Project Topic

THE RELATIONSHIP BETWEEN STOCK PRICES AND COMMERCIAL BANKS LENDING RATES IN NIGERIA

Project Attributes
 Format: MS word ::   Chapters: 1-5 ::   Pages: 49 ::   Attributes: Questionnaire, Data Analysis,Abstract  ::   1449 people found this useful

Project Department

Project Body

1.0 REVIEW OF RELATED LITERATURE

In Nigeria, two broad monetary policy regimes could be distinguished since the establishment of the Central Bank in 1959. The first regime was characterized by direct administrative controls on credit and interest rates, while the other dwells on the era where credit to the private sector was competitively distributed. In the first period (until 1986) banks were assigned mandatory guidelines on how much credit to make to preferred sectors of the economy. More so, minimum cash ratios were stipulated and special deposits were used to control free reserves with banks. Thus, banks made most loans essentially in order to meet government regulations and not necessarily based on the expected returns. The wisdom of this era was to stimulate growth of the domestic economy by delivering credit at low interest rates, while pegging and defending the naira from wanton depreciation, mainly motivated by the need to avoid foreign inflation in imported intermediate and capital goods. According to Nnanna (2001) historically, the Central Bank of Nigeria via its monetary policy circulars had directly controlled the volume and cost of credit in the economy, until the era of financial sector liberalization in the mid-80s. He also found in the same study that distortions in the pricing of loans caused by the administrative intervention in the market rendered financial intermediation by the deposit money banks ineffective. In the era of liberalized interest rates beginning from 1993, deposit money banks engaged in diligent credit packaging and risk analysis before making loans in order to reduce carrying non-performing assets in their books. Consequently, loans were advanced based on the computed returns to investment and the relative risk in the borrowing sector. The major consequences of liberalization have been increases in the volatility of interest rates and increased sensitivity of the exchange rate to domestic economic developments and external shocks, which eventually affected prices and consequently, increased the sensitivity of exchange rate to the interest rate as stable exchange rate helped lock-in inflation. From 1970 to 1985 (Figure 1), which marked the period of strict administrative controls, the standard deviation of the prime lending rate was 1.8 and rose to 6.8 in the period between 1986 and 1992. During this period the economy had become liberalized to a large extent, but interest rate liberalization only came in 1993. However, the extent of dispersion of interest rate slowed to 5.3% in the years since 1993. The wide dispersion in the years after direct controls is indicative of the effects of market interactions and administrative frictions in the policy break point, while the relative convergence after 1993 could be explained by the numerous entries in the banking industry and improved efficiency of the intermediation process. Nnanna and Dogo (1998) have shown that financial liberalization has led to increased credit to the private sector of the economy. However, evaluating the sectoral distribution of loans by the deposit money banks in Nigeria, it could be observed that the real sectors of the economy have not benefited proportionately. This situation could be attributed to the relative high risk and the long period of pay back associated with the sector. Foreign exchange movements have been pivotal in the supply of money in the Nigerian economy, particularly, since the commercial exploration of crude oil. Foreign exchange policies have essentially sought to ensure a healthy balance of payments and the attainment of a stable exchange rate. Before deregulation of the economy external sector policies depended on foreign exchange allocations and administered exchange rates. In the circumstance, the levels of money supply flowing from net foreign earnings were fairly stable and predictable

Traditional literature links the extent of bank lending to the level of economic activity and interest rate. Growth in the Gross Domestic product calls forth greater investment and greater demand for bank loans. While low interest rates encourage consumption and grow loans. Greene and Villanueve (1991) show strong negative correlation between real interest rates and private investments. Following are other factors that influence the lending behaviour of banks. By regulation, the single obligor limits for lending connects the size of a bank’s balance sheet to the volume of loans it can make to an enterprise. Literature on capital adequacy and other prudential guidelines are extensive and their links to the lending pattern of banks have been well documented in the literature including (Kashyap and Stein, 2000; Benanke and Gertler, 1987) Both studies infer that in situations of credit constraints, the level of capital will determine the extent of bank lending. However, it has been argued that though it might appear apparent that the level of banks’ capital does matter for the volume of lending, what is less clear is whether the trends in bank loans are caused by variations in capital or by cheer changes in the level of demand for bank loans (Sharpe, 1995). Financial liberalization, particularly external finance liberalization and stock price movements also influence bank lending. Financial liberalization unleashes mixed impact for economic agents in the loans market. For companies, while the international sources of capital are opened including share offers and the relative price of foreign capital tends to decline. For banks, competition heightens and the interest rate spread diminishes. Financial liberalization also improves the ability of domestic banks in developing markets to improve on credit packaging and risk assessment. Overall, the loan base for banks is enlarged. Olaf et al. (2007) summarizes the arguments for financial liberalization in Thailand by stating that the good news is that liberalization makes borrowing cheaper and easier as it decreases interest rate spreads and reduces collateral requirements. Moreover, it modernizes the financial system by enlarging the power of market forces at the cost of traditional institutions, here reflected in a declining importance of collateral based and relationship lending. However, one of the downsides of financial liberalization is the fact that more risky ventures could be financed by banks and in the face of reduced collateral requirements. Banking becomes more risky by greater interest rate and exchange rate changes (Stiglitz, 2004). Equity price fluctuations may affect the lending behaviour of banks, especially where banking regulations do not impose explicit ceilings on lending. In such operating environments, loans flow freely to sectors where return on investment is higher and risk is well understood and could be managed. The connection between equity price fluctuations and lending behaviour of banks could be traced in diverse dimensions. In jurisdictions that banks hold equities in their portfolio of investments, an increase in share values will boost the size of banks’ balance sheets and encourage increased lending. The reverse will play out when equity prices dip, other factors held steady. There might be a twist in this line of linkage especially for lending in developing countries. Banks in these countries have very shallow avenues for investment such that the stock market acts as an active competitor for investments vis-à-vis loans and advances. In this circumstance, expectation might be that an increase in stock prices will attract funds away from loans in favour of incremental outlay on stock investments. Declining stock prices also weakens borrowers’ collaterals held in equities, thus shrinking the demand for loans. Kim and Ramon (1994) evaluated stock prices and bank lending behaviour in Japan and found that changes in stock prices positively correlated with loans advanced, particularly, in the period after the loans market had been deregulated. Mansor (2006) applied the VAR technique to discern the effects which stock market fluctuations could have on the volume of loans and whether bank loans propagate financial shocks to the real economy in Malaysia. The variables included were bank loans, stock prices, consumer price index, Gross Domestic Product, Interest rate and exchange rate. The result indicated a positive response of bank loans to innovations on stock prices. However, there was no evidence of feed back from bank loans to stock prices. The latter finding led to the conclusion that bank health may depend crucially on the stock market, but that the attempt to invigorate the stock market through increased lending is futile. The other ancillary finding is that despite much hype on the currency mismatching of bank assets and liabilities, there seems to be no effect of exchange rate on bank loans. The exchange rate may only affect bank loans indirectly through its effect on stock prices and real output – which are dampened by currency depreciation. To identify the under currents in the fluctuation of loan supplies, an alternate approach has emerged which links loan supply to macroeconomic shocks. Potential sources of macroeconomic shocks are exchange rate changes, interest rate fluctuations, changing monetary policy stance, financial market volatilities and fiscal actions of governments. Degirmen (2007) applied the vector autoregression (VAR) to determine that public borrowing in Turkey crowded out private loans. From a policy perspective, the lending view of monetary policy transmission is anchored on the hypothesis that reduction in bank reserves squeezes their loan making capabilities. Mbutor (2007) has documented that monetary tightening, signaled by an increase in the monetary policy rate, reduces bank lending in Nigeria. The outcome is explained by many factors including the divestment from loans and advances to investment in government securities and other short term inter bank outlets. Azis and Thorbecke (2002) show that positive interest rate and exchange rate shocks decrease both capital and loan growth in domestic banks relative to foreign banks in Indonesia. Generally, the nature of the macroeconomic environment influences the lending behaviour of banks. A booming economy provokes expectations that future flow of income streams are assured, thus, encouraging demand and supply for loans. As asserted by Talavera et al. (2006), banks make out more loans during periods of boom and reduced level of macroeconomic uncertainty and curtail lending when the economy is in recession. Studies focused on the effects which macroeconomic stability might have on the lending behaviour of banks in Nigeria have received limited attention. However, Somoye and Ilo (2009) in a recent study have indicated measures of macroeconomic stability- including changes in money supply, exchange rate and inflation- impact on bank loans only in the long run perspective: while in the short run, the total deposit and capital base of banks play very important roles in influencing the ability of banks to make loans. The mixed nature of outcomes of studies regarding the impact the stock prices and exchange rate dynamics might have on the lending pattern of banks lends necessity to this study of how bank lending in Nigeria reacts to exchange rate and equity price fluctuations. Asset prices, broadly defined, would have been more reflective of portfolio adjustments, but data are mainly available in high frequencies only on the equity front.

1.1 THE DATA

YEAR

LOA

VOD

LR

GDP

DPR

1985

17598

25460

25.6

11.4

55.6

1986

18500

31750.5

26.4

13.7

54.7

1987

21350.5

33554.8

28.6

14.8

55.6

1988

22500

36750

32.4

14.91

56.8

1989

23450

37456.9

35.8

18.85

62.3

1990

26000

38777.3

44.3

19.4

66.5

1991

31306.2

53208.7

38.6

29.0

80.4

1992

42735.8

75047.7

29.1

39.4

66.5

1993

65665.3

110453.60

42.2

63.2

59.8

1994

66127.3

142537.5

48.5

67.8

55.2

1995

114668.9

178962.10

33.1

78.2

42.9

1996

169437.10

214357.8

43.1

85.8

60.9

1997

385551.0

280028.7

40.2

85.7

73.3

1998

272895.5

314303.5

46.8

89.6

72.9

1999

322764.9

476350.9

61

89.9

76.6

2000

508302.2

702104.5

64.1

86.8

74.4

2001

796164.80

928326.9

52.9

87.6

54.6

2002

954628.8

1100710.30

50.9

87.7

51.0

2003

1210033.10

1294492.8

50.5

87.9

65.6

2004

1519242.7

1606174.7

50.2

90.3

62.8

2005

1847822.6

989791.2

51.7

90.5

61.9

2006

19270556.7

990787.6

52.8

91.5

62.7

2007

1956306.4

1166750.6

52.9

93.14

66.10

2008

1200567.1

1285750.5

57.8

93.7

63.52

2009

1211766.6

1386678.7

49.01

95.6

60.05

2010

1323667.9

1297876.7

48.72

96.6

59.75

2011

1526647.7

977665.7

51.93

98.6

61.23

2012

1583677.4

967675.5

55.60

98.9

66.75

2013

1697696.7

878766.5

48.71

98.45

68.76

2014

2905764.1

890725.7

46.02

89.9

51.02

2015

2978988.2

870743.1

49.03

88.65

50.07

CBN statistical bulletin 2016

H0: there is no significant relationship between bank lending and GDP

1.2 METHODLOGY

1.2.1 INTRODUCTION

This chapter describes the various methods and techniques used to collect and analyze the data gathered for the study to gain a deeper understanding of the topic under study.

The data collection stage is important since the result of the analysis is dependent on the quality of the data obtained. Therefore, the method selected for data collection must be the most appropriate to assist in achieving the objectives of the study:

1.2.2 MODEL SPECIFICATION

The model for the study comprises of two constructs as described below:

MODEL 1

GDP= α+β1 LOA+ e-------------------------------------- (1)

 

Where

GDP signifies gross domestic product

LOA signifies Loans and advances

α is the equation’s constant.

β1 the coefficient of loan and advances

e Is the error term of the equation

1.2.3 MODEL LIMITATION

The model is limited to only the above variable due to the unavailability of other statistical bulletin for other variables; but the research work was able to evaluate the determinants of banks lending behavior.

1.2.3.1 Reason for the adoption of the model: The model was adopted because it will elicit information on the nature of the kind of relationship between the dependent variable and the rest of the independent variables as listed above.

1.2.4 THE VARIABLES

Dependent variables: The dependent variable is variable that other variables are dependent on; for the first model the dependent variable is the GDP (Gross domestic products

Independent variable: these are those variables that are not dependent on any other variable; the independent variable for the model above is the loan and advances

1.2.4.1 Data requirements

The study will make use of the data from the statistical bulletin from CBN statistical bulletin 2016; the data required for the study include the data for the LOA, GDP, VOD, LR and DPR

1.2.4.2 Definition of variables:

LR: Liquidity ratio Comptrollers Handbook (1998), states that lending is the principal business activity for most commercial banks. The loan portfolio is typically the largest asset and the predominate source of revenue. As such, it is one of the greatest sources of risk to a bank’s safety and soundness. Since loans are illiquid assets, increase in the amount of loans means increase in illiquid assets in the asset portfolio of a bank. According to Pilbeam (2005, p. 42), in practice the amount of liquidity held by banks is heavily influenced by loan demand that is the base for loan growth. If demand for loans is weak, then the bank tends to hold more liquid assets (i.e. short term assets), whereas if demand for loans is high they tend to hold less liquid assets since long term loans are generally more profitable. Therefore loans and advances have negative impact on banks liquidity and vice versa.

LOA: Loans refers to a debt provided by a financial institution for a certain period while Advances are the funds provided by the banks, which needs to be payable within one year. After a detailed research on the two terms we are here compiling an article, in which you will find all the necessary differences between loans and advances.

DR: Deposits rate is the frequency at which money is going into the bank.
 

GDP: Gross domestic product of Nigeria

1.2.4.3 Sources of data

The source of data of this research is mainly secondary data. These are obtained from various source such as the world bank and the central bank of Nigeria (CBN), publication, e.g. statistically bulletin, statement of account, annual report, bulletin e.t.c.) Publication from the National Bureau of Statistic (NBS) and International Journal of the International finance statistics year book.

1.2.4.4 Unit of measurement

The data collected for the study from the central bank statistical bulletin measured in naira.

1.2.5 Method of analysis 

The data collected will be analyzed using ordinary least square regression method (OLS) to test for the nature of the relationship between the dependent variables and the rest of the independent variables.

1.2.5.1 Evaluation methods 

The techniques employed in this research is the simple linear the relationship between the dependent and independent variable(s).

        The equations are filled with a least square to determine the parameters of the model. The models are estimated using annual Nigeria data for the period of 1985-2015. Although there is no consensus on which available econometric models is the most suitable for empirical studies but the parameter estimates obtained by ordinary least square (OLS) have some optional properties and the computational procedures of the OLS is fairly simple as compared with the other econometric techniques. Similarly, the OLS method has been used as a wide range of economic relationship with fairly satisfactory results.

1.3 DATA ANALYSIS AND INTERPRETATION

RESEARCH HYPOTHESIS

Hypothesis to be tested

H0: there is no significant relationship between bank lending and GDP

H1: there is a significant relationship between bank lending and GDP

Level of significance (α=0.05)    

Decision Rule: reject H0 if p-value is less than the level of significance; otherwise accept the null hypothesis

Table 1

ANOVAa

Model

Sum of Squares

df

Mean Square

F

Sig.

1

Regression

1860.712

4

465.178

65.948

.000b

Residual

169.288

24

7.054

 

 

Total

2030.000

28

 

 

 

a. Dependent Variable: GDP

b. Predictors: (Constant), LOA

 

CONCLUSION BASED ON THE DECISION RULE

 

From the result in table 1 above, the p-value is 0.00 which is less than the level of significance; we therefore reject the null hypothesis and conclude that there is a significant relationship between bank lending and GDP

Table 2

Model Summary

Model

R

R Square

Adjusted R Square

Std. Error of the Estimate

1

.879a

.778

.646

4.90291

a. Predictors: (Constant), LOA

 

From table 2 above, the multiple correlation indexes is 0.879, this shows that there is a strong positive relationship between bank lending and GDP

The value of R2 is 0.778; this means that the explanatory variable was able to explain 77.8% of the dependent variable while the unexplained 22.2% was captured as the error term

The value of the adjusted R2 is 0.646; this means that the explanatory variable was able to explain 64.6% of the dependent variable while the unexplained 35.4% was captured as the error term

Table 3

Coefficients

Model

Unstandardized Coefficients

Standardized Coefficients

t

Sig.

B

Std. Error

Beta

1

(Constant)

67.216

6.025

 

11.156

.000

LOA

2.581E-006

.000

.279

1.566

Get the complete project »

Can't find what you are looking for?
Call 0906 809 7513

  • Subscribe to Free Job Alert
    Enter your email below and click subscribe

    LATEST JOB VACANCIES


    We require the services of an experienced Business Development Manager with a wide range of business clientele and a network of c... Read more

    Regulate day-to-day operations of unit in conjunction with Departmental Heads, Manager, Executive Chef, Security. Cordinate and l... Read more

    FINANCE OFFICER

    LEAD Enterprise Support Company Limited in (Lagos State)
    Job Objective: The Finance Officer will assist the Finance/Admin Manager in the implementation of the HMO’s accounting policies an... Read more

    Copyright © 2024 All Right Reserved CVClue
    A Subsidiary of EMINENT INFO TECH VENTURES