Money supply, inflation and economic growth of Sri Lanka: co-integration and causality analysis

Abstract

Purpose

GDP growth, money growth and inflation are essential to an economy's macroeconomic stability and have a direct impact on the policymaking process. Sri Lanka is currently concerned about high inflation. Inflation is a monetary phenomenon. Inflation has been caused by monetary policy in several nations. According to the economic theories of Karl Marx, Irving Fisher and Milton Friedman, a continuous increase in the money supply causes inflation. This paper aims to investigate the relationship between Sri Lanka's GDP growth, money growth and inflation.

Design/methodology/approach

An econometric model and the economic theories of Fisher and Friedman are used to figure out how money supply, inflation and economic growth are linked. Between 1990 and 2021, data were gathered from secondary sources.

Findings

The increase in the money supply is found to cause inflation. Inflation has negative effects on both short- and long-term economic growth. Long-term, the increase in money supply has a negative effect on economic growth.

Research limitations/implications

According to research, the money supply and inflation are inextricably linked, and the money supply has a direct impact on economic growth. As a result, the government should have an appropriate monetary policy and proposals to control inflation levels and stimulate economic growth.

Originality/value

The paper adds to the existing literature in two ways. First, it fills in the lack of studies in Sri Lanka, where there are no papers on this important relationship, especially with a modern econometric study. Second, it tries to shed light on the asymmetric shocks (both positive and negative shocks and changes) between the three variables, which was not done in previous studies.

Citation

Madurapperuma, W. (2023), "Money supply, inflation and economic growth of Sri Lanka: co-integration and causality analysis", Journal of Money and Business, Vol. 3 No. 2, pp. 227-236. https://doi.org/10.1108/JMB-08-2022-0039

1. Introduction

The primary goals of macroeconomic policy are to expand the economy and maintain price stability. Growth in the money supply ensures appropriate economic growth; however, inflation will result from insufficient money supply growth in the absence of corresponding macroeconomic policy. The correct amount of money in circulation is required for both economic expansion and the maintenance of stable prices. According to the Department of Census and Statistics, As shown in Figure 1, the Consumer Price Index (CPI) in Sri Lanka increased to 21.5% in March 2022 from 17.5% in February 2022. Food inflation increased from 24.7% in February to 29.5% in March as a result. Official data show that Sri Lanka printed 505.9bn rupees in 2020; however, the amount of liquidity injected into the banking system was significantly greater.

It is widely acknowledged that countries with more autonomous central banks achieve their primary goal of maintaining price stability with greater success. This is because central banks that are directly influenced by the government, such as in Sri Lanka, are required to bridge the unfinanced portion of the budget deficit by using their discretionary monopoly power of creating “fiat money” to lend to the government. This is because the government has direct influence over the central banks. Seigniorage refers to the revenue generated by the government through the creation of such money. This causes an increase in the central bank's monetary base, which causes a multiple expansion of the aggregate money supply, which causes inflation. This is exactly what has happened in Sri Lanka since the beginning of the previous year as a direct result of poor monetary policy formulation, which has resulted in inflation reaching double digits.

Sri Lanka's economy grew at a slower-than-expected 1.8% in the fourth quarter of the fiscal year 2021, bringing total year growth to 3.7%. According to analysts, inflationary pressures, which increased in the final three months of the year, as well as an unexpected tightening of rates by Sri Lanka's central bank, could be contributing factors to the negative economic growth. This paper will argue whether Sri Lanka's economic crisis should be declared over because of an excess of money supply. This paper investigates why Sri Lanka's inflation is caused by an excessive money supply. The purpose of this study is to add to the existing literature by analyzing the relationship between GDP growth, money supply and inflation in Sri Lanka as it transitions out of the current economic recession.

This study helps policymakers make monetary policy decisions. The government should not make monetary policy mistakes in order to control inflation and economic recession. The remainder of the paper is structured as follows: Section 2 explains the economics data and methodology; Section 3 summarizes the results; Section 4 discusses the findings; and Section 5 provides conclusions.

2. Literature review

Friedman (1970) believed a high money supply would, in the long run, result in a high rate of inflation. However, there is no correlation between money supply and inflation over a short period of time. According to the theory proposed by McCandless and Weber (1995), there is a very strong correlation between the amount of money in circulation and the rate of inflation, with the correlation coefficient falling somewhere in the range of 0.92–0.96, which is very close to the value of 1. Additionally, an increase in the money supply will, over the course of time, eventually result in the same degree of an increase in inflation. This is because money supply and inflation are directly correlated with one another. Tobin (1970) concluded, after conducting empirical research, that changes in the money supply will have an influence on fluctuations in output in the short term. Barro (1978) advanced the hypothesis that the anticipated rate of monetary expansion has no bearing on total economic output. In spite of this, Komendi and Meguire (1984), Ngoc (2020), Yuhan and Sohibien (2018) and Dinh, (2019) found that the money supply does, in fact, have an effect on output.

Empirical research Xie et al. (2009), Su et al. (2016) and Crowder's (1998) findings indicate that the amount of money in circulation does not influence actual output over the course of an extended period of time. In a nutshell, empirical studies (Makin et al., 2017; Jones and Manuelli, 1995; Ma, 2006; Alexandrov et al., 2021) carried out in western nations revealed that there was a great deal of unpredictability regarding the connection between the expansion of the money supply and the rate of economic growth and that there was no clearly defined connection between the money supply and the rate of economic growth as of yet. These findings were uncovered Fang, (2012), He and Liu (2011) and Liu and Pang (2011) held the view that an excessive amount of money was being produced, which they believed to be the primary cause of China's inflation (Fang, 2012). According to the findings of Lu et al. (2017), the amount of money that is circulating in the economy does not have a significant co-integration relationship with the rate of inflation. They concluded that an increase in the money supply did not necessarily result in inflation during the process of monetizing the economy because the money supply was already inflated prior to the increase in the money supply.

The marginal had been absorbed by the economy during the process of monetization. Huang and Deng (2000) and Hu and Zhang (2009) believed that the money supply had a non-neutral effect on output, regardless of whether the impact of the money supply was anticipated or not. They came to this conclusion even though the impact of the money supply was anticipated. To put it another way, the total amount of money that was available remained an essential component in the operation of China's economy. Tin (2017) held the belief that shifts in the money supply would have an effect on output in the short term, but that these shifts would not have an effect on output in the long term.

As we can see from the analysis of the pertinent literature that was just carried out, the results of the various studies do not concur with one another in any way. This phenomenon can be largely attributed to the selection of the appropriate sample spacing and modeling techniques, both of which were discussed earlier. In this article, we use the co-integration analysis to determine whether there is a balanced relationship between the country of Sri Lanka's money supply, economic growth and inflation. In addition, we investigate the factors that contribute to inflation in Sri Lanka by employing a test known as the Granger causality test.

3. Methodology

In this paper, we use the time series data of broad money (M2), gross domestic product (GDP) and GDP deflator from 1990 to 2021 to look at the relationship between money supply, economic growth and inflation in Sri Lanka. All of the information comes from the Central Bank of Sri Lanka's and the Department of census and Statistics' annual reports. First, the stationary test was done to see if all of the time series data had unit roots. The data is processed with Eviews 5.0. Money supply can be broken up into three administrative levels based on how liquid it is: M0 (cash in circulation), M1 (currency) and M2 (currency and Quasi money). In this study, M2 is used as a stand-in for the amount of money in circulation. The level of economic growth is shown by the gross domestic product (GDP). The consumer price index (CPI), the producer price index (PPI) and the GDP deflator can all be used to figure out how much prices have gone up. In this study, inflation is measured by the GDP deflator. The General Multiple Linear Regression was chosen because it met the following requirements.

(1)GDP =F (M2,INF)

From equation (1), the econometric form of the equation is specified in natural logarithm as;

(2)lnGDPt =β0 +β1lnINFt +β2lnM2t+Ut

where; INF = Inflation, GDP = Gross Domestic Product, M2 = Money Supply Growth and U_t = Error Term β₀ is a constant and β₁to β₂ are the parameters to be estimated.

4. Analysis of empirical results and discussion

The empirical investigation begins with a descriptive/summary statistics analysis of the investigated variables. The mean was employed as a measure of central tendency, while the standard deviation was employed as a measure of dispersion. Table 1 demonstrates that all variables have a positive mean and standard deviation, with the standard deviation being greatest for money supply and lowest for GDP. Skewness and Kurtosis represent the measures of Skewness and Peakedness, while the maximum and minimum values for each variable are also calculated (Table 1) (see Table 2).

To find out if the model had a problem with multicollinearity, a correlation matrix was used to find out how much the variables under study were related to each other. Correlation shows how changes in one variable can be caused by changes in another. Multicollinearity can happen when there is a strong link between a lot of different things. In the table below, you can see what the matrix came up with. The general rule is that multicollinearity between two variables is a cause for concern if it is 80% or higher. But the current study doesn't show any major cases of multicollinearity between the variables, as the correlation between inflation and money supply is only 62%. This proves that there is no multicollinearity among the model's variables.

4.1 Unit root test analysis

The Unit Root Test is used to see if a set of data is stationary. This step is very important because if non-stationary variables aren't found and used in the model, it will lead to a problem called “spurious regression.” In spurious regression, the results show that there is a statistically significant and meaningful relationship between and among the variables in the stated regression model when in reality, there is only a correlation, not a meaningful causal relationship. The Augmented Dickey-Fuller and Phillips-Perron tests were done, and the results are shown in Table 3. The result of the unit root test shows that none of the model's variables is stationary at their levels, but they all become stationary after the first difference. This means that co-integration analysis should be used since the idea of co-integration says that variables must be stable after at least one difference (1).

5. Analysis of the Co-integration test

The stationary linear combination is called the co-integrating equation, and it can be thought of as a relationship between the variables in which they are in long-term equilibrium. The goal is to find the stationary linear combination of the time series variables that are being looked at. So, the co-integration method developed by Johansen and Juselius (1988, 1990) has been used to find stable long-term relationships between inflation, GDP growth and money supply growth in Sri Lanka. Both the Trace and Maximum-Eigen tests have been used to find these relationships. In Table 4, the results are shown. At the 5% significance level, both the trace test and the maximum-eigen test show that there is a co-integrating equation. So, the idea that there is not a co-integrating equation is not true. As a result, we can say that there is a strong long-term link between the given variables. An error correction model (ECM) is used to separate this effect because variables can have either long-term or short-term effects.

Table 5 displays the long-run regression results. There is a clear and significant negative relationship between GDP, inflation and money supply. Money supply and inflation both have a negative effect on GDP. It implies that a 1% change (increase) in inflation results in a 66% decrease in economic growth (GDP) and a 1% increase in the money supply results in a 12% decrease in economic growth. The model's long-run co-integrating relationship shows that inflation and money supply growth have a significant negative impact on economic growth in Sri Lanka.

5.1 Short-run dynamics (ECM)

The coefficient of money supply has a positive and significant effect on short-run economic growth, indicating a direct relationship between money supply and short-run economic growth. This suggests that a 1% increase in money supply can result in an economic growth increase of approximately 0.54%. This result is consistent with theories and findings from previous research suggesting a positive relationship between the two variables. However, the results IN Table 6 indicate that an increase in Sri Lanka's money supply will result in inflation. This suggests that a 1% increase in money supply can result in an inflation increase of approximately 0.46% in Sri Lanka. The first lag's inflation coefficient is −0.361, and the second lag's coefficient is −0.3667; both values are statistically significant. The relationship between GDP growth and inflation is inverse. This indicates that a one percent increase in prices results in a 0.36% decline in economic growth in Sri Lanka (see Table 7).

The coefficient of the error correction term indicates the rate at which the deviation from the long-run equilibrium is eliminated. It indicates how long it would take the economy to reach equilibrium in the long run. Its coefficient, 0.601, is statistically significant. This indicates that the rate of adjustment is approximately 0.601%, indicating that if there is a deviation from equilibrium, approximately 0.60% of GDP is adjusted annually as the variable moves toward restoring equilibrium. The results indicate that the deviation from the long-term growth path caused by a particular shock is adjusted annually by sixty percent. This suggests that economic growth is capable of readjusting to long-term equilibrium following each short-term "shock" caused by inflation and money supply. The adjusted R-squared value is 0.632, which indicates that approximately 63.2% of the variation in the GDP is explained by the independent variables, indicating a very good fit. The Durbin–Watson statistic is high, indicating that there is no autocorrelation of the first order. The overall equation is statistically significant, as indicated by the F-statistic probability value (0.008407).

5.1.1 Granger causality test

Granger causality tests are performed to confirm the causal relationship between money supply, economic growth and inflation. Table 7 outlines the test results. Granger causality exists between the money supply and inflation.

Money supply can contribute to inflation. Inflation can be caused by money supply, and the p-value is 0.0490. This indicates that the expansion of the money supply is the primary cause of inflation in Sri Lanka. According to the quantity theory of money, the excessive growth of money supply will inevitably result in inflation.

The results of diagnostic tests are depicted in Table 8. It was verified that the error terms of the short-run models lack heteroscedasticity, lack serial correlation and are normally distributed. In addition, the Durbin–Watson statistic was found to be greater than R2, indicating that the short-run models are not spurious.

The stability of the long run parameters was tested using CUSUM of recursive squares (CUSUMQ). The CUSUM test results are presented in Figure 2.

The CUSUM test results are presented in Figure 1. The graphs presented above provide clear evidence that the model fits well within the confident level line. Therefore, based on the CUSUM test graph, we can conclude that the model utilized for the analysis is the most suitable. Subsequently, results fail to reject the null hypothesis at 5% level of significance because the plots of the tests fall within the critical limits. Therefore, it can be realized that the selected VECM model is stable.

6. Concluding remarks and future research

In this paper, an attempt was made to investigate the dynamics of GDP, inflation and money supply in Sri Lanka using a multivariate autoregressive framework for the period of 1990–2021. The time period covered by this study is from 1990 to 2021. Only after performing a first order difference are all of the variables confirmed to be stationary, confirming the existence of a long run equilibrium relationship between the variables. The cointegration test also hints at a long-run equilibrium relationship, which inspires an interest in learning more about the dynamics of the variables under investigation in this research. The co-integration result demonstrates that there is at least one linear combination in the long run, and as a consequence, there is a long run equilibrium relationship between the variables in the model. This indicates that money supply and inflation have a negative effect on the GDP growth rate in the economy. According to the results of the error correction, both the positive and negative signs for gross domestic product and their respective values are highly significant. At the 5% level of significance, the result of Granger causality reveals that there is a unidirectional causality running from money supply to inflation. This causality runs in the same direction as the money supply.

The following are the two primary points that are supported by this study:

1. There is significant statistical evidence to support the contention that inflation has a dampening effect, both in the short run and in the long run, on the rate at which economic growth occurs.

2. The availability of money has a beneficial effect on economic expansion in the short run, but a deleterious effect over the course of a full economic cycle. In light of the findings of the empirical research, the authors suggest the following policy implications: To begin, the authorities in charge of the economy need to switch between an expanded monetary policy and a contractionary monetary policy on a regular basis. Because increasing the money supply will assist in promoting a higher rate of growth, but extreme policies regarding the money supply may be detrimental to growth. Second, inflation has a deleterious effect on economic growth in the case of Sri Lanka, and high inflation will have the effect of destroying economic activities. The government should strive to keep inflation at low and stable rates in order to achieve high and sustainable growth rates. As a result, the government ought to be in possession of the pertinent monetary policy in order to grow the economy as well as proposals to make monetary policy, control inflation levels and stimulate economic growth.

Figures

Figure 1

Money supply and inflation, 2015–2021

Figure 2

Stability test

Table 1

Descriptive statistics

	ln (GDP)	ln(INF)	ln(M2)
Mean	1.638485	2.051932	13.88344
Median	1.694709	2.145931	13.83754
Maximum	2.213161	3.114848	15.84683
Minimum	0.813228	0.758484	12.46629
Std. Dev	0.321019	0.569644	1.182842
Skewness	−0.381120	−0.258559	0.209819
Kurtosis	2.923117	2.688635	1.639784
Jarque-Bera	0.709197	0.440268	2.448426
Probability	0.701455	0.802411	0.293989
Sum	47.51606	59.50604	402.6199
Sum Sq. Dev	2.885483	9.085828	39.17523

Source(s): Author calculation, 2021

Table 2

Correlation analysis

	ln(GDP)	ln(INF)	ln(M2)
ln(GDP)	1.000000	0.416066	−0.267096
ln(INF)	0.416066	1.000000	−0.628544
ln(M2)	−0.267096	−0.628544	1.000000

Source(s): Author calculation, 2021

Table 3

Results of unit root tests

Variables	ADF(0)	ADF(1)	PP(0)	PP(1)
Ln(GDP)	−1.02319	−7.0386***	−0.70230	−8.13942***
Ln(CPI)	−1.01689	−5.13179***	−0.92136	−12.2556***
Ln(M2)	−0.18774	−2.437***	−0.65509	−3.73498***

Note(s): The numeric values in cells are t-statistic. Probability values for rejection of the null hypothesis are employed at the 5% significant level (**, p-value <0.05 and***, p-value <0.01)

Source(s): Author calculation, 2021

Table 4

Results of Johansen cointegration test

Number of cointegration (r)	Trace statistic	Maximum eigenvalue statistic
r = 0	14.45508 (29.79707)	12.25598(21.13162)
r ≤ 1	65.4292 (69.818)	14.26460 (0.9970)
r ≤ 2	15.49471 (0.9918)	3.841466 (0.4437)

Note(s): Series: ln(GDP) ln(M2) ln(INF)

Cointegrating equations are significant at the 0.05 level (**, p-value <0.05 and ***, p-value <0.01)

Trace test indicates no cointegration at the 0.05 level; Max-eigenvalue test indicates no cointegration at the 0.05 level

Source(s): Author calculation, 2021

Table 5

Long run test Results

Dependent variable = lnGDP
Long term Results
Regressor	Coefficient	Standard Error	T-statistics
Constant
LnM2	−0.116574*	0.10669	−1.09262
lnINF	−0.662348*	0.22161	−2.98876
constant	1.298223
R-Squared	0.631535
DW test

Note(s): *, **, *** represent 1, 5 and 10% significance levels, respectively

Source(s): Author calculation, 2021

Table 6

Error correction model. Dependent variable dlnGDP

Error correction	D(ln(GDP))	D(ln(INF))	D(ln(M2))
ECT	−0.601176 **	1.317188	0.033387
	(0.34564)	(0.57875)	(0.04790)
	[−1.73930]	[ 2.27594]	[ 0.69695]
D(ln(GDP(−1)))	0.005464	−0.343730	−0.023811
	(0.31799)	(0.53245)	(0.04407)
	[ 0.01718]	[−0.64557]	[−0.54028]
D(ln(GDP(−2)))	0.043731	−0.210192	−0.039300
	(0.23388)	(0.39161)	(0.03241)
	[ 0.18698]	[−0.53674]	[−1.21242]
D(ln(INF(−1)))	−0.360628**	−0.115706	0.029349
	(0.18278)	(0.30605)	(0.02533)
	[−1.97301]	[−0.37806]	[ 1.15858]
D(ln(INF(−2)))	−0.366716*	−0.059357	−4.87E-06
	(0.14433)	(0.24166)	(0.02000)
	[−2.54087]	[−0.24562]	[−0.00024]
D(ln(M2(−1)))	5.354260*	4.693967**	0.776025
	(1.81365)	(3.03678)	(0.25136)
	[ 2.95219]	[ 1.54570]	[ 3.08729]
D(ln(M2(−2)))	−4.805445**	−4.842874*	−0.068086
	(1.67670)	(2.80746)	(0.23238)
	[−2.86602]	[−1.72500]	[−0.29300]
C	−0.166737	−0.098760	0.037290
	(0.14054)	(0.23532)	(0.01948)
	[−1.18640]	[−0.41968]	[1.91448]
R-squared	0.631535	0.601040	0.668526
Adj. R-squared	0.459585	0.414858	0.513838
Sum sq. resids	1.190990	3.339075	0.022877
S.E. equation	0.281779	0.471810	0.039053
F-statistic	3.672775	3.228247	4.321776

Note(s): *, **, *** represent 1, 5 and 10%

Source(s): Author calculation, 2021

Table 7

Pairwise granger causality tests

Null hypothesis	Obs	F-statistic	Prob
LOG(INF) does not Granger Cause LOG(GDP)	27	0.20479	0.6549
LOG(GDP) does not Granger Cause LOG(INF)		3.06181	0.0929
LOG(M2) does not Granger Cause LOG(GDP)	27	1.74413	0.1991
LOG(GDP) does not Granger Cause LOG(M2)		1.02233	0.3221
LOG(M2) does not Granger Cause LOG(INF)	31	4.23597	0.0490
LOG(INF) does not Granger Cause LOG(M2)		0.70109	0.4095

Source(s): Author calculation, 2021

Table 8

Short run diagnostics

Diagnostic test results
Test	F-statistics	p-value
Normality	1.1389	0.5658
Heteroskedasticity	0.8457	0.5601
Serial Correlation	0.2452	0.7844

Source(s): Author calculation, 2021