Outsourcing Decisions: Theory and Evidence

By

Aekapol Chongvilaivan* and Shandre M. Thangavelu**

8 January 2013

Preliminary Draft

Abstract

This paper attempts to theoretically and empirically examine outsourcing decisions. We

develop a partial equilibrium model of outsourcing decisions in which heterogeneous firms

face choices between disintegration and internalisation. We show that more productive firms

are inclined to contract out production activities at arm’s length. The key driver of this result

rests with the spirit of Grossman and Helpman (2002) who posited that an outsourcing

decision involves a trade-off between efficiency gains from core competences and

specialisation and transaction costs in terms of searching for potential suppliers and

contractual enforcement. Having established the theoretical proposition, we then turn

attention to an empirical investigation of international outsourcing determinants using the

establishment-level data of Thailand’s manufacturing industries. We employ the Probit model

of international outsourcing decisions taking into account various econometric issues. These

include the endogeneity bias problem, unobserved industry-specific characteristics, and

heteroscedasticity. The consistent estimates reveal a positive correlation between productivity

levels and probability that the outsourcing decision is made, and therefore sheds light on our

theoretical prediction. In addition to productivity levels, the empirical framework highlights

several establishment-specific characteristics as a key determinant of international

outsourcing. These comprise their export statuses, foreign ownership

                                                            *

Department of Economics, National University of Singapore, AS2 Arts Link 1, Singapore 117570, Tel: +65-6516-4524. Email address: ecsac@nus.edu.sg. **

Corresponding author: Department of Economics, National University of Singapore, AS2 Arts Link 1, Singapore 117570, Tel: +65-6516-6853. Email: ecssmt@nus.edu.sg.

1 Introduction

Outsourcing has become an essential part of business activities in the era of globalisation.

Firms for the time being contract out what they once did for themselves with an ever

expanding set of production activities, including the upstream stages, such as material and

service inputs, human resource department, and research and development activities, and

downstream stages like delivery, marketing and packaging. This trend of organisational

disintegration has attracted substantial research inquiries into its economic impacts. For

instance, much research done by Feenstra and Hanson (1996, 1999), Anderton and Brenton

(1999), Geishecker (2002), and Hsieh and Woo (2005), among others, has been devoted to

the impacts of international outsourcing on the shifts in the relative demand for skilled

workers. Recent studies such as Egger and Egger (2006), Görzig and Stephan (2002), and

Görg and Hanley (2004, 2005) in contrast weigh in on the relationships between international

outsourcing and firm performance. Nevertheless, little has been done to push forward the

understanding of international outsourcing decisions. The insight into this subject is equally

indispensable as it offers the policy tools to harness the economic impacts of international

outsourcing.

This paper is among the very first attempts to theoretically and empirically examine

outsourcing decisions. We develop a partial equilibrium model of outsourcing decisions in

which heterogeneous firms face choices between disintegration and internalisation. We show

that more productive firms are inclined to contract out production activities at arm’s length.

The key driver of this result rests with the spirit of Grossman and Helpman (2002) who

posited that an outsourcing decision involves a trade-off between efficiency gains from core

competences and specialisation and transaction costs in terms of searching for potential

suppliers and contractual enforcement.

Having established the theoretical proposition, we then turn attention to an empirical

investigation of international outsourcing determinants using the establishment-level data of

Thailand’s manufacturing industries. We employ the Probit model of international

outsourcing decisions taking into account various econometric issues. These include the

endogeneity bias problem, unobserved industry-specific characteristics, and

heteroscedasticity. The consistent estimates reveal a positive correlation between productivity

levels and probability that the outsourcing decision is made, and therefore sheds light on our

theoretical prediction. In addition to productivity levels, the empirical framework highlights

several establishment-specific characteristics as a key determinant of international

outsourcing. These comprise their export statuses, foreign ownership, and new entrance to the

markets, on top of market size, and capacity utilisation.

The organisation of this paper can be structured as follows. Section 3.2 elaborates a

partial equilibrium model of outsourcing decisions incorporating firm heterogeneity. The

solutions and analyses are given at the end of this section. Section 3.3 examines various

characteristics and international outsourcing activities in Thailand’s manufacturing industries

and introduces the empirical framework of outsourcing decisions. Section 3.4 presents and

discusses the empirical evidence. Section 3.5 concludes.

2 The Model of Outsourcing Decisions

This section attempts to develop a simple, partial equilibrium model of outsourcing decisions

– the arrangements that entail contractual costs but allow firms to tap on specialised skills

possessed by external contractors. The theoretical framework builds upon the spirit of

Grossman and Helpman (2002), incorporating firm heterogeneity in terms of differences in

productivity levels. A testable hypothesis the model highlights is that more productive firms

tend to resort to outsourcing than do less productive ones. The econometric framework and

estimations introduced later in this paper aims to empirically investigate this base-line

theoretical result.

2.1 Demand

Consider an economy in which a continuum of varieties i is produced. As in Dixit and Stiglitz

(1977), the love-of-variety preference of a representative consumer in each country is

assumed to take a constant elasticities of substitution (CES) functional form whereby

consumption spans over a continuum of goods indexed by i.

( )[ ] ρρ 1

∫ Ω∈=

idiiqU , (1)

where ( )1,0∈ρ and the set of Ω represents the mass of available varieties. It is noteworthy

that the assumption ( )1,0∈ρ implies goods Ω∈i are substitutes, and thus an elasticity of

substitution between two goods lying in the set of Ω is straightforwardly represented

by ( ) 111 >−= ρσ . Accordingly, an aggregate price or price index in the economy can be

portrayed as:

( )[ ] σσ −Ω∈

−∫= 11

1

idiipP . (2)

With this aggregate price, we can derive a demand and expenditure functions for each good i

by employing the utility maximization conditions. Let E denote an aggregate expenditure in

each country. The Walrasian demand function can be derived as:

( )σ−

⎟⎠⎞

⎜⎝⎛=

Pip

PEiq )( . (3)

It is straightforward to see that the demand for the good i in (3) well behave since it is

negatively correlated with its own price p(i) and positively correlated with the aggregate

expenditure E. In addition, it also captures substitution between the good i and other goods j,

where Ω∈j and ij ≠ , now that an increase in the aggregate price P results in an outward

shift in the demand for the good i. Using the Walrasian demand function (3), we can also

derive the expenditure spent on good i denoted by r(i).1

σ−

⎟⎠⎞

⎜⎝⎛=

1)()(PipEir , (4)

where ∫ Ω∈=

idiirE )( .

2.2 Production

A continuum of firms operates in the monopolistic competition market without entry barriers.

For simplicity, we assume that the production technology is linear and requires only one

factor of production – labour. Firms are heterogeneous with respect to their inherent

productivity, indexed by ( )∞∈ ,0θ . Firm productivity is captured by the constant marginal

cost of producing q units of final outputs, whereby more productive firms can produce the

same amount of final outputs at lower marginal cost than less productive firms. In particular,

the amount of labour needed to produce q units of goods by a firm θ is represented as

                                                            1 Since each producer manufactures differentiated product indexed by the good i, the expenditure spent on each variety is also revenue for each heterogeneous firm.

( )θ

θ qvfql vvv +=, , (5)

where 0, >vv vf . The constant parameter vf captures the fixed production cost and thus

represents the number of workers a firm has to employ, regardless of the final goods

produced. The θvv workers are needed to produce an additional unit of final goods. Ones

may observe that a more productive firm (with a higher value ofθ ) has lower marginal cost

of production than a less productive firm (with a lower value ofθ ), ceteris paribus.

As Grossman and Helpman (2002) pointed out, two conceptual frameworks account

for outsourcing decisions. The first is the transaction costs theory by Coarse (1937), which

has dominated the literature on outsourcing decisions, such as Williamson (1975 and 1986),

Klein, et al. (1978), Monteverde and Teece (1982), and Murray, et al. (1995). This theoretical

proposition articulates that outsourcing pertains to the prohibitive costs of turning to the

market which in turn induce firms to carry out production activities in-house, instead of

disintegrating production. This problem, according to the transactions cost theory, emanates

from either: (1) fewness of firms in the market, ex ante; or (2) asset specificity that causes a

lock-in problem between buyer and seller, ex post. The former implies that a decision to

contract out production at arm’s length incurs costs of searching potential providers. The

latter in contrast obliges firms costly investment in commitment technology, such as cost of

running more complex organisations and enforcing contracts,  since lock-in effects and

transaction-specific capital put quasi-rents at risk of being appropriated ex post by their

contractual partners. These transaction costs are independent of the amount of goods

produced and thus taken by c.

The other has to do with the core competences approach to outsourcing decisions,

where firm competitiveness arises from outsourcing (Grant, 1991; and Gainey and Klaas,

2003). This theoretical explanation points out that improved productivity performance in a

firm can be attained through specialisation in production activities that provide core

competences and contracting out the rest. The productivity enhancement stems from the

extent to which outsourcing enables firms to procure what they do not know how to do and at

the same time develop in-house what they do better than the suppliers (Argyres, 1996). The

productivity gains from a decision to outsource are captured by the reduction of marginal cost

by g.

Accordingly, the labour demand function that takes into consideration transaction

costs and enhancement of productivity performance associated with a decision to rope in

outsourcing can be expressed as:

( )θ

θ qvfql ooo +=, , (6)

where vvo fcff >+= and vvo vgvv <−= .

Since the preference of a representative consumer is assumed to be a CES utility

function over a continuum of varieties, it is apparent from (3) that the elasticities of demand

for the variety i uniquely produced by the firm θ is equal to σ . Hence, ones can derive the

mark-up pricing condition using the profit maximization condition as follows:

ρθθ k

kv

p =)( , (7)

where ovk ,= . By substituting (7) into (4), the revenue function of a firm θ can be shown as:

1

)(−

⎥⎦

⎤⎢⎣

⎡=

σρθθk

k vPEr . (8)

The exogenously given wage rate is normalised to unity. Accordingly, by using (5), (6), and

(7) and simple manipulations, ones can show that a firmθ ’s profit function is:

kk

k fv

PE−⎥

⎦

⎤⎢⎣

⎡=

−1

)(σ

ρθσ

θπ , (9)

where ovk ,= .2 

2.3 Outsourcing Decisions

Having solved for the firm’s profit functions, we then turn attention to the solutions to

optimal outsourcing decisions. The firm’s objective is to choose the production structure (k)

such that its profits are maximised. Specifically, the firm’s objective function can be written

as follows.

                                                            2 In fact, the aggregate expenditure (E) and aggregate price (P) are exogenous to the firms in a partial equilibrium setting like ours. However, in the general equilibrium setting, both variables are endogenously determined by free exit/entry and stability conditions.

( )⎪⎭

⎪⎬⎫

⎪⎩

⎪⎨⎧

−⎥⎦

⎤⎢⎣

⎡−⎥

⎦

⎤⎢⎣

⎡=

−−

=o

ov

vkovk

fv

PEfv

PEMaxMax11

,,,0

σσρθ

σρθ

σθπ (10)

It should be highlighted that the outsourcing decisions in (10) depend upon the endowed

productivity levels (θ ). If vk = , the production takes place in-house (no outsourcing). If

ok = , the firm decides to contract out production activities. In the latter case, the firm bears

transaction costs and enjoy productivity improvements. The solutions to the profit

maximisation problem (10) are characterised by the cut-off productivity levels that help

identify the optimal outsourcing decisions for each firmθ .

[Insert Figure 3.1 here]

Figure 3.1 illustrates the first equilibrium in which some firms internalise production

in-house while the others make use of outsourcing. Two cut-off productivity levels are

concerned with this equilibrium. One is the zero-profit cut-off ( vθ ) where the form is

indifferent between market entry and exit. Thus, vθ is the firm’s productivity level such

that ( ) 0=θπ v . The solution to this equality takes the following expression.

11−

⎟⎠⎞

⎜⎝⎛=

σσρ

θE

fPv

vv

v (11)

The other is the cut-off productivity level ( oθ ) above which the firms downsize their

production structure by outsourcing, instead of integrating production. This cut-off must

therefore satisfies ( ) ( )θπθπ ov = and can be shown as

11

1111)(1

−

−− ⎪⎪

⎭

⎪⎪

⎬

⎫

⎪⎪

⎩

⎪⎪

⎨

⎧

⎥⎦

⎤⎢⎣

⎡−

−=

σ

σσ

σρ

θ

vo

voo

vvE

ffP

.      (12)

The existence of vθ and oθ is not straightforward, nevertheless. The reason is that

some firms, which find outsourcing more profitable than internalising the whole production

process, may not be able to stay active in the market unless they are sufficiently productive.

To rule out which case, ones can show that the following condition must be satisfied.3

( ) cg +<+ − 11 1σ (13)

If productivity gains from outsourcing (g) are minimal, relative to transaction costs

(c), then some firms choose to outsource, while some other firms internalise production.

While less productive firms θ , where [ ]ov θθθ ,∈ , refrain from outsourcing activities and

undertake all production stages in-house, more productive firms θ , where )[ ∞∈ ,oθθ , resort

to contracting out production. Intuitively, a decision to outsource has to do with transaction

costs in terms of ex ante outlay of searching potential suppliers and ex post contractual

enforcement, which are not related to production scale. If the reduction in marginal

production costs as a result of outsourcing is comparatively limited, the firms must be

sufficiently productive such that the huge production scale enables them to overcome the

transaction costs.

[Insert Figure 3.2 here]

Figure 3.2 portrays the other characterisation of the equilibrium. If productivity gains

from outsourcing (g) are sufficiently large such that the condition (13) is violated, the

positive correlation between firm productivity and the decision to downsize production

structure still holds, but all firms find outsourcing more profitable than in-house production.

The intuition is straightforward. If outsourcing considerably plunges marginal costs, then

even the least productive firms that remain in the market can tap on contracting out

production activities. In this case, the only relevant cut-off productivity level isθ in which a

decision to outsource produces zero profit. The expression of θ can be easily shown as

11−

⎟⎠⎞

⎜⎝⎛=

σσρ

θE

fPv

oo . (14)

As represented in Figure 2, when efficiency gains from outsourcing are pronounced,

only outsourcing firms θ , where [ )∞∈ ,θθ , exist in the market. Again, only productive

firms can enter the market and make use of the outsourcing strategy.                                                             3 As shown in Figure 3.1, the existence of vθ and oθ requires that ov θθ < . This holds only if the intersection

point between the oθ nexus and the horizontal axis lies in the right hand side of vθ .

To summarise, not all firms find outsourcing options profitable. Our simple, partial

equilibrium model articulates that more productive firms tend to leverage on outsourcing. A

key catalyst of this theoretical proposition rests with the transaction costs (e.g. costs of

searching potential providers and contract enforcement) and efficiency gains (e.g. improved

competitiveness and ability to tap on core competences) associated with outsourcing

activities. To find an outsourcing decision profitable, a firm must be sufficiently productive

such that a large production scale allows gains from declines in marginal costs to dominate

the incurred transaction costs.

3.3 The Empirical Framework

This section develops an empirical test of our theoretical model in which a firm’s inherent

productivity serves as a crux determinant of outsourcing decisions. It should be highlighted

that our focus in this paper abstracts from an attempt to empirically investigating a causal

relationship between outsourcing and firm-level productivity as a relationship between

outsourcing decisions and firm productivity is two-folded. On the one hand, our theoretical

framework establishes that more productive firms are inclined to make use of outsourcing.

Much research, on the other hand, points to the reversed causality – the effects of outsourcing

on firm performance, such as Görzig and Stephan (2002), Girma and Görg (2004), Egger and

Egger (2006), and Amiti and Wei (2009). This section zeroes in on examining determinants

of outsourcing decisions.

Another noteworthy issue pertains to the scope of the outsourcing definition. While

the notion of outsourcing in the theoretical framework is loosely defined as downsizing the

in-house production scale and thus includes both domestic and international outsourcing, this

section puts emphasis specifically on the decisions on international outsourcing since the

primary objective of this book is to deepen the understanding of international outsourcing in

the context of the Southeast Asian economies.

3.1 Overview of Data

The empirical framework employs the establishment-level data retrieved from the report of

the Manufacturing Industry Survey in 2003 provided by the National Statistical Office

(NSO), Thailand.4 The dataset includes basic information on manufacturing establishments,

such as the numbers of employees, raw materials, parts and components, sales values,

                                                            4 From this point onwards, we shall substitute ‘firm’ by ‘establishment’.

inventory and fixed assets, among others. The establishments under the scope of this survey

are those engaged primarily in manufacturing industries (Category D of the International

Standard Industrial Classification – ISIC Rev.3). The dataset classifies the sample

establishments into 6 sizes based on the numbers of employees.5 They are located in the

premises throughout Thailand, including Bangkok and Pattaya, on top of municipal areas and

non-municipal areas that form as a sub-district administration organization.

Data collection weighs in on interviews. The enumerators who are permanent and

temporary staff of NSO were sent out to interview the owners or the entrepreneurs of the

sample manufacturing establishments. According to the survey, manufacturing industries are

classified according to the 4-digit ISIC levels. The dataset covers the 64 types of

manufacturing activities (4-digit codes) in 21 industries (2-digit codes) and builds upon a

Stratified Systematic Sampling method. The selection of sample establishments is carried out

separately and independently in each type of manufacturing activities.

[Insert Table 1 here]

Table 1 summarises the sample establishments classified by regions. The total

numbers of the sample establishments amount to 8,730 establishments. The manufacturing

establishments are intensively located in the vicinity and central (38.8 percent) areas followed

by the southern areas (20.7 percent), northern areas (17.7 percent), northeast areas (12.5

percent), and Bangkok (10.4 percent).

3.2 Econometric Model

This sub-section develops an econometric model that conveys empirical evidence that

establishes a relationship between establishment-level productivity and outsourcing

decisions. In so doing, we consider the following econometric specification:

( ) ( )iiiiiiiiii CapuSizeXCapuSizeXOut λβββθβαθ +++++Φ== 4321,,,1Pr ,             (15) 

where ( )•Pr denotes probability, and ( )•Φ is the cumulative distribution function (CDF). The

binary variable of outsourcing decisions ( iOut ) takes the value of unity if an establishment i

deals with international outsourcing; it takes the nil value otherwise. iθ represents the

establishment-level productivity. iX is a vector of establishment-specific characteristics. We

                                                            5 Specifically, the sizes of establishments are segregated into those with 1-15, 16-25, 26-30, 31-50, 51-200, and more than 200 persons.

control for market size and capital utilisation designated by iSize and iCapu , respectively.

The vector of 2-digit ISIC industry dummies ( iλ ) enters the specification to control for the

industry-specific effects.

It should be highlighted that the econometric model (15) can also be seen as a latent

variable model where a latent (unobserved) variable ( *iOut ) is regressed on the independent

variables.

iiiiiii CapuSizeXOut ελβββθβα ++++++= 4321* , (16)

where⎪⎩

⎪⎨

⎧

≥

<=

01

00

*

*

i

i

i

Outif

OutifOut , and the error term ( iε ) is assumed to be independently and

identically distributed (iid).

[Insert Tables 2 and 3 here]

Table 2 portrays detailed information on outsourcing decisions in Thailand’s

manufacturing industries. Out of 8,730 establishments, 1,677 establishments thrived on

international outsourcing. This table also breaks down the sample establishments according to

their intensity of international outsourcing activities. The breakdown indicates that the

intensity of international outsourcing activities is evenly distributed among Thailand’s

manufacturing establishments as approximately 6-15 percent of samples are found in all

ranges of outsourcing intensity. Table 3 reveals the number and percentage of establishments

that contract out their production to foreign providers. It can be observed that international

outsourcing seems to be the most prevalent among the relatively advanced industries in

which approximately 40-50 percent of establishments are concerned with international

outsourcing. These industries include communication equipment and apparatus, refined

petroleum products, electronic machinery and apparatus, and office, accounting and computer

machinery. These figures are much higher than the traditional, labour-intensive sectors where

the proportion of outsourcing establishments takes the value merely 10-15 percent, such as

food and beverages, textiles, wearing apparel (including dressing and dyeing of fur),

publishing, printing and reproduction of recorded media, and non-metallic mineral products.

This observation sheds clearer light on our hypothesis that international outsourcing decisions

have to do with establishment-level productivity now that the advanced industries tend to

exert higher productivity.

Central to our analysis is productivity levels of establishments ( iθ ). We measure the

establishment-level productivity by outputs per worker, e.g. iii ly=θ , where iy is the value

of outputs produced, and il is the number of workers employed in-house. This index of

establishment productivity is identical to that of Bernard and Jensen (1999), Fixler and Siegel

(1999), Girma and Görg (2002), and Egger and Egger (2006). Our theoretical framework

developed in Section 2 establishes that outsourcing decisions hint upon productivity levels.

We argued that a decision to procure intermediate inputs from markets reduces production

costs (in terms of the amount of labour required to produce the same units of final outputs),

thanks to increased core competences and specialisation. However, reliance on the market, as

opposed to in-house production, incurs transaction costs in terms of searching potential

providers and contract enforcement. In this sense, only sufficiently productive producers find

outsourcing profitable as their large market shares allow a plunge in production costs to

dominate the arising transaction costs.

Three establishment-specific characteristics in the vector iX are taken into account,

including foreign ownership ( iFhold ), exporters ( iExport ) and new entry to the markets

( iNew ). These characteristics enter the specification (16) as dummies. iFhold , iExport

and iNew take the value of unity if the establishments are foreign-owned, exporting, and new,

respectively; they take the nil value otherwise. The first two dummies aim to control for

exposure to international markets in that foreign-owned and exporting establishments tend to

find international outsourcing easier than do domestically owned and non-exporting

establishments, thanks to their superior knowledge and/or information on foreign markets and

input providers (Markusen, 1995 and Girma and Görg, 2002). The last characteristic – new

entry to the markets – may hinder outsourcing decisions due to inferior knowledge about the

markets and production technologies, compared to incumbent establishments.

[Insert Table 3.4 here]

Table 3.4 reveals various attributes of establishments in Thailand’s manufacturing

industries that may influence their outsourcing decisions. The first column reports the

numbers of foreign-owned establishments. The numbers of establishments that carry out

exporting activities are shown in the second column. The last column exposes the numbers of

new establishments. A new establishment is defined as one which has operated for less than

12 months. In these columns, the percentages of the total numbers of establishments appear in

parentheses.

Out of 8,730 samples, there were 934 foreign-owned establishments. Industrial

decomposition further demonstrates that advanced manufacturing industries tend to have

relatively high foreign ownership, particularly office, accounting and computing machinery

(42.86 percent), communication equipment and apparatus (41.84 percent), electrical

machinery and apparatus (29.08 percent), and motor vehicles, trailers and semi-trailers (27.27

percent). Nevertheless, the figures are relatively low for labour-intensive industries including

wood and cork (3.60 percent), tobacco (4.26 percent), publishing, printing and reproduction

of recorded media (4.31 percent), and food and beverages (5.24 percent).

Excluding publishing, printing and reproduction of recorded media and tobacco,

Thailand’s manufacturing industries are export-oriented with the proportion of exporting

firms ranging from 12.80 percent in the machinery and equipment industry to 44.68 percent

in the communication equipment and apparatus industry. The particularly large proportions of

exporting establishments in some advanced industries, like communication equipment and

apparatus (44.68 percent), office, accounting and computing machinery (35.71 percent) and

electrical machinery and apparatus (33.33 percent), suggest that comparative advantages of

Thailand’s manufacturing sector have departed from traditional, labour-intensive production,

like food and beverages and textiles, and rested with sophisticated, capital-intensive

production.

The other establishment-specific characteristic our empirical framework attempts to

control for is whether an establishment is a new entrant in the market. In our dataset, the new

entrants on average amount to roughly 15 percent of sample establishments. The breakdown

by industries further reveals that the new entrants are concentrated intensively in labour-

intensive industries like food and beverages (34.16 percent) and textiles (13.39 percent).

These figures are much higher than those for relatively advanced industries, like electrical

machinery and apparatus, communication equipment and apparatus, medical precision and

optical instruments, motor vehicles, trailers and semi-trailers, and transportation equipment,

where the new entrants make up less than one percent.

Market size denoted by iSize is measured by the ratio of an establishment i’s asset

values to total asset values of all establishments in the same 4-digit industries. It is meant to

control for scale economies adjusted by the degree of competitiveness in the final output

markets. Intuitively, market size may have a say in international outsourcing decisions since

the magnitude of efficiency gains is dependent upon their market size. In addition, large

establishments are able to exploit scale economies to compensate the incurred transaction

costs associated with international outsourcing. In this sense, ones would expect that large

establishments are more likely to tap on international outsourcing to enhance their

competiveness.

Last but not least, the rate of capacity utilisation ( iCapu ) proxied by the ratio of fuel

and electricity costs to total values of machinery is introduced to account for the possibilities

that production capacity may be varied at least in short run. If this is the case, a short-run

swing of production capacity may alter outsourcing decisions. For instance, a positive

demand shock to a final output market may force an establishment to temporarily over-utilise

the exiting production factors. As the production scale expands, a decision to contract out

production activities would offer greater efficiency gains. Hence, we hypothesises that a

relationship between capacity utilisation and international outsourcing decisions exhibits a

positive sign.

3.3.3 Model Estimations

Since the dependent variable in the base-line econometric specification (15) is binary, the

simplest approach to the model estimation is perhaps to employ the Probit model in which the

estimators are obtained by the Maximum Likelihood (ML) estimation. Nevertheless, at least

three concerns regarding consistency of the estimates arise and must be addressed.

First and most important, even though the establishment productivity level ( iθ ) is

assumed to be exogenous in the theoretical model, it is at best questionable whether this

assumption indeed holds empirically. As is well-known, if the regressors are endogenously

determined by other unobservables, the ML estimates are biased and inconsistent. The

potential endogeneity problem seems imperative as much research has identified a wide array

of drivers of establishment-level productivity performance, such as investment in human

capital, organisational management, and uses of information technology (Black and Lynch,

2001; Griliches, 1997; and Oliner and Sichel, 1994, among others). To tackle this issue, the

productivity levels ( iθ ) is instrumented by two instrumental variables (IVs): skill intensity

( iSkill ) and value-added contributed by workers ( iVadd ). The ratio of skilled to total workers

serves as a proxy of skill intensity while the latter is measured by the ratio of an

establishment’s value-added to total workers.

The instrumental variable (IV) estimations of the Probit model are not

straightforward, nevertheless. We employ Newey’s two-step minimum chi-square estimation

to obtain the efficient estimators of the Probit model with endogenous regressors (Newey,

1987). Based on this approach, the minimum chi-square (MCS) estimators make use of the

initial estimates of the reduced form parameters and the restrictions imposed by the model

structure to estimates all structural parameters. Ferguson (1958) shows that the MCS

estimators are asymptotically efficient in the absence of the restrictions implied by the

structure, and thus the MCS estimation is an efficient method in which reduced form

parameters are employed to estimate the structural parameters. In addition, we also produce

the Wald test of endogeneity using Newey’s two-step estimators. The statistic is chi-square

distributed under the null of no endogeneity. A rejection of the null implies that the ML

estimates are biased and inconsistent, and thus Newey’s MCS estimation is imperative.

The other econometric issue is concerned with the industry-specific fixed effects. Our

dataset covers a wide range of manufacturing activities and industries, and therefore the

absence of other unobservable industry-specific effects may convey biased estimators. A

vector of two-digit ISIC industry dummies ( iλ ) enters the model specification to account for

the industry-specific effects.

The final econometric issue we attempt to address is the heteroscedasticity problem.

We concern that the assumption that the stochastic error term ( iε ) is homoscedastic, is

violated due to the variation of establishment size in our dataset. If this is the case, the

standard estimators of the Probit model are biased. To obtain consistent estimates, we tap on

White’s (1980) heteroscedasticity-robust standard error procedure in the estimation of the

base-line specification (15).

4 Empirical Results

Table 3.5 presents the estimation of the Probit model where all regressors are treated as

exogenous. The parameter estimates are portrayed in terms of the marginal effects, ( )•Φ′ .6

The second column (Probit) reports the maximum likelihood (ML) estimates of the Probit

model without the 2-digit industry dummies, while the last column (FEProbit) reveals the ML

estimates where the 2-digit industry dummies are included. The standard error estimators are

heteroscedasticity-robust for all estimations. The econometric specification (15) is adequate

as Wald’s chi-squared appears to be statistically significant at the 1 percent level across all

estimations and hence reject the null hypothesis that the parameter estimates are jointly equal

to zero. The predictability of the model is satisfactory since the correctly specified rates

amount to 84.69 and 85.3 percent for the Probit and FEProbit results, respectively.7

[Insert Tables 3.5 and 3.6 here]

Table 3.6 produces the Instrumental Variable (IV) estimates based on Newey’s two-

step minimum chi-squared method. The second column denoted by IV presents the IV

estimates of the Probit model without the 2-digit industry dummies whereas the last column

indicated by IVFE refers to the IV estimations in which the proxy of establishment

productivity ( iθ ) is instrumented by two instrumental variables – skill intensity ( iSkill ) and

value-added contributed by workers ( iVadd ). These two variables potentially serve as valid

IVs in that they are strongly correlated with iθ and are likely to be exogenous to the

establishments at least in short run.8 The Wald test of exogeneity, however, fails to reject the

null of no endogeneity bias problem as the statistics are statistically insignificant across all

estimations. This result points to the fact that the ML estimation of the Probit model in Table

5 suffices to convey consistent estimates. The finding that the endogeneity bias problem does

not pose serious challenges to consistency of the estimates seems plausible given the casual

                                                            6 For the establishment-specific dummies represented by the vector iX , the interpretation of the parameter estimates is a discrete probability change of the dummies from 0 to 1. 7 An outsourcing decision is correctly specified when the predicted probability of outsourcing is greater than 0.5 for an outsourcing establishment and is lower than 0.5 for a non-outsourcing establishment. 8 The rationale is that skill intensity and labour contribution to value-added typically count on production technology an establishment ropes in. Since the production technology is unlikely to be volatile at least in short run, these two IVs are arguably exogenous to the establishments.

observation that the parameter estimates are qualitatively unchanged when the establishment

productivity ( iθ ) is treated as endogenously determined.9

The key findings of our empirical exercise based on Table 5 can be summarised as

follows.

First and foremost, we find strong evidence supporting our theoretical proposition that

more productive establishments tend to make use of international outsourcing than do less

productive firms since the coefficients of establishment productivity ( iθ ) are positive and

statistically significant at the 1 percent level across all estimations. Intuitively, even though

contracting out production stages at arm’s length enables producers to tap on core

competences and thus enhance overall production efficiency, reliance on the markets as

opposed to in-house production incurs transaction costs in terms of searching for potential

providers and enforcing contracts. Therefore, outsourcing is not always optimal for all

establishments. The establishments must be sufficiently productive such that their large

market size allows gains from improved production efficiency to dominate incurred

transaction costs.

Exposure to international trade matters to the use of international outsourcing. The

coefficient of iExport appears to be positive and statistically significant at the 1 percent level

across all estimations. Intuitively, exporters have advantage of information on international

contractors and markets over non-exporters and thus stand in good stead to exploit the global

production networks. The positive relationship between the export status and the international

outsourcing decisions is consistent with Feenstra and Hanson (1997) who presented the

evidence from Mexico’s plants that exporters are more likely to contract out production

stages at arm’s length than non-exporters.

Foreign-owned establishments tend to contend with international outsourcing now

that foreign ownership ( iFhold ) has a positive effect on the tendency of international

outsourcing. As manifested in Table 5, the coefficient of iFhold is, again, persistently

positive and statistically significant at the 1 percent level. This finding substantiates the

previous study by Girma and Görg (2002). They produce evidence that nationality of

ownership has a say in the use of international outsourcing. A possible explanation of this                                                             9 We spot slight sensitivity of the parameter estimates in the IVFE column. Although still positive, the coefficient of iRsize turns out to be statistically significant.

evidence is that foreign-owned establishments may have superior knowledge of foreign input

suppliers, better access to the global production networks, and exceptional managerial

technology and organisational structure. These advantages enable foreign-owned firms to

trim their in-house production and procure intermediate inputs from the markets.

Another establishment-specific characteristic that explains international outsourcing

decisions among Thai manufacturers is new entry. The coefficient of the new entry dummy

( iNew ) turns out to be negative and statistically significant at the 1 percent level. Our

estimations therefore point to the evidence that newcomers, compared to incumbent

establishments, are less likely to take on international outsourcing. The insights into this

finding perhaps rest with several advantages possessed by incumbent establishments. These

include richer knowledge about the markets and better access to intermediate input suppliers.

Another possible explanation is that new entrants may have limited capability to contract out

production stages due to the lack of core competences and specialisation. Over time, they

realise where their comparative advantages lie, and are ultimately thrive on international

outsourcing.

Our empirical estimations also control for market size. We find that establishments

with larger market shares are more likely to contract out production activities since the

coefficient of iRsize in Table 5 is consistently positive and statistically significant at least at

the 5 percent level. Intuitively, a decision to downsize in-house production on the one hand

enhances an establishment’s core competencies and thus lower production costs. The decision

forces the establishment to invest in transaction costs in terms of contractual arrangement and

enforcement, on the other. Therefore, the establishment must attain the sufficiently large

market share if it is to tap on international outsourcing.

Last, capacity utilisation seems to be positively correlated with an international

outsourcing decision as the coefficient of iCapu exhibits a positive, statistically significant

sign. The positive correlation between capacity utilisation and outsourcing decisions may be

explained by the fact that higher capacity utilisation implies larger amount of output

produced. The expansion of the production scale makes the use of international outsourcing

desirable to the establishments as improvement of production efficiency rendered by

international outsourcing becomes more pronounced.

5 Policy Conclusions

This paper examines the theory and evidence of international outsourcing decisions. The

simple, partial equilibrium model of outsourcing decisions developed in this paper builds

upon that of Grossman and Helpman (2002). They articulated that contracting out production

stages at arm’s length on the one hand enhances production efficiency as the outsourcing

firms leverage on core competences and improved specialisation. Outsourcing activities, on

the other hand, come at cost in that reliance on the markets, as opposed to in-house

production, incurs transaction costs, in particular costs of searching prospective providers and

investment in contractual enforcement technology. Given this trade-off, our model of

outsourcing decisions incorporating firm heterogeneity shows that more productive firms

tend to rope in outsourcing than less productive ones.

We then empirically investigate the determinants of international outsourcing

decisions, using establishment-level data of Thailand’s manufacturing industries. Addressing

several econometric issues such as the potential endogeneity bias, industry-specific

characteristics, and heteroscedasticity, we find strong evidence that establishment

productivity is positively correlated with probability of international outsourcing in

Thailand’s manufacturing sector. This finding therefore substantiates our theoretical

proposition.

Our empirical framework sheds further light on several establishment-specific

characteristics that shape outsourcing decisions. The estimates reveal that exporters and

foreign-owned establishments have higher tendency that the international outsourcing

decision is made, than do non-exporters and locally owned establishments. This may be

attributable to superior international contacts and information about foreign markets and

contractual partners. Interestingly, new entrants tend to internalise production activities and

hence refrain from outsourcing activities, due probably to the lack of core competences and

specialisation. The establishments must take time to learn what they are good at in order to

contract out parts and components others can do better.

The effects of market size and capacity utilisation are also of expected sign. Our

estimates show that establishments with larger market size and higher capacity utilisation

prone to undertake international outsourcing. The intuition is that market size and capacity

utilisation have to do with the production scale. In this sense, large establishments and those

intensively utilising their fixed capital are able to overcome the arising transaction costs of

and realise considerable production efficiency gains from international outsourcing.

References

Abraham, K.G. and S.K. Taylor (1996), “Firms’ Use of Outside Contractors: Theory and Evidence”, Journal of Labor Economics, Vol. 14, No. 3, pp. 394-424.

Ahn S., K. Fukao, and K. Ito (2008) “Outsourcing in East Asia and Its Impact on the Japanese and Korean Labor Markets”, OECD Trade Policy Working Papers, No. 65, OECD.

Amiti, M. and J. Konings (2007), “Trade Liberalization, Intermediate Inputs, and Productivity: Evidence from Indonesia”, American Economic Review, Vol. 97, No. 5, pp. 1611-1638.

Amiti, M. and S-J Wei (2009), “Service Offshoring and Productivity: Evidence from the US”, The World Economy, Vol. 32, No. 2, pp. 203-220.

Anderton, B. and P. Brenton (1999), “Outsourcing and Low-skilled Workers in the UK”, Bulletin of Economic Research, Vol. 51, No. 4, pp. 267-285.

Antràs, P. (2003), “Firms, Contracts, and Trade Structure”, Quarterly Journal of Economics, Vol. 118, No. 4, pp.1375-1418.

Antràs, P. (2005), “Incomplete contracts and the product cycle”, American Economic Review, Vol. 95, No. 44, pp. 1054-1073.

Antràs, P., and E. Helpman (2004), “Global Sourcing”, Journal of Political Economy, Vol. 112, No. 3, pp. 552-580.

Argyres, N. (1996), “Evidence on the Role of Firm Capabilities in Vertical Integration Decisions”, Strategic Management Journal, Vol. 17, No. 2, pp. 129-150.

Arellano, M. and S. Bond (1991), “Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations”, Review of Economic Studies, Vol. 58, No. 2, pp. 277-297.

Arndt, S. (1999), “Globalization and Economics Development”, The Journal of International Trade & Economic Development, Vol. 8, No. 3, pp. 309-318.

Aw, B-Y, S. Chung, and M.J. Roberts (2000), “Productivity and Turnover in the Export Market: Micro-level Evidence from the Republic of Korea and Taiwan (China)”, World Bank Economic Review, Vol. 14, No. 1, pp. 65-90.

Baldwin, J.R. and W. Gu (2003), “Export-market Participation and Productivity Performance in Canadian Manufacturing”, Canadian Journal of Economics, Vol. 36, No. 3, pp. 634-657.

Berman, E., J. Bound, and Z. Griliches (1994), “Changes in the Demand for Skilled Labor within U.S. Manufacturing: Evidence from the Annual Survey of Manufactures”, Quarterly Journal of Economics, Vol. 104, No. 2, pp. 367-398.

Bernard, A.B. and J.B. Jensen (1999), “Exceptional Export Performance: Cause, Effect, or Both?”, Journal of International Economics, Vol. 47, No. 1, pp. 1-25.

Bhagwati, J., A. Panagariya, and T.N. Srinivasan (2004), “The Muddles over Outsourcing”, Journal of Economics Perspectives, Vol. 118, No. 4, pp. 93-114.

Black, S.E. and L.M. Lynch (2001), “How to Compete: The Impact of Workplace Practices and Information technology on Productivity”, Review of Economics and Statistics, Vol. 83, No. 3, pp. 434-445.

Calabrese, G. and F. Erbetta (2005), “Outsourcing and Firm Performance: Evidence from Italian Automotive Suppliers”, International Journal of Automotive Technology and Management, Vol. 5, No. 4, pp. 461-479.

Clerides, S.K., S. Lach, and J.R. Tybout (1998), “Is Learning by Exporting Important? Micro-dynamic Evidence from Colombia, Mexico, and Morocco”, Quarterly Journal of Economics, Vol. 113, No. 3, pp. 903-948.

Criscuolo C. and M. Leaver (2005), “Offshore Outsourcing and Productivity”, mimeo.

Deardorff, A.V. (2001), “Fragmentation in Simple Trade Models”, North American Journal of Economics and Finance, Vol. 12, No. 2, pp. 121-137.

Dell’mour, R., P. Egger, K. Gugler, and M. Pfaffermayr (2000), “Outsourcing of Austrian Manufacturing to Eastern European Countries: Effects on Productivity and the Labor Market”, in Fragmentation of the Value-Added Chain, S. Arndt, H. Handler, and D. Salvatore (Eds.), Vienna: Austrian Ministry for Economic Affairs and Labour, pp. 249-302.

De Loecker, J. (2007), “Do Exports Generate Higher Productivity? Evidence from Slovenia”, Journal of International Economics, Vol. 73, No.1, pp. 69-98.

Egger, H. and P. Egger (2006), “International Outsourcing and the Productivity of Low-skilled Labour in the EU”, Economic Inquiry, Vol. 44, No.1, pp. 98-108.

Ethier, W. (1982), “National and International Returns to Scale in the Modern Theory of International Trade”, American Economic Review, Vol. 72, No. 3, pp. 389-405.

Evenson, R. and L. Westphal (1995), “Technological Change and Technology Strategy”, in Handbook of Development Economics, Volume 3, T.N. Srinivasan and J. Behrman (Eds.), Amsterdam: North-Holland.

Feenstra, R.C. and G. Hanson (1996), “Globalization, Outsourcing, and Wage Inequality”, American Economic Review, Vol. 86, No. 2, pp. 240-245.

Feenstra, R.C. and G. Hanson (1997), “Productivity Measurement and the Impact of Trade and Technology on Wages: estimates for the U.S., 1972-1990”, NBER Working Papers No. 6052.

Feenstra, R.C. and G. Hanson (1999), “The Impact of Outsourcing and High-technology Capital on Wages: Estimates for the United States”, Quarterly Journal of Economics, Vol. 114, No. 3, pp. 907-940.

Feenstra, R.C. and B.J. Spencer (2005), “Contractual versus Generic Outsourcing: The Role of Proximity”, NBER Working Paper No. 11885.

Ferguson, T.S. (1958), “A Method of Generating Best Asymptotically Normal Estimates with Application to the Estimation of Bacterial Densities”, Annals of Mathematical Statistics, Vol. 29, No. 4, pp. 1046-1062.

Fitzenberger, B. (1999), Wages and Employment Across Skill Groups, Heidelberg: Physica-verlag.

Fixler, D.J. and D. Siegel (1999), “Outsourcing and Productivity Growth in Services”, Structural Change and Economic Dynamics, Vol. 10, No. 2, pp. 177-194.

Gainey, T.W. and B.S. Klaas (2003), “The Outsourcing of Training and Development: Factors Impacting Client Satisfaction”, Journal of Management, Vol. 29, No. 2, pp. 207-229.

Garcia, J., P. Hernandez, and A. Lopez-Nicolas (2001), “How Wide is the Gap? An Investigation of Gender Wage Differences Using Quantile Regression”, Empirical Economics, Vol. 26, No. 1, pp. 149-167.

Geishecker, I. (2002), “Outsourcing and the Relative Demand for Low-Skilled Labour in German Manufacturing: New Evidence”, German Institute for Economic Research DIW-Berlin Discussion Papers, No. 313.

Geishecker, I. and H. Görg (2005), “Do Unskilled Workers Always Lose From Fragmentation?”, North American Journal of Economics and Finance, Vol. 16, No. 1, pp. 81-92.

Geishecker, I. and H. Görg (2008), “Winners and Losers: A Micro-level Analysis of International Outsourcing and Wages”, Canadian Journal of Economics, Vol. 41, No. 1, pp. 243-270.

Girma, S. and H. Görg (2002), “Foreign Ownership, Returns to Scale and Productivity: Evidence from UK Manufacturing Establishments”, CEPR Discussion Papers No. 3503.

Girma, S. and H. Görg (2004), “Outsourcing, Foreign Ownership, and Productivity: Evidence from UK Establishment-Level Data” Review of International Economics, Vol. 12, No. 5, pp. 817-832.

Görg, H. and A. Hanley (2004), “Does Outsourcing Increase Profitability”, The Economic and Social Studies, Vol. 35, No. 3, pp. 267-288.

Görg, H. and A. Hanley (2005), “International Outsourcing and Productivity: Evidence from the Irish Electronics Industry”, North American Journal of Economics and Finance, Vol. 16, No. 2, pp. 255-269.

Görg, H., A. Hanley, and E. Strobl (2008), “Productivity Effects of International Outsourcing: Evidence from Plant-level Data”, Canadian Journal of Economics, Vol. 41, No. 2, pp. 670-688.

Görzig B. and A. Stephan (2002), “Outsourcing and Firm-level Performance”, DIW Berlin Discussion Papers, No. 309.

Grant, R.M. (1991), “The Resource-Base Theory of Competitive Advantage: Implications for Strategy Formulation”, California Management Review, Vol. 33, No. 3, pp. 114-135.

Griliches, Z. (1997), “Education, Human Capital and Growth”, Journal of Labor Economics, Vol. 15, No. 1, pp. S330-344.

Grossman, S.J. and O.D. Hart (1986), “The Cost and Benefit of Ownership: A Theory of Lateral and Vertical Integration”, Journal of Political Economy, Vol. 94, No. 4, pp. 691-719.

Grossman, G.M. and E. Helpman (2002), “Integration versus Outsourcing in Industry Equilibrium”, Quarterly Journal of Economics, Vol. 117, No. 1, pp. 85-120.

Grossman, G.M. and E. Helpman (2005), “Outsourcing in A Global Economy”, Review of Economic Studies, Vol. 72, No. 1, pp. 135-159.

Hanson, G.H. and A.E. Harrison (1999), “Trade, Technology, and Wage Inequality”, Industrial and Labor Relation Review, Vol. 52, No. 2, pp. 271-288.

Hart, O.D. and J. Moore (1990), “Property Rights and the Nature of the Firm”, Journal of Political Economy, Vol. 98, No. 6, pp. 1119-1158.

Head, K. and J. Ries (2002), “Offshore Production and Skill Upgrading by Japanese manufacturing Firms”, Journal of International Economics, Vol. 58, No. 1, pp. 81-105.

Head K., J. Ries, and B.J. Spencer (2004), “Vertical Networks and U.S. Auto Parts Exports: Is Japan Different?”, Journal of Economics and Management Strategy, Vol. 13, No. 1, pp. 37–67.

Helg, R. and L. Tajoli (2005), “Patterns of International Fragmentation of Production and Implications for the Labour Markets”, Gerald R. Ford School of Public Policy Discussion Paper No. 503.

Helpman, E. and P. Krugman (1985), Market Structure and Foreign Trade, Cambridge: MIT Press.

Hijzen, A., H. Görg, and R.C. Hine (2005), “International Outsourcing and the Skilled Structure of Labour Demand in the United Kingdom”, Economic Journal, Vol. 115, No. 506, pp. 860-878.

Hsieh, C-T and K.T. Woo (2005), “The Impact of Outsourcing to China on Hong Kong’s Labor Market”, American Economic Review, Vol. 95, No. 5, pp. 1673-1687.

Johnson, M. (2003), “Riding the Offshoring Wave”, Computerworld Daily, September 15, 2003.

Jones, R. (2000), Globalization and the Theory of Input Trade, Cambridge: MIT Press.

Jones, R. and Kierzkowski (1990), “The Role of Services in Production and International Trade: A Theoretical Framework”, in The Political Economy of International Trade, R. Jones and A. Krueger (Eds.), Oxford: Blackwell Publishing, pp. 31-48.

Jones, R.W. and H. Kierzkowski (2001), “A Framework for Fragmentation”, in Fragmentation: New Production Patterns in the World Economy, S.W. Arndt and H. Kierzkowski (Eds.), New York: Oxford University Press.

Kimura, F. (2002), “Subcontracting and the Performance of Small and Medium Firms in Japan”, Small Business Economics, Vol. 18, No. 3, pp. 163-175.

Klein, B., R.G. Crawford, and A.A. Alchian (1978), “Vertical Integration, Appropriate Rents, and the Competitive Contracting Process”, Journal of Law and Economics, Vol. 21, No. 2, pp. 297-326.

Koenker, R. and K.F. Hallock (2001), “Quantile Regression”, Journal of Economic Perspectives, Vol. 15, No. 4, pp. 143-156.

KPMG International (2006),”Asian Outsourcing: The Next Wave”, available at:    http://www.kpmg.com/Global/en/WhatWeDo/Industries/Electronics-Software-

Services/Documents/Outsourcing-on-the-rise-in-Asia.pdf

Kraft, K. (1994), “Wage Differentials between Skilled and Unskilled Workers”, Review of World Economics, Vol. 130, No. 2, pp. 329-349.

Krugman, P. (1981), “Intraindustry Specialization and the Gains from Trade”, Journal of Political Economy, Vol. 89, No. 5, pp. 959-974.

Lyons, B.R. (1995), “Specific Investment, Economies of Scale, and the Make-or-buy Decision: A Test of Transaction Cost Theory”, Journal of Economic Behavior and Organization, Vol. 26, No. 2, pp. 431-443.

Machin S., A. Ryan, and J. van Reenen (1996), “Technology and Changes in Skill Structure: Evidence from A Panel of International Industries”, CEPR Discussion Papers No. 1434.

Maddala, G.S. (1977), Econometrics, New York: McGraw-Hill Book Company.

Markusen, J.R (1995), “The Boundaries of Multinational Enterprises and the Theory of International Trade”, Journal of Economic Perspectives, Vol. 9, No. 2, pp. 169-189.

Mazzola, F. and S. Bruni (2000), “The Role of Linkages in Firm Performance: Evidence from Southern Italy”, Journal of Economic Behavior and Organization, Vol. 43, No. 2, pp. 223-238.

McKinsey Global Institute (2003), “Offshoring: Is it a Win-win Game?”, available at: http://www.mckinsey.com/mgi/reports/pdfs/offshore/Offshoring_MGI_Perspective.pdf

McLaren, J. (2000), “‘Globalization’ and Vertical Structure”, American Economic Review, Vol. 90, No. 5, pp. 1239-1254.

Melitz, M. J., (2003), “The Impact of Trade on Intra-industry Reallocations and Aggregate Industry Productivity”, Econometrica, Vol. 71, No. 6, pp. 1695-1725.

Monteverde, K. and D.J. Teece (1982), “Appropriable Rents and Quasi-Vertical Integration”, Journal of Law and Economics, Vol. 25, No. 2, pp. 321-328.

Morrison, C.J. (1988), “Quasi-Fixed Inputs in U.S. and Japanese Manufacturing: A Generalized Leontief Restricted Cost Function Approach”, Review of Economics and Statistics, Vol. 70, No. 2, pp. 275-287.

Morrison, C.J. and D.S. Siegel (2001), “The Impacts of Technology, Trade and Outsourcing on Employment and Labour Composition”, Scandinavian Journal of Economics, Vol. 103, No. 2, pp. 241-264.

Murray, J.Y., A.R. Wildt, and M. Kotabe (1995), “Global Sourcing Strategies of US Subsidiaries of Foreign Multinationals”, Management International Review, Vol. 35, No. 4, pp. 307-324.

Newey, W.K. (1987), “Efficient estimation of Limited Dependent Variable Models with Endogenous Explanatory Variable”, Journal of Econometrics, Vol. 36, No. 3, pp. 231-250.

Oliner, S.D. and D.E. Sichel (1994), “Computer and Output Growth Revisited: How Big Is the Puzzle?”, Brookings Papers on Economic Activity, Vol. 25, No. 2, pp. 273-334.

Qiu, L. D. and B.J. Spencer (2002), “Keiretsu and Relationship-specific Investments: Implications for Market-opening Policy”, Journal of International Economics, Vol. 58, No. 1, pp. 49–79.

Rae, S. (2002), “Offshore Resourcing: Once Adventurous, Now Essential for Financial Services Firms”, IBM Business Consulting Services.

Schultz, T.P. and G. Mwabu (1998), “Labor Unions and the Distribution of Wages and Employment in South Africa”, Industrial and Labor Relations Review, Vol. 51, No. 4, pp. 680-703.

Shapiro, C. and J. Stiglitz (1984), “Equilibrium Unemployment as a Worker Discipline Device”, American Economic Review, No. 74, No. 3, pp. 433-444.

Siegel, D. and Z. Griliches (1992), “Purchases Services, Outsourcing, Computers, and Productivity in Manufacturing”, in Output Measurement in the Service Sector, Z. Griliches (Ed.), Chicago: University of Chicago Press, pp. 429-460.

Spencer, B. J. and L. Qiu (2001), “Keiretsu And Relationship-specific Investments: A Barrier to Trade?”, International Economic Review, Vol. 42, No. 4, pp. 871–901.

Strauss-Kahn, V. (2004), “The Role of Globalization in the Within-Industry Shift Away From Unskilled Workers in France”, in Challenges to Globalization: Analyzing the Economics, R.E. Baldwin and L.A. Winters (Eds.), Chicago: University of Chicago Press, pp. 209-234.

Terdiman, R. and F. Karamouzis (2002), “Going Offshore to Globally Source IT Services”, Gartner Research, available at: http://www.gartner.com/resources/103900/103916/103916.pdf.

Tomiura, E. (2005), “Foreign Outsourcing and Firm Characteristics: Evidence from Japanese

Manufacturers”, Journal of the Japanese and International Economies, Vol. 19, No. 2, pp. 255-271.

Tomiura, E. (2007), “Foreign Outsourcing, Exporting, and FDI: A Productivity Comparison at the Firm Level”, Journal of International Economics, Vol. 72, No. 1, pp. 113-127.

White, H. (1980), “A Heteroskedasticity-Consistent Covariance Matrix Estimator and A Direct Test for Heteroskedasticity”, Econometrica, Vol. 48, No. 4, pp. 817-838.

Williamson, O.E. (1975), Markets and Hierarchies: Analysis and Antitrust Implications, New York: Free Press.

Williamson, O.E. (1986), “Vertical Integration and Related Variations on A Transaction-Cost Theme”, in J.E. Stiglitz and G.F. Mathewson (eds.), New Developments in the Analysis of Market Structure, Cambridge: MIT Press, pp. 149-174.

Figure 1: Cut-off Productivity Levels with the Presence of Non-outsourcing Firms.

 

Source: Authors’ estimation.

Figure 2: Cut-off Productivity Level without the Presence of Non-outsourcing Firms.

Source: Authors’ estimation.

Table 1: Sample Manufacturing Establishments by Regions.

Region Number of Establishments Bangkok 909

Vicinity and 3,383 Northern 1,545

Northeastern 1,089 Southern 1,804

Whole Kingdom 8,730 Source: The 2003 Manufacturing Industry Survey, National Statistical Office, Thailand.

Table 2: International Outsourcing Decisions in Thailand’s Manufacturing Industries.

Outsourcing Decisions Number of Establishments Percentage YES 1,677 19.2 NO 7,053 80.8

< 10 % 261 15.6 10 % - 19 % 228 13.6 20 % - 29 % 199 11.9 30 % - 39% 173 10.3 40 % - 49 % 102 6.1 50 % - 59 % 134 8.0 60 % - 69 % 114 6.8 70 % - 79 % 129 7.7 80 % - 89 % 135 8.1 90 % - 100 % 202 12.1

Source: The 2003 Manufacturing Industry Survey, National Statistical Office, Thailand.

Table 3: International Outsourcing Decisions by Industries.

Industry Number of Establishments

Percentage

Food and beverages 211(9.37) 9.37Tobacco n.a. n.a.Textiles 101 (12.95) 12.95Wearing apparel; dressing and dyeing of fur 46 (13.65) 13.65Leather footwear products 72 (25.90) 25.90Wood and cork, except furniture 63 (18.92) 18.92Paper and paper products 50 (26.74) 26.74Publishing, printing and reproduction of recorded media 27 (11.64) 11.64

Refined petroleum products 16 (44.44) 44.44Chemicals and chemical products 127 (33.96) 33.96Rubber and plastics products 122 (27.73) 27.73Other non-metallic mineral products 84 (11.62) 11.62Basic metals 34 (23.78) 23.38Fabricated metal products 163 (20.87) 20.87Machinery and equipment 105 (33.55) 33.55Office, accounting and computing machinery 6 (42.86) 42.86Electrical machinery and apparatus 61 (43.26) 43.26Communication equipment and apparatus 75 (53.19) 53.19Medical precision and optical instruments 37 (38.54) 38.54Motor vehicles, trailers and semi-trailers 65 (36.93) 36.93Other transport equipment 40 (23.12) 23.12Furniture manufacturing 172 (20.65) 20.95Recycling n.a. n.a.Total manufacturing 1677 19.2Source: The 2003 Manufacturing Industry Survey, National Statistical Office, Thailand.

Table 4: Characteristics of Manufacturing Establishments.

Industry Foreign Ownership

Exports New Entry

Food and beverages 118 (5.24) 309 (13.73) 439 (34.16)Tobacco 2 (4.26) 3 (6.38) n.a.Textiles 62 (7.95) 121 (15.51) 172 (13.39)Wearing apparel; dressing and dyeing of fur 19 (5.64) 64 (18.99) 58 (4.51)Leather footwear products 29 (10.43) 64 (23.02) 48 (3.74)Wood and cork, except furniture 12 (3.60) 67 (20.12) 64 (4.98)Paper and paper products 27 (14.44) 46 (24.60) 13 (1.01)Publishing, printing and reproduction of recorded media 10 (4.31) 8 (3.45) 10 (.78)

Refined petroleum products 8 (22.22) 9 (25.00) 4 (.31)Chemicals and chemical products 70 (18.72) 80 (21.39) 55 (4.28)Rubber and plastics products 82 (18.64) 147 (33.41) 65 (5.06)Other non-metallic mineral products 45 (6.22) 99 (13.69) 110 (8.56)Basic metals 19 (13.29) 28 (19.58) 15 (1.17)Fabricated metal products 78 (9.99) 100 (12.80) 62 (4.82)Machinery and equipment 55 (17.57) 72 (23.00) 30 (2.34)Office, accounting and computing 6 (42.86) 5 (35.71) n.a.Electrical machinery and apparatus 41(29.08) 47 (33.33) 5 (.39)Communication equipment and apparatus 59 (41.84) 63 (44.68) 4 (.31)Medical precision and optical instruments 21 (21.88) 24 (25.00) 2 (.16)Motor vehicles, trailers and semi-trailers 48 (27.27) 48 (27.27) 6 (.47)Other transport equipment 26 (15.03) 25 (14.45) 12 (.93)Furniture manufacturing 97 (11.64) 219 (26.29) 111 (8.64)Recycling n.a. n.a. n.a.Total manufacturing 934 (10.70) 1,648 (18.9) 1,285 (14.72)Source: The 2003 Manufacturing Industry Survey, National Statistical Office, Thailand. Note: Percentages in parentheses.

Table 5: Probit Model Estimates.

Dependent Variable: ( )1Pr =iOut

Independent Variable Probit FEProbit

iθ ( )4408.0244. *** ( )0049.0241. ***

iExport ( )0176.3224. *** ( )0183.3294. ***

iFhold ( )0232.2744. *** ( )0231.2407. ***

iNew ( )0141.0832. ***− ( )0149.0618. ***−

iRsize ( )4408.258.1 *** ( )3872.8135. **

iCapu ( )0012.0058. *** ( )0012.0062. ***

Number Obs. 6,981 6,981

Wald Chi-squared ***26.589,1 ***63.654,1

Pseudo R-squared .3046 3313.

Log Pseudolikelihood 66.538,2− 28.441,2−

Correctly Specified Rate %69.84 %3.85

Note: 1) ** and *** statistically significant at 5 and 1 percent, respectively; 2) The parameter estimates are the marginal effects of the Probit model; and 3) Robust standard errors are in parentheses. Table 6: Instrumental Variable (IV) Estimates of the Probit Model.

Dependent Variable: ( )1Pr =iOut

Independent Variable IV IVFE

iθ ( )0218.2092. *** ( )0240.2239. ***

iExport ( )0513.9930. *** ( )0541.030.1 ***

iFhold ( )0640.8128. *** ( )0656.7339. ***

iNew ( )0917.3474. ***− ( )0939.2134. **−

iRsize ( )284.1965.2 ** 466.1 ( )192.1

iCapu ( )0055.0125. ** ( )0059.0131. **

Number Obs. 5,145 5,145

Wald Chi-squared ***18.340,1 ***26.378,1

Wald Test of Exogeneity 07. .81

Note: 1) ** and *** statistically significant at 5 and 1 percent, respectively; 2) The parameter estimates are based on Newey’s two-step minimum chi-squared estimation; 3) Robust standard errors are in parentheses; and 4) The Wald statistic is chi-squared distributed based on the null of no endogeneity.