The paper examines how real exchange rate volatility affects productivity growth and how this relationship depends on a country's level of financial development. It finds that the impact of exchange rate volatility on growth varies according to a country's financial development - volatility has a more negative impact in countries with lower financial development. The authors develop a theoretical model to explain this finding, where financial development determines firms' ability to finance innovation costs during exchange rate fluctuations. However, some comments note issues with the empirical specifications and question aspects of linking the model to the empirical results.
1. Exchange Rate Volatility
and Productivity Growth:
The Role of Financial Development
Aghion, Bacchetta,
Ranciere, and Rogoff
Comments by Valerie Cerra
2. Main point
Real exchange rate volatility, and thus
the exchange rate regime affect growth
Impact depends on the country’s level of
financial development
3. Related Literature on exchange
rate regimes and growth
Ghosh, Gulde & Wolf: hard pegs work
best
Sturzenegger & Levy-Yeyati: floats are
best
Reinhart-Rogoff: lntermediate flexibility
performs best
4. Correlations Among Regime
Classification Schemes
(Frequency of outright coincidence, in %, given in parentheses)
IMF Ghosh LY-S R-R
1.00
IMF
(100.0)
0.60 1.00
Ghosh
(55.1) (100.0)
0.28 0.13 1.00
LY-S
(41.0) (35.3) (100.0)
0.33 0.34 0.41 1.00
R-R
(55.1) (35.2) (45.3) (100.0)
(Frequency of outright coincidence, in %, given in parenthesis.)
Sample: 47 countries. From Frankel, “Experience of and Lessons from Exchange Rate
Regimes in Emerging Economies,” in Monetary and Financial Integration in East Asia,
ADB, 2004. Table 3, prepared by Marina Halac & Sergio Schmukler.
6. Empirics: robustness tests
Alternative measures: real ex rt volatility, real
ex rt overvaluation, Reinhart Rogoff regimes,
Gosh et al regimes, Levy-Yeyati and
Sturzenegger
Time windows
Measures of financial development
Crisis dummies
Continents
Nonlinearities
7. Theoretical mechanism
Productivity growth occurs through innovation
that requires payment of liquidity cost
If country has low level of financial
development, liquidity cost must be paid
through operating profits
Nominal wages are sticky one period and
Pt = St due to PPP assumption
Operating profits depend on price (thus ex rt)
surprises that generate real wage losses
8. Theoretical mechanism
Unexpected nominal appreciation
lower price level and higher real wage
than expected
High operating profits fund the liquidity
cost
higher productivity growth
Unexpected nominal depreciation
lower productivity growth
9. Theoretical mechanism:
Exchange rate volatility and growth
Probability of innovation profits
Financial development
Concave distribution
of liquidity cost shocks Cost of innovation
10. Comments: Empirics
Useful finding that exchange rate
regime has macro impact
But is it a story about real exchange rate
volatility rather than regime?
Robustness of regime classification: LYS
not significant—free fall or dual ex rt?
11. Comments: Empirics
All relevant control variables present?
• Savings or investment rates
• Political crises and wars
• Capital account openness
• Govt investment, esp infrastructure
• Govt deficits or debt overhang
Proper lags?
• Secondary enrollment
12. Measure of financial
development
Private credit to GDP – often used to measure financial
development for long-term growth
Hardy and Pazarbasioglu (1998) – boom and bust pattern in
advance of banking crises. Kaminsky and Reinhart (1996) –
growth in domestic credit to GDP accelerates steadily and
markedly as the crisis approaches, peaking at the time the
crisis erupts.
Sachs, Tornell, and Velasco (1996), Radelet and Sachs
(1998) and Corsetti, Pesenti, and Roubini (1998) argue that
this measure proxies for financial fragility, as the quality of
bank loans is likely to deteriorate significantly when bank
lending grows at a rapid pace in a relatively short period of
time.
Five year averages and within country estimation – relates to
short-term changes, thus possibly financial fragility rather
than development
13. Use of 5 year growth averages
To “filter out business cycle fluctuations”
Standard, but only true if there are business cycle
fluctuations
Cerra and Saxena, 2005, “Growth Dynamics: The
Myth of Economic Recovery”
• No “cycles” – Shocks are permanent
Nominal wage rigidity much shorter than 5 years
Model relies on operation profits changing due to
business cycle mechanism, so why filter?
Measure of financial development argues for
between country regressions
14. Empirics
Verify whether there is a weak
instrument problem using lags of
variables as instruments in GMM?
Heterogeneity in the estimates?
Try direct measure of corporate
profitability or dependence on external
finance
15. Theoretical model
Robustness of cost distributions and
production technologies?
• Prob of innovation may be convex in S t
Uniform cost Convex if labor share
distribution is greater than 1/2
16. Model’s link to empirics
Model assumes PPP, providing link between
ex rt and price surprise,
• P=S
but empirics show result depends on changes
in real exchange rate
Model requires unexpected price (ex rt)
changes: if workers expect deprec or apprec, it
would be built into wage contract
17. Model’s link to empirics
Show evidence that nominal exchange rate appreciation
leads to significant rise in real labor costs and decline in
operating profits, or that appreciation leads to fall in
productivity
• Maybe other mechanism: fluctuations in operating profits due to
pricing to market
Countries far from technological frontier may upgrade by
importing, so appreciation may improve liquidity and thus
productivity growth
Balance sheet effects of appreciation could relax credit
constraints if net wealth is considered as collateral
Private credit to GDP and output per worker are strongly
correlated, so are results on financial development proxy
for convergence relationship?
• Which interaction predominates if both included in regression?
18. Model
Stabilizing role of flexible exchange rates
analyzes nominal ex rt and productivity
volatility using Taylor rule
• Reaction to productivity shock is not clear—how is it
assumed to affect the output gap?
• One for one passthrough is counter to evidence,
especially for developed countries, that the paper find
benefit from volatility
• Taylor rule may have forward looking component.
Prices may be stable despite nominal ex rt changes,
especially if noise in risk premium is temporary
19. Empirics (previous version)
Higherlevels of regulation have negative
impact on growth as exchange rate is
more flexible
• Less flexible labor market, product markets
and entry fixed
Counters OCA theory – ex rt flexibility
substitutes for other types of flexibility in
adjusting to real shocks