Does Monetary Policy Affect Economic Growth in Jordan? Evidence from Ordinary Least Square Models

The main objective of this paper is to analyze equilibrium and dynamic causality relationships between monetary policy tools and economic growth in Jordan for the period (1990-2017). For this purpose, it considers the autoregressive distributed lag (ARDL) and vector error correction (VEC) models estimations. The results of ARDL approach show that monetary policy variables (i.e., real interest rate and money supply) have positive impact on economic growth in long-run and short-run except inflation rate. In addition, the results of VECM indicate bidirectional causal relationships between economic growth and monetary policy variables in long-run and short-run.


Introduction
The aim of this paper is to examine equilibrium and dynamic causality relationships between monetary policy instruments and economic growth in Jordan. However, Jordan is an upper middle income developing kingdom with a population of 9.904 million (World Bank, 2018). Appendices A, B, C, and D show that economic growth rate (gross domestic product (GDP)), real interest rate (IR), money supply (M2), and inflation rate (IFR) registered an annual growth rate of 5%, 8.5%, 9.6%, and -2%, respectively. Recently, the government of Jordan has implemented Jordan economic growth plan 2018-2022 to stabilize the macroeconomic indicators. This will be achieved through a 5% growth rate (USD 1.8 billion of growth per year) in top five contributors sectors to GDP which are finance, government services, transport, manufacturing, and tourism & hospitality (Jordan Economic Growth Plan Report 2018-2022. Different studies examined the impact of monetary policy tools on economic growth. The results of some studies confirmed that there is no impact of monetary policy on economic growth (see for example Kamaan, 2014;Monteil et al., 2012;Lashkary & Kashani, 2011;Buigut, 2009). On the other hand, the findings of other studies confirmed that monetary policy is vital for economic growth (i.e., Havi & Enu, 2014;Fasanya et al., 2013;Kareem et al., 2013;Vinayagathasan, 2013;Chaudhry et al., 2012;Coibion, 2011;Amarasekara, 2009;Suleiman et al., 2009;Ali et al., 2008;Rafiq & Mallick, 2008;Khabo & Harmse, 2005).
The efficiency of monetary policy is a key of an energetic financial system. The larger the financial system, the more sensitive interest rate will be in production and cumulative demand. Thus, economic growth stability is based on monetary policy instruments (i.e., IR, M2, and IFR). The leading contribution of this study differs from other studies in Jordan (see for example Obeid & Awad, 2017;Soufan, 2013). That is, in this study, effects of the monetary policy on economic growth are investigated using vector error correction (VEC) and autoregressive distributed lag (ARDL) models. In addition, the hypotheses of equilibrium (long-run and short-run) relationships and dynamic causality analysis between monetary policy variables (i.e., IR, M2, and IFR) and economic growth are examined. In this study, after the introduction, data and empirical model are presented. Following methodology and results estimates, conclusions and policy recommendations are explained.

Data and Empirical Model
This paper analyses the equilibrium and dynamic causality relationships between economic growth and monetary policy instruments in Jordan, particularly interest rate, money supply and level of inflation. Due to the availability of data, the range of the analysis will be annual time-series data for the  period, establishing a small sample size of 28 observations. The GDP in local currency unit (i.e., Billions JD) is used to measure economic growth in Jordan.
Interest rate (IR) is a major monetary policy instrument and measured by real interest rate. The money supply is measured by the broad monetary aggregate (M2, Billions JD) as a proxy of the total monetary aggregates available in the Jordan's economy. Inflation rate (IFR) refers to consumer prices (annual %). Data used in this study is taken from the World Bank (2018). The variables are transformed into a natural logarithmic form (i.e., LogGDP, LogIR, LogM2, and LogIFR).

Unit Root, Co-integration, and Equilibrium Relationships
This paper employs the ARDL approach proposed by Pesaran et al. (2001) to examine equilibrium relationships between monetary policy variables (i.e., IR, M2, and IFR) and GDP. Before estimating the ARDL approach, it is necessary to test unit root and co-integration. Table 1 shows that the variables are stationary at the first difference (i.e., I(1)). The advantage of the ARDL approach is that it allows the nonlinear relationship among the GDP, the IR, the M2, and the IFR. The model is specified as follows: Where ∆ is a lag operator; α 0 is constant; ε t is assumed to be N (0, ζ 2 ); α 1 α 2 α 3 α 4 are for long-run estimates, while other coefficients: α 5 α 6 α 7 α 8 are for short-run estimates. For ARDL approach, the first step is to test Eq. (2) by ordinary least square. The null hypothesis of no co-integration is H 0 : α 1 = α 2 = α 3 = α 4 = 0 against alternative H 1 : α 1 ≠ α 2 ≠ α 3 ≠ α 4 ≠ 0. The bounds F-statistics test is used to check the existence of co-integration among variables. Critical values for F-statistics test are given by Pesaran and Pesaran (2009). Pesaran et al. (2001) argued that if the calculated F-statistics value is higher than the upper bound, i.e., I(1), then, the H 0 of no co-integration is rejected and concludes the existence of co-integration. If the calculated F-statistics value is below the lower bound, i.e., I(0), then, the H 0 of no co-integration is accepted. Thus, there are no long-run relationships among variables. If the calculated F-statistics value is between I(0) and I(1), then results are unsettled.
The vector autoregressive (VAR) model is used to find lag order that was two through Akaike information criterion and Schwarz Bayesian criterion. Table 2 shows the results of co-integration test. The calculated F-statistics value (5.20) is higher than the upper bound (5.12) at 1% significance level, showing a co-integration between economic growth and monetary policy instruments. The result of co-integration is consistent with result obtained for Nigeria (Sulaiman & Migiro, 2014). (1) *, **, *** represent the significance at 1%, 5%, and 10% levels, respectively.
(2) The analysis of unit root test is conducted using Augmented Dickey-Fuller test.
(3) Source: Author's estimations using E-views software package 9.0.  Once co-integration is well-known, the next step is to estimate equilibrium relationships between economic growth and monetary policy variables. In Table 3, the long-run and short-run results show that all variables have positive impact on economic growth except inflation rate. These results are similar to the results obtained for Vietnam (Anwar and Nguyen, 2018), Canada (Champagne and Sekkel, 2018), and Turkey (Varlik and Berument, 2017). Notes.
Inflation is a double-edge sword, that it can be beneficial to economic growth but at the same time, it may adversary affect. Thus, it's necessary to follow the strategy of moderation to control inflation in order to have a positive effect on economic growth. The mechanisms used by central bank are money supply and interest rate. That is, money supply is the main economic engine, while interest rate is considered an important tool to motivate investors. Results of diagnostic tests in Table 4 confirm that Eq. (2) is free from heteroskedasticity problem, correctly specified, and residuals are normally distributed.

The VECM Granger Causality
Once co-integration is confirmed among variables, it is necessary to employ the VECM instead of VAR model. The negative and significant error correction term (Ect) shows long-run causal relationships, while the F-statistics test illustrates the short-run causality. Eq.

Conclusions and Policy Recommendations
The current paper analyzes equilibrium and dynamic causality relationships between economic growth and monetary policy variables (i.e., real interest rate, money supply, and inflation rate). It employs ARDL and VEC models over the  period. The results show the existence of co-integration between economic growth and monetary policy variables. While the results of ARDL approach show that monetary policy variables have positive impact on economic growth in long-run and short-run except inflation rate. In addition, the results of VECM indicate bidirectional causal relationships between economic growth and monetary policy variables in long-run and short-run. A rising from these findings, this paper recommends the following:


The growth of the economy should be the topmost consideration when implementing monetary policy. Strong macroeconomic policies (i.e., Jordan economic growth plan 2018-2022) should be followed to stabilize the economy.


The financial sector in Jordan should be more regulated and supervised by ministry of finance and central bank in order to achieve the efficiency of monetary policies.