Crisis, Austerity and Automatic Stabilization

Transcription

1 Comments on: Crisis, Austerity and Automatic Stabilization by Mathias Dolls, Clemens Fuest, Andreas Peichl and Christian Wittneben Roberto Ramos Banco de España 6 June 2016 Fiscal Sustainability, XXI Century Banco de España and Barcelona GSE The views expressed in this presentation are those of the author and do not necessarily correspond to the views of Banco de España or the Eurosystem

2 The Paper The paper estimates the automatic scal stabilization properties of national EU tax-benet systems in i.e. how much a shock to market income translates into a shock on households' disposable income. τ I = 1 Y D Y M where: Y D = Y M ( T + S B ) The results show that automatic stabilizers are heterogeneous across countries (both in levels and changes). Cross-section estimates range from less than 25% to almost 60%. Mostly increases in automatic stabilization during the nancial crisis stemming from austerity measures. 1 / 11

3 The Paper Still preliminary, but very relevant question, transparent methodology and results. Some questions/comments: Interpretation of short term eects of policy adjustments. Accuracy of EUROMOD. Other. 2 / 11

4 On the Short-Term Eects of Policy Adjustments One novelty of the paper is to construct a stabilization coecient (θ t+1 ) accounting for the short-term eects of policy adjustments. Interpretation is: τ I t built-in exibility, stabilization capacity in the long-term. θ I t additional short-run burden of new policy. I found this interpretation confusing: τt I does not take ino account either behavioral changes or general equilibrium eects (it seems rather a day-after stabilization measure, i.e. short-run). 3 / 11

5 On the Short-Term Eects of Policy Adjustments Interpretation is: τ I t long-term stabilization capacity. θ I t additional short-run burden of new policy. I found this interpretation confusing: Also: θ T t+1 = T 1 t+1 T 0 t Y M t 4 / 11

6 On the Short-Term Eects of Policy Adjustments Interpretation is: τt I long-term stabilization capacity. θt I additional short-run burden of new policy. I found this interpretation confusing: Also: θ T t+1 = T 1 t+1 T 0 t Y M t = T 1 t T 0 t Y M t + T 1 t+1 T 1 t Y M t = τt T + T t+1 1 T t 1 Yt M }{{} additional burden? 4 / 11

7 On the Short-Term Eects of Policy Adjustments More on interpretation: If Tt+1 1 = Tt 0 θt+1 T = 0 government did not let automatic stabilizers work. If Tt+1 1 = Tt 1 θt+1 T = τt T automatic stabilizers work to full eect. If Tt+1 1 > Tt 0 after shock. θ T t+1 < 0 government needs to increase revenue But in most of the countries θ T t+1 > τ T t, which means T 1 t > T 1 t+1 : To give the intuition of this case would help interpret the results. 5 / 11

8 Comments on EUROMOD EUROMOD is used to estimate changes in disposable income after shocks to gross income. Great (and unique) tool to ensure cross-country comparability. But two potential concerns. 6 / 11

9 Comments on EUROMOD To what extent does EUROMOD model the specicities of each tax-benet system? High marginal tax rates imply high scal stabilization coecients. But the existence of tax credits and deductions if not accounted by EUROMOD could bias the results. 7 / 11

10 Comments on EUROMOD Higher marginal tax rates are not always associated to more revenue. Figure: PIT Marginal Rates and Revenue - Selected Group of Countries Marginal Tax Rate Single person at 100% of average earnings, no child SI EL HU NL ES FR UK NO Revenue on Income and Profits of Individuals % of GDP SE FI 8 / 11

11 Comments on EUROMOD This suggests that there could be signicant variations in tax benets across countries. If not accounted by EUROMOD and correlated to marginal rates, it could aect the cross-country comparability. Suggestion: compare EUROMOD estimates with external data and assess if a relationship with stabilization coecients exists. 9 / 11

12 Comments on EUROMOD Unemployment benets are an important component of income stabilization (after an unemployment shock). In Dolls, Fuest and Peichl (2012) they account for 40% of the income stabilization index. In many countries, unemployment benets depend on the contribution history of individuals unobserved by EUROMOD. How are they imputed? Again, an external validation would be useful. 10 / 11

13 Other Comments Stabilization coecients could vary a lot along the income distribution. Eects of positive/negative income shocks may not be symmetric. The paper is silent about output uctuations and demand stabilization. The interpretation of stabilization coecients and performance during the crisis is not obvious. Is it possible to compute standard errors of stabilization coecients? 11 / 11