∗ (0.0204)(0.0348)(0.1669)(0.1452)(0.1318)(0.1261)(0.1188)(0.1165)(0.1234)(0.1394)(0.1342) deficit/moneyt0.0687∗∗∗0.0626∗∗∗-0.00260.00250.0073∗0.0105∗∗0.0134∗∗∗0.0142∗∗0.0178∗∗0.0259∗∗∗0.0263∗ (0.0039)(0.0034)(0.0063)(0.0044)(0.0044)(0.0044)(0.0050)(0.0058)(0.0077)(0.0082)(0.0135) deficit/moneyt−10.0105∗∗∗ 0.0372∗∗∗ 0.00480.0073∗∗ 0.0073∗∗ 0.0074∗∗∗ 0.0090∗∗∗ 0.0114∗∗∗ 0.0129∗∗ 0.01190.0448∗∗∗ (0.0037)(0.0036)(0.0058)(0.0031)(0.0031)(0.0027)(0.0030)(0.0043)(0.0051)(0.0095)(0.0135) deficit/moneyt−2-0.0403∗∗∗ -0.0166∗∗∗ 0.00710.00300.00460.00360.00420.00150.00440.0121∗∗∗ 0.0145∗∗ (0.0037)(0.0034)(0.0044)(0.0031)(0.0033)(0.0032)(0.0034)(0.0038)(0.0041)(0.0045)(0.0059) deficit/moneyt−30.00490.00210.00000.00370.00450.00560.00460.0098∗∗ 0.0102∗∗ 0.0161∗∗ 0.0272∗∗∗ (0.0040)(0.0034)(0.0068)(0.0039)(0.0038)(0.0039)(0.0045)(0.0049)(0.0050)(0.0068)(0.0083) P3 j=0deficit/moneyt−j0.0438∗∗∗ 0.0853∗∗∗ 0.00940.0165∗∗ 0.0236∗∗∗ 0.0272∗∗∗ 0.0312∗∗∗ 0.0369∗∗∗ 0.0452∗∗∗ 0.0660∗∗∗ 0.1128∗∗∗ (0.0080)(0.0074)(0.0105)(0.0075)(0.0079)(0.0087)(0.0104)(0.0127)(0.0152)(0.0203)(0.0257) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. aDynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table6:Dynamicpanelresultsover1960–2006I(B) dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.0046∗∗ 0.0157∗∗∗ 0.66640.67530.67750.68170.68380.68640.69210.69940.7189 (0.0021)(0.0019)(2.6058)(2.6058)(2.6057)(2.6057)(2.6057)(2.6057)(2.6057)(2.6057)(2.6058) inflationt−10.7228∗∗∗ 0.4166∗∗∗ 0.1993∗ 0.2777∗∗∗ 0.3070∗∗∗ 0.3029∗∗∗ 0.3376∗∗∗ 0.3868∗∗∗ 0.4134∗∗∗ 0.4447∗∗∗ 0.4506∗∗∗ (0.0160)(0.0207)(0.1115)(0.0962)(0.0908)(0.0820)(0.0761)(0.0874)(0.0820)(0.0746)(0.0856) deficit/moneyt0.0688∗∗∗0.0635∗∗∗-0.00330.00200.0080∗0.0106∗∗0.0134∗∗∗0.0153∗∗0.0183∗∗0.0308∗∗∗0.0224 (0.0039)(0.0035)(0.0064)(0.0043)(0.0048)(0.0046)(0.0051)(0.0064)(0.0089)(0.0088)(0.0160) deficit/moneyt−10.0106∗∗∗ 0.0258∗∗∗ 0.0120∗ 0.0081∗∗ 0.0084∗∗ 0.0106∗∗∗ 0.0109∗∗∗ 0.0111∗∗ 0.0147∗∗ 0.0184∗ 0.0456∗∗∗ (0.0036)(0.0034)(0.0063)(0.0034)(0.0034)(0.0034)(0.0042)(0.0049)(0.0058)(0.0108)(0.0145) deficit/moneyt−2-0.0406∗∗∗ -0.0270∗∗∗ 0.0073∗ 0.0046∗ 0.00390.00320.00290.0019-0.0001-0.00060.0043 (0.0036)(0.0033)(0.0038)(0.0028)(0.0032)(0.0031)(0.0035)(0.0042)(0.0049)(0.0076)(0.0130) deficit/moneyt−30.00330.0013-0.00710.00040.00250.00320.00320.00610.0112∗∗ 0.0163∗ 0.0413∗∗∗ (0.0039)(0.0034)(0.0084)(0.0039)(0.0040)(0.0041)(0.0045)(0.0050)(0.0055)(0.0096)(0.0121) P3 j=0deficit/moneyt−j0.0421∗∗∗ 0.0636∗∗∗ 0.00890.0151∗∗ 0.0228∗∗∗ 0.0277∗∗∗ 0.0305∗∗∗ 0.0344∗∗∗ 0.0441∗∗∗ 0.0650∗∗∗ 0.1137∗∗∗ (0.0077)(0.0070)(0.0104)(0.0073)(0.0080)(0.0088)(0.0102)(0.0123)(0.0149)(0.0206)(0.0259) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. aDynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1and∆πit−2. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2,πit−3and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2,∆πit−3and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table7:Dynamicpanelresultsover1960–2006I(C) dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.0066∗∗∗ 0.0099∗∗∗ 0.59640.60610.60880.61110.61300.61550.62050.63040.6517 (0.0020)(0.0019)(3.4057)(3.4059)(3.4058)(3.4059)(3.4058)(3.4059)(3.4058)(3.4058)(3.4056) inflationt−10.7197∗∗∗ 0.5998∗∗∗ 0.2420∗∗ 0.3064∗∗∗ 0.3258∗∗∗ 0.3553∗∗∗ 0.3947∗∗∗ 0.4459∗∗∗ 0.4718∗∗∗ 0.4609∗∗∗ 0.4626∗∗∗ (0.0141)(0.0181)(0.1151)(0.0950)(0.0828)(0.0780)(0.0818)(0.0865)(0.0769)(0.0774)(0.0691) deficit/moneyt0.0653∗∗∗0.0670∗∗∗-0.00090.00320.0077∗0.0118∗∗∗0.0128∗∗0.0188∗∗∗0.0198∗∗0.0324∗∗∗0.0273 (0.0038)(0.0037)(0.0069)(0.0042)(0.0044)(0.0042)(0.0050)(0.0065)(0.0085)(0.0081)(0.0180) deficit/moneyt−10.0090∗∗ 0.0164∗∗∗ 0.0120∗∗ 0.0111∗∗∗ 0.0105∗∗∗ 0.0099∗∗ 0.0096∗ 0.01000.0149∗∗ 0.01420.0438∗∗∗ (0.0036)(0.0035)(0.0058)(0.0035)(0.0038)(0.0045)(0.0057)(0.0063)(0.0076)(0.0137)(0.0146) deficit/moneyt−2-0.0417∗∗∗ -0.0361∗∗∗ 0.0017-0.00090.00200.00300.0037-0.00030.00290.00470.0128 (0.0036)(0.0034)(0.0047)(0.0034)(0.0041)(0.0049)(0.0055)(0.0059)(0.0071)(0.0106)(0.0116) deficit/moneyt−30.00180.0000-0.00160.00110.00280.00220.00400.00800.00780.0143∗ 0.0286∗∗∗ (0.0038)(0.0036)(0.0082)(0.0043)(0.0041)(0.0044)(0.0048)(0.0055)(0.0055)(0.0081)(0.0087) P3 j=0deficit/moneyt−j0.0345∗∗∗ 0.0473∗∗∗ 0.01120.0145∗∗ 0.0229∗∗∗ 0.0268∗∗∗ 0.0301∗∗∗ 0.0364∗∗∗ 0.0454∗∗∗ 0.0656∗∗∗ 0.1124∗∗∗ (0.0074)(0.0071)(0.0105)(0.0072)(0.0074)(0.0083)(0.0105)(0.0127)(0.0146)(0.0193)(0.0265) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. aDynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1,∆πit−2and∆πit−3. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2,πit−3,πit−4and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2,∆πit−3,∆πit−4and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Figure 2: Dynamic panel results over 1960–2006 I The estimates of P3
j=0deficit/moneyt−j. The upper is estimated with one instru-ment. The middle is estimated with two instruments. The lower is estimated with three instruments. The black solid line: the coefficients of quantile regression. The black dotted line: the 95% confidence interval of quantile regression. The grey solid line: the coefficient of D-GMM L. The grey dotted line: the coefficient of D-GMM D.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table8:Dynamicpanelresultsover1960–2006II(A) dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.00310.0031∗∗∗ 0.35090.36090.36600.36960.37280.37620.38120.38970.4033 (0.0023)(0.0023)(6.1837)(6.1838)(6.1838)(6.1837)(6.1837)(6.1837)(6.1838)(6.1840)(6.1838) inflationt−10.6993∗∗∗ 0.1419∗∗∗ 0.20080.2834∗∗ 0.2978∗∗ 0.3123∗∗ 0.3331∗∗∗ 0.3673∗∗∗ 0.4029∗∗∗ 0.4347∗∗∗ 0.5848∗∗∗ (0.0198)(0.0323)(0.1646)(0.1399)(0.1221)(0.1246)(0.1224)(0.1153)(0.1206)(0.1425)(0.1897) deficit/GDPt0.4234∗∗∗0.3326∗∗∗-0.00480.01620.01730.05040.04910.05070.04670.04180.0261 (0.0261)(0.0227)(0.0422)(0.0289)(0.0283)(0.0331)(0.0355)(0.0386)(0.0430)(0.0553)(0.1826) deficit/GDPt−10.2308∗∗∗ 0.3506∗∗∗ 0.0822∗∗ 0.0592∗∗ 0.0590∗∗∗ 0.0644∗∗∗ 0.0739∗∗∗ 0.0907∗∗∗ 0.1104∗∗ 0.1355∗ 0.3374∗∗ (0.0250)(0.0225)(0.0419)(0.0278)(0.0229)(0.0245)(0.0265)(0.0317)(0.0439)(0.0776)(0.1715) deficit/GDPt−2-0.2843∗∗∗ -0.0791∗∗∗ 0.0487∗ 0.0394∗ 0.0472∗∗∗ 0.01700.0105-0.00630.03240.0449∗∗∗ 0.0526 (0.0253)(0.0234)(0.0281)(0.0205)(0.0141)(0.0180)(0.0198)(0.0247)(0.0244)(0.0235)(0.0617) deficit/GDPt−30.00990.01950.04430.03900.0499∗∗ 0.0523∗ 0.0601∗ 0.0858∗∗ 0.0948∗ 0.1430∗∗ 0.1801∗ (0.0270)(0.0229)(0.0303)(0.0248)(0.0242)(0.0304)(0.0350)(0.0415)(0.0503)(0.0564)(0.0993) P3 j=0deficit/GDPt−j0.3798∗∗∗ 0.6236∗∗∗ 0.1704∗∗∗ 0.1537∗∗∗ 0.1734∗∗∗ 0.1841∗∗∗ 0.1936∗∗∗ 0.2208∗∗∗ 0.2843∗∗∗ 0.3651∗∗ 0.5961 (0.0609)(0.0536)(0.0654)(0.0583)(0.0566)(0.0639)(0.0718)(0.0838)(0.0946)(0.1487)(0.3877) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. aDynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table9:Dynamicpanelresultsover1960–2006II(B) dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.0048∗∗ 0.0065∗∗∗ 0.62640.61860.61450.61040.60880.60490.60430.60760.6158 (0.0023)(0.0019)(3.3176)(3.3176)(3.3177)(3.3176)(3.3175)(3.3176)(3.3176)(3.3177)(3.3176) inflationt−10.6498∗∗∗ 0.01700.2159∗ 0.3452∗∗∗ 0.4150∗∗∗ 0.4789∗∗∗ 0.4956∗∗∗ 0.5839∗∗∗ 0.6215∗∗∗ 0.6332∗∗∗ 0.6759∗∗∗ (0.0242)(0.0400)(0.1169)(0.1104)(0.1137)(0.1176)(0.1199)(0.1096)(0.1079)(0.1120)(0.1405) deficit/GDPt0.3602∗∗∗0.2440∗∗∗0.03080.03380.02390.03910.03990.08020.09190.1208∗0.0626 (0.0252)(0.0220)(0.0256)(0.0258)(0.0301)(0.0369)(0.0485)(0.0596)(0.0651)(0.0667)(0.0607) deficit/GDPt−10.2521∗∗∗ 0.3006∗∗∗ 0.06410.1030∗∗ 0.1108∗∗ 0.1297∗∗∗ 0.1381∗∗∗ 0.1248∗∗∗ 0.1024∗∗ 0.08590.2213∗∗∗ (0.0237)(0.0201)(0.0460)(0.0513)(0.0539)(0.0447)(0.0351)(0.0398)(0.0509)(0.0600)(0.0774) deficit/GDPt−2-0.1885∗∗∗ -0.0495∗∗ -0.0389-0.0430-0.0061-0.0171-0.0219-0.0446-0.0228-0.0196-0.0570 (0.0239)(0.0213)(0.0489)(0.0586)(0.0631)(0.0596)(0.0536)(0.0458)(0.0394)(0.0483)(0.0685) deficit/GDPt−30.0951∗∗∗ 0.0422∗∗ 0.0587∗∗ 0.0535∗ 0.02600.03130.03250.05010.04550.03360.0147 (0.0260)(0.0215)(0.0261)(0.0291)(0.0300)(0.0321)(0.0368)(0.0401)(0.0389)(0.0396)(0.0616) moneygrowtht0.1854∗∗∗0.2322∗∗∗0.0780∗∗0.1224∗∗∗0.1655∗∗∗0.1996∗∗∗0.2482∗∗∗0.2930∗∗∗0.3446∗∗∗0.4176∗∗∗0.5198∗∗∗ (0.0117)(0.0106)(0.0322)(0.0418)(0.0533)(0.0627)(0.0724)(0.0790)(0.0870)(0.0944)(0.0988) moneygrowtht−1-0.0230∗ 0.1581∗∗∗ 0.0730∗∗∗ 0.1114∗∗∗ 0.1328∗∗∗ 0.1671∗∗∗ 0.1949∗∗∗ 0.2150∗∗∗ 0.2484∗∗∗ 0.2684∗∗∗ 0.2703∗∗∗ (0.0123)(0.0144)(0.0267)(0.0243)(0.0225)(0.0252)(0.0281)(0.0306)(0.0319)(0.0310)(0.0442) moneygrowtht−2-0.1106∗∗∗ 0.0362∗∗∗ 0.0509∗∗ 0.0893∗∗∗ 0.1128∗∗∗ 0.1318∗∗∗ 0.1492∗∗∗ 0.1681∗∗∗ 0.1915∗∗∗ 0.1963∗∗∗ 0.1994∗∗∗ (0.0113)(0.0122)(0.0203)(0.0178)(0.0180)(0.0180)(0.0183)(0.0192)(0.0203)(0.0259)(0.0346) moneygrowtht−3-0.1110∗∗∗ -0.0308∗∗∗ 0.0268∗∗ 0.0568∗∗∗ 0.0842∗∗∗ 0.0883∗∗∗ 0.0999∗∗∗ 0.1009∗∗∗ 0.1054∗∗∗ 0.1187∗∗∗ 0.1465∗∗∗ (0.0106)(0.0095)(0.0125)(0.0167)(0.0137)(0.0130)(0.0149)(0.0212)(0.0279)(0.0350)(0.0464) P3 j=0deficit/GDPt−j0.5190∗∗∗0.5372∗∗∗0.1147∗∗∗0.1473∗∗∗0.1545∗∗∗0.1829∗∗∗0.1885∗∗∗0.2104∗∗∗0.2170∗∗∗0.2208∗∗∗0.2416∗∗ (0.0585)(0.0489)(0.0439)(0.0424)(0.0439)(0.0469)(0.0500)(0.0565)(0.0588)(0.0675)(0.1104) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. a DynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Figure 3: Dynamic panel results over 1960–2006 II
The estimates of P3
j=0deficit/GDPt−j with instruments ∆πit−1. The upper is the result without other controlled variables. The lower is the result of controlling money growth rates. The black solid line: the coefficients of quantile regression. The black dotted line: the 95% confidence interval of quantile regression. The grey solid line:
the coefficient of D-GMM L. The grey dotted line: the coefficient of D-GMM D.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table10:Dynamicpanelresultsover1960–2006III(A) dependentvariable:inflation quantilea D-GMMLbD-GMMDc0.10.20.30.40.50.60.70.80.9 intercept0.0082∗∗∗0.0239∗∗∗1.08361.09241.09511.09821.10171.10771.11441.12081.1346 (0.0022)(0.0021)(2.7128)(2.7129)(2.7130)(2.7130)(2.7130)(2.7130)(2.7129)(2.7131)(2.7132) inflationt−10.7281∗∗∗0.2139∗∗∗0.25270.3026∗∗0.3325∗∗0.3518∗∗∗0.3735∗∗∗0.3573∗∗∗0.3833∗∗∗0.4460∗∗∗0.6085∗∗∗ (0.0203)(0.0344)(0.1652)(0.1409)(0.1298)(0.1327)(0.1264)(0.1187)(0.1302)(0.1340)(0.1438) deficit/moneyt0.0670∗∗∗0.0616∗∗∗-0.00180.00030.00600.0096∗∗0.0130∗∗∗0.0141∗∗0.01310.0163∗0.0254∗ (0.0039)(0.0034)(0.0058)(0.0045)(0.0043)(0.0040)(0.0048)(0.0059)(0.0082)(0.0096)(0.0136) deficit/moneyt−10.0090∗∗0.0365∗∗∗0.00590.0067∗∗0.0063∗0.0069∗∗0.0069∗∗0.0087∗∗0.0126∗∗0.0162∗0.0411∗∗∗ (0.0037)(0.0036)(0.0053)(0.0030)(0.0033)(0.0031)(0.0030)(0.0042)(0.0054)(0.0093)(0.0136) deficit/moneyt−2-0.0404∗∗∗-0.0165∗∗∗0.00620.00360.00360.00360.00340.00370.0073∗∗0.0143∗∗∗0.0167∗∗∗ (0.0036)(0.0034)(0.0042)(0.0030)(0.0032)(0.0035)(0.0035)(0.0035)(0.0033)(0.0040)(0.0058) deficit/moneyt−30.00560.0027-0.00020.00460.0068∗0.0061∗0.00600.0088∗0.0099∗0.0167∗∗0.0288∗∗∗ (0.0040)(0.0034)(0.0068)(0.0039)(0.0037)(0.0037)(0.0048)(0.0046)(0.0054)(0.0071)(0.0104) growthofrealGDPpercapita-0.4001∗∗∗-0.2630∗∗∗-0.0795-0.1028∗∗∗-0.1065∗∗∗-0.1286∗∗∗-0.1560∗∗∗-0.1982∗∗∗-0.2430∗∗∗-0.3062∗∗∗-0.5654∗∗∗ (0.0474)(0.0404)(0.0489)(0.0250)(0.0278)(0.0341)(0.0381)(0.0441)(0.0538)(0.0671)(0.1352) P3 j=0deficit/moneyt−j0.0412∗∗∗0.0843∗∗∗0.01010.0153∗∗0.0227∗∗∗0.0262∗∗∗0.0294∗∗∗0.0353∗∗∗0.0429∗∗∗0.0635∗∗∗0.1120∗∗∗ (0.0079)(0.0073)(0.0101)(0.0076)(0.0082)(0.0089)(0.0109)(0.0128)(0.0159)(0.0212)(0.0258) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. a DynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2,∆witand∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2,∆witand∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table11:Dynamicpanelresultsover1960–2006III(B) dependentvariable:inflation quantilea D-GMMLbD-GMMDc0.10.20.30.40.50.60.70.80.9 intercept0.00020.0160∗∗∗0.25400.25860.26360.26740.27190.27840.27990.28480.3041 (0.0020)(0.0019)(12.3507)(12.3507)(12.3506)(12.3506)(12.3506)(12.3506)(12.3505)(12.3505)(12.3507) inflationt−10.6547∗∗∗0.2023∗∗∗0.20560.3011∗∗0.3124∗∗0.3195∗∗∗0.3050∗∗0.2771∗∗0.3511∗∗∗0.4105∗∗∗0.5083∗∗∗ (0.0184)(0.0314)(0.1621)(0.1475)(0.1291)(0.1183)(0.1188)(0.1082)(0.0944)(0.0907)(0.1529) deficit/moneyt0.0575∗∗∗0.0542∗∗∗-0.00420.00650.0089∗∗0.0111∗∗∗0.0133∗∗∗0.0168∗∗∗0.0178∗∗∗0.0213∗∗∗0.0191∗ (0.0035)(0.0031)(0.0060)(0.0044)(0.0041)(0.0039)(0.0051)(0.0059)(0.0067)(0.0065)(0.0103) deficit/moneyt−10.0090∗∗∗0.0337∗∗∗0.00660.0066∗∗0.0066∗0.0067∗∗0.0087∗∗∗0.0082∗∗0.00760.0135∗0.0219∗ (0.0033)(0.0033)(0.0053)(0.0027)(0.0039)(0.0034)(0.0029)(0.0034)(0.0047)(0.0082)(0.0120) deficit/moneyt−2-0.0341∗∗∗-0.0139∗∗∗0.0072∗0.00490.00260.00470.00510.00550.0075∗∗0.0113∗∗∗0.0107∗ (0.0033)(0.0031)(0.0040)(0.0032)(0.0030)(0.0036)(0.0038)(0.0034)(0.0033)(0.0029)(0.0058) deficit/moneyt−30.0072∗∗0.00350.00120.00430.0070∗0.00630.00580.00750.0159∗∗∗0.0190∗∗∗0.0276∗∗∗ (0.0036)(0.0031)(0.0057)(0.0035)(0.0041)(0.0040)(0.0050)(0.0058)(0.0056)(0.0052)(0.0098) growthofrealGDPpercapita-0.3282∗∗∗-0.2222∗∗∗-0.0606-0.0931∗∗∗-0.1194∗∗∗-0.1261∗∗∗-0.1439∗∗∗-0.1861∗∗∗-0.2130∗∗∗-0.2715∗∗∗-0.3953∗∗∗ (0.0426)(0.0370)(0.0451)(0.0273)(0.0260)(0.0241)(0.0371)(0.0348)(0.0372)(0.0574)(0.0947) oilpriceinflation0.1704∗∗∗0.1422∗∗∗0.0481∗∗∗0.0391∗∗∗0.0449∗∗∗0.0595∗∗∗0.0739∗∗∗0.1010∗∗∗0.1319∗∗∗0.1875∗∗∗0.3059∗∗∗ (0.0063)(0.0057)(0.0111)(0.0078)(0.0088)(0.0104)(0.0149)(0.0197)(0.0321)(0.0488)(0.0823) P3 j=0deficit/moneyt−j0.0396∗∗∗0.0775∗∗∗0.01080.0223∗∗∗0.0251∗∗∗0.0288∗∗∗0.0329∗∗∗0.0380∗∗∗0.0489∗∗∗0.0651∗∗∗0.0793∗∗∗ (0.0071)(0.0067)(0.0095)(0.0075)(0.0079)(0.0088)(0.0107)(0.0126)(0.0144)(0.0164)(0.0187) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. a DynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. bDynamicGMMofArellanoandBond(1991)withinstrumentsπit−2,∆witand∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2,∆witand∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Figure 4: Dynamic panel results over 1960–2006 III
The estimates of P3
j=0deficit/moneyt−j with instruments ∆πit−1. The upper is the results of controlling the growth of real GDP per capita. The lower is the results of controlling the growth of real GDP per capita and oil price inflation. The black solid line: the coefficients of quantile regression. The black dotted line: the 95%
confidence interval of quantile regression. The grey solid line: the coefficient of D-GMM L. The grey dotted line: the coefficient of D-GMM D.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table12:Dynamicpanelresultsover1960–2006IV(A) dependentvariable:inflation quantilea D-GMMLbD-GMMDc0.10.20.30.40.50.60.70.80.9 intercept0.0278∗∗∗0.0215∗∗0.49290.50240.50810.51350.52020.52630.52990.53280.5534 (0.0105)(0.0091)(5.7715)(5.7715)(5.7715)(5.7715)(5.7716)(5.7715)(5.7715)(5.7713)(5.7714) inflationt−10.6503∗∗∗0.2026∗∗∗0.19960.2910∗∗0.3223∗∗0.3137∗∗∗0.2803∗∗0.2792∗∗0.3255∗∗∗0.4161∗∗∗0.5082∗∗∗ (0.0185)(0.0314)(0.1576)(0.1467)(0.1282)(0.1169)(0.1218)(0.1113)(0.0976)(0.0902)(0.1538) deficit/moneyt0.0569∗∗∗0.0541∗∗∗0.00010.00540.0070∗0.0110∗∗∗0.0124∗∗0.0163∗∗∗0.0146∗∗0.0215∗∗∗0.0172 (0.0035)(0.0031)(0.0052)(0.0040)(0.0038)(0.0040)(0.0050)(0.0058)(0.0072)(0.0072)(0.0110) deficit/moneyt−10.0089∗∗∗0.0336∗∗∗0.0096∗∗0.0079∗∗∗0.0089∗∗0.0070∗∗0.0085∗∗0.0077∗∗0.0089∗0.01190.0236∗∗ (0.0033)(0.0033)(0.0041)(0.0030)(0.0040)(0.0036)(0.0035)(0.0035)(0.0049)(0.0085)(0.0119) deficit/moneyt−2-0.0340∗∗∗-0.0139∗∗∗0.0056∗0.00380.00190.00480.00470.0061∗0.0084∗∗0.0114∗∗∗0.0104∗ (0.0033)(0.0031)(0.0033)(0.0030)(0.0031)(0.0036)(0.0036)(0.0032)(0.0035)(0.0030)(0.0057) deficit/moneyt−30.0068∗0.00340.00130.00420.00650.00580.00610.00770.0165∗∗∗0.0191∗∗∗0.0268∗∗∗ (0.0036)(0.0031)(0.0050)(0.0038)(0.0041)(0.0039)(0.0051)(0.0056)(0.0058)(0.0051)(0.0099) growthofrealGDPpercapita-0.3251∗∗∗-0.2220∗∗∗-0.0903∗∗-0.0942∗∗∗-0.1237∗∗∗-0.1278∗∗∗-0.1441∗∗∗-0.1797∗∗∗-0.2152∗∗∗-0.2750∗∗∗-0.4011∗∗∗ (0.0425)(0.0370)(0.0409)(0.0287)(0.0260)(0.0251)(0.0416)(0.0378)(0.0388)(0.0606)(0.0959) oilpriceinflation0.1738∗∗∗0.1429∗∗∗0.0521∗∗∗0.0401∗∗∗0.0473∗∗∗0.0608∗∗∗0.0752∗∗∗0.1008∗∗∗0.1329∗∗∗0.1879∗∗∗0.3034∗∗∗ (0.0064)(0.0058)(0.0109)(0.0082)(0.0092)(0.0101)(0.0149)(0.0190)(0.0323)(0.0493)(0.0849) openness-0.0791∗∗∗-0.01570.0042-0.0030-0.0070-0.0100-0.0118-0.0162-0.0174-0.0148-0.0188 (0.0296)(0.0257)(0.0134)(0.0102)(0.0098)(0.0097)(0.0099)(0.0106)(0.0120)(0.0132)(0.0166) P3 j=0deficit/moneyt−j0.0386∗∗∗0.0772∗∗∗0.0166∗∗0.0213∗∗∗0.0243∗∗∗0.0286∗∗∗0.0317∗∗∗0.0379∗∗∗0.0484∗∗∗0.0638∗∗∗0.0780∗∗∗ (0.0071)(0.0067)(0.0085)(0.0074)(0.0080)(0.0087)(0.0108)(0.0124)(0.0148)(0.0173)(0.0196) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. a DynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2,∆witand∆xit−j,wherej=0,...,3. cDynamicGMMofArellanoandBond(1991)withinstruments∆πit−2,∆witand∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table13:Dynamicpanelresultsover1960–2006IV(B) dependentvariable:inflation quantilea D-GMMLbD-GMMDc0.10.20.30.40.50.60.70.80.9 intercept-0.00500.00090.33120.34060.34110.34450.34980.35060.35470.35470.3600 (0.0117)(0.0103)(4.8439)(4.8439)(4.8439)(4.8439)(4.8439)(4.8439)(4.8438)(4.8436)(4.8436) inflationt−10.6251∗∗∗0.2105∗∗∗0.4662∗∗∗0.4164∗∗∗0.4656∗∗∗0.4673∗∗∗0.4321∗∗∗0.4250∗∗∗0.3997∗∗∗0.4369∗∗∗0.5170∗∗∗ (0.0189)(0.0324)(0.0666)(0.0707)(0.0604)(0.0606)(0.0654)(0.0536)(0.0465)(0.0491)(0.1061) deficit/moneyt0.0595∗∗∗0.0583∗∗∗0.00030.00700.0119∗∗∗0.0151∗∗∗0.0182∗∗∗0.0203∗∗∗0.0182∗∗∗0.0209∗∗∗0.0129 (0.0037)(0.0033)(0.0048)(0.0059)(0.0042)(0.0043)(0.0048)(0.0049)(0.0045)(0.0058)(0.0107) deficit/moneyt−10.0096∗∗∗0.0361∗∗∗0.00110.0074∗0.0083∗∗0.0080∗∗0.00550.00490.00470.00380.0204∗ (0.0035)(0.0036)(0.0044)(0.0039)(0.0039)(0.0038)(0.0044)(0.0046)(0.0058)(0.0084)(0.0110) deficit/moneyt−2-0.0376∗∗∗-0.0166∗∗∗0.0105∗∗0.0067∗∗0.00480.00640.00490.00560.0117∗∗∗0.0116∗∗∗0.0096∗∗ (0.0035)(0.0034)(0.0044)(0.0033)(0.0036)(0.0040)(0.0041)(0.0037)(0.0034)(0.0038)(0.0045) deficit/moneyt−30.0083∗∗0.00490.00500.00170.00550.00250.00690.00730.00900.0176∗∗0.0264∗∗ (0.0038)(0.0033)(0.0048)(0.0047)(0.0042)(0.0048)(0.0056)(0.0058)(0.0063)(0.0076)(0.0111) growthofrealGDPpercapita-0.3366∗∗∗-0.2412∗∗∗-0.0419-0.0339-0.1078∗∗∗-0.1147∗∗∗-0.1311∗∗∗-0.1326∗∗∗-0.1630∗∗∗-0.1896∗∗∗-0.2798∗∗∗ (0.0487)(0.0429)(0.0358)(0.0331)(0.0328)(0.0293)(0.0316)(0.0335)(0.0431)(0.0581)(0.1076) oilpriceinflation0.1738∗∗∗0.1434∗∗∗0.0474∗∗∗0.0416∗∗∗0.0511∗∗∗0.0596∗∗∗0.0739∗∗∗0.0894∗∗∗0.1137∗∗∗0.1585∗∗∗0.2452∗∗∗ (0.0067)(0.0062)(0.0093)(0.0098)(0.0102)(0.0117)(0.0140)(0.0172)(0.0279)(0.0492)(0.0943) openness-0.0857∗∗∗-0.0221-0.0136-0.0188-0.0205∗-0.0232∗∗-0.0276∗∗-0.0299∗∗-0.0335∗∗∗-0.0338∗∗∗-0.0382∗∗∗ (0.0320)(0.0283)(0.0136)(0.0120)(0.0115)(0.0116)(0.0116)(0.0116)(0.0117)(0.0118)(0.0140) exchangerateregime0.0157∗∗∗0.0101∗∗∗0.0050∗∗∗0.0075∗∗∗0.0088∗∗∗0.0098∗∗∗0.0116∗∗∗0.0148∗∗∗0.0171∗∗∗0.0203∗∗∗0.0248∗∗∗ (0.0021)(0.0019)(0.0018)(0.0020)(0.0021)(0.0024)(0.0027)(0.0027)(0.0026)(0.0027)(0.0040) P3 j=0deficit/moneyt−j0.0398∗∗∗0.0826∗∗∗0.0168∗∗0.0229∗∗0.0306∗∗∗0.0321∗∗∗0.0355∗∗∗0.0382∗∗∗0.0437∗∗∗0.0539∗∗∗0.0693∗∗∗ (0.0051)(0.0064)(0.0081)(0.0089)(0.0081)(0.0095)(0.0113)(0.0123)(0.0137)(0.0154)(0.0220) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. a DynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2,∆witand∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2,∆witand∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Figure 5: Dynamic panel results over 1960–2006 IV
The estimates of P3
j=0deficit/moneyt−j with instruments ∆πit−1. The upper is the results of controlling the growth of real GDP per capita, oil price inflation and openness. The lower is the results of controlling the growth of real GDP per capita, oil price inflation, openness and the exchange rate regime. The black solid line: the coefficients of quantile regression. The black dotted line: the 95% confidence interval of quantile regression. The grey solid line: the coefficient of D-GMM L. The grey dotted line: the coefficient of D-GMM D.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table14:Dynamicpanelresultsofhigh-incomecountriesover1960–2006 dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.0098∗∗∗ 0.0163∗∗∗ 0.08640.08920.08960.09060.08960.08730.08980.09280.1106 (0.0012)(0.0013)(0.6272)(0.6273)(0.6274)(0.6273)(0.6275)(0.6273)(0.6273)(0.6274)(0.6276) inflationt−10.3611∗∗∗ 0.01130.14310.13810.20890.28030.4194∗ 0.6381∗∗ 0.7149∗∗∗ 0.8199∗∗∗ 0.6461∗∗ (0.0334)(0.0427)(0.1555)(0.1712)(0.1779)(0.1990)(0.2464)(0.2622)(0.2527)(0.2565)(0.2790) deficit/moneyt0.0166∗∗∗0.0073∗∗0.00040.00690.0109∗∗0.0123∗0.0170∗0.0253∗∗0.0286∗∗∗0.0296∗∗∗0.0325∗∗∗ (0.0030)(0.0029)(0.0082)(0.0059)(0.0056)(0.0072)(0.0089)(0.0103)(0.0100)(0.0091)(0.0100) deficit/moneyt−10.0347∗∗∗ 0.0375∗∗∗ 0.00570.0083∗∗ 0.0075∗∗ 0.0072∗ 0.00560.00870.01180.01120.0172 (0.0030)(0.0027)(0.0059)(0.0037)(0.0035)(0.0044)(0.0063)(0.0083)(0.0098)(0.0120)(0.0139) deficit/moneyt−2-0.00280.0104∗∗∗ -0.0009-0.00030.0003-0.0032-0.0014-0.0026-0.00110.00860.0068 (0.0033)(0.0031)(0.0060)(0.0055)(0.0043)(0.0047)(0.0053)(0.0043)(0.0056)(0.0069)(0.0094) deficit/moneyt−3-0.0092∗∗∗ 0.00080.00420.00380.00450.00610.00600.00340.00490.00870.0148 (0.0031)(0.0029)(0.0075)(0.0067)(0.0072)(0.0073)(0.0077)(0.0087)(0.0115)(0.0147)(0.0194) P3 j=0deficit/moneyt−j0.0392∗∗∗ 0.0559∗∗∗ 0.00940.0188∗ 0.0232∗ 0.02240.02720.03490.0441∗ 0.0581∗∗ 0.0712∗∗ (0.0046)(0.0044)(0.0091)(0.0108)(0.0127)(0.0158)(0.0191)(0.0219)(0.0237)(0.0257)(0.0321) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. aDynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table15:Dynamicpanelresultsofmiddle-andlow-incomecountriesover1960–2006 dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.0089∗∗ 0.0276∗∗∗ 0.60300.61280.61800.62250.63040.63510.64250.64700.6593 (0.0035)(0.0033)(9.2791)(9.2790)(9.2789)(9.2789)(9.2791)(9.2791)(9.2793)(9.2793)(9.2794) inflationt−10.7366∗∗∗ 0.2332∗∗∗ 0.19760.2946∗ 0.3354∗∗ 0.3784∗∗ 0.3308∗∗ 0.3521∗∗∗ 0.3801∗∗∗ 0.4809∗∗∗ 0.6304∗∗∗ (0.0253)(0.0445)(0.2305)(0.1780)(0.1556)(0.1481)(0.1357)(0.1254)(0.1346)(0.1572)(0.2077) deficit/moneyt0.0767∗∗∗0.0752∗∗∗0.00220.00340.00840.00910.00950.01090.01120.02080.0255 (0.0054)(0.0046)(0.0090)(0.0066)(0.0068)(0.0065)(0.0068)(0.0076)(0.0101)(0.0158)(0.0397) deficit/moneyt−10.00420.0386∗∗∗ 0.00990.0105∗∗ 0.0090∗ 0.00670.0097∗∗ 0.00690.00980.01930.0598∗∗∗ (0.0052)(0.0052)(0.0082)(0.0042)(0.0053)(0.0047)(0.0038)(0.0062)(0.0100)(0.0173)(0.0165) deficit/moneyt−2-0.0495∗∗∗ -0.0223∗∗∗ 0.0126∗∗ 0.0065∗ 0.00690.0075∗ 0.0078∗ 0.00770.0083∗ 0.0155∗∗∗ 0.0205∗∗ (0.0051)(0.0048)(0.0051)(0.0039)(0.0047)(0.0043)(0.0042)(0.0049)(0.0050)(0.0054)(0.0086) deficit/moneyt−30.00290.0002-0.00380.00300.00580.00570.00580.0114∗ 0.0141∗ 0.0177∗ 0.0305∗ (0.0055)(0.0047)(0.0072)(0.0048)(0.0045)(0.0047)(0.0058)(0.0067)(0.0073)(0.0092)(0.0177) P3 j=0deficit/moneyt−j0.0343∗∗∗ 0.0917∗∗∗ 0.02100.0235∗∗ 0.0301∗∗ 0.0291∗∗ 0.0328∗∗ 0.0369∗∗ 0.0434∗ 0.0733∗∗ 0.1362∗∗ (0.0120)(0.0113)(0.0151)(0.0119)(0.0127)(0.0130)(0.0146)(0.0175)(0.0236)(0.0363)(0.0568) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. aDynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
High-income countries
Middle- and low-income countries
Figure 6: Dynamic panel results of high-income and middle- and low-income coun-tries over 1960–2006
The estimates of P3
j=0deficit/moneyt−j with instruments ∆πit−1. The upper is the results of high-income countries. The lower is the results of middle- and low-income countries countries. The black solid line: the coefficients of quantile regression. The black dotted line: the 95% confidence interval of quantile regression. The grey solid line: the coefficient of D-GMM L. The grey dotted line: the coefficient of D-GMM D.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table16:DynamicpanelresultsofOECDcountriesover1960–2006 dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.0136∗∗∗ 0.0336∗∗∗ 0.06670.06340.06230.06400.05930.05710.06310.07450.0889 (0.0013)(0.0018)(0.5340)(0.5338)(0.5340)(0.5341)(0.5340)(0.5340)(0.5339)(0.5340)(0.5345) inflationt−10.5189∗∗∗ -0.2104∗∗∗ 0.09830.33750.4828∗∗∗ 0.5311∗∗ 0.8181∗∗∗ 1.0395∗∗∗ 0.9909∗∗∗ 0.8751∗∗∗ 0.7365∗∗∗ (0.0368)(0.0626)(0.2632)(0.2201)(0.1618)(0.2096)(0.2575)(0.2117)(0.1628)(0.1820)(0.2154) deficit/moneyt0.00620.0070∗0.0259∗∗∗0.0262∗∗∗0.0232∗∗∗0.0242∗∗∗0.0337∗∗∗0.0379∗∗∗0.0397∗∗∗0.0400∗∗∗0.0477∗∗ (0.0046)(0.0038)(0.0059)(0.0045)(0.0066)(0.0093)(0.0106)(0.0108)(0.0112)(0.0140)(0.0214) deficit/moneyt−1-0.0062-0.0027-0.0066-0.00870.00190.00160.00440.0025-0.00170.00280.0034 (0.0048)(0.0038)(0.0086)(0.0085)(0.0059)(0.0058)(0.0079)(0.0097)(0.0112)(0.0115)(0.0143) deficit/moneyt−2-0.0029-0.00500.00380.00830.00650.00530.00500.00860.01170.0123-0.0029 (0.0048)(0.0039)(0.0072)(0.0070)(0.0048)(0.0051)(0.0067)(0.0080)(0.0086)(0.0087)(0.0120) deficit/moneyt−3-0.0023-0.0064∗ -0.0090∗ -0.0041-0.0101-0.0086-0.0142∗ -0.0125-0.0085-0.0173-0.0046 (0.0047)(0.0039)(0.0051)(0.0049)(0.0064)(0.0061)(0.0079)(0.0101)(0.0127)(0.0130)(0.0158) P3 j=0deficit/moneyt−j-0.0052-0.00710.0142∗∗ 0.0217∗∗∗ 0.0215∗∗∗ 0.0224∗∗ 0.0288∗∗ 0.0365∗∗∗ 0.0412∗∗∗ 0.0378∗∗∗ 0.0436∗ (0.0051)(0.0045)(0.0070)(0.0054)(0.0069)(0.0091)(0.0112)(0.0136)(0.0126)(0.0138)(0.0241) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. aDynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table17:Dynamicpanelresultsofnon-OECDcountriesover1960–2006 dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.00430.0245∗∗∗ 0.22810.23620.24220.24570.25100.25500.26160.26800.2881 (0.0030)(0.0028)(3.2408)(3.2410)(3.2407)(3.2407)(3.2407)(3.2406)(3.2406)(3.2405)(3.2403) inflationt−10.7265∗∗∗ 0.1971∗∗∗ 0.17020.2792∗ 0.3065∗∗ 0.3411∗∗ 0.3279∗∗∗ 0.3449∗∗∗ 0.3778∗∗∗ 0.4543∗∗∗ 0.5580∗∗∗ (0.0236)(0.0400)(0.1857)(0.1539)(0.1384)(0.1347)(0.1246)(0.1165)(0.1210)(0.1403)(0.1660) deficit/moneyt0.0715∗∗∗0.0659∗∗∗-0.00220.00210.00560.00870.00890.01150.01260.0233∗∗0.0234 (0.0046)(0.0040)(0.0083)(0.0056)(0.0055)(0.0054)(0.0062)(0.0076)(0.0097)(0.0107)(0.0252) deficit/moneyt−10.0119∗∗∗ 0.0412∗∗∗ 0.01010.0102∗∗∗ 0.0082∗ 0.00670.0109∗∗∗ 0.00890.01200.01970.0490∗∗∗ (0.0044)(0.0043)(0.0070)(0.0036)(0.0043)(0.0042)(0.0042)(0.0056)(0.0073)(0.0135)(0.0128) deficit/moneyt−2-0.0408∗∗∗ -0.0149∗∗∗ 0.0102∗∗ 0.00470.00680.0072∗ 0.00610.00530.0078∗ 0.0148∗∗∗ 0.0169∗∗ (0.0043)(0.0041)(0.0049)(0.0036)(0.0042)(0.0042)(0.0040)(0.0040)(0.0043)(0.0049)(0.0075) deficit/moneyt−30.00500.00240.00000.00760.0080∗ 0.00760.00800.0130∗∗ 0.0157∗∗ 0.0170∗∗ 0.0255∗∗ (0.0047)(0.0040)(0.0073)(0.0051)(0.0046)(0.0048)(0.0053)(0.0059)(0.0063)(0.0074)(0.0104) P3 j=0deficit/moneyt−j0.0475∗∗∗ 0.0946∗∗∗ 0.01810.0246∗∗ 0.0286∗∗ 0.0302∗∗ 0.0339∗∗ 0.0386∗∗ 0.0480∗∗ 0.0748∗∗∗ 0.1148∗∗∗ (0.0096)(0.0089)(0.0142)(0.0114)(0.0126)(0.0137)(0.0153)(0.0176)(0.0219)(0.0279)(0.0358) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. aDynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
OECD countries
non-OECD countries
Figure 7: Dynamic panel results of OECD and non-OECD countries over 1960–2006
The estimates of P3
j=0deficit/moneyt−j with instruments ∆πit−1. The upper is the results of OECD countries. The lower is the results of non-OECD countries. The black solid line: the coefficients of quantile regression. The black dotted line: the 95% confidence interval of quantile regression. The grey solid line: the coefficient of D-GMM L. The grey dotted line: the coefficient of D-GMM D.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table18:Dynamicpanelresultsover1970–2006 dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.0044∗∗ 0.0244∗∗∗ 0.48780.49780.50090.50360.50670.51080.51460.52400.5373 (0.0022)(0.0024)(4.2169)(4.2171)(4.2173)(4.2173)(4.2173)(4.2174)(4.2173)(4.2172)(4.2174) inflationt−10.7225∗∗∗ 0.2693∗∗∗ 0.3303∗ 0.3403∗∗ 0.3597∗∗∗ 0.3717∗∗∗ 0.3799∗∗∗ 0.3920∗∗∗ 0.4400∗∗∗ 0.4448∗∗∗ 0.5488∗∗∗ (0.0211)(0.0378)(0.1922)(0.1599)(0.1375)(0.1292)(0.1179)(0.1226)(0.1364)(0.1292)(0.1439) deficit/moneyt0.0560∗∗∗0.0495∗∗∗-0.00120.00210.00650.0091∗∗0.0096∗0.01110.0167∗0.0228∗∗0.0216 (0.0037)(0.0032)(0.0058)(0.0045)(0.0041)(0.0044)(0.0052)(0.0068)(0.0086)(0.0090)(0.0161) deficit/moneyt−10.0082∗∗ 0.0262∗∗∗ 0.00310.00470.00400.0050∗ 0.0079∗∗ 0.0111∗∗ 0.00850.01280.0457∗∗∗ (0.0035)(0.0033)(0.0046)(0.0028)(0.0025)(0.0027)(0.0032)(0.0045)(0.0052)(0.0113)(0.0160) deficit/moneyt−2-0.0292∗∗∗ -0.0140∗∗∗ 0.00050.00310.00430.0056∗ 0.00330.00180.00630.0148∗∗∗ 0.0151∗∗ (0.0034)(0.0031)(0.0040)(0.0028)(0.0029)(0.0032)(0.0031)(0.0041)(0.0046)(0.0043)(0.0069) deficit/moneyt−30.0103∗∗∗ 0.0087∗∗∗ 0.00150.00200.00300.00260.00570.00750.0125∗∗ 0.0130∗ 0.0248∗∗ (0.0038)(0.0032)(0.0061)(0.0043)(0.0034)(0.0042)(0.0044)(0.0052)(0.0055)(0.0067)(0.0096) P3 j=0deficit/moneyt−j0.0453∗∗∗ 0.0704∗∗∗ 0.00390.01190.0178∗ 0.0222∗∗ 0.0266∗∗ 0.0315∗∗ 0.0440∗∗ 0.0634∗∗∗ 0.1072∗∗∗ (0.0075)(0.0067)(0.0101)(0.0092)(0.0092)(0.0102)(0.0113)(0.0144)(0.0177)(0.0215)(0.0280) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. aDynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table19:Dynamicpanelresultsover1980–2006 dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.0069∗∗ 0.0266∗∗∗ 1.16701.17901.18231.18281.18611.18921.19221.19771.2079 (0.0033)(0.0031)(85.9528)(85.9526)(85.9527)(85.9527)(85.9527)(85.9527)(85.9525)(85.9526)(85.9526) inflationt−10.6952∗∗∗ 0.1312∗∗∗ 0.18220.18530.20290.2421∗ 0.2238∗ 0.2187∗ 0.2445∗ 0.2975∗∗ 0.4931∗∗∗ (0.0249)(0.0441)(0.1660)(0.1442)(0.1378)(0.1323)(0.1327)(0.1320)(0.1358)(0.1518)(0.1888) deficit/moneyt0.0726∗∗∗0.0665∗∗∗-0.0035-0.0020-0.00230.00130.00430.00250.00020.00400.0105 (0.0052)(0.0044)(0.0060)(0.0043)(0.0034)(0.0034)(0.0034)(0.0042)(0.0063)(0.0089)(0.0233) deficit/moneyt−10.0129∗∗∗ 0.0457∗∗∗ 0.00090.00260.0048∗ 0.0053∗ 0.0059∗∗ 0.00600.0102∗ 0.01750.0495∗∗∗ (0.0049)(0.0048)(0.0044)(0.0034)(0.0029)(0.0028)(0.0030)(0.0040)(0.0056)(0.0111)(0.0178) deficit/moneyt−2-0.0472∗∗∗ -0.0169∗∗∗ -0.00280.00160.00260.0000-0.00100.00250.0083∗∗ 0.0103∗∗ 0.0123 (0.0048)(0.0046)(0.0045)(0.0029)(0.0024)(0.0027)(0.0036)(0.0037)(0.0036)(0.0049)(0.0082) deficit/moneyt−30.00320.00040.0041∗∗ 0.00290.00470.0073∗ 0.0089∗∗ 0.0110∗∗∗ 0.0104∗∗ 0.0130∗∗ 0.0238∗∗ (0.0051)(0.0043)(0.0063)(0.0046)(0.0041)(0.0041)(0.0041)(0.0036)(0.0051)(0.0073)(0.0138) P3 j=0deficit/moneyt−j0.0414∗∗∗ 0.0958∗∗∗ -0.00140.00510.00980.01400.0182∗ 0.0219∗∗ 0.0291∗∗ 0.0448∗ 0.0961∗∗ (0.0111)(0.0102)(0.0119)(0.0092)(0.0087)(0.0087)(0.0097)(0.0107)(0.0144)(0.0237)(0.0409) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. aDynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
1970–2006
1980–2006
Figure 8: Dynamic panel results over 1970–2006 and 1980–2006
The estimates of P3
j=0deficit/moneyt−j with instruments ∆πit−1. The upper is the results over 1970–2006. The lower is the results over 1980–2006. The black solid line:
the coefficients of quantile regression. The black dotted line: the 95% confidence interval of quantile regression. The grey solid line: the coefficient of D-GMM L. The grey dotted line: the coefficient of D-GMM D.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table20:Dynamicpanelresultsover1990–2006 dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.0117∗∗∗ 0.0286∗∗∗ 0.26690.27030.27730.28280.28510.28950.29370.30080.3076 (0.0035)(0.0030)(0.4808)(0.4807)(0.4807)(0.4807)(0.4807)(0.4807)(0.4808)(0.4807)(0.4806) inflationt−10.4818∗∗∗ 0.01070.39550.47190.36390.28330.26560.18690.14320.11180.2473 (0.0381)(0.0421)(0.3536)(0.3114)(0.2930)(0.2678)(0.2419)(0.2140)(0.1720)(0.1706)(0.2082) deficit/moneyt0.0008-0.0004-0.0092∗-0.0107∗∗-0.0073∗∗-0.0084∗∗∗-0.0072∗∗-0.0066∗-0.0032-0.0048-0.0031 (0.0066)(0.0055)(0.0051)(0.0042)(0.0035)(0.0032)(0.0030)(0.0035)(0.0039)(0.0061)(0.0109) deficit/moneyt−10.0194∗∗∗ 0.0195∗∗∗ -0.0050-0.0048-0.0038-0.0012-0.0010-0.0018-0.0032-0.00180.0037 (0.0057)(0.0047)(0.0052)(0.0040)(0.0033)(0.0027)(0.0026)(0.0025)(0.0032)(0.0049)(0.0079) deficit/moneyt−20.0092∗ 0.0138∗∗∗ -0.0058-0.0011-0.0013-0.0007-0.0021-0.0009-0.00010.00200.0080 (0.0054)(0.0044)(0.0052)(0.0034)(0.0029)(0.0028)(0.0030)(0.0032)(0.0035)(0.0054)(0.0099) deficit/moneyt−3-0.00420.0001-0.0034-0.0032-0.0015-0.0016-0.00020.00220.00270.00300.0104 (0.0049)(0.0041)(0.0064)(0.0041)(0.0030)(0.0029)(0.0035)(0.0042)(0.0052)(0.0072)(0.0074) P3 j=0deficit/moneyt−j0.0253∗∗ 0.0330∗∗∗ -0.0234∗∗ -0.0198∗∗ -0.0139∗∗ -0.0118∗ -0.0106∗ -0.0071-0.0037-0.00160.0190 (0.0126)(0.0103)(0.0107)(0.0084)(0.0070)(0.0064)(0.0062)(0.0068)(0.0071)(0.0087)(0.0124) Standarderrorsinparenthesis.∗∗∗ ,∗∗ and∗ indicatethesignificantlevelof1%,5%and10%repectively. aDynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table21:Dynamicpanelresultsover1990–2000 dependentvariable:inflation quantilea D-GMMLb D-GMMDc 0.10.20.30.40.50.60.70.80.9 intercept0.0663∗∗ 0.0977∗∗∗ 0.36190.37940.39900.39920.39880.39870.41360.41380.4362 (0.0267)(0.0260)(1.3529)(1.3526)(1.3527)(1.3526)(1.3527)(1.3528)(1.3528)(1.3529)(1.3525) inflationt−10.7361∗∗∗-0.4563∗∗1.13830.92320.67730.71500.75670.78630.53830.59620.5617 (0.0500)(0.1832)(1.5171)(1.1562)(0.9095)(0.7730)(0.7261)(0.6910)(0.6725)(0.6799)(0.7880) deficit/moneyt0.0267-0.0210-0.0147-0.0117-0.0140∗ -0.0096-0.0082-0.0072-0.0073-0.0043-0.0097 (0.0241)(0.0233)(0.0108)(0.0085)(0.0080)(0.0070)(0.0067)(0.0073)(0.0090)(0.0133)(0.0310) deficit/moneyt−10.03100.0288-0.0206∗ -0.0205∗ 0.00540.00850.0113∗∗ 0.0127∗∗ 0.01170.02020.0518∗ (0.0216)(0.0204)(0.0114)(0.0115)(0.0076)(0.0057)(0.0053)(0.0056)(0.0071)(0.0123)(0.0287) deficit/moneyt−2-0.1141∗∗∗ -0.0466∗∗ -0.00520.0061-0.0004-0.0005-0.0039-0.00210.0016-0.00520.0041 (0.0212)(0.0199)(0.0113)(0.0083)(0.0060)(0.0053)(0.0058)(0.0068)(0.0078)(0.0117)(0.0230) deficit/moneyt−30.0017-0.01860.0194∗ 0.00910.0031-0.0007-0.0003-0.0039-0.0037-0.0032-0.0269 (0.0121)(0.0115)(0.0116)(0.0113)(0.0098)(0.0097)(0.0101)(0.0103)(0.0109)(0.0143)(0.0244) centralbankindependence-0.1009∗∗ -0.03340.00250.0025-0.0037-0.0047-0.0057-0.0062-0.0079-0.0071-0.0251 (0.0513)(0.0490)(0.0154)(0.0071)(0.0061)(0.0060)(0.0060)(0.0061)(0.0067)(0.0089)(0.0247) P3 j=0deficit/moneyt−j-0.0547-0.0574∗ -0.0210-0.0169-0.0059-0.0023-0.0010-0.00060.00230.00750.0193 (0.0355)(0.0332)(0.0157)(0.0176)(0.0176)(0.0173)(0.0172)(0.0175)(0.0181)(0.0187)(0.0239) Standarderrorsinparenthesis.∗∗∗,∗∗and∗indicatethesignificantlevelof1%,5%and10%repectively. a DynamicpanelquantileregressionofLin(2010)withinstruments∆πit−1. b DynamicGMMofArellanoandBond(1991)withinstrumentsπit−2and∆xit−j,wherej=0,...,3. c DynamicGMMofArellanoandBond(1991)withinstruments∆πit−2and∆xit−j,wherej=0,...,3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
1990–2006
1990–2000
Figure 9: Dynamic panel results over 1990–2006 and 1990–2000
The estimates of P3
j=0deficit/moneyt−j with instruments ∆πit−1. The upper is the results over 1990–2006. The lower is the results of controlling central bank independence over 1990–2000. The black solid line: the coefficients of quantile regression. The black dotted line: the 95% confidence interval of quantile regression.
The grey solid line: the coefficient of D-GMM L. The grey dotted line: the coefficient of D-GMM D.
‧
Ahking, F. W. and Miller, S. M. (1985). The relationship between government deficits, money growth and inflation. Journal of Macroeconomics, 7 ,447–467.
Aisen, A. and Veiga, F. J. (2008). The political economy of seigniorage. Journal of Development Economics, 87, 29–50.
Alesina, A. and Summers, L. H. (1993). Central bank independence and macroeconomic perfor-mance: some comparative evidence. Journal of Money, Credit and Banking, 25, 151–162.
Alfaro, L. (2005). Inflation, openness, and exchange-rate regimes: the quest for short-term com-mitment. Journal of Development Economics, 77, 229–249.
Amemiya, T. (1982). Two stage least absolute deviations estimators. Econometrica, 50, 689–711.
Anderson, T. W. and Hsiao, C. (1981). Estimation of dynamic models with error components.
Anderson, T. W. and Hsiao, C. (1981). Estimation of dynamic models with error components.