Transcript Latin America’s Road to Inflation Targeting

Saving for Growth:
New Views on an Old Latin American Debate
Augusto de la Torre and Alain Ize
Center for Global Development
Washington, DC
February 2015
Chief Economist Office
Latin America and the Caribbean Region
The World Bank
1
After seven magnificent years, LAC is back to low growth
LAC: Real GDP Growth
2
To address the new reality, LAC will need to save more…
 We find aggregate demand composition (saving) to matter for growth…
Growth is not just about supply (investment prospects, productivity…)
 Foreign and domestic saving are not perfect substitutes

 …through three key channels
The real exchange (ER) rate channel (external competitiveness effect)
 The real interest rate (IR) channel (sustainability effect)
 The endogenous saving (ES) channel (saving-follows-growth effect)

 These findings, which deviate from the conventional neoclassical view,…
That saving matters neither for growth nor for the real exchange rate
 That growth leads to lower saving

 …reflect frictions that limit factor (especially capital) mobility and help
reconcile macro theory with well established observations

Hence, this is essentially a middle income country feature
 The policy implications are relevant to LAC … but fraught with tensions
3
We are dealing with an old yet largely unsettled debate
 The correlation between domestic saving and growth is undisputed…
Faster growth generally requires higher investment (it is not all TFP…)
 Higher investment implies higher domestic saving – as per the Feldstein-Horioka
puzzle (1980) of observed high correlation of S & I at country level

 … what is hotly debated is the direction of causality…
“… no support for the view that domestic saving is the binding constraint to
economic growth…” [saving follows growth] (Rodrik, 2000)
 “Countries whose productivity falls behind are countries that ‘tax’ saving…”
[rationale for the “allocation paradox”] (Gourinchas & Jeanne, 2011)

 … and how to reconcile theory with observation and policy concerns
Consistent with observation (Prasad et al., 2007), policy makers worry that low
saving and CA deficits can undermine growth
 But conventional open-economy growth models treat domestic and foreign
savings as perfect substitutes and, hence, capital as perfectly mobile…
 … contrary to the observed large and persistent productivity dispersion

4
Exploring the links between saving and growth
5
ER channel: works via the current account, external
competitiveness, and tradable vs. nontradable production
Positive learning
externalities
bs
SD
•
•
•
•
•
•
•
bg
e
be
be
bi
I
d
g
Suboptimal saving (excess expenditure over income) appreciates the real exchange rate because T (imports)
respond elastically at given world prices while the price of NT rises, given inelastic supply.
The real appreciation lasts because frictions limit inter-sectorial and inter-national factor mobility.
The real appreciation hinders growth inasmuch as a smaller T sector generates less learning externalities.
National saving is suboptimal because externalities are not internalized.
NB1: The frictions that limit factor mobility also limit the substitutability between foreign and domestic
saving, which is why domestic saving matters for the real exchange rate and, hence, growth.
NB2: SD and e reinforce each other (endogeneity and multiplier effect).
NB3: Whether the current account deficit widens due to low saving or high investment matters for growth—
the difference depends on d.
6
IR channel: works via the capital account, domestic vs.
foreign saving, and the sovereign risk rating
Positive learning
externalities
e
SD
gr
ge
I
gr
gi
r
d
g
gg
Negative crises
externalities
•
•
•
Ex ante: sub-optimal saving raises the interest rate via BOP deficits and the sovereign risk premium
• The risk premium rises as collateral constraints increase default risk, thereby limiting capital mobility
• A higher risk premium depresses growth through a higher cost of capital
Ex post: sub-optimal saving endangers external debt viability => BOP crisis => negative growth externalities
NB1: sub-optimal saving drags down growth even though it depreciates the real exchange rate
7
ES channel: works through the multiplier effect of the
responsiveness of domestic saving to growth
Positive learning
externalities
bg
e
SD
be
be
bi
ge
gr
I
gr
gi
r
d
g
gg
Negative crises
externalities
a
•
•
•
Multiplier effect
Saving is growth elastic because frictions (e.g., collateral constraints) prevent inter-temporal smoothing.
The responsiveness of saving to growth introduces a multiplier effect.
8
If the multiplier effect is large enough (ad >1) => growth would self-propel and SD policy would not matter
Empirical results: growth effects of saving are stronger
for middle-income countries with BOP deficits
 Strong fit for MICs, weaker fit for
HICs and LICs

Saving matters more at middle
income stages of development
 Significant asymmetries
CA deficits matter more for e and r
than surpluses
 Saving for growth implies avoiding
chronic CA deficits

 Sizable impacts under IV estimates
+ 10 pp SD => 1.5 pp higher g
 Should be even higher for CA < 0!

Structural Form
High Income Middle Income
0.0387**
0.0283**
Low Income
0.00788
IV
CA>0
0.07**
0.107***
0.0567+
1.05
0.0296
-0.349**
-0.781
-0.4+
-0.328***
-0.0332
-0.320***
0.0330
(-0.3)
-1.411***
-1.411***
-1.645*
-1.286*
0.146
-1.4+
γr
0.0222
-0.0913***
-0.0367
-0.121***
0.0257
(0)
γe
0.218***
0.218***
0.388***
0.137*
0.143
(0)
γI
0.266***
0.266***
0.167
0.270***
0.305+
(0)
γg
1.049***
1.049***
2.174+
0.884*
0.836
2.8**
δ
3.114***
3.114***
1.482**
3.256***
2.237*
(3.1)
0.08
0.15
0.04
0.19
0.00
1.54
α
CA>0
0.0227**
CA<0
0.0227**
βe
0.0804+
0.182***
0.0119
βS
-0.379***
-0.379***
βI
-0.328***
βg
Impact
*** p<0.01, ** p<0.05, * p<0.1, + p<0.2
Notes: Statistical significance is reported for country clustering. The standard deviations for the IV estimation are
calculated using the suest and nlcom Stata routines. Impact measures the growth respnse (in percentage points) to a
10 percent of GDP increase in the saving ratio.
9
On average, the world as a whole follows an ER pattern,
but deviations likely reflect an IR pattern
Domestic Saving and Real Exchange Rate Gaps, 1981-2011
10
ER pattern: high savers are undervalued and grow faster
Under-saving
Saving Gap
Over-saving
Average Gaps by Quadrant for the Whole Sample
(1981-2011)
Undervalued
Overvalued
Real Exchange Rate Gap
11
IR pattern: under-savers with BOP sustainability problems
(as proxied by the country ratings) are undervalued
Under-saving
Saving Gap
Over-saving
Average Gaps by Quadrant for the Whole Sample
(1981-2011)
Undervalued
Overvalued
Real Exchange Rate Gap
12
Undervaluation is good for growth, overvaluation is not
Under-saving
Saving Gap
Over-saving
Average Gaps by Quadrant for the Whole Sample
(1981-2011)
Undervalued
Overvalued
Real Exchange Rate Gap
13
Saving also matters for future growth (investment)
Under-saving
Saving Gap
Over-saving
Average Gaps by Quadrant for the Whole Sample
(1981-2011)
Undervalued
Overvalued
Real Exchange Rate Gap
14
Where is LAC?
15
LAC stayed for a long time under the IR spell (under-saver
– undervalued)…
Domestic Saving and Real Exchange Rate Gaps, 1981-2011
2
1
National Saving
EAP
ECAMNA
0
HI
LAC1 SSA
LAC2
-1
-2
-1
-0.5
0
Real Exchange Rate
0.5
1
●LAC1 Countries per Period ♦LAC1 1990-2011 Average ●LAC2 Countries per Period ♦LAC2 1990-2011 Average
♦Other groups of countries 1990-2011 Average
●Other countries per Period
16
…which reflected LAC’s past history of low sovereign
ratings and chronic macro-financial problems
Domestic Saving and Sovereign Rating Gaps, 1981-2011
2
1
National Saving
EAP
ECA MNA
0
SSA
LAC1
HI
LAC2
-1
-2
-1
-0.5
0
Country Rating
0.5
1
●LAC1 Countries per Period ♦LAC1 1990-2011 Average ●LAC2 Countries per Period ♦LAC2 1990-2011 Average
♦Other groups of countries 1990-2011 Average
●Other countries per Period
17
Thus, LAC grew slowly despite competitive exchange rates…
Real Exchange Rate and Growth Gaps, 1981-2011
10
GDP per capita Growth
5
ECA EAP
LAC1
0
MNA SSA
LAC2
HI
-5
-10
-1
-0.5
0
Real Exchange Rate
0.5
1
●LAC1 Countries per Period ♦LAC1 1990-2011 Average ●LAC2 Countries per Period ♦LAC2 1990-2011 Average
●Other countries per Period
♦Other groups of countries 1990-2011 Average
18
… but there was substantial heterogeneity within LAC…
Saving and Real Exchange Rate Gaps for LAC1 Countries, 1990-2011 Averages
0.5
VEN
ER/HS
CHL
IR/HS
PAN
ARG
MEX
PER
0
ER / LS
National Saving
ECU
URY
IR/LS
COL
BAH
BRA
TTO
-0.5
BAR
ER/LS
CRI
-1
-0.4
-0.2
0
0.2
Real Exchange Rate
19
…resulting in contrasting macroeconomic performances of
the ER low and high savers, matching the ER priors
Policy-Adjusted Gaps for High and Low Saver LAC1 Countries, 1981-2011
Panel A. National Saving Gaps
Panel C. Real Exchange Rate Gaps
Panel G. GDP per capita Growth Gaps
Panel D. Country Rating Gaps
20
LAC moved to the ER pattern in the last decade, reflecting
large real appreciations and improved country ratings
Saving and Exchange Rate Gaps for LAC1 Countries, 2004-2011 Averages
0.5
ARG
VEN
PAN
TTO
0
CRI
National Saving
PER CHL
ECU MEX
BAH URY
COL
BRA
-0.5
-1
BAR
-1.5
-0.5
0
Real Exchange Rate
0.5
21
Final thoughts
22
A saving mobilization agenda should make sense for many
LAC countries
 Demand appears to also matter for growth, but not in the
Keynesian way


LAC countries with substantial, recurrent CA deficits should save more
to grow faster…
… which should in now way distract from sorely needed supply-side
and enabling environment (rule of law) reforms
 Raising saving is within the reach of policy…

It can be done via fiscal, financial, and safety net reforms
23
However, it won’t be easy
 It will take time, requiring perseverance
 Difficult tensions will need to be managed

Macro conflict between short-term and long-term growth objectives
• Switch towards a “tighter fiscal looser monetary” policy mix
• Tilt public investment in favor of investment

Distributional conflict – who will consume less?
 Current weak world demand and easy liquidity make things
even trickier

The combination of higher country ratings in LAC and markets willing
to finance CA deficits can boost the adverse impacts of the ER channel
24
Thank you
25