economic resilience: trade-offs between growth and ... · efficiency channel fragility channel...
TRANSCRIPT
ECONOMIC RESILIENCE:
TRADE-OFFS BETWEEN
GROWTH AND ECONOMIC
FRAGILITY
Filippo Gori Economics Department, OECD Policy Challenges in the Global Economy NERO Meeting, Paris, 19 June 2017
2
Financial crises and deep recessions have been frequent
0
2
4
6
8
10
12
14
16
18
20
1970 1975 1980 1985 1990 1995 2000 2005 2010
Banking, currency or sovereign debt crises Severe recessions
Note: The chart refers to crisis and severe recession episodes for 35 OECD countries over the period 1970Q1-2010Q4. Crisis
episodes are from Babecky et al. (2012).
Source: OECD calculations based on Babecky et al. (2012) data and Hermansen and Röhn (2015) for severe recessions.
120 episodes of
financial crisis and
over 100 cases of
severe recessions
Number of OECD countries affected during each episode
0
2
4
6
8
10
12
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
Unemployment rate
Japan United States European Union (28 countries)
3
Crises might have long-lasting effects
Per
cen
t
Source: OECD.stat
Pre-crisis Levels
GDP growth
• Measures should be taken to reduce economic fragility.
Risk mitigation is key, but….
4
Economic
fragility
• However we should be aware of the growth consequences of risk reducing policies
Economic
fragility
Fragility
GDP
Growth
• Risk-mitigation needs to be balanced against potential cost of economic growth
Benefits and costs of risk mitigating policies should be considered
5
Crisis
Risk
• Banking crisis
• Currency crisis
• Twin crisis
vs GDP growth
extreme negative
GDP growth rates
vs
mean GDP growth
Paper I Paper II
Tail
Risk
6
0 GDP growth
Extreme negative growth
Tail risk = the risk of an extreme negative growth shock
GDP-at-Risk = quantile of the
GDP growth distribution
i.e. max GDP loss with 95% prob.
Approach I: growth and GDP tail risk
Theoretical quantiles of normal distribution
Extreme negative growth events not uncommon...
United States United Kingdom
7
Netherlands
Extreme
negative
growth rates
Extreme
positive
growth rates
𝑄𝜏 𝑦𝑡|𝑋𝑡 = 𝑋𝑡 𝛽(𝜏)
𝛃 𝛕 : effect on the τth quantile of y due to a marginal change in X (e.g. for 𝜏 = 0.05 => the impact on the 5-percent GDP-at-risk).
𝒚: quarterly GDP growth.
𝑿: factors that can influence the GDP growth distribution:
1) country characteristics (size, stage of development, trade openness)
2) macro policies (size of automatic stabiliser, real short-term interest rate)
3) structural policies
8
Conditional quantile regression
Unbalanced panel with a maximum of 42 countries, mostly OECD, over the period 1970-2014
Approach I: growth and GDP tail risk
9 Policy variable
Positive effect on growth
GD
P g
ro
wth
Approach I: growth and GDP tail risk
Conditional quantile regression an illustration
Increase in tail risk
x1 x2
10
Conditional quantile regression an illustration
The diamonds depict the point estimates and the thick bars a confidence interval of +/- 1 standard deviation. The horizontal blue line depicts the OLS estimate. Source: Authors’ calculation based on Caldera Sánchez and Röhn (2016).
The relationship between government effectiveness, mean growth and GDP tail risk
Approach I: growth and GDP tail risk
11
Conditional quantile regression an illustration
The diamonds depict the point estimates and the thick bars a confidence interval of +/- 1 standard deviation. The horizontal blue line depicts the OLS estimate. Source: Authors’ calculation based on Caldera Sánchez and Röhn (2016).
The relationship between active labour market policies, mean growth and GDP tail risk
Approach I: growth and GDP tail risk
12
Policies
Growth
Crisis Risk
Possible fragility channel
Growth Channel (efficiency channel)
Financial liberalisation
Capital account openness
Trade openness
Exchange rate regime
Product Market Regulation (PMR)
Tax settings
…
Approach II: growth and crisis risk
• A growth regression
Growthiτ= policyit + controlsit + εiτ
13
Growth
Growth Channel
Crisis Risk
Possible fragility channel • Nested with a probability model:
Pr(Crisisiτ)= policyit + controlsit + μiτ
(Similar to Ranciere Tornell Westermann (2006) – a treatment effect model (Maddala 1983)
Sample: panel of 100 countries, 1970-2010)
Growthiτ= policyit + controlsit + Pr (𝐶𝑟𝑖𝑠𝑖𝑠𝑖𝜏)+ εiτ
2 SLS estimation
Crisis Risk
Systemic banking crisis
Currency crisis
Twin crisis
Approach II: growth and crisis risk
14
Approach II: growth and crisis risk G
row
th (
perc
enta
ge p
oint
s)
Crisis risk (red line) percent
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
Financial marketsopenness
Capital accountopenness
(Less) Reg. in 7 non-manufacturingnetwork-ind.
Adjusted statutorycorporate income tax
rate
Efficiency and fragility channel
Direct growth effect (efficiency channel) Crisis risk (fragility channel) - right axis
Growth and crisis risk, an illustration
RESULTS
15
Lower growth and lower fragility
Higher growth and lower fragility Higher growth and fragility
Lower growth and higher fragility -0.50
-0.30
-0.10
0.10
0.30
0.50
-0.88 -0.68 -0.48 -0.28 -0.08 0.12 0.32 0.52 0.72
Gro
wth
(Y
axi
s)
Fragility (X axis)
Note: The X axis plots the effect on fragility, defined as a higher likelihood of financial crises (currency, banking or twin crises) or a higher GDP (negative) tail risk., where tail risk is defined as the effect on the bottom 10% of the distribution of quarterly GDP growth. The Y axis plots the effect of a policy variable on growth. Source: OECD calculations in Caldera Sanchez and Gori (2016) “Can Reforms promoting growth increase financial fragility? An empirical assessment” and Caldera Sanchez and Roehn (2016) “How do policies influence tail risks?”
A growth-fragility space…
Policy areas outside the financial sector do not involve growth-fragility trade-offs
17
The X axis plots the effect on fragility; Fragility is defined as higher likelihood of financial crises (polices with red outline) or a higher GDP (negative) tail risk.
Three types of financial crises are considered: Currency, banking and twin crises. Tail risk is defined as the effect on the bottom 10% of the distribution for
quarterly GDP growth. For each policy, the Y axis plots the average (overall) growth effect.
Source: Authors’ calculation based on Caldera Sánchez and Gori (2016) and by Caldera Sánchez and Röhn (2016).
Pro-growth
product and labour
market policies
have little or no
impact on fragility
Better-quality
institutions associated
with higher growth
and lower fragility
More trade-offs are found in the area of financial market policies
18
The X axis plots the effect on fragility; Fragility is defined as higher likelihood of financial crises (polices with red outline) or a higher GDP (negative) tail risk.
Three types of financial crises are considered: Currency, banking and twin crises. Tail risk is defined as the effect on the bottom 10% of the distribution for
quarterly GDP growth. For each policy, the Y axis plots the average (overall) growth effect.
Source: Authors’ calculation based on Caldera Sánchez and Gori (2016) and by Caldera Sánchez and Röhn (2016).
Macro-prudential
policies can reduce
fragility, but may also
lower growth
Growth benefits from
financial market
liberalisation offset
by higher crisis risk
Capital account openness increases both growth and economic fragility..
Composition of capital flows matters
19
Growth benefits from
capital account
openness outweigh
crises risks
Fragility linked to
portfolio debt
inflows
The X axis plots the effect on fragility; Fragility is defined as higher likelihood of financial crises (polices with red outline) or a higher GDP (negative) tail risk.
Three types of financial crises are considered: Currency, banking and twin crises. Tail risk is defined as the effect on the bottom 10% of the distribution for
quarterly GDP growth. For each policy, the Y axis plots the average (overall) growth effect.
Source: Authors’ calculation based on Caldera Sánchez and Gori (2016)
• Product and labour market policies
• Enhance growth, little impact on crisis risk.
• However a growth-fragility trade- off exists in other policy
areas: Financial market liberalisation
• Pro-growth impact of reduced by crisis risks, which are driven by
=> Excess private credit growth
Misalignments in real estate markets
Capital account openness
• Pro-growth effects of greater outweigh crisis risk
=> Risk mainly associated with portfolio debt flows.
Macro prudential necessary but may entail some costs.
Concluding remarks
20
Lower growth and lower fragility
Higher growth and lower fragility Higher growth and fragility
Lower growth and higher fragility -0.50
-0.30
-0.10
0.10
0.30
0.50
-0.88 -0.68 -0.48 -0.28 -0.08 0.12 0.32 0.52 0.72
Gro
wth
(Y
axi
s)
Fragility (X axis)
Financial Markets
Quality of institutions
Labour Markets
Product Markets
Macro prudential
policies
Note: The X axis plots the effect on fragility, defined as a higher likelihood of financial crises (currency, banking or twin crises) or a higher GDP (negative) tail risk., where tail risk is defined as the effect on the bottom 10% of the distribution of quarterly GDP growth. The Y axis plots the effect of a policy variable on growth. Source: OECD calculations in Caldera Sanchez and Gori (2016) “Can Reforms promoting growth increase financial fragility? An empirical assessment” and Caldera Sanchez and Roehn (2016) “How do policies influence tail risks?”
Sum-up: evidence on growth-fragility
OECD Resilience website
http://www.oecd.org/economy/growth/economi
c-resilience.htm
BACK UP
22
23
Approach II: growth and crisis risk
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
Capital accountopenness
FDI Portfolio equity Portfolio debt Other flows
Effect of capital flow on growth and crisis risk
Efficiency channel Fragility channel Overall growth Crisis risk
Capital flows composition
Gro
wth
(pe
rcen
tage
poi
nts)
Crisis risk (red line) percent
Growth and crisis risk, an illustration
24
Conditional quantile regression an illustration
The diamonds depict the point estimates and the thick bars a confidence interval of +/- 1 standard deviation. The horizontal blue line depicts the OLS estimate. Source: Authors’ calculation based on Caldera Sánchez and Röhn (2016).
The relationship between active labour market policies, mean growth and GDP tail risk
Approach I: growth and GDP tail risk