Foreign Direct Investment, Exports and Aggregate Productivity
Joel Rodrigue∗
Department of Economics, Vanderbilt University, Nashville, TN, United States
Abstract
This paper presents and estimates a model of foreign direct investment (FDI) and exports with
heterogeneous firms. The model highlights the interaction between firms’ location and export
decisions and their effect on aggregate productivity and welfare. The model is estimated using
plant-level Indonesian manufacturing data. The results are broadly consistent with the pattern
of productivity, exports and FDI across plants. Counterfactual experiments suggest that there
are substantial productivity gains due to international trade and FDI. They emphasize that the
nature of firm-level trade and FDI interactions is crucial to determining the aggregate effects of
trade and FDI policy across industries.
KEYWORDS: exports; foreign direct investment; firm heterogeneity; aggregate productivity; resource allocation; Indonesia
∗
Contract Information.
Mailing Address: Department of Economics, Vanderbilt University, VU Station B #351819, 2301 Vanderbilt Place, Nashville, TN 37235-1819; Tel.: +1 615 322 2871; fax: +1 615 343 8495.
E-mail address: [email protected]
1. Introduction
It is well known that there are large differences in total factor productivity across developed
and developing countries. Recent research by Restuccia and Rogerson (2008) and Hsieh and
Klenow (2009) suggest that the misallocation of resources across firms can have important effects on aggregate total factor productivity, particularly in developing economies. Likewise, a
number of papers have provided estimates of the gains associated with trade in goods (e.g Eaton
and Kortum, 2002; Arkolakis, Costinot, and Rodrı̀guez-Clare, 2010), multinational production
(e.g. McGrattan and Prescott, 2009) or both (Ramondo and Rodrı̀guez-Clare, 2010; Irarrazabal,
Moxnes, and Opromolla, 2010) across heterogeneous and countries.1 We add to this literature
by re-examining this question in a developing country context where we can observe the impact
that the entry and exit of foreign firms have on the reallocation of resources and aggregate
productivity in specific industries. This paper builds a model of international trade and foreign
direct investment (FDI) with heterogeneous firms. The model highlights how differences in production, location and export costs affect the structure of international trade. Using Indonesian
plant-level manufacturing data, the model is structurally estimated and used to perform counterfactual experiments that assess the positive and normative effects of barriers to trade and
FDI.
This paper is motivated by a set of stylized facts regarding the nature of trade and multinational production across heterogeneous firms. First, as has been often reported, exporting firms
and multinational tend to be much more productive than domestic firms.2 Second, empirical
studies have repeatedly confirmed that trade can substantially impact resource allocation across
firms and aggregate productivity. Pavcnik (2002) and Trefler (2004) find that increasing trade
openness induces productivity gains among exporters in Chile and Canada, while Pavcnik (2002)
also finds increases in aggregate Chilean productivity. Third, recent evidence suggests that the
volume and direction of trade are intimately related to multinational production decisions. For
instance, it is estimated that total sales from multinational firms accounts for 60% of world GDP
and over 35% of world trade in 2001.3
Despite the growth in research examining FDI and exporting in an environment with heterogeneous firms, few studies have empirically estimated the role of FDI on resource allocation,
trade flows and their effect on aggregate productivity. Further, while there are a number of
results which study the effects of trade liberalization on the entry and exit decisions of domestic firms, few papers disentangle the effect of trade or FDI policy change on foreign entry and
1
A subset of recent papers which also attempt to measure the gains from trade in trade-only models would
include Alvarez and Lucas (2007), Fieler (2011) and Waugh (2010). Similarly, Ramondo (2011) and Burstein
and Monge-Naranjo (2009) make important contributions to the literature evaluating the gains to foreign direct
investment in models with foreign direct investment alone.
2
See Helpman, Melitz and Yeaple (2004) or the results in Section 5.
3
See Ramondo (2011).
1
exit decisions or provide a quantitative estimate changes in aggregate productivity induced by
policy-led reallocation. This paper adds to this literature by structurally estimating a model
heterogeneous firms with FDI and exporting. The estimated model captures productivity differences across foreign and domestic firms in terms of their entry, exit and export decisions
and recover estimated costs in dollar values associated with both exporting and FDI. Finally,
using the structurally estimated parameters we use counterfactual experiments in order to assess the implications of changes in FDI or trade policy separately on movements in aggregate
productivity in a developing country.
Recently there have been a number of papers re-examining the impact of foreign direct
investment (FDI) on productivity growth. McGrattan and Prescott (2007) and Ramondo and
Rodriguez-Clare (2010) study and quantify the impact of foreign direct investment on aggregate
productivity growth within a country and its welfare implications. In both cases they find that
inward foreign capital flows can potentially induce large increases in aggregate productivity and
welfare. Their results are particularly striking when the presence of foreign firms encourages
technology and productivity upgrading among domestic firms. This paper will similarly reexamine the impact of inward FDI on aggregate productivity and welfare, exploiting observed
productivity differences across domestic and multinational firms. However, in contrast to the
above literature it focuses on the role of resource reallocation in response to changes in the trade
or FDI policy environment.4
A separate but related set of empirical evidence suggests that among firms producing in a
given country, multinational firms are the most productive.5 This paper confirms these productivity differences across foreign and domestic firms in Indonesia and, based on its empirical
findings, extends the international trade and FDI framework of Helpman, Melitz and Yeaple
(2004) by allowing firms to offshore production in a foreign country. In this environment, firms
may set up plants in foreign countries for two reasons. First, as in Melitz, Helpman and Yeaple
(2004), firms can set up plants solely to access the local market in foreign countries.6 Relative
4
The paper also abstracts from the notion of knowledge or productivity spillovers from foreign to domestic
firms. The first reason for this assumption is that due to the short time period we examine it is unlikely that we
will be able to identify significant knowledge spillovers across firms. Secondly, among studies that have examined
this feature of micro-data there is little empirical evidence for within-industry spillovers; see Rodrik (1999), Aitken
and Harrison (1999), Javorcik (2004), Ramondo (2009) and the references therein. Most importantly, Blalock and
Gertler (2008) examine horizontal productivity spillovers using the same data and conclude that “none of [their]
models provide any evidence of horizontal technology transfer” between foreign and domestic firms. Downstream
spillovers from FDI, as investigated by Javorcik (2004) and Blalock and Gertler (2008), require a model of interindustry production which is beyond the scope of this paper.
5
Helpman, Melitz and Yeaple (2004) and Arnold and Javorcik (2005) compare multinationals, exporters and
non-exporters. Roberts and Tybout (1997), Clerides, Lach and Tybout (1998), Bernard and Jensen (1999), Aw,
Chung, and Roberts (2000), Bernard et. al. (2003) and Eaton, Kortum, and Kramarz (2004) compare exporters
and non-exporters.
6
Both Chor, Foley and Manova (2008) and Irarrazabal, Moxnes, and Opromolla (2010) recently report that
the majority of multinational affliate sales are intended for the destination market and conclude that market
access is a key motive for foreign direct investment. Earlier results can be found in Brainard (1997), Markusen
2
to a model without FDI, such as Melitz (2003), FDI provides an additional avenue for firms
to make sales to foreign consumers. Second, in this model firms can also set up plants in a
foreign country in order to export back to the country of origin.7 Thus, this paper constructs a
parsimonious model that simultaneously allows firms to engage in FDI for both market access
and export-platform production. The advantage of this approach is that all of the model’s predictions can be readily tested. Moreover, the model is used to empirically assess the influence
of policy on plant-level decisions, and thus provides a framework for evaluating economic policy
and predicting changes in aggregate productivity, exports and FDI.
In a closely related paper Irarrazabal, Moxnes and Opromolla (2010) structurally estimate
a model of multinational production and trade across OECD countries using multinational production decisions by Norwegian multinational firms. Their model also extends the Helpman,
Melitz and Yeaple (2004) model to examine the nature of FDI across a wide set of countries
whereas this paper focusses on inward FDI within a developing country. In particular, Irarrazabal, Moxnes and Opromolla (2010) present a model that seeks to explain the gravity of FDI
across geographically distant countries, while we are interested solely in the impact of inward
FDI to Indonesia. A key difference between this paper and similar papers which quantify the
gains from openness to trade or FDI, is that this paper provides specific estimates for individual
industries in a developing country context. While McGrattan and Prescott (2009), Ramondo and
Rodriguez-Clare (2010) and Irarrazabal, Moxnes, and Opromolla (2010) all provide insightful
estimates for the importance of trade and FDI flows across countries, these studies are not able
to distinguish the differential impact of these flows across industries. In particular, this study is
able to provide disaggregated estimates of the gains from FDI or trade and demonstrates that
the observed differences in gains from trade or FDI depend crucially on industrial differences
trade costs, FDI costs and the distribution of productivity across industries. Moreover, we are
able to recover export and FDI entry costs and examine how changes in specific policy related
parameters may influence resource reallocation and aggregate productivity across heterogeneous
firms in this specific context.
Another important related contribution would include the work of Feinberg and Keane (2006)
who build a structural model of U.S. multinational corporations and Canadian affiliates.8 Although their model only captures the degree to which the existing multinational firms engage in
export activities, they find that technical change, particularly improvements in logistics, have
an important effect on the export decisions of multinational firms.
Last, this paper is also related to papers which investigate whether FDI and trade are substitutes or complements at the firm-level. Again, the evidence here is mixed, though most studies,
and Maskus (1999), Markusen and Maskus (2002) and Blonigen, Davies and Head (2003).
7
See Antras and Helpman (2004), Grossman, Helpman and Szeidl (2006) and Garetto (2010) for examples.
8
Surprisingly, they report that many U.S. multinational and affiliates engage in interfirm trade, which is similar
the patterns observed in the Indonesian context.
3
such as Svensson (1996), Clausing (2000) and Head and Ries (2001), find that FDI and home activity are complements which is consistent with the model presented in this paper. In contrast,
using product-level data from the Japanese auto industry, Blonigen (2001) finds substantial
evidence affiliate production and exports can act as either substitutes or complements.
In contrast to much of the preceding FDI literature, the model presented in this paper is
estimated using a panel of Indonesian manufacturing plants. The estimated model captures
the pattern of productivity, exports and market share across firms with different ownership and
export status. The results document that the mean of domestic productivity at the steady state is
substantially higher than the estimated mean at entry, indicating that endogenous exit decisions
play an important role in determining aggregate productivity. Moreover, this paper performs a
number of counterfactual policy experiments demonstrating that restrictions on FDI and trade
can potentially have widely different effects on aggregate productivity. In industries where many
multinational firms use Indonesia as a location for exporting we observe that restricting trade
has a very large impact on aggregate productivity since restricting trade also discourages FDI.
In contrast, when the primary reason for FDI is access to the domestic Indonesian market we
observe smaller declines in aggregate productivity in response to heightened trade restrictions
since many foreign producers choose to stay in Indonesia despite the change in trade policy.
Failing to account for these differences will lead to biased estimates of the impact of trade
restrictions. Last, the welfare impact on Indonesia of simultaneous trade and FDI restrictions
are estimated to be approximately 1 percent across industries. However, when trade or FDI are
individually restricted, the welfare impacts are often very small since trade (FDI) flows provide
some insurance against FDI (trade) restrictions.
The next section outlines the Indonesian policy environment faced by foreign and domestic
producers. Section 3 presents a model of exports and FDI with heterogeneous firms and countries, while Sections 4 and 5 describe the estimation methodology and the data used to estimate
the model. Section 6 presents empirical results and Section 7 concludes.
2. Foreign investment and trade policies in Indonesian manufacturing
Indonesia’s manufacturing sector is an attractive setting for research examining the plantlevel trade and FDI decisions. With the fourth largest national population worldwide and a rich
endowment of natural resources the country supports a large array of domestic and foreign-owned
manufacturing firms across a wide set of industries. Due to the large changes in trade policy
over a period of time when the Indonesian government has carefully collected an exceptionally
rich data set on plant-level manufacturing activity, it is not surprising that a number of scholars
have recently chosen to study the causes and consequences of international trade in Indonesia:
Amiti and Konings (2007), Bernard and Sjöholm (2003), Blalock and Gertler (2004), and Amiti
and Davis (2008) are only a subset of recent examples.
4
During the 1970s and early eighties Indonesia pursued an explicit import substitution policy. In the mid-eighties Indonesia began to shift from a policy of import substitution to export
promotion. In particular, by 1986 there were a number of policies in place aimed at reducing import tariffs, reforming customs administration and introducing a more generous duty-drawback
scheme.9 Despite the initial policies changes, in many cases, the tariffs rates and protection
remained high well in the 1990s. In 1995 Indonesia became a member of the World Trade
Organization agreeing to reduce all bound tariffs to 40 percent or less over a ten year period
beginning that year.10
Similarly since the 1960s government regulation of foreign direct investment has shifted from
a policy of discouraging inward flows of investment from abroad to one which actively encourages inward FDI (Wie, 1994; Hill, 1998; Pangestu, 1996). After achieving independence from
the Netherlands in 1945 the Sukarno government nationalized many of the formerly Dutch manufacturing firms and continued to discourage inward foreign investment through weak property
rights and the threat of further nationalization. The “New Order” economic reform package of
the Suharto government instituted gradual reforms which allowed foreign investment in most
industries but still required large minimum levels of initial and long-term Indonesian ownership
in new ventures. When oil prices fell sharply in the mid-1980s the government pursued foreign investment more actively by providing exemptions to the foreign investment policy in the
island of Batam (closed region to Singapore), underdeveloped eastern provinces, government
sponsored research parks, and industries which were capital-intensive, technology-intensive or
export-oriented. Finally, in 1994, the government lifted nearly all equity restrictions on foreign investment. With few exceptions, multinational enterprises are allowed to establish and
maintain operations with 100 percent of equity held by foreign investors.
3. A Model of FDI and Exports
3.1. Environment
There are two countries, Home and Foreign, which are endowed with non-depreciating stocks
of labour, L. All foreign country variables are starred. Consumers supply labour inelastically.
Their preferences are defined by a Cobb-Douglas utility function over the agriculture product
R
δ/α
z and a continuum of manufactured goods indexed by v: U = z 1−δ v∈V q(v)α dv
. The
elasticity of substitution between different varieties of manufactured goods is given by ε =
1/(1 − α) > 1.
In each country there are two sectors: a homogeneous good sector (agriculture) and a differentiated good sector (manufacturing). The agricultural sector is characterized by a continuum
9
See Blalock and Gertler (2004) for details.
The policy was subject to an exclusion list of products for which it did not apply. A list of the excluded
products is available at: http://www.wto.org/english/tratop e/schedules e/goods schedules e.htm
10
5
of potential firms that can enter freely and produce a homogeneous agricultural product, z, with
linear technology, z = φl l. Producers hire labour on perfectly competitive markets which pins
down the relative wage across countries. It is assumed that the foreign agricultural technology
is more productive than the technology employed in the home country, φ∗l > φl , so that wages
in the foreign country are greater than those in the home country.
The manufacturing sector is the primary focus of this paper. It is characterized by a continuum of monopolistically competitive firms that produce horizontally differentiated goods.
Individual manufacturing firms are indexed by i and the endogenous measure of manufacturing
producers be denoted by M .
There exists an unbounded measure of ex-ante identical potential entrants. To enter the
market each potential entrant must pay a sunk cost, fe . Once the entry cost is paid, each firm
receives a productivity draw a from the distribution, Ga (a), and an extreme cost shock with
constant probability ξ. A firm’s productivity level is constant over the life of the firm. If the
firm suffers the extreme cost shock it is forced to exit the industry. Conditional on survival, each
firm can decide to exit immediately or to produce according to the linear production function
q = l/a where l is labour hired on competitive markets. To keep the model as transparent
as possible, this paper focuses on the case in which both foreign and domestic exporters and
non-exporters are present. Each country has four types of firms: non-exporters, exporters,
multinational non-exporters and multinational exporters.
Firms differ with respect to their productivity level a, their current FDI and export status and their past FDI and export status. Denote a firm’s decision by dit ≡ (dfitd , dxit ) ∈
{(0, 0), (0, 1), (1, 0), (1, 1)} where dfitd is the firm’s decision to maintain a plant abroad and dxit is
the firm’s decision to export. If a firm decides to engage in FDI or export that decision takes a
value of 1, and is 0 otherwise. All new entrants are assumed to enter the market without any
FDI or export experience, dit = (0, 0).
To produce domestically, each firm must pay a non-stochastic fixed overhead cost f each
period. If the firm also decides to export abroad it bears an additional one-time suck cost
cx , per-period fixed cost fx and iceberg transport costs τ > 1 per unit shipped to the foreign
country. As in Helpman, Melitz and Yeaple (2004) the firm may choose to set up production
in the foreign country rather than export to that market. While the firm saves on the export
costs, cx and fx , and on the transport costs, τ , by choosing to produce abroad, it incurs the
an additional fixed overhead cost, ff d and a one-time sunk cost cf d associated with the initial
plant set-up abroad.
Firms may choose not to produce domestically at all. In this case, each firm sets up a plant
abroad and exports back to its country of origin. The firm then incurs the fixed costs f , fx and
ff d and the transport cost τ each period, along with the sunk costs cx and cf d in the initial
period. If cf d > 0 and ff d > 0, any firm that produces abroad to export back home incurs
higher fixed and sunk costs and the same transport costs of a home exporter. Thus, to give
6
firms an incentive to produce abroad and export there must be some difference in factor prices
(wages) across countries.
Given a firm’s current and past FDI and export decisions they incur the following set of
fixed and sunk costs in a given period:


f



 f + f + c (1 − dx )
x
x
i,t−1
F (dit , di,t−1 ) =
∗
∗ + c∗ (1 − df d )

f
+
f
+
f

i,t−1
fd
fd


 f ∗ + ζ f ∗ + f ∗ + [ζ c∗ (1 − dx
fx x
fd
i,t−1 )
cx x
for (dfitd , dxit ) = (0, 0)
for (dfitd , dxit ) = (0, 1)
for (dfitd , dxit ) = (1, 0)
d
+ c∗f d ](1 − dfi,t−1
) for (dfitd , dxit ) = (1, 1)
where starred variables denote costs incurred in the foreign country. The sunk and fixed export
costs are interpreted as forming and maintaining distribution and service networks in export
markets. It is important to note that firms must pay this cost even if they are exporting back to
their home market, but the sunk and fixed cost may be reduced through the parameters ζcx and
ζfx . These parameters capture the complementarity between foreign ownership and exporting
to foreign markets. The fixed and sunk FDI costs include the distribution, servicing, overhead
and set up costs of operating a subsidiary abroad.
The size of the export and FDI costs plays an important role in determining each firm’s
optimal production and export decisions. In turn, these decisions influence productivity and
welfare outcomes across countries. For instance, if foreign firms must incur large sunk costs
in order to produce in Indonesia, highly productive firms from abroad will choose to produce
elsewhere. As such, larger amounts of production will be allocated to less productive domestic
firms and aggregate productivity will be consequently reduced. A key benefit of our approach
is the ability to pin down the magnitude costs in a context where they are likely to have a
particularly important role in determining aggregate outcomes.
Regardless of the magnitude of FDI or export costs, it is necessary to allow for stochastic
elements in the model. As noted by Rust (1987) this stochastic element of the model is necessary
in order to generate a positive likelihood for some of the observed behaviour in the data. For
example, without these shocks the model would be inconsistent with the large number of small
firms observed in the data. We interpret these shocks as capturing differences the fixed and sunk
costs associated with FDI and exporting across firms as in Das, Roberts and Tybout (2007).
We will denote the stochastic of fixed production costs, export and FDI costs as χit , xit and fitd ,
respectively. The production, export and FDI cost shocks are all independently drawn from
extreme-value distributions with the respective scale parameters %χ , %x and %f d .11
It is well known that the Dixit-Stiglitz (1977) framework generates a demand function
δRP 1−ε /pε for each variety where R is the total revenue earned in each country, P is an index of
11
We allow for variation in these draws across foreign and domestic firms. See the Appendix for details.
7
manufacturing prices and p is the price chosen by each individual producer.12 Since demand is
exogenous to each individual producer the optimal pricing rule for each firm depends only on the
firm-specific productivity level p(a) =
aw
α
where w is the wage in the home country. The revenue
earned by a plant with productivity level ai producing for market j = h, f can be written as
j
rit
= (mci B j )1−ε
(1)
where mci = w/ai if the firm is not exporting to market j and mci = wτ /ai if the firm is
exporting to market j and w is wage in the country where production occurs.
3.2. FDI & Exports
In order to build intuition we first consider a simplified framework with only fixed (perperiod) costs and no sunk costs.13 We require that each productive plant in Indonesia always
serves the Indonesian market since casual empiricism suggests that few firms produce solely for
export markets.14 In equilibrium, I rule out the possibility that multinational firms produce
solely for the foreign market or that any firm will serve the same market by both exports and
FDI.15
Consider the production possibilities facing a firm that originates in Indonesia.16 The operating profits for any firm with productivity level a producing solely for domestic consumption
are
π(a) = (awB)1−ε − f,
(2)
where B = [(1 − α)γwL]1/(1−ε) /(αP ). Indonesian exporters earn both profits from export sales
in addition to those earned at home.
πx (a) = π(a) + (τ awB ∗ )1−ε − fx .
(3)
Similarly, firms that produce in both countries for each market earn the additional operating
profits from their plants abroad. The operating profits for a multinational firm which does not
export17 are
πmn (a) = π(a) + π ∗ (a) − ff d .
(4)
Lastly, consider multinationals which do not produce at home for the home market, but instead
12
RR = wL where L is the total amount of labour and w is the wage in the home country. The price index is
P = [ V p(v)1−ε dv]1/(1−ε)
13
Alternatively, we can consider a scenario where sunk costs are amortized over time and built into annual fixed
payments.
14
It would be possible to generate production solely for export by allowing fixed costs to vary across countries.
15
Rob and Vettas (2001) use uncertainty in a dynamic setting to explain this phenomenon.
16
For notational convenience we implicitly assume that the fixed costs are identical across countries. We will,
however, revisit this issue when estimating the model.
17
We might think of this as horizontal FDI as in Helpman, Melitz and Yeaple (2003).
8
produce abroad for both home and foreign markets.18 The operating profits for an exporting
multinational are
πmx (a) = πx∗ (a) − ff d .
(5)
Equation (??) implies that, conditional on productivity, the firm which chooses to use the foreign
country as an export platform will earn similar profits to an exporter in that country up to the
fixed foreign direct investment cost. An identical set of equations apply to firms which originate
in the foreign country with country indicators reversed.
Equations (??)-(??) indicate that export and investment decisions will depend primarily
on both firm-specific characteristics (productivity) and the differences across countries (wages,
fixed costs, size). Suppose that the foreign country is a large, high-wage country, while the home
country is relatively small and characterized by low-wages. Consider the export and investment
decisions facing foreign country firms. The first decision is whether to produce domestically
and export abroad versus producing abroad and exporting back to the country of origin. This
decision will clearly depend upon the cost of exporting in either direction. However, even
if transport costs are equal there will be substantial differences in these two production and
export decisions. When transport costs are equal, the advantage of exporting from the foreign
country to the home country (Indonesia) is that the foreign firm incurs lower fixed overhead
costs. The disadvantage is that labour costs are high.19 Thus, foreign firms that produce in the
home country must be productive enough to afford the higher fixed costs.
Similarly, consider the foreign firm deciding whether to produce all units in the home country
and export back to the foreign country versus opening a plant in each country. By producing
all units in the home country the firm incurs the lowest marginal costs on each unit of output
and saves the extra fixed costs from operating multiple plants. However, by producing all units
in the home country the firm incurs the transport cost on each unit exported.
The model has several intuitive implications. First, since exporting is costly, firms that export
must be productive enough to afford these extra costs. Second, higher fixed costs associated with
investing abroad suggests that multinational firms will be more productive than their domestic
counterparts. Finally, if country size is equal across countries, only firms from the foreign (high
wage) country will invest in the home (low wage) country and export back to the foreign country.
In other words, the model predicts that vertical multinationals will originate in the high wage
country and produce in the low wage country. This does not mean that FDI only flows from the
foreign country to the home country. Highly productive home firms may still want to invest in
the foreign country in order to save the transport costs from exporting. It will never, however,
be profit maximizing for home firms to use the foreign country as an export platform for home
18
In contrast to the preceding return function we may think of this as a type of vertical multinational structure.
Country size also plays an important role here. Firms will be more likely to enter larger markets by FDI,
ceteris paribus. However, to focus on the effect of factor prices I abstract from country size differences here.
19
9
country sales.
It is well known that multinational firms are more productive than domestic exporters and
that domestic exporters are more productive than domestic non-exporters.20 We will present
evidence in section 5 which is consistent with the suggestion that there is no clear productivity
ranking across foreign exporters and non-exporters. Following Melitz it is possible to derive
fixed cost bounds that match either scenario. We focus on two sets conditions that must hold
in order for multinational non-exporters to be more productive than multinational exporters,
(C1a)-(C3a), or for the opposite to be true, (C1b)-(C3b).21
fx
(wτ B ∗ )1−ε
<
h
1−ε
ff d
(wBµ )
− (w∗ B ∗ µ∗ )1−ε
ff d
<1−
fx − f
wBµ
w ∗ B ∗ µ∗
τ wf < wh
wτ 1−ε
fx
<
f + ff d
w∗
(C1a)
1−ε
ff d
<1−
fx − f
(C2a)
w ∗ B ∗ µf
wBµ
w < τ w∗
(C3a)
(C1b)
1−ε
(C2b)
(C3b)
The conditions (C1a) and (C1b) imply that the fixed costs of production abroad must be large
enough to ensure that not all exporting firms are more profitable by offshoring production. The
second set of conditions, (C2a) and (C2b), provide restrictions on the fixed costs in relation to
wages, transport costs and country sizes which distinguish which type of organizational structure
is profit maximizing. Condition (C2a) is required for non-exporting multinational to be more
productive than exporting multinationals, while (C2b) is the condition required for the opposite
ranking to hold.22 Note that the second term in the RHS of each condition is the inverse of the
other. Conditional on country size, increases in transport costs will discourage export platform
operations in either case. That is, among firms which have lower transport costs we would be
more likely to observe foreign firms using Indonesia as an export platform irrespective of the
fixed cost ranking. The last set of conditions are required in order to for all four types of firms
to simultaneously exist; conditional on transport costs, these conditions are necessary for no one
type of multinational production to be dominant for all firms.
There are a number of other production and export possibilities that I have purposely omitted
here. For instance, it is possible that foreign investors may open plants in the home country
exclusively for sales to the home country consumers without operating a plant in any other
country. As noted in Ramstetter and Sjoholm (2006) almost all of the foreign-owned plants in
20
See Helpman, Melitz and Yeaple (2004), for example.
In either case, domestic exporters are restricted to be more productive than domestic non-exporters and
multinational firms are restricted to be more productive than domestic firms. Since these bounds are very similar
to those reported in Melitz (2003) and Helpman, Melitz and Yeaple (2004) we relegate them to the Supplemental
Appendix.
1
1−ε
1−ε
22
w∗ τ
The parameter µ is defined as µ ≡
−
1
and µ∗ is identical with country superscripts reversed.
w
21
10
Indonesia are part of larger multinational firms.23 Table 1 documents the production and sales
combinations that I have purposely restricted from the set possibilities by placing bounds on
the fixed and transport costs. The first row of Table 1 shows that firms can locate in the home
country and produce only for the home country or both markets, but not just the foreign market.
As in Melitz (2003) I exclude this possibility since less than two percent of all Indonesian plants
receive all revenues from export sales. Similarly, in the second row I exclude all foreign plants
that receive all revenues from export sales. Although seven percent of foreign firms strictly
export all output, it is likely that these plants do so to protect technological advantages, patents
or product features from local rivals.24 I exclude this possibility since it is not observable in
the Indonesian manufacturing data.25 The last row simply indicates that if a firm produces in
both countries, then it must sell in both countries. Since firms must pay all of the fixed costs to
produce in a given country and the Dixit-Stiglitz framework implies some residual demand for
each variety, selling to local consumers can always increase profits.26
Table 1: Production and Sales Combinations
Production
Sales Location
Location
Home
Foreign
Both
Home
Yes∗
No
Yes
Foreign
No
Yes∗
Yes
Both
No
No
Yes
Only if the ownership is the same as the country where production is located.
3.3. Exit & Entry Decisions
Consider the dynamic component of a manufacturer’s decision problem. At the beginning of
each period, a producer’s state is given by their productivity level ai and their previous FDI and
export decisions, di,t−1 . The Bellman equations for a producer originating in the home country
23
Further, the data does not reveal separate plants which are multinationals and those which stand-alone but
have foreign ownership. However, Indonesia is a small country relative to its major sources of FDI (Japan, the
United States and Europe). If foreign investors are able to profitably operate a plant which sells exclusively to
Indonesian consumers, it is likely they are also able to the same in their home countries.
24
Since the large majority of plants (93%) have positive amounts of domestic sales it is not likely that these
firms are subject to the same restrictions.
25
See Antras and Helpman (2004) for a model with incomplete contracts and vertical integration.
26
Over 98% of Indonesian producers and 93% of foreign producers sell to the local Indonesian market.
11
are written as follows:
Z
V (ai , di,t−1 ) =
max{χ (0), W (ai , di,t−1 ) + χ (1)}dH χ (χ ),
(6)
Z
max{J(ai , di,t−1 , 0) + f d (0), J(ai , di,t−1 , 1) + f d (1)}dH f d (f d ) (7)
Z x0
0
x x0
J(ai , di,t−1 , 0) =
max
π(ai , di,t−1 , d ) + βV (ai , d ) + (d ) dH x (x )
(8)
d x0
Z ∗
∗
x0
0
x∗ x0
J(ai , di,t−1 , 1) =
max
π(ai , di,t−1 , d ) + βV (ai , d ) + (d ) dH x (x )
(9)
0
W (ai , di,t−1 ) =
dx
where H k represents the cumulative distribution of k , k ∈ {χ, x, f d}. The parameter β ∈ (0, 1)
is the discount factor. Thus, the value function V (·, ·) characterizes the firm’s exit decision,
while W (·, ·) describes the firm’s decision whether to engage in FDI. Similarly, J(·, ·, 1) is the
continuation value of a firm which has decided to maintain a plant abroad, while J(·, ·, 0) is
the continuation value of a firm that only produces in its country of origin. Note that export
decisions are made after the FDI decision. The advantage of this decision structure is that
it allows for natural separation between foreign and domestic export decisions. A potential
drawback is that firms make an FDI decision before realizing their export shock.27
3.4. Equilibrium
I focus on a stationary equilibrium where aggregate variables and the distribution of productivity are constant over time. Denote the stationary distribution of productivity across
incumbent firms by µ(a, d). Although firms exit home and foreign markets as they receive the
various firm-specific shocks outlined above, these changes are exactly offset by new entrants of
each type in equilibrium. For convenience, the time subscript is dropped and the firm’s state is
denoted as (ai , di ).
Under free entry the expected value of the firm must be equal to the fixed entry cost fe :
Z
V (a, (0, 0))ga (a)da = fe
(10)
where ga (a) is the distribution of initial productivity draws.
A stationary equilibrium requires that the number of exiting firms must equal the number
of successful new entrants. Specifically,
Z
M
!
X
Z
χ
P (χ = 0|a, d)µ(a, d) da = Me
P χ (χ = 1|a, (0, 0))ga (a)da,
(11)
d
27
Similar empirical results are found using alternative timing structures. Decision trees are included in the
Supplemental Appendix.
12
where M is the mass of incumbents, Me is the total mass of entrants that attempt to enter
the market, P χ (χ = 0|a, d) is the probability of a firm of type (a, d) exiting the market and
P χ (χ = 1|a, d) = 1 − P χ (χ = 0|a, d) is the probability of a firm (a, d) choosing to not exit. Using
the properties of extreme-value distributed random variables, the probability of producing this
period (χ = 1) is calculated as:
χ
P (χ = 1|a, d) = (1 − ξ)
exp(W (a, d)/%χ )
exp(0) + exp(W (a, d)/%χ )
(12)
where ξ is the exogenous probability of exit and W (·, ·) is defined as in equation (??).
The conditional choice probabilities for all possible FDI and export decisions follow the
nested logit formula.28 Conditional on operating, the probability of maintaining a plant abroad
is given by
0
P f d (df d = 1|a, d, χ = 1) = P
exp(J(a, d, (1, dx )/%f d )
fd
¯f d x
d¯f d exp(J(a, d, (d , d ))/% )
(13)
while the conditional probability of producing everything in the country of origin is given by
0
0
P f d (df d = 0|a, d, χ = 1) = 1 − P f d (df d = 1|a, d, χ = 1) and J(·, ·, ·) is given in equations
(??)-(??). Similarly, conditional on operating and choosing to maintain a plant abroad, the
firm’s export choice probability can be calculated as
0
0
∗
exp([π(a, d, (dx , 1)) + βV (a, d0 )]/%x )
.
x0
0
x∗
dx0 exp([π(a, d, ((d , 1)) + βV (a, d )]/% )
0
P x (dx |a, d, χ = 1, df d = 1) = P
(14)
The choice probability for exporting conditional on not engaging in FDI is analogous to (??)
where the scale parameters are replaced with their unstarred counterparts.
The final condition required for a stationary equilibrium is that the measure of firms of with
state (a, d) is constant over time. Define the probability of FDI and export as
d
P d (dit |ai , di,t−1 ) = P f d (dfitd |ai , di,t−1 )P x (dxit |ai , di,t−1 , dfi,t−1
)
and P f d (dfitd = 1|ai , di,t−1 ) = 0 for Indonesian plants by assumption. The equilibrium condition
can therefore be written as
M µ(a, d) = M P χ (χ = 1|a, d)
X
P d (d|a, d0 , χ = 1)µ(a, d)
d0
χ
+Me P (χ = 1|a, (0, 0))P d (d|a, (0, 0), χ = 1)ga (a)
(15)
The first term on the right-hand side of (??) is the measure of incumbent producers who do not
exit and have state (a, d) in the current period. The second term is the measure of new entrants
28
See McFadden (1978).
13
with the same state.
4. Structural Estimation
This paper extends Rust’s (1987) nested logit dynamic programming framework to examine
the nature of exit, export and FDI decisions in the presence of stochastic fixed costs. The following subsections describe the data, construct the likelihood function, and discuss the identification
of the model’s parameters.
4.1. The Likelihood Function
It is assumed that total revenue is measured with error and that exogenous technological
change occurs at rate ρ. By modifying the profit functions to include measurement error and a
time trend, equation (??) can be used to write the logarithm of observed revenue for any plant
i as
ln rit = ρt t + ln ϕb (1 − dxit ) + ln[ϕb + ϕw exp(ϕτ )]dxit − ln ai + ν1,it
(16)
where rit is observed revenue, ϕw measures the ratio of wages across countries (w∗ /w)1−ε , ϕb
is a function of the country sizes and prices (B ∗ /B)1−ε and ν1,it is the associated measurement
error. Equation (??) highlights an important limitation of the data: the data only documents
the revenue, exports and ownership from plants located in Indonesia. Estimating the model
requires imposing some consistency across plants located in different countries. In particular,
it is assumed that every foreign non-exporting plant also produces in a separate plant located
in the foreign country with the same firm-specific productivity level as the plant located in
Indonesia. Ramstetter and Sjoholm (2006) analyze the same Indonesian data and indicate that
foreign owned plants in Indonesia are typically part of multinational corporations. Moreover, to
the extent that the model captures the decision of foreign plants to enter Indonesia, the model’s
implications for Indonesia will remain valid.29
Define ϕτ to be a function of iceberg shipping costs, ϕτ ≡ (1 − ε) ln τ . We assume that ϕτ is
firm-specific, constant over the life of the firm and drawn from a normal distribution with mean
µτ and variance στ2 .30 The ratio of export to total sales (among exporters) can then be written
as
ln
f
rit
rit
!
=
exp(ϕτ )ϕb
1 + exp(ϕτ )ϕb
29
+ ν2,it
(17)
Other papers use data that directly connects parents and subsidiary plants (e.g. Helpman, Melitz and Yeaple
(2004)). However, Baily, Hulten and Campbell (1992) show that accounting for productivity differences across all
plants is particularly important when evaluating policy. Last, the OECD (2006) reports that Indonesia received
14,352 million US dollars of FDI from 1993-1996, while it supplied 40 million US dollars worth of FDI.
30
Allowing transport costs to vary across foreign and domestic plants makes little difference to the final results.
14
f
where rit
is the plant’s export sales and ν2,it is measurement error in export intensity. We
assume that νit ≡ (ν1,it , ν2,it ) is randomly drawn from a mean zero normal distribution with
covariance matrix Σν . The density function is given by gν (·) and parameterize Σν using the
unique Cholesky decomposition as Σν = Λν Λ0ν . Let λj,k represent the (j, k)-th element of Λν .
A firm’s detrended net profit may be expressed as:
π(a, di,t−1 , dit ) = r(a, dit ) − F (di,t−1 , dit )
(18)
where the value of π(·) depends on the vector of parameters to be estimated by maximum
likelihood, θ:31
∗
θ = (ρ, ϕb , f, fx , ff d , cx , cf d , ζfx , ζcx , ξ, ε, %χ , %x , %x , %f d , µτ , στ , σa , σa∗ , λ11 , λ21 , λ22 ).
Define the variable χ̄it be a variable that takes a value of 0 if a plant exits the data and
1 otherwise. Thus, the probability of χ̄ = 1 for domestic plants is simply the probability that
χ = 1, Pθχ̄ (χ̄it = 0|ai , di,t−1 ) = Pθχ (χit = 0|ai , di,t−1 ). However, a foreign plant may exit because
of a shock that causes the firm to die entirely or because it chooses to produce in its country of
origin instead:
∗
Pθχ̄ (χ̄∗it = 0|ai , di,t−1 ) = Pθχ (χit = 0|ai , di,t−1 ) + Pθχ (χit = 1|ai , di,t−1 )Pθf d (dfitd = 0|ai , di,t−1 ).
Denote Ti,0 as the first year the firm appears in the data. Then, conditional on ai and di,t−1
the likelihood contribution of plant i in year t > Ti,0 is

χ̄

 Pθ (χ̄it = 0|ai , di,t−1 )
Pθχ̄ (χ̄it = 1|ai , di,t−1 ) P d (dit |ai , di,t−1 , χit = 1)
Lit (θ|ai , di,t−1 ) =
{z
}
{z
}|

 |
FDI/Export
Stay/Exit
gν (ν̃it (ai ))
| {z }
for χ̄it = 0
for χ̄it = 1
Rev./Exp. Intensity
where starred values replace unstarred values for foreign plants. Note that the endogeneity of
the export, FDI and exiting decisions are controlled by simultaneously considering the likelihood
contribution from each decision. Since only plants that chose to stay in the market are observed
their likelihood contribution of these plants in the initial year is calculated conditional on χit = 1,
Lit (θ|ai , di,t−1 ) = P d (dit |ai , di,t−1 , χit = 1)gν (ν̃it (ai )).
Let Ti,1 denote the last year plant i appears in the data. Then, the likelihood contribution
31
The detrended firm’s problem uses a trend-adjusted discount factor β exp(ρ) when solving the Bellman’s
equation. The discount factor is not estimated and is set to 0.96. It is difficult to identify the discount factor β
in dynamic discrete choice models. Also, the mean of the distribution of initial productivity draws is assumed to
be the same for Indonesian and foreign plants.
15
from each plant i is
Ti,1
Li (θ|ai , di,Ti,0 ) =
Y
Lit (θ|ai , di,t−1 ).
t=Ti,0 +1
The joint distribution of unobserved productivity ai and export/FDI status di,t−1 depend on
whether entering plants are observed in the initial year of the sample (Heckman (1981)). In the
initial year of the sample it is assumed that the pair (ai , di,t−1 ) are drawn from the stationary
distribution µ(a, d) since we cannot determine if the firm is a new entrant or an incumbent
firm. In subsequent years, ai is assumed to be drawn from the distribution of initial draws upon
successful entry into Indonesia for all entering firms
gae (a) = R
P χ̄ (χ̄ = 1|a, dt−1 )
ga (a).
P χ̄ (χ̄ = 1|a0 , dt−1 )ga (a0 )da0
(19)
where dt−1 = (0, 0) in the plant’s initial year of entry for all domestic plants. Equation (??) is
used along with the choice probabilities (??)-(??) to build the likelihood function.
The likelihood contribution from each plant i is calculated by numerically integrating out
unobserved plant-specific productivity ai as
( R
Li (θ) =
R
Li (θ|a0 , di,Ti,0 )µ(a0 , di,Ti,0 )da0
for Ti,0 = 1993,
Li (θ|a0 , di,Ti,0 )Pθd (di,Ti,0 |a0 , di,Ti,0 −1 )gae (a0 )da0
for Ti,0 > 1993,
where the starred distributions are used in place of the unstarred distributions for foreign
firms and the stationary distribution of foreign plants is conditional on entering Indonesia,
µ∗ (a0 , di,Ti,0 |dfi,Td i,0 = 1). The parameter vector θ can then be estimated by maximizing the
logarithm of the likelihood function
L(θ) =
N
X
ln Li (θ).
(20)
i=1
The evaluation of the log-likelihood function involves solving the dynamic programming
problem that approximates the Bellman equations (??)-(??) by discretization of the state space.
For each candidate choice of parameter vector, the discretized dynamic programming problem
(??)-(??) is solved and used to calculate the conditional choice probabilities (??)-(??) and the
stationary distributions. Using the conditional choice probabilities and the stationary distributions, it is possible to evaluate the log-likelihood function (??). Searching over the parameter
space of θ, the estimates are found by maximizing (??).
16
5. Data
The primary source of data is the Indonesian manufacturing census between 1993 and 1996.
We focus on these years due to the fact that it is a period of relative calm in the Indonesian
economic and policy environment. This is a key feature of the data as the model described is
stationary in nature.32 Collected annually by the Central Bureau of Statistics, Budan Pusat
Statistik (BPS), the survey covers the population of manufacturing plants in Indonesia with at
least 20 employees.33 The data capture the formal manufacturing sector and record detailed
plant-level information on over 100 variables covering industrial classification (5-digit ISIC),
revenues, intermediate inputs, labour, capital, energy, trade and foreign ownership. Nominal
values of total sales are converted to the real value added, rit , using the manufacturing output,
input, and export price deflators at the industry level.34
An advantage of this data set relative to many other plant-level data sets is that it includes
the percentage of foreign ownership for each individual plant. A foreign plant is defined as any
plant where at least 10 percent of total equity is held by foreign investors. In over 66% of the
foreign plants in the sample, foreign investors own at least 50% of the equity, while foreign
investors own at least 25% of foreign plants in 95% of the sample.35 Similarly, exporters are
identified as plants that receive any positive revenues from export sales.36
5.1. Exports and Foreign Ownership
In this section we describe the Indonesian manufacturing industry and the differences across
domestic and foreign owned firms. Multinational and/or foreign-owned firms are typically the
largest plants in the Indonesian manufacturing sector between 1993-1996. Although only six
percent of all firms have any foreign ownership, foreign firms account for more than one quarter
of total output and over one third of all exports in each year. Moreover, foreign firms are
not solely export oriented but also capture one quarter of the Indonesian domestic market for
32
There is substantial turmoil in the Indonesian economic and policy environment due to the Asian crisis. It
would be inappropriate, for instance, to extend the estimation to the subsequent years in 1997 and 1998 where
a number of institutional changes and shocks which are not captured by the model may affect the parameter
estimates and the model’s quantitative predictions. Ito and Satio (2006) document the change in volatility during
the Asian crisis which began in 1997. Agung et al. (2001) describe the nature of lending constraints faced by
Indonesian before and after the financial crisis.
33
Very few foreign firms are near this threshold in any year, including the year immediately before exiting.
34
Price deflators are constructed as closely as possible to Blalock and Gertler (2004).
35
A description of the Indonesian manufacturing sector, data coverage over this period, and a discussion of
foreign ownership in Indonesia can be found in Blalock and Gertler (2005).
36
The data do not reveal the export destinations of each plant. However, as shown in Table 11 in the Appendix,
Indonesian industries that earn a higher percentage of revenues from exports are more likely to export to developed
countries. Along with the discussion in Section 2, this would suggest that foreign exporters in export-intensive
industries are likely to export to developed countries. To be conservative all foreign exporters are assumed to
export exclusively to developed countries.
17
manufactured goods. This suggests that Indonesia may also be an important market for foreign
plants.
A common explanation for these findings is that multinational firms are substantially larger
and more productive than their domestic counterparts.37 Comparing the second and third rows
with the fourth and fifth rows of Table 2 it is evident that foreign plants are also more capitalintensive, use a higher fraction of skilled employees and have greater output and value added
per worker than their domestic counterparts.
Similarly, it is often found that domestic exporters are relatively large, capital and skillintensive, and more productive relative to domestic non-exporters.38 The bottom two rows of
Table 2 are consistent with this result for the Indonesian manufacturing sector. A surprising
feature of Table 2 is that foreign non-exporters report higher mean values of labour productivity,
value-added per worker, capital-intensity and skill-intensity. This may reflect differences in the
allocation of exporting plants across industries. Moreover, it is not clear that the differences
across ownership or export status are statistically significant.39
Table 2: Descriptive Statistics
All
Plants
Foreign
Exporters
Foreign
Non-Exporters
Domestic
Exporters
Domestic
Non-Exporters
Total
Salesa
43.85
(293.49)
169.79
(388.76)
170.54
(366.30)
87.89
(518.86)
21.77
(191.92)
Labor
Capital
220.82
(678.55)
704.81
(1,057.50)
402.59
(637.71)
477.64
(1,242.39)
128.42
(378.73)
7.32
(24.17)
30.46
(47.23)
29.27
(49.24)
15.56
(38.90)
3.29
(11.32)
Export
Intensitya,b
0.69
(0.33)
0.70
(0.34)
—
—
0.69
(0.33)
—
—
Skill
Ratioc
0.16
(0.13)
0.16
(0.14)
0.27
(0.18)
0.15
(0.11)
0.16
(0.13)
K/L
Ratioa
0.03
(0.03)
0.06
(0.05)
0.09
(0.06)
0.03
(0.03)
0.02
(0.02)
Output/
Worker
0.13
(0.37)
0.34
(0.57)
0.46
(0.82)
0.16
(0.38)
0.09
(0.31)
V-A/
Workerd
0.04
(0.15)
0.12
(0.23)
0.17
(0.36)
0.05
(0.12)
0.03
(0.13)
No. of
Obs.
57,414
—
2,444
—
1,736
—
9,796
—
43,438
—
Notes: Reported numbers are sample means with standard deviations in parentheses. (a) In millions of Indonesian Rupiahs.
(b) Computed using the sample of exporting plants. (c) The ratio of skilled to total workers. (d) Value-added per worker.
Following Bernard and Jensen (1999) and Kasahara and Lapham (2007), the export and
ownership premia are estimated using a pooled ordinary least squares regression over the 19931996 period:
ln Xit = α0 + α1 dxit (1 − dfitd ) + α2 dxit dfitd + α3 (1 − di tx )dfitd + Zit β + λt + ηi + it ,
37
(21)
See Helpman, Melitz and Yeaple (2004) for an example.
See Kasahara and Lapham (2006) for an example.
39
It may also be possible that non-exporting plants are part of a multi-plant firm where another plant is
responsible for exports of finished products. However, BPS Indonesia estimates that 95% of plants in our sample
are not attached to other plants in Indonesia.
38
18
where Xit is a vector of plant attributes such as total employment, domestic sales, output per
worker, average wages, the percentage of non-production workers and capital per worker. A
firm’s export status is captured by the dummy variable dxit , while dfitd is a dummy variable
capturing whether any of the plant’s equity is held by foreign investors. Last, Zit is matrix of
control variables including dummy variables for plant-age, while the year-dummies and plantlevel fixed effects are added to control for the year and plant-specific effects. The coefficient
αj where j = {1, 2, 3} captures the average log point difference between the group of firms
associated with aj and domestic non-exporters (the excluded group).
Table 3: Export & Ownership Premia
Export/Ownership Status
Output per Worker
Value-Added per Worker
Solow Residual
Average Wage
Non-Production/Total Workers
Capital per Worker
Domestic Sales
Total Sales
Total Employment
Domestic Exporters
0.047
(0.011)
0.090
(0.014)
0.005
(0.009)
0.047
(0.008)
-0.019
(0.009)
0.141
(0.011)
-0.736
(0.017)
0.151
(0.014)
0.084
(0.006)
No. of Observations
Pooled OLS: 1993-1996
Foreign Non-Exporters
0.245
(0.029)
0.177
(0.035)
0.065
(0.024)
0.113
(0.022)
0.114
(0.023)
0.551
(0.029)
0.386
(0.042)
0.402
(0.036)
0.168
(0.01)
53808
Foreign Exporters
0.312
(0.030)
0.197
(0.036)
0.067
(0.025)
0.113
(0.022)
0.003
(0.024)
0.467
(0.030)
-0.501
(0.046)
0.494
(0.036)
0.192
(0.015)
Notes: Standard errors are in parentheses.
The results in Table 3 confirm that domestic exporters tend to demonstrate higher premia
than domestic non-exporters. Similarly, foreign plants are generally more productive than their
domestic counterparts. However, the ranking across export status is much less clear for foreign
plants. The last two columns indicate that while foreign exporters demonstrate higher average
labour productivity and value added per worker, foreign non-exporters pay higher averages, use
capital and skill more intensively, and have much higher domestic sales.40 Moreover, it appears
that the mean differences are rarely statistically different. As such, a model of FDI in Indonesia
must reconcile both the existence of foreign exporters and non-exporters in the same narrowly
defined industries while at the same time generating similar productivity levels across these
groups. The following section presents one possible model that captures these features of the
40
The Solow residual for each plant is calculated using the Olley-Pakes (1996) production function estimation
routine. Details can be found in the Supplemental Appendix.
19
Indonesian data.41
5.2. Data Used in the Structural Estimation
The structural estimation in this paper focuses on Indonesian food, textile and manufactured
metals industries as they are among Indonesia’s largest industries and receive substantial foreign
direct investment.42 Moreover, as shown Das, Roberts and Tybout (2007) the magnitude of
sunk and fixed entry costs often vary substantially across industries. The food, textile and
manufactured metals industry data consists of unbalanced panels of 6,042, 4,491 and 2,497
plants, respectively.
This paper focuses the following four variables for each plant i in year t: χit , rit , dxit and dfitd .
Intermediate materials are subtracted from total revenues to account for differential intensities
in intermediate materials.43 The binary variables dxit and dfitd are constructed by checking the
value of export sales and foreign ownership in each year. The entry/exit decision, χit is identified
by checking whether plants employed a positive number of workers in each year. Descriptive
statistics for all four variables are found in Table 4. Table 4 reveals substantial variation across
variables and high exit and entry rates over the sample period. These are particularly important
for identifying the choice probabilities as well as the initial distribution of productivity shocks.
Table 4: 1993-96 Descriptive Statistics
Food
Metalsg
Textiles
Value
Addeda
11.78
(290.42)
16.30
(122.27)
17.63
(129.56)
Export
Intensitya,b
52.5
(32.1)
43.7
(30.7)
59.7
(30.3)
Labour
106.64
(703.44)
168.87
(325.05)
245.09
(730.25)
Mark-Up
Ratec
0.26
(0.18)
0.33
(0.19)
0.26
(0.16)
% Foreign
Plantsd
0.02
—
0.10
—
0.04
—
Entry
Ratese
0.15
—
0.16
—
0.15
—
Exit
Ratesf
0.09
—
0.06
—
0.08
—
No. of
Plants
6,042
—
2,497
—
4,491
—
Notes: Reported numbers are sample means with standard deviations in parentheses. (a) In millions of Indonesian Rupiahs.
The percentage is calculated as the mean percentage across plants. (b) Computed using the sample of exporting plants. (c)
The mark-up rate is computed as (revenue-variable cost)/revenue where variable cost is measured as the sum of materials,
energy, fuel and the wages paid to production workers. (d) The average is computed as the percentage of plants with foreign
ownership. (e) The number of new entrants divided by the total number of plants in 1993. (f ) The number of exiting plants
divided by the total number of plants. (g) Metals refers to manufactured metals rather than basic metals.
41
See the Supplemental Appendix for a discussion on the possible impact of transfer pricing, differences in
tax-rates across countries or mark-ups across heterogeneous firms.
42
Approximately 58% of all foreign plants in the Indonesian manufacturing census are in the food, textiles
and manufactured metals industries (2-digit ISIC codes). A total of 729 plants that are owned entirely by the
Indonesian government are omitted.
43
The author thanks Richard Rogerson for this suggestion.
20
5.3. Identification
It is not possible to identify all of the parameters of the model. Equation (??) is a reducedform specification where the reduced-form parameters represent the following structural parameters:
ϕB =
wB
w∗ B ∗
1−ε
ϕW =
w 1−ε
.
w∗
It is important to note that policy changes may affect the value of reduced-form parameters if
the underlying structural parameters change. For instance, any change to the aggregate price
level P will lead to a change in B = [(1 − α)δE]1/(1−ε) /(αP ) and thus ϕB . The counterfactual
experiments in this paper explicitly account for equilibrium price changes on the reduced-form
coefficients using the relationship between the reduced-form coefficients and the aggregate prices.
The scale of the profit function also cannot be identified because multiplying the profit function
by a constant leads to the same optimal choice. Thus, for identification the profit function (??)
is normalized by κ = ε/(w∗ B ∗ )1−ε .44
The identification of the revenue function (??) parameters follows from the within-plant variation in revenue and export status. The fixed cost and scale parameters are similarly identified
by relating the variation in productivity to the variation in export and FDI probabilities.
The sunk cost parameters are identified from two different sources of variation. The sunk
export cost is identified from the differences in export frequencies across plants with similar
productivity levels but different export histories. The sunk FDI costs are identified off of the
entry and exit dynamics of foreign firms, given the parametric assumption on the distribution
of productivity. Specifically, a model without sunk costs cannot match the distribution of
productivity among foreign owned firms in Indonesia, because high fixed costs and low cost
shocks are both required to match the higher productivity and low exit rates across foreign
owned plants. This results in too few low productivity, foreign owned plants. A model with
sunk costs can do a better job of simultaneously matching these features since the previous
investment in Indonesia will encourage low productivity foreign plants to stay in Indonesia even
in the presence of large cost shocks.
Due to limitations of the data it is not possible to identify the parameters ϕw , f ∗ , fx∗ ,
and c∗x . The first parameter represents differences in wages across countries. It is calibrated
using cross-country manufacturing wage data from International Labour Organization Bureau
of Statistics. The last three parameters are fixed and sunk cost parameters in the foreign country
and are calibrated using data from the World Bank’s Doing Business Report. The fixed cost
of operating in a foreign country, f ∗ is calibrated to be f ∗ = 0.51f . Similarly, the fixed and
sunk costs of exporting from the foreign country, fx∗ = 0.38fx and c∗x = 0.38cx . The calibrated
44
Specifically, the parameters κf , κfx , κff d , κ%χ , κ%f d κ%x and κ%∗x are estimated instead of f , ff d , fx , fx∗ ,
%χ , %f d , %x and %∗x .
21
parameters are constructed by comparing Indonesia’s foreign wage, index of business rigidity
and required export processing days to an average values for the rest of the world where the
share of FDI in Indonesia are used as weights. To test for possible misspecification around the
fixed cost parameters, the robustness of the results with regards to this calibration is checked by
estimating the model under various alternative fixed and sunk cost assumptions. In particular,
the model is re-estimated assuming that the fixed operation costs are equal across countries and
that fixed costs are much higher in Indonesia.45
6. Estimation Results
Table 5 presents the maximum likelihood parameter estimates of the model along with
the associated asymptotic standard errors. The standard errors are computed using the outer
product of gradients estimator. The parameters are evaluated in millions of Indonesian rupiahs
in 1983. The implications and interpretation of the estimated parameters are discussed in the
following subsections.
Table 5: Structural Estimates
Industry
κ%χ
κ%x
κ%x∗
κ%f d
κcx
κcf d
κf
κfx
κff d
ζfx
ζcx
ρ
ϕB
µτ
σa
σa∗
στ
ξ
ε = 1/mark-up
log-likelihood
No. of Obs.
Food
73.742
(9.171)
2.609
(1.528)
13.535
(2.098)
2.357
(0.775)
8.174
(4.594)
20.884
(7.482)
4.762
(1.864)
1.414
(1.135)
0.085
(3.104)
0.012
(1.378)
0.003
(0.337)
0.019
(0.008)
0.355
(0.010)
-7.651
(0.577)
1.234
(0.009)
1.217
(0.014)
0.213
(0.119)
0.009
(0.003)
3.8
-30575
17,786
Metals
79.376
(12.969)
20.965
(3.959)
18.997
(7.556)
3.797
(0.421)
71.135
(12.548)
48.525
(5.148)
0.441
(2.161)
0.353
(2.278)
2.772
(6.219)
0.141
(5.700)
0.009
(0.036)
0.005
(0.011)
0.777
(0.028)
-4.330
(0.225)
1.197
(0.017)
0.849
(0.012)
0.156
(0.140)
0.007
(0.002)
3.0
-14269
7,549
Textiles
115.037
(42.359)
5.891
(1.507)
25.779
(19.343)
2.942
(0.887)
16.664
(4.111)
58.391
(18.159)
1.161
(4.804)
1.437
(0.734)
0.623
(14.805)
0.127
(3.134)
0.007
(0.278)
0.002
(0.019)
0.771
(0.031)
-4.819
(0.248)
0.716
(0.019)
0.570
(0.036)
0.295
(0.087)
0.001
(0.003)
3.8
-29901
13,287
Notes: Standard errors are in parentheses. The parameters are evaluated in units of millions of Indonesian Rupiahs in 1983.
Metals refers to manufactured metals.
45
Robustness results are presented in the Appendix. Also, see the Supplemental Appendix for a brief description
of the results from the estimation of a model without sunk costs.
22
6.1. Productivity
The model predicts that only the most productive domestic firms are able to produce profitably in Indonesia. Table 6 confirms that domestic incumbents are on average 6 to 37 percent
more productive than new domestic entrants across industries. This pattern is even more striking among foreign firms where incumbents 46 to 113 percent more productive than domestic
plants across industries. This pattern is consistent with the finding that only the productive
firms find it profitable to maintain foreign affiliates.
Table 6: Average Productivity
Mean at successful entry in Indonesia
Mean at steady state in Indonesia
Non-Exporters
Exporters
Food
Indonesian
1.108
1.215
1.174
1.958
Foreign
1.099
2.201
3.465
0.495
Metals
Indonesian
Foreign
1.208
1.840
1.650
3.921
1.590
4.280
2.385
2.452
Textiles
Indonesian
Foreign
1.035
1.156
1.099
1.684
1.039
1.534
1.416
1.956
Note: Metals refers to manufactured metals.
As shown in Section 2 domestic exporters tend to have higher productivity than domestic
non-exporters, and foreign plants are more productive than domestic plants. This pattern is
confirmed in each of the three industries documented in Table 6. There is no consistent ranking
of exporters and non-exporters within foreign plants.46
Our theory suggests that industries where shipping costs may be higher (or with smaller
export markets), we might expect to find the most productive firms engaging in horizontal FDI.
In contrast, in industries where large export markets (captured by a low τ ), we might expect
to find the opposite pattern. This will be a key feature of the counterfactual experiments as
exporting and non-exporting foreign plants will react differently to changes in policy across
industries.
6.2. Sunk and Fixed Costs
The estimates of sunk export and FDI costs across industries are large. The estimated
sunk cost of export for a domestic exporter ranges from 154 thousand 1983 US dollars in the
food industry to 537 and 572 thousand US dollars in the textiles and manufactured metals
industries, respectively. Across industries, the sunk costs are greater than one year’s worth of
export revenue. In contrast, foreign plants are estimated to only pay 1% of the sunk export
costs faced by their domestic counterparts. Sunk FDI costs are also estimated to be very large.
Across industries the sunk FDI costs range from 1883 thousand 1983 US dollars (textiles) to
572 thousand 1983 US dollars (metals). However, these costs also account for a substantial
46
Moreover, in the food industry we see that foreign exporters are less productive than domestic non-exporters.
Lu (2010) also reports that highly export-intensive firms are often found to much less productive in similar setting.
23
Table 7: Distribution of Plants by Ownership/Export Status
Actual
Food
Metals
Textiles
Predicted
Food
Metals
Textiles
Domestic Non-Exporters
0.931
0.824
0.839
Domestic Exporters
0.046
0.071
0.114
Foreign Non-Exporters
0.016
0.058
0.025
Foreign Exporters
0.008
0.047
0.023
0.925
0.827
0.801
0.051
0.068
0.152
0.018
0.063
0.020
0.006
0.042
0.028
Note: Metals refers to manufactured metals.
percentage of total revenue; across industries the sunk FDI costs account for 156% (food) to
1245% (textiles) of annual total revenue earned by foreign plants in Indonesia. Although the
average sunk costs are typically a large percentage of revenue it is important to recall that the
sunk costs incurred are substantially lower since the estimated model predicts many plants only
export or invest abroad when they receive beneficial cost shocks.
Fixed costs are typically much smaller than the estimated sunk costs. The average fixed
cost of operation in Indonesia ranges from 4 to 5 thousand 1983 US dollars in the textiles
and manufactured metals industries, while it is 90 thousand dollars in food industry. The
average fixed export cost ranges from 4 thousand 1983 US dollars in the manufactured metals
industry to 26 thousand 1983 US dollars in the food industry. As with the sunk costs, foreign
firms are estimated to only pay a small percentage of the fixed export costs (1 to 15% across
industries). The fixed FDI costs range from less than 1 thousand dollars in the manufactured
metals industry, to 2 thousand dollars in the food industry and 19 thousand 1983 US dollars
in the textiles industry. It is important to note that there exists substantial uncertainty in the
estimates of most fixed costs since the standard errors are usually large and the estimates are
often insignificant.47
6.3. Exports and FDI
The estimates of the parameters ϕτ and ϕB jointly capture the impact of exporting at the
plant-level. The estimates also imply that on average 21 and 43 percent of a domestic exporters’
revenues are from export sales in the food and manufactured metals industries, while exporters
in the textiles industry receive 70 percent of revenues from export sales. Table 7 documents the
predicted distribution of plants export and ownership status for all industries.48 The model’s
predictions match the distribution of plants across export and ownership status very closely
47
Government policy may also drive differences in the estimates across industries. For example, it is well known
that the manufactured metals industry received large subsidies during this period.
48
The estimated model does not provide a prediction in terms of the total number of foreign firms relative to
the total number of domestic firms. As such, the percentage of foreign firms in the data is taken as given.
24
Table 8: Export/Domestic Market Share by Ownership/Export Status
Actual
Food
Metals
Textiles
Predicted
Food
Metals
Textiles
Export Share
Domestic
Foreign
0.837
0.163
0.458
0.542
0.717
0.283
0.968
0.515
0.686
Domestic Non-Exp.
0.551
0.462
0.602
0.032
0.485
0.314
0.576
0.514
0.551
Domestic Market Share
Domestic Exp.
Foreign Non-Exp.
0.289
0.126
0.099
0.300
0.238
0.097
0.208
0.109
0.300
0.209
0.262
0.048
Foreign Exp.
0.032
0.138
0.062
0.007
0.115
0.103
Note: Metals refers to manufactured metals.
in all industries. Table 8 reports the models predicted domestic market and export shares
across export and ownership status. The model also matches the distribution of sales relatively
well. While one would expect the model to predict the dominant role of foreign plants in the
domestic and export markets, Table 8 demonstrates that it often overstates their importance
in both domestic and export markets. Because of the productivity differences between foreign
and domestic plants, small differences in the distribution of plants or the predicted productivity
differences can have large impacts at the aggregate level.
6.4. Dynamics
Table 9 documents the actual and predicted transition probabilities of FDI, export and exit
in the textiles industry.49 It is noteworthy that despite the model’s restriction that no domestic
plant can become a foreign plant, the model captures many of the transition probabilities across
FDI and export status relatively well. Although the model generally captures the persistence
in export and ownership status well, it does slightly underpredict the degree persistence for
domestic exporters.50
6.5. Counterfactual Experiments
This section presents the results from a series of counterfactual experiments intended to
examine the effect of trade and FDI barriers for the textiles and manufactured metals industries.
The counterfactual results for the food industry are presented in the Appendix as the results a
very similar to those found in the manufactured metals industry. To determine the quantitative
implications of barriers to trade and FDI, the following three counterfactual experiments are
conducted by manipulating three parameters:
1. Autarky: fx , fx∗ , ff d → ∞,
49
Similar tables can be found in the Supplemental Appendix for the food and manufactured metals industries.
Das, Roberts and Tybout (2007), Kasahara and Lapham (2007) and Ruhl and Willis (2007) all find that sunk
costs are necessary in order to capture the dynamic behaviour of entry and exit in international trade markets.
50
25
Table 9: Transition Probabilities - Textiles
Actual
Dom. Non-Exporters at t
Dom. Exporters at t
For. Non-Exporters at t
For. Exporters at t
Predicted
Dom. Non-Exporters at t
Dom. Exporters at t
For. Non-Exporters at t
For. Exporters at t
Dom. Non-Exporters
0.945
0.309
0.138
0.027
Dom. Exporters
0.051
0.679
0.031
0.066
For. Non-Exporters
0.003
0.001
0.581
0.178
For. Exporters
0.001
0.011
0.256
0.732
Exit
0.088
0.068
0.030
0.016
0.902
0.501
—
—
0.098
0.499
—
—
—
—
0.436
0.391
—
—
0.564
0.609
0.088
0.058
0.021
0.016
2. No Trade: fx , fx∗ → ∞,
3. No FDI in Indonesia: ff d → ∞,
The experiments are chosen to demonstrate to key features of policy change that have been
relatively unexamined elsewhere in the literature. First, we examine how trade and FDI policy
changes can induce different reactions by both foreign and domestic firms. Second, the experiments highlight that the expected outcome from a change in trade or FDI policy depends
greatly on the degree to which foriegn firms use Indonesia as a platform for exporting. In order
to determine the full impact of trade or FDI barriers on the Indonesian economy it is important
to consider the effect policy changes have on the aggregate price level. The experiments impose
the free entry conditions (??) and solve for the new price levels in Indonesia and the rest of the
world which satisfy (??) under the policy change.51
Table 10 presents the results from the three counterfactual experiments in the manufactured
metals and textiles industry. The effect of trade and FDI on average productivity can be best
understood by comparing the steady state level of average productivity between the estimated
(base) model and that of the counterfactual experiments in the first row. The second row
compares welfare across the experiments and will be left to the end.
As expected, eliminating both trade and FDI induces the largest reduction average productivity across industries. In the textiles industry aggregate productivity falls by 32.8 percent,
while in the manufactured metals industry we observe a smaller fall in productivity by only 17.7
percent. Column 3 examines the change in the structure of the Indonesian economy when trade
is restricted by FDI is permitted. In the textiles industry restricting trade causes productivity to fall by 31 percent, which is almost as severe as autarky. In contrast, restricting trade
only causes aggregate productivity to fall by less than 2 percent in the manufactured metals
industry. The difference across industries is attributable to the differential response of foreign
firms. In the textiles industry almost all foreign firms exit when trade is restricted. This is not
51
See the Supplemental Appendix for details.
26
Table 10: Counterfactual Experiments
Avg. Productivityb
−(ε − 1)δ∆ ln P
Exit/Entry Rate of Foreign
Firms in Indonesia
% of For. Non-Exporters
% of Dom. Exporters
%∆ in Exports
Mkt.
Mkt.
Mkt.
Mkt.
Shr.
Shr.
Shr.
Shr.
of
of
of
of
Dom. Non-Exp.
Dom. Exp.
For. Non-Exp.
For. Exp.
Base
5.040
—
Metalsa
Autarky
No Trade
4.145
4.971
−δ5.2
−δ2.7
No FDI
4.327
−δ0.1
Base
1.915
—
Textiles
Autarky
No Trade
1.286
1.309
−δ13.1
−δ13.1
No FDI
1.762
−δ0.1
—
-100
-82.2
-100
—
-100
-99.8
-100
6.3
6.8
—
0
0
-100
1.0
0
-100
0
7.6
-39.5
2.3
15.2
—
0
0
-100
0.1
0
-100
0
15.9
-31.4
51.4
10.9
26.2
11.5
100
0
0
0
77.5
82.5
17.5
55.1
29.8
4.8
10.3
100
0
0
0
99.9
64.9
35.1
0
0
22.5
0.1
Notes: a) Metals refers to manufactured metals. b) Average productivity of all plants located in Indonesia in the steady
state and is calculated using plant-level revenue shares as weights.
particularly surprising in this industry since the export market is relatively large and as such
the option value of exporting from Indonesia is highly valued even among firms that are not
currently exporting. Moreover, the most productive firms in the textiles industry are foreign
owned exporters. Restricting trade thus greatly reduces the value of maintaining a plant in
Indonesia and provides a strong for the most productive firms to stop producing in Indonesia.
In constrast, not all foreign plants choose to exit the manufactured metals industry in response to the change in trade policy. There is still a strong incentive to exit as 82.5 of existing
plants choose to stop producing, but the 17.5 percent of the most productive foreign plants
choose to continue production despite the trade restrictions. It is worth noting that in this
industry, in general, the most productive foreign plants were not exporting in the first place. As
such, they are less directly affected by the change in policy and have less incentive to leave the
country. Also, the export market in manufactured metals industry is estimated to be substantially smaller than the textiles industry and as such it valued less by foreign plants in Indonesia.
The remaining foreign non-exporters further grow into the vacated markets left by the exiting
plants which in turn mitigates the fall in productivity. In sum, restricting trade was almost
equivalent to autarky in the textiles industry, its effect on aggrgate productivity is substantially smaller in the manufactured metals industry. This difference can largely be attributed to
differences in foreign firm characteristics and exit decisions.
The fourth column presents the results from a counterfactual experiment where trade is
allowed but FDI is not permitted. While this policy reduces aggregate productivity in both
industries by encouraging foreign firms to leave Indonesia, the relative effect of restricting FDI
in comparison to exports varies strikingly across industries. In the textiles industry restricting
27
FDI induces an 8 percent fall in aggregate productivity, which is substantially smaller than the
31 percent reduction induced by restrictions on trade. In the textiles industry, both policies
encourage the exit of essentially all foreign firms, but only the change in trade policy causes
productive domestic exporters to shrink. Since many foreign firms generate a large number of
sales from exporting in Indonesia, trade restrictions act as both an impediment to trade and
FDI in this industry.
In the manufactured metals industry, restricting FDI causes aggregate productivity to fall
by 14 percent, which is substantially larger than the effect of trade on aggregate productivity in
this industry. This result is due to the fact that the industry is dominated by a small number
of large foreign firms who are as sensitive to the change in trade policy. As such, the restriction
on FDI forces the most productive firms out of industry and induces a relatively large fall in
aggregate productivity. By comparison, the most productive firms in Indonesia do not leave
the country in response to the change in trade policy leaving aggregate productivity much less
affected.
The welfare results are reported in the second row of Table 10. Welfare is measured as
the change in the inverse price level since increases in prices reduce the purchasing power of
consumers. The parameter δ captures the size of the textiles sector in the Indonesian economy.
Blalock and Gertler (2005) estimate that manufacturing composes approximately one quarter
of the Indonesian and the textiles industry accounts for almost one quarter of manufacturing
which implies δ ≈ 0.055. Although this will greatly reduce the size of the welfare impact it is
clear from Table 10 that if similar changes occur across all Indonesian manufacturing sectors
there would be significant drop in overall welfare.
Table 10 documents a relatively large fall in welfare under autarky and trade, but smaller
changes to welfare when there are restrictions to FDI. The intuition behind this result is that
trade (FDI) flows partially “insure” FDI (trade) flows in the presence of FDI (trade) restrictions.
For instance, a restriction to FDI is insured by the trade flows into Indonesia so that Indonesian
consumers can continue to access foreign goods through trade. Similarly, trade restrictions are
insured by foreign firms that enter and continue to produce in Indonesia. Since the foreign firms
tend to be much larger and more productive they tend to greatly reduce the fall in welfare. In
the textiles industry trade restrictions cause almost as large a fall in welfare as autarky, but
this is again due to the fact trade restrictions strongly discourage FDI in that industry. In sum,
the estimates imply that the model’s predictions broadly match the features of the Indonesian
manufacturing data. Moreover, the estimates confirm the ranking of productivity across plants
with different ownership and export status above.
It is natural to question the extent to which our results depend on the model’s assumptions.
To examine this issue I test the extent to which the calibration of fixed export and FDI costs
across countries affect the quantitative results. In particular, to test for possible misspecification
around the fixed cost parameters I re-estimate the model under various alternative fixed and
28
sunk cost assumptions. Here, I report two cases. First, the model is re-estimated under the
assumption that all fixed and sunk costs are equal across countries. Second, we consider a
scenario where the all foreign fixed and sunk costs are half as large in the foreign country as
they are in the benchmark model. The results from these exercises are presented in Appendix
and document the same qualitative results across industries and very similar quantitative results
across estimation assumptions.
7. Conclusions
This paper presents and estimates a model of foreign direct investment and exports with
heterogeneous firms. The model is structurally estimated using a panel of Indonesian manufacturing plants. Counterfactual policy experiments are employed to assess the impact of a change
in FDI or trade policy on aggregate productivity.
The model’s empirical predictions broadly match the features of the Indonesian manufacturing data. The model captures export decisions at the plant level and documents the differential
export behaviour across foreign and domestic firms. Moreover, the estimates confirm the ranking
of productivity across plants with different ownership and export status. The model emphasizes
that accounting for FDI flows is essential to recovering accurate estimates of the impact of trade
on aggregate productivity. In particular, the counterfactual experiments imply that the impact
of trade on productivity is greatly mitigated by FDI flows.
The counterfactual experiments imply that trade and FDI restrictions cause a large reduction
in aggregate productivity through the reallocation of resources from highly productive foreign
plants to less productive domestic plants. However, the size and magnitude varies across industries and depends crucially on whether exporting or market access are the primary reason for
foreign direct investment. These results are particularly important for policymakers in developing countries where the interaction between trade and foreign direct investment policy has been
largely unexamined.
The welfare implications differ substantially across international integration policies. Autarky causes substantial reductions in welfare, while trade and FDI restrictions cause much
smaller welfare impacts on an economy such as Indonesia. The model suggests that trade and
FDI act as substitutes for each other and reduce the welfare impact of trade or FDI restrictions.
29
Appendix
Table 11: Counterfactual Experiments: Food Industry
Avg. Productivitya
−(ε − 1)δ∆ ln P
Base
3.286
—
Autarky
2.833
−δ4.1%
Food
No Trade
3.197
−δ4.0%
No FDI
2.941
−δ1.3%
—
-100
-54.6
-100
1.8
5.1
—
0
0
-100
1.1
0
-100
0
5.2
-2.7
57.6
20.8
20.9
0.7
100
0
0
0
81.5
0
18.5
0
73.7
26.3
0
0
Exit/Entry Rate of Foreign
Firms in Indonesia
% of For. Non-Exporters
% of Dom. Exporters
%∆ in Exports
Dom.
Dom.
Dom.
Dom.
Mkt.
Mkt.
Mkt.
Mkt.
Shr.
Shr.
Shr.
Shr.
of
of
of
of
Dom. Non-Exp.
Dom. Exp.
For. Non-Exp.
For. Exp.
Table 12: Indonesian Export Destinations
Industry
Wood
Textiles
Food
Manufactured Metals
Minerals
Chemicalsb
Basic Metals
Paper
Export
Intensitya
0.67
0.60
0.52
0.43
0.41
0.35
0.33
0.27
% of Industry Exports
to Developed Nations
58%
75%
64%
49%
47%
47%
32%
19%
No. of
Obs.
1,860
1,816
956
892
275
1,065
144
147
Notes: Data compiled from the United Nations Commodity Trade Statistics 1994. (a) Export Intensity is the mean export
intensity of all exporting firms in the industry. (b) Firms in the petroleum industry are omitted as they were large outliers.
This resulted in the removal of twelve plant-year observations.
30
Table 13: Counterfactual Experiments - Robustness
Experiment
Autarky
No Trade
No FDI
Industry
Assumption
% ∆Ā
% ∆W̄
% ∆Ā
% ∆W̄
% ∆Ā
% ∆W̄
A1
-13.8
-4.1
-2.7
-4.0
-10.5
-1.3
Food
A2
-10.5
-3.2
0.8
-3.1
-8.4
-1.0
A3
-17.1
-2.0
-4.3
-1.9
-3.5
-1.3
A1
-17.8
-5.2
-1.4
-2.7
-14.1
-0.1
Metals
A2
-38.7
-5.0
-21.9
-0.1
-38.6
-5.0
A3
-5.7
-7.3
-4.4
-7.2
-0.1
0
A1
-32.8
-13.1
-31.6
-13.1
-8.0
-0.1
Textiles
A2
-9.1
-14.1
-3.7
-14.1
-1.7
-12.2
A3
-19.5
-15.3
-18.7
-15.3
-12.2
0
Notes: a) The percentage change in average productivity for all plants in the steady state. Calculated using plant-level
revenue shares as weights and evaluated relative to the estimated model. b) The welfare impact must be multiplied by
industry share δ. The parameter estimates that generated these results are available in the Supplemental Appendix.
References
[1] Agung, J., B. Kusmiarso, B. Pramono, E.G. Hutapea, A. Prasmuko, N. Joko Prastowo,
“Credit Crunch In Indonesia In the Aftermath of the Crisis: Facts, Causes and Policy
Implications,” Directorate of Economic Research and Monetary Policy. (2007) Bank Indonesia.
[2] Aitken, B. and A. Harrison, “Do domestic firms benefit from direct foreign investment?
Evidence from Venezuela,” American Economic Review, 89(1999) 605-618.
[3] Alvarez, F., and R.E. Lucas, “General Equilibrium Analysis of the Eaton-Kortum Model
of International Trade,” Journal of Monetary Economics, 54(2007) 1726-1768.
[4] Amiti, M. and J. Konings, “Trade liberalization, intermediate imputs and productivity:
Evidence from Indonesia,” American Economic Review, 97(2007) 1611-1638.
[5] Antras, P. and E. Helpman, “Global Sourcing,” Journal of Political Economy, 112(2004)
552-580.
[6] Arkolakis, C., A. Costinot, and A. Rodrı̀guez-Clare, “New Trade Models, Same Old
Gains?” American Economic Review, (2010) forthcoming.
[7] Arnold, J.M. and B.S. Javorcik, “Gifted Kids or Pushy Parents? Foreign Acquisitions and
Plant Performance in Indonesia,” Journal of International Economics, 79(2009) 42-53.
[8] Aw, B.Y., S. Chung, and M.J. Roberts, “Productivity and Turnover in the Export Market: Micro-level Evidence from the Republic of Korea and Taiwan (China),” World Bank
Economic Review, 14(2000) 65-90.
[9] Baily, M.N., C. Hulten, and D. Campbell, “Productivity Dynamics in Manufacturing
Plants,” Brookings Papers on Economic Activity: Microeconomics, (1992) 67-119.
[10] Bernard, A.B. and J.B. Jensen, “Exceptional Exporter Performance: Cause, Effect, and
or Both?” Journal of International Economics, 47(1999) 1-25.
31
[11] Bernard, A.B., J. Eaton, J.B. Jensen, and S. Kortum, “Plants and Productivity in International Trade,” American Economic Review, 93(1999) 1268-1290.
[12] Blalock, G. and P.J. Gertler, “Learning from Exporting Revisited in a Less Developed
Setting,” Journal of Development Economics, 75(2005) 397-416.
[13] Blalock, G. and P.J. Gertler, “Welfare gains from Foreign Direct Investment through
technology transfer to local suppliers,” Journal of International Economics, 74(2008) 402421.
[14] Blonigen, B.A., “In search of substitution between foreign production and exports,” Journal of International Economics, 53(2001) 81-104.
[15] Blonigen, B.A., R.B. Davies, and K. Head, “Estimating the Knowledge-Capital Model of
the Multinational Enterprise: Comment,” American Economic Review, 93(2003) 980-994.
[16] Brainard, S.L., “An Empirical Assessment of the Proximity-Concentration Trade-off Between Multinational Sales and Trade,” American Economic Review, 87(1997) 520-544.
[17] Burstein, A., and A. Monge-Naranjo, “Foreign Know-How, Firm Control, and the Income
of Developing Countries,” Quarterly Journal of Economics, 124(2009) 149-195.
[18] Chor, D., F. Foley, and K. Manova, “Host Country Financial Development adn MNC
Activity,” Working Paper Stanford University.
[19] Clausing, K.A., “Does multinational activity displace trade?” Economic Inquiry, 38(2000)
190-205.
[20] Clerides, S., S. Lach, and J.R. Tybout, “Is Learning by Exporting Important? MicroDynamic Evidence from Columbia, Mexico and Morocco,” Quarterly Journal of Economics, 113(1998) 903-947.
[21] Doing Business: Comparing Business Regulations, World Bank. (2006).
http://www.doingbusiness.org/.
[22] Das, S., M.J. Roberts, and J.R. Tybout, “Market Entry Costs, Producer Heterogeneity,
and Export Dynamics,” Econometrica, 75(2007) 837-873.
[23] di Giovanni, J., and A. Levchenko, “Firm Entry, Trade, and Welfare in Zipf’s World,”
(2010) NBER Working Paper 16313.
[24] Eaton, J. and S. Kortum, “Technology, Geography and Trade,” Econometrica, 70(2002)
1741-1779.
[25] Eaton, J., S. Kortum and F. Kramarz, “Dissecting Trade: Firms, Industries, and Export
Destinations,” American Economic Review, Papers and Proceedings, 94(2004a) 150-154.
[26] —–, “An Anatomy of International Trade: Evidence from French Firms,” Econometrica,
79(2011) 1453-1498
[27] Fieler, A.C., “Non-Homotheticity and Bilateral Trade: Evidence and a Quantitative Explanation,” Econometrica, 79(2011): 1069-1101.
32
[28] Feinberg, S.E., and M.P. Keane, “Accounting for the Growth of MNC-Based Trade Using
a Structural Model of U.S. MNCs,” American Economic Review, 96(2006) 1515-1558.
[29] Garetto, S. “Input Sourcing and Multinational Production,” mimeo, (2010) Boston University.
[30] Heckman, J.J., “The Incidental Parameters Problem and the Problem of Initial Conditions
in Estimating A Discrete Time-Discrete Data Stochastic Process and Some Monte Carlo
Evidence,” in: C.F. Manski and D.L. McFadden (Eds.), Structural Analysis of Discrete
Data With Econometric Applications, M.I.T. Press, 1981.
[31] Grossman, G.M., E. Helpman and A. Szeidl, “Optimal Integration Strategies for the Multinational Firm,” Journal of International Economics, 70(2006) 216-238.
[32] Head, K. and J. Ries, “Overseas Investment and Firm Exports,” Review of International
Economics, 9(2001) 108-122.
[33] Helpman, E., M.J. Melitz and S.R. Yeaple, “Export versus FDI with Heterogeneous
Firms,” American Economic Review, 94(2004) 300-316.
[34] Hsieh, Chang-Tai and Peter J. Klenow. “Misallocation and Manufacturing TFP in China
and India,” Quarterly Journal of Economics, 124(2009) 1403-1448.
[35] Hill, H., “Foreign Investment and Industrialization in Indonesia,” East Asian Social Science
Monographs, Oxford University Press, Singapore, 1988.
[36] International Labour Organization Bureau of Statistics. (2007). http://ilo.laborsta.org/.
[37] Irarrazabal, A., A. Moxnes and L.D. Opromolla, “The Margins of Multinational Production and the Role of Inter-firm Trade,” (2010) Working Paper.
[38] Ito, T. and K. Sato, “Exchange Rate Changes and Inflation in Post-Crisis Asian
Economies: VAR Analysis of the Exchange Rate Pass-through,” NBER Working Paper
12395, (2006).
[39] Javorcik, B.S. “Does Foreign Direct Investment Increase the Productivity of Domestic
Firms? In Search of Spillovers through Backward Linkages,” American Economic Review,
93(2006) 605-627.
[40] Kasahara, H. and B. Lapham, “Productivity and the Decision to Import and Export:
Theory and Evidence,” University of Western Ontario Working Paper. (2006).
[41] Lu, D., “Exceptional Exporter Performance?
Firms,” University of Chicago, (2010).
Evidence from Chinese Manufacturing
[42] Markusen, J.R. and K.E. Maskus, “Multinational Firms: Reconciling Theory and Evidence,” NBER Working Paper 7163, (1999)
[43] —–, “Discriminating Among Alternative Theories of the Multinational Enterprise,” Review
of International Economics, 10(2002) 694-707.
33
[44] McFadden, D., “Modeling the Choice of Residential Location,” in: A. Karlquist et al.
(Eds.), Spatial Interaction Theory and Residential Location, North- Holland, Amsterdam,
1978.
[45] McGrattan, E.R., and E.C. Prescott, “Openness, Technology Capital, and Development,”
Journal of Economic Theory, 144(2007) 2454-2476.
[46] Melitz, M.J., “The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity,” Econometrica, 71(2003) 1695-1725.
[47] International Direct Investment Statistics Online Database, OECD. (2006)
http://www.sourceoecd.org/.
[48] Pangestu, M., “Economic Reform, Deregulation and Privitization: The Indonesian Experience” Centre for Strategic and International Studies, Jakarta, 1996.
[49] Pavcnik, N. “Trade Liberalization, Exit, and Productivity Improvements: Evidence From
Chilean Plants,” Review of Economic Studies, 69(2006) 245-276.
[50] Ramondo, N. “A Quantitative Approach to Multinational Production,” mimeo. (2011).
https://sites.google.com/site/nataliaramondo/home/research.
[51] Ramondo, N. “Foreign Plants and Industry Productivity: Evidence from Chile,” Scandanavian Journal of Economics. 111(2009): 789-809.
[52] Ramondo, N. and A. Rodrı́guez-Clare, “Trade, Multinational Production, and the Gains
from Openness,” mimeo. (2010) http://www.econ.psu.edu/∼aur10/research.htm.
[53] “Ramstetter, E.D., and F. Sjoholm, “The Issues: Multinationals and the questions they
raise,” in: E.D. Ramstetter and F. Sjoholm (Eds.), Multinational Corporations in Indonesia and Thailand, Palgrave MacMillan, New York, 2006.
[54] “Restuccia, Diego and Richard Rogerson, “Policy Distortions and Aggregate Productivity
with Heterogeneous Plants,” Review of Economic Dynamics, 11(2008), 707-720.
[55] Rob, R. and N. Vettas, “Foreign Direct Investment and Exports with Growing Demand,”
Review of Economic Studies, 70(2003) 629-648.
[56] Roberts, M.J. and J.R. Tybout, “Decision to Export in Columbia: An Empirical Model
of Entry with Sunk Costs,” American Economic Review, 87(1997) 545-564.
[57] Rodrik, D., “The new global economy and developing countries: making openness work,”
Policy Essay 24, Overseas Development Council. Johns Hopkins University Press, Washington, DC.
[58] Ruhl, K.J., and J.L. Willis, “Convexities, Nonconvexities, and Firm Export Behaviour,”
mimeo. (2007). http://www.vanderbilt.edu/econ/sempapers/Ruhl.pdf
[59] Rust, J., “Optimal replacement of GMC bus engines: an empirical model of Harold
Zurcher,” Econometrica, 55(2006) 999-1033.
34
[60] —–, “Estimation of Dynamic Structural Models, Problems and Prospects: Discrete Decision Processes,” in: C.A. Sims (Eds.), Advances in Econometrics Sixth World Congress,
Volume II, Cambridge University Press, New York, 1994.
[61] Svensson, R., “Effects of overseas production on home country exports: Evidence based
on Swedish multinationals,” Review of World Economics, 132(1996) 304-329.
[62] Trefler, D., “The Long and Short of the Canada-U.S. Free Trade Agreement,” American
Economic Review, 94(2004) 870-895.
[63] United Nations, Commodity Trade Statistics 1994, United Nations Publishing Division,
New York, 1995.
[64] Waugh, M., “International Trade and Income Differences,” American Economic Review,
100(2010) 2093 - 2124.
[65] Wie, T.K., “Intra-regional investment and technology transfer in Indonesia,” in: K. Yanagi
(Ed.), Symposium on Intra-Regional Investment and Technology Transfer, Asian Productivity Organization, Kuala Lumpur, 1994.
35