Enquiry in the characteristics, theoretical impact, and empirical effectiveness of capital controls

(1)

University of Amsterdam Faculty of Economics and Business

MSc in Economics

Master's thesis

Enquiry in the characteristics, theoretical impact, and

empirical effectiveness of capital controls

Supervisor: D. Veestraeten 2nd reader: N.J. Leefmans Submitted at: 14 March 2016

Maarten Koster Student No. 5645328

(2)

1 Introduction

Liberalizing the capital account is thought to be beneficial for several reasons. In economies where capital is relatively scarce capital inflows can have a positive impact on economic growth. The additional capital can push down the interest rate and enable investment opportunities previously not feasible due to financing restraints, for example. In economies where capital is relatively abundant liberalization can also have a positive impact. Domestic residents can invest their capital in foreign firms where a higher return can be attained, for example. This can have a positive impact on domestic welfare. Also, liberalization may enable foreign firms to enter the domestic market, for instance

manufacturers via Foreign Direct Investment (FDI). This can increase competition and efficiency of domestic manufacturers with higher economic growth as a result, for example (Henry, 2006; Levine, 2001).

Nonetheless, capital controls – i.e. measures that restrict the flow of capital – are gaining acceptance, for example at the IMF. While in the 1990s the IMF expressed a negative view on capital controls, since the turn of the century several IMF supported programs included capital controls, for example in Iceland in 2008 (IMF, 2010). In addition, in 2012 the IMF proposed a framework that should guide IMF policy advice in relation to capital flows. It states that while the removal of capital controls is still a primary objective, certain circumstances may call for temporary capital controls (IMF, 2012b).

Under certain circumstances capital controls may be appropriate. For example, following the demise of three major banks in Iceland in 2008 excessive capital outflows threatened the stability of the domestic economy. There was a high risk of foreign creditors repatriating their funds. This could lead to an excessive exchange rate depreciation, causing problems for the private sector which was holding large debts denominated in foreign currencies. As part of an IMF approved economic program capital controls were implemented that aimed at stabilizing the exchange rate and thereby easing the negative effects of the problems in the banking sector (IMF, 2010; IMF, 2012a).

(4)

However, the debate on the effects of capital controls is not settled. According to Eichengreen (2001), the theoretical impact of controls on economic welfare depends on one's point of view. For example, capital controls may limit the efficient working of the financial system and have a negative impact on economic welfare. The arguments discussed in the first paragraph of this introduction in favour of liberalization correspond with this theoretical view. However, in the presence of market failures, such as

information asymmetries, capital controls may very well be a “second best” solution (Eichengreen 2001). Neither does the empirical literature provide a clear cut case. According to Versteeg (2008) different papers produce contradictory results or find no impact at all of controls on economic growth. Nonetheless, this paper shows that

differences in the characteristics of applied capital controls may result in different effects (Versteeg, 2008).

This paper takes stock of the current state of the literature on capital controls. This inquiry is divided in three parts. First, what are the dimensions on which capital controls can be differentiated? Second, what are the theoretical effects of capital restrictions? Third, how effective are capital controls at attaining their objectives?

The research question is: What are the characteristics, theoretical impact and empirical effectiveness of capital controls?

To answer this question a literature review is conducted. This review is divided in three parts. The first part reviews the dimensions on which capital controls can differ, thereby identifying characteristics and partially answering the research question. However, also the objectives of restrictions are examined. While maximising economic growth may be an objective of government policies such as capital account restrictions, there are also other (primary) objectives. This is of importance in determining the effectiveness of capital controls at a later point. In addition, the view of the IMF is examined on capital controls and in what situations they are considered an appropriate measure. Capital flows themselves are also examined. This shows that capital flows can be differentiated based on type and characteristics. Some characteristics, and thus types of flows, are more helpful in attaining objectives than others. This first part is covered in

(5)

chapter 2.

The second part reviews the theory on capital controls. What impact do capital controls (or the inverse, capital mobility) have on economic growth, among others. This second part corresponds with chapter 3. In the third part the empirical literature is

examined. How effective are capital controls in attaining their objectives? This third part is covered in chapter 4. In chapter 5 this paper is summarized and concluded.

2 Capital flows and capital controls

This chapter conducts several enquiries. First, the dimensions on which capital controls can differ are explored, thereby identifying characteristics. That in part provides a partial answer to the research question stated in the introduction. This topic is dealt with in section 2.2.1. Second, the objectives of capital controls are investigated in section 2.2.2. Understanding why capital controls are implemented is necessary to determine whether controls can be effective in attaining those objectives. Further elaborating on the

objectives of controls, in section 2.2.3 the new IMF framework is discussed that guides IMF policy advice on the usage of capital restrictions. This section provides insights in when capital controls are considered an appropriate measure by the IMF. Third, the types and characteristics of capital flows are explored. This shows that some

characteristics, and thus types of flows, are more helpful in attaining objectives than others. This is addressed in section 2.1.

2.1 Private capital flows: Types and characteristics

All flows in the capital account relate to international transactions involving financial assets and liabilities. Nonetheless, two categories of international capital flows can be distinguished, namely public and private capital flows. The former includes all flows that relate to the public sector, such as IMF credits, government bonds, and reserve transactions. The latter is the topic of this section (IMF, 2009; Combes, Kinda, and Plane, 2011).

(6)

their potential volatility most noticeably during the Asian crisis of 1997-1998. This led a number of countries to take preventive measures through capital controls restricting inflows and thereby aiming to prevent a crisis (IMF, 2011).

In the next three sections, the three major categories of private capital flows are discussed, namely FDI, 'Portfolio investment', and 'Other investment'. This provides clarity on what differentiates these categories from each other and how that may affect their potential volatility (IMF, 2009).

2.1.1 Foreign Direct Investment

According to the IMF Balance of Payments Manual (BMP6) an FDI relationship exists between a direct investor and an investment enterprise when the former can exercise significant influence over the latter and both entities reside in a different economy. This relationship arrises when the direct investor makes an investment in the investment enterprise that entitles her to voting power of at least 10 per cent. With voting power between 10 and 50 per cent the direct investor has a 'significant degree of influence' over the investment enterprise, thereafter she is said to have 'control'. This category includes all capital flows that arise within such a relationship (IMF, 2009).

FDI differentiates itself from other flows in a number of ways. FDI is often

considered a more stable capital flow. According to Chuhan, Perez-Quiros, and Popper (1996) FDI is associated with physical capital which is marked by its lack of mobility and durable nature. This is thought to enhance stability. Turner (1991) points out a number of differences between FDI and 'Portfolio investment' discussed next. First, FDI is often engaged in for a longer term. The returns are usually backloaded. For example, setting up a factory requires a large initial investment and generating returns takes some time. Second, because a FDI often involves illiquid assets, it is more difficult to sell the investment and repatriate the returns. These factors seem to increase FDI's stability (Chuhan, Perez-Quiros, and Popper, 1996; Turner, 1991).

Nonetheless, different capital flows within the category FDI can carry different interpretations. IMF (2009) distinguishes three different types of flows. There are flows

(7)

from the FDI investor to an investment enterprise, vice versa, and between fellow investment enterprises residing in different economies. The first type can constitute an investment that gives a direct investor influence, as described above. The second type of flow can be the result of a divestment. The last can be a consequence of a firm's financial structure, for example, an investment enterprise arranging the financing for the FDI investor (IMF, 2009).

As a result, IMF (1993) makes no distinction on the basis of maturity for this

category. Capital flows within this category can have different meanings. Because of the significant influence an FDI investor has on the investment enterprises the investor can have arbitrary control over the characteristics of capital flows within this relationship, e.g. long versus short term (IMF, 1993).

2.1.2 Portfolio investment

This paper adheres to the definition of 'Portfolio investment' as seen in the previous edition of BPM6, namely BPM5. In BPM6 the category what was previously 'Portfolio investment' is split in two, namely: 'Portfolio investment' and 'Financial derivatives (other than reserves) and employee stock options' (IMF, 1993; IMF, 2009).

The former concerns international transactions and positions of debt and equity securities, excluding those that are covered under FDI. Securities are marked by their negotiability and are thus tradable. This tradability reduces the significance of their maturity. For debtors that means that short term bonds can be rolled over and be used as a long term finance vehicle. Likewise for creditors. A long term bond can be sold and bought on a secondary market allowing it to be a short term investment if desired. It follows that the creditor or debtor is not necessarily limited by the (original) maturity of a specific debt instrument. As a result this category does not differentiate on the basis of maturity (IMF, 1993; IMF, 2009).

The category 'financial derivatives (other than reserves) and employee stock options' distinguishes itself from other categories because it is not concerned with the transfer of funds, other resources, or ownership but with that of risk. They exist in two broad

(8)

categories of contracts: forward-type and option. The first is an agreement to buy or sell a specific item at a predefined date and price. The second grants the purchaser the right to buy or sell a specific item at a predefined price before or at a certain date. Employee stock options, in addition, also serve as remuneration (IMF, 1993; IMF, 2009).

The general opinion seems that 'Portfolio investment' is more volatile than FDI. This is discussed in the previous section to some extent. Additional arguments come from empirical research. Turner (1991) measures the variability of different categories of capital flows. In accordance with the previous section this paper finds that the volatility of the category 'Portfolio investment' is indeed higher than that of FDI. In terms of variability, 'Portfolio investment' is only surpassed by 'Short term bank loans', discussed in the next paragraph. In addition, Turner

(1991) measures the correlation between yearly differences in the level of various categories of capital flows and in the funds required to finance current account

imbalances over the period 1975-89. Assuming that a change in the financing requirement results in a change in the

return on capital, this gives an indication of the sensitivity of capital flows to changes in their returns. The results are very similar compared to that above. This paper finds that 'Portfolio investment' is more accommodating to the external financing requirement, and thus more volatile, than FDI but less than 'Short term banking loans' (Turner, 1991).

Similarly, Graph 1 shows the differences in volatility between the three different capital flows. In the two decades that have passed since Turner (1991) was published the volatility of 'Portfolio inflows' has more than doubled. In comparison, the volatility of FDI and 'Other inflows' has remained at the same level. While the volatility of the category 'Other inflows' does increase during times of crisis, it does seem to return to its original level in stable times (IMF, 2011).

Contrary to the findings above, Claessens, Dooley, and Warner (1995) finds no

Graph 1: IMF (2011), p. 15. Volatility of capital flows: Coefficient of variation (per cent of annual GDP)

(9)

evidence of differences between categories of capital flows. This paper is a response to Turner (1991) and performs several tests to verify claims about individual categories in both developed and developing countries. Claessens, Dooley, and Warner (1995) attempts to measure volatility, persistence, and predictability of flows with the

coefficient of variation, autocorrelation, and autoregression. It finds that the flows show very little consistent behaviour between the different countries. In addition, there is some degree of substitution between flows. Changes in one type of flow can be compensated for by other flows. As a result, the volatility of the capital account can be lower than that of the individual categories and render the variability of individual flows less relevant (Claessens, Dooley, and Warner, 1995).

2.1.3 Other investment

This last category is a residual of transactions that do not fit in the other categories. For example, it includes equity that entitles the owner to less then 10 per cent voting rights and is not negotiable. This excludes it from both the categories of 'Portfolio investment' and FDI. In addition, it includes (bank) loans, trade credit and advances, and annuity and pension entitlements, among others (IMF, 2009).

Contrary to the categories discussed above maturity is considered a relevant factor. The assets included in this category are not easily tradable. Also, there does not exist an FDI relationship. Therefore, contrary to the category 'Portfolio investment', the available options to mitigate restrictions that maturity poses are limited. In addition, maturity cannot be an arbitrary decision as in the case of FDI. Consequently, the residual maturity of a loan can be important to determine the liquidity position of, for example, a financial institution (IMF, 1993).

In the papers reviewed in this chapter the category 'Other investment' is not discussed as such. Rather are its subcategories considered, short and long term flows. The

characteristics of these two types of flows diverges widely from each other. As shown earlier, Turner (1991) finds that short term flows are more volatile than 'Portfolio investment'. However, long term flows are found to be more stable than FDI. It seems

(10)

that retaining the distinction between original maturity for this category of flows is justified (Turner, 1991).

IMF (2011) is an exception and does consider the aggregate of the category 'Other investment'. In Graph 1, it shows that the coefficient of variation for the category 'Other investment' is considerably lower than that for 'Portfolio investment'. However, around times of crisis, such as during the Asian crisis at the end of the 1990s, the coefficient of variation for 'Other investment' increases by more than 100 per cent. The variability shown in this graph is likely to be lower for capital outflows. Turner (1991) finds somewhat lower numbers for outflows compared to inflows (IMF, 2011; Turner, 1991).

As shown in the previous section, Claessens, Dooley, and Warner (1995) finds no such differences between the categories. This result is confirmed by Chuhan, Perez-Quiros, and Popper (1996). Similar to Claessens, Dooley, and Warner (1995), Chuhan, Perez-Quiros, and Popper (1996) performs an autoregression. Using this method a variable is regressed on it past values to determine whether a variable is correlated with its past values, if it is persistent, and if so, to what extent? From this perspective, Chuhan, Perez-Quiros, and Popper (1996) finds no indication that one flow has different

characteristics than another albeit FDI seems to be slightly more persistent than other flows when focussing on non-G7 countries. However, comparing the interaction between flows results in a different image. Therefore, a vector autoregression is used which allows multiple variables to be estimated simultaneously using their own past values as explanatory variables. Chuhan, Perez-Quiros, and Popper (1996) focusses in particular on short term investment and FDI. This paper finds that the variance in short term investment is explained for almost three quarters by the variance of other flows while for FDI this number is at most one third. In addition, in this paper Granger causality is measured to determine whether a causal relationship exists between capital flows in different countries. A strong causal relation is found between short term

investment flows in different countries. However, for FDI such a relationship is weak. In conclusion, Chuhan, Perez-Quiros, and Popper (1996) finds that short term investment can amplify the volatility of other domestic flows and thus the capital account as a whole.

(11)

In addition this type of flow can cause financial turmoil in foreign markets to spread to the domestic market (Chuhan, Perez-Quiros, and Popper, 1996).

Overall, there seems to be evidence that the types of capital flows discussed have distinct features. 'Portfolio investment' can be considered less stable than the other types of flows. However, this category seems to have gained in volatility considerably in the past decades. This may explain why earlier research, such as Claessens, Dooley, and Warner (1995) did not find a significant difference between these types of flows. FDI on the other hand seems to be relatively stable and independent of the variation in other types of flows. 'Other investment' appears to have similar characteristics as FDI. However, in times of crisis its variability multiplies. Perhaps that can be explained by the finding of Chuhan, Perez-Quiros, and Popper (1996) that this category responds rather strong to the variability of other types of flows. In tranquil times a low variability is observed and vice versa in turbulent times.

2.2 Capital controls: Characteristics and objectives

To handle capital flows several policy tools are available. Broadly they can be divided into two categories, namely Capital Flow Management Measures (CFMs) and non-CFMs. The latter consists of measures that do not aim primarily at influencing capital flows. They are often of a structural nature, do not discriminate based on residency, and usually do not distinguish between currencies. This includes measures that strengthen financial institutions or the institutional framework, such as reserve requirements and loan-to-value ratios. Contrary to non-CFMs, CFMs intend to directly influence capital flows. They can be used to manipulate the exchange rate and/or influence the amount and/or type of capital flows entering or exiting the economy. This category can be further subdivided in 'residency based CFMs' and 'other CFMs'. The latter includes measures such as differentiated reserve requirements for deposits in foreign currencies and minimum holding periods for specific financial assets and does not discriminate on the basis of residency. The former, also known as capital controls, is the topic of this chapter (IMF, 2011).

(12)

The remainder of this chapter focusses on capital controls and thus the residency based CFM's. The dimensions on which controls differentiate and the reasons for imposing them are discussed in section 2.2.1 and 2.2.2 respectively. In section 2.2.3 the new IMF policy framework is discussed that aims to guide IMF policy advice regarding the use of capital controls (IMF, 2011).

2.2.1 Characteristics: Four dimensions

No single legal definition of capital controls exists according to Ostry et al. (2011). According to IMF (2011) capital controls are “measures affecting cross-border financial activity that discriminate on the basis of residency” (p. 40). In IMF (2010) further clarification is provided, albeit based on somewhat outdated methodology of the IMF from the 1990s. It identifies a capital control as a measure that “discriminates between domestic and international transactions, and the difference is not justified by differences in circumstances related to international transactions” (p. 28). A transaction is

international when either of the parties is a non-resident or the asset traded is a foreign capital asset. According to this definition, a measure differentiating between a similar domestic and international transaction, of which the latter entails more risk, is not necessarily a capital control. The increased risk associated with the international transaction may justify the implemented measure (IMF, 2010).

The remainder of this section explores four dimensions along which capital controls can differ.

Scope

Capitol control regimes can differ in their scope. In a comprehensive regime the capital account is effectively closed, possibly with a small list of exceptions. However,

according to Noy and Rajan (2009) it is more common to have a selective regime. Contrary to a comprehensive regime the capital account here is fairly open. Selective regimes can target specific flows. First, it can restrict transactions based on the category of capital flows. Some types of flows are thought to be more volatile than others for example. Volatile capital flows can have an adverse impact on an economy as is shown

(13)

in section 2.2.2. Categories of capital flows are discussed in more detail in chapter 2.1. Second, it can be directed at a specific sector, preventing foreign ownership of financial institutions or firms in strategic sectors for instance (Ostry et al., 2011).

Direction

Capital controls can discriminate flows on the basis of their direction. From this perspective three variants can be distinguished. First, restrictions can focus on capital inflows. This category includes legislation that requires a fraction of any capital inflow to be deposited at the central bank without remuneration for instance. This raises the cost of investment and discourages capital inflows. Second, capital control regimes can target capital outflows. A country can limit convertibility of domestic into foreign currency and thereby limit deinvestment of FDI or repatriation of profits for example. The last variant includes a regime that limits both inflows and outflows (Noy and Rajan, 2009; Reinhart and Smith, 1998).

Nonetheless, the net effects of these variants are uncertain. Ostry et al. (2010) gives an example on the effect of capital outflow controls. The introduction of restrictions on outflows can increase net inflows by limiting capital outflows while retaining the level of inflows. However, limitations may also reduce attractiveness as a destination for investment and reduce capital inflows. The final result is ambiguous. Similarly for capital controls on inflows, or controls in general. According to Quinn and Toyoda (2008) restrictions on the capital account can be a stimulus for firms to invest.

Competition in a country employing capital controls is likely to be lower compared to countries with an open capital account. This increases the payoff of such an investment in potentia and can be a stimulus to circumvent controls in order to make that investment (Ostry et al., 2010; Quinn and Toyoda, 2008).

Duration

The duration of episodes with capital controls can vary. Exceptional circumstances can be the rationale for authorities to institute temporary limitations on capital flows.

(14)

First, the authorities that maintain and enforce the restrictions can gain by the

perpetuation and expansion of controls as these can increase their power and welfare. Second, entities that have invested in a system that diverts the restrictions can gain by maintaining controls and keep an advantage over their (potential) competitors. A third argument is provided by Reinhart and Smith (2002). This paper finds that temporary controls are often left in place longer than required and that this may be due to a fear that the situation that gave rise to the restrictions might return (Dooley, 1996; Reinhart and Smith, 2002).

However, if, for example, capital controls are imposed to address a market failure in the economy or economic system permanent capital controls may (thought to) be necessary. An example of a market failure is that ever increasing the number of economies with a liberalized capital account can reduce the incentive for investors to gather country specific information. As more countries liberalize, an efficiently

diversified portfolio becomes more dispersed over economies. The incentive to collect country specific information can be reduced as each country's representation in an efficient portfolio is lower. Each country has too little weight to justify an investment in information. Under such circumstances a rumour could cause a capital outflow. The harmful effects of such an outflow are discussed in section 2.2.2. The lack of incentive to collect information could be considered a market failure and capital controls could mitigate the harmful effects by curbing capital outflows for example. This example is considered in more detail in chapter 3.2.1 (Noy and Rajan, 2009; Ostry et al., 2010; Calvo and Mendoza, 2000).

Type of restraint

The restraint placed on capital flows can be price or quantity based. With price controls the targeted transactions are taxed to reduce their profitability and thereby their

attractiveness to investors. Therefore they are also called market based controls. An example of such a targeted transaction is converting foreign currency into domestic short term debt or equity. Controls based on quantities are also called administrative controls. They allow authorities to limit certain flows or place an outright ban. For instance, firms

(15)

may require permission before exchanging domestic currency or can only acquire foreign currency to purchase foreign input materials that are unavailable domestically (Gallagher, 2011; Noy and Rajan, 2009; Ostry et al., 2011).

Quantity based controls have several disadvantages. Compared to price based

controls they lack transparency, are prone to bribery, introduce high administration costs, and stimulate black market activities. In addition, price based controls adjust more easily to the economic cycle and also generate revenues for the authorities. Therefore, in general price based controls are preferred from an economic point of view (Noy and Rajan, 2009; Ostry et al., 2011).

However, in certain situations quantity based controls can be more appropriate. This is particularly relevant where information asymmetries and uncertainties are present. For instance, informational asymmetries on the quality of assets creates uncertainty about their actual profitability and makes it difficult to set the right tax rate for a desired specific effect. A tax rate too low may produce little effect and thereby allow unfettered access to assets deemed too risky. This may lead to excessive risk taking and reduced financial stability. Another example is that uncertainty about the private sector's reaction to imposed price measures can lead to (minor) judgemental errors on the part of the authorities. A misjudgement that alters the exchange rate may have a significant effect on institutions with large open FX positions. In such a situation, a cap or an outright ban on some transactions can be a safer option (Ostry et al., 2011).

2.2.2 Objectives

In this section the objectives of capital controls are explored. This topic is addressed in three themes. First, the leeway authorities have in forming their macroeconomic policies and the role capital controls can play in them is discussed. The second theme focusses on hot money which can be a rationale for controls on both inflows and outflows. The last theme discusses sequencing reforms. Overoptimism or scepticism about reform plans, for instance, may cause excessive inflows or outflows that can undermine the reforms that gave rise to these sentiments. Capital controls during a period of transition can

(16)

provide the stability needed for changes to take place.

The impossible trinity

The 'impossible trinity' limits macroeconomic policymakers in their freedom. This principle states three policies that countries may find desirable but of which only two are mutually compatible at any time. The first 'corner' of this trinity is a stable exchange rate. In countries with less developed capital markets a higher volatility of the exchange rate can lead to lower productivity growth, according to Aghion et al. (2009). This is

rationalized with a model that assumes that wages are sticky and that the source of growth is innovation. While short term costs are constant an appreciation can reduce current earnings. An appreciation makes export products less attractive as their prices denominated in foreign currency increases. With well developed markets such a firm may borrow to keep up innovation. However, lacking those resources a firm might be forced to reduce innovation and productivity growth in anticipation of a potential

appreciation. Calvo and Reinhart (2002) provides another argument in favour of a stable exchange rate, in particular relevant for emerging markets. Often these countries have large foreign currency denominated liabilities. Debt servicing becomes increasingly difficult when the currency devaluates or depreciates (Aghion et al., 2009; Calvo and Reinhart, 2002).

The second 'corner' is monetary independence. According to Taylor (1995),

authorities can use monetary policy to influence the level of real GDP in the short run by intervening in the money market. Price rigidities allows authorities to manipulate

demand. Take for example raising the short term nominal interest rate. Due to price rigidities this also increases the real interest rate as inflation responds with a lag due to the rigidities. With the increased real interest rate the price of current consumption also increases relative to future consumption due to interest foregone. This has a negative impact on current consumption. However, for investments, such as real estate, the long term real interest rate seems more relevant. Long term interest rates are determined by the expected weighted average of short term rates in the future. To elaborate further on the previous example, by raising the expectation that the short term interest rate

(17)

increases over an extended period, the long term interest rate can be pushed in that same direction. However, as rigid prices adjust to the new money supply, the contractionary effect fades, and the economy returns to its initial level of production, or differently real GDP (Taylor, 1995).

The last 'corner' is capital mobility. The relation between capital mobility and economic growth is discussed in chapter 3 and as such is not discussed here.

Of these three 'desirable' policies only two are mutually consistent. Take monetary independence for example. An intervention in the money market affects the interest rate and can influence economic activity. However, when capital mobility is given this intervention also has an impact on the nominal exchange rate. Interest rate parity ensures that the difference between domestic and foreign interest rates is equal to the expected change in the exchange rate. A fixed exchange rate is thus not possible. As the economy returns to its fundamentals so does the exchange rate and the interest rate. This example shows that combining these three policies is mutually inconsistent. Combining monetary independence with capital mobility comes at the cost of an adaptive exchange rate. An alternative would be to restrict capital flows and keep the exchange rate stable (Taylor, 1995; Obstfeld, Shambaugh and Taylor, 2005).

Managing hot money

Hot money is a concept that refers to short term capital inflows susceptible to a reversal in their direction. While large capital inflows in general pose challenges for authorities, a potential reversal creates additional issues. According to Reinhart and Reinhart (2008) the increased demand for the domestic currency as a result of the capital inflows often causes an upward pressure on the exchange rate. As argued in the previous section an appreciation of the exchange rate can reduce the demand for export goods. These inflows can also increase domestic demand. According to Reinhart, Calvo, and Leiderman (1994) short term funds are often deposited at banks, increasing their liquidity and lending. This results in a maturity mismatch if lending goes toward long term investments. In addition, assuming a part of this lending is for non-traded goods, such as real estate, and the price of these goods is not perfectly elastic, their price will

(18)

increase. This can cause a bubble to develop. A central bank may try to mitigate these effects and sterilize the inflow. However, that also comes with a risk. Assuming the domestic interest rate is higher than the foreign rate, sterilizing capital inflows can cause a quasi-fiscal deficit according to Reinhart, Calvo and Leiderman (1994b). When

sterilizing the inflow, the central bank acquires foreign reserves, the capital inflow, in exchange for domestic currency. To sterilize the inflow the central bank sells treasury bills to reduce the domestic money supply. Since the domestic interest rate is higher than the foreign rate, the central bank makes a loss on this procedure. In addition, by

maintaining the quantity of money in circulation sterilization tends to sustain the interest rate differential that might underly the capital inflows. Overall, the fiscal cost of this procedure can be considerable (Calvo, Leiderman, and Reinhart, 1993; Reinhart and Reinhart, 2008; Reinhart, Calvo, and Leiderman, 1994b).

These episodes of short term capital inflows rarely last. A period of large capital inflows can continue for quite a while, usually between 2 and 4 years. Nonetheless, these inflows often end with a 'sudden stop'. Due to the short maturity of these (past) inflows, a sudden stop causes a reversal of capital flows. This reversal can have a damaging effect on the economy and, for example, on production and employment, according to Reinhart and Reinhart (2008). A capital outflow causes a deficit on the capital account which will have to be compensated for by an increased surplus on the current account. Since the current account is the difference between domestic production and aggregate demand, either the former has to increase or the latter has to decrease. As production is unlikely to increase in the short term, demand has to decrease. This affects both

tradables and non-tradables. While the loss in demand for the former can be

compensated by increasing exports, that is not the case for non-tradables. To restore demand for non-tradables its relative price has to fall, either by decreasing its nominal price or increasing the (domestic) price of tradables through a devaluation of the

exchange rate. However, both 'solutions' can push firms into insolvency, create a loss of output, and increase unemployment. Decreasing the prices of non-tradables lowers a firm's income while liabilities remain the same. A devaluation causes foreign currency

(19)

denominated debt levels and servicing to rise which can push firms in insolvency (Reinhart and Reinhart, 2008; Reinhart and Calvo, 2000).

Capital inflows can have external and internal drivers. The former are drivers in the external environment that make an economy relatively more attractive for investment. For instance a decline in foreign interest rates and reduced opportunities for investment in other economies. As such, these flows carry a high risk of reversal according to Reinhart, Calvo, and Leiderman (1994). If the domestic economy loses its appeal, due to an increase in foreign interest rates for example, capital might flow out. However, according to Bacchetta and Wincoop (1998) the internal environment is a more relevant factor than the external environment. Internal drivers are liberalization of the capital account and financial markets and macroeconomic stabilization among others. This paper finds that countries that have liberalized their capital account and financial sector experienced larger overshootings of their long run level of capital inflows (Reinhart, Calvo, and Leiderman, 1994b; Bacchetta and Wincoop, 1998).

Sequencing reforms

The timing of liberalization of the capital account within a reform or stabilization program can affect its success. According to Johnston and Ryan (1994) capital account liberalization should be implemented at the beginning of such a program. This paper provides a number of arguments in favour of liberalizing the capital account concurrent with the liberalization of the financial system. First, increased capital mobility can stimulate the formation of an efficient financial system. Foreign institutions can take over domestic banks and improve their performance. Similar arguments are discussed in more detail in chapter 3.1.2. Second, lifting controls on credit as part of liberalizing the financial system often leads to credit expansion and if implemented in isolation can reduce the liquidity of domestic financial institutions. Low liquidity of financial institutions increases the risk of instability as there is less room for manoeuvre. A financial setback or an above average number of depositors withdrawing their funds can render the bank illiquid. Liberalizing the capital account simultaneously allows flight capital to return and reduce that strain. Last, limitations on the capital account are often

(20)

circumvented through unofficial channels. Opening the capital account would not change actual capital flows. It would merely increase accuracy of statistical reporting on capital flows, improve interpretation thereof, and thereby result in improved policy decisions. Nonetheless, Johnston and Ryan (1994) does address some of the risks of scheduling liberalization of the capital account before reforming the financial system. For example, liberalizing the capital account may create an incentive to borrow abroad if credit controls are still in use. The risks of foreign borrowing were discussed in the previous section (Johnston and Ryan, 1994).

Lack of credibility can endanger an otherwise sound program. In case of trade

liberalization it may result in overborrowing. For example, consumers may take on loans to import goods and take advantage of 'temporary' low prices if they believe the reforms are to be reversed in the short term. This seems to have been the case at the end of the 1970s in Chile, according to McKinnon and Pill (1996). Another example, trade liberalization can render some domestic industries unprofitable. If it is believed that liberalization is reversed in short term, it may be profitable for affected firms to finance losses with debt until the tide changes. According to Edwards (1989) this happened in Argentina at around the same time. In case of a stabilization program, an open capital account can result in an outright capital flight according to Grilli and Milesi-Ferretti (1995). Both overborrowing and a capital flight can undermine a program. The former can result in an appreciation of the domestic currency harming exporters and

jeopardising the trade liberalization. Also overborrowing can attract hot money which adds a new set of problems, as discussed in the previous section. The latter can result in insolvencies, among others. Irrespective of whether the lack in credibility is justified or not these examples show how sentiment can undermine such programs (Edwards, 1989; Grilli and Milesi-Ferretti, 1995; McKinnon and Pill, 1996).

Nonetheless, also a credible program can be undermined. For example, according to Grilli and Milesi-Ferretti (1995) an inflation stabilization program often produces high interest rates. This attracts foreign capital and causes the exchange rate to appreciate, thereby harming the export sector and impairing trade liberalization, if applicable.

(21)

Another example is given by McKinnon and Pill (1996). This paper finds that high credibility of bank supervision and a reform program to liberalize the capital account can lead to overborrowing and demonstrates this mechanism with a 2 period model. In a nutshell, because banks are 'custodians of the monetary system' the authorities are bound to have either an implicit or explicit bail-out mechanism. Financial institutions have a monopoly on information about macroeconomic risk. Both households and the

authorities rely on banks for their risk assessment, which they consider credible. Because banks' tail risk is reduced (due to the bail-out mechanism), risk is estimated too low. In combination with consumption smoothing this results in too much consumption in the first period. Only in the second period when the payoff is realized do all market participants know whether the financial sector was 'well behaved', whether there was overoptimism. Overoptimism thus leads to overborrowing relative to the payoff, or differently income. According to McKinnon and Pill (1996) this is often the case (Grilli and Milesi-Ferretti, 1995; McKinnon and Pill, 1996).

These considerations can be a motivation for capital controls. However, McKinnon and Pill (1996) does not propose outright controls. Realizing that overborrowing is mostly realised with consumption rather than investment, it is impossible to differentiate capital inflows based on their destination. Rather, this paper proposes to limit consumer borrowing, for example credit cards and mortgages (McKinnon and Pill, 1996).

2.2.3 The IMF's policy stance on capital controls

In the past decade the IMF has changed its policy stance toward capital controls. The IMF's Articles of Agreement explicitly allow for capital account restrictions.

Nonetheless, it has not always been a proponent thereof. According to IMF (2010) the pro capital account liberalization view of the IMF in the 1990s was expressed through their policy advice, research, and speeches. In addition, in 1997 the Interim Committee of the IMF even proposed to extend the IMF's jurisdiction to the capital account and make liberalization thereof an important part of its activities. Later it criticized the controls imposed by Malaysia in the 1998, calling “any restrictions imposed on the movement of capital (are) not conducive to building investor confidence” (IMF, 2010;

(22)

Fischer, 2004; Kaplan and Rodrik, 2001: 10).

In the past decade that stance seems to have changed. The IMF supported programs that included controls on capital outflows in Argentina (2002) and Iceland (2008) (IMF, 2010). In 2012 the IMF proposed a policy framework that should guide IMF policy advice in relation to capital flows and capital controls (IMF, 2012).

This framework aims to provide coherence in the approach to the capital account. While trade in goods and services is bound by a globally accepted set of rules, this is not the case for capital flows. Instead, many different bilateral and multilateral agreements exist, each with their own reach. These agreements are often not designed with stability of the financial system in mind. For example, these agreements may push the pace of capital account liberalization beyond what is appropriate, as discussed in section 2.2.2 on sequencing. In addition, due to the variety of agreements, the possibility arrises that restrictions on the capital account differentiate between members of IMF. Also, this diversity in approaches frustrates research evaluating measures influencing capital flows (IMF, 2010).

The framework consists of three lines of defence against excessive capital inflows. The first line is non-CFM measures. As discussed in section 2.2 they are of a structural nature and do not intend to influence capital flows. These measures are preferred because they increase the resilience of the economy and its ability to absorb capital inflows. In addition, it ensures that macroeconomic policies are not accommodating inflows by undervaluing the exchange rate or setting the interest rate at high levels, for example. Measures in this category may reduce, or even eliminate, the need to manage capital flows (IMF, 2011).

The second line of defence is CFM measures that do not discriminate based on residency. These measures can influence the size and type of flows and have the

advantage that they can be directed at specific types of flows. However, these measures can have international repercussions. If combined with an under or overvalued exchange rate it may continue global imbalances. For example, restricting capital flows to

(23)

maintain an undervalued exchange rate may sustain a competitive advantage in trade and as a result a surplus on the current account. Also, the flows may be redirected to other countries, merely shifting the problem (IMF, 2011; Blanchard and Milesi-Ferretti, 2011).

The last line of defence are CFM measures that discriminate on the basis of residency, also known as capital controls. Members of the IMF expect a level of fairness with respect to their citizens, therefore these measures should be used as a last resort. The focus on residents versus non-residents (rather than citizenship) is based on the fact that the IMF deals itself with the balance of payments of its members which is comprised of transactions between residents and residents. These measures, including

non-discriminatory CFM measures, entail administrative costs which are likely to increase as loopholes in the CFM regulations are found (and subsequently closed by the authorities). Nonetheless, IMF (2011) finds that certain situations do legitimize measures in this category. For example, flows that are not intermediated by regulated intermediaries (regulation in the first line of defence would not have much effect on those flows) or if CFM measures can more directly target the problem at hand, among others (IMF, 2011).

The approach to problematic capital outflows is similar to the framework outlined above. The three lines of defence still apply, less discriminatory measures are preferred. However, during a crisis the first and second line of defence may have no impact at all on capital outflows, leaving capital controls as an only option. IMF (2012b) stresses that such measures should be temporary and combined with a broader set of reforms

increasing the resilience of the economy, thereby reducing the chances on crises in the future (IMF 2012b).

2.3 Chapter summary

This chapter conducted several enquiries. In section 2.1 the characteristics of three major types of private capital flows are explored, namely FDI, 'Portfolio investment' and 'Other investment'. Evidence is found that these three types of capital flows are not equal in terms of the volatility. Portfolio flows seem to be more volatile than flows in other categories. Short term flows in the category 'Other investment' tend to amplify the

(24)

behaviour of other flows. In section 2.2.1 the dimensions on which capital controls can differ are examined. Controls can differ in their scope (are all flows targeted or only a selection?), direction (are inflows, outflows, or both restricted?), duration (are controls temporary or permanent?), and restraint (are flows outrightly forbidden or taxed?). Section 2.2.2 the objectives of controls are discussed. Capital controls may have the purpose to increase monetary independence and/or increase the stability of the exchange rate. However, controls can also be deemed necessary to maintain the stability of capital flows or the economy during reform plans. In section 2.2.3 the view of the IMF on capital controls is discussed. The IMF considers controls to be justified in certain circumstances. However, only as a last resort, of a temporary nature, and part of a broader reform program. The current chapter discussed capital controls and some of their (perceived) benefits. The next chapter reviews the theoretical benefits (and costs) of capital mobility (the inverse of capital controls).

3 Theoretical perspectives on the relationship between

capital mobility and economic growth

This chapter reviews the theoretical impact of capital mobility – capital mobility being the inverse of capital controls. The results of this chapter give a partial answer to the research question posed in chapter 1. In addition, the knowledge gained in this chapter may assist in interpreting the results from the empirical literature review in the next chapter.

The next two sections approach the relationship between capital, capital mobility, and growth from two perspectives. First, the Neoclassical perspective is discussed. This includes a broad range of theories, though all with a common denominator. According to Weintraub (1993) all Neoclassical theories are based on three central assumptions:

1. “People have rational outcomes among preferences.” Simply put if an individual prefers A to B and B to C, that individual also prefers A to C.

(25)

3. “People act independently on the basis of full and relevant information.” In the second part two papers are discussed that diverge from these three principles, in particular the last, namely people acting on the basis of full information.

3.1 Neoclassical perspective

Within the Neoclassical framework economic growth can be disaggregated into three components. Growth accounting relates economic growth to changes in productivity A, capital K, and labour L. As a basis often a Cobb-Douglas production function is used to describe output in terms of these three inputs respectively: 𝑌𝑌 = 𝐴𝐴 ⋅ 𝑓𝑓(𝐾𝐾, 𝐿𝐿) = 𝐴𝐴𝐾𝐾𝛼𝛼𝐿𝐿1−𝛼𝛼. Therefore, the impact on economic growth resulting from a change in each of these three factors can be approximated by equation 3.1 (Jones, 2002: 45).

The next two sections investigate how capital mobility can affect economic growth through two of these three channels, namely capital K and productivity A. Increased capital mobility eases the import and export of capital. Therefore capital K can have a direct effect on economic growth. This is the topic of the next section. Thereafter the relationship between capital mobility and productivity is discussed. Through FDI, technology and knowledge can be transferred and have a positive effect on productivity A and thus growth for example (Varma, 2009).

3.1.1 Capital

The Neoclassical Solow growth model can provide insight in the relationship between capital and growth. The basis of this model consists of two equations. One is a

production function, for example a Cobb-Douglas production function similar to the one used above for growth accounting, 𝑌𝑌 = 𝐾𝐾𝛼𝛼(𝐸𝐸𝐿𝐿)1−𝛼𝛼. Compared to the production

function defined previously, here productivity is more narrowly defined as to include only labour productivity, denoted by E. Defining Y and K in terms of output and capital per unit of effective labour EL subsequently, this equation can be written as 𝑦𝑦� = 𝑘𝑘�𝛼𝛼. The tilde signifies their new definition. The second equation describes how capital

𝑌𝑌˙ 𝑌𝑌

=

𝐴𝐴˙ 𝐴𝐴

+ 𝛼𝛼

𝐾𝐾˙ 𝐾𝐾

+ (1 − 𝛼𝛼)

𝐿𝐿˙ 𝐿𝐿

(3.1)

(26)

accumulates over time 𝑘𝑘�̇ = 𝑠𝑠y� − (𝑛𝑛 + 𝑔𝑔 + 𝑑𝑑)𝑘𝑘�. While (invested) savings sỹ adds to the capital stock per unit of effective labour, population growth rate n, depreciation rate of the capital stock d, and the rate of growth of labour productivity g diminishes capital per effective unit of labour (Jones, 2002).

With this model the long-run steady state rate of growth can be found. This can be derived by setting the capital accumulation equation to 0 and solving it for ỹ.

Multiplying this outcome with E(t) gives output per capita y. This results in the equation 3.2. Observing that n, g, and d are constants, the steady state output per capita is growing at a rate equal to the growth rate of labour augmenting 'E', or differently at a rate 'g'. The Solow model does not predict a long term effect on economic growth per capita resulting from a permanent or temporary increase (decrease) of the savings rate, for example due to higher capital inflows (outflows) resulting from capital account liberalization (Jones, 2002).

Nonetheless, in the short run the output growth rate can diverge from this steady state. An increase in 's' causes the capital accumulation equation to diverge from its steady state zero growth outcome. More capital is saved than is required to maintain the level of capital per unit of effective labour. Therefore capital per unit of effective labour

increases. Considering the relationship between capital and output set in the production function defined above, output per unit of effective labour must also increase. Output per capita must thus temporarily grow at a higher rate than the steady state rate 'g' (Jones, 2002).

The Neoclassical framework and the Solow model allow for capital account liberalization to have an effect on capital flows. Within an economy where firms maximize profit and under perfect competition, the interest rate will equal the marginal product of capital (MPC). With a Cobb-Douglas production function as defined earlier in this section, this will be 𝑟𝑟 = 𝛼𝛼𝑌𝑌

𝐾𝐾. If capital is inhibited from flowing across borders

𝑦𝑦

*

_{= 𝐸𝐸(𝑡𝑡) �}

𝑠𝑠

𝑛𝑛+𝑔𝑔+𝑑𝑑

�

𝑎𝑎 (1−𝑎𝑎)⁄

(27)

and take advantage of differences between domestic and foreign interest rates, thereby equalizing these rates and by definition also MPC, differences may occur across countries. When liberalizing the capital account, if the world interest rate r* is lower than the domestic rate, capital will flow in until the rates are equalized, and vice versa. Therefore, the effect on capital flows within this framework is expected to be temporary (Henry, 2006; Jones, 2002).

However, capital mobility is not necessarily a requirement for factor prices to equalize within the Neoclassical framework. The Factor-Price equalization theorem provides an alternative mechanism, according to Krugman (1993). According to this theorem international trade can lead to convergence of the prices for factors of production (Krugman, 1993).

As countries start to trade goods on the world market their prices converge. Otherwise an opportunity for arbitrage would exist. Conditional on that countries are not too

dissimilar and produce the same set of goods (albeit in different quantities) and share (and use) the same technology factor prices must converge also. Under these conditions the world price for goods also defines the price of its inputs (Jones, 2008).

According to this theory trade in goods can be seen as a substitute for trade in factors of production. A country with abundant capital can produce capital intensive goods and export these to countries that are relatively less abundant in capital. Compared with a situation of autarky, the capital abundant country will have a higher demand for capital to fulfil the demand for export goods and as a result will face a higher interest rate. The country facing relative capital scarcity can import these capital intensive goods and will not have to produce these goods themselves. Compared to the situation in autarky, the country will have to produce less capital intensive goods and will therefore also face a lower demand for capital. As a result its interest rate will be lower. The end result will be that the interest rate in the two countries is equalized (Krugman and Obstfeld, 2006).

However, the assumptions underlying this theory are challenging. For example, countries may not be too dissimilar. If they differ too much, it may not be possible to

(28)

produce the same set of goods. In that case factor prices may not equalize. Also, both countries need to use the same production technology. If two countries differ

substantially and use either a labour or capital intensive technology depending on their relative abundance, this condition does not hold (Jones, 2008).

Regardless of whether liberalization affects net capital flows, be it temporarily or permanently, a change in capital flows seems to have only a relatively modest impact on growth. Capital is thought to have diminishing returns to scale. So without an

accompanying increase in the factor (effective) labour, an increase in capital yields less and less output per additional unit of capital. The impact of capital can be demonstrated by applying the aforementioned method of capital accounting (Jones, 2008).

Using capital accounting the impact of additional capital on economic growth can be estimated. Some data is required for this exercise, namely the capital share α, the capital inflow, and the capital-output ratio. The share of capital in national income α relates growth of capital to economic growth. The capital-output ratio translates the rate of capital inflows, which is denominated in economic output, to the rate it adds to the existing capital stock. Krugman (1993) uses numbers of the early 1990s of Mexico when it faced large capital inflows. These numbers are consecutively 1/3, 5, and 3. Entering these numbers into the growth accounting equation, equation 3.1, results in a mere 0,56 addition to economic growth (Krugman, 1993).

In conclusion, from a theoretical perspective the impact of capital on economic growth seems limited. While there is some ground to expect increased capital inflows as a result of capital account liberalization, even significant flows have only a modest impact.

3.1.2 Total Factor Productivity

As shown in the previous section long term growth is driven by increases in labour productivity. This is shown by equation 3.2 where labour augmenting E is the only variable present. Rewriting the Solow model to include Total Factor Productivity (TFP), denoted by A, rather than labour productivity E produces a very similar outcome. To this

(29)

aim the production function for growth accounting is used. This results in equation 3.3 where A is the level of TFP at time t. From this equation it follows that TFP sets the per capita growth rate in the long run.

Regardless of its prominent position in the Solow model not much insight is provided

in the nature of TFP. While TFP is the driver of long term growth, productivity itself is an exogenous variable not explained within this model. That is, unless time itself is counted as an explanation, as noted by Arrow (1962).

Nonetheless, in Neoclassical literature TFP is not ignored. TFP is empirically derived with the growth accounting equation, equation 3.1. After accounting for growth in output by changes in production factors capital and labour, the residual is attributed to TFP. Increasing the productivity of labour via education or capital via technological progress increases the productivity of these two factors without necessarily changing their quantity for example. However, a less obvious example is improved government regulation easing the entrance into certain markets, stimulating competition, and improving efficiency of firms in that market (Mankiw, 2007).

In the remaining part of this section three possible channels are discussed through which liberalization can influence TFP.

Diversification of portfolios

Enhanced possibilities for diversification of investment portfolios, enabled through capital account liberalization, can affect the interest rate. In the previous section it is indicated that the interest rate may change following liberalization due to (temporarily) increased net capital inflows or outflows. However, even without a change in net capital flows the interest rate can change. This is an outcome of the Neoclassical Capital Asset Pricing Model (CAPM), which is discussed in the next paragraph. In turn, this can have an effect on the composition of investment and the contribution thereof on TFP growth (Shiller, 2006; Harvey, 2001).

𝑦𝑦

*

_{= 𝐴𝐴(𝑡𝑡) �}

𝑠𝑠

𝑛𝑛+𝑑𝑑

�

𝑎𝑎 (1−𝑎𝑎)⁄

(30)

CAPM provides a framework for calculating the risk premium on a capital asset. In the context of capital assets, risk can be viewed from two angles. Assuming the return on an individual asset is normally distributed, the risk on that asset can be measured by the variance or standard deviation of its return. That is the first angle. However, the risk on a portfolio of capital assets is not their averaged risk. Assuming the return on some assets behaves differently than others under the same changing conditions, the combined portfolio risk is lower. Movements in the value of individual assets partially offset each other. From the second perspective, risk on an asset can be viewed as the variance they add to the existing portfolio rather than their individual variance. Within CAPM, this second angle is of essence (Brealey, Meyers, and Allen, 2006).

As a portfolio becomes more diversified the individual movements of an asset or a set of assets becomes irrelevant. Unique risk associated with these assets is diversified away. What is left is market risk. An example of market risk is a global financial crisis

affecting most assets negatively. The variance (or risk) of such an efficiently diversified portfolio, or market portfolio, can be approximated by the weighted sum of covariances between all individual assets (Brealey, Meyers, and Allen, 2006).

In an efficient market the interest rate is driven to its competitive level. Since diversification can reduce the risk of an asset, in a competitive market, the interest rate for such an asset reflects that reduced level of risk. Therefore, risk on such an asset is measured by its covariance with the market portfolio. This portfolio carries only the undiversifiable market risk and the covariance reflects the responsiveness of that asset to market risk. By dividing that covariance with the market portfolio its variance, beta β is produced. When β is 0, that asset is irresponsive to changes in the market. It is virtually risk free. Hence, it requires an interest rate that reflects that level of risk. An example of such an asset is a government bond. When β is 1, it moves in equal proportion with the market portfolio and deserves an interest rate equal to that of the market portfolio. According to CAPM, for any asset with a β between or above those numbers their remuneration is in linear proportion to those two benchmarks (Brealey, Meyers, and Allen, 2006).

(31)

This outcome has consequences for the interest rate in a country. In an economy with complete capital controls a portfolio cannot contain assets from abroad. As a result, geographical risk, political risk, and the domestic business cycle cannot be diversified away. Therefore that portfolio cannot be as diversified compared to a portfolio that is efficiently diversified on a global level. Undiversifiable risk, and thereby interest rate, is therefore likely to be higher without an accompanying higher expected return (Harvey, 2001).

This has also consequences for projects with high expected returns. For example, projects that require large capital investments for the development of new production techniques. Such projects are likely to be associated with high risk. Since risk cannot be properly diversified, the cost of capital for these projects is also high, perhaps

prohibitively high. By enhancing diversification through liberalization the associated risk with this project, from a global perspective, might become lower, sufficiently so to render a project profitable. This can positively influence growth in TFP and thereby GDP. By enabling high return investment, capital can be shifted away from low profit projects to growth enhancing projects (Harvey, 2001; Levine, 2001).

Foreign Direct Investment

In the Neoclassical Solow model as presented above all capital is treated equally. Capital is denoted by a cumulative K. However, as presented in chapter 2.1 different types of private capital flows can be identified, namely FDI, 'Portfolio investment', and 'Other investment'. According to Kojima (1975) not all types of capital have the same impact on the domestic economy. Some forms of capital can have externalities bringing

knowledge and novel technologies to an economy for example. This can have a positive impact on productivity and economic growth (Kojima, 1975).

Kojima (1975) differentiates between two types of capital. The first is 'money capital'. This is capital in a monetary form and is relatively free to be utilized in any sector of the economy and any production factor, such as real capital, discussed next, or labour. Capital flows in this category have an important function in equalizing interest rates between economies. The second type is 'real capital'. This category corresponds closely

(32)

to the private capital flow FDI discussed in section 2.1. 'Real capital' often has an impact on the activities and performance of firms operating in the sector the flow is targeted at (Kojima, 1975).

According to Graham and Krugman (1991) FDI is likely to have a positive influence on TFP. For a firm to invest in a foreign economy it must have a substantial advantage. By definition, domestic firms have better knowledge of and access to their 'own' market. A foreign firm must be able to offset that disadvantage and produce at a lower cost or higher quality, perhaps due to a higher productivity (as cited in Borensztein, Gregorio, and Lee, 1995).

In Adhikary (2011) several channels are discussed through which FDI can influence TFP in the Neoclassical theory (although these arguments are not exclusive to

Neoclassical theory). First, FDI can bring new technologies to the recipient country. By setting up a production facility it can bring new techniques and more efficient

production methods to that country for example. Second, new foreign owned firms or (partial) foreign takeovers bring (at least to some extent) foreign managers to manage these investments. These are likely to be experienced managers, skilled at organizing firms in an efficient and cost effective manner. Through their local employees

knowledge on these novel production techniques, methods, and managing skills can get diffused through the domestic industry improving productivity of domestically owned firms as well (Adhikary, 2011).

Stock market liquidity

Increased stock market liquidity can have a positive impact on TFP through two channels. First, improved possibilities for purchasing and selling stocks reduces the liquidity premium required as a compensation for illiquid assets. Second, these improvements also allow domestic participants to benefit from information on

discrepancies between actual and potential performances and valuations of listed firms. Increased ease of transactions on the stock market can have a beneficial impact on TFP. The reduced transaction costs translate in a lower liquidity premium and thereby

(33)

lower expected return on these assets. This allows more investments to qualify for financing on the stock market, possibly crowding out less profitable projects and increasing domestic average productivity. As in the section on diversification of portfolios above, an increase in net capital flows is not required. The larger number of stock market participants, due to the newly acquired access by foreign traders, increases the ease of transactions without necessarily expanding net capital inflows (Mishkin, 2006; Henry, 2000).

By the same token, more liquid stock markets enable participants to profit from information. In a more liquid stock market it is easier to acquire and part with assets. Therefore market participants have a greater incentive to research, purchase, and improve under performing domestic firms and thereby increase productivity (Levine, 2001).

3.2 Alternative perspectives.

In the following sections two papers are discussed that challenge part of the Neoclassical premises. Both papers challenge the third assumption on complete information. Calvo and Mendoza (2000) finds information frictions to be a 'key distinctive feature of global markets'. Bartolini and Drazen (1997a) argues that information asymmetries exist between investors and governments (Weintraub, 1993; Calvo and Mendoza, 2000; Bartolini and Drazen, 1997a).

The aim of this presentation is not to give an entire overview of the literature on asymmetric information. The aim is to give insight in how incomplete information can cause government policy to have counterintuitive results from a Neoclassical perspective 3.2.1 Diversification of portfolios

As argued above enhanced portfolio diversification can stimulate investment in more productive projects. However, Calvo and Mendoza (2000) shows that ever increasing the number of liberalized economies can also have a drawback. Calvo and Mendoza (2000) presents a model that shows that in a market with incomplete information enhanced diversification can stimulate herd behaviour in portfolio allocation decisions. Herd

Enquiry in the characteristics, theoretical impact, and empirical effectiveness of capital controls

Enquiry in the characteristics, theoretical impact, and

empirical effectiveness of capital controls

Table of Contents

1

Introduction

2

Capital flows and capital controls

3

Theoretical perspectives on the relationship between