Capital formation in transition economies: empirical evidence on the effect of openness and other determinants.
Capital has long been recognized for its main contribution to the production process at large, enhancing the productivity of other factors of production, being instrumental in the growth of an economy, giving meaning to the technological process and advancement. Many studies have been conducted to explore these influences capital has on economic performance. In this study the focus is to empirically estimate determinants of capital formation in countries in transition, especially those in Central and Eastern Europe. These countries have been working for almost two decades toward market economy. They adopted market led mechanisms, after a long (more than half a century) "experiment" with central planning type of command and control mechanism, for allocating countries resources. Therefore the role of openness in capital formation is considered to be of prime interest. Among chief determinants are rate of interest, demand for capital goods, expectation about economic growth and economic progress, available credit to private firms, openness of the economy toward foreign capital, and business practices.
The purpose of this study is threefold. First, I would like to test the theoretical findings with regard to capital formation. Second, the findings of the paper will be contrasted and elaborated for possible policy prescriptions. Third, the paper aims adding to the existence stock of knowledge about the determinants of capital formation. Conventional (what capital formation theory suggests) and non-conventional (institutional factors) determinants and their effect on the capital formation are estimated. Among the conventional factors rate of interest, economic growth, and saving have been analyzed and included in the statistical work. While inflation rate, openness of country, and credit availability to private sector are considered to be more of institutional factors.
It is critical that countries with low capital per worker ratios, like those in transition on which this study is written, sustain economic growth. Economic growth enables these economies to increase total domestic production, output per capita, transforming economic landscape, and more importantly making progress toward market oriented economy. Physical capital that countries acquire via spending on capital goods enhances workers' ability to increase their productivities; therefore, making a substantial contribution on economic growth (Romer, 2007). This is suggested by neoclassical economic theory as well as by countless writings in economic development and growth. The role of capital formation in development and progress cannot be ignored on at least two counts. First capital per worker is crucial and it directly affects production in general as well as the economy's output growth. Second, more capital brings with it new technology, new techniques, and knowledge. In this way importance is enhanced via technological progress (Kukeli, Fan, & Fan, 2006).
It is plausible to think that at a low level of capital per worker and capital endowment countries started transitioning (almost two decades ago) toward market economy marginal productivity of capital increases as the amount of physical capital via capital formation increases as well. This alone is not sufficient reason to justify the enormous importance of capital in these countries once one considers spillover effects and the virtual cycle of growth that it can bring to the economy (Blanchard, 2011).
Capital formation is very crucial in economic development as it not only increases labor productivity but it also contributes to economic growth and technological advancement of the country. New capital brings also new rules of game, making substantial contribution to the enhancement of market economy institutions. Capital in its classical form is defined to represent physical structures, buildings, tools, machineries, etc. (Kukeli, 2007).
This study is different from other studies done on capital formation in at least three ways. First, it looks specifically at a very special and unique region, that of central and Eastern Europe which has courageously taken draconian steps toward market economy after a long command a control resource allocation mechanisms. Second, it incorporates not only determinates known by the neoclassical models, such as the rate of interest, and saving rates, bur also looks at the institutional factors such as openness. Third, policy factors are considered recognizing that they play a very important role in enhancing economy's ability to absorb its own as well as foreign capital accumulation. Study makes use of cross-sectional combined over time data (version of panel data).
A number of previous studies have identified that capital formation is determined by growth of GDP (Blomstrow, Lipsey, & Zejan 1996), rate of return (Romer, 2007; Baswarth, Colins, & Reinhart, 1999), flow of FDI (Desay, Foley, & Hines, 2005; and Feldstein, 1995), saving rate, inflation, and expectation about economy's performance at large (Mishkin, 2010). From production function point of view the cost of capital relative to the cost of labor (and other conceivable used resources) should be an important factor determining the amount of capital formation in the economy.
Baswarth, Collins, and Reinhart (1999) have stressed the importance of interest rate in the economic activity. With reference to the mid-1997 Asian crises they have stated "Increase in interest rates after mid-1997 Asian currency crises has decreased economic activity." They argue "For many developing countries, the ability to draw upon an international pool of financial capital offers large potential benefits". The given convincing argument is "Low levels of capital/worker in these countries have long held the output down." They also find that FDI affects investment (via which capital formation takes place). The role of interest rate is crucial when it comes to the demand for capital goods. It also affects the level of GDP growth and the down spiral continues until economy recovers or policy actions are designed and implemented in such a way that will mitigate the recovery process.
Capital formation is linked with consumption--investment and saving investment decisions individuals and societies make in long run. These decisions are driven by known factors coming from inter-temporal choice as well as by expectations about the economy at large and policy agenda. This is in line with what Baswarth, Solow, and Summers (1982) suggests that "A program to expand domestic capital formation requires attention to all aspects of the saving-investment decision, including expectation of future growth in demand, interest rates, and tax policy".
The role of foreign capital in domestic capital formation is crucial in domestic capital formation. It is Desai, Foley, and Hines (2005) concludes that "... greater foreign investment is associated with higher level of domestic investment. This estimated complementarity implies that firms combine home production with foreign production to generate final output at lower cost than would be possible with production in just one country." This effect is justified not only with augmentation of output from adding foreign capital into the production process but also from the augmentation of productivity of domestic capital as well.
Soundness of government policies is detrimental in long-run planes to add capital to the production process via saving and investments. A Hall (1980) State that estimates policies would effect capital formation as well as level of output. He recognizes the effect that monetary expansion would have had on capital formation in the presence of inflation when he asserts: "... stimulus to capital formation from monetary expansion offsets part of the inflationary influence of the extra money".
Nabende and Slater (2003) have studied the short-run and long-run effects of determinants of private investments and have made recommendations for policy makers. Their sample includes ASEAN (Association of Southeast Asian Nations) countries for twenty nine years. They conclude that output is chief determinant in private investments whereas monetary policy variables are less effective.
Institutions have indispensable role in reducing the risk of economic entities and the society at large (North, 1990). Transition economies went through institutional set-up and adaptation that will effect a great deal capital formation. Ruggles (1993) calls for taking into account institutional aspects of modern economic system as the process of saving is not adequate in explaining capital formation. There are at least three variables in our model that capture both intuitional state of the economies and the soundness of policy, namely, credit availability to the private sector and inflation level (proxy for price and economic stability.
THEORETICAL MODEL AND HYPOTHESES
This paper follows the widely held position that the economy's production function is of Cobb-Douglas. From the production function it can be extrapolated that capital formation is the cumulative function of factors influencing the size of spending on these kinds of goods that will be used to produce consumable goods. Therefore the capital accumulation is a product of factors like openness, credit available to the economy, growth rate of the economy's GDP, saving rate and the level of inflation. In short, both conventional (neoclassical production function) and non conventional (those from institutional view of decision making process) factors will be responsible for changes in the capital formation. The main objective is to test whether fluctuations in the gross capital formation (as a percentage of GDP) are driven by conventional and non-conventional factors mentioned earlier in this paper. Therefore the equation to be evaluated for this purpose is:
k = f(c, g, s, m, p, r),
k= gross capital formation as percentage of GDP, c= domestic credit to private sector as a percentage of GDP, g= annual percentage growth rate of GDP, s= gross saving rate, m= volume of imports to GDP, used as a proxy for the index of openness, p= level of annual inflation, r= real interest rate.
The main hypotheses to be tested are:
H1: Gross capital formation is positively and significantly effected by the domestic credit to private sector.
Availability of credit to the private sector is crucial in enabling entrepreneurs to acquire capital. In transition economies international lines of credit are difficult to be obtained by countries governments and much more difficult if impossible for small businesses which need to grow. It would be interesting to see the behavior of data in this case.
H2: Within our theoretical framework the claim is that increase in GDP will positively and significantly affect gross capital formation.
Studies have shown (Boswarth, Solow, and Summers, 1982) that capital formation has affected positively the rate of economic growth. In this instance, the GDP was used as a proxy for demand for capital goods as well as other goods. Therefore a positive relationship is expected to be revealed by the data analysis.
H3: There is a positive and significantly relationship between capital formation and gross saving rate.
Expectations are that a high level of saving rate will increase the amount of recourses that households and firms could devote to capital formation. At a higher rate of saving the interest rates should decrease and that would increase the amount of funds to be used to buy capital goods.
H4: There is a positive and significant correlation between the volume of import to GDP, the proxy for the index of openness, and the capital formation.
Testing this relationship is important as if it holds in this specific scenario then it would be interesting for both policymakers and economists to understand the transmission mechanism of such an effect. As the volume of imports increases, via free of tariff entry of new capital (proxy for the index of openness) increases the capital formation should increase, hence a positive effect.
H5: There is a negative and significant relationship between capital formation and the level of annual inflation.
There should be noted that inflation may transmit its effect indirectly via real interest rate, however we will first test for this negative relationship and then might skip interest rate from the equation and test only the effect of inflation (call it Tobin's effect, Feldstein, 1980). This is consistent with Hall's and Feldstein assertion (Hall 1980, Feldstein 1980).
H6: There is a negative and significant relationship between interest rate and capital formation.
The authors will include real interest rate in the regression equation along with annual inflation first and then alone without inflation.
Sample and Data Collection
Data are collected for 13 countries of Central and East Europe (Albania, Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Hungary, Macedonia (FYR), Montenegro, Poland, Romania, Serbia, Slovak Republic, and Slovenia). The reason for bringing together in a study of this nature is justified based on the common background: experience of socialism and transition from command and control to market led economy that all of them has undergone now for close to two decades. Getting data on national level is very difficult if impossible due to different accounting and statistical practices these countries follow, use of different currencies, and reliability.
Therefore relying on the data from World Bank Development Indicators (World Bank, 2011) has been the only solution. Data are attempted to be used for a relatively long period of time from 1990 to 2010; 21 years (this is the time period where data are available for all countries and most of the variables used). This period of time is valid if and only if we use unbalanced panel data method. While doing this would increase the size of the sample and we will capture the full effect of determinants with more information, it has the disadvantage that missing data for specific country is just as taking the whole country away from the model.
Due to the fact that data are over time and small countries like Albania, Slovenia are included in the same sample with relatively large countries like Czech Republic and Hungary the use of data normalized by countries' GDP takes care of heteroskedasticty. Another feature of this study has to do with the fact that most of the variables used in the model are measured as the fraction to GDP in percentages. This technique is highly suggested (Baltagi, 2001; Dielman, 1989) to cope with heteroskedasticity when dealing with sections (countries) of different size of GDP level and consequently of other variables. A regression analysis for the data (pooled cross sectional balanced data) is carried out and the output is given below.
The variables used are the following: domestic credit to private sector (c), GDP growth rate (g), gross capital formation (k), gross saving (s), ratio of total import to GDP as proxy for openness (m), inflation (p), and real interest rate (r). All variables except inflation and real interest rate are expressed as a percentage of GDP. Inflation and interest rates are in percentages measured annually.
RESULTS OF THIS STUDY
The descriptive statistics are given in Table 1.
A correlation matrix is obtained for all variables in the model (Table 2). It reveals a strong relationship between capital formation and saving rate, capital formation and openness toward imports, capital formation and growth, credit availability to the economy, and inflation rate. It shows very small to no relationship between capital formation and real interest rate. This is not surprising in literature but nevertheless it needs to be further explored and explanations as to why there is no correlation should be provided.
The first regression run is to estimate the following equation:
k = [b.sub.0] + [b.sub.1]c + [b.sub.2]g + [b.sub.3]s + [b.sub.4]m + [b.sub.5]p + [b.sub.6]r + e.
k = 12.83+ 0.095c + 0.44g + 0.11s + 0.086m - 0.0067p + 0.0093 r + e. (1)
The output of this regression, presented in table 3, is characterized by adjusted R square of 0.42, suggesting that 42 percent of variation in capital formation is explained by variations in all other variables included into the model. The signs of coefficients are as expected for all independent variables. Coefficients are statistically different from zero (at least at 5 percent significance level), suggesting their effect on capital formation, except for real interest rate.
Looking at coefficients well interesting results need to be noticed. For example an increase by one percent of domestic credit to private sector will lead to seven percent increase in capital formation. Growth rate of GDP and saving rate have a large impact on capital formation, as well as significant effect on it has openness.
Why it is that the interest rate is not influencing the rate at which capital is formed through investment? This is a very unusual result and needs some explanations. One way to explain it is to think in terms of nominal interest rate as being eroded by the large fluctuations in the price levels. It could be suggested that interest rate becomes a redundant variable when price-level (inflation) is included into the regression analysis.
An expected thing to do to deal with the fact that interest rate not only is statistically not significant but also the marginal effect is very small, is to exclude it from the regression analysis. Therefore a second regression is run and the results are presented in table 4. The following estimated equation is obtained:
k = 12.43 + 0.077c + 0.36g + 0.10s + 0.11m - 0.0069p + e. (2)
Results are reported in Table 4. The striking results show that adjusted R square does not change (0.42). That means that the power of the regression did not diminish by taking out real interest rate factor. Also, coefficients have slightly to no changes compared to regression with real interest rate as an independent variable. At the same time it is noticeable that t-stats have not changed.
All of these (adjusted R square, coefficients, and t-stats) direct our attention to the fact that real interest rate should not have been in the regression model. One might argue that in this case the real interest rate could be viewed as a redundant variable. There are at least three statistics that have not changed from exclusion of interest rate from the regression equation: adjusted R square, t-statistics, the coefficients and their signs.
This study has several implications pertaining to policy prescription for countries that undergo changes in the economic system: from command and control to price system, market economy. From the results of this study one can infer that it is good policy to increase the credit availability to private entrepreneurs as it directly will increase the capital formation. Equally important is to encourage the saving from both public as well as private sector and to employ tax free imports on capital goods. From the direct soundness of policy, maintaining stable prices (keeping inflation under control) will directly affect capital formation.
LIMITATIONS OF THIS STUDY
The limitations of this study rest on the ad hock model used to estimate the determinants of capital formation. Future studies can be undertaken to employ other model(s) that comes from grounded theory. This study has another limitation that has to do with data. For empirical purposes and constrained from the estimation techniques we used aggregated data. It would be interested to estimate and further explore capital formation at country level. This will be possible with the improvement on the data and their availability.
This paper has empirically explored the effect of several variables into the capital formation in transition economies. Cross-sectional, time series data from former transition economies are employed. Among factors included in our analysis, only real interest rate has no statistical significance. Robustness of the results is maintained when real interest rate is excluded from the model. Growth rate of GDP, saving rate, credit availability to the private sector, openness of the economy and price level are all important factors that drive capital formation in these countries.
Policy makers should maintain fiscal and monetary actions that aim at sustaining the positive growth rate of the GDP, increase credit availability to the private sector, keep stable prices, promote saving from consumers in the economy, and encourage free trade with the rest of the world.
Baltagi, B. H. (2001). Econometric Analysis of Panel Data, John Wiley & Sons. Blanchard, O. (2011). Macoeconomics. Boston: Pearson
Blomstrom, M., Lipsey, R., & Zejan, M. (1996). Is fixed Investment the Key to Economic Growth? The Quarterly Journal of Economics, 11, (1), 269-276.
Boswarth, B., Solow, R., & Summers, L. (1982). Capital Formation and Economic Policy. Brooking Papers on Economic Activity, No. 2, 273-326.
Boswarth, B., Collins, S., & Reinhart, C. (1999). Capital Flows to Developing Economies: Implications for Saving and Investment. Brooking Papers on Economic Activity, No. 1. 143-180.
Desai, M., Foley, C., & Hines Jr., J. (2005). Foreign Direct Investment and the Domestic Capital Stock. AEA Papers and Proceedings, 95, (2) 33-38.
Dielman, T. (1989). Pooled Cross-Sectional and Time Series Data Analysis. Marcel Dekker, Inc., New York.
Feldstein, M. (1980). Fiscal Policies, Inflation, and Capital Formation. The American Economic Review, 70, (4) 636-650.
Hall, R. (1980). Stabilization Policy and Capital Formation. The American Economic Review, 70. (2) 157-163.
Kukeli, A., Fan, L., & Fan, C. (2006). FDI and Growth in Transition Economies: Does the Mode of Transition Make a Difference. International Review of Economics, Springer, vol. 53(3), 302-322
Kukeli, A. (2007). Foreign Direct Investment in Transition Economies: An Institutional Approach. International Business and Economics Research Journal, 6, (12) 53-61.
Mishkin, F. S. (2010). The Economics of Money, Banking & Financial Markets. Boston: Addison-Wesley.
Nabende-Bende, A., & Slater, J. (2003). Private Capital Formation: Short-run and Long-run Crowding-in (out) Effects in ASEAN, 1971-99. Economics Bulletin, 3, (27) 1-16.
North, D. (1990). Institutions, Institutional Change and Economic Performance. New York: Cambridge University Press.
Romer, D. (2007). Advanced Macroeconomics. McGraw-Hill, New York, NY
Ruggles, R. (1993). Accounting for Saving and Capital Formation in the United States, 1947-1991. Journal of economic Perspectives, 7, (2) pp. 3-17.
World Bank. (2011). World Development Indicators, Washington D.C.
University of New York Tirana
About the Author:
Agim Kukeli holds an MA and PhD in Economics from Colorado State University (Fort Collins, Colorado, USA). He has been teaching instructor for three academic years at Colorado State University and two academic years as Visiting and Assistant Professor at Colorado Mesa University (Colorado, USA). He has served as Founding Rector (President) of University "A. Moisiu" Durres (Albania), Professor of Economics and Dean at University of New York in Tirana (Albania) where he is currently. He has published articles on FDI and economic growth with special emphasis on transition economies and has lectured in MBA program in Prague, Czech Republic and State University of Tetovo, FYR of Macedonia.
Table 1 Summary Statistics, using the observations 1:01-13:21 (missing values were skipped) Variable Mean Median Minimum Maximum Std. Dev. k 23.2944 23.5964 5.19999 42.0000 6.00440 c 37.7173 35.2016 3.49767 94.3816 20.2191 g 2.39388 3.85042 -30.5085 88.9577 8.69347 s 12.7700 17.1062 -71.8218 31.9867 14.9092 m 54.1051 54.0266 19.6519 98.3634 17.0699 p 40.7703 6.65285 -1.27929 1500.00 143.892 r 7.05350 5.79900 -71.2053 374.309 29.7717 Variable C.V. Skewness Kurtosis k 0.257762 -0.031482 0.587349 c 0.536071 0.491214 -0.329076 g 3.63153 3.25839 39.4296 s 1.16752 -1.76851 4.72729 m 0.315494 0.156770 -0.568437 p 3.52934 7.20234 60.1918 r 4.22084 8.99468 109.663 Table 2 Correlation coefficients, 5% critical value (two-tailed) = 0.1187 for n = 273 k c g s m p r 1.0000 0.4311 0.3640 0.2801 0.3273 -0.2840 -0.0022 k 1.0000 -0.1777 0.2134 0.4492 -0.0368 -0.0050 c 1.0000 -0.2100 0.2361 -0.3325 -0.1501 g 1.0000 -0.1805 -0.0191 0.0848 s 1.0000 -0.1061 0.0810 m 1.0000 -0.2124 p 1.0000 r Table 3 Regression results I Fixed-effects, using 194 observations; Included 13 cross-sectional units Dependent variable: k Coefficient Std. Error t-ratio p-value const 12.8336 1.17983 10.8775 <0.00001 c 0.0951266 0.0195021 4.8778 <0.00001 g 0.435518 0.0872726 4.9903 <0.00001 s 0.104104 0.0295016 3.5288 0.00053 m 0.0861154 0.0242381 3.5529 0.00049 p -0.00671629 0.00261219 -2.5711 0.01097 r -0.00926178 0.0270239 -0.3427 0.73222 Adjusted R-squared 0.423050 Table 4 Regression results II Fixed-effects, using 202 observations; Included 13 cross-sectional units Dependent variable: k Coefficient Std. Error t-ratio p-value const 12.4254 1.114 11.1538 <0.00001 c 0.0773548 0.0186429 4.1493 0.00005 g 0.35851 0.0807546 4.4395 0.00002 s 0.101551 0.0291738 3.4809 0.00062 m 0.108306 0.0233939 4.6297 <0.00001 p -0.00694408 0.00241471 -2.8757 0.00451 Adjusted R-squared 0.414421
|Printer friendly Cite/link Email Feedback|
|Publication:||International Journal of Business and Economics Perspectives (IJBEP)|
|Date:||Mar 22, 2012|
|Previous Article:||Money neutrality/non-neutrality controversy: empirical extension based on the Nigerian evidence.|
|Next Article:||International trade and labor productivity in U.S. manufacturing industries.|