Which Structural Reforms Matter for economic growth: PREZENTĀCIJAS NOSAUKUMS Evidence from Bayesian Model Averaging Olegs Krasnopjorovs (Latvijas Banka) 2 nd Lisbon Conference on Structural Reforms 06.07.2017
MOTIVATION Structural reforms matter for economic growth. But we cannot "improve everything" at the same time => on which reforms we should focus on? By how much this reforms will enhance economic growth of the EU? Classical regression approach: we can estimate: Real GDP per capita growth = f (Global Competitiveness Index components) But: Model uncertainty is a well-known problem in empirical literature on economic growth. With > 100 potential growth determinants, it is easy to find several statistically significant combinations of GCI components, but it is not clear: 1. Which model to use (and which factors indeed are growth determinants)? 2. What is the impact of these factors on growth (since every model gives different coefficient or even opposite sign). 2
ALTERNATIVE: BAYESIAN MODEL AVERAGING Brock, Durlauf, West (2003): "in many cases, model selection is actually inappropriate, as conditioning policy evaluation on a particular model ignores the role of model uncertainty." "In contrast, we advocate the use of model averaging methods, which represent a formal way through which one can avoid policy evaluation that is conditional on a particular economic model". BMA estimates all combinations of potential growth determinants (=2^X), and constructs weighted average of all models. 2 potential growth determinants: 4 models; 10 potential growth determinants : > thousand models 20 potential growth determinants : > million models etc. 3
DATA Sample: 191 country (IMF database). Exclude: o micro-states (<0.25 mill.population); countries with missing data; o natural resource exporters (pure rent >20% of GNI; World Bank data); => 106 countries. Theoretical foundation: conditional beta-convergence model (initial income included). To mitigate endogeneity issues: Cross-sectional model: Global Competitiveness indices for 2005 => Real GDP per capita growth 2006-2015. To remove business cycle impact: HP filtered real GDP per capita growth 2006-2015 (apply HP filter to 2000-2019 data to mitigate end-point problem) 4
BMA: EMPIRICAL RESULTS Posterior inclusion probability (share of models that include this factor) Share of models with "+" sign (1) / "-" sign (0) Source: author's calculations 5
initial income level proxy for health system quality Investments quantity (% of GDP) Wage flexibility; reflects also labour market flexibility GCI pillar 3 "Macroeconomic Stability" Investments quality (FDI new technology transfer) FDI restrictions Total tax rate, % of profits exports; proxy for trade openness Strength of investor protection (Low) costs of crime and violence on businesses Efficient use of talent / human capital proxy Quality of education system proxy Blue: increase GDP per capita growth; Red: decrease GDP per capita growth; Bold: might be changed with structural reforms Source: authoŗ's calculations 6
2 EU COUNTRIES WITH THE WEAKEST ECONOMIC GROWTH (GR, IT) ARE WELL-IDENTIFIED WITH THIS MODEL Application of global model to the EU countries We cannot deny the role of countryspecific factors (not included properly in GCI), but still largest part of cross-country growth differences may be represented with our BMA model. Source: authoŗ's calculations 7
IMPACT ON GDP PER CAPITA GROWTH: (EXAMPLE: INVESTOR PROTECTION) Top-3 EU countries with the highest investor protection (as measured by GCI 2006-07): Ireland (8.3); UK (8.0); Belgium (7.0). Top-3 average: 7.77; BMA coefficient: 0.035 => Latvia: 5.7 (drag on growth: 0.07 pp.); Greece: 3.0 (drag on growth: 0.17 pp.). 8
DRAG ON GROWTH FROM INSTITUTIONAL WEAKNESSES (PP; ANNUAL REAL GDP PER CAPITA; PEER: EU TOP-3 AVERAGE) Math quality costs of crime investor protection health system quality labour market flexibility & efficiency FDI restrict., tech. transfer Baltics: Estonia 0.03 0.04 0.06 0.08 0.02 0.13 0.36 Latvia 0.05 0.06 0.07 0.16 0.06 0.42 0.82 Lithuania 0.03 0.09 0.09 0.13 0.05 0.55 0.94 Weak growth EU countries: Greece 0.05 0.03 0.17 0.02 0.59 0.59 1.44 Italy 0.07 0.17 0.10 0.02 0.60 0.66 1.61 Cyprus 0.04 0.04 0.27 0.02 0.42 0.50 1.29 Portugal 0.09 0.01 0.06 0.02 0.35 0.23 0.75 SUM Source: authoŗ's calculations 9
Italy Greece Bulgaria Croatia Cyprus Slovenia Austria Germany Poland Finland Sweden Romania Netherlands Belgium EU28 Lithuania Spain Malta Latvia Luxembourg France Portugal Denmark Hungary Ireland Czech Rep. Slovak Rep. Estonia UK 0.47 0.46 0.36 0.34 1.32 1.30 1.29 1.24 1.23 1.18 1.12 1.11 1.07 1.05 1.05 1.00 0.97 0.94 0.87 0.86 0.82 0.82 0.80 0.75 0.68 0.63 0.61 1.44 1.61 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 IF EU BEST INSTITUTIONAL PRACTICES WOULD HAVE ADOPTED 10 YEARS AGO, EU INCOME ANNUAL GROWTH WOULD BE BY 1.0 PP HIGHER DURING THE PREVIOUS DECADE Drag on 2006-15 growth from institutional weaknesses (pp; annual real GDP per capita) Source: authoŗ's calculations 10
Italy Greece Bulgaria Austria Croatia Slovenia Hungary Finland Cyprus Romania Germany Poland EU28 Slovak Rep. Spain Latvia Belgium Sweden Netherlands France Denmark Portugal Malta Lithuania Czech Rep. Luxembourg Estonia UK Ireland 0.28 0.24 1.23 1.20 1.19 1.13 1.06 1.05 1.02 1.01 0.91 0.86 0.83 0.81 0.79 0.76 0.76 0.76 0.75 0.74 0.67 0.64 0.63 0.61 0.51 0.46 0.41 1.49 1.38 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 IF APPLY BMA COEFFICIENTS TO GCI 2016-17: DRAG ON ANNUAL INCOME GROWTH DURING THE NEXT DECADE MIGHT BE 0.8 PP Drag on 2016-25 growth from institutional weaknesses (pp; annual real GDP per capita) Source: authoŗ's calculations 11
CONCLUSIONS By introducing EU best institutional practices to all EU countries, EU real GDP per capita annual growth rate may be raised by about 1 percentage point. ~ 1 pp. growth dividend is it worth reforming? Structural reforms is not a magic stick: cannot reach Luxembourg income level in few years; But impact on growth is significant: with reforms we will be 10% richer in 10 years. Improve quality of education and health systems; Combat crime and violence; Strengthen the protection of investors, including trust/quality of judicial system and insolvency procedures; Improve health system quality; Raise labour market flexibility and particularly wage flexibility; Decrease FDI restrictions and increase FDI technology transfer. 12
13