Transcript The Dynamics of Trade and Competition

The Dynamics of Trade and
Competition
Natalie Chen (Warwick & CEPR)
Jean Imbs (Lausanne & CEPR)
Andrew Scott (London Business
School & CEPR)
Motivation
• Academic audiences attribute decline in
global inflation to improvements in central
bank practice
• Business audiences tend to attribute the
decline to globalisation and technology
You can see why….
I argue that the most important and most
unusual factor supporting worldwide
disinflation has been the mutually reinforcing
mixture of deregulation and globalization, and
the consequent significant decrease in
monopoly pricing power.
K. Rogoff, 2003
An issue worth investigating….
Globalisation and Inflation – What
are the links?
• Substitution towards cheaper imports brings down price
level and during transition lowers inflation
• Increasing competition narrows markups and lowers
price levels and lowers inflation during transition
• Increasing competition spurs productivity growth,
reduces costs and lowers inflation during transition
• Increasing competition restrains wage growth and
lowers inflation
• Increasing openness increases importance of exchange
rates and reduces effectiveness of inflation surprises
• Increasing competition reduces “output gap” and
reduces inflation bias
Globalisation and Inflation – Focus
of this paper
• Substitution towards cheaper imports brings down price
level and during transition lowers inflation
• Increasing competition narrows markups and lowers
price levels and lowers inflation during transition
• Increasing competition spurs productivity growth,
reduces costs and lowers inflation during transition
• Increasing competition restrains wage growth and
lowers inflation
• Increasing openness increases importance of exchange
rates and reduces effectiveness of inflation surprises
• Increasing competition reduces “output gap” and
reduces inflation bias
What this paper does
Outlines a theoretical model with rich microeconomic
channels through which trade exerts pro-competitive effects
on productivity, prices and mark ups
Combines model with EU sectoral data and includes control
for aggregate nominal influences (and in particular monetary
policy) to isolate micro pro-competitive effects.
Difference in Differences estimation
Differentiates between short run and long run effects.
Drastically different in theory
Contribution
Model implied observable variables, model implied
specifications.
Two-country version of Melitz-Ottaviano (2005) with
international differences in productivity, in wages and in
trading costs.
Openness (import penetration) has:
* negative and significant impact on
manufacturing prices
* positive and significant impact on
manufacturing productivity (truncation effect)
* negative and significant impact on margins
(pro-competitive effect)
Effects revert in the long run: non-liberalizing country
becomes an attractive base camp from which to export to
liberalized economy.
Plan
• Theory
• Estimation Strategy
• Data (markups)
• Main Results
Plan
• Theory
• Estimation Strategy
• Data (markups)
• Main Results
Theory
Objectives:
Introduce theoretical channels between prices,
productivity and mark ups
Motivate our measures and our estimation.
Ingredients:
Imperfect competition with elasticity of demand
depending on number of firms [Ottaviano, Tabuchi and
Thisse (2002)]. Then mark ups depend on number of
firms as well.
Firms with heterogeneous productivity, and fixed cost
of entry. Productivity is revealed after cost is paid, and
non-productive firms exit. [Melitz (2003)]
Mechanism:
Liberalizing domestic economy lowers tariff. Import share rises as more foreign
firms export to domestic market.
Rising import share leads to increase in number of firms.
Immediately lowers mark ups.
Also increases productivity as, with low prices, fewer firms make the cut.
Both channels reduce prices.
In long run, firms can choose where to locate. Closed economy attractive,
because more protected. Also, has become cheaper to export to domestic
market from there. Firms relocate abroad
.
Number of firms now falls, with opposite end effects on prices, margins and
productivity.
Inspiration:
Extension of Melitz (2003) and Melitz and Ottaviano (2005).
Demand
Inverted demand for variety u in sector i:
Implies total demand for variety u in sector i:
where N denotes total number of firms (domestic and foreign), and
L is market size (number of consumers). * denotes foreign country.
Supply
Labor is sole input, with unit cost c, unknown ex-ante, different across
countries.
τ denotes cost of foreign export to domestic market –
τ* cost of domestic export to foreign market.
Domestic profit maximization implies
Key Melitz-Pareto simplification: Assume c follows Pareto distribution in
[0,cM], with parameter s.
We further assume c* follows Pareto with parameter k in [0,c*M],
c*M ≠ cM.
Optimal pricing and distributional assumptions give average sectoral price
and costs:
Where cD is cost for marginal firm still in activity, i.e. the one that verifies
p(cD) = cD
By definition,
Equilibrium
Need to solve for cD and the number of firms.
Marginal firm still in business is pricing at cost, and is also the
one with highest price (lowest productivity). Nonnegativity
constraint on demand binding for this form and so
Thus
Negative, downward sloping relation between number of firms
supported by market N and threshold cost level. High costs means
high prices, limited demand and few varieties.
Short Run Supply
No location decision in the short run. The number of firms in each
country is given – but firms can still choose to participate in each
market, i.e. choose to produce for domestic and/or for foreign market.
In other words, the number of firms operating in each market is
endogenous (since decision to export is endogenous) – but number of
firms located in each market exogenous.
By definition:
Traces upward sloping relation between N and cD. The larger costs, the
larger the number of firms that choose to operate
A fall in τ increases N for a given level of cD. A fall in trading cost
means more firms will be operating in the domestic market, as foreign
exporters become active there.
In equilibrium, N increases and cD falls: prices, costs and markups fall.
Long Run Supply
Long run by definition means location decisions are endogenous,
i.e. so is the number of firms in each country.
Free entry conditions in both countries:
Simplifies (under Pareto assumption):
Now cD is independent on N or N*. Falling trading cost τ means higher
cD. I.e. higher prices, costs and markups. Relocation effect.
Relocation means bilateral trade liberalisation has anticompetitive effects in the long run
Plan
• Theory
• Estimation Strategy
• Data (markups)
• Main Results
Openness
Introduce import share θ
We have
By symmetry
Useful to rewrite:
From Theory to Estimation
Prices
Markups
Productivity
Econometrics Issues
• Intercepts
Estimation is differences in differences, i.e. international differences in
sectoral growth rates.
Country pair/sector specific intercepts
• Nominal Prices:
Model is one of real prices. Control for *aggregate* prices as well, and
thus for any (aggregate) influence on nominal prices.
• Lagged Dependent Variables
How long does the short run last? Aren’t prices sluggish?
Include lagged dependent variables. Not crucially affecting conclusions.
(Correct for bias induced by lagged dependent variables with fixed effects using
Arellano-Bond)
• Stationarity
• Endogeneity:
Instruments for import shares
1) Ratio of imports weights to their value, across countries, sectors
and over time.
2) Gravity inspired variable:
where ωjk denotes the (inverse of) distance between countries j
and k.
3) Transport costs, as measured by differences between CIF and
FOB values.
Taken together, instruments deliver R2 above 40%.
4) Dummies Single Market 1992 and Italian Lira re-entry 1996.
Plan
• Theory
• Estimation Strategy
• Data (markups)
• Main Results
Data
Data cover manufacturing sectors only.
7 countries, 10 sectors, 1989-1999.
Belgium, Denmark, France, Germany, Italy, the Netherlands, Spain
Sectoral PPI from Eurostat
Labor productivity (Real Value Added per Worker) from OECD STAN
Mark up data from
Bank for the Accounts of Companies Harmonized (BACH).
Homogeneous layout for balance sheets, profit and loss accounts,
investment and depreciation.
where Variables Costs = materials, consumables, staff
Plan
• Theory
• Estimation Strategy
• Data (markups)
• Main Results
Summary
Developed simple theory suggesting import shares should
affect prices negatively, via increased productivity and lower
markups.
Showed conjecture is supported by the data. Rising import
shares lower prices, because they increase productivity and
lower margins.
Effects of foreign openness on domestic variables, and of
relative numbers of firms are consistent with theory.
Crucial implication of model is that effects are opposite in the
long run. Surprisingly strong evidence supporting that
conjecture.
Robustness
Nominal Exchange Rates
Factor Endowments
GMM estimators
Benchmark (Italy) as a treatment effect
Origin of Imports
What about Globalisation and
Inflation?
• We ignored the macro channels through
which openness affected inflation
• Don’t examine labour market and impact
through wage restraint
• Focus on how cheaper imports, lower
markups and lower costs/greater
productivity contribute to lower inflation as
openness increases
What about Globalisation and
Inflation?
• Impact of greater openness in EU during
this period has contributed to lower
inflation
• Direct effect surprisingly small – around
0.1-0.2% per annum
• If believe the long run reversal effect then
can also expect this effect to unwind and
lead to higher inflation
I guess it’s the central bankers that did it!