Towards the growth of domestic credit in Tanzania : Does foreign capital flow really matter ?

This study aims to examine the relationship between international capital flows, and domestic credit expansion in Tanzania during the period between 2004 and 2012. The data used in this study is extracted from World Bank database except credit regulation quality index (CRINDEX), which is taken from the Fraser Institute Index of Economic Freedom. The study disintegrated the variable capital flow into debt and equity flows, and examined the relationship between the two sub-variables and the domestic credit. The findings of this study reveal that the general current account balance is not influential enough to determine the empirical relationship between international capital flows and domestic credit expansion; rather the component of international capital flow, net debt flow, is reported to have more significant relationship with domestic credit. The perceptible empirical relationship reported in this study between net debts flows and domestic credit development brings forward the need for analytical models which can explain this relationship. Particularly, it is imperative to gain a better understanding of both the positive and negative relationships between international debt flows and domestic credit growth. In essence due to the current East African Community Common Market Agreement, the financial integration and free mobility of capital among country members will have a serious effect on productive allocation of bank credit via the rise of inflows into the non-banking sector which crowd out domestic loans to non-financial business sector. This twist in credit allocation may result into real estate booms, financial vulnerability, and poor economic growth. Therefore, creating more investment opportunities could significantly alleviate the adverse effects of capital inflows.

Rwanda, Malawi, Zambia, Zimbabwe, Swaziland etc.Among the relief which these countries get include removal of exchange restrictions in all current account transactions some capital transactions such as long-term, non-debt-creating, and foreign direct investments (FDI).The purpose of this effort was to expand the domestic and foreign investment environments so as to improve, and encourage participations of many countries in capital markets.The fact that the world is now interconnected than some years back cannot go unnoticed although the recent financial crisis has reviewed the global factors which governs the financial conditions in the world (Lane and McQuade, 2013), The development of countries globally may be governed by the flow of capital from external sources and within the local financial systems.Such flow supports the domestic credit conditions in the particular country.International capital flow has been proven to support the economic growth.It is a known fact that the growth of a nation's economy depends to a great extent on a smooth flow of capital to businesses, and therefore foreign capital flow has significant impact on domestic credit (Reisen and Soto, 2001).
Several empirical studies have identified that foreign capital flow has a significant impact in economic growth as reported by Aizenman et al. (2013).According to the authors, the relationship between the two has been complex and mixed.Also, it is reported that the use of only FDI as a focal capital component has been a common practice and this has created several gaps in the literature.The literature forgets that foreign capital can be obtained in the third world countries not only through FDI but also through equity inflow and debt inflow and bank lending.
Studies which focused on FDI include Javorcik (2004), who reports positive impact of FDI on productivity; Li and Liu (2005) found out the positive direct relationship between FDI and economic growth, and indirect relationship via the human capital; Kose et al. (2009) show a significant evidence of the positive relationship between equity inflow and FDI.
Also, Choong et al. (2010) report the effect of private capital flow on economic growth.Opposing to these findings, Davis (2015) show that micro economic variables are affected more by changes in debt-based capital but not equity-based.Furthermore, Durham (2004) finds the insignificant relationship between either FDI or equity portfolio, and economic growth.Similarly, in their African study, Gui-Diby and Renard (2015) fail to get the significant relationship between international capital inflows in the form of FDI and industrialization.The authors suggested the weaker government policies which create FDI environment as the cause of failure of industrial development using foreign capital.
The developing countries have been currently affected by slow or poor movement of foreign capital due to the recent financial crisis which hit the banking industry of the Lotto 141 donors such as America and Europe.Lane and Milesi-Ferretti (2011) have documented that the size of recession variation during 2008 to 2009 was considerably related to the scale of domestic credit growth during the 2003 to 2008 periods, and the size of outstanding current account imbalances.
Relatedly, Lane and Milesi-Ferretti (2012) show that above-normal current account deficits during the precrisis period were significantly associated with major declines in domestic demand, and sharp reversals in private capital flows over 2008 to 2010.
The financial crisis is reported to be the result of two twin factors, balance sheet problems and a rapid increase in credit growth in some countries as put forward by Lane and McQuade (2013).The importance of these two factors brings the question of whether there is a significant relationship between domestic credit growth and international capital flows.
If the two variables are determined together this should develop the analytical framework which will ultimately guide theoretical and policy analysis.On one side, it would indicate that international capital flows should be a key theme in the mushrooming literature that tries to understand the dynamics of domestic credit growth.On the other hand, it would indicate that the domestic credit channel is a key channel in understanding the relation between international capital flows and domestic macroeconomic, and financial variables.
During pre-crisis period, Europe experienced substantial cross-country variation in domestic credit growth and cross-border capital flows.Lane and MCQuade (2013) investigated the inter-relations between domestic credit growth and international capital flows during boom period, and established that domestic credit growth in European countries is strongly related to net debt inflows but not to net equity inflows.
The development of the cross-border initiative framework facilitates the cross-border financial flows which can influence domestic credit growth through multiple channels.At a macroeconomic level, current account imbalances can affect macroeconomic variables such as the rate of output growth, the level of domestic spending, exchange rates, inflation and asset prices which can all influence equilibrium credit growth in a range of macrofinancial models.
This study, therefore, examines the relationship between international capital flows and domestic credit expansion in Tanzania during the period between 2004 and 2012.Tanzania is relevant for the study of this nature because it is involved in Cross-Border Initiative framework where the main agenda was to eliminate exchange restrictions on current account transactions without discrimination and to relax certain types of capital transactions.This agenda opens up the door for smoother flow of capital which is believed to have a great impact on domestic credit provided by banking sector in the region.
In particular, the study separately identifies net capital flows and domestic credit growth as important sources of macroeconomic imbalances, such that it is highly relevant to understand any inter-connections between such variables

METHODOLOGY Data and variables
The sample used in this study covers the period 2004 to 2012.This period is chosen because we want to understand the impact of the financial crisis of 2008/2009 on the growth of domestic credit in Tanzania.The period represents 3 years before 2008 to 2009 world financial crisis, and 3 years after the crisis.All data is extracted from World Bank database except credit regulation quality index (CRINDEX), which is taken from the Fraser Institute Index of Economic Freedom.The variables are defined as in the database as follows:

Dependent variable
Domestic credit provided by banking sector (% of GDP): This is defined as in the World Bank's data catalog as follows all gross credit facilities extended to the different sectors except central government.

Independent variables
Current account balance: This is the total of export less primary and secondary income measured in US Dollars.

Net flows on external debt
This is payments on long-term external debt net of principal repayments of LT external debt and IMF repurchases up to 1984, measured in US Dollars

Portfolio equity, net inflows
Portfolio equity includes net inflows from equity securities other than those recorded as direct investment and including shares, stocks, depository receipts (American or global), and direct purchases of shares in local stock markets by foreign investors.Data are in current U.S. dollars.

Gross domestic product (GDP) per capita (current US$)
This gross value of all contributed by all resident producers including any product taxes excluding any subsidy which are not part of the products value

Credit market regulation index
This is credit regulation quality component of the Fraser Institute's Indicators of Economic Freedom dataset.

Empirical analysis
In this study, the empirical analysis includes two simple Ordinary Least Square (OLS) regressions of the dependent variable domestic credit on the explanatory variable, and international capital flow.Before the simple OLS estimation was applied its basic assumptions (heteroscedasticity and multicollinearity) were tested.The results of the two tests conducted render the use of OLS valid.
In relation to international capital flows, we use the aggregate current account balance (net capital flows).We also split aggregate net flows between net debt flows and net equity flows.In this case we run two different OLS regressions; one between the aggregate net flows (current account balance) and domestic credit, and the second one between disaggregated net flows (net debt flows and net equity flows) and domestic credit.The two models are presented below.
The first model involves the independent variable as the aggregate international capital flow measured as the current account balance as a percentage of GDP while the second equation disaggregates the international capital flows into net debt capital flow and net equity capital flows both measured as the percentage of GDP DMCDTi,t= α + β1*CABGDPi,t + β2*LnGDPCi, t + β3*CMRIND i, t + ei.

Descriptive statistics
The analysis starts by examining the descriptive statistics.Starting with domestic credit, we theoretically know that the mean domestic credit reflects the average financial sector development.
It is reported in Table 1 that the mean value of domestic credit is 16.6.When compared to other East African Countries, it lags behind so much as, for instance, Kenya has 43.6 and Burundi 25.9.The country with the least growth of domestic credit is Uganda with the average value of domestic credit 11.8.Tanzania and Rwanda have closely similar average values of domestic credit as Rwanda has 15.6 as indicated in Table 1.
This shows, in general terms, that Tanzania's financial sector has an average lower growth rate for period from 2004 to 2012 compared to Kenya and Burundi but the financial sector growth of Tanzania is better than that of Uganda, and marginally that of Rwanda.
The other variable is current account balance.In this study, we found that Tanzania has an average deficit current account balance of -10.5 as presented in Table 5.This is the highest in East Africa compared to other four countries.The country with average lower deficit balance is Kenya with the average balance of -4.9.This is followed by Rwanda with the average deficit balance of -5.3, and then Uganda with the balance of -6.9.Burundi is closer to Tanzania with the balance of -9.4.The study further examines the domestic credit growth during 2004 to 2012.The motive of doing this is to see the impact of financial crisis of 2008 to 2009.To understand the impact of the crisis on the growth of domestic credit, the line graph is developed and the trend studied.Figure 1 show that Tanzania experienced significant growth of the domestic credit after the financial crisis.Figure 1  According to Financial Stability Report (2013), the banking sector which is the source of domestic credit growth continued to expand and remained profitable, highly liquid and adequately capitalized.The volume of deposits increased by 17.9% to TZS 14,175.57billion during the year ending March 2013 from the level recorded in the corresponding period in 2012.The number of banking institutions increased from 49 in March 2012 to 51 in March 2013, while that of branches rose from 521 to 559 during the same period.In as far as capital adequacy is concerned, the banking sector was adequately capitalized in aggregate terms during the year to March 2013.The industry's ratio of core capital to total risk-weighted assets increased from 17.9% in March 2012 to 18.9% in the year to March 2013, well above the minimum regulatory ratio of 10.0%.
Many factors may be contributed to the domestic credit levels.In their study on emerging markets, Giray and Kutay (2013) found that loose monetary policy in the domestic market, differences between domestic and global lending rates and real trade openness positively contribute to domestic credit levels.Their findings also show that external balance and perceptions of global tail risk negatively affect domestic credit levels.
Figure 2 shows how current account balance of Tanzania has been changing between 2004 to 2012.During 2007, the current account deficit widened by 3.7% to US$ 679.0 million from a deficit of US$ 654.5 million recorded during 2006.According BOT Economic Bulletin, (2007), the widening deficit follows a significant increase in imports of goods and services that could not be financed by an 8.9% increase in exports.
Also, the shortfall in disbursements of official transfers added pressure on the current account balance.The dismal performance of the goods account was mainly attributed to the decline in traditional exports as it was offseason for most of the traditional export.
Furthermore, during 2010, current account deficit narrowed to USD 797.3 million from a deficit of USD 898.4 million recorded in the corresponding period in 2009, as per BOT Monetary Policy Statement (2011), largely due to the rise in exports of goods and services and official current transfers.According to the statement, export of goods and services amounted to USD 3,356.4 million, which were 21.2% higher than the amount recorded in the corresponding period in 2009 .The value of exports of goods was 25.0% higher compared with values recorded in the same time frame a year ago.The policy statement further reveals that the higher values were attributed to increases in the export volumes of coffee, tobacco, and cashew nuts.In addition, export values of manufactured goods recorded an increase during the period, with much of the increase being recorded in export of plastic items, textile apparels and manufactured tobacco.
During the year ending December 2012, current account deficit narrowed to USD 3,438.0 million compared to a deficit of USD 3,977.1 million recorded in the corresponding period in 2011.BOT Monetary Policy Statement (2011) associates this development primarily with improved industrial production associated with stability in power supply, increase in international tourist arrivals, and increase in the volumes of cotton, coffee and tobacco following good weather.
According to the monetary policy statement (2013), the slowdown in growth of imports also contributed to the narrowing of the current account deficit.
Before processing the data for analysis we checked whether the data is reliable and valid so that the results extracted from such data are unbiased and accurate.We used Cronbach's alpha as a measure of validity that is the extent to which a scale records the "true" value or score of the concept you're trying to measure without capturing any unintended characteristics.According to -16 - Boermansab and Kattenbergb (2011), a reliable measure has a zero or very little random measurement error which might introduce arbitrary distortion into the measurement process, causing inconsistent measurements.
When the Cronbach's alpha test was run, it was found in Table 2 that Cronbach's alpha of the overall domestic credit scale is 0.7985 when the scores of all variables are combined in a scale under homogeneous weighting.This suggests the internal reliability of the scale is very high.Therefore, we have no doubt of data reliability and validity (Table 2).
Also prior to running linear regression several test should be done to confirm whether it is viable to run OLS.These tests include heroscedasticity and multicolinearity.

Testing for heteroscedasticity
The sample is said to have heteroskedasticity if the variance of the error term is not homogenous that is to say the variance of the error term is constant, and this is one of the assumptions on which OLS is built.The sources of heteroskedasticity include, among others, measurement errors, subpopulation difference for instance in our study the effect of domestic credit on international capital flow differ for different countries.Heteroskedasticity can also be caused by model misspecification using logarithms of some variables like in our case par capital income.
According to Long and Ervin (2000), when heteroscedasticity is moderate, OLS standard errors behave quite well.However, when heteroscedasticity is severe, ignoring it may render standard errors and pvalues biased, the direction of which depends on the pattern of heteroscedasticity.In some cases the form of the heteroscedasticity is clear and can be easily modeled.More commonly, though, heteroscedasticity is a trouble that can't be modeled because its source is not clearly understood.When the Breusch-Pagan test is run the results show that heteroskedasticity is not a problem because the variance of the error term is not constant.Table 3 shows that the hypothesis that the variance of the error term is constant is rejected, and therefore it is imperative to believe that the effect of heteroscedasticity does not exist in our case.

Testing for multicollinearity
We also check the possibility of multicollinearity which might have an influence on the study regression results.According to Wooldridge (2006), multicollinearity increases the variance of beta although it strictly does not violate OLS assumptions.
According to Wooldridge (2006), the level of multicollinearity is directly related to the size of the standard errors in in the study regressions.This test checks whether there is a need to disregard the simple OLS results, and renders them biased and inconsistent as previously reflected in Demsetz and Villalonga (2001) and Cho (1998).
To test whether there is a potential multicollinearity, we use VIF.run, the average VIF value was only 5.27 as presented in Table 3 indicating no threat of multicollinearity as this value is far more below the recommended threshold of 10 as previously suggested by Belsley et al. (1980) (Table 4).
Both tests indicate that assumptions of OLS about multicolinearity and heteroscedasticty do hold, and hence we can use OLS.

Regression results
The analysis begins by considering the international capital flow aggregately as current account balance in a regression setting.Because the current account balance is the composition of net debt and equity flows we expect different effects of such components on domestic credit, therefore, the study recognized this potential difference and therefore considered this disaggregation of current account balance to see how its components do differ in impacting the domestic credit.
The regression analysis begins by considering only GDP per capita and domestic credit, and results of these two variables may be seen in Table 5.The results show that GDP per capita is strongly significant at 1% significant level, and has the unexpected negative sign as implied in Bezemer et al. (2014).This result shows that in Tanzania, during 2004 to 2012, domestic credit did not have a promising level of growth (Table 5).
We extend the original regression equation by including the index which reflects liberalization of the credit market as discussed in the methodology and expand the model specification to include the credit market liberalization index.This index is introduced in the equation so as to take care of the features which relate to country financial systems.For example, a country whose credit market framework is conducive and more liberal may be more likely to adopt and bear rapid growth in credit when there is low risk aversion.
When this variable was added, the result was insignificant and negative as presented in regression always faster under more liberal regulatory regimes, and this would have been indicated by significantly positive relationship.Unfortunately, an opposite relationship is reported in this study suggesting that the credit growth in Tanzania is not faster as Tanzania is not one of the more liberal regulatory regimes.We further introduced the international capital flow variables in the model.We start by including the average current account balance.Regression in Table 7 shows that current account balance is significantly negative at 5% significant level showing that Tanzania was running current account deficits during this period.
We also examined whether the components of the current account balance, net international debt flows and net international equity flows, do have different patterns relationship with domestic credit growth.The results confirm a remarkable difference: It is reported that net debt flows are highly significant at 1% significant level but net equity flows are insignificant as presented in regression Table 8.This, according to Lane and McQuade (2013), shows that the aggregate current account balance cannot better explain the relationship which exists between the international capital flow and domestic credit but the components of the current account balance do actual give a very remarkable difference.This significant relationship between international net debt flows and credit growth is a unique result because in many studies the current account balance is taken as an aggregate variable.In this study case, disaggregating the current account balance has provided a very important insight that current account balance is not a reliable measure to uncover the relationship between international capital flows and domestic credit expansion.

Conclusion
This study aims to examine the relationship between international capital flows and domestic credit expansion in Tanzania during the period between 2004 and 2012.The findings of this work validate that the current account balance is not a reliable measure to uncover the relationship between international capital flows and domestic credit expansion.
The study disintegrated the variable capital flow into debt and equity flows, and examined the relationship between the two sub-variables and the domestic credit.The results of the study concluded that the general current account balance has got no influence in determining the empirical relationship between international capital flows and domestic credit expansion.The results of this study show a more significant relationship between international net debt flows and domestic credit.
This perceptible empirical relationship reported in this study between net debts flows and domestic credit development brings forward the need for analytical models which can explain this relationship.Particularly, it is imperative to gain a better understanding of both the positive and negative relationships between international debt flows and domestic credit growth.
In essence, due to the current East African Community Common Market Agreement, the financial integration and free mobility of capital among country members will have a serious effect on productive allocation of bank credit via the rise of inflows into the non-banking sector which crowd out domestic loans to non-financial business sector.This twist in credit allocation may result into real estate booms, financial vulnerability, and poor economic growth.
Therefore, creating more investment opportunities could significantly alleviate the adverse effects of capital inflows.

CONFLICT OF INTERESTS
The author has not declared any conflict of interests.
provided by banking sector as a % of GDP CABGDP= Current Account Balance as the % of GDP LnGDPC= Natural Logarithm of GDP per capita CMRIND = Credit Market Regulations Index NDEBTGDP = Net Debt flows as a % of GDP NEQUITYGDP =Net Equity flows as a % of GDP

Figure 1 .
Figure 1.A graph showing domestic credit growth for Tanzania from 2004-2012.
shows a consistent increase in domestic credit.The trend has been in an increasing side from 2006 to 2011 although a very slight decrease is observed between 2008 and 2009.On the other hand, current account balance fell rapidly before the financial crisis from 2005 to 2007.It then increased sharply between 2008 and 2009 before it remained constant in 2009 to 2010.It then finally dropped rapidly between 2010 and 2011, and again started increasing between 2011 and 2012.

Figure 2 .
Figure 2. A Graph showing current account balance for Tanzania from 2004-2012.

Table 1 .
The summary of mean domestic credit and current account balances.

Table 2 .
Data reliability and validity.

Table 5 .
The 1/VIF (tolerance factor) gives us what proportion of variance of an explanatory variable is independent of all the other explanatory variables.A VIF above 10 indicates potential trouble.When this test was A regression showing the relationship between dependent variable domcredit (domestic credit) and independent variables lngdppc (GDP per capita).

Table 6 .
A regression showing the relationship between dependent variable domcredit (domestic credit) and independent variables; cmrind (credit market regulation index) and lngdppc (GDP per capita).

Table 7 .
Table 6 contrary to findings ofGiray and Kutay (2013).According to Giray and Kutay (2013), credit growth is A regression showing the relationship between dependent variable domcredit (domestic credit) and independent variables; cmrind (credit market regulation index); lngdppc (GDP per capita) and capgdp (current account balance).