Behavioral Approaches to Optimal FDI Incentives

Author(s): M. Rosenboim, I. Luski and T. Shavit

Source: Managerial and Decision Economics, Vol. 29, No. 7 (Oct., 2008), pp. 601-607
Published by: Wiley
Stable URL: http://www.jstor.org/stable/25151633
Accessed: 24-04-2018 16:44 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

Wiley is collaborating with JSTOR to digitize, preserve and extend access to Managerial and
Decision Economics

This content downloaded from 181.65.173.6 on Tue, 24 Apr 2018 16:44:29 UTC
All use subject to http://about.jstor.org/terms
MANAGERIAL AND DECISION ECONOMICS
Manage. Decis. Econ. 29: 601-607 (2008)
Published online in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/mde. 1435

Behavioral Approaches to Optimal FDI

Incentives
M. Rosenboima, I. Luskib and T. Shavitcd*
a Department of Economics, Sapir College, Sdearot, Israel
Department of Economics, Ben-Gurion University of the Negev, Beer-Sheva, Israel
c Department of Management, Ben-Gurion University of the Negev, Beer-Sheva, Israel
d College of Management, Business School, Israel

Countries attempt to attract foreign investors by offering them a set of incentives. The most
common types of foreign direct investment incentives are grants and tax relief. Although the
amount of the grant is independent of future situations, the value of a tax relief depends on
future profits. Our study used the behavioral approach to test experimentally the preferences
of managers regarding the desired types of incentives under various conditions. We found,
'Regret Effect', 'Statues Quo Bias', and 'Insurance Effect' in subjects' decision making. A
country can improve the incentives it offers by considering the various behavioral biases of the
companies' managers. Copyright ? 2008 John Wiley & Sons, Ltd.

INTRODUCTION officials may improve the government's choice of

Investment by a multinational corporation (MNC)
is accompanied by various substantial benefits for
the host country's economy. Governments that
view the MNC as an engine for growth and
development, and may offer various incentives in
order to attract the investment of MNCs. Our
paper focuses on the process of choosing the
optimal set of incentives given by the government
of the host country. The novelty of this paper is in
the inclusion of a behavioral theory and approach,
considering the optimal set of foreign direct
investment (FDI) incentives. We hypothesized that
an MNC is governed by a CEO and other senior
officers, and their considerations for choosing the
right location for their investment follows the well
known biases of behavioral economics. We argue
that understanding the considerations of the MNC
""Correspondence to: College of Management, 7 Rabin Avenue,
Rishon-Le'Zion, Israel. E-mail: shavittal@gmail.com
an optimal set of incentives.
The globalization processes that have taken
place for the past two decades, as well as the
development of MNCs, have led to a search for
new sites for investment. The process of choosing a
site for this purpose may be divided into two
stages. In the first stage, the MNC examines
various possible sites, focusing on the conditions
of the site by considering its economic stability, the
availability of professional labor, the political and
security situation, the legal system, modern infra
structures, trade agreements and trade treaties, etc.
A short list of suitable sites is created.
In the second stage, the MNC looks for the most
profitable site for investment by examining the
short list of countries. Usually, the MNC negoti
ates with the appropriate governments looking for
the optimum set of incentives. The countries that
are in the short list having similar economic
characteristics may compete among themselves

Copyright ? 2008 John Wiley & Sons, Ltd.

This content downloaded from 181.65.173.6 on Tue, 24 Apr 2018 16:44:29 UTC
All use subject to http://about.jstor.org/terms
 602 M. ROSENBOIM ET AL.

by improving their offer of incentives for winning investment location. Both the countries and the
the foreign investment. A host country that is MNC base their decision on the present value of
competing for the foreign investment may offer the flows of future revenues and outlays. Future
foreign company financial incentives in order to revenues depend on the productivity in the host
raise its attractivity as the best site for the country and the efficiency of the entering MNC.
investment (UNTCAD, 2000). For example, Intel Information on the level of productivity may be
established a factory in Costa Rica by investing available to either the host country or the MNC.
600M US$. This factory employs about 2000 Asymmetric information plays a major role in the
workers. The Costa Rica government gave Intel process of choosing the optimal set of incentives.
a grant of 300M US$. Another example is Canon, When there is complete information on the
which sought a place to establish a factory in Asia. reliability of the foreign company as well as on
The Vietnamese government offered Canon a tax the efficiency of the host country, the specific type
relief for 10 years, which was a major factor in of incentive (grant or tax relief) is of little
Canon's decision to establish its factory in importance to either the country or the MNC.
Vietnam (Bjorvatn and Eckel, 2006). However, under incomplete information on pro
The two most common types of incentives are ductivity, the mix of grant and tax relief is
grants and tax relief. The impact of these important and has a major impact on the MNC
incentives on the decision of where the company and the country's decision. In that case, on the one
will be located has increased over time (Taylor, hand, the country fears that the foreign company
2000) and, as a result, an incentive competition has will exploit a one-time grant, and may within a
emerged. This competition may cause a shift from short time terminate its investment and leave the
the host country to the foreign company of the country. On the other hand, the country may fear
major part of the surplus from the activities of the that substantial tax relief will lead to exaggerated
foreign company (Oman, 2000). Blomstrom and benefits for the MNC when profits are high.
Kokko (2003) suggest avoiding this competition Conversely, the foreign company that invests in
by having agreements between the countries the project may worry that the host country
regarding the 'rules of the game' of FDI incentives, will not fulfill future commitments for the tax
especially between countries in specific regions in relief. In this case, the foreign company is already
which the competition might be fierce. 'bound' to the country by the sunk costs, and
A number of models have been developed for switching to another site is not relevant. Therefore,
examining FDI incentives and their efficiencies. under risk aversion and increasing risk, the MNC
Barros and Cabral (2000) examined the influence prefers a one-time grant whereas the countries
of the size of a country and its unemployment rate prefer tax relief.
on the incentives given to an MNC. According to The literature on individual decision making
their model, subsidy competition may lead to and choice has long ago acknowledged that human
optimal FDI location, whereas a policy of no agents violate many of the fundamental assump
subsidy may distort it. Two reasons can support tions of economic theory (see the survey by
this proposition. First, the competition among Camerer, 1995). Empirical evidence accumulated
countries compels them to commit to improving in classroom experiments and economic labora
their respective infrastructure and macroeconomic tories has motivated new generalized theories (e.g.
system and, second, that competition can bring an Kahneman and Tversky's 1979 Prospect Theory)
optimal allocation of FDI to those countries which that may be utilized to explain the experimental
benefit more from FDI incentives (see Bjorvatn evidence and gain insight into various aspects of
and Eckel, 2006). economic reality. Behavioral effects have become a
Kaplan et al. (2003) studied the influence of significant part of explanations regarding human
information on the optimal mix of fixed grant and decision making under uncertain conditions. Stu
tax relief to be offered to an MNC. They showed dies in the behavioral economics and finance fields
that in equilibrium, the competing countries tend use behavioral effects to explain various contra
to use grants as the main tool to attract FDI. The dictions to known theories. Results from economic
participants in this game are countries that have to experiments are used to support behavioral effects.
offer a set of incentives to attract FDI, and an Kanheman and Tversky (1979) developed an
MNC that has to make a decision regarding its alternative model to the expected utility theory,

Copyright ? 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 601-607 (2008)
DOI: 10.1002/mde

This content downloaded from 181.65.173.6 on Tue, 24 Apr 2018 16:44:29 UTC
All use subject to http://about.jstor.org/terms
 OPTIMAL FDI INCENTIVES 603

the Prospect Theory. They suggest that subjects' The questionnaire included 20 questions that
value function V should be treated as a function in examined the investors' preferences between the
two arguments: the asset position that serves as a investment grants and the tax relief as dependency
reference point and the magnitude of the change in the risk embodied in the expected cash flow.
(positive or negative) from that reference point. Furthermore, some of the questions considered the
Kanheman and Tversky proposed that the value possibility of reducing the grant or the tax relief in
function is (1) defined based on deviations from exchange for receiving insurance in case of loss.
the reference point, (2) generally concave for gains The rest of the paper is organized as follows:
and commonly convex for losses, and (3) steeper First, in the second section we describe the
for losses than for gains so that v(x) < ? v(?x). experimental procedure. Next, in the third section,
They also suggest that people overweigh outcomes we present the primary results and provide
that are considered certain, relative to outcomes some possible explanations. Finally, we summarize
that are merely probable, and named this effect the the conclusions and present some policy implica
'Certainty Effect'. tions.
Kachelmeier and Shehata (1992) conducted a
series of laboratory experiments in China in
order to elicit people's certainty equivalents for
EXPERIMENTAL PROCEDURE
a sequence of lotteries. They found that the
average ratios of certainty are equivalent to
the expected values for the high-prize trials, and The participants in the experiment were 102 MBA
were systematically lower than the ratios for students from Ben-Gurion University and the
the low-prize trials across win percentages. There
College of Management. However, we used the
was a marked trend from risk-loving or risk results of 86 subjects (55 men and 31 women;
neutral preferences to risk-averse as payoffs average age 33.6 years), because 16 subjects either
increased. did not answer all the questions or omitted some
data, and their results were out of the range of two
Benartzi and Thaler (2004) investigated the
value of freedom of choice. In their study, they standard deviations. The experiment took place
presented retirement investors with information during a regular lesson, and lasted approximately
half an hour.
about the distribution of outcomes that they could
expect to obtain from the portfolios they had We presented the subjects with four different
selected, and from the median portfolio selected by
cash flow scenarios with equal means but different
risks, and confronted them with two initial states
their peers. A majority of participants actually
as follows: First we asked them for the minimum
preferred the median portfolio, and this provided
some evidence for the fact that investors do not grant they were willing to accept in order to
concede the tax relief.
have well-defined preferences.
The purpose of this study is to test the impact of
Second, a similar question was then asked about
behavioral biases on the decision process of the minimum tax relief they were willing to get in
professional managers and financial experts when order to concede the grant.
they consider their preference for the grant or the
The fourth scenario (the insurance case) con
tax relief under various risky situations. We sidered the possibility of losses, and the subjects
examined experimentally the influence of beha were asked to state the amount they were willing to
pay in order to be insured if losses occur.
vioral effects on the investors' decision making
regarding incentives. We used a questionnaire to
determine the optimal mix of tax relief and grants
Description of the Scenarios
while taking into consideration various psycholo
gical effects. The questionnaire was distributed to Table 1 presents the possible cash flow in each
a large number of MBA students with prior scenario.
managing experience, some of them were accoun In each scenario, subjects were asked to choose
tants. Their responses led us to propose an between receiving 100000 NIS grant and having
improved mix of incentives that can reduce the their tax reduced from 35 to 25% (10% tax relief).
country's cost of incentives such that the attrac The subjects were told that in the case of a loss
tiveness of the country is not hurt. (scenario 4) they would get tax refund.

Copyright ? 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 601-607 (2008)
DOI: 10.1002/mde

This content downloaded from 181.65.173.6 on Tue, 24 Apr 2018 16:44:29 UTC
All use subject to http://about.jstor.org/terms
 604 M. ROSENBOIM ET AL.

Table 1. The Possible Scenarios of Cash Flow Table 3 presents the results
comparison between the min
tax relief of the different scen
Scenario Expected value
A 1 000 000 NIS with no risk 1 000 000 NIS
Probabilities 50% 50%
B 500000 NIS 1500 000 NIS 1000 000 NIS Table 3. Comparing Minimu
C 0 NIS 2 000 000 NIS 1 000 000 NIS mum Tax Relief Betwe
D -100 000 NIS 2100 000 NIS 1000 000 NIS statistics and (/rvalue)]
NIS = New Israeli Shekels.
Test between Minimum gr
scenarios difference relief difference
AandB 7.68(0.03) 1.19(0.00)
B and C 24.77 (0.00) 0.07 (0.89)
In scenario D subjects faced the possibility of
loss. In this case they were asked for theC and D -12.42 (0.10) -0.91 (0.01)
compensation they are willing to get in the case
of a loss in order to reduce the grant or the tax
relief. In this case we tested the insurance premiumTable 3 shows that increasing the ris
of the subjects. of the cash flow raise the minimum grant
The questions were presented as follows: According to the expected utility theory, a
the risk increases the subjective certainty
1. The grant of 100 000 NIS is reduced to 50 000
equivalent of risk-averse subjects decreases, an
NIS and you are compensated in the case of a
vice versa for risk seekers. The minimum grant i
loss. What is the minimum compensation this case is the subjective certainty equivalent
required in the case of a loss so that you would
meaning that the subjects show risk-seekin
accept the reduced grant? behavior.
2. The tax relief would be 5% instead of 10%
Although scenario D is riskier than scenario C,
(reduction from 35 to 30% tax rate), and in the
we found that the minimum grant for scenario D is
case of a loss you can receive compensation.lower. In scenario D the tax relief also has a
What is the minimum compensation you are negative effect, as in the case of a loss (-100 000
willing to receive in case of a loss in order for
NIS) tax relief reduces the taxation refund.
you to accept the reduced tax relief?
Although the expected value of the tax relief is
the same as in the other scenarios, the possibility
RESULTS of a loss has higher effect as losses have higher
effect than profits (loss aversion) (Kahneman and
Table 2 presents the following results:
Tversky, 1979). Table 2 also shows that the
required minimum tax relief increases as the risk
of the cash flow is increased (but the difference
1. The average minimum grant the subjects were
willing to accept in order to give up 10%between
tax scenarios B and C was not found to be
relief (expected value is 100 000 NIS for each
significant).
scenario). The certain alternative, the grant, is the same in
2. The average minimum tax relief the subjects all the scenarios. An increase in the required
were willing to get in order to give up 100 000 minimum tax relief between scenarios A and B is
NIS grant. typical to risk-averse subjects. The same tax relief
in scenarios B and C is typical to risk-neutral
subjects. These results are inconsistent with the
Table 2. Minimum Grants and Minimum Tax
Reliefs risk attitude that was found when the question
focused on the required minimum grant. The
Scenario Minimum grant Minimum increase in the tax relief between scenarios C and
(thousand NIS) tax relief
D is consistent with the decrease in the grant
A 113.7 11.73% between these two scenarios caused by the loss
B 124 12.88% aversion.
C 152.9 12.98% It seems that the subjects' preferences changed
D 141.1 14.02%
when we switched from a grant to tax relief. When

Copyright ? 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 601-607 (2008)
DOI: 10.1002/mde

This content downloaded from 181.65.173.6 on Tue, 24 Apr 2018 16:44:29 UTC
All use subject to http://about.jstor.org/terms
 OPTIMAL FDI INCENTIVES 605

subjects face an initial state in which they have tax these two cases is not significant (^-statistics is
relief with an expected value of 100000 NIS they 0.042 and significance is 0.967).
ask for a grant above 100000 NIS (7Mest The effect of the insurance on the expected value
significance <0.01 for the hypothesis that the of the project is significant, and subjects were
average grant is lower than 100000 NIS in all willing to reduce their expected value in order to
scenarios), as we expected from risk-seeking avoid the possibility of loss. This behavior is
subjects. However, when they faced a state in consistent with the loss-aversion behavior we saw
which they can get a sure grant of 100 000 NIS in scenario D.
they asked for tax relief above 10% (T-test As mentioned earlier, some of the subjects were
significance <0.01 for the hypothesis that the accountants who were studying for their MBA,
average tax relief is lower than 10% NIS in all and their decisions were not the result of
scenarios). A tax relief above 10% gives the misunderstanding the meaning of tax relief.
subjects an expected discount above 100000 NIS Table 4 presents a comparison between accoun
as expected from risk-averse subjects. tants and non-accountants.
A possible explanation for the inconsistent
results is that subjects pay more attention to the Table 4. Minimum Grant and Minimum Tax
tax relief if he/she deems the project as having Relief
high cash value. In this case the increase in the
high-valued cash flow (which is also an increase Scenario Minimum grant Minimum tax relief (%)
in the cash flow's risk) increases the minimum Accountants Others Accountants Others
grant. The behavioral bias, which explains
this behavior, is the 'Regret Effect' (Loomes & A 109 118 11.57 11.86
B 116 130 11.91 13.73
Sugden, 1987; Loomes, 1988). According to this C 125* 176 13.12 12.87
effect subjects are motivated by the possible D 129* 151 13.55 14.44
regret that they will experience if they give up the *T-test significant <0.05 for the hypothesis that the average
tax relief and the cash flow will be in its higher minimum grant (or minimum tax relief) for accountants is not
value. different than that of others.

Next we analyze the insurance case in which

subjects can insure their losses (the fourth scenar
io). In the first case of the insurance, the grant of
100000 NIS was reduced to 50000 NIS and in the
case of a loss the subjects get a refund. The average
refund the subjects were willing to accept in the
case of a loss in order to accept the grant reduction
was 69 468 NIS. A refund of 69468 NIS gives the
subjects an expected value of 34 734 NIS (69 468*
0.5), because the refund is given only in the case of
a loss. The expected value of 34 734 NIS is lower
than the grant reduction by 50000 NIS (T-test
significance <0.01).
In the second case of insurance, the tax relief of
10%o was reduced to 5% and in the case of a loss
the subjects get a refund. The average refund the
subjects were willing to accept in the case of a loss
in order to reduce the tax relief was 68 829 NIS. A
refund of 68 829 NIS gives the subjects an expected
value of 34415 NIS (68 829*0.5), because the
refund is given only in the case of loss. The
expected value of 34415 NIS is lower than the
expected value of the tax relief reduction by 50 000
NIS (r-test significance <0.01). It turns out that
the difference in the average expected value of
Although accountants are more experienced in
tax calculations and decision making concerning
incentives, overall we did not find significant
differences between them and the other subjects
(except for grant in scenarios C and D). This result
is consistent with the evidence that people with
experience or expertise are not immune to framing
effects (e.g. Roszkowski & Snelbecker, 1990; Loke
& Tan, 1992).
DISCUSSION AND CONCLUSIONS
Our results show that investors tend to make
decisions in a way that sometimes contradicts the
expected utility theory. First, it was found that an
increase in the risk of the expected cash flow
enlarges the minimal grant required by investors in
order to waive tax relief, a finding that contradicts
the expected utility theory. This result can be
explained by the 'Regret Effect' (Loomes &
Sugden, 1987; Loomes, 1988), which states that
an individual tends to over value tax relief when

Copyright ? 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 601-607 (2008)
DOI: 10.1002/mde

This content downloaded from 181.65.173.6 on Tue, 24 Apr 2018 16:44:29 UTC
All use subject to http://about.jstor.org/terms
 606 M. ROSENBOIM ET AL.

the cash flow is uncertain, as there is fear of incentives that raise the attractivity of a country to
regretting the possibility of lower tax relief in case foreign investors. We suggest that by considering
the investment brings in a high cash flow. behavioral biases of decision makers, a govern
Second, it was found that investors demand ment can save costs and raise its attractivity.
compensation for changing tax relief into a grant.
This result also contradicts the expected utility
theory, because the grant is certain and is
REFERENCES
supposed to improve the investors' utility when a
tax benefit is replaced by a grant. This result can
Barros PP, Cabral L. 2000. Competing for foreign direct
be explained by the 'Status Quo Bias' (Kahneman investment. Review of International Economics 8:
et al, 1991), which claims that individuals tend to 360-371.
remain in existing situations to postpone better Benartzi S, Thaler RH. 2004. Save more tomorrow:
alternatives. using behavioral economics to increase employee
Third, it was found that decision makers in firms savings. Journal of Political Economy 112(1): 164-187.
Bjorvatn K, Eckel C. 2006. Policy competition for
are influenced by the way the incentives are foreign direct investment between asymmetric
presented. Sometimes similar incentive baskets countries. European Economic Review 50(7):
are presented in a different manner, and this might 1891-1907.
lead to different decisions. Hence, countries that Blomstrom M, Kokko A. 2003. The Economics of
Foreign Direct Investment Incentives. CEPR, February
compete for the arrival of foreign investors and 2003.
offer sets of incentives in order to draw in the
Camerer C. 1995. Individual decision making. In Hand
investors must take into consideration the ways in book of Experimental Economics, Kagel JH, Roth AE
which possible incentive baskets are presented. A (eds). Princeton University Press: Princeton, NJ;
587-703.
country can raise its level of attractiveness for
Kachelmeier SJ, Shehata M. 1992. Examining risk
foreign investment by changing the structure of the
preferences under high monetary incentives: experi
incentives basket, without changing the basket's mental evidence from the People's Republic of
cost. This behavior can be found with ordinary China. The American Economic Review 82(5): 1120?
subjects as well as with experienced subjects (such 1141.
as accountants). Kahneman D, Knetsch JL, Thaler RH. 1991. Anoma
lies: the endowment effect, loss aversion and status
We also considered the possibility that a specific
quo bias. Journal of Economic Perspectives 5: 193-206.
incentive will be activated only in specific cases. Kahneman D, Tversky A. 1979. Prospect theory: an
For example, an investor receives a grant only if analysis of decisions under risk. Econometrica 47:
he/she suffers losses. We examined what is the 263-291.
required amount of compensation to be deducted Kaplan T, Luski I, Wettstein D. 2003. Government
policy towards multinational corporations. Economics
from the grant and be received only in the case of
Bulletin 6(3): 1-8.
loss. This incentive method serves as a kind of an
Loke WH, Tan KF. 1992. Effects of framing and
'abandon option' for the investor, in case the missing information in expert and novice judgment.
investment does not live up to its expectations Bulletin of the Psychonomic Society 30: 187-190.
(insurance of losses). We found that the expected Loomes G. 1988. Further evidence of the impact of
regret and disappointment in choice under uncer
amount of insurance required by investors is lower
tainty. Economica 55(217): 47-62.
than the value of the offered incentive. These
Loomes G, Sugden R. 1987. Testing for regret
results are explained by the 'Insurance Effect' and disappointment in choice under uncertainty.
(Slovic et al, 1982), which states that as losses are Economic Journal 97, Supplement: Conference Papers
118-129.
more intimidating than costs, the individuals are
Oman C. 2000. Policy Competition for Foreign Direct
willing to pay a higher premium than the expected Investment: A Study of Competition Among Govern
damage in order to avoid the possibility of loss. A ments to Attract FDI. OECD: Paris.
government that wishes to reduce the cost of Roszkowski MJ, Snelbecker GE. 1990. Effects of
incentives can achieve this by offering an incentive 'Framing' on measures of risk tolerance: financial
with insurance characteristics. However, we must planners are not immune. The Journal of Behavioral
Economics 19(3): 237-246.
caution that this type of incentive may raise the
Slovic P, Lichtenstein S. 1971. Comparison of Bayesian
chances of moral hazards and adverse selection.
and regression approaches to the study of information
Our results can be used to improve the utilization processing in judgment. Organizational Behavior and
of government budgets by helping to plan a set of Human Performance 6: 649-744.

Copyright ? 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 601-607 (2008)
DOT. 10.1002/mde

This content downloaded from 181.65.173.6 on Tue, 24 Apr 2018 16:44:29 UTC
All use subject to http://about.jstor.org/terms
 OPTIMAL FDI INCENTIVES 607

Slovic P, Fischhoff B, Lichtenstein S. 1982. Response Taylor CT. 2000. The impact of host country govern
mode, framing, and information-processing effects in ment policy on US multinational investment deci
risk assessment. In New Directions for Methodology of sions. World Economy 23: 635-648.
Social and Behavioral Science: Question Framing and UNCTAD. 2000. Tax incentives and foreign direct
Response Consistency, Hograth R (ed.). Jossey-Bass: investment a global survey. A SIT Advisory Studies,
San Francisco; 21-36. No. 16k, United Nations: New York and Geneva.

Copyright ? 2008 John Wiley & Sons, Ltd. Manage. Decis. Econ. 29: 601-607 (2008)
DOI: 10.1002/mde

This content downloaded from 181.65.173.6 on Tue, 24 Apr 2018 16:44:29 UTC
All use subject to http://about.jstor.org/terms