The Basis for Balances of Payments - Research Paper Example

Assignment March Executive Summary Economic theory in most cases is used in order to come up with tangible policies to propel a country to sustainable development. In most cases, modeling is used to come up such policies in cases were data is not always available. In this case, GDP growth is influenced by many factors which include exports and imports which form the basis for balances of payments. The aim of the research paper was to develop a model to enable in the forecasting of GDP if the other variables are known in addition to determining its (GDP) response to changes in one of more of the variables. The model had GDP as the dependent variable and energy production (kt of oil equivalent); exports of goods and services (% of GDP); gross savings (% of GDP); imports of goods and services (% of GDP); labour participation rate, total (% of total population ages 15+); life expectancy at birth, total (years); market capitalization of listed companies (% of GDP); real interest rate (%); research and development expenditure (% of GDP); stocks traded, total value (% of GDP); Bulgaria dummy variable; Hungary dummy variable; Lithuania dummy variable; year 1995 dummy variable; 1996D; 1997D; 1998D; 1999D; 2000D; 2001D; 2002D; 2003D; 2004D; 2005D; 2006D; 200D7 and 2008D as the independent variables. The model results demonstrated that a percentage increase in energy production (kt of oil equivalent), ceteris paribus, leads to 0.0003 percentage increase in GDP while a percentage increase in the average of exports and imports of goods and services (% of GDP), ceteris paribus, causes the GDP to grow by 0.017 percent. Further, a percentage increase in gross savings (% of GDP), ceteris paribus, cause the GDP to grow by 0.77 percent while an increase of one percent in labor participation rate total (% of total population ages 15), ceteris paribus, causes a decrease in GDP growth of 0.35 percent. A percentage increase in life expectancy at birth, total (years), ceteris paribus, causes the GDP to increase by 2.9 percent while market capitalization of listed companies (% of GDP) increase by one percent causes a 0.21 percentage increase in GDP. Real interest rate (%) increase by one percent causes a GDP decrease of 0.06 percent while an increase in research and development expenditure (% of GDP) by one percent decreases GDP growth by 0.62 percent. An increase of one percent in stocks traded, total value (% of GDP) causes the GDP to decrease by 0.02 percent. Further, the GDP for Bulgaria was found to be higher than that for Poland by 0.32 percent while that for Hungary was found to be higher than that for Poland by 0.25%. The GDP growth for Lithuania was found to be higher than that for Poland by 0.4 percent. Using 1995 as the base year, the results further indicated that the GDP for the four countries was better in 1996 as compared to 1996 by 0.237 percent while 1997 was better than 1995 by 0.206 percent. GDP in 1998 was better by 0.226, 1999 by 0.206, 2000 by 0.219, 2001 by 0.04, 2002 by 0.196, 2003 by 0.208, 2004 by 0.185, 2005 by 0.163, 2006 by 0.149, 2007 by 0.154 and 2008 by 0.167. The model was found to be effective and could be used to forecast GDP values if the independent variables are known. Statement of the Problem Economic theory in most cases is used in order to come up with tangible policies to propel a country to sustainable development. In most cases, modeling is used to come up such policies in cases were data is not always available. In this case, GDP growth is determined/influenced by many factors which include exports and imports which form the basis for balances of payments. In this regard, there is no model which can determine the impacts of GDP and relative differences in GDP and thus we could run a regression model with the various variables and several dummies controlling for other variables in order to determine their influence in GDP growth. Such a model will enable in the forecasting of GDP if the other variables are known in addition to determining how GDP responds to changes in one of more of the variables. The GDP responses will be different for different countries as consumers in different countries have different tastes and preferences compared to those in other countries. Economic Theory The model to be estimated is Y(GDP)= Energy production (kt of oil equivalent) +Exports of goods and services (% of GDP)+ Gross savings (% of GDP)+ Imports of goods and services (% of GDP)+ Labor participation rate, total (% of total population ages 15+)+Life expectancy at birth, total (years)+ Market capitalization of listed companies (% of GDP)+ Real interest rate (%)+Research and development expenditure (% of GDP)+ Stocks traded, total value (% of GDP)+BD (Bulgaria dummy variable)+ HD (Hungary dummy variable)+ LD (Lithuania dummy variable)+ 1995D (year 1995 dummy variable) + 1996D +1997D +1998D+ 1999D+2000D+2001D+2002D+2003D+2004D+2005D. In the model, dummy variables are included. They are used in this model to represent sub-categories in the study. Generally, dummy variables take the values 1 or 0 depending on the researcher. According to Koop (2000), dummy variables assist in regression analysis since they give a single regression equation and represent many sub-categories. Miles and Shevlin (2001) sums up by saying that dummy variables make a researcher obtain meaningful results. For example in this case, the indicator dummy variables will give the performance of the GDP relative to that of one of the model years. This means that the researcher will be able to tell the performance of the other years as compared to the dummy year i.e. 1995. The model has GDP as the dependent variable and energy production (kt of oil equivalent); exports of goods and services (% of GDP); gross savings (% of GDP); imports of goods and services (% of GDP); labour participation rate, total (% of total population ages 15+); life expectancy at birth, total (years); market capitalization of listed companies (% of GDP); real interest rate (%); research and development expenditure (% of GDP); stocks traded, total value (% of GDP); Bulgaria dummy variable; Hungary dummy variable; Lithuania dummy variable; year 1995 dummy variable; 1996D; 1997D; 1998D; 1999D; 2000D; 2001D; 2002D; 2003D; 2004D; 2005D; 2006D; 200D7 and 2008D as the independent variables. The model will be fitted and then tests to confirm regression analysis assumptions1. This included Homoskedasticity, autocorrelation and endogenous tests. The model employs the use of panel data, data followed up for a period of time, which provides several observations making comparisons easier. Panel data gives the research more points and increases statistical degrees of freedom thus reducing collinearity in the explanatory variables. Further, panel data takes care of un-observed heterogeneity. However, panel data is seriously affected by effects of attrition. Energy is a key factor of production and high energy production means that production is cheaper and thus more investors, more jobs and more exports. The rate at which a country bridges the gap between imports and exports gives a clear direction where the economy is going. Such balances are maintained if exports surpass the imports in stable population. Gross savings on the other hand translates to investments which drive the cost of loanable goods low creating an avenue for borrowing and thus more investments. Population growth has a bearing on labour participation rates. If the age specific death rates are high between 15 and 65 years, then the labour force is not able to meet the demand. The strength of the capital market is another key indicator in the determination of the best model since it has a bearing on the real interest rates. Research and development expenditure as compared to the GDP is key. Major economies are build on research and the recognition of innovations and awarding them hence giving the country and its citizens a reason to be innovative. Innovations, research and technological advancement affords a country competitive edge which translates to more exports as compared to imports and thus positive balance of payment. An economy consists of a supply side and a demand side according to Damodar (2003). An economy changes overtime depending on the factors of production and the transition paths the country has followed over time. Peoples’ consumption power and saving rates are very important determinants of growth. This is so because, more consumption in inelastic goods leads to more revenues as the consumption is not affected by prices while more savings translates to more investments thus more jobs and more exports meaning that the balance of payment is made positive. Another very important factor is population growth. High population growth means that the capital stock is spread out to many individual thinly leading to negative accumulation of stock and thus reduced growth. At the same time, it is assumed that high population growth translates to more labour force. In a steady economic growth with a steady population growth, the development curves are in such a way that they show steadiness. However, with population growth, the steadiness curves moves negatively. According to Wooldridge (2003), the Solow model intuitively predicts that economies with high saving rates, those with steady population growth rates and those that have low depreciation rates are likely to develop at a faster rate. Results The model to be estimated is; MODEL: Y(GDP)= Energy production (kt of oil equivalent) +Exports of goods and services (% of GDP)+ Gross savings (% of GDP)+ Imports of goods and services (% of GDP)+ Labor participation rate, total (% of total population ages 15+)+Life expectancy at birth, total (years)+ Market capitalization of listed companies (% of GDP)+ Real interest rate (%)+Research and development expenditure (% of GDP)+ Stocks traded, total value (% of GDP)+BD (Bulgaria dummy variable)+ HD (Hungary dummy variable)+ LD (Lithuania dummy variable)+ 1995D (year 1995 dummy variable) + 1996D +1997D +1998D+ 1999D+2000D+2001D+2002D+2003D+2004D+2005D. Heteroskedasticity Test The hettest test was used to test for constant variance which is a condition set in regression analysis of the results are to be valid and reliable (Verbeek 2008; Baum (2006). The results indicate that, the variances are not constant. Breusch-Pagan / Cook-Weisberg test for heteroskedasticity with the null hypothesis as Ho: Constant variance [Chi-Square (1) = 15.96; p = 0.0001; p10 while the mean VIF is 29.46 and some tolerance values are less than 0.1. This means that some of the model variables are highly correlated. Extreme multi-collinearity violates the assumption made on ordinary least squares. Huge multi-collinearity translates to large standard errors making confidence intervals to very wide. In this case, type II error occurs where the null is accepted when it is actually false. In order to correct, the collinearity menace, then correlations values of the independent variables are sought and the highly correlated variables are combined or eliminated. Outlier Analysis The model has outliers as observed from the above box plot. These are some of the causes of heterogeneity. According to Dougherty, C. (2002), endogenous test are used to determine difference in contribution to the model between the two variables i.e. exports of goods and services (as a % of GDP) and imports of goods and services (as a % of GDP), the results indicate that the two are not significantly different. [egs - igs = 0; F (1, 32) = 3.79; Prob > F = 0.0604]. ( 1) ep - egs = 0 ( 2) ep - gs = 0 ( 3) ep - igs = 0 ( 4) ep - lpr = 0 ( 5) ep - le = 0 ( 6) ep - mc = 0 ( 7) ep - ri = 0 ( 8) ep - rde = 0 ( 9) ep - st = 0 F( 9, 49) = 3.56 Prob > F = 0.0018 Endogenous tests for all the independent variables shows that they are different [F (9, 49) = 3.56, p = 0.0018, p Read More