How local are IT services markets: Proximity and IT outsourcing Ashish Arora, Chris Forman1 ...

33 Pages·2007·0.93 MB·English

How local are IT services markets: Proximity and IT outsourcing

Ashish Arora, Chris Forman1
Carnegie Mellon University
[email protected], [email protected]

March 2007

Abstract

We examine the question of which services are tradable within a concrete setting: the outsourcing of IT
services across a broad cross-section of establishments in the US. If markets for IT services are local, then
we should expect increases in local supply should increase the likelihood of outsourcing by lowering the
cost of outsourcing. If markets are not local then local supply will not affect outsourcing demand. We
analyze the outsourcing decisions of a large sample of 99,775 establishments in 2002 and 2004, for two
types of IT services: programming and design and hosting. Programming and design projects require
communication of detailed user requirements whereas hosting requires less coordination between client
and service provider than programming and design. Our empirical results bear out this intuition: The
probability of outsourcing programming and design is increasing in the local supply of outsourcing, and
this sensitivity to local supply conditions has been increasing over time. This suggests there is some non-
tradable or “local” component to programming and design services that cannot be easily removed. In
contrast, the decision to outsource hosting is sensitive to local supply only for firms for which network
uptime and security concerns are particularly acute.

1 Corresponding author. Authors are listed in alphabetical order. We thank the associate editor and two anonymous
referees for very helpful comments. We also thank the track co-chairs (Eric Clemons, Rajiv Dewan, and Robert
Kauffman) and three anonymous reviewers for the minitrack on Competitive Strategy, Economics, and IS for the
40th Annual Hawaii International Conference on System Sciences. We thank the Alfred P. Sloan Foundation for
financial support, and Harte Hanks Market Intelligence for providing essential data. All errors are our own.
1. Introduction

The outsourcing and offshoring of services in the US is an important and growing phenomenon
that has recently attracted widespread attention. At present, widely varying projections of the number of
jobs “at risk” have been presented, mostly by consulting firms [31]. Ultimately, these estimates turn on
the extent to which IT services are tradable.2
There have been two prevailing views on the tradability of IT services. One view emphasizes the
role of information technology in reducing the costs of performing services at a distance. Under this view,
IT reduces the costs of coordinating economic activity over long distances. Proponents of this view argue
that all services are potentially tradable. A second view argues that humans work best in physical
proximity to one another, and that some face-to-face interaction is required for the execution of many
types of services. Proponents of this view argue that offshoring is fraught with hidden costs arising from
inexperienced services personnel, differences in language and culture, and time differences between
vendor and client site [29]. Though a great deal of case study work has examined offshore project
decisions and governance in a variety of situations (e.g., [10]), this is ultimately a question not of what is
possible but rather what is predominant.
In this paper, we examine the extent to which markets for purchased IT services are local. If some
elements of IT services delivery must be supplied locally, then suppliers must be located near customers,
and so are not tradable. By contrast, if the markets for IT services are not local, then suppliers need not be
collocated, and so IT services are tradable. Our focus is on the tradability of purchased IT services: we do
not consider the tradability of services that are performed within the boundaries of the firm. That is, our
results do not inform which services can be offshored within a firm. However, at present the vast majority
of offshored services are also outsourced, and captive offshoring is usually restricted to large firms that
have IT as a central part of their value creation strategy [34].
We examine the IT outsourcing decisions of a large cross-section of establishments in the US

2 In this paper we follow the international trade literature and in particular Jensen and Kletzer (2005) in using the
label ‘tradable services’ to refer to those that can be conducted at a distance.
1
during two years: 2002 and 2004. We investigate the extent to which the outsourcing decision depends
upon the local supply of outsourcing firms. Our major hypothesis is that if some components of IT
services need to be delivered locally, then increases in local supply should increase the likelihood of
outsourcing by lowering the cost of outsourcing. We extend this analysis by examining how this
relationship has changed over time, how it differs across IT tasks, and how it is moderated by the
characteristics of the establishment.
We examine the decisions of firms to outsource two types of IT services: programming and
design and hosting. Programming and design refers to the decision to outsource programming tasks or the
planning and designing of information systems. To be successful, these outsourcing projects typically
require communication of detailed user requirements. Hosting involves management and operation of
computer and data processing services for the client, as well as hosting of Internet services and web
servers. After an initial set-up period, the requirements of such hosting services will be relatively static
and will require less coordination between client and service provider than programming and design.
As a preview to our results, we show that:
1. The probability of outsourcing programming and design is increasing the local supply of
outsourcing.
2. The sensitivity to local supply conditions appears to have increased over time. This suggests
that there is some irreducible non-tradable or “local” component to programming and design
services that cannot easily be removed through advancement in software development
practices or through improvements in communications capabilities engendered by
advancements in IT.
3. Outsourcing of hosting services is less sensitive to variance in local supply than is the
decision to outsource programming and design, though the sensitivity of the hosting decision
has also increased over time. This increase in sensitivity appears to be concentrated among
those firms for which network uptime and security concerns are particularly acute.
2. Related Literature
2
This paper is related to three areas of prior research. First, we contribute to recent work that has
examined which types of service work are most effectively conducted offshore. Second, we advance work
in the IT outsourcing and nascent IT offshoring literatures. Third, we contribute to recent work that seeks
to understand the geographic dispersion in the location of high technology industries.
We view our research as building upon recent attempts to understand which services are tradable
across a broad cross-section of the economy. Jensen and Kletzer [22] examine which services are tradable
by examining geographic concentration in economic activity. Tradable industries will be geographically
concentrated to take advantage of economies of scale and favorable location factors. By contrast, non-
tradable industries must locate where demand is and thus must have a geographic distribution that is
similar to that of overall economic activity. Our approach is complementary: If a service is easily
tradable, demand decisions will not be sensitive to whether the service is locally available (or the extent
of its availability).
Ono [33] examines manufacturing firm decisions to outsource white-collar services. She
examines how the outsourcing decision varies with potential demand as proxied by total population and a
demand shifter. Like Ono [33], we examine how the decision to outsource services depends upon local
market conditions, however our analysis focuses on identifying which IT services are tradable and we
focus on a broader cross-section of industries. We also explicitly model local supply, and treat it as
endogenous in the sense of potentially depending upon aggregate local demand.
We also contribute to recent field research conducted in other industries that has examined the
operational risks of outsourcing services that require intensive coordination or transfer of tacit knowledge
between buyer and supplier (e.g., [3], [12], [19]). In contrast to this research that relies on small samples
or case studies, we provide systematic evidence on which IT services can most easily be offshored using a
broad cross-section of industries in the U.S. economy.
Prior research on the IS outsourcing decision has often focused on variation in establishment- or
firm-level factors. These could be economic explanations such as the desire of firms to obtain cost
advantages through economies of scale or scope (e.g., [2], [25]) or the role of transaction costs on the
3
sourcing decision [2]. Other work has focused on political factors [24] or on strategic responses to
institutional influences [1]. To our knowledge, no prior work has addressed our primary question: the
conditions under which purchased IT services need to be conducted locally. 3
We also contribute to recent research that has examined the spatial distribution of economic
activity in high-technology industries ([7], [23]). Much of this prior literature demonstrates that
technology-intensive industries concentrate for one of three reasons: thicker labor markets, the
availability of complementary resources, or knowledge spillovers [28]. In our research we provide one
explanation for the geographic dispersion in software production in the U.S.
3. Framework and Hypotheses for the Decision to Outsource
We examine whether the decision to outsource is increasing in the local supply of outsourcing
firms. Prior work in media richness theory (e.g., [15], [16]) suggests that transmission of information that
is equivocal, uncertain, or complex is better accomplished through the use of richer media such as face-to-
face communications. Software development tasks typically require frequent small adjustments among
co-workers, particularly in early stages of software development [32]. In other words, insofar as IT
outsourcing tasks require transmission of information that is complex or equivocal, or require frequent
adjustments among developers, the theory suggests that the costs of outsourcing will be lower when
providers of outsourcing services are nearby. It is well established in other nontradable services markets
that prices are declining in the number of entrants, so long as labor supply is sufficiently elastic [9]. Thus,
since the price of local IT services is decreasing in the number of local suppliers, we expect the likelihood
of outsourcing to be increasing in the supply of local IT services firms.
Hypothesis 1: Outsourcing is increasing in the supply of local firms, other things equal.
We also examine how the sensitivity to local supply depends upon the outsourcing task. Case
study research on globally distributed software research suggests that activities that are difficult to

3 We also advance recent work that has investigated the factors influencing the performance of IT outsourcing
engagements (e.g., Gopal et al 2003; Gopal and Sivaramakrishnan 2006). In contrast to prior work which has
focused on characteristics of the contract or the engagement, we examine how proximity to outsourcing firms
influences the perceived value of outsourcing.
4
conduct at a distance include those that require customer interaction, customer proximity, and deep
domain knowledge, while those that are easier to conduct at a distance include activities like coding and
maintenance, or infrastructure maintenance [11].
In our study we examine the outsourcing of two sets of activities: programming and design and
hosting. The decision to outsource programming and design tasks, due to the requirement of transmitting
equivocal information (particularly at early stages of the software development lifecycle) and to the need
for frequent coordination between supplier and customer, will be more sensitive to changes in local
supply than will the decision to outsource hosting services.
Hypothesis 2: The decision to outsource programming and design will be more sensitive to variations in
local supply than will the decision to outsource hosting.
A long line of information systems research has suggested that IT reduces communication and
coordination costs within and between firms (e.g., [26], [13], [17], [18]). To examine how IT influences
the importance of proximity to suppliers, we examine how the relationship between the decision to
outsource and local supply changes over time. If technological improvements in IT are decreasing the
coordination costs of sourcing at a distance, then the sensitivity of the outsourcing decision to changes in
local supply will decline over time. However, it is also possible that sensitivity to local supply will
increase over time. For example, increases in concerns over information security may require increasing
proximity for hosting providers if security incidents require immediacy in communications between
clients and service providers. Though we state our hypothesis in terms of the sensitivity of local supply
increasing over time, we treat it as an empirical question as to which hypothesis is supported by the data.
Hypothesis 3: The sensitivity of outsourcing to local supply is increasing over time.
In our last set of hypotheses, we examine how the sensitivity to local supply varies based upon
the characteristics of the firm. In particular, we examine the effects of variation in firms’ IT infrastructure
environment on the relationship between local supply and the decision to outsource hosting. We have
5
conducted a similar set of analyses for programming and design, but we focus on the results for hosting
services for the sake of brevity. We examine five classes of application infrastructure within a firm:
enterprise applications, functional applications, network applications, security applications, and other
applications.
For hosting services, one major benefit of proximity is the immediacy of replies in face-to-face
communications. This will be particularly valuable for firms with heavy investments in network and
security applications. For firms with large investments in network infrastructure, reliability of the network
is particularly important and network downtime must be minimized. Network problems, when they occur,
must be handled quickly. Firms with large investments in security applications will be particularly
concerned with information security issues. When security issues arise, they must be handled quickly, and
web servers and firewalls may need to be accessed physically.4 That is, in both cases, immediacy of
communication between service provider and client is particularly important. This suggests that proximity
to suppliers will be relatively more important for firms with greater investments in security and network
infrastructure software.
Hypothesis 4a: The relationship between outsourcing hosting and local supply is stronger for firms with
greater investments in security software.
Hypothesis 4b: The relationship between outsourcing hosting and local supply is stronger for firms with
greater investments in network infrastructure software.

4. Framework and Econometric Model
We motivate our econometric model with a simple theoretical framework that examines an
establishment’s decision to staff IT projects with internal or external IT staff. Establishments face the
following maximization problem:

4 This is vividly illustrated in the HBS case study, The iPremier Company: Denial of Service Attack, HBS 9-601-
114 [6], in which an online retailer under denial of service attack was unable to access the hosting facility remotely
and an IT staffer had to physically visit the hosting site to address the issue.
6
maxπ(x ,x ,z)−wx −w x
1 2 1 1 2 2
x,x
1 2
where x and x represent the number of external or internal IT employees hired (respectively), w and w
1 2 1 2
represent the wage or cost of hiring an additional external or internal worker, and z represents other
characteristics that condition profits (and hence, demand for outsourcing). The function π(⋅) represents
the value of IT projects. To decide upon the optimal level of outsourcing and IT employment, we take
first order conditions:
dπ(x ,x ,z)
1 2 −w =0
dx 1
1
dπ(x ,x ,z)
1 2 −w =0
dx 2
2
leading to the following optimal levels of outsourcing and internal IT employment:
x = f(w,x ,z) (1a)
1 1 2
x = f(w ,x ,z) (1b)
2 2 1
The focus of our analysis will be on the optimal level of outsourcing x .
1
To econometrically estimate the outsourcing decision embedded in these first order conditions in
(1a), we must make a number of additional assumptions. First, we assume that the cost of hiring external
workers is a linear function of local supply, w =g(os)+η, where os represents local supply and η is the
1
error term. This assumption is motivated by the arguments presented prior to hypothesis 1: (1) The
provision of some IT services is most effective and less costly when it is performed in close proximity to
clients and (2) If markets for goods and services are local, then their price is declining in local supply [9].
While we recognize that wages may depend on factors other than local supply, these other factors will be
included in the error term η. Indeed, to the degree that markets are not local, local supply will not matter
for the outsourcing decision. Thus, this specification merely makes concrete our hypotheses.
We do not observe the true quantity of employees outsourced, but only a binary variable
indicating whether outsourcing had been used. Thus, the number of outsourcing employees hired will be a
latent variable x*. Thus, the decision we observe for establishment i will be
1
7
x* = f(os +η,x ,z ,ν ).
1i i i 2i i i
Assuming that x* is linear in parameters gives us
1i
x* =αos +βx +γz +η+ν (2)
1i i 2i i i i
We assume that η and ν are distributed normally and we estimate a linear probability model of the
i i
decision to outsource. Our major interest is in testing whether α>0, that is, whether the decision to
outsource is increasing in local supply.
Of course, as (2) is a cross-sectional regression, one may be concerned that os may be correlated
i
with unobserved location-specific factors η that increase the likelihood of outsourcing. For example,
i
outsourcing firms may prefer to locate in places with a more highly skilled workforce, which may also
lower the costs of outsourcing. In this case, estimates of αwill be inconsistent. Further, x may be
2
correlated with unobservables that increase the value of outsourcing at an establishment. To address these
issues, we use instrumental variable (IV) techniques, as described later.
5. Data
The data we use come from the Harte Hanks Market Intelligence CI Technology database
(hereafter CI database). The CI database contains establishment-level data on (1) establishment
characteristics, such as number of employees, industry and location; (2) use of technology hardware and
software, such as computers, networking equipment, printers and other office equipment; and (3) use of
outsourcing. Harte Hanks collects this information to resell as a tool for the marketing divisions at
technology companies. We obtained data from the CI database for two years: 2002 and 2004. Interview
teams survey establishments throughout the calendar year; our sample contains the most current
information as of December 2002 and 2004 respectively.
We focus on the establishment rather than the firm as the unit of analysis because establishment-
level data will enable us to more precisely measure the impact of changes in local supply on the costs of
outsourcing. Most software investment decisions in our data are made at the establishment level. For
8
instance, 80% of the establishments in our sample that responded to the question stated that IT investment
decisions were made at the establishment rather than the firm level. Moreover, we control for multi-
establishment firms in our estimation. Our sample from the CI database contains all commercial
establishments with over 100 employees: there were 91,129 such establishments in 2002 and 89,776 in
2004. We dropped establishments which did not answer questions on outsourcing or had other crucial
data missing, which resulted in samples of 52,191 establishments in 2002 and 47,584 in 2004. In general,
excluded establishments were smaller and had fewer investments in IT. Thus, establishments with
relatively heavy investments in IT are over-represented in our sample. However, since these are more
likely to be able to reach out beyond local suppliers, our estimates are likely lower bounds on the true
effects.
5.1 Identifying Decisions to Outsource
Our dependent variable is x , the extent of outsourcing by establishment i. This variable x is
1i 1i
latent. We observe only discrete choices: whether or not the establishment chooses to outsource a
particular service or not, with the observed decision assuming a value of either one or zero, respectively.
Harte Hanks tracks 20 separate binary measures of outsourcing services that an establishment
may use. We aggregate these 20 different outsourcing services into two categories that have similar
production technologies. These two categories will comprise the dependent variables for our baseline
model. Later we report on robustness checks in which we explore alternative classifications.
The first variable measures an establishment’s decision to outsource programming or design
services. An establishment is considered to have outsourced programming and design if it answers yes to
outsourcing any of the following services: application design; contract programming; application
development; package software implementation; or Internet/web application development.
The second variable measures an establishment’s decision to outsource hosting services. An
establishment is considered to have outsourced hosting services if it answers yes to outsourcing any of the
following services: Internet/web server hosting; Internet routers; web site management; Internet firewalls;
LAN client/server; LAN network management; or LAN maintenance.
9