J Ind Compet Trade (2018) 18: 303–3 18

DOI 10.1007/s10842-017-0259-y

Hidden Costs of Offshore Outsourcing: an Analysis

of Offshoring Decisions

Deeparghya Mukherjee1,2

Received: 4 April 2017 / Revised: 7 August 2017 /

Accepted: 24 August 2017 / Published online: 21 September 2017
© Springer Science+Business Media, LLC 2017

Abstract Offshore outsourcing has grown as a form of industrial organisation to increase

profitability of firms. However, offshoring may be less lucrative than envisaged, due to
the presence of hidden costs. We study the strategic interaction amongst onshore Cournot
firms in the decision to offshore when they receive signals about offshore hidden costs.
The analysis helps suggest policy implications for countries which are potential offshoring
locations. We find the precision of signals and the range of possible hidden costs to be
crucial in determining offshoring destinations. Updating of information about hidden costs
leads to different equilibria including herding in offshoring.

Keywords Offshore outsourcing · Hidden costs · Bayesian nash equilibrium · Herding

JEL Classification D23 · D83 · L13

1 Introduction

Offshore outsourcing- a phenomenon that has gained precedence towards the close of the
previous century has attracted the attention of media, politicians, economists and sooth say-
ers. Amidst wide ranging claims of cost savings and employment effects, media reports and
academicians have highlighted that the benefits of offshoring may fall below expectations

A version of this paper is available as an Indian Institute of Management Bangalore Working Paper
No.: 370.

 Deeparghya Mukherjee
deeparghya@iimnagpur.ac.in

1 Visiting Research Fellow, Institute of South Asian Studies, National University of Singapore,
29 Heng Mui Keng Terrace. #08-06 Block B., Singapore 119620, Singapore
2 Economics Department, Indian Institute of Management Nagpur, IIM Nagpur, VNIT Campus,
South Ambazari Road, Ambazari, Nagpur, Maharashtra 440010, India
304 J Ind Compet Trade (2018) 18: 303–3 18

due to the existence of hidden costs. Offshoring jobs in the presence of asymmetric infor-
mation on hidden costs have strategic implications for inter-firm competition. This decision
to offshore and the choice of destination is analysed in this paper through a Bayesian game
showing possibilities of herding in the choice of destination amongst firms. Policy implica-
tions for countries aiming to improve their attractiveness as offshoring destinations are also
addressed.
Trading costs and difference in costs of production across geographies have influ-
enced production decisions especially for multinational companies over time (Antras 2003;
Navaretti and Venables 2006). Technological achievements, especially in information tech-
nology (IT) have allowed greater fragmentation of production process while reducing
trading costs. Together they have fostered intra and inter-firm trade in intermediate inputs
(Yeats 2001), (Borga and Zeile 2004). Offshore outsourcing, a specific form of vertical
integration have emerged as forms of industrial organisation.1
Countries like India are amongst the most preferred offshoring destinations today. How-
ever, several research reports suggest the existence of hidden costs of offshore production
that reduce the gains from offshoring. Surprisingly, TPI surveys put India and east European
economies to be equally competitive for offshoring although the production costs are higher
for the latter. Despite this, India continues to be a leading offshore destination. Our results
throw light on how India can be the leading offshore destination when other countries are
equally competitive with information asymmetries on hidden costs playing a key role.

2 Related Literature

Hidden costs in offshore outsourcing siphon out cost advantages which form the prime
motivation to offshore. Overby (2003) points out that hidden costs tend to reduce the cost
benefits from offshoring to India to about 20% of the planned benefits in some cases.
Barthelemy (2001) offers four different sources of hidden costs, namely: Vendor search
and contracting; transitioning to the vendor; managing the effort and transitioning after
outsourcing. Apart from these, the quality of offshore employees (Davison 2004), geograph-
ical distance (Goolsby 2002), cultural differences (Karamouzis 2002; Qu and Brocklehurst
2003) add to hidden costs. Finally, employee morale in the firm seeking to offshore part
of its production process poses yet another hidden cost and can affect the final perfor-
mance of offshore employees (Baruch and Hind 2000). The discomfiture among employees
has been known to attract protests by labour unions and to result in negative publicity
(DiamondCluster 2004) which affects the brand image of the firm.
Although the existence of the above hidden costs in offshore outsourcing is recognised,
there is a lack of treatment of the same in the literature in terms of its role in determining
offshore destinations. Offshoring and FDI decisions (make or buy) in the presence of trans-
action costs have been studied by Grossman and Helpman (2002, 2003). However, hidden
costs cover a range of costs including transaction costs and we focus on the strategic interac-
tion in the offshoring decision. Asymmetric information and investment decisions have been
studied by Chamley and Gayle (1994). In our attempt to address the decision to offshore we
introduce signals of imperfect precision (as in Zhang 1997) in the context of hidden costs to
study the strategic interactions between firms and the possibility of a herding equilibrium.

1 Offshoreoutsourcing refers to sourcing of inputs for production from a vendor located in a different
country(Sabherwal 1999)
J Ind Compet Trade (2018) 18: 303–3 18 305

We contribute to the literature on industrial organization, arguing that information asym-

metry on hidden costs of offshore production across firms leads to strategic interactions
between them influencing offshoring decisions. In philosophy, we add to the stream of
literature in industrial organisation dealing with information asymmetries and strategic
interaction (Milgrom and Roberts 1982) and investment under uncertainty (Chamley and
Gayle 1994). To the best of our knowledge the issue of hidden costs is not addressed in the
offshoring literature in the way we proceed.

3 The Model

We consider a simple Cournot Duopoly in a two period framework. The market demand
function is given by:

P = P (Q); P  < 0; P  ≤ 0; (1)

where P is the price level of the good and Q is the total industry output. Thus Q = q1 + q2
where q1 and q2 are the respective output levels of the two firms in the duopoly.
The firms have the option to produce domestically or offshore production.

3.1 Domestic Production

The domestic marginal cost of production is Cd i.e. Cdi = Cd ∀i, where i ∈ {1, 2} (constant
returns to scale).
Firm i’s profit equation is given by:


2
πi = [P ( qi ) − Cd ]qi (2)
i=1

Reaction functions:

q1 = R(q2 , Cd ) (3)

q2 = R(q1 , Cd ) (4)

The market equilibrium outputs: q1∗ = q2∗ = q ∗

Q∗ = 2q ∗ and P ∗ = P (Q∗ ).
πi∗ = (P ∗ − Cd )qi∗ and π ∗ = 2πi∗ gives the total industry profit.

3.2 Offshore Production

The offshore location has two components of marginal cost of production. (1)A known cost
C̄ and (2) an unknown(hidden) cost C ∈ {Cl , Ch }. We assume:

C̄ + Ch > Cd > C̄ + Cl

In other words, if the offshore location has high(low) hidden costs, the total marginal cost of
production is more(less) than the onshore costs. The exact hidden cost prevailing offshore
306 J Ind Compet Trade (2018) 18: 303–3 18

is unknown to both the firms and both receive signals about the possible hidden cost. The
timeline below highlights how the model proceeds:
⎧
⎪
⎪ 1. Players receive signals about hidden costs
⎪
⎪
⎪
⎪ θi = {Cl , Ch } i ∈ {1, 2}
⎨
2. Either or both players choose to offshore based on
Period 1:
⎪
⎪ expected costs
⎪
⎪
⎪ 3. If only firm i has offshored firm j  = i observes output
⎪
⎩
and produces as a Stackelberg follower
⎧
⎪
⎪ 1. firm j updates belief about hidden costs from firm i’s
⎨
output if only firm i offshored in period 1
Period 2:
⎪
⎪ 2. firm j decides to offshore or not
⎩
3. firms compete on quantity
Step 3 in Period 1 above follows because, after firm i has offshored, producing as a
Cournot producer for firm j without the knowledge of actual costs faced by firm i could lead
to producing more(less) when firm i faces low(high) hidden costs offshore. Hence becoming
a Stackelberg follower is rational.
The model is solved through backward induction. To elucidate on the above, in period
1, firms receive signals on the unknown component of foreign marginal cost of production.
The signal received by a firm i is given by θi = {Cl , Ch } i ∈ {1, 2}. Each firm has a privately
known precision pi = P r[C = Ch |θi = Ch ] = P r[C = Cl |θi = Cl ].2 Given the signal,
each firm calculates the expected profits of offshore production.
In period 1, the decision to move offshore or stay domestic rests on the expected prof-
itability over two periods and the description of the equilibrium in the next section brings
that out. If both players move offshore then they compete as Cournot duopolists with off-
shore production costs. If both players stay domestic then the outcome is the same as with
onshore production. The interesting case occurs when one firm has moved offshore and the
other stays domestic at the end of period 1. We assume without loss of generality, that firm
1 offshores in period 1 and firm 2 stays domestic. Since the true value of hidden costs faced
by firm 1 is unknown, firm 2 has to wait for firm 1’s output choice to decide on its own
production. Firm 2 thus is a Stackelberg follower and firm 1 becomes a Stackelberg leader. 3
In period 2, firm 2 observes firm 1’s action to move and the output choice made by firm 1 to
form an updated belief of the hidden cost of production offshore. Firm 2 then decides to move off-
shore or not and the firms compete on quantity. In this period in the case of a separating equilib-
rium, if both firms have gone offshore they compete as Cournot firms else firm 2 remains a
stackelberg follower. There are no costs of late movement, i.e. firm 2 does not incur added
costs for moving second to the offshore location. We detail the solutions of the model below:

3.3 Separating Equilibrium

In a separating equilibrium, the actions of firm 1 in period 1 completely reveals the level of
hidden cost that would be faced offshore. We provide a list of notations first:

2 Theprecision gives the probability that the signal is correct. The signals for the two firms are independent
and hence the precisions do not have common priors. The signals are exogenous to the model.
3 Producing in a Cournot framework, firm 2 may end up producing more(less) than the market demand when
firm 1 faces low(high) hidden costs and hence may bear a loss. Hence it is optimal for firm 2 to be a Stackel-
berg follower. We show that in this period firm 1 may choose to reveal the true costs of production or to pool
so as not to reveal the true costs depending on profitability and firm 2’s beliefs.
J Ind Compet Trade (2018) 18: 303–3 18 307

Notations:
F (Ch , Cd ) = P H S − (C̄ + Ch ): Profit per unit of output to firm 1 when firm 1 has
offshored to the destination with high hidden cost as a Stackelberg leader, while firm 2
produces onshore as a Stackelberg follower.
G(Ch , Cd ) = P H C (Cd , Ch ) − (C̄ + Ch ): Profit per unit of output to firm 1 when firm
1 has offshored to the destination with high hidden cost in Cournot duopoly, while firm 2
produces onshore.
F (Cl , Ch , Cd ) = P H S − (C̄ + Cl ): Profit per unit of output to firm 1 when firm 1 has
offshored to the destination with low hidden cost as a Stackelberg leader and mimics the
high hidden cost type, while firm 2 produces onshore as a Stackelberg follower.
F (Ch , Cl , Cd ) = P LS − (C̄ + Ch ): Profit per unit of output to firm 1 when firm 1 has
offshored to the destination with high hidden cost as a Stackelberg leader and mimics the
low hidden cost type, while firm 2 produces onshore as a Stackelberg follower.
G(Ch ) = P H C (Ch ) − (C̄ + Ch ) : Profit per unit of output if both players have offshored
to the destination with high hidden cost in Cournot duopoly.
q1H C (Ch ): Cournot output of firm 1 when both players have offshored with high hidden
costs of production.
q1H C (Ch , Cd ): Cournot Output of firm 1 when firm 1 has offshored to a high hidden cost
destination while firm 2 has not.
q1LC (Cl , Cd ): Cournot Output of firm 1 when firm 1 has offshored to a low hidden cost
destination while firm 2 has not.

Proposition 1 In a separating equilibrium, where firm 2’s beliefs are given by:
P[C = Cl |q1 ≥ q1LS ]=1 and
P[C = Cl |q1 < q1LS ]=0
the equilibrium outputs of firm 1 when she faces high offshore hidden cost q1H S and when
she faces low offshore hidden cost q1LS are related as
F (Ch , Cl , Cd )q1LS + δ[G(Ch )]q1H C (Ch ) − δ[G(Ch , Cd )]q1H C (Ch , Cd )
q1H S ∈ [ ,
F (Ch , Cd )
F (Cl , Cd )q1LS + δ[G(Cl )]q1LC − δ[G(Cl , Cd )]q1LC (Cl , Cd )
]
F (Cl , Ch , Cd )
(5)

Proof In the separating equilibrium, at the end of period 1, firm 1 on moving offshore
chooses to reveal the correct hidden cost. Given the revelation principle, firm 1 can either
choose to produce q1H S or q1LS and reveals only when her incentives out of revelation are
satisfied. The incentive compatibility(IC) conditions for each cost type firm are written
below.
If the first firm faces high hidden cost on moving offshore, the IC for her to produce the
output corresponding to the high hidden cost is given by:
IC for firm 1 when she faces high hidden costs (ICH):
[P H S − (C̄ + Ch )]q1H S + δ[P H C (Cd , Ch ) − (C̄ + Ch )]q1H C (Ch , Cd )
≥ [P LS − (C̄ + Ch )]q1LS + δ[P H C (Ch ) − (C̄ + Ch )]q1H C (Ch ) (6)
We note that P H C (Cd , Ch ) is the prevailing price in the Cournot Game when firm 1 uses
offshore production and firm 2 produces domestically. δ is the discount factor. P H C (Ch ) is
the prevailing price when both players are using offshore production in the Cournot game
308 J Ind Compet Trade (2018) 18: 303–3 18

of period 2. Similarly, the respective outputs are q1H C (Ch , Cd ) and q1H C (Ch ). The equation
can be interpreted simply as follows: the discounted value of profits from revealing the true
type given the beliefs of firm 2 should outweigh the discounted value of profits if firm 1
was to mimic the other type.
Using the notations developed, one can write Eq. 6 simply as:
F (Ch , Cd )q1H S + δ[G(Ch , Cd )]q1H C (Ch , Cd ) ≥ F (Ch , Cl , Cd )q1LS + δ[G(Ch )]q1H C (Ch )
Thus:
F (Ch , Cd )q1H S ≥ F (Ch , Cl , Cd )q1LS + δ[G(Ch )]q1H C (Ch ) − δ[G(Ch , Cd )]q1H C (Ch , Cd )
(7)
From here we obtain a lower bound for q1H S .
One can similarly write down the IC for firm 1 when she faces low hidden costs as(ICL):
[P LS − (C̄ + Cl )]q1LS + δ[P LC (Cl ) − (C̄ + Cl )]q1LC (Cl )
≥ [P H S − (C̄ + Cl )]q1H S + δ[P LC (Cd , Cl ) − (C̄ + Cl )]q1LC (Cl , Cd ) (8)
Once again one can simplify Eq. 8 to write it as:
F (Cl , Cd )q1LS + δ[G(Cl )]q1LC ≥ F (Cl , Ch , Cd )q1H S + δ[G(Cl , Cd )]q1LC (Cl , Cd )
Thus we get an upper bound for q1H S for any given q1LS as:
F (Cl , Cd )q1LS + δ[G(Cl )]q1LC − δ[G(Cl , Cd )]q1LC (Cl , Cd ) ≥ F (Cl , Ch , Cd )q1H S (9)
Combining the expressions from Eqs. 7 and 9 for the lower and the upper bound of q1H S we
get the expression in the proposition.

It is easy to check that the above inequalities increase as Cl is decreased and Ch is

increased. This gives the first observation of this paper.

Observation 1 Firm 1’s incentive to reveal the hidden cost faced by her increases as the
difference (Ch − Cl ) approaches ∞ .

Proof in Appendix A.
So, when the difference between the possible hidden costs is high the first firm has no
incentive to deviate from revealing his type. So in such cases the equilibrium Stackelberg
output i.e. q1H S is produced by firm 1 when she faces high hidden costs correctly revealing
the type to firm 2. Firm 2 in period 1 produces q2H S and they produce q1H C (Ch , Cd ) and
q2H C (Ch , Cd ) respectively at the end of period 2 to maximise profits. Similarly, they produce
for the case when they face low hidden costs of production.

3.4 The decision to offshore

Firm 1 decides to offshore knowing the unraveling of the separating equilibrium from the
values of possible hidden costs. The decision to offshore takes the above into account and
offshoring occurs when the expected profits of offshore production over two periods out-
weigh those of domestic production. In other words, if the signal is Cl with precision p1
then firm 1 offshores only if:
p1 [F (Cl , Cd )q1LS + δ[G(Cl )]q1LC ]
+(1 − p1 )[F (Ch , Cd )q1H S + δ[G(Ch , Cd )]q1H C (Ch , Cd )] > (1 + δ)[π1 ∗]
J Ind Compet Trade (2018) 18: 303–3 18 309

Similarly one can write down the case when the signal is Ch .
The equilibrium output combinations for each period output would be the Stackel-
berg outputs in period 1 and Cournot Outputs in period 2. When hidden cost is Ch , the
Stackelberg outputs are respectively:
q1H S = q1 (J, Cd , Ch ) where J = P −1 and q1 (J ) > 0; q1 (Cd ) > 0 and q1 (Ch ) < 0.
q2H S = q2 (J, Cd , Ch ) where J = P −1 and q2 (J ) > 0; q2 (Cd ) < 0 and q2 (Ch ) > 0.
QH S = q1H S + q2H S
P H S = P (QH S )
The period 2 outputs:
q1H C = q1 (J, Cd , Ch ) where J = P −1 and q1 (J ) > 0; q1 (Cd ) > 0 and q1 (Ch ) < 0.
q2H C = q2 (J, Cd , Ch ) where J = P −1 and q2 (J ) > 0; q2 (Cd ) < 0 and q2 (Ch ) > 0
QH C = q1H C + q2H C
P H C = P (QH C )
When hidden cost is Cl , in period 1, the outputs are respectively:
q1LS = q1 (J, Cd , Cl ) where J = P −1 and q1 (J ) > 0; q1 (Cd ) > 0 and q1 (Cl ) < 0.
q2LS = q2 (J, Cd , Cl ) where J = P −1 and q2 (J ) > 0; q2 (Cd ) < 0 and q2 (Cl ) > 0
QLS = q1LS + q2LS
P LS = P (QLS )
The period 2 outputs:
q1LC = q1 (J, Cd , Cl ) where J = P −1 and q1 (J ) > 0; q1 (Cd ) > 0 and q1 (Cl ) < 0.
q2LC = q2 (J, Cd , Cl ) where J = P −1 and q2 (J ) > 0; q2 (Cd ) < 0 and q2 (Cl ) > 0
QLC = q1LC + q2LC
P LC = P (QLC )

3.5 Pooling equilibrium

In the pooling equilibrium firm 2 imperfectly updates belief about the hidden costs of the
offshore location. This leads to the second proposition:

Proposition 2 In a pooling equilibrium when the belief structure of firm 2 is:

P [C = Cl |q1 = q p ] = p̄
P [C = Cl |q1 = q p ] = 1
player 1 chooses to produce q p in period 1 irrespective of the hidden cost faced and
qp
≥ [F (Ch , Cd )q1H S + δG(Ch )q1H C (Ch ) − (1 − p̄){δG(Ch , Cd )q1H C (Ch , Cd )}
−p̄{δG(Ch )q1H C (Ch )}]/[P P − (C̄ + Ch )] (10)

Proof First, we note that p̄ is the threshold level of precision which equalises the domestic
and the offshore production costs. In effect this means that firm 2 updates his belief about
310 J Ind Compet Trade (2018) 18: 303–3 18

the foreign location having low hidden costs only to the extent that she remains indifferent
between offshoring and staying domestic.
In this case, irrespective of the hidden cost structure being faced by firm 1, she produces
q p in order to protect her monopoly profits from operating in the offshore destination. This
can essentially happen when the discounted profits of operating alone in the offshore loca-
tion net of two periods outshines the profits if the second firm follows in the second period.
We look at the incentive compatibility conditions firm 1 for both cases where she faces high
hidden costs(ICH) or low hidden costs(ICL):
IC for the high hidden cost type to pool (ICH):

[P P − (C̄ + Ch )]q p
+(1 − p̄)[δ{P H C (Cd , Ch ) − (C̄ + Ch )}q1H C (Ch , Cd )]
+p̄[δ{P H C (Ch ) − (C̄ + Ch )}q1H C (Ch )]
≥ [P HS
− (C̄ + Ch )]q1H S + δ[P H C (Ch ) − (C̄ + Ch )]q1H C (Ch ) (11)
IC for the low hidden cost type to pool (ICL):

[P P − (C̄ + Cl )]q p
+(1 − p̄)[δ{P LC (Cd , Cl ) − (C̄ + Cl )}q1LC (Cl , Cd )]
+p̄[δ{P LC (Cl ) − (C̄ + Cl )}q1LC (Cl )]
≥ [P LS − (C̄ + Cl )]q1LS + δ[P LC (Cl ) − (C̄ + Cl )]q1LC (Cl ) (12)
Now, when Cl is reasonably close to Ch the above two ICs are simultaneously satisfied
when the ICH is satisfied. This is shown in Appendix B. Hence we get the expression for
q p as outlined in proposition 2 from Eq. 11.

It can also be checked that as Cl decreases for a given value of Ch the IC of the low
hidden cost facing firm 1 nears equality and finally the IC seizes to be satisfied and the
inequality sign reverses. Hence, this gives the second observation of this paper.

Observation 2 The incentive to pool increases as the difference (Ch −Cl ) approaches zero.

Proof in Appendix B.
This can be explained intuitively. When the hidden costs are sufficiently low, there is little
to lose from the entry of a second firm in terms of profits foregone for operating together in
that location. Hence firm 1’s incentive to protect monopoly decreases.
The pooled case outputs for each period can be written as:
q1 = q1 (J, Ch , Cl , Cd ) where J = P −1 and q1 (J ) > 0; q1 (Cd ) > 0;
p

q1 (Cl ) < 0 and q1 (Ch ) < 0;

q2 = q2 (J, Cd , q1 ) where J = P −1 and q2 (J ) > 0; q2 (Cd ) < 0 and q2 (q1 ) < 0;
p p p

p p
Qp = q1 + q2
P p = P (Qp )
We propose a solution for output q p in the example below which supports the belief of firm
2 that the expected cost of offshore production is same as the domestic cost of production.
J Ind Compet Trade (2018) 18: 303–3 18 311

In this equilibrium firm 2 chooses not to follow and produces as a Stackelberg follower
in period 2. Firm 1 reveals its true type and produces the optimal output to maximise profits
i.e. q1H S or q1LS depending on whether it is facing high or low costs.

4 An Example

We consider a linear example where demand is given by:

P = a − bQ (13)

where a, b are positive parameters, P is the price level of the good and Q is the total
industry output. We assume a to be large enough such that the derived expressions for
outputs and profit are all positive. Thus Q = q1 + q2 where q1∗ and q2∗ are the respective
output levels of the two firms in the duopoly.
The domestic marginal cost of production is assumed to be Cd . We first outline the
pre-offshore production and profit outcomes.
In the case of domestic production, firm i’s profit equation:


2
πi = [a − b( qi )]qi − Cd qi
i=1

In the above equation i represents the firm and i ∈ {1, 2}. The reaction functions:

q1 = (a − Cd )/2b − q2 /2 (14)

q2 = (a − Cd )/2b − q1 /2 (15)

The market equilibrium outputs can thus be expressed as q1∗ = q2∗ = (a − Cd )/3b = q ∗ ,
total output as Q∗ = 2(a − Cd )/3b, price as P ∗ = (2Cd + a)/3. The profits accruing
to each firm and the industry profits thus obtained are π1 = π2 = (a − Cd )2 /9b and
π = 2(a − Cd )2 /9b.
Now, with the possibility of offshore outsourcing: In period 1 when both players offshore,
production decisions are taken based on whether it is a high or low cost destination. The
model continues to be in the Cournot framework as no firm enjoys a first mover advantage.
Under high hidden costs, the output is q1hc = q2hc = (a − (C̄ + Ch ))/3b. It thus follows that
Qhc = (2a − 2(C̄ + Ch ))/3b and P hc = (2(C̄ + Ch ) + a)/3. The profits fall along with
consumer surplus. This lowers total welfare in the society.
When firms face low hidden costs, the output q1lc = q2lc = (a − (C̄ + Cl ))/3b It thus
follows that Qlc = (2a − 2(C̄ + Cl ))/3b and P lc = (2(C̄ + Cl ) + a)/3. The profits increase
along with consumer surplus. This increases total welfare in the society.
When only one firm offshores, say firm 1 has moved offshore and firm 2 stays domestic.
In this case one would have a separating or pooling equilibria as detailed in the previous
section.

4.1 Separating Equilibrium

In a separating equilibrium, the actions of firm 1 in period 1 completely reveal the level of
hidden cost that would be faced offshore.
312 J Ind Compet Trade (2018) 18: 303–3 18

Firm 2 believes that if firm 1 produces output q1H S (q1LS ) in period 1 then the offshore
destination is high(low) cost.
So firm 2’s beliefs are:
P [C = Cl |q1 ≥ q1LS∗ ] = 1and
P [C = Cl |q1 < q1LS∗ ] = 0.
In period 1, when the firms face high hidden costs, the solutions are: q1H S∗ = ((a + Cd )/2 −
(C̄ +Ch ))/b; q2H S∗ = (a −3Cd )/4b +(C̄ +Ch )/2b; QH S∗ = (3a −Cd )/4b −(C̄ +Ch )/2b;
P H S∗ = (2(C̄ + Ch ) + Cd + a)/4; π1H S = (a + Cd − 2(C̄ + Ch ))2 /8b; π2H S = (a − 3Cd +
2(C̄ + Ch ))2 /16b.
We note here that the comparisons with the Cournot case are not obvious as described in
the previous section.
While facing low hidden costs the solutions are: q1LS∗ = ((a + Cd )/2 − (C̄ + Cl ))/b;
q2 = (a − 3Cd )/4b + (C̄ + Cl )/2b; QLS∗ = (3a − Cd )/4b − (C̄ + Cl )/2b; P LS∗ =
LS∗

(2(C̄+Cl )+Cd +a)/4; π1LS = (a+Cd −2(C̄+Cl ))2 /8b; π2LS = (a−3Cd +2(C̄+Cl ))2 /16b
The second period outputs of each firm, price faced by consumers and profits are given
by: For high hidden costs: q1H C = (a + Cd − 2(C̄ + Ch ))/3b; q2H C = [(a − Cd ) +
(C̄ + Ch − Cd )]/3b; QH C = (2a − Cd − C̄ − Ch )/3b; P H C = (a + Cd + C̄ + Ch )/3;
π1H C = (a+Cd −2(C̄+Ch ))2 /9b; π2H C = (a−2Cd +(C̄ +Ch ))2 /9b; π H C = π1H C +π2H C .
In case of low hidden costs: q1LC = (a − C̄ − Cl )/3b; q2LC = (a − C̄ − Cl )/3b;
Q LC = 2(a − C̄ − Cl )/3b; P LC = (a + 2(C̄ + Ch ))/3; π1LC = (a − C̄ − Cl )2 /9b;
π2 = (a − C̄ − Cl )2 /9b; π l = 2(a − C̄ − Cl )2 /9b;
LC

Given the above one can also check the validity of observation 1 in the context of the
example. Here:
a + Cd + 2(C̄ + Cl )
F (Ch , Cl , Cd ) = [ − (C̄ + Ch )]
4
q1LS = (a + Cd )/2b − (C̄ + Cl )/b
Now,
FCl ()q1LS + F ()q1C
LS
l

1 (a + Cd ) (C̄ + Cl ) 1 (a + Cd + 2(C̄ + Cl )
= [ − ]− [ − (C̄ + Ch )]
2 2b b b 4
(Ch − Cl )
= >0
b
Hence as Cl decreases keeping Ch constant, the IC of the high type holds more strongly and
hence his incentive to reveal improves.

4.2 Pooling Equilibrium

Firm 2 imperfectly updates belief from firm 1’s move in period 1 and the updated belief is
a result of firm 1’s decision to move offshore and the output choice. So the updated belief
at the end of period 1 for firm 2 is
P [C = Cl |q1 = q p ] = p̄
P [C = Cl |q1 = q p ] = 1
J Ind Compet Trade (2018) 18: 303–3 18 313

p̄ makes firm 2 indifferent between moving offshore and remaining onshore. The candidate
solution for q p is the Stackelberg leader output of firm 1 had firm 1 been a leader while
operating domestically. This level of output satisfies the pooling ICs for a selection of
values of Cl and Ch . This keeps expected costs of production offshore exactly equal to
domestic production for firm 2.
p
Hence outputs and price in period 1: q1 = q1S = (a − Cd )/2b; q2S = (a − Cd )/4b;
Q = 3(a − Cd )/4; P = (a + 3Cd )/4;
p p

In period 2 firm 1 produces as per the revealed costs and firm 2 serves the residual
market. In the following section an extension to the model is offered. The model remains
identical after the movement of the first firm in this extension. However, the extension adds
light into the choice of offshoring destination.

5 Two Locations

We have two competing destinations where jobs may be offshored, A and B. The con-
tracted costs are C̄ A and C̄ B . The hidden costs are C A ∈ {ClA , ChA } for destination A and
C B ∈ {ClB , ChB } for destination B. The two firms get signals about hidden costs for both the
j
locations and the precisions for each location for each firm is given by pi where i ∈ {1, 2}
and j ∈ {A, B}. Two interesting cases are analysed below.

5.1 Offshoring to the Location with Higher Contracted Costs

Both players get signals of lower hidden cost from both locations. Firms offshore to the
location with the lowest expected costs of production. However we claim and prove below
that the destination with higher known costs of offshore production may be unambiguously
preferred (more profitable) under specific conditions.

Claim An offshore destination with similar or higher known costs of production may turn
out to be the more lucrative destination if :
a) The precision of signal from the destination is higher with similar range of hidden costs
as its competitor or
b) The range of hidden costs in that location is lower given that the lower bound(Cl ) of
both the locations are identical and the precisions are same.
In both cases the absolute difference in expected hidden costs must outshine the absolute
difference in known costs of offshoring.

Proof A and B are the two destinations where jobs could be offshored. We assume C̄ A ≥
C̄ B . It is assumed that the signals received are those of low hidden cost in both the locations.
Now destination A is more profitable when:
[P (Q) − C̄ A − piA ClA − (1 − piA )ChA ]q1A > [P (Q) − C̄ B − piB ClB − (1 − piB )ChB ]q1B (16)
then the firms offshore to destination A. In the above expression, q1A and q1B are outputs cal-
culated out of expected costs of offshore production in the respective destinations. Writing
piA ClA + (1 − piA )ChA as EA and piB ClB + (1 − piB )ChB as EB and using the linear example
notations we rewrite the above equation as:
[a − b(q1A + q2A ) − C̄ A − EA]q1A > [a − b(q1B + q2B ) − C̄ B − EB]q2B (17)
314 J Ind Compet Trade (2018) 18: 303–3 18

C̄ A +EA A +EA C̄ A +EA

i.e. [a − b( a+C
2b −
d
b + C̄ 2b + a−3C
4b ) − C̄ − EA]( 2b −
d A a+Cd
b ) > [a −
a+Cd C̄ B +EB C̄ B +EB a−3Cd a+Cd C̄ B +EB
b( 2b − b + 2b + 4b ) − C̄ − EB]( 2b −
B
b )
a+Cd −2C̄ A −2EA a+Cd −2C̄ A −2EA a+Cd −2C̄ B −2EB a+Cd −2C̄ B −2EB
i.e. ( 4 )( 2b )>( 4 )( 2b )
i.e. (a + Cd )2 − 4(a + Cd )(C̄ A + EA) + 4(C̄ A + EA)2 > (a + Cd )2 − 4(a + Cd )(C̄ B +
EB) + 4(C̄ B + EB)2
i.e. (C̄ B + EB − C̄ A − EA)[a + Cd − (C̄ B + EB + C̄ A + EA)] > 0
Now the above inequality holds when both terms are negative or both are positive. When
both are negative, it follows that destination A’s total expected unit costs of production are
unambiguously higher than in B. Hence it would be unprofitable to offshore there in a linear
cost framework. Thus for the above inequality to hold otherwise, EA < EB is a necessary
condition and EB −EA > C̄ A − C̄ B is a sufficient condition. From the necessary condition:
pB ClB + (1 − pB )ChB > pA ClA + (1 − pA )ChA (18)
i.e. ChB − ChA > pB (ChB − ClB ) − pA (ChA − ClA )
(a) When ChA = ChB and ClA = ClB The above inequality holds only when pA > pB .
This is part (a) of the claim.
(b) When pA = pB and ClA = ClB The inequality follows if ChB > ChA . This is part (b)
of the claim.
From the sufficient condition it emerges that the absolute difference in expected hidden
costs should be higher than the absolute difference in known costs of offshoring.

The lower bound of possible hidden costs Cl may be interpreted as costs that accrue
while outsourcing to any destination (unavoidable costs incurred at the minimum). Hence
destinations would vary in terms of the range of possible hidden costs driven by the higher
value (Ch ). The above two cases have been illustrated to show: in case (a) the role of preci-
sion or correctness of the signal in case of destinations with similar range of hidden costs;
and in case (b) Under similar precision of signals, a lower range of possible hidden costs
play a part in facilitating offshoring to locations which otherwise have high contractible
costs(C̄). 4

5.2 Herding

We consider the case where firms receive signals to believe different locations to be prof-
itable i.e. firm 1(2) believes location A(B) to be profitable. We outline an interesting case
which entails herding behaviour by one of the firms below:
Let C̄ A = C̄ B
But,
C̄ + p1A ClA + (1 − p1A )ChA < Cd < C̄ + p1B ClB + (1 − p1B )ChB (19)
and
C̄ + p2A ClA + (1 − p2A )ChA > Cd > C̄ + p2B ClB + (1 − p2B )ChB (20)
We consider the specific case of Cl < Cl and Ch = Ch with a separating equilibrium.
A B A B

Then firm 1 chooses to move in the first period. But firm 2 may choose to wait given

4 This is what can be seen in reality in the case of German firms offshoring activities to Austria rather than
India. India with its lower fixed(contractible) costs of production turns out to be less competitive than Austria
given that cultural and language differences push up the higher bound of its hidden costs.
J Ind Compet Trade (2018) 18: 303–3 18 315

irreversible offshoring. Depending on the first period output and the updated beliefs of firm
2, firm 2 may choose to follow firm 1 in period 2 to maximise profits over the two periods
in the way described in the game for the separating equilibrium case. This happens when
the following holds true for firm 2:

[P LSA − Cd ]q2LSA + δ[P LCA (ClA ) − (C¯A + ClA )]q2LCA (ClA )

≥ (1 + δ)[P LSAB − (C¯B + p B C B + (1 − p B )C B )]q LSAB
2 l 2 h 2 (21)

In the above expression: P LSA = The price charged in the first period where firm 1 has
offshored to destination A while firm 2 is producing in the home country.
q2LSA = the Stackelberg follower’s output produced by firm 2 when firm 1 has offshored
to destination A and has revealed low hidden costs in the location under a separating equilibrium.
P LCA = the Cournot equilibrium price charged in period 2 when both players are
producing under a low hidden cost framework in destination A in period 2.
q2LCA (ClA )= the Cournot output produced by firm 2 in period 2 when both are producing
in destination A under the lower hidden cost framework.
P LSAB = the expected price charged when firm 1 has offshored to destination A(revealing
low hidden costs) and firm 2 contemplates offshoring to destination B under the expected
costs highlighted in the equation.
q2LSB = The planned output that firm 2 would produce when it offshores to destination B
with firm 1 offshoring to destination A where low hidden costs are revealed.
This case of firm 2 choosing to ignore his own signals and information so as to follow
firm 1 forms the case for herding in offshoring. We look at the implications of the results
and intuitions derived so far in this paper in the next section.

6 Conclusion

We have looked at the micro aspects of offshore outsourcing decisions in the presence of
hidden costs and signals of imperfect precision. Formulating the analysis as a Bayesian
game, we looked at the offshoring decision in a two period framework where two firms
interact strategically, leading to interesting outcomes. These are based on the precision of
signals, updating due to observation of moves by the first mover and the range of hidden costs.
We find that a destination that has a lower dispersion of hidden costs or a high precision
of signals in favour of low hidden costs would be a more preferred offshore destination.
Even if the known costs of offshore production are higher for the location, higher precision
of signals i.e. lower perceived hidden costs make it more attractive.
We also show that herding in favour of an offshoring destination is a possibility. A partic-
ular case of the firms receiving opposing signals is analysed to this effect. Here we showed
that a higher degree of imperfect information(lower precision) may lead a firm to follow the
first mover when the first mover’s destination is revealed more competitive than the other.
This is the case of herding behaviour in offshoring as the second firm completely overlooks
his own information.
In terms of policy implications, the analysis would suggest that governments of countries
aiming to be lucrative offshoring destinations may want to lower the range of hidden costs
involved in working in their geography. Reduction in the lower limit of hidden costs as much
as possible or increasing the precision in favour of lower hidden costs is a necessity. Initia-
tives of the government may include advertising, tax holidays, upgrading human capital etc.
At the firm level, highlighting the ability to deliver world class solutions on diverse projects
316 J Ind Compet Trade (2018) 18: 303–3 18

around the world, brand image and reputation, reduce uncertainties related to the potential
of offshore business partner. Interviews with senior executives of Indian BPO companies
conducted by the author revealed the above aspects to be crucial. Finally, in the Indian case,
the role organisations like NASSCOM in promoting the IT industy in foreign countries and
thereby reducing perceived hidden costs cannot be denied.

Acknowledgements The author acknowledges the help of Prof. Soumyanetra Munshi in guiding to
formalise this paper.

Appendix A: Proof of Observation 1

We simplify equation 8 and write it as:

F (Ch , Cd )q1H S + δ[G(Ch , Cd )]q1H C (Ch , Cd ) ≥ F (Ch , Cl , Cd )q1LS + δ[G(Ch )]q1H C (Ch )
where, F (Ch , Cd ) = [P H S − (C̄ + Ch )] and P H S is a function of Cd and Ch . G(Ch , Cd ) =
[P H C (Cd , Ch ) − (C̄ + Ch )] and similarly for the others.
Now, that gives the lower bound for q1H S i.e.
F (Ch , Cd )q1H S ≥ F (Ch , Cl , Cd )q1LS + δ[G(Ch )]q1H C (Ch ) − δ[G(Ch , Cd )]q1H C (Ch , Cd )
Differentiating this expression wrt Cl we get:
FCl ()q1LS + F ()q1C
LS
l

Now, for a given demand function and the costs of production Ch , q1H S is the profit max-
imising Stackelberg output. For any increases in output the revenues increase less than the
costs. Hence at the given cost Ch it becomes more and more unprofitable to produce more
output. Hence for conventional demand functions with decreasing marginal revenue this
expression comes out to be positive.
So, as Cl decreases, the RHS reduces in value. Hence the inequality holds more strongly.
Thus the higher is (Ch − Cl ) the stronger the possibility of the high type to reveal his true
hidden cost. Hence observation 1 stands proved.

Appendix B: Proof of Observation 2

Qualitative proof of one IC binding in the pooling equilibrium: ICH:

[P P − (C̄ + Ch )]q p
+(1 − p̄)[δ{P HC
(Cd , Ch ) − (C̄ + Ch )}q1H C (Ch , Cd )]
+p̄[δ{P H C (Ch ) − (C̄ + Ch )}q1H C (Ch )]
≥ [P H S − (C̄ + Ch )]q1H S + δ[P H C (Ch ) − (C̄ + Ch )]q1H C (Ch ) (22)
i.e.:
[P P − (C̄ + Ch )]q p
≥ [P H S − (C̄ + Ch )]q1H S + δ[P H C (Ch ) − (C̄ + Ch )]q1H C (Ch )
−(1 − p̄)[δ{P H C (Cd , Ch ) − (C̄ + Ch )}q1H C (Ch , Cd )]
−p̄[δ{P H C (Ch ) − (C̄ + Ch )}q1H C (Ch )] (23)
J Ind Compet Trade (2018) 18: 303–3 18 317

ICL:
[P P − (C̄ + Cl )]q p
+(1 − p̄)[δ{P LC (Cd , Cl ) − (C̄ + Cl )}q1LC (Cl , Cd )]
+p̄[δ{P LC (Cl ) − (C̄ + Cl )}q1LC (Cl )]
≥ [P LS − (C̄ + Cl )]q1LS + δ[P LC (Cl ) − (C̄ + Cl )]q1LC (Cl ) (24)
i.e.:

[P P − (C̄ + Cl )]q p
≥ [P LS − (C̄ + Cl )]q1LS + δ[P LC (Cl ) − (C̄ + Cl )]q1LC (Cl )
−(1 − p̄)[δ{P LC (Cd , Cl ) − (C̄ + Cl )}q1LC (Cl , Cd )]
−p̄[δ{P LC (Cl ) − (C̄ + Cl )}q1LC (Cl )] (25)
Now it may be noted that LHS of the ICH is less than LHS of ICL. Similarly RHS
of ICH is less than RHS of ICL. Thus, the IC of the high type alone guarantees the low
type IC as well. It is easy to note that as Cl increases profitability of the low hidden cost
type falls. Hence leaving Ch unchanged, one gets a situation where both inequalities are
simultaneously satisfied.

Proof of Observation 2: Following on from above:

[P P − (C̄ + Ch )]q p ≥ [P LS − (C̄ + Cl )]q1LS + δ[P LC (Cl ) − (C̄ + Cl )]q1LC (Cl )

−(1 − p̄)[δ{P LC (Cd , Cl ) − (C̄ + Cl )}q1LC (Cl , Cd )]
−p̄[δ{P LC (Cl ) − (C̄ + Cl )}q1LC (Cl )] (26)
For a lower Ch , the LHS increases and we have strict inequality. Hence the ICs become
stronger. [Proved]

References

Antras P (2003) Firms, contracts, and trade structure. Q J Econ 118(4):1375–1418

Barthelemy J (2001) The hidden costs of IT outsourcing. MIT Sloan Management Review
Baruch Y, Hind P (2000) Survivor syndrome - a management myth? J Manag Psychol Emerald 15(1):29–45
Borga M, Zeile W (2004) International fragmentation of production and intrafirm trade of US multinational
comapnies. Bureau Of Economic Analysis Working paper No. WP2004-02
Chamley C, Gayle D (1994) Information revelation and strategic delay in a model of investment. Economet-
rica 62(5):1065–1085
Davison D (2004) Top 10 risks of offshore outsourcing. CIO
DiamondCluster (2004) 2004 Global IT outsourcing study. DiamondCluster International
Goolsby K (2002) Offshore is not offhand. Recommendations for effective offshore outsourcing, White
Paper. Everest Group, Dallas
Grossman G, Helpman E (2002) Integration versus outsourcing in industry equilibrium. Q J Econ 117(1):85–
120
Grossman G, Helpman E (2003) Outsourcing vs FDI in industry equilibrium. J Eur Econ Assoc 1(2-3):317–
327
Karamouzis F (2002) Debunking the myths of offshore IT service offerings. ID number DF-15-5315 (March
21). Gartner Inc., Stamford
318 J Ind Compet Trade (2018) 18: 303–3 18

Milgrom P, Roberts J (1982) Predation, reputation, and entry deterrence. J Econ Theory 27(2):280–312
Navaretti GB, Venables AJ (2006) Multinational firms in the world economy. Economics Books. Princeton
University Press, Princeton
Overby S (2003) The hidden costs of offshore outsourcing. CIO Magazine
Qu Z, Brocklehurst M (2003) What will it take for China to become a competitive force in offshore
outsourcing? An analysis of the role of transaction costs in supplier selection. J Inf Technol 18:53–67
Sabherwal R (1999) The role of trust in outsourced IS development projects. Commun ACM 42(2):80–86
Yeats AJ (2001) Just how big is global production sharing?. In: Arndt SW, Kierzkowski H (eds) Frag-
mentation: new production patterns in the world economy. Oxford University Press, Oxford, pp 108–
143
Zhang J (1997) Strategic delay and the onset of investment cascades. RAND J Econ 28(1):188–205