credit growth: Differences between the financial
channel and the export channel
Master thesis
International Economics & Business
Abstract
This thesis examines the effects of a local currency appreciation on domestic credit growth. This study makes a distinction in an appreciation of the local currency against the Euro and against the US Dollar. Results show that an appreciation of the local currency has a significant effect on domestic credit growth only when it’s examined jointly with foreign liabilities or exports. Furthermore, results show that an appreciation of the local currency against the US Dollar has a more significant impact via the financial channel than via the exports channel. However, appreciations against the Euro have more impact via the exports channel.
Key words: Domestic credit growth, Appreciations, Crisis
Author: Dion Wesselink
Student number: S2374498
Supervisor: Prof. Dr. D.J. Bezemer
2
Table of contents
1. Introduction ... 3
2. Literature review ... 7
Exchange rates and the financial channel ... 10
Exchange rates and the exports channel ... 14
3. Data and methodology ... 16
4. Results ... 21
Depreciation of the exchange rate ... 24
The financial channel ... 26
Export channel ... 29
Difference in impact of both channels ... 32
5. Conclusion ... 34
Limitations ... 35
Future research ... 36
References ... 37
3 1. Introduction
In the past financial pools outside the national boundaries were hardly available. If firms or citizens were looking for capital, they would go to local capital markets (Galai et al. 2011). Outstanding loans of banks were mostly in domestic currency. Even when foreign banks entered the market in the 1990’s, most loans were still in domestic currencies. The foreign banks behaved as domestic banks and used the deposits of customers as a source of loans (Frömmel et al. 2016). After a while this behaviour changed and the foreign subsidiaries started to borrow more money from their parent banks to finance loans.
Since the 1990s, barriers are falling as a result of developments in electronic trading, exchange of information across borders and falling transaction costs (Coeurdacier et al. 2013). These falling barriers have led to a large increase in cross-border asset trade (Lane et al. 2003). More and more countries allow their citizens and firms to borrow in foreign currency (Özsöz et al. 2015). The idea behind this is to diversify the portfolio in such a way that imperfect correlated productivity or output shocks across countries have less impact (Coeurdacier et al. 2013). Credit is a large influencer of a country’s economic performance. The ratio of credit borrowed to private firms has a significant positive effect on GDP growth and thus on economic growth (King et al. 1993). Furthermore, they found a significant positive effect of the size of the financial sector and the importance of banks relative to the central bank on GDP growth. Arcand et al. (2012) and Law et al. (2014) agree that credit increases economic growth, although, they state that too much finance can be harmful. They argue that finance increases economic growth to a certain threshold. Beyond that threshold finance will have a negative impact on economic growth. They argue that more finance beyond this threshold increases the probability of potential misallocated resources and the probability of a large economic crash. Gourinchas et al. (2012) argue that prior to crises domestic credit grew rapidly together with large appreciations of the local currency.
4 Not only bank assets had increased substantial, also the amount of domestic credit in emerging European countries had increased substantial. They state that this substantial increase in domestic credit is tied to the integration process of these countries in the European Union. As a result of the substantial increase in credit, the currencies of these countries appreciated. They also state that countries that experience both, a credit boom and a currency appreciation, simultaneously are more vulnerable to a crisis. They found that a credit boom is often financed by foreign borrowing and that in the crisis of 2008 European countries financed its credit boom by short term foreign borrowing. Short term foreign borrowing is even more risky when you take the exchange rate risk into consideration.
Not only debt flows increased as a result of a more integrated world, the flow of goods and services also increased. Similar to the increased debt flows, exchange rates are an important factor in terms of trade. An appreciation of the national currency could lead to increased imports and decreased exports, which will negatively affect the trade balance. A depreciation of the national currency could lead to increased exports and decreased imports, which will affect the trade balance positively. Do et al. (2007) state that increased exports lead to higher demand for external finance, which indicates that an increase in exports increases the demand for domestic credit.
5 Figure 1: Turnover of OTC foreign exchange instruments (daily averages, in billions of US dollars). Source: BIS.org: Table D11.3.
Therefore, the exchange rate is an important factor in terms of the economic well-being of a country. An appreciation of the local currency can at the same time benefit and harm the economic performance of a country. Former research focused mostly on the relation between the local currency and the US Dollar. This paper will link the effects of an appreciation of the local currency to both the Euro and the US Dollar. This paper will try to find the effect of an appreciation of the local currency on the domestic credit growth of that country. A distinction will be made in its effect between the financial channel and the export channel. The financial channel is the influence of the exchange rate on the foreign liabilities of a country and the export channel is the influence of the exchange rate on the exports of a country. This paper will determine which channel has the most significant impact on domestic credit. Therefore, the research questions are as follows:
1. Does an appreciation of the local currency has an effect on the growth of domestic credit?
2. What is the influence of the financial channel and the export channel on this relationship?
6 the US Dollar. This thesis examines the time period of 2004 till 2016, because this is the largest coverage period available. The reasons for this is that the Euro is introduced as a means of payments in 2002 and the dependent variables in this thesis are lagged by one period. In 2002 the Euro experienced large appreciations which could bias the results. Furthermore, a distinction will be made between countries on continetnal Europe and countries that are not on continental Europe, to examine whether the effect of an appreciation of the local currency against the Euro has a larger effect on countries that are located on continental Europe.
This paper finds that an appreciation of the local currency has only a significant effect on domestic credit growth when it is combined with foreign liabilities or exports. Results show that an appreciation of the local currency against the US Dollar has a positive effect on domestic credit growth, when that country has a relatively high share of US Dollar denominated liabilities. Furthermore, this paper finds that an appreciation of the local currency against the Euro has a negative effect on domestic credit growth, when it has a high share of Euro denominated liabilities. The last result can be explained by the fact that countries have on average more US Dollar denominated liabilities than Euro denominated liabilities. Furthermore, an appreciation of the local currency of a country against the Euro has a negative effect on domestic credit growth, when it has a relatively high share of exports to GDP. This result is the same in case of an appreciation against the US Dollar. Next to that, the results show that the export channel has the largest effect on domestic credit growth when a local currency appreciates against the Euro. The financial channel has the largest effect on domestic credit growth when a local currency appreciates against the US Dollar.
7 appreciation or depreciation of the Euro or the US Dollar could also influence a country’s economy.
The structure of this paper is as follows, section 2 discusses the existing literature on domestic credit growth and the relation of the exchange rate together with the financial channel and the exports channel. Furthermore, hypothesis are formed and discussed. Section 3 describes the data collection and methodology. Section 4 discusses the results of the regression model and section 5 concludes the paper.
2. Literature review
Since the introduction of the Bretton Woods system in 1944, the US Dollar became one of the major trading currencies in the world. In the Bretton Woods system the US Dollar became the only currency that had a fixed value compared to gold, while other currencies could fluctuate by 2% upwards or downwards without the requirement to intervene (D’Arista, 2009). After the Second World War, the US was responsible for 60% of the world output and it owned 60% of the world’s gold reserves. Some governments hold no foreign exchange as reserves unless they were convertible into gold. As a result of the Second World War, Europe and Japan had not enough gold to back up their currency, so nobody accepted their currency as a means of payment. They took loans and grants from the US to build up their foreign exchange reserves. The US Dollar had the function as a unit of account. However, without the competition of other currencies that were convertible to gold, the US Dollar also became a unit of exchange and a unit of storage (D’Arista, 2009) and consequently the global leading vehicle currency.
8 As presented in Figure 1, the Euro and the US Dollar are the two most important currencies in the global financial system, especially with the increasing financial integration. Financial integration has several advantages. Countries experiencing a crisis can borrow from abroad, financial markets are becoming more efficient, risks can be dispersed and economic growth is often stimulated (Obstfeld, 1998). Claessens et al. (2001) state that financial integration leads to increased competition in the banking system and that banks have to start operating more efficiently. Not only the banking sector becomes more efficient, also the industries that experience more competition as a result of financial integration, become more efficient (Melitz, 2012). Firms can now operate in a larger market, which increases at the same time the competition and the number of potential new customers. To reach these new customers, firms have to innovate, enhance their output and become the most efficient producer in the market. These innovations are often financed with credit, which increases the demand for domestic credit.
Lane et al. (2014) states that financial integration has increased international capital flows and that as a result these capital flows, especially debt flows, increases domestic credit growth significantly. They state that banks have more access to different sources of capital such as, foreign investors, foreign depositors, international bonds and interbank money markets. Frömmel et al. (2016) found a positive correlation between international capital flows and domestic credit growth. Furthermore, Furceri et al. (2012) found that the credit-to-GDP ratio increases with 2% within 2 years after a capital inflow surge. This effect is mostly driven by debt capital inflows.
9 However, financial integration also has a few disadvantages. Firstly, bank regulation and supervision becomes much harder when countries are increasingly more financially integrated (Cetorelli et al. 2011, 2014). Secondly, Demirgüc-Kunt et al. (1998) argue that as a result of international capital flows the growth of credit in the receiving countries increased substantial. For several countries debt sustainability has become a pressing issue (Engel, 2014). Engel (2014) states that countries may be forced to reduce spending, because of their debt obligations, when they have acquired too much debt. He explains that a currency depreciation can increase the debt burden when the foreign credit is denominated in foreign currencies. A depreciation will increase the value of the debt obligation, which results in a higher debt burden. An increase in the debt burden means that citizens, firms and banks have to use more local currency to repay their debt in foreign currency. This results in lower investments, which negatively affects economic growth (Ramzan et al. 2014; Sawada. 1994). An increasing debt burden can result in a banking crisis.
This brings us to the third disadvantage of financial integration: contagion. Considerable literature has been devoted to this topic after the financial crisis of 2008. According to Cetorelli et al. (2011) a banking crisis can be transmitted directly or indirectly. Cross-border lending is the direct channel and the indirect channel is through the internal capital markets of globalized banks. Moreover, Allen et al. (2000) conducted research on this topic, they state that the overlapping claims of different banks is the dominant factor of contagion. A crisis in one region can easily be transmitted to other regions, when the banks in the crisis region are not able to repay their debt. The banks who have claims on the banks in the crisis region have a higher ratio of non-performing loans, which negatively affects their balance sheets. These negative effects of the balance sheets can pass on the crisis to the home country of these banks.
10 be disastrous. Dell’arricia et al. (2012) argue that firms cannot invest in new profitable projects or they have to reduce their output, when they have no access to finance. The decrease in lending and the decrease in investments can lead to a downward spiral, where bank problems and economic problems strengthen each other.
Calvo et al. (2013) links a crisis to the international market. They argue that in times of crisis international capital flows decrease. Also the international demand for products decreases, which decreases the amount of imports and exports. To restore the situation, multiple countries have to depreciate their currency. However, as also stated by Engel (2014) and Bruno et al. (2012) a depreciation of the local currency leads to an increased debt burden. Calvo et al. (2013) acknowledge that this could be a problem, but they also state that such a depreciation can actually help a country recover from a crisis. In a former study of Calvo (Calvo et al. 2006), they study this phenomenon. They find that as a result of a local currency depreciation, the average output of countries increases with 25%, which mitigate the crisis. When the output of firms is growing again, they are investing more, which leads to a growth in domestic credit.
The literature provides arguments for an increase and a decrease of domestic credit growth when the local currency appreciates. On the one hand, an appreciation of the local currency may lead to positive effects on domestic credit growth, when that country has a relatively high level of foreign denominated debt. On the other hand, an appreciation of the local currency may lead to negative effects on domestic credit growth, when that country has a relatively high share of exports to GDP. This leads to the following hypotheses:
H1: Local currency appreciations against the Euro have an effect on domestic credit growth
H2: Local currency appreciations against the US Dollar have an effect on domestic credit
growth
Exchange rates and the financial channel
11 Figure 2: Cross border positions per currency (source: BIS.org: Table 1A)
12 turn to foreign currencies, because they expect that these currencies will lose less value. These two consequences of currency creation partially limit the central bank in credit expansion as well. If the supply of credit in a country cannot fulfil the demand, foreign capital can offer a solution.
De Nicolo et al. (2005) and Levy-yeyati (2006) found that citizens want to save in foreign currency and banks want to provide loans that are denominated in a foreign currency when the quality of institutions is low. According to Neanidis et al. (2009) quality of institutions can be perceived as very low when there is: a relatively high level of corruption, a relatively inefficient government, political instability, relatively low regulatory quality, relatively low rule of law and relatively low accountability. The study by Neanidis et al. (2009) found that low quality of institutions is often linked to the credibility of monetary policy and they confirm the findings from De Nicolo et al (2005) and Levy-Yeyati (2006). To strengthen the quality of institutions countries have to strengthen the legal rights of creditors and the quality and accountability of the government (De Nicolo et al. 2003).
Another reason for international capital flows is the rate of return. A great influencer on the rate of return is the exchange rate (Ishioro, 2014, Claassen et al, 1997). When the local currency is likely to appreciate, citizens and firms want to borrow in foreign currency to buy local assets, because this could lead to a higher rate of return. The other way around is also an option. When the local and the foreign assets have the same rate of return, but the exchange rate is likely to depreciate, citizens want to buy foreign assets with local currency (Dullien, 2009; Ishioro, 2014). Another reason why local citizens want to hold foreign assets is because these are often more liquid. Dullien (2009) states that in dollarized countries, the local currency loses liquidity value relative to foreign assets, because unexpected costs will occur in foreign currency. He argues that the citizens of a dollarized country are not only exposed to higher exchange rate risk, but the switch to foreign assets also leads to larger depreciations of the local currency.
13 with US Dollars. When the US Dollar appreciates, the local borrower has to pay a relatively higher amount in his local currency to repay the debt that is denominated in US Dollars. Not all local borrowers can manage this, which increases the default rate of the loans set out by the banks. Bruno et al. (2015) found that, because of this, global financial conditions tighten when the US Dollar appreciates.
Figure 3: cross-border bank lending in US Dollars. From Bruno et al. (2015)
In the example of Bruno et al. (2015) local banks borrow from global banks and they lend this money to local borrowers. In the case of an appreciation of the local currency, the assets of the local borrowers with a currency mismatch increases relative to their liabilities, which increases their lending capacity. As a result, the demand for loans increase, which gives the local bank the chance to build up leverage. Bruno et al. (2015) state that appreciations of local currencies leads to greater risk taking by banks. The banks have more trust that their borrowers will not default, so they will increase their lending capacity. More debt can create economic growth, but it could also deteriorate financial stability.
14 currency depreciates, the local collateral will be worth less relative to foreign collateral, which will decrease domestic lending and thus decrease domestic credit growth.
According to Özsöz et al. (2015) the bank transforms the exchange rate risk to default risk. When banks provide loans in foreign denominated currencies, their balance sheet comprises the same assets and liabilities, but the local borrower has a currency mismatch, as Figure 3 presented. The local borrower has borrowed money in foreign currency and with that money they acquired assets in local currency. A depreciation of the local currency increases the probability of default for this local borrower, which impacts the balance sheet of the banks. In this example the local borrower did buy local assets with foreign money, but the banks themselves could also do this, which will lead to a currency mismatch on the balance sheet of the banks, which in turn will increase the exchange rate risk. Furthermore, banks that accept foreign denominated deposits from local citizens take on foreign exchange risk. Even if they provide these foreign deposits as foreign loans to domestic firms, they are still exposed to foreign exchange rates. Appreciations of the foreign currency or depreciations of the local currency can lead to higher default rates (Dullien, 2009).
The literature argues that a currency depreciation increases the relative value of foreign denominated debt, which will increase the relative value of the liabilities of the banks. Furthermore, it is stated that the chance of borrowers defaulting on loans will increase, which has a negative effect on the assets of banks. Therefore, the value of the liabilities will increase and the value of the assets will decrease, this will make banks more cautious. More cautious behaviour of banks results in lower credit supply, which will decrease the credit growth of the country. Currency appreciations have the opposing effect, which lead to an increase in credit growth. In this line of reasoning, the following hypotheses are formulated:
H3: The amount of foreign liabilities has a positive effect on the relation between
appreciation of the local currency against the Euro and domestic credit growth.
H4: The amount of foreign liabilities has a positive effect on the relation between
appreciation of the local currency against the US Dollar and domestic credit growth.
Exchange rates and the exports channel
15 of products, they are willing to import products from other countries. Countries can benefit from this, by specializing in one variety and achieve economies of scale and import the other variety. The benefit for the country is that the firm is more specialized, which increases the output, and that the consumers have more variety in products. The second benefit is shifting labour from less productive firms to relatively more productive firms. They argue that there is more competition on the world market than in the domestic market. More competition means that relatively less productive firms are pushed out of the market. The capital and the labour in those firms can then be shifted to more productive sectors. The third benefit from exports is gains from innovation. Melitz et al. (2012) argues that firms that are competing on the world market, have a higher potential number of customers than in their home country. When these firms innovate they increase their output and are able to reach more customers, leading to higher profits.
As Figure 1 presented in the introduction of this paper, international trade has grown massively in the last decade. One major influencer of international trade is the exchange rate of a country. Levy-Yeyati et al. (2001) states that there is a price stability-growth dilemma in countries. He argues that a fixed exchange rate regime can reduce inflation and reduce the volatility of the exchange rate, money growth, interest rate and GDP growth. A floating exchange rate can reduce output fluctuations and can improve growth performance during a crisis. He finds that countries with a floating exchange rate experience higher inflation, but also higher growth performance compared to a fixed exchange rate. This thesis focuses only on countries that have a floating exchange rate, therefore, higher inflation levels and also higher growth performance are expected.
16 2015). The country as a whole will increase exports and will decrease its imports, which results in a positive effect on the current account balance.
Therefore, a depreciation of the local currency leads to a higher demand for the products from that country. To fulfil this demand, firms have to invest to increase their output. Most firms don’t have enough internal capital to finance these investments, and will have to rely on external finance (Do et al. 2007). An increase in exports results in domestic credit growth. An appreciation of the local currency has the opposite effect. Local currency appreciations result in lower demand, lower output and thus less required investments. This leads to lower domestic credit growth. In this line of reasoning the following hypotheses are formed:
H5: Exports have a negative effect on the relation between appreciation of the local currency
against the Euro and domestic credit growth.
H6: Exports have a negative effect on the relation between appreciation of the local currency
against the US Dollar and domestic credit growth.
3. Data and methodology
This thesis investigates the data of the following countries: Australia, Brazil, Canada, Chile, China, Colombia, Czech Republic, Hungary, India, Indonesia, Israel, Japan, Korea, Malaysia, Mexico, New Zealand, Norway, Poland, Russia, Saudi Arabia, South Africa, Sweden, Switzerland, Turkey, Thailand, the United Kingdom and the United States. All of these countries have a floating exchange rate, which means that exchange rate fluctuations against the Euro and the US Dollar can occur. From these 27 countries, 9 are located on continental Europe. These countries have a higher share of Euro denominated liabilities than countries which are not located on continental Europe. This indicates that an appreciation of the local currency against the Euro should have a larger effect. This thesis examines the time period of 2004 till 2016, because this is the largest coverage period available. The reason for this is that the Euro is introduced as a means of payments in 2002 and the dependent variables in this thesis are lagged by one period. In 2002 the Euro experienced large appreciations which could bias the results.
17
RER= Log ( nominal exchange rate * (Euro CPI / local CPI)) (1) The formula for the real exchange rate compared to the US Dollar is as follows:
RER= Log (nominal exchange rate * (U.S. CPI / local CPI)) (2) Where CPI stands for consumer price index. The nominal exchange rate data and the CPI data are from the international financial statistics database from the IMF. In this database they weight the currency of each country to a basket of other major currencies (the US Dollar, Euro, Japanese Yen, Pound Sterling and the Chinese Renminbi) to calculate the nominal exchange rate. The nominal exchange rate is expressed as the amount of the local currency you require to buy one unit of the basket currency. Which means that an depreciation of the local currency represents an increase in the RER.
Domestic credit growth is calculated as the percentage change between the total domestic credit provided to the non-financial sector by banks in period t and the total domestic credit provided to the non-financial sector by banks in period t-1. The data for domestic credit provided to the non-financial sector by banks is from the database of the Bank of International Settlements (BIS). The variables Euro denominated liabilities and US Dollar denominated liabilities are calculated as follows: amount of Euro(US Dollar) denominated liabilities divided by the total liabilities in period t. The data for the Euro denominated liabilities, US Dollar denominated liabilities and the total liabilities is from Table A6 from the database of the BIS. The variable exports to GDP is calculated as the total amount of exports of a country divided to the GDP of that country. Where exports is the total value of goods and other market services provided to the rest of the world. The data is from the World Development Indicators database from the Worldbank.
Furthermore, control variables are added to the regression model. First, several bank factors are added. The definition and sources of these bank factors are listed in Table 10 in the Appendix. The following effects are predicted for these bank factors:
- Bank Leverage: A higher leveraged bank is more profitable and a more profitable bank is more eager to expand its credit (Ryoo, 2013).
18 - Ratio of non-performing loans: Banks face losses when they have a relatively high share of non-performing loans, which will undermine their solvency (Sorge, 2004). Lower solvency means a lower supply of credit, which will decrease the growth of domestic credit.
- Cost-to-income ratio: A higher cost-to-income ratio means that the bank generates lower profits, which means that they are able to provide fewer loans, which lowers the growth of domestic credit.
- Return on equity: This factor measures the profitability of a bank. A more profitable bank is more eager to expand its credit.
- Credit-to-deposits ratio: Banks with a relatively high credit-to-deposit ratio are less liquid and have a higher probability to default when a sudden shock occurs. Such a shock can cause a bank run, which means that many customers withdraw their money in a short time period. Banks with a relatively high ratio of credit-to-deposits are reluctant in giving out new loans, which decreases the growth of domestic credit.
Additional to the bank factors, country factors are included. The definition and sources of these bank factors are listed in Table 11 in the Appendix. The following effects are predicted for these country factors:
- Lending interest rate: A lower lending interest rate increases the demand for loans in a country, because citizens have to pay less interest.
- ∆M2: M2 is the broader definition of money supply than M1. M1 includes cash and deposits, while M2 also includes savings deposits, money market securities and mutual funds. An greater amount of M2 means that banks have more money, which increases the supply of loans.
- ∆GDP: An increase in GDP means that the economy of that country is growing, which could result in a higher demand for credit.
- Inflation: An increase in inflation means that consumers have to pay relatively more money for a product than before. As a result consumers can consume less which will decrease the economic growth of that country, if everything else stays the same. A higher rate of inflation has as a result that the demand for credit will decline.
19 result, consumers are able to spend less which in turn will decrease the economic growth of that country, which will decrease the demand for credit (Modigliani, 1961). - Financial account: A higher surplus in the financial account means that more money
is given out to other countries, this means that there is less money available in the domestic market. Which decreases the supply of credit.
GDP growth is chosen instead of GDP per capita growth, because the level of multicollinearity was too high with other independent variables if GDP per capita growth was included.
A panel regression with country fixed effects, time fixed effects and clustered standard errors at the country level will be used, to determine what the effects of the different variables are in the regression model. The following panel regression formula is used to determine the effect of the real exchange rate on domestic credit growth.
∆L(c,y)= ß0+ßi*∆RER(c,i)(y-1) + ßj*foreignliabilities(c,j)(y-1)
+ ßk*exportstoGDP(c,k)(y-1) + ßl*bankfactors(c,l)(y-1)
+ ßm*countryfactors(c,m)(y-1)+e(c,q) (3)
Where:
∆L(c,y) is the percentage growth in domestic credit in country c and in year y, as given
by the yearly difference in the domestic credit of BIS reporting country banks in country c between year y-1 and y.
∆RER(c,i)(y-1)is the change in the real exchange rate in country c and in year y-1 as
described in the data section.
∆Foreignliabilities(c,j)(y-1) is the share of Euro (US Dollar) liabilities in country c in
year y-1, as described in the data section.
ExportstoGDP(c,k)(y-1) is the share of exports to GDP in country c and in year y-1 as
described in the data section.
Bank factors l in country c encompass the bank leverage, ∆bank leverage, ratio
non-performing loans, cost-to-income ratio, return on equity ratio and credit-to-deposits ratio in year y-1, as described in the data section.
Country factors m in country c encompass the interest rate, m2, GDP, inflation,
public debt and financial account in country c in year y-1, as described in the data
20 All independent variables are lagged by one year to reduce endogeneity concerns and maximize the period coverage. The following panel regression formula is used to determine the effect of the financial channel on the relation between appreciation of the local currency and domestic credit growth.
∆L(c,y)= ß0+ßi*∆RER(c,i)(y-1) + ßj*foreignliabilities(c,j)(y-1)
+ ßk*exportstoGDP(c,k)(y-1)
+ ßl*∆RER(c,l)(y-1)*foreignliabilities(c,l)(y-1) + ßm*bankfactors(c,m)(y-1)
+ ßn*countryfactors(c,n)(y-1)+e(c,q) (4)
Where:
∆RER(c,l)(y-1)*foreignliabilities(c,l)(y-1)is the interaction factor of the amount of
liabilities in country c in year y-1 on the variable real exchange rate change in country
c in year y-1.
The following panel regression formula is used to determine what the effect of the export channel is:
∆L(c,y)= ß0+ßi*∆RER(c,i)(y-1) + ßj*foreignliabilities(c,j)(y-1)
+ ßk*exportstoGDP(c,k)(y-1)
+ ßl*∆RER(c,l)(y-1)*exportstoGDP(c,l)(y-1) + ßm*bankfactors(c,m)(y-1)
+ ßn*countryfactors(c,n)(y-1)+e(c,q) (5)
Where:
ßl*∆RER(c,l)(y-1)*exportstoGDP(c,l)(y-1) is the interaction effect of the share of exports
to GDP in country c in year y-1 on the variable real exchange rate change in country c in year y-1.
The following panel regression formula is used to determine what the effect of both the channels are and which channel has the most impact.
∆L(c,y)= ß0+ßi*∆RER(c,i)(y-1) + ßj*foreignliabilities(c,j)(y-1)
+ ßk*exportstoGDP(c,k)(y-1)
+ ßl*∆RER(c,l)(y-1)*foreignliabilities(c,l)(y-1)
+ ßm*∆RER(c,m)(y-1)*exportstoGDP(c,m)(y-1) + ßn*bankfactors(c,n)(y-1)
21 4. Results
This section will describe the results from the panel regression with country fixed effects and clustered standard errors at country level. Table 1 presents the descriptive statistics for all variables. The table includes the number of observations, the mean, the standard deviation, the maximum and the minimum. Most variables have 351 observations. However, in some databases there was no data available for all years for all countries, therefore some variables have less observations.
Table 1: Descriptive statistics
Variable N mean sd max min
∆Domestic credit 351 .1132449 .1007288 .5704066 -.0929108
∆Real exchange rate (Euro) 351 -.0030787 .0316772 .1029252 -.1235275
∆Real exchange rate (US Dollar) 338 -.0032718 .0319739 .1039588 -.1248058
Share of Euro liabilities 351 .1866651 .1376199 .7011279 .0298791
Share of US Dollar liabilities 338 .5585989 .2102262 .9541302 .0732693
Exports to GDP 350 .3787263 .2090903 1.153.731 .0903752
Bank leverage 294 .0820189 .0276087 .1488041 .035
∆Bank leverage 268 .0224884 .0992159 .5660377 -.3333333
Non-performing loans 346 .032413 .0313946 .212 .00082
Cost to income ratio 324 .5483143 .1641045 218.087 .0324779
Return on equity ratio 324 .1340447 .0972641 .7706489 -.5023155
Credit to deposits ratio 301 1.165.938 .4974552 2.997.889 .4115216
Lending interest rate 298 .0862052 .0881969 .6708333 .005
∆M2 339 .1041251 .0838011 .5782277 -.4791319
∆GDP 351 .0356609 .0302675 .1423139 -.0782088
Inflation 351 .0359165 .0321521 .2529637 -.0134672
∆Public debt 351 .0098762 .1666646 2.195.278 -.4065917
Financial account 348 3,357,493 7,544,843 532,941 -1,193,751
22 Table 2: VIF values Euro model Table 3: VIF values Dollar model
Variable VIF Variable VIF
∆Real exchange rate 1.21 ∆Real exchange rate 1.25
Share of Euro liabilities 1.99 Share of US Dollar liabilities 2.50
Exports-to-GDP 1.85 Exports-to-GDP 1.74
Bank leverage 1.57 Bank leverage 2.02
∆Bank leverage 1.15 ∆Bank leverage 1.16
Non-performing loans 1.68 Non-performing loans 1.58
Cost-to-income ratio 1.67 Cost-to-income ratio 1.57
Return on equity ratio 1.63 Return on equity ratio 1.59
Credit-to-deposits ratio 1.36 Credit-to-deposits ratio 1.29
Lending interest rate 1.61 Lending interest rate 1.64
∆M2 1.73 ∆M2 1.78
∆GDP 2.32 ∆GDP 2.29
Inflation 1.68 Inflation 1.64
Share of public debt 1.46 Share of public debt 1.47
Financial account 1.44 Financial account 1.37
Mean VIF 1.62 Mean VIF 1.66
Additional to the VIF values, a correlation table is used analyse whether there are multicollinearity issues and to examine which variables feature the highest correlations. According to Farrar et al. (1967) correlations that are smaller than 0.8 or 0.9 are acceptable. In Table 4 and Table 5 the correlation values of the different variables are presented. The dependent variable domestic credit growth is also included, to obtain an indication of the relationship between domestic credit growth and the independent variables. The first correlation table is for the Euro model and the second table is for the Dollar model. In the Euro model the correlation between ∆M2 and ∆Domestic credit has the highest value, namely 0.6343. The highest correlation between independent variables is between the variables
Lending interest rate and Inflation, with a correlation of 0.4750 This correlation is below 0.8,
∆Domestic credit 1.00
∆Real exchange rate -0.2708 1.00
Share of Euro liabilities -0.2433 0.0407 1.00
Exports-to-GDP -0.2478 0.0084 0.4266 1.00
Bank leverage 0.3813 -0.0125 -0.2150 -0.1760 1.00
∆Bank leverage -0.0962 0.1420 -0.0092 0.0875 0.0945 1.00
Non-performing loans -0.0908 0.1309 0.3616 0.2727 0.2011 0.1280 1.00
Cost-to-income ratio -0.0985 0.0009 0.2625 -0.1027 0.1308 -0.0759 0.1286 1.00
Return on equity ratio 0.4119 -0.1013 -0.1653 -0.0981 0.1049 -0.0211 -0.2834 -0.4150 1.00
Credit-to-deposits ratio 0.1541 -0.0617 -0.0517 -0.0459 -0.0690 -0.0283 -0.1384 -0.2171 0.2162 1.00
Lending interest rate 0.4034 -0.0782 -0.1498 -0.3038 0.3244 -0.0542 0.0910 0.0328 0.2090 -0.0574 1.00
∆M2 0.6472 -0.1715 -0.1389 -0.1882 0.3156 -0.0584 -0.0399 0.0233 0.3018 0.1781 0.3256 1.00
∆GDP 0.5846 -0.3270 -0.3646 -0.0528 0.0917 -0.0554 -0.0652 -0.3110 0.3163 0.3004 0.0810 0.4596 1.00
Inflation 0.4355 -0.0310 -0.0497 -0.2160 0.3298 -0.1459 0.1560 -0.0061 0.2194 -0.0334 0.4750 0.4086 0.1955 1.00
Share of public debt -0.3535 0.1960 0.1014 -0.0557 -0.1118 0.0658 0.0559 -0.0712 -0.1422 0.1578 -0.1445 -0.2783 -0.3906 -0.1949 1.00
Financial account -0.2288 0.0777 -0.0511 -0.3021 0.1064 -0.1354 -0.1266 0.1109 -0.1222 -0.1610 -0.1765 -0.0966 -0.1505 -0.1642 0.0696 1.00 Table 4: Correlation table in the Euro model
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
∆Domestic credit 1.00
∆Real exchange rate -0.2632 1.00
Share of Dollar liabilities 0.4508 -0.0212 1.00
Exports-to-GDP -0.3206 0.0273 -0.4294 1.00
Bank leverage 0.4736 -0.0134 0.5220 -0.1141 1.00
∆Bank leverage -0.0969 0.1534 0.0029 0.0940 0.0912 1.00
Non-performing loans -0.0921 0.1352 -0.1725 0.2717 0.2160 0.1190 1.00
Cost-to-income ratio -0.0902 -0.0035 -0.0343 -0.0921 0.1216 -0.0723 0.1329 1.00
Return on equity ratio 0.3973 -0.0932 0.1945 -0.1264 0.1398 -0.0130 -0.2832 -0.4144 1.00
Credit-to-deposits ratio 0.1155 -0.0516 0.1053 -0.1092 -0.0063 -0.0302 -0.1505 -0.2106 0.2032 1.00
Lending interest rate 0.3913 -0.0778 0.3729 -0.3509 0.3805 -0.0545 0.0903 0.0397 0.2005 -0.0888 1.00
∆M2 0.6427 -0.1725 0.3340 -0.2368 0.3824 -0.0577 -0.0414 0.0301 0.2960 0.1507 0.3124 1.00
∆GDP 0.5731 -0.3095 0.4015 -0.0947 0.1400 -0.0382 -0.0653 -0.3101 0.3020 0.2817 0.0646 0.4564 1.00
Inflation 0.4269 -0.0393 0.2705 -0.2601 0.3900 -0.1342 0.1582 -0.0010 0.2109 -0.0641 0.4672 0.3979 0.1789 1.00
Share of public debt -0.3422 0.1834 -0.2397 -0.0430 -0.1353 0.0515 0.0488 -0.0732 -0.1297 0.1757 -0.1381 -0.2735 -0.3813 -0.1880 1.00
24 Depreciation of the exchange rate
The results of the panel regression to determine what the effect of the real exchange rate is, are presented in Table 6. This panel regression includes country fixed effects, time fixed effects and clustered standard errors at the country level (due to limited space, the values of the time fixed effects of the regression models are moved to the appendix). The first two columns feature the results of the Euro model. In the Euro model the real exchange rate is calculated as explained in Formula 1. The variable Share of liabilities is the share of Euro liabilities to total liabilities. In the Dollar model the real exchange rate is calculated as explained in Formula 2. The variable Share of liabilities is the share of US Dollar liabilities to total liabilities.
Interestingly, the results in both models indicate that the ∆real exchange rate has no significant effect on ∆domestic credit. This means that the regression provides no evidence to support hypothesis H1, which predicted that an appreciation of the local currency against the Euro would have an effect on the growth of domestic credit, and hypothesis H2, which predicted that an appreciation of the local currency against the US Dollar would have an effect on the growth of domestic credit. These results indicate that an appreciation of the local currency does not result in an increase in domestic credit growth. Another interesting result is that the effects of the variable Share of liabilities and the variable exports to GDP are not significant. These results indicate that these variables individually do not lead to domestic credit growth. In both models the same control variables are significant and all have the predicted sign. These variables are: ratio non-performing loans, credit to deposits ratio,
25 Table 6: Regression model with as dependent variable domestic credit growth.
Euro model Dollar model
1 2
∆Real exchange rate -0.118 -0.104
(0.380) (0.453) Share of liabilities -0.0164 0.0243 (0.852) (0.741) Exports-to-GDP -0.0174 -0.0265 (0.833) (0.741) Bank leverage -0.109 -0.114 (0.806) (0.812) ∆Bank leverage -0.0123 -0.0139 (0.635) (0.623)
Ratio non-performing loans -1.033*** -1.032***
(0.000) (0.000)
Cost-to-income ratio -0.0646 -0.0678
(0.194) (0.199)
Return on equity ratio 0.0290 0.0267
(0.677) (0.698)
Credit-to-deposits ratio -0.113*** -0.108**
(0.009) (0.023)
Lending interest rate -0.303** -0.294**
(0.019) (0.033) ∆M2 0.175* 0.173* (0.083) (0.089) ∆GDP 0.602*** 0.576** (0.008) (0.012) Inflation -0.261 -0.288 (0.184) (0.149)
Share of public debt -0.0523 -0.0490
26 The financial channel
The results of the panel regression to determine the effect of the financial channel, are presented in Table 7. The first two columns feature the results of the Euro model. In the Euro model the real exchange rate is calculated as explained in Formula 1. The variable Share of
liabilities is the share of Euro liabilities to total liabilities. The variable ∆RER*Share of liabilities is thus the real exchange rate from Formula 1 * the share of Euro liabilities. The last
two columns present the results of the Dollar model. The Dollar model calculated the real exchange rate as explained in Formula 2. The variable Share of liabilities is the share of US Dollar liabilities to total liabilities. The variable ∆RER*Share of liabilities is in this case the real exchange rate from formula 2 * the share of US Dollar liabilities.
Column 1 shows that the ∆real exchange rate is negatively correlated with the factor domestic
credit growth and that this result is significant with a 10% significance level. A negative
effect means that an appreciation of the local currency against the Euro leads to an increase in domestic credit growth. In this case, an appreciation of the local currency of 1% against the Euro leads to an increase of 0.320% in domestic credit growth. Furthermore, the interaction
term has a significant positive effect on ∆domestic credit, at a significance level of 5%. This
result can be interpreted as follows: When the local currency appreciates 1% against the Euro and the country has a relatively high share of Euro liabilities then the domestic credit growth in that country will decrease by 1.320%. This result is contradictory to the hypothesis. This paper explains later why this could be the case. Additionally to these two variables the analysis shows significant results for non-performing loans, credit to deposits ratio, lending
interest rate, ∆M2 and ∆GDP. All these variables have the predicted sign.
In column 2 the same variables are included as in column 1, only now they are tested for the countries that are on continental Europe. These countries have a higher share of Euro denominated liabilities, which implies that the effect of the interaction term should be larger. The regression model shows that the negative effect of the variable ∆real exchange rate is not significant. The positive effect of the interaction term on ∆Domestic credit is still significant, although now at a significance level of 10%. However, the effect has increased from 1.320 to 5.617. This indicates that an 1% appreciation of the local currency against the Euro in a country which has a relatively high share of Euro denominated liabilities, leads to a decline in domestic credit growth of 5.617%. Furthermore, the variables non-performing loans, credit to
27 contradictory to the predictions. The effect of ∆GDP is not significant in this model. The results from this column must be interpreted with care, because the number of observations is low.
Column 3 shows that the ∆real exchange rate has a significant positive effect on ∆domestic
credit. An depreciation of 1% of the local currency against the US Dollar leads to an increase
of 0.493% in domestic credit growth. Furthermore, the interaction effect has a significant negative effect on ∆domestic credit at a significance level of 1%. This result is in line with the predictions and indicates that an appreciation of the local currency against the US Dollar, jointly with a relatively high share of US Dollar denominated liabilities, leads to an increase in domestic credit growth. In this case, an appreciation of 1% of the local currency against the US Dollar in a country with a relatively high share of US Dollar denominated liabilities leads to an 0.998% increase in domestic credit growth. Additionally to these two variables, the variables non-performing loans, credit to deposits ratio, lending interest rate and ∆M2 are significant and all have the predicted sign.
Column 4 includes the same variables as column 3, only now they are tested for countries not located on continental Europe. These countries are to a lesser extent influenced by European countries, which indicates that they have less Euro denominated liabilities and thus more US Dollar denominated liabilities. Similar to the Euro model, the effect of the variable ∆real
exchange rate is not significant and the interaction term has an increased effect on ∆domestic credit, although at a lower significance level. In this case the effect of an appreciation of 1%
of the local currency against the US Dollar with a relatively high share of US Dollar denominated liabilities leads to an increase of 1.145% in domestic credit growth. Furthermore, the model shows that the variables non-performing loans, credit-to-deposits,
lending interest rate and ∆GDP maintain significance. Additionally, similar to the Euro
28 Table 7: Regression model with as dependent variable domestic credit growth.
Euro model Dollar model
1 2 3 4
∆Real exchange rate -0.320* -1,758 0.493** 0.597
(0.092) (0.239) (0.024) (0.102) Share of liabilities -0.0344 -0.119 0.0199 -0.0345 (0.704) (0.148) (0.782) (0.652) Exports-to-GDP -0.0215 -0.449 -0.0360 -0.0544 (0.808) (0.173) (0.669) (0.512) ∆RER*share of liabilities 1.320** 5.617* -0.998*** -1.145** (0.041) (0.053) (0.010) (0.013) Bank leverage -0.0637 0.187 -0.180 -0.533 (0.884) (0.893) (0.688) (0.420) ∆Bank leverage -0.00678 -0.0543 -0.00335 -0.0163 (0.798) (0.521) (0.915) (0.629) Non-performing loans -1.069*** -1.117*** -1.069*** -0.710** (0.000) (0.005) (0.000) (0.045) Cost-to-income ratio -0.0597 -0.0136 -0.0609 0.0176 (0.220) (0.803) (0.241) (0.879)
Return on equity ratio 0.0372 -0.199*** 0.0274 0.119*
(0.588) (0.002) (0.698) (0.052)
Credit-to-deposits ratio -0.109** -0.236*** -0.104** -0.0933**
(0.011) (0.003) (0.030) (0.013)
Lending interest rate -0.321** -0.856 -0.323** -0.264**
(0.016) (0.321) (0.017) (0.043) ∆M2 0.170* 0.504*** 0.174* 0.117 (0.091) (0.000) (0.084) (0.129) ∆GDP 0.697*** 0.658 0.597*** 0.433** (0.002) (0.370) (0.006) (0.027) Inflation -0.258 0.902 -0.254 -0.330* (0.182) (0.305) (0.183) (0.062)
Share of public debt -0.0501 -0.111 -0.0428 -0.0246
(0.446) (0.593) (0.474) (0.659) Financial account -0.000107 -0.000335 -0.000159 0.000405 (0.374) (0.180) (0.195) (0.205) Constant 0.331*** 0.639*** 0.333*** 0.312*** (0.000) (0.000) (0.003) (0.000) N 190 38 180 142 Adj. R-sq 0.599 0.937 0.597 0.521
29 The results of the Dollar model are in line with the predictions of hypothesis H4. These results agree with the findings of Bruno et al. (2014) and Frömmel et al. (2016). They argued that an appreciation of a currency of a country with a relatively high share of US Dollar liabilities leads to an increase in domestic credit growth. However, the results of the Euro model are not completely in line with the predictions, since the regression model presents no evidence to support hypothesis H3. This hypothesis predicts that an appreciation of a currency of a country with a high share of Euro liabilities leads to an increase in domestic credit growth. The results of the panel regression indicate that an appreciation of the currency of a country with high shares of Euro liabilities leads to a decrease in domestic credit growth. One explanation could be that the difference between the US Dollar and the Euro in terms of acceptance as a global vehicle currency is still too considerable. Beck et al. (2008) states that most countries choose the US Dollar as their safe haven currency, despite the upcoming relevance of the Euro. This is in line with Figure 1 and Figure 2, which show that the US Dollar is a more common currency to use in trade and cross border lending than the Euro. Furthermore, there is a negative correlation between the variables Euro liabilities and US Dollar liabilities. This means that an increase in value of one variable results in a decrease in value of the other. The countries in the sample have a higher share of US Dollar liabilities than Euro liabilities on average. The descriptive statistics show that the countries in the sample have 18.67% of Euro liabilities and 55.86% of US Dollar liabilities on average.
An interesting result is that the variable ∆real exchange rate is significant in both models. In the Euro model the effect of the ∆real exchange rate is negative and in the Dollar model the effect of the ∆real exchange rate is positive. An explanation of the positive effect of the ∆real
exchange rate in the Dollar model is that the interaction effect controls for the effects of the
financial channel. The negative effect of the financial channel is filtered out of the effects of the exchange rate. This results in a positive effect, which is mainly driven by the export channel. The negative effect of the ∆real exchange rate in the Euro model is ambiguous, likely other factors which are not in the model have an effect on this variable.
Export channel
30 Column 1 shows that the ∆real exchange rate is negatively correlated with the variable
∆domestic credit and that this result is significant at a 10% significance level. A negative
effect indicates that an appreciation of the local currency against the Euro leads to an increase in domestic credit growth. In this case, an appreciation of the local currency of 1% against the Euro leads to an increase of 0.413% in domestic credit growth. Furthermore, the interaction
term has a significant positive effect on ∆domestic credit, at a significance level of 5%. This
result can be interpreted as follows: When the local currency appreciates 1% against the Euro and the country has a relatively high share of exports to GDP then the domestic credit growth in that country will decrease by 1.018%. This result is in line with the predictions. Additional to these two variables, the variables non-performing loans, credit-to-deposits ratio, lending
interest rate, ∆M2 and ∆GDP are significant. These variables all have the predicted sign. In
column 2, where only countries located on continental Europe are included, no significant effects of the main variables are examined. Again, the results from this column must be interpreted with care, because the number of observations is low.
Column 3 shows that the ∆real exchange rate has a negative effect on ∆domestic credit at a significance level of 10%. This means that an appreciation of the local currency against the Euro of 1% leads to an increase in domestic credit growth of 0.374%. Furthermore, the
interaction effect has a positive effect on ∆domestic credit at a significance level of 5%. This
means that an appreciation of a country’s currency of 1% against the US Dollar and that country has a relatively high share of exports to GDP leads to a decrease in domestic credit growth of 0.922%. This result is in line with the predictions. Additional to these two variables, the variables non-performing loans, credit-to-deposits ratio, lending interest rate,
∆M2 and ∆GDP are significant. Column 4 shows a significant negative effect of the ∆real exchange rate on ∆domestic credit, however, there is no significant effect of the interaction
term.
31 effect of the real exchange rate in both models is negative. The explanation for this is that the regression model controls for the positive effect of the export channel.
Table 8: Regression model with as dependent variable domestic credit growth.
Euro model Dollar model
1 2 3 4
∆Real exchange rate -0.413** 0.746 -0.374* -0.385**
(0.030) (0.473) (0.064) (0.038) Share of liabilities -0.0210 -0.0970 0.0343 -0.0161 (0.813) (0.492) (0.656) (0.828) Exports-to-GDP -0.0430 -0.407 -0.0492 -0.0638 (0.616) (0.519) (0.574) (0.452) ∆RER*Exports-to-GDP 1.018** -0.476 0.922** 0.895 (0.013) (0.856) (0.027) (0.123) Bank leverage -0.141 0.677 -0.145 -0.508 (0.744) (0.567) (0.755) (0.441) ∆Bank leverage 0.00137 -0.0649 -0.00118 -0.0148 (0.959) (0.743) (0.968) (0.644) Non-performing loans -1.105*** -1.572*** -1.099*** -0.772** (0.000) (0.000) (0.000) (0.025) Cost-to-income ratio -0.0537 -0.0223 -0.0564 0.0279 (0.282) (0.848) (0.284) (0.824)
Return on equity ratio 0.0233 -0.209*** 0.0209 0.119
(0.729) (0.009) (0.758) (0.109)
Credit-to-deposits ratio -0.116*** -0.302** -0.110** -0.0979***
(0.009) (0.023) (0.025) (0.009)
Lending interest rate -0.326** -0.402 -0.315** -0.257*
(0.011) (0.772) (0.021) (0.056) ∆M2 0.171* 0.362* 0.169* 0.104 (0.086) (0.081) (0.092) (0.161) ∆GDP 0.670*** -0.530 0.632*** 0.566** (0.003) (0.378) (0.006) (0.011) Inflation -0.175 0.165 -0.206 -0.293* (0.388) (0.518) (0.318) (0.063)
Share of public debt -0.0490 -0.369 -0.0454 -0.0293
(0.438) (0.258) (0.464) (0.618) Financial account -0.000122 -0.000416 -0.000176 0.000411 (0.320) (0.259) (0.172) (0.254) Constant 0.343*** 0.727*** 0.322*** 0.289*** (0.000) (0.006) (0.005) (0.001) N 190 38 180 142 Adj. R-sq 0.600 0.884 0.592 0.514
32 Difference in impact of both channels
The results of the panel regression to determine which channel has the most considerable impact on an appreciation of the local currency against the Euro and which channel has the most impact on an appreciation of the local currency against the US Dollar, are presented in Table 9. The table is divided into two columns, the first column features the Euro model and the second column presents the Dollar model.
Column 1 shows a negative effect of the ∆real exchange rate on ∆domestic credit, at a significance level of 5%. This result indicates that an appreciation of the local currency of 1% against the Euro leads to an increase of 0.476% in domestic credit growth. Furthermore, there is no significant effect for the interaction effect real exchange rate * Euro liabilities, but there is a significant effect for the interaction term real exchange rate * exports on ∆domestic
credit. The interaction term real exchange rate * exports has a positive effect on ∆domestic credit at a significance level of 10%. The results in this model indicate that an appreciation of
a local currency against the Euro has a more significant effect on domestic credit growth when the country has a relatively high share of exports to GDP than when the country has a relatively high share of Euro liabilities. Column 2 shows only a significant effect of the
interaction term real exchange rate * US Dollar liabilities on ∆domestic credit. There is no
significant effect of the interaction term real exchange rate * exports on ∆domestic credit. The results in this model indicate that an appreciation of the local currency against the US Dollar has a more significant effect on domestic credit growth when that country has a relatively high share of US Dollar liabilities than a relatively high share of exports to GDP.
33 Table 9: Regression model with as dependent variable domestic credit growth.
Euro model Dollar model
1 2
∆Real exchange rate -0.476** 0.248
(0.019) (0.380) Share of liabilities -0.0318 0.0260 (0.728) (0.724) Exports-to-GDP -0.0395 -0.0465 (0.650) (0.592) ∆RER*share of liabilities 0.873 -0.830** (0.269) (0.047) ∆RER*Exports to GDP 0.771* 0.493 (0.092) (0.214) Bank leverage -0.103 -0.186 (0.813) (0.678) ∆Bank leverage 0.00172 0.00168 (0.950) (0.957)
Ratio non-performing loans -1.111*** -1.099***
(0.000) (0.000)
Cost-to-income ratio -0.0531 -0.0559
(0.284) (0.291)
Return on equity ratio 0.0302 0.0242
(0.656) (0.730)
Credit-to-deposits ratio -0.113** -0.106**
(0.011) (0.031)
Lending interest rate -0.332** -0.329**
(0.012) (0.016) ∆M2 0.168* 0.172* (0.090) (0.087) ∆GDP 0.716*** 0.623*** (0.001) (0.005) Inflation -0.194 -0.216 (0.341) (0.291)
Share of public debt -0.0483 -0.0419
(0.453) (0.483) Financial account -0.000114 -0.000163 (0.347) (0.189) Constant 0.338*** 0.330*** (0.000) (0.004) N 190 180 adj. R-sq 0.600 0.597
34 The results show that the economic risks of an appreciation of the local currency against the Euro are less severe. The results indicate that in case of an appreciation of the local currency against the Euro, the export channel has a more pronounced effect than the financial channel. This means that an appreciation of the local currency against the Euro leads to lower domestic credit growth, if that country has a relatively high share of exports to GDP. The demand for products of the country decreases, which decreases the output of the exporting firms. Consequently, these firms do not have to innovate, which decreases the domestic credit growth.
5. Conclusion
As a result of a more integrated world, international borrowing increased massively and this has a positive effect on domestic credit growth. Domestic credit growth is an important subject for research. Not only is domestic credit growth important for economic growth (Ernst et al. 2016; King et al. 1993; Goldsmith, 1969), it is also a robust predictor for a financial crisis (Bridges et al. 2017; Frömmel et al. 2016; Gourinchas et al. 2012). They state that countries that experience both a credit boom and a currency appreciation at the same time are more vulnerable to a crisis. There are two channels that jointly with currency appreciations affect the growth of domestic credit. The financial channel increases domestic credit growth when the local currency appreciates and the export channel decreases domestic credit growth when the local currency appreciates. Most of the literature estimates the impact of one channel or it is about appreciations against the US Dollar (Bruno et al. 2014; Frömmel et al. 2016).
This paper used a panel regression with country fixed effects, time fixed effects and clustered standard errors to answer the following research questions:
1. Does an appreciation of the local currency has an effect on the growth of domestic credit?
2. What is the influence of the financial channel and the export channel on this relationship?
35 domestic credit growth only when it is examined jointly with foreign liabilities or exports. Results show that an appreciation of the local currency of a country against the US Dollar increases domestic credit growth, when that country has a relatively high share of US Dollar liabilities. Furthermore, this paper finds that an appreciation of the local currency against the Euro decreases domestic credit growth, when that country has a relatively high share of Euro liabilities. This result can be explained by the fact that countries have more US Dollar liabilities than Euro liabilities on average. Furthermore, an appreciation of the local currency of a country against the Euro decreases domestic credit growth, when that country has a relatively high share of exports to GDP. This result is similar in case of an appreciation against the US Dollar. Additionally, the results show that the export channel has a more significant effect on domestic credit growth when a local currency appreciates against the Euro. The financial channel has the most significant effect on domestic credit growth when a local currency appreciates against the US Dollar.
The results contribute to the existing literature on the risk of financial integration and domestic credit growth. This paper shows the negative effects of a more financially integrated world. An increasing number of countries have large amounts of foreign denominated liabilities. It is often forgotten that these foreign denominated liabilities could harm the economy if the local currency of a country, with a relatively high share of foreign denominated debt, depreciates (Bruno et al. 2014., Frömmel et al. (2016). Furthermore, the results highlight that there are differences in the impact of the financial channel and the export channel when a currency appreciates against the Euro or the US Dollar. In addition, the paper emphasizes that not only a local currency appreciation or depreciation has major influences on a country. An appreciation or depreciation of the Euro or the US Dollar could also influence a country’s economy.
Limitations
36 Future research
37 References
Agiomirgianakis, G. M., Sfakianakis, G., & Voulgaris, F. (2016). Determinants of economic growth revisited: is competitiveness and investment the solution for Greece?. International
Economics and Economic Policy, 13(3), 359-367.
Andrade, J. S., & Duarte, A. P. (2015). Optimum currency areas, real and nominal convergence in the European Union. Notas económicas, (42), 7-28.
Allen, F., & Gale, D. (2000). Financial contagion. Journal of political economy, 108(1), 1-33.
Arcand, J. L., Berkes, E., & Panizza, U. (2015). Too much finance?. Journal of Economic
Growth, 20(2), 105-148.
Beck, T. (2003). Financial dependence and international trade. Review of International
Economics, 11(2), 296-316.
Beck, R., & Rahbari, E. (2011). Optimal reserve composition in the presence of sudden stops. Journal of International Money and Finance, 30(6), 1107-1127.
Bridges, J., Jackson, C., & McGregor, D. (2017). Down in the slumps: the role of credit in five decades of recessions. Bank of Engeland, 659, 1-31
Bruno, V., & Shin, H. S. (2014). Cross-border banking and global liquidity. The Review of
Economic Studies, 42, 1-45
Calvo, G. A., Izquierdo, A., & Talvi, E. (2006). Phoenix miracles in emerging markets: recovering without credit from systemic financial crises. National Bureau of Economic
Research. 221, 1-42
Calvo, G. (2013). The Mayekawa Lecture: Puzzling over the Anatomy of Crises: Liquidity and the Veil of Finance. Monetary and Economic Studies, 31, 39-64.
Cetorelli, N., & Goldberg, L. S. (2011). Global banks and international shock transmission: Evidence from the crisis. IMF Economic Review, 59(1), 41-76.
Cetorelli, N., McAndrews, J., & Traina, J. (2014). Evolution in bank complexity. Federal
Reserve Bank of New York Economic Policy Review. 20(2), 85–106.
