OPEN ECONOMY AND FINANCIAL BURDEN OF
CORRUPTION: THEORY AND APPLICATION TO ASIA
Hrishikesh D. Vinod,
Professor of Economics and
Director of the Institute of Ethics and Economic Policy, Fordham University,
Bronx, NY 10458, E-Mail: Vinod@fordham.edu
ABSTRACT
We discuss why corruption remains high and show that corruption contributes to the Banking distress and to the rapid transmission across international stock and currency markets. Undeveloped ‘derivative securities’ markets make the risk from stress-induced volatility difficult to manage. Vinod’s (1999) closed economy model is extended to include the ‘international diversification puzzle’ and the effect of corruption on the value at risk (VaR). Our theory predicts that foreign direct investment (FDI) will be low and cost of capital will be high in corrupt countries, which is supported by Asian data. We include some policy recommendations regarding financial institutions and markets.
JEL Classification: F30, H82, O19
Key Words: Asian crisis, Value at Risk, Derivative securities, Diversification, Hedging, Generalized Pareto.
Author’s
footnote:
Professor
of Economics and Director of the Institute of Ethics and Economic Policy,
Fordham University, Bronx, NY 10458, Tel. 718-817-4065
Fax:
718-817-3518,
E-Mail:
Vinod@fordham.edu ,
1.
INTRODUCTION
Corruption
is defined as abuse of public office for private gain and it is often symptomatic
of wider governance problems. From Bardhan’s (1997) review, Vinod (1999) and
their references it is clear that corruption hurts economic development. However, typical arguments in this
literature are based on closed economy models largely ignoring open economy
aspects and both domestic and international financial sectors. This paper fills the gap. This section
covers the financial sector and the next section discusses how corruption leads
to some open economy market failures. These failures and inefficient financial
markets explain how corruption lowers foreign direct investments (FDI) and increases
cost of capital in developing countries.
Although the world of finance may seem complicated,
technical, and mysterious, this paper explains how it is hurt by corruption. Transparency
International (TI) report released on October 16, 2001 notes that the September
11 attacks in
the US clarified the high stakes in money laundering, and that fighting
corruption means fighting terrorism.
Let us first
note that corruption remains high, especially in developing countries. Even a cursory review of corruption
perception index (CPI) by TI confirms this.
The question is: Why? We suggest that it is because of missing or
inadequate consensus against corruption in three specific areas due to
conflicts of interest.
(1)
Economists define that “middle class consensus” exists in a
country when the share of income for the middle class is high and the degree of
ethnic polarization is low. Such a consensus is known to favor greater
democracy, more health, better infrastructure, and above all, greater success
in economic development. What is the role of corruption in such consensus? In corrupt countries if the income share of
the middle class is high, it may be due to bribery, money laundering, tax
evasion and corruption. We need greater education to convince the middle class
that long run costs of corruption are high.
(2) In 1999, TI started publishing Bribe Payers Perceptions Index (BPI), which
ranks 19 leading exporting countries in terms of the degree to which their
corporations are perceived to be paying bribes abroad. Many
rich countries such as Japan, which are not corrupt in terms of CPI, are shown
to be big bribe payers abroad. There is
no strong consensus in rich countries that they need to open their own markets,
especially to products in which the poorest countries have a comparative
advantage; to prosecute those who pay bribes abroad; to forego bribery-infested
export promotion, which encourages wasteful military buildups in developing
countries.
(3) The Banking privacy laws are used by corrupt entities for
money laundering. Since money laundering can be profitable to the Banks, they
have little incentive to report them and lose their business. In July 2000, the G7 organization of seven leading industrialized nations announced a
new campaign to deter money laundering. Their Financial Action Task Force
(FATF) on money laundering named fifteen countries protecting money
launderers. DJN (2001) notes that about $1 trillion per
year is laundered in an increasingly borderless world, aided by criminal
development of new and more sophisticated methods for moving money, even as
countries develop counter measures.
Exploiting vulnerabilities in the financial system is an area of rising
concern as Internet Banking transactions are quick, easy, and anonymous. Inadequate customer identification and account
monitoring procedures allow money laundering.
September 2001 terrorist attacks have brought money laundering
to the forefront and is likely to help create a new consensus against this
third item, despite the efforts of Bank lobbyists. The missing consensus
regarding the first two items arises from a conflict between self-interest and
public good. Hence, the invisible hand
of market forces is unlikely to weed out corruption. The consensus building requires better understanding of how
corruption increases the cost of capital, slows economic growth and hurts
everyone. We postpone till section 3 our evidence regarding how corruption
increases the cost of capital. The
following four subsections discuss various aspects of domestic and international
finance and corruption.
1.1
Corruption and Competition in Banking
Consider some
basic causes of corruption in the context of finance. Primitive
societies had reciprocal barter exchanges often based on mutual trust rather
than organized markets. The initial roots of corruption are found in the
informal barter. With the advent of
modern markets there was no strong need for mutual trust, which was important
for barter. Laws enforced by the
government replaced the trust. Hence,
good governance became very important for law enforcement when trust is lost.
Some countries made no investment in social capital to build trust in good
governance, and corrupt rent seeking became the norm. Ades et al’s (1999)
theory suggests that high rents need not always cause corruption. However, an
empirical study by these authors finds that high level of rents does indeed
lead to high corruption. Also, less competition means bureaucrats extract larger
rents from the monopolists, and this too is shown to lead to greater
corruption. Greater competition imposes
a better control on bureaucrats than stricter supervision. In the context of finance, foreign capital and foreign Banks can
provide the necessary competitive pressures. In the sequel we use relevant
Economics and Finance theory to show why it is useful to encourage foreign
direct investments and expansion of sophisticated markets in futures and
derivative securities to better manage all risks.
There
are three ways in which financial institutions can get hurt by corruption. First,
attacks on physical assets. We note that March 1993 terrorist attacks on Bombay
stock exchange was helped by corrupt customs officials who permitted the import
of RDX explosives in India. Also, corruption in the form of money laundering
did have a role in the September 2001 terrorist attack on physical assets at
the World Trade Center. Second, corruption can attack the goodwill of
financial institutions. Svetlana Kudryavtsev accepted $500 a month to monitor
money-laundering accounts at the Bank of New York and entered guilty plea for
it (NY Times, March 29, 2000). As a result of involvement in Russian money
laundering the Bank lost goodwill. Third, stock manipulation and fraud
through corruption hurts ordinary investors. For example, Rs 680 crore ONGC
corporate funds were credited in the notorious financier Harshad Mehta's
account (no 1028 with the UCO Bank, Hanuman Street, Mumbai) by a corrupt
official named Deosthali (Hindustan Times, Delhi, February 17, 2001). Mehta used these funds to defraud Indian
investors to the tune of nearly a billion dollars. Some lost their life savings and their example hurt the
confidence in stock markets and Banks.
Banking distress is defined in
terms of: (i) falling growth rates of real GDP, (ii) large fluctuations of
inflation rate, credit expansion rate and capital inflow rate, (iii) sharp declines
in exchange rates, (iv) adverse shock in international trade, (v) sharp rise in
real interest rate, Hardy and Pazarbasioglu (1999). Vinod (1999) notes that corruption can reduce economic growth and
discourage savings. Clearly, reduced supply of savings can lead to higher price
in the form of higher interest rates. Corruption infested governmental waste
and policies for fighting inflation can lead to large swings in inflation rate
and exchange rate. If bribe demands lead to interruption or cancellation of
major international trade transactions, they can create trade shocks. Thus, corruption contributes to each of the
five factors causing Banking distress.
How
does a localized Banking distress turn into a financial contagion affecting
large and far-flung geographical regions? The answer may be found in routine
portfolio rebalancing by investors in response to any loss or distress.
We discuss some portfolio theory regarding risk and hedging in Section 2. An
adverse shock affecting only one asset (one country) can lead international
investors to reduce their investment levels in all risky assets (all
neighboring countries) when they rebalance the portfolios. Schinasi and Smith
(2000) note that such behavior is particularly strong if the investor is
leveraged (uses borrowed funds) and that several traditional
portfolio-rebalancing rules, including VaR rules, can induce such behavior. Of
course, rebalancing remains a valid and rational response. Chang, and Velasco (1998) attribute the contagion to
international Banking illiquidity and self-fulfilling pessimism, both of which
are made worse by corruption.
IMF, World Bank and major multinational Banks can help fight the
financial contagion before it occurs. They have begun to focus on removing the
distress factors including corruption. To promote worldwide prosperity, IMF
identified “promoting good governance in all its aspects, including ensuring
the rule of law, improving the efficiency and accountability of the public
sector, and tackling corruption” (September, 1996). World Bank also has
indicated similar initiatives against corruption. More recently, IMF is trying
to eliminate the opportunity for bribery, corruption, and fraudulent activity
in the management of public resources. IMF directives seek to limit the scope
for ad hoc decision making, for rent seeking, and for undesirable
preferential treatment of individuals or organizations. However, IMF needs to
be more proactive against misuse
of official foreign exchange reserves, abuse of power by Bank supervisors, and
similar areas where information should be available at the IMF. However, unless the consensus is developed
on two fronts mentioned above, strong action by IMF or World Bank employees
against the corrupt cannot be expected. After all, these can be sensitive
issues of local laws and personal safety of World Bank/ IMF staff and their
families.
1.3
Corruption and Slow Growth of Derivative Markets.
This
subsection first notes that that very large amounts of money (A trillion US
dollars) transfer between various countries on a daily basis. These short-term funds are called “hot
money” transfers. Since these transfers create new kinds of risks, financial
institutions need new tools (e.g., interest swaps) for managing them. This subsection shows that corruption slows
the growth of these tools and thereby hurts developing countries.
A derivative is defined as a security
whose value depends on the price of some other asset. In modern finance, put and call options, financial futures,
warrants and swaps are called derivative securities. For example, consider a
manager at a large securities firm where the supply of funds is short term
while the demand for funds longer term (or vice versa), which causes a
mismatch. There is also a mismatch in
timing of the need for funds. Before
the derivatives were developed, the main sources of funds for solving the
mismatch were Bank overnight loans and commercial paper. The former tend to disappear in panic
situations and the latter cannot be easily rolled over.
Thanks to the
derivatives, the solution is simply to float say 10-year paper with fixed rate
and swap it into a floating interest rate.
Since this eliminates interest rate risk, the global market for swaps is
near five trillion dollars. Moreover,
since the interest swaps have a market value reflecting the interest
rate risk, both assets and liabilities move up and down correctly. While derivatives may contribute to greater
volatility of markets, they do create new avenues for people to take risk. For sophisticated players, they provide an
opportunity to keep cost of borrowing low.
Thinness of the market, lack of scale economies, lack of trust in local
financial institutions and corruption are among the reasons for inadequate
derivative markets in developing countries.
If Mr. Harshad Mehta could bribe Bank officials to defraud thousands of
Indian investors in standard stock market, the potential for fraudulent
manipulation loom even larger in thin derivative markets. Thus, corruption risk
is difficult to manage since corruption itself prevents access to the modern
tools of managing this risk by preventing growth of derivative markets.
Accordingly, FDI is reduced in corrupt developing countries.
The derivative markets have focused on three different types
of risk. (i) Credit risk refers to the ability of the borrower to generate
revenue to pay back the debt. (ii) Default risk is with reference to collecting
when default occurs. (iii) Transaction risk arises from possible problems with
international electronic transfer of funds and enforcement of foreign exchange
contracts. Although corruption risk is
currently considered as a part of one of these three, it may be helpful to have
a separate category for corruption. La Porta et al (1998) find that common-law countries (US,
UK, India) provide superior legal protection to investors compared to French
civil law countries. However, the CPI
data used in Vinod (1999) suggests that common-law countries providing good
legal protection of investors do not enjoy generally less corruption. France and her former colonies do not, in
general, have any greater corruption than UK and her former colonies. This supports the notion that corruption
imposes a somewhat unique type of risk burden with a distinct probability
distribution than for the risk associated with currency fluctuations, taxes and
investments.
1.4
Corruption and Private Credit Rating Agencies.
We view private rating agencies, which need to maintain their
reputation by making right calls as a positive force against corruption. Their term “country risk” refers to creditworthiness
of sovereign governments who have a large leeway in structuring debts and payment
schedules than private businesses. However, global money traders (hot money
transfer agents) gather background information from rating agencies (S&P
and Moody’s) regarding actual deficits in different countries. Since rating
agencies can potentially negate domestic fiscal policies of sovereign
countries, policy makers often resent the power of money traders and such
rating agencies. On the other hand, this creates a balance of power and forces
some fiscal discipline on various countries without involving the IMF. For example, Thailand took on debt to build
super highways and was punished for it by declining Thai Baht. In short, the
positive contribution of rating agencies and hot money is that they create a
countervailing power and they discourage corruption.
We must also mention the negative role of hot money transfers
when the investors rely on private rating agencies. A trillion dollars going in
and out of different countries daily can obviously increase the volatility of
markets. Changes in credit ratings lead
to herd behavior by investors, even if the ratings themselves are in response
to valid new information. The herd
behavior creates intrinsically self-fulfilling short-run instabilities. When
ratings go down, cost of borrowing for that country increases, precisely at the
time when a country in trouble can least afford it. Sometimes, investors lose money even before the country ratings
are lowered. If so, they try to sell investments in other similar countries
(same geographic region or similarly situated). Investors need only a limited
amount of information about impending trouble, not detailed information. This
can create a rush to beat the rating agencies. In any case, pessimism is
self-fulfilling and can lead to further declines in balance of payments.
On balance and in the long run, rating agencies play a mildly
positive role by inviting market discipline and thereby reducing corruption. If
countries do not discipline themselves, world markets do it for them through
hot money transfers. Since markets
treats everyone equally, market discipline is politically more bearable than
edicts from IMF bureaucrats, who can sometimes make wrong policy recommendations. Since IMF has little expertise in tax
deficits, corruption, and weaknesses of local Banks in all countries, IMF did
make some wrong calls in response to the Asian contagion.
The
outline of the remaining sections is as follows. Section 2 discusses two
results from an open economy theoretical model for studying corruption and
comments on the financial burden of corruption. It also outlines the
transmission mechanism by which corruption hurts. Subsection 2.1 focuses on the diversification puzzle and
subsection 2.2 focuses on value at risk computation. Section 3 studies
empirical data and policy implications.
Section 4 has concluding remarks.
2. OPEN
ECONOMY MODEL FOR BURDEN OF CORRUPTION
It is
useful to briefly mention some results of the closed economy model of corruption
before discussing market failures arising in an open economy allowing international
trade and finance. Capital accumulation is one of the most important features
of economic development. Vinod (1999)
uses the fundamental differential equation of the neoclassical “closed economy”
growth model to show that corruption reduces efficiency of capital and
exponentially reduces the rate of capital accumulation. Erlich and Lui (1999) discuss corruption
in an endogenous growth model. Using marginal excess burden calculation,
similar to deadweight loss, in the context of applied general equilibrium
model, Vinod (1999) estimated that a dollar reduction in corruption will
benefit developing countries at least $1.67. He argued that compounding raises
$1.67 even higher in just few years. Top
five actions recommended for reducing corruption in order of importance are:
(1) reduce red tape, (2) increase efficiency of the judiciary, (3) increase per
capita GNP, (4) increase economic freedom and schooling, and (5) reduce income
inequality. Although some had speculated that corruption could improve
efficiency by speeding up the bureaucrats, Vinod (1999) refers to recent theory
proving the opposite. In short, the
burden of corruption from closed economy models is already shown to be high.
The following discussion shows that open economy models, which include trade
and international commerce, make the burden even heavier. We present some “results” from
the extended theory in subsections 2.1 and 2.2, which should be regarded as supplemental
to the earlier closed economy results.
The financial burden of corruption is felt by ordinary citizen and
arises from various sources discussed in subsections 1.1 to 1.4 above.
2.1
Corruption and the Market Failure Induced by the Home Bias:
When
the presence of international trade is introduced in a traditional study of the
tradeoff between risk and return, we note that the tradeoff does not smoothly
extend to investments abroad. There
appears to be a market failure, since investors in developed countries
overwhelmingly exhibit a so-called “home bias.” This subsection explains various
factors causing the home bias or “diversification puzzle,” including
corruption. More importantly, the
impact of home bias is asymmetric with respect to rich and poor countries. The
underlying incentives encourage flight of capital out of poor and corrupt countries. Hence policy makers in poor countries impose
capital controls to prevent flight of capital, which in turn encourage
monopolies and corruption.
French
and Poterba (1991) were the first to discuss home bias and estimate its
size. They assume that a representative
agent in each country has one of the simplest exponential decay constant-relative-risk-aversion
(CRRA) utility functions. It is defined with respect to wealth W as the
argument (instead of consumption) as:
(1) U(W)= -exp(-l W/W0),
where W0
is initial wealth. The wealth increases
by investing in portfolios consisting of possibly risky international assets.
Each portfoilio is defined by a vector w of weights. If the probability distribution of asset returns is Normal, only
the mean and variance matter. Denote the mean vector by m
and covariance matrix by S. Now, the expected utility is:
(2) E(U)= -exp
(-l{wm - 0.5 lw¢S w}).
Assuming
expected utility theory (EUT), maximization of (2) yields a first order condition
involving a simple analytic formula. If w* denotes optimal portfolio weights,
we have the necessary condition:
(3) m =l w*¢S
French
and Poterba bypass the need for historical data on international equity returns
m and compute estimates of S for US, Japan, UK, France,
Germany and Canada. They ask what set of “optimal” expected equity returns m*
would justify the observed pattern of international holdings. They find, for example, that UK investors
need about 500 basis points higher return in domestic market to justify not
investing in US stocks. In general, investors in developed countries hold too
high a proportion of their portfolio in domestic securities (94% for US and in
excess of 85% for UK and Japan). French and Poterba (1991) argue that within
the set of six rich countries, there are few institutional barriers against
international investments. Yet the
“puzzle” lies in the empirically observed lack of adequate international
diversification. Baxter and Jermann
(1997) argue that the puzzle is “worse than you think.” If we remember that all
of our own human capital is concentrated in the home country, hedging would
require us to invest even larger proportion of our wealth abroad.
The
relevance of hedge component has been known at least since Merton (1973) who
shows that the conditional excess market return is a linear function of its
conditional variance (the risk component) and its covariance with investment
opportunities. We now discuss the hedge component in Merton’s model. Let J(W;
F; t) denote the indirect utility function with subscripts denoting partial
derivatives, let W denote wealth as before, and let F denote a vector of state
variables that describes investment opportunities.
(4) E
(excess return) = [-JWWW/JW ]Es2 + [-JWF/JW ]E(covariance),
where
[-JWWW/JW ]
measures the constant relative risk aversion CRRA. Clearly, the first term of
(4) represents risk and the covariance term measures the hedge component. This theory can incorporate corruption as
one of the state variables in the vector F.
How do
we solve the puzzle or explain the home bias? Let us first consider theoretical
explanations and then institutional ones. The CRRA utility function (1) used by
French and Poterba is unrealistic. In the consumption context, Carroll and
Kimball (1996) and Vinod (1998), among others, argue that CRRA utility should
be rejected, because it does not lead to a concave consumption function, as it
should. These authors propose using a
hyperbolic or diminishing absolute risk aversion (HyDARA) utility
function. Another way of stating this
issue is that CRRA is unrealistic because it gives inadequate importance to
uncertainty. We can incorporate HyDARA
in (4) by making [-JWWW/JW ] term
a declining function over time.
A
second theoretical explanation of home bias is that EUT assumed above is
unrealistic. A need for non-EUT models is explained in the survey by Starmer
(2000). Vinod (2001) considers risk aversion, stochastic dominance and non-EUT
in the context of portfolio theory.
These modifications lead to a revision of (3) involving a nonlinear
transformation of weights for greater realism. A third theoretical explanation
is that the model lumps all risk in the variance term based on the distribution
of returns. Since the currency devaluation risk applies only for foreign
investments, it can be large enough to cancel the hedging term. A fourth theoretical explanation of home
bias is that it is derived by assuming that return distribution is Normal.
Nonnormal distributions require skewness, kurtosis and higher moments which
when included can reduce the home bias.
Now we
note some institutional explanations for home bias in investments. First, investing
in home country avoids the risk associated with enforcement of property
rights in a foreign country. Even
if no language barrier exists and the property laws are essentially similar,
such as between US and UK, enforcement costs can be large. The observed difference
of 500 basis points noted by French and Poterba (1991) may be reflecting such
costs including attorney fees. Second,
corruption levels can be different in different countries at different times
for different industrial sectors. This can mean uncertainty regarding size of
bribes and financial and time costs of getting things done abroad can be
larger.
Serrat
(2001) develops a two-country exchange economy dynamic equilibrium model to
attribute the home bias to the role of nontraded goods. He shows that
risk-adjusted expected growth rates of home and foreign endowments of tradables
and nontradable goods are relevant.
Since the mean and variance of home and foreign endowments of tradables
need not be constant, hedging demands causing home bias become less important
in their dynamic setting with intertemporal substitution. In corrupt countries
the hedging motive is likely to retain the home bias if variance of home endowments
is larger.
The new
point here is not only that corruption can explain home bias, but also that
institutional explanations of the home bias do not apply symmetrically for
investors in rich and poor countries.
With lower corruption and better legal and insurance protections for
small investors in rich countries, the investors in poor countries are not as
reluctant to invest abroad, i.e., less subject to home bias.
Note
that it is also helpful to think of “international diversification puzzle” not
as something to be solved, but a pedagogical device to learn economic behavior. A similar use of six consumption puzzles is
advocated in Vinod’s (1996) appendix. Accordingly the useful insight from this
puzzle is summarized as:
Result
1:
Other
things remaining the same, all investors in different countries will be
generally better off in terms of diversification and hedging if they invest a
large proportion of their wealth in foreign assets. In practice, investors exhibit home bias. Various theoretical
generalizations of the model and institutional barriers partially explain the
home bias. However the bias against investment in poor and corrupt countries is
justifiably stronger than the bias against investing in rich and less corrupt
countries.
From
the viewpoint of the developing country, it is desirable to let as much of the
domestic savings fund domestic investments.
Accordingly, developing countries often impose controls on the outflow
of capital to prevent flight of scarce capital resources to foreign countries,
thereby helping domestic capital accumulation. Clearly, the government can
readily identify the domestic resident investor and punish any export of
capital by outlawing capital account transfers. Since there are subtle ways of
exporting capital, capital flight control can manifest itself in at least five
categories.
IMF
(2001) annual reports consider five control categories designed to discourage
flight of capital. The data from several such reports are in the form of a
“yes” or “no” answer to the following five questions: (1) Do multiple exchange
rates exist? (2) Are there payment arrears and bilateral payment arrangements?
(3) Are there controls on current transfers and payments for invisible
transactions? (4) Are there controls on proceeds from export or invisible
transactions? (5) Are there controls on capital account transactions? Appendix
I of IMF(2001) has dots for some two to thirteen parts within these five
categories. We assign weight 1 for each of the five categories so that our
index of capital controls ranges from 0 to 5 with zero indicating no controls.
If there are K subcategories within each category, we simply assign the weight
(1/K) for each and add. Table 1, which reports our index values, will be
discussed later in our empirical Section 3.
Unfortunately,
these controls themselves can encourage corruption and reduce “economic
freedom” of entrepreneurs to create wealth around the world. Often, the first
immediate effect of controls on out-going capital is to discourage in-coming
investment in the form of FDI. The
policy makers in developing countries have to consider quantitative information
on the extent of the following related issues: (1) Can the domestic saver
illegally invest abroad anyway? (2) Do these exchange controls influence the
foreign non-resident owner of capital? (3) Are the investments in the form of
equity or debt? (4) Do the controls encourage corruption and encourage
inefficient allocation of resources? (5) If domestic industries rely solely on
domestic capital, do they strive to become competitive in global marketplace?
(6) Are the foreign investors likely to be too fickle with a very short-term
focus? (7) Does the international flow of funds (hot money) destabilize the exchange
rate?
The
quantitative importance of answers to these questions will vary from country to
country. There is two-way causation
between capital flight controls and corruption. We have shown that there is a tradeoff
between capital flight controls and discouragement of FDI. Hence the optimal policy for
different countries will be different.
2.2
Corruption and the Market Failure Induced by the Value at Risk (VaR) Criterion:
This
subsection discusses the role of VaR methods in discouraging FDI in corrupt countries.
In recent Finance literature, value at risk has become a popular tool for
practical portfolio choices. By
definition, the value at risk at time t for horizon t is the upper bound of the
one-sided confidence interval:
(5) Pr[DP(t) <
-VaR] =
1-a¢,
where a¢ is the confidence level and DP(t) =DPt (t), is the relative change (return)
in the
portfolio
value over the time horizon t defined below in (6). Let t denote the current time, S(t)
denote the portfolio value at t, and denote P(t) = logS(t). If we let the
entire range of the time period be [t, T], the remaining time horizon is: t =T-t. Now we define the relative
return as:
(6) DPt (t) = P(t+t)-P(t).
The
essence of the VaR computations is estimation of low quantiles in the portfolio
return distributions. Expected
utility theory (EUT) leads to equal weight on all returns in computation of
risks. Starmer (2000), Vinod (2001) and others show that the weights are high
on extreme observations when an investor behaves according to the non-EUT. This
means that worst-case scenarios are more important to a non-EUT investor
leading to larger weight on perceived costs of corruption. In other words, more realistic non-EUT
worsens the impact of VaR calculations, by further lowering the lower limit of
the confidence interval.
2.2.1
What is the effect of the assumed distribution on VaR?
Stable
Paretian distributions are useful in modeling market and credit risks. In
general, stable distributions do not have closed form expressions for the
density and distribution functions. Stable random variables are commonly
described by their characteristic functions and by four parameters: tail index a, skewness b, location m, and scale s. Modeling with such parameters
will depict fat tails and skewness of these distributions. Estimates of α for Japan, Singapore and
Hong Kong markets are around 1.50 for almost every sum size, indicating that
these markets have high probability of large returns, Omran (2001).
A
linear combination of independent stable (or jointly stable) random variables
with stability index a is
again a stable random variable with the same a. Hence, any stable random variable can be decomposed into
the “symmetry” and “skewness” parts. If s is the volatility, X*=(z1-a¢ s) is the limiting return at a
given confidence level a¢, and
if Y0 denotes the initial value of the portfolio, then value at risk
is: VaR= -Y0X*.
Thus, modeling Value at Risk (VaR) by stable Pareto distributions permits
convenient decomposition of the distribution into three parts. The mean or
centering part, skewness part and dependence (autocorrelation) structure,
Rachev et al (2001).
Longin
and Solnik (2001) extend stable Pareto to generalized Pareto and show with examples
that cross-country equity market correlations increase in volatile times and become
particularly important in bear markets. These correlations are neither constant
over time nor symmetric with respect to the bull and bear markets. This means
that one must reject multivariate normal as well as multivariate GARCH with
time varying volatility. The estimation
of VaR is related to estimating the worst-case scenario in terms of the
probability of very large losses in excess of a threshold q. The so-called “positive q-exceedances”
correspond to all observed losses that exceed q (say 10%). Login and Solnik show that the generalized
Pareto distribution is a suitable probability distribution, and estimate its
parameters using 38 years of monthly data.
As q increases, the correlation across markets of large losses
does not converge to zero, but increases. This is ominous for an investor who seeks to diversify across
various countries, including corrupt developing countries. It means that losses in one country will not
cancel with gains in another country. Rather, all similar countries might
suffer extreme losses at the same time.
2.2.2
Do open economy international markets worsen VaR calculations?
Consider
an international investor whose portfolio consists of many assets. The computation
of VaR for any portfolio is computed by decomposing it into “building blocks,”
which depend on some risk factors. For example, currency fluctuation and corruption
are risk factors with open economy international investments. Risk professionals
first use risk categories for detailed separate analysis. They need complicated
algorithms to obtain total portfolio risk by aggregating risk factors and their
correlations.
A
well-known barrier to foreign investment is that exchange rates (currency
values) fluctuate over time and can mean a loss when the return is converted
into the investor’s home currency. The
individual investor would need to allow for potentially unfavorable timing of
currency conversion. However, financial
markets have derivative instruments including forward and future exchange rate
markets to hedge against such risks.
Hence derivative securities linked to exchange rates at a future date
can mitigate if not eliminate the exchange rate risk. Arbitrage activities by traders can be expected to price the
foreign investments appropriately different from domestic investments to take
account of exchange rate risk. However,
these hedging activities need free and open markets in target currencies. For
developing countries like India, which have exchange control, there are black
markets with a fluctuating premium over the official exchange rate. This
increases the exchange rate risk and related costs even higher due to
corruption.
This
paper suggests corruption as an additional risk factor. Corruption can suddenly
lead to a cancellation of a contract or sudden and unexpected increases in the
cost of doing business. Depending on the magnitude of corrupt practices, the
cost can vary considerably. When
we consider almost the entire world, enforcing property rights, especially in
developing countries, can be expensive and time consuming. Furthermore, the
presence of corruption and lack of transparency among public institutions add
to the cost of investing abroad, even if the foreign investor is not directly
affected by corruption. Since corruption
increases the cost of enforcement of all property rights, it is obviously an
additional burden. Assume that the
investor knows his worst-case scenario (lower limit of a 100(1-a¢)%
confidence interval) return.
Assume
that this VaR is r and
is based on all other considerations except corruption. When the cost of
corruption is considered, it is prudent to consider the worst-case scenario. Since worst-case corruption cost can be
substantial, it discourages FDI in corrupt developing countries. We may
conclude that open economy international markets do worsen VaR calculations. We summarize implications from
various models above as:
Result
2:
Assume
that investors consider worst-case scenario by methods similar to the value at
risk (VaR). Since worst-case corruption cost can be substantial, and since
cross-country correlations can be unfavorable, corruption substantially increases cost associated with
international investments and discourages investment in corrupt countries.
In
summary, the theoretical extension of closed economy corruption models to open
economy in this section predicts that investment risks are high in corrupt
countries implying a need for a higher compensation (risk premium). Hence our theory predicts that the cost of
capital is high and foreign direct investment is low in corrupt countries. Even our modest excursion into an open
economy shows that the burden of corruption suggested by closed economy models
in the literature becomes heavier under an open economy.
3. EMPIRICAL RESULTS AND POLICY IMPLICATIONS
This section
first discusses our empirical study and then mentions some policy implications.
To verify that corruption increases the cost of capital we cite
the following direct evidence. An international accounting firm, Price-Waterhouse-Coopers, PWC
(2001), surveyed chief financial officers (CFO) and PWC personnel. The survey
asks information about the percentage penalty in terms of cost of capital due
to lack of transparency and corruption in 34 countries. We summarize their results by computing a
weighted average (using midpoints of their intervals weighted by frequencies)
and report it in Table 1 for some Asian countries under the heading “cost of
capital % penalty.” For example, corruption increases the cost of capital by
over 4.4 % per year in Indonesia. When one considers the large borrowings in
billions of dollars the 4.4% becomes huge and avoidable. This is our first empirical result.
Table 1
reports some empirical data on corruption (CPI), gross domestic product (GDP),
foreign direct investments (FDI), ratio (FDI/GDP), and the ratio (Trade/GDP).
We create an index based on IMF (2001) Appendix I to measure the extent of
control on international capital transactions. In Section 2.2 subsequent to
Result 1 we mention five categories of controls and our index ranging between 0
to 5 designed to prevent the flight of capital. A column entitled “capital flow
control index” of Table 1 reports the index ranging from 0.077 for Hong Kong to
4.00 for Viet Nam. We include the following major Asian countries: Bangladesh,
Indonesia, Pakistan, Vietnam, India, Philippines, Thailand, China South Korea,
Malaysia, Taiwan, Japan, Hong Kong and Singapore. Since China threatens to not provide any data whatsoever
if Taiwan and Hong Kong data are reported separately, the data sources such as
the World Bank report that for Taiwan and Hong Kong data are “missing.” However, we can compute simple correlation
coefficients between countries after excluding missing rows.
Note
that the correlation between corruption purity CPI and (FDI/GDP) is 0.5221.
Since this is statistically significant at conventional levels (one tail, 5%),
our table supports our proposition that FDI go to purer (less corrupt)
countries. Similar correlation between CPI and (Trade/GDP) is even higher at
0.7341 and more significant. This
supports the notion that foreign trade exposes the country to foreign
competition and discourages corruption. Conversely, lower corruption creates
greater foreign trade opportunities.
The
correlation coefficient between the CPI and our “capital flow control index” is
-0.7522, with t-statistic of -3.786. The negative sign arises
because the CPI measures purity. These estimates suggest high and statistically
significant correlation, consistent with the two-way causation between
corruption and controls designed to curb the flight of capital mentioned earlier
at the end of Section 2.1.
We have
constructed a suitably weighted average of the percentage penalties in terms of
cost of capital from opinion surveys of knowledgeable persons in various countries
by Price-Waterhouse-Coopers accounting firm. It is reported in a column entitled “cost of capital %
penalty” in Table 1. It is clear that
the penalty is high for corrupt countries, namely 4.4 % for Indonesia. Note
that the penalty steadily decreases as corruption decreases. It is only 0.021 %
for Singapore. The correlation
coefficient between the CPI and this “cost of capital % penalty” is -0.827,
which is also statistically significant, as seen from the t-statistic of -3.8919
reported in Table 1.
The
open economy theory of Section 2 predicts that corruption discourages foreign
direct investments. An obvious implication from these empirical results
summarized in Table 1 is that the theoretical prediction is supported by the
data. The closed economy models have
already noted that corruption imposes serious burden on the economies of developing
countries. Open economy models suggest
an even heavier burden of corruption and imply that any policy actions designed
to reduce corruption and improve governance will yield rich long-term
dividends. The general public and
policy makers at all levels need a greater appreciation of this result.
What is
the policy implication of home bias result? The home bias is stronger for investors
in rich countries than for corruption-ridden developing countries. This means
that capital flight from developing countries is larger than incoming FDI. To the extent that investors must overcome
institutional and other impediments before achieving optimal diversification
and hedging on their own, this is an example of market failure. To correct the imbalance, investors in rich
countries may need incentives to invest in developing countries. The main point
of Result 1 is that both sets of investors can be better off with safer and
easier international investments. A
long-term solution is to develop a financially well-integrated global economy
with similar property laws for resolution of disputes and common accounting
standards. Despite the tendency to support herd behavior leading to
self-fulfilling pessimism, private independent credit and investment rating agencies need to be
expanded and encouraged, since they promote efficient allocation of global
resources. Thus, greater financial integration and transparency will benefit
everyone (Pareto superior).
During
the years of socialist dogma in India, the central planners had rightly been
afraid that the Indian rich will invest most of their capital assets in US and
UK, if permitted to do so. Despite liberalization of 1990s India still retains
controls on outflow of capital. A massive flight of capital from India can
create balance of payment problems and hurt domestic capital formation. The evidence from Asian contagion suggests
that Indian Rupee was largely immune to the volatility induced by the contagion
due to her capital controls. Mexico
does not have controls similar to India and has always faced the problem of
flight of Mexican capital to US.
Mexican Peso fluctuates more than Indian Rupee. However, there is far
greater direct investment by Americans in Mexico than in India and faster
poverty reduction in Mexico.
Empirical
evidence suggests that as capital flow controls increase foreign direct investments
decrease. We favor less controls for following reasons. Controls on capital
flows introduce bureaucratic delays, room for corruption and discourage all
foreign trade. Reduced foreign trade means less competition for domestic
suppliers leading to monopolistic practices, rent seeking and corruption. This slows
innovation and growth of world-class competitive products, needed for long-term
prosperity in a global marketplace.
Another
policy implication of our analysis is to recommend a greater role for international
agencies similar to the IMF, World Bank, and others. Blanket guarantees of invested
principal amounts granted to the international lenders have been shown to
create a moral hazard and encourage reckless investments. It is necessary,
therefore, to develop some prudent forms of encouragement for international
investing and lending. Legal infrastructure, including property laws against
abuse, confiscation, corruption and fraud need to be strengthened.
International court needs to be established to enforce property laws and
contracts. A better international allocation of savings and investments can be
possible if the cost of enforcement of property laws in diverse jurisdictions
can somehow be shared equitably, some of which should be borne by international
agencies. Partial loan guarantees and
small subsidies for foreign direct investments (FDI) in equities are also
worthy of consideration. We suggest
subsidized (not free) insurance for FDI and active help in enforcing laws
against corrupt abuses of international investors.
Foreign
aid to developing countries can be effective if adopted in the form of encouragement
for FDI. The advantages of FDI compared
to short-term loans (which comprise “hot money transfers” of international funds)
are obvious from the viewpoint of developing countries. The FDI funds are not withdrawn too readily
on a whim, and help long-term investments.
Foreign aid could be more beneficial if it is given in the form of
matching grants to encourage honest entrepreneurs from disadvantaged countries
create wealth. An effective aid program
should have accountability and explicit punishments for corruption. The foreign
aid dollars and technical assistance with modern surveillance equipment for
local law enforcement against corruption are also worth a serious consideration.
Some aid dollars should be allocated to educate the general public and policy
makers in poor countries about the serious harm from corruption.
4.
CONCLUDING REMARKS
Reviews
of corruption literature show that typical arguments use closed economy models
largely ignoring both domestic and international financial sectors and open
economy aspects. This paper fills these
gaps. Subsections 1.1 to 1.4 discuss the financial sector. Section 2 discusses
how corruption leads to some open economy market failures stated as two “results.”
Our first result essentially states that investors are better off with large
FDI in terms of diversification and hedging, although institutional barriers,
corruption and non-EUT risk perceptions discourage them. Our second result is
that value at risk (VaR) worst-case scenario method tends to
discourage foreign
direct investments in corrupt countries.
Generalized Pareto distribution is perhaps the most realistic
distribution for investment returns, and it suggests that simultaneous large
losses are possible in various countries.
Hence the worst-case scenario is that corruption can cause large simultaneous
losses to international investors.
We list many such mechanisms by which corruption can
decrease FDI. Whether it significantly
decreases FDI in Asian countries is formulated as an empirical question. We use Transparency
International’s empirical data on corruption (CPI). We construct ratios of
“total international trade” and FDI to gross domestic product (GDP) and compute
correlation coefficients with CPI. Our
analysis suggests that correlations are statistically significant, suggesting
that FDI is indeed discouraged by corruption.
We show
with a new summary measure using Price-Waterhouse-Coopers’ data that corruption increases
the cost of capital. We provide specific examples to show that corruption can
damage the physical assets of financial institutions as bribery and money laundering
facilitate terrorist attacks. Other examples show how corruption hurts bank goodwill
and creates significant losses to ordinary investors. We then discuss five
factors causing Banking distress and show how corruption can worsen each of
them. We show that in response to the distress in one country, (e.g., Thailand
in 1997) routine portfolio rebalancing by investors can lead to financial
contagion affecting several countries.
While capital flow controls can immunize a country from such contagions,
they are shown to encourage corruption and monopolies. The correlation
coefficient between a new index based on IMF data and CPI is used for this.
We
discuss the potential role of derivative markets (swaps) in helping better
management of risk associated with international investments, including corruption
risk. Encouraging competition to
domestic Banks by admitting foreign Banks and encouraging the development of
sophisticated futures and derivative markets are shown to be potentially
useful. We note the positive contribution of independent rating agencies (e.g.,
S&P and Moody’s) in instilling discipline on national economies and in
encouraging FDI. We also note their negative contribution in creating
self-fulfilling pessimism.
Worldwide corruption persists unabated, because the middle
classes in poor countries often benefit from corruption. The policy makers in
rich countries are unwilling to import goods produced in poor countries in
large quantities. Also, they do not
actively eliminate bribes designed to win contracts involving exports to poor countries,
since both employment and living standards in rich countries are linked to such
exports. We cite
Bribe Payers Perceptions Index (BPI) to show that many rich countries indulge
in harmful bribery in poor countries for export promotion.
It would take strong leadership and education to build effective consensus in
both rich and poor countries against corruption.
This
paper represents an initial step in filling the gap in the corruption
literature regarding open economy and financial institutions. The recent
terrorist attacks have focused on money laundering. An important component of any long-term solution to the terrorism
problem must involve several anticorruption efforts supporting good governance.
Our novel theoretical and empirical results and our policy recommendations
deserve further study and are subject to the usual caveats.
References:
Ades, Alberto and Rafael Di Tella (1999) Rents competition and corruption, American Economic Review, 89(4), 982–993.