References
Import Competition and the
Composition of Firm Investments
Philippe Fromenteau
Jan Schymik
Jan Tscheke
University of Munich
June 2016
Introduction
Investment decisions are determined by a trade-off between current and
expected future profits
If foreign competition reduces expected future profits, firms will channel
financial resources towards the present and away from the future
This has potentially important economic implications on
1
credit demand and financing cost, because each dollar invested needs to be
refinanced more frequently
2
firm-size distribution over time, because heterogeneous firms are
differentially affected by import competition
3
long-run economic growth, because firms might not fully exploit long-term
potential (short termism)
“We need an economy where companies plan for the long run [. . . ] - leading
to higher productivity, better service, and larger profits.” (Clinton 2016)
What we do in this paper
We study whether competition from abroad incentivizes firms to shift
investments from long-term into short-term assets
We model the investment decision of a representative firm
firm maximizes current and future profits
firm investment reduces marginal costs today (short-term investment) or
tomorrow (long-term investment)
competitive pressure diminishes future price-cost margins, reducing rents on
long-term investment
We use the theoretical structure to derive
testable predictions
within-firm difference-in-differences estimator
We test the model predictions with US firm data for 1995-2009
We quantify the effect on financing costs and the differential effect for
small firms
Related Literature
Impact of international trade on level of investments into technology
Access to foreign markets induces investments in technology upgrading, see
Lileeva and Trefler (2010), Bustos (2011)
Chinese import competition increases productivity and the absolute volume
of innovation, see Bloom et al. (2016)
Technology upgrading: IT spending/usage, R&D investments, etc.
Impact of international trade on corporate finance
Import competition leads to market share increase for firms with large cash
holdings, increases the cost of debt, induces firms to reduce debt-equity ratio
see Fresard (2010), Valta (2012), Xu (2012)
Causes for short-term investment behavior
Economic uncertainty, credit crunches, investor pressures or agency problems
see Aghion et al. (2010), Garicano and Steinwender (2016), Terry (2015),
Garicano and Rayo (2016) and Bénabou and Tirole (2016)
Theoretical Framework
Demand and Industry Structure
economy exists for two time periods t ∈ {0, 1}
each period t: Lt consumers
linear-quadratic utility function following Melitz and Ottaviano (2008)
firms face a linear demand:
qit = At −
Lt
pit ,
γ
intercept is given by
At ≡
linear demand → ptmax =
zero
ηNt Lt
αLt
+
p¯t
ηNt + γ
ηNt + γ γ
αγ
ηNt +γ
+
ηNt
ηNt +γ p¯t
at which demand is driven to
ptmax : measure for toughness of competition
Nt : number of firms; p¯t : average price level; γ: degree of differentiation
Production and Investment Decision
investments: profit maximizing firms can opt for two types of investment in
order to reduce their marginal costs of production c ∗ :
short-term investments k: reduce the unit costs of production in period 0
θ
c0 = c ∗ − (c ∗ ) k 0.5
long-term investments z: reduce the unit costs of production in period 1
θ
c1 = c ∗ − ϕ (c ∗ ) z 0.5 ,
ϕ>1
magnitude of cost reduction depends on firm productivity c ∗ and the
parameter θ
unit of short-term investment k and long-term investment z are equally
costly and require r units of labor
Production and Investment Decision
firms compete on a monopolistically competitive market and take the
average price level p¯t as well as the number of firms Nt as given
per-period profits are given as:
π (ct ) =
2
Lt D
ct − ct
4γ
firm with p ctD = ctD = ptmax earns zero profits
ctD : reflects intensity of competition in the economy
reduction in ctD implies a rise in the toughness of competition
firms with lower marginal costs ct charge lower prices, generate larger
demand, face a lower price elasticity of demand, set higher mark-ups and
earn higher profits
Production and Investment Decision
Choice between short- and long-term investments:
taking the size of the market Lt and the level of competition ctD as given,
the firm optimizes profits discounted with a factor δ ∈ (0, 1) over time:
max π (c0 ) + (1 − δ) π (c1 ) − rk − rz.
k,z
determining the first order conditions with respect to short- and long-term
investments and solving for the optimal level of k and z yields
k 0.5
z
0.5
=
−1
4γr
2θ
− (c ∗ )
L0
4γr
2θ
− ϕ (c ∗ )
L1 (1 − δ) ϕ
=
θ
c0D − c ∗ (c ∗ )
−1
θ
c1D − c ∗ (c ∗ ) .
we are going to consider the ratio of these optimal investments (in logs)
intuitively: firms choose the investment composition such that the marginal
returns of short- and long-term investments are equalized
Impact of Import Competition on Investment Composition
increase in competition (c1D < c0D ) reduces firms’ profits in period 1
this diminishes the marginal return on long-term investments
firms adjust their investment composition towards short-lived investments
𝑀𝑅𝑘
𝑀𝑅𝑧
∆𝑐𝐷 < 0
𝑧
𝑘∗
𝑘
𝑧∗
𝑘∗
𝑧∗
Impact of Import Competition on Investment Composition
Derivation of the Diff-in-Diff:
compare the investment composition in the following two scenarios
increase in import competition: c1D < c0D (open economy)
no increase in import competition: c1D = c0D (closed economy)
consider relative investments (in logs) for each of the two scenarios:
Scenario 1: c1D < c0D
[ln (k) − ln (z)]open
=
2
nh i
ln c0D − c ∗ − ln c1D − c ∗
4γr
4γr
− (c ∗ )2θ − ln
− ϕ (c ∗ )2θ
− ln
L0
L1 (1 − δ) ϕ
Scenario 2: c1D = c0D
4γr
4γr
[ln (k) − ln (z)]closed = −2 ln
− (c ∗ )2θ − ln
− ϕ (c ∗ )2θ
L0
L1 (1 − δ) ϕ
Impact of Import Competition on Investment Composition
subtracting provides the difference-in-differences equation:
open
[ln (k) − ln (z)]
− [ln (k) − ln (z)]
closed
= ln c0D − c ∗ − ln c1D − c ∗
|
{z
}
>0
identifies the shift in the relative composition of investments induced by
import competition
Prediction 1 Import competition increases the amount of short-term relative to
long-term investments.
Heterogeneous Investment Responses across Firms
Shift in the composition of investments depends on firm productivity c ∗ :
∂ ln c0D − c ∗ − ln c1D − c ∗
>0
∂c ∗
while all firms loose profits and market power, the relative decrease in
profits is larger for less productive firms
therefore: high-cost firms shift their composition of investments relatively
more towards short-term investments
empirically, we capture productivity by size
Prediction 2 Import competition increases the relative amount of short-term
investments less for larger firms.
Impact of Market Size on Investment Composition
Trade liberalization implies higher import competition but also larger
export markets
an increase in market size L1 > L0 generates additional demand such that
the marginal return on long-term investments increases
accordingly, firms shift investments towards the long-run
a market size effect can therefore offset the competition effect
We need to control for confounding effects through the market size effect
Identification and Data
Identification
Difference-in-Differences specification:
ln (Iisct ) = β0 + β1 × ln (ImpCompst ) × Short-Termc + X0isct ζ + λc + λit + εisct
Iisct denotes investments by a firm i in investment category c at time t
ImpCompst measures import competition in industry s and year t
Short-Termc reflects the depreciation rate of investment category c
X0isct is a vector of control variables
λc are investment type and λit firm-year fixed effects
β1 : effect of import competition on the relative composition of investments:
open
β1 = [ln (k) − ln (z)]
closed
− [ln (k) − ln (z)]
Prediction 1 : we expect import competition to change the composition of
investments towards short-term assets (β1 > 0)
Identification
Introducing firm heterogeneity yields a triple difference estimator:
ln (Iisct ) = β0 +β2 ×ln (ImpCompst )×Short-Termc ×Sizei +X0isct ζ +λc +λit +εisct
where Sizei is a measure of firm size
β2 measures the differential impact across the firm-size distribution
β2
=
n
o
open
closed
[ln (k) − ln (z)]
− [ln (k) − ln (z)]
large firm
n
o
open
closed
− [ln (k) − ln (z)]
− [ln (k) − ln (z)]
small firm
∗
∗
with c (small firm) > c (large firm)
Prediction 2 : we expect that import competition leads to a stronger
compositional effect towards short-term assets for smaller firms (β2 < 0)
Data
Population of stock listed manufacturing firms in the US (1995-2009)
manufacturing firms: primary US SIC code (2000 to 3999)
source: Compustat and Worldscope
sample: 4,428 firms
Short-Termc as in Garicano and Steinwender (2016):
use expenditures into different asset classes as measure of investments
rank investments according to their time to payoff, by
1
depreciation rates suggested by Garicano and Steinwender (2016)
2
simple ranking, following the ordering of depreciation rates
consider: Land (0%), Buildings (3%), Machines (12%), Transportation
(16%), R&D (20%) Computer (30%), Advertising (60%)
Measures of size
total assets, employment and sales
we use average firm size (1995-1999)
Data
Foreign competition (3-dig US SIC) following:
ImpCompst =
Impst
,
Prodst + Impst − Expst
Impst and Expst : value of total US imports and exports (UN Comtrade)
Prodst : value of US domestic shipments (NBER Manufacturing)
Alternative trade effects
export market: importance of exports relative to domestic consumption
degree of openness: total US imports and exports over domestic shipments
ExpMarketst =
Expst
Prodst + Impst − Expst
Opennessst =
Impst + Expst
Prodst
Data
Fixed effects λit
control for firm-specific demand shocks, credit shocks or technology shocks
as long as they do not affect short- and long-term investments differently
identification is based on variation within a firm, for a given year and across
investment categories
Interacted controls at the firm level
Uncertainty: annual standard deviation of daily stock returns
Finance: current ratio, external financial dependence, capital cost, recession
Interacted controls at the sectoral level
capital and skill intensity: capital stock per worker and share of
non-production worker wages in total compensation
total factor productivity and value added to control for sector specific
productivity and size
Empirical Results
Baseline Results
Dependent Variable: log(Investment)
(1)
(2)
(3)
(4)
0.0347***
(0.00886)
0.0330***
(0.00906)
0.0333***
(0.00903)
Panel A: Measure of Depreciation: Ordering
log(ImpComp) * Depreciation
0.0455***
(0.00813)
Panel B: Measure of Depreciation: Depreciation rate
log(ImpComp) * Depreciation
0.143
(0.0901)
0.252***
(0.0968)
0.237**
(0.0994)
0.248**
(0.0998)
Industry Controls * Depreciation
sd(Stock Return) * Depreciation
no
no
yes
no
yes
yes
yes
yes
Investment FE
Firm-Year FE
yes
yes
yes
yes
yes
yes
yes
yes
89,735
3,308
2,163
89,436
3,308
2,163
81,912
3,308
2,163
72,064
3,308
2,163
Observations
Firm Clusters
Industry-Year Clusters
Baseline Results
Specification (3) is our preferred specification: β̂1 = 0.033
β1 is positive and statistically significant, confirming Prediction 1
Increasing import competition by 10%, increases investments into short-term
category A by 0.33% more than investments into neighboring long-term
category B
Thought experiment to evaluate economic significance of our estimates:
we calculate the average depreciation rate in the sample
pie chart
we consider the average increase in import competition over the sample
period 1995-2009: 60% over the 15-year period
estimates provide us with the differential change in investments
we pin down the absolute change in one base-category (dln(Land) = 0)
the new portfolio of assets results in a new average depreciation rate
estimates suggest that the average increase in import competition has
reduced the lifespan of firm assets by 72 days on average
corresponds to 4.6% of the average asset lifespan
presuming a refinancing rate of 3%, this would impose an additional cost of
6$ for each 1000$ invested
Heterogeneous Investment Responses
Dependent Variable: log(Investment)
Employment
(2)
Measure of Size
Sales
(3)
Assets
(4)
0.0396***
(0.0101)
-0.000599
(0.000385)
-0.000736
(0.000950)
0.0403***
(0.00985)
-3.42e-06*
(1.74e-06)
-3.00e-06
(3.97e-06)
0.0401***
(0.00966)
-3.32e-06***
(1.22e-06)
-4.83e-06
(3.36e-06)
0.237**
(0.0994)
0.321***
(0.109)
-0.00647*
(0.00371)
0.00421
(0.0102)
0.335***
(0.106)
-3.86e-05**
(1.85e-05)
2.51e-05
(4.54e-05)
0.326***
(0.105)
-3.58e-05***
(1.23e-05)
-2.82e-06
(3.62e-05)
yes
yes
yes
yes
81,912
2,866
2,381
yes
yes
yes
yes
73,136
2,866
2,381
yes
yes
yes
yes
75,578
2,866
2,381
yes
yes
yes
yes
75,660
2,866
2,381
(1)
Panel A: Measure of Depreciation: Ordering
log(ImpComp) * Depreciation
0.0330***
(0.00906)
log(ImpComp) * Depreciation * Size
Depreciation * Size
Panel B: Measure of Depreciation: Depreciation rate
log(ImpComp) * Depreciation
log(ImpComp) * Depreciation * Size
Depreciation * Sales
Industry Controls * Depreciation
sd(Stock Return) * Depreciation
Investment FE
Firm-Year FE
Observations
Firm Clusters
Industry-Year Clusters
Heterogeneous Investment Responses
β2 is negative and statistically significant, confirming Prediction 2
Invoking an analogous thought experiment to evaluate economic
significance of our estimates:
we compare a firm at the 10th percentile with a firm at the 90th percentile
of the firm size distribution (in terms of assets)
we find that the lifespan of assets decreases by about 15 days more for the
small firm when compared to the large firm
Alternative Channels
1
Control for financial frictions
External financial dependence: capital expenditure net of cash flow, divided
by capital expenditure
Current ratio: total of current assets over current liabilities
Capital cost: capital expenditures over total liabilities
Financial crisis: dummy for years 2007-2009
2
Control for access to export markets
Alternative Channels
Dependent Variable: log(Investment)
(1)
(2)
(3)
(4)
(5)
0.0302***
(0.00904)
0.00963***
(0.00223)
0.0319***
(0.00919)
0.0313***
(0.00912)
0.0331***
(0.00906)
Panel A: Measure of Depreciation: Ordering
log(ImpComp) * Depreciation
0.0330***
(0.00906)
Current Ratio * Depreciation
External Dependence * Depreciation
0.000198**
(9.86e-05)
Capital Cost * Depreciation
-0.200***
(0.0233)
Crisis * Depreciation
-0.00795
(0.0187)
Panel B: Measure of Depreciation: Depreciation rate
log(ImpComp) * Depreciation
0.237**
(0.0994)
Current Ratio * Depreciation
0.233**
(0.0987)
0.00579
(0.0247)
External Dependence * Depreciation
0.229**
(0.100)
0.221**
(0.0989)
0.00159*
(0.000821)
Capital Cost * Depreciation
-2.116***
(0.290)
Crisis * Depreciation
Industry Controls * Depreciation
sd(Stock Return) * Depreciation
Investment FE
Firm-Year FE
Observations
Firm Clusters
Industry-Year Clusters
0.241**
(0.0993)
-0.260
(0.209)
yes
yes
yes
yes
81,912
3,358
2,441
yes
yes
yes
yes
79,083
3,358
2,441
yes
yes
yes
yes
78,956
3,358
2,441
yes
yes
yes
yes
78,963
3,358
2,441
yes
yes
yes
yes
81,912
3,358
2,441
Alternative Channels
Dependent Variable: log(Investment)
(1)
Panel A: Measure of Depreciation: Ordering
log(ImpComp) * Depreciation
0.0330***
(0.00906)
log(ExpMarket) * Depreciation
(2)
(3)
-0.0360***
(0.00908)
0.0767***
(0.0123)
-0.0814***
(0.0118)
log(Openness) * Depreciation
0.0146*
(0.00763)
Panel B: Measure of Depreciation: Depreciation rate
log(ImpComp) * Depreciation
0.237**
(0.0994)
log(ExpMarket) * Depreciation
-0.489***
(0.0994)
log(Openness) * Depreciation
Industry Controls * Depreciation
sd(Stock Return) * Depreciation
Investment FE
Firm-Year FE
Observations
Firm Clusters
Industry-Year Clusters
(4)
yes
yes
yes
yes
81,912
3,358
2,441
yes
yes
yes
yes
81,912
3,358
2,441
0.644***
(0.126)
-0.847***
(0.117)
0.0894
(0.0840)
yes
yes
yes
yes
81,912
3,358
2,441
yes
yes
yes
yes
81,912
3,358
2,441
A Quasi-Natural Experiment: China’s WTO Accession
Idea: Use China’s accession to WTO in 2001 as a quasi-natural
experiments, isolating an import competition effect
China’s annual TFP growth in manufacturing vastly exceeded TFP growth
in the US (8% vs 3.9%, see Autor et al. (2013))
Prospect of formal WTO accession stimulated restructuring of the Chinese
manufacturing industry (see Autor et al. (2016))
US imports from China vastly exceeded US exports to China and China has
a comparative advantage in industrial goods (see Autor et al. (2016))
⇒ Surge can be treated as mostly exogenous to the US market
⇒ China’s growth resulted in a large supply shock for manufacturing goods
Identification: Use average pre-trade-agreement level of tariffs to identify
firms affected by trade liberalization (see Guadalupe and Wulf (2010))
change in China’s WTO membership status led to a substantial reduction in
expected US imports tariffs on Chinese goods (see Pierce and Schott (2016))
A Quasi-Natural Experiment: China’s WTO Accession
New triple difference estimator:
ln (Iisct ) = β0 +β3 ×Post2000t ×Pre-WTO-Tariffs ×Short-Termc +X0isct ζ +λc +λit +εisct
where Pre-WTO-Tariffs is the average (1995-1999) effectively applied
import tariff in industry s (US SIC, 3 dig.)
the effectively applied tariff is defined as the lowest available tariff, given by
preferential tariffs if existent and MFN tariffs otherwise
we restrict sample period to the years 1999 to 2003 (WTO entry: 2001)
β3 measures increase in relative short-term investments after the Chinese
WTO accession in 2001, comparing high-tariff to low-tariff firms
we expect β3 > 0 (Prediction 1 )
A Quasi-Natural Experiment: China’s WTO Accession
Dependent Variable: log(Investment)
(1)
(2)
(3)
(4)
0.00636**
(0.00290)
-0.0359***
(0.0134)
0.00489
(0.00344)
0.00550*
(0.00301)
-0.0241*
(0.0144)
0.00541
(0.00352)
0.00539*
(0.00300)
-0.0219
(0.0139)
0.0181***
(0.00371)
0.00430
(0.00313)
-0.0123
(0.0148)
0.0184***
(0.00378)
0.0642**
(0.0272)
-0.642***
(0.146)
0.105***
(0.0327)
0.0548*
(0.0282)
-0.577***
(0.154)
0.104***
(0.0337)
0.0549**
(0.0270)
-0.500***
(0.147)
0.134***
(0.0351)
0.0447
(0.0282)
-0.376**
(0.157)
0.155***
(0.0364)
no
no
yes
yes
30,949
2,379
no
yes
yes
yes
27,804
2,379
yes
no
yes
yes
30,517
2,379
yes
yes
yes
yes
27,428
2,379
Panel A: Measure of Depreciation: Duration rank
Post2000 * Pre-WTO-Tariff * Depreciation
Post2000 * Depreciation
Pre-WTO-Tariff * Depreciation
Panel B: Measure of Depreciation: Depreciation rate
Post2000 * Pre-WTO-Tariff * Depreciation
Post2000 * Depreciation
Pre-WTO-Tariff * Depreciation
Industry Controls * Depreciation
sd(Stock Return) * Depreciation
Investment FE
Firm-Year FE
Observations
Firm Clusters
A Quasi-Natural Experiment: China’s WTO Accession
using the results from specification (3) in panel B
we compare 2 firms, one at the 25th percentile and one at the 75th
percentile of the pre-2000 tariff distribution
we find that after 2001, the average investment duration increased by
roughly 106 days less for a firm at the 75th percentile of the distribution
Further Robustness
1
Alter/omit/bundle investment categories
2
check whether results hinge on the specific ordering of categories
Further Robustness
Dependent Variable:
log(Investment)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
0.0330***
(0.00906)
0.0194**
(0.00918)
0.0236*
(0.0128)
0.0334***
(0.00935)
0.0634***
(0.0166)
0.104***
(0.0281)
0.0896***
(0.0273)
-0.0133
(0.00870)
Industry Controls * Depreciation
sd(Stock Return) * Depreciation
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
Investment FE
Firm-Year FE
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
none
R&D
R&D/
Advertising
Transportation/
Computer
none
none
none
none
Measure of Depreciation: Ordering
log(ImpComp) * Depreciation
Excluded categories
Number of categories
7
6
5
5
4
3
3
7*
Observations
Firm Clusters
Industry-Year Clusters
81,912
3358
2441
58,441
3358
2441
49,572
3358
2441
76,822
3358
2441
81,912
3358
2441
81,912
3358
2441
81,912
3358
2441
81,912
3358
2441
The ordering of categories resembles the ordering of depreciation rates in specification (1)-(4). Specification (5) groups Land, Buildings and
Machinery into one category and R&D and Computer into another. Specification (6) additionally takes Transportation into the category with
Land, Buildings and Machinery, while specification (7) takes it into the category with R&D and Computer. In specification (8), R&D is ordered
as the most long term investment. Investment expenses are either derived from balance sheet data on assets (Land, Buildings, Machines,
Transportation and Computer) or taken from the income statement (R&D and Advertising).
Thank you!
Average Shares of Investment Categories in Total
Investments
Transportation
Land
Buildings
R&D
Computer
Advertising
Machines
back
References
References I
Aghion, Philippe, George-Marios Angeletos, Abhijit Banerjee, and Kalina
Manova. Volatility and Growth: Credit Constraints and the Composition of
Investment. Journal of Monetary Economics, 57(3):246–265, 2010.
Autor, David H., David Dorn, and Gordon H. Hanson. The China
Syndrome: Local Labor Market Effects of Import Competition in the United
States. American Economic Review, 103(6):2121–68, 2013.
Autor, David H., David Dorn, and Gordon H. Hanson. The China Shock:
Learning from Labor Market Adjustment to Large Changes in Trade. Working
Paper 21906, National Bureau of Economic Research, 2016.
Bénabou, Roland and Jean Tirole. Bonus Culture: Competitive Pay,
Screening, and Multitasking. Journal of Political Economy, 124(2):305–370,
2016.
Bernard, Andrew B., J. Bradford Jensen, and Peter K. Schott. Survival of
the Best Fit: Exposure to Low-Wage Countries and the (Uneven) Growth of
U.S. Manufacturing Plants. Journal of International Economics,
68(1):219–237, 2006.
References
References II
Bloom, Nicholas, Mirko Draca, and John Van Reenen. Trade Induced
Technical Change? The Impact of Chinese Imports on Innovation, IT and
Productivity. Review of Economic Studies, 83(1):87–117, 2016.
Bustos, Paula. Trade Liberalization, Exports, and Technology Upgrading:
Evidence on the Impact of MERCOSUR on Argentinian Firms. American
Economic Review, 101(1):304–340, 2011.
Fresard, Laurent. Financial Strength and Product Market Behavior: The Real
Effects of Corporate Cash Holdings. Journal of Finance, 65(3):1097–1122,
2010.
Garicano, Luis and Luis Rayo. Why Organizations Fail: Models and Cases.
Journal of Economic Literature, 54(1):137–92, 2016.
Garicano, Luis and Claudia Steinwender. Survive Another Day: Using
Changes in the Composition of Investments to Measure the Cost of Credit
Constraints. Review of Economics and Statistics, forthcoming, 2016.
Guadalupe, Maria and Julie Wulf. The Flattening Firm and Product Market
Competition: The Effect of Trade Liberalization on Corporate Hierarchies.
American Economic Journal: Applied Economics, 2(4):105–27, 2010.
References
References III
Lileeva, Alla and Daniel Trefler. Improved Access to Foreign Markets Raises
Plant-level Productivity ... For Some Plants. Quarterly Journal of Economics,
125(3):1051–1099, 2010.
Melitz, Marc J. and Gianmarco I. P. Ottaviano. Market Size, Trade, and
Productivity. Review of Economic Studies, 75(1):295–316, 2008.
Pierce, Justin R. and Peter K. Schott. The Surprisingly Swift Decline of US
Manufacturing Employment. American Economic Review, forthcoming, 2016.
Terry, Stephen J. The Macro Impact of Short-Termism. Discussion Papers
15-022, Stanford Institute for Economic Policy Research, 2015.
Valta, Philip. Competition and the Cost of Debt. Journal of Financial
Economics, 105(3):661–682, 2012.
Xu, Jin. Profitability and Capital Structure: Evidence from Import Penetration.
Journal of Financial Economics, 106(2):427–446, 2012.