"... In PAGE 13: ... Overall, as the impulse responses line up well with standard theory, the structural shocks appear to be properly identi ed. Table4 reports variance decompositions for each country. In ation shocks themselves account for most of the variability in in ation for each country.... In PAGE 21: ...Table4 : Variance decompositions Percent of in ation variance Percent of interest rate variance Percent of real output variance attributable to: attributable to: attributable to: In ation Saving Technology In ation Saving Technology In ation Saving Technology Lag shock shock shock shock shock shock shock shock shock Australia 1 63 (17) 0 (6) 37 (16) 0 (8) 97 (10) 3 (7) 26 (17) 10 (12) 64 (18) 2 70 (13) 3 (7) 27 (12) 4 (11) 94 (12) 2 (4) 32 (18) 5 (8) 63 (17) 3 80 (9) 2 (5) 18 (8) 8 (13) 91 (13) 1 (2) 29 (17) 4 (7) 67 (16) 4 83 (8) 2 (4) 15 (7) 9 (14) 90 (14) 1 (2) 27 (17) 3 (7) 70 (16) 8 89 (6) 1 (3) 9 (5) 10 (15) 89 (15) 0 (1) 26 (17) 2 (7) 72 (17) Austria 1 71 (19) 0 (7) 29 (19) 9 (14) 78 (17) 13 (12) 50 (19) 7 (12) 43 (19) 2 84 (11) 3 (5) 13 (8) 18 (17) 77 (17) 5 (5) 46 (19) 3 (7) 51 (18) 3 89 (7) 2 (3) 9 (6) 23 (18) 74 (18) 3 (3) 43 (20) 2 (6) 55 (18) 4 92 (6) 1 (3) 7 (5) 25 (18) 72 (19) 2 (3) 41 (20) 2 (7) 56 (18) 8 96 (3) 1 (2) 4 (3) 28 (19) 71 (19) 1 (1) 39 (20) 2 (7) 59 (19) Belgium 1 62 (17) 0 (7) 38 (16) 2 (11) 98 (12) 1 (5) 28 (17) 16 (16) 56 (18) 2 80 (11) 3 (5) 17 (8) 1 (11) 98 (11) 0 (2) 28 (16) 8 (11) 64 (17) 3 88 (7) 2 (4) 11 (5) 1 (11) 99 (12) 0 (1) 28 (17) 5 (9) 66 (17) 4 91 (5) 2 (3) 8 (4) 1 (12) 99 (12) 0 (1) 29 (17) 4 (8) 68 (17) 8 96 (3) 1 (1) 4 (2) 1 (12) 99 (12) 0 (1) 29 (18) 2 (8) 69 (18) Canada 1 85 (13) 3 (7) 12 (10) 0 (6) 97 (8) 3 (6) 11 (12) 11 (12) 79 (15) 2 90 (9) 4 (5) 7 (5) 4 (9) 94 (9) 2 (3) 19 (14) 5 (6) 76 (14) 3 93 (6) 2 (3) 4 (3) 6 (10) 93 (11) 1 (2) 18 (14) 5 (6) 77 (14) 4 95 (4) 2 (3) 3 (3) 7 (11) 92 (11) 1 (1) 17 (14) 6 (7) 77 (14) 8 97 (2) 1 (1) 2 (1) 9 (12) 91 (12) 0 (1) 15 (14) 6 (9) 78 (15) Denmark 1 87 (10) 0 (4) 13 (9) 10 (11) 90 (11) 0 (3) 0 (7) 1 (6) 99 (9) 2 92 (6) 1 (3) 7 (5) 13 (12) 87 (12) 0 (2) 3 (9) 11 (11) 86 (13) 3 94 (5) 1 (2) 5 (4) 13 (13) 86 (13) 0 (1) 4 (10) 14 (12) 83 (14) 4 95 (4) 1 (2) 4 (3) 14 (13) 86 (13) 0 (1) 4 (10) 15 (13) 81 (15) 8 98 (2) 0 (1) 2 (2) 14 (13) 86 (13) 0 (0) 4 (11) 16 (14) 79 (16) France 1 60 (19) 5 (10) 35 (19) 9 (14) 89 (15) 2 (6) 27 (18) 2 (9) 71 (19) 2 77 (11) 4 (5) 19 (10) 12 (15) 86 (16) 1 (4) 28 (18) 1 (7) 71 (18) 3 85 (8) 3 (4) 12 (7) 14 (16) 85 (16) 1 (2) 27 (18) 1 (7) 72 (18) 4 89 (6) 2 (3) 9 (5) 16 (16) 84 (17) 0 (2) 27 (19) 1 (8) 72 (18) 8 95 (3) 1 (1) 4 (3) 18 (17) 82 (17) 0 (1) 25 (19) 2 (10) 73 (19) Ireland 1 61 (18) 35 (18) 4 (12) 11 (19) 72 (24) 17 (20) 16 (16) 10 (18) 74 (23) 2 66 (15) 26 (13) 8 (11) 25 (21) 65 (23) 10 (14) 12 (16) 3 (12) 85 (19) 3 70 (13) 23 (9) 7 (9) 23 (21) 70 (22) 7 (10) 11 (16) 4 (11) 85 (18) 4 77 (11) 18 (8) 5 (7) 25 (21) 70 (22) 5 (8) 11 (17) 4 (12) 86 (18) 8 87 (7) 10 (6) 3 (4) 29 (22) 68 (22) 3 (5) 11 (18) 4 (13) 85 (19)... ..."

"... In PAGE 13: ... Overall, as the impulse responses line up well with standard theory, the structural shocks appear to be properly identi ed. Table4 reports variance decompositions for each country. In ation shocks themselves account for most of the variability in in ation for each country.... In PAGE 22: ...Table4 : Variance decompositions (continued) Percent of in ation variance Percent of interest rate variance Percent of real output variance attributable to: attributable to: attributable to: In ation Saving Technology In ation Saving Technology In ation Saving Technology Lag shock shock shock shock shock shock shock shock shock Italy 1 84 (15) 3 (10) 14 (12) 7 (14) 90 (15) 3 (6) 17 (20) 7 (12) 76 (18) 2 89 (11) 4 (7) 8 (7) 15 (17) 84 (17) 2 (3) 8 (14) 4 (9) 88 (15) 3 91 (9) 3 (6) 6 (6) 17 (18) 82 (18) 1 (1) 6 (11) 6 (11) 88 (16) 4 93 (7) 2 (5) 5 (5) 17 (18) 82 (18) 1 (1) 4 (10) 7 (12) 88 (16) 8 96 (5) 1 (3) 3 (3) 18 (18) 82 (18) 0 (1) 2 (9) 8 (15) 90 (17) Japan 1 50 (20) 0 (8) 50 (20) 0 (6) 87 (13) 13 (12) 5 (11) 11 (14) 84 (17) 2 67 (15) 1 (6) 32 (13) 3 (8) 90 (11) 7 (6) 4 (11) 8 (12) 88 (15) 3 75 (12) 1 (4) 24 (11) 4 (9) 92 (10) 4 (4) 4 (11) 6 (12) 90 (15) 4 79 (11) 1 (4) 20 (10) 4 (10) 93 (10) 3 (3) 4 (11) 6 (12) 91 (15) 8 88 (7) 0 (2) 11 (6) 5 (11) 94 (11) 2 (1) 3 (11) 5 (12) 92 (15) Netherlands 1 67 (20) 11 (14) 22 (17) 9 (14) 82 (17) 9 (13) 34 (20) 20 (18) 47 (21) 2 74 (14) 11 (9) 15 (11) 8 (13) 88 (15) 4 (7) 35 (21) 9 (13) 55 (20) 3 82 (11) 7 (7) 11 (8) 10 (14) 87 (15) 3 (5) 35 (21) 5 (11) 59 (20) 4 86 (8) 6 (5) 8 (6) 11 (15) 87 (15) 2 (3) 35 (21) 4 (11) 61 (20) 8 93 (5) 3 (3) 4 (3) 13 (16) 86 (16) 1 (2) 35 (21) 2 (10) 64 (20) New Zealand 1 91 (12) 3 (7) 6 (9) 17 (15) 64 (18) 19 (15) 1 (10) 3 (11) 95 (14) 2 94 (7) 2 (4) 4 (5) 28 (16) 63 (17) 9 (7) 1 (10) 3 (10) 96 (14) 3 96 (5) 1 (3) 3 (4) 31 (16) 63 (17) 6 (4) 1 (11) 2 (10) 96 (14) 4 97 (4) 1 (2) 2 (3) 33 (17) 63 (17) 4 (3) 1 (11) 2 (10) 97 (14) 8 99 (2) 0 (1) 1 (1) 35 (18) 63 (18) 2 (1) 1 (11) 2 (11) 97 (15) Sweden 1 70 (21) 0 (7) 30 (20) 4 (12) 96 (13) 0 (7) 22 (19) 0 (9) 78 (20) 2 82 (13) 3 (6) 14 (11) 17 (16) 83 (16) 0 (5) 23 (19) 1 (9) 76 (19) 3 87 (9) 2 (4) 10 (7) 18 (16) 81 (17) 0 (4) 17 (18) 3 (10) 80 (18) 4 89 (8) 2 (4) 9 (6) 18 (16) 82 (17) 0 (3) 14 (17) 4 (11) 82 (17) 8 93 (5) 1 (3) 5 (4) 17 (17) 83 (17) 0 (2) 10 (17) 5 (12) 84 (18) United Kingdom 1 36 (21) 23 (18) 40 (21) 0 (10) 70 (19) 30 (18) 47 (20) 5 (11) 48 (21) 2 43 (18) 18 (7) 39 (16) 18 (17) 64 (18) 18 (12) 37 (19) 2 (7) 61 (19) 3 64 (11) 14 (6) 22 (10) 29 (18) 60 (18) 10 (8) 31 (18) 3 (8) 66 (18) 4 72 (9) 10 (4) 18 (7) 35 (18) 57 (18) 8 (6) 28 (18) 3 (9) 68 (17) 8 83 (7) 7 (3) 11 (5) 41 (19) 55 (18) 5 (4) 27 (18) 3 (9) 70 (18) United States 1 89 (12) 0 (5) 11 (11) 13 (14) 86 (14) 1 (6) 6 (12) 2 (8) 92 (14) 2 89 (10) 2 (4) 9 (8) 16 (15) 84 (15) 1 (3) 2 (8) 3 (7) 94 (11) 3 92 (8) 2 (3) 6 (6) 16 (15) 84 (15) 0 (2) 1 (8) 7 (9) 92 (11) 4 93 (6) 1 (3) 5 (5) 16 (15) 84 (15) 0 (2) 1 (8) 8 (11) 91 (12) 8 96 (4) 1 (2) 3 (3) 16 (15) 84 (15) 0 (1) 1 (8) 10 (13) 89 (14)... ..."

"... In PAGE 6: ... Figure 4 also shows the first and last terrain maps in their aligned positions. Table2 shows the quantitative comparison of ground truth vs. estimated motion with the quantities defined in Table 1.... ..."

"... In PAGE 26: ...11): the aggregate sub-regional unobserved group effects are correlated with the training status. In column (2) of Table5 we report the estimates of the constant treatment effect model. In contrast, with the variable treatment model, the point estimate of the mean selection bias is large and negative.... In PAGE 26: ...post participation. In Appendix D.1 we derive how this affects our regression specification. The estimation results can be found in column (3) of Table5... In PAGE 27: ... We derive in Appendix D.2 how this affects the specification of the variable treatment effects. Column (4) of Table5 reports the findings of this model. The point estimates are very different.... ..."

"... In PAGE 16: ... (This may be compared with the analysis reported in Clark et al. (1996, Table2 ). They reported a larger log relative risk, namely log(1:14) = 0:131, but this is less signi cant (logrank test P = 0.... ..."

"... In PAGE 15: ...uch an approach is 0.956218, very close to unity. We should mention here that although the ECM with GARCH (1,0) errors performs statistically better than the simple error correction representation, it did not manage to increase hedging effectiveness, as measured by using the adjusted R2, and hence the simple error correction representation is considered superior. Table5 summarizes the comparisons of the optimal hedge ratios estimated using alternative methods. In terms of risk reduction, the appropriate method for estimating optimal hedge ratios is the ECM.... In PAGE 15: ... 5.4 In-sample analysis Table5 reports also the RMSEs, MAEs and MAPEs for each model. The results indicate that the error-correction specification outperforms all the other models since ... ..."

"... In PAGE 15: ...14 How does this degree of mobility manifest itself in terms of mobility measures based on correlations of functions of the household expenditure variable? This is seen in Table3 . As long as incomes are not negatively correlated over time, correlations will lie between 0 (complete mobility in the sense that incomes in period 1 and period 2 are uncorrelated) and 1 (no mobility).... In PAGE 15: ... Thus all mobility measures based on correlation of functions of the income variable will lie between 1 (complete mobility) and 0 (no mobility). The mobility measures in Table3 range from 0.278 and 0.... In PAGE 15: ...obility). The mobility measures in Table 3 range from 0.278 and 0.395, which in general indicates substantial mobility although it is farther from complete mobility than from complete immobility. The main point of Table3 is to show how the mobility seen in the transition matrices of Table 2 is measured by these mobility indices. With one exception, the different mobility measures give similar results.... In PAGE 15: ... This is seen in Table 4 for the simple correlation coefficient. The first line shows the correlation coefficients for per capita expenditures in the two years for both samples of households, which is simply one minus the associated mobility index given in Table3 . The second line shows (the OLS ... In PAGE 16: ... The first row of Table 5 shows estimates of economic mobility when per capita expenditures are instrumented using household income. As expected, the estimated mobility is much lower than the uncorrected estimates given in Table3 . The figures in brackets show the IV-corrected estimates as a percentage of the uncorrected estimates.... In PAGE 17: ... This implies that these estimates can be thought of as upper bounds of the true amount of mobility. Thus nearly half, and perhaps even more than half, of the mobility shown in Table3 is due to measurement error and is therefore spurious. Because use of income as an instrumental variable is likely to overestimate mobility, it is useful to estimate mobility is using other plausible instrumental variables.... In PAGE 17: ...118 for the 50% threshold sample. Taken at face value, these estimates suggest that almost two thirds of the observed mobility in Vietnam seen in Table3 is purely due to measurement error in the expenditure variable. Yet there are conceptual problems with durable goods as an instrumental variable.... ..."

"... In PAGE 14: ... Estimates appear robust to a variety of starting values. Table1 presents the ma ximum likelihood estimates for our model and the country pairs a) US and Canada, b) US and France, c) US and Germany, d) US and Italy, e) US and Japan, and f) US and UK, respectively. The most important thing to notice about these estimates is that, for every country pair, the adjustment of prices to a transitory shock is much faster than the adjustment of the exchange rate.... In PAGE 16: ... Of course, we should be careful about interpreting the estimates of k too literally since they are not significant. The remaining estimates in Table1 are of the long-run inflation rates in each country and the normalizing initial values for the unobserved equilibrium prices and exchange rate. It is encouraging to note that the estimates for all of the parameters associated with US prices only (i.... ..."

"... In PAGE 21: ... A desire to move motivated by employment reasons has thus a quantitatively large and positive effect on the probability of moving house. In Table11 we report the results from a bivariate probit allowing an employment status change and migration decision to be modelled simultaneously.23 Again the standard errors 20 It can be argued that, because mortgage holders typically face higher transaction costs when moving house, they would be more willing to commute longer distances.... ..."