European Journal of Economic Studies, 2017, 6(2)

EUROPEAN of Economic

Journal Studies

Has been issued since 2012. ISSN 2304-9669. E-ISSN 2305-6282 2017, 6(2). Issued 2 times a year

EDITORIAL BOARD

Dr. Vidishcheva Evgeniya – Sochi State University, Sochi, Russian Federation (Editor-in-Chief) Dr. Simonyan Garnik – Scientific Research Centre of the Russian Academy of Sciences, Sochi, Russian Federation Dr. Levchenko Tatyana – Sochi State University, Sochi, Russian Federation Dr. Tarakanov Vasilii – Volgograd State University, Volgograd, Russian Federation Dr. Balatsky Evgeny – Central and Mathematics Institute (RAS), Moscow, Russian Federation Dr. Dinh Tran Ngoc Huy – Banking University HCMC Viet Nam – GSIM, International University of Japan, Japan Dr. Gerasimenko Viktor – Odessa State Economic University, Odessa, Ukraine Dr. Gvarliani Tatjana - – Sochi State University, Sochi, Russian Federation Dr. Gunare Marina – Baltic International Academy, Riga, Latvia Dr. Kryshtanovskaya Olga – Institute of Sociology of the Russian Academy of Sciences, Moscow, Russian Federation Dr. Minakir Pavel – Economic Research Institute of the Far Eastern Branch Russian Academy of Sciences, Khabarovsk, Russian Federation Dr. Papava Vladimir – Ivane Javakhishvili Tbilisi State University, Tbilisi, Georgia Dr. Prokopenko Olga – Sumy State University, Sumy, Ukraine Dr. Vishnevsky Valentine – Institute of Industrial Economics of the National Academy of Sciences of Ukraine, Donetsk, Ukraine ournalEconomic of Studies J

Journal is indexed by: CrossRef (UK), EBSCOhost Electronic Journals (USA), Electronic scientific library (Russia), 1 (Australia), Open Academic Journals (Russia), ResearchBib (Japan), ULRICH’s WEB (USA).

uropean

All manuscripts are peer reviewed by experts in the respective field. Authors of E the manuscripts bear responsibility for their content, credibility and reliability. Editorial board doesn’t expect the manuscripts’ authors to always agree with its opinion. 2017

Postal Address: 1367/4, Stara Vajnorska Release date 01.09.17. str., Bratislava, Slovakia, Nove Mesto, Format 21  29,7/4. 831 04 Is. Website: http://ejournal2.com/ Headset Georgia. 2 E-mail: [email protected]

Founder and Editor: Academic Publishing Order № 112

House Researcher s.r.o. А

© European Journal of Economic Studies, 2017

75

European Journal of Economic Studies, 2017, 6(2)

EUROPEAN of Economic

Journal Studies

Издается с 2012 г. ISSN 2304-9669. E-ISSN 2305-6282 2017, 6(2). Выходит 2 раза в год.

РЕДАКЦИОННЫЙ СОВЕТ

Видищева Евгения – Сочинский государственный университет, Сочи, Российская Федерация (Гл. редактор) Левченко Татьяна – Сочинский государственный университет, Сочи, Российская Федерация Симонян Гарник – Сочинский научно-исследовательский центр Российской академии наук, Сочи, Российская Федерация Тараканов Василий – Волгоградский государственный университет, Волгоград, Российская Федерация Балацкий Евгений – Центральный экономико-математический институт РАН, Москва, Российская Федерация Вишневский Валентин – Институт экономики промышленности Национальной академии наук Украины, Донецк, Украина Гварлиани Татьяна – Сочинский государственный университет, Сочи, Российская Федерация Герасименко Виктор – Одесский государственный экономический университет, Одесса, Украина Гунаре Марина – Балтийская международная академия, Рига, Латвия Динь Чан Нгок Хай – Банковский университет Хошимин Вьетнам - GSIM, Международный университет Японии, Япония Минакир Павел – Институт экономических исследований ДВО РАН, Хабаровск, Российская Федерация Крыштановская Ольга – Институт социологии РАН, Москва, Российская Федерация Папава Владимир – Тбилисский государственный университет имени Иване Джавахишвили, Тбилиси, Грузия Прокопенко Ольга – Сумский государственный университет, Сумы, Украина

ournalEconomic of Studies

J Журнал индексируется в: CrossRef (Великобритания), EBSCOhost Electronic Journals Service (США), Global Impact Factor (Австралия), Научная электронная библиотека (Россия), Open Academic Journals Index (Россия), ResearchBib (Япония), ULRICH’s WEB (США).

Статьи, поступившие в редакцию, рецензируются. За достоверность сведений, изложенных в статьях, ответственность несут авторы публикаций. Мнение редакции может не совпадать с мнением авторов материалов.

uropean Адрес редакции: 831 04, Словакия, Дата выпуска 01.09.17

E г. Братислава, Нове Место, ул. Стара Формат 21  29,7/4. Вайнорска, 1367/4

Гарнитура Georgia. Сайт журнала: http://ejournal2.com/ E-mail: [email protected] 2017 А

Учредитель и издатель: Academic Publishing Заказ № 112 House Researcher s.r.o. Is. 2

© European Journal of Economic Studies, 2017

76 European Journal of Economic Studies, 2017, 6(2)

C O N T E N T S

Articles and Statements

Current Account Dynamics of Central European Countries K. Halil Arıç, M. Tuncay, S. Kun Sek ...... 78

Does Government Size Affect in Developing Countries? Evidence from Non-stationary Panel Data M. Cetin ...... 85

The Role of and its Targeting for Low-Income Countries (Lessons from Post-Communist Georgia) V. Charaia, V. Papava ...... 96

Study on Client-Satisfaction Factors in Construction Industry M. Duljevic, M. Poturak ...... 104

Enablers of Successful Knowledge Sharing Behavior: KMS, Environment and Motivation A. Özlen ...... 115

The Relationship between Short-Run Rate and its Economic Determinants: Consumer Index, Industrial Production Index, Consumption and Exchange Rate. An Empirical Research for the Four Most Developed Countries J. Rigas, G. Theodosiou, G. Rigas, A. Goulas …………………………………………………………. 124

77 European Journal of Economic Studies, 2017, 6(2)

Copyright © 2017 by Academic Publishing House Researcher s.r.o.

Published in Slovak Republic European Journal of Economic Studies Has been issued since 2012. ISSN: 2304-9669 E-ISSN: 2305-6282 2017, 6(2): 78-84

DOI: 10.13187/es.2017.6.78 www.ejournal2.com

Articles and Statements

Current Account Dynamics of Central European Countries

Kıvanç Halil Arıç a , * , Merve Tuncay a , Siok Kun Sek b a Cumhuriyet University, Sivas, Turkey b Universiti Sains Malaysia, Minden, Penang, Malaysia

Abstract "Current account" has been considered as an important variable in forecasting an economic crisis. Therefore; specifying the determinants of current account is a substantial topic for policy makers. The aim of this study is to examine the current account dynamics in the scope of Central European countries (Poland, Hungary, Czech Republic, Slovakia Republic and Croatia) during the 1997 – 2015. Panel data analysis was used in the methodology. According to the analysis results, growth of gross domestic product has no significant effect on current account. Real exchange rate, foreign direct and importation affect the current account negatively. However, exportation and government expenditure have positive effects on current account in Central European countries. Keywords: current account, European Countries, balanced panel data.

1. Introduction Current account represents the situation of macroeconomic policies and the behavior of economic agents. It is considered as an important variable in the international perspective. Current account cannot be considered as a target variable such as rate and inflation also it cannot be taken as a policy variable such as and supply. Mainly, current account has effects on the decision of lenders and borrowers in the global economy. If the current account deficit keeps up in a continuous path, it identifies the inadequate creditworthiness of a country in the global economy context. In this circumstance, the country may face the risk of bankruptcy (Hassan et al., 2015: 190). In the 1990s financial liberalization expanded in the global level and monetary integration process of Europe was started in 2001 as a of euro. After the implementation of the euro in the European countries, it has contributed to the credit expansion and a decrease in private , and these factors led to high current account deficits for some European countries (Brissimis et al., 2010). Monetary integration brings different implementations on and units labor costs in Europe and, therefore; current account positions diverge between European

* Corresponding author E-mail addresses: [email protected] (H.A. Arıç), [email protected] (M. Tuncay), [email protected] (S.K. Sek)

78 European Journal of Economic Studies, 2017, 6(2) countries. For instance Germany and some smaller northern European countries' economy policies generate current account surpluses, however western periphery, eastern and most of the southern countries of Europe exhibit current account deficits (Belke, Schnabl, 2013). Before the 2009 global economic crisis, there had been excessive current account deficits in the periphery of Europe. During the pre-crisis period, these deficit problems led to economic shrinking, depredation sovereign creditworthiness, and problems in banking systems in the periphery regions. Also the decline in aggregated demand and losses on foreign asset holdings in the periphery had negative effects on current account surplus in European countries. In this respect, the control of current account imbalances has been seen as a prior policy for European policymakers (Lane, Pels, 2012). The aim of this study is to examine the current account dynamics for Central European Countries (Poland, Hungary, Czech Republic, Slovakia Republic and Croatia) in the period of 1997- 2015. Panel data was used in the analysis process. There are several studies which examine the current account dynamics in the European countries. These studies selected numerous countries in the analysis process. However; this study focuses on a limited number of countries for the specific results for the Central Europe region. There are five main sections in this study. First section includes literature review, second section represents data and methodology, third section elaborates on the analysis process, fourth section infer analysis results and finally, fifth section is the conclusion of the study.

2. Literature Review There are numerous studies in the literature in the respect of current account dynamics. Some of these studies examine this subject for OECD countries (Gosse, Serranito, 2014; Cavdar and Aydin, 2015; Bertola and Prete, 2015; Karras, 2016) and some of them analyze different country groups (Erauskin, 2015; Kim, 2015; Martin, 2016; Tan et al. 2015; Moral-Benito and Roehn, 2016). In order to determine the scope of this study, we particularly examine current account dynamics literature for the European countries. Aristovnik (2006) investigated the current account dynamics for the Eastern Europe and the former Soviet Union countries in the period of 1992-2003 by using dynamic panel data analysis. He concludes that economic growth has a negative effect on current account balance. This result reflects that the economic growth is relevant with increasing of domestic investment instead of domestic . Public budget shocks move together with current account breakdown and this circumstance indicates the twin deficit conditions in the region. Rise in of the real exchange rate and deterioration of the terms of also affects the affect the current account balance negatively. Gehringer (2015) examines current account dynamics for all European countries, except Luxemburg, during the 1995 and 2010 period by using panel data method. He concludes that excessive private and public consumption cause current account deficits. Additionally, credit variable, growth of GDP per capita, real exchange rate and construction sector variables have negative effects on current account balance in European Union economies. Bollano and Ibrahimaj's (2015) study on the current account dynamics of Central and Eastern European countries in the period of 2015:1 to 2014:4 by using panel data methodology. They find out that GDP growth and fiscal deficit have a negative effect on current account. However, depreciation of the real effective exchange rate affects current account positively. Zorzi et al. (2009) make a comprehensive survey about current account benchmarks for Central and Eastern European countries. They use external sustainability approach â la Lane and Milesi-Ferretti (LM) and structural current accounts literature (SCA) which is based on panel data methodology. According to LM approach they provide the importance of sensitivity of outcome to the external indebtedness and the consideration to exclude the foreign direct investment subcomponent from the foreign assets aggregate. In respect to SCA approach they analyze the sensitivity of outcome to various combinations of fundamentals. Brissimis et al. (2010) examined the determinants of current account for Greece during the 1960 to 2007 by using co-integration analysis in the long and the short run. They conclude that the current account balance could be established when the ratio of private sector financing to GDP counts as an indicator for financial liberalization in the model. Kang and Shambaugh (2016) also study current account deficits for countries in the Euro area and the Baltics which faced the global financial crisis with significant current account deficits.

79 European Journal of Economic Studies, 2017, 6(2)

Accordingly, large current account balances prior to the crisis is the best predictor of a sharp drop in output during the crisis. They suggest supportive macro policies to moderate the adjustment process and to keep overall euro inflation at or above target level are necessary. Kollmann et al. (2015) investigate the determinants of German’s current account surplus and its effects on Euro Area for the period of 1995 and 2013 in which they find factors like positive shocks to German saving rate, world’s demand for German exports, German labor reforms and other positive German aggregated supply shocks have effect on German current account surplus and negatively affect Euro Area net exports. The research also discusses that exchange rate regime may have a first order effect on current account dynamics.

3. Data and Methodology We used current account balance as a percentage of GDP (CA) for dependent variable. Independent variables are GDP growth rate (GDP), real effective exchange rate index (RER), exports of and services as a percentage of GDP (EXP), imports of as a percentage of GDP (IMP), foreign direct investment as a percentage of GDP (FDI) and general government final consumption expenditure growth (GOV) respectively. Data was collected from “World Development Indicators” in the website. This data was collected in the respect of five Central European Countries (Poland, Hungary, Czech Republic, Slovakia Republic and Croatia) in the period of 1997-2015. According to theoretical perspective, in the emerging markets, growth of the economy leads to an increasing expectation of incomes and, correspondingly, an increasing on workers’ consumption. Therefore, it can be expected that GDP growth has a negative effect on current account (Zorzi et al., 2009; Bollano and Ibrahimaj, 2015; Gehringer, 2015). Real exchange rate adjustment is the most effective indicator on current account adjustment than other adjustment instruments such as income, output and expenditure. Relative price movements lead to matching expenditure between domestic goods and foreign goods (Gervais et al., 2016). If an appreciation occurs for the real exchange rate, it leads to an increase in the purchasing power of household with respect to imported goods, as well as an increase the in the value of the property assets of domestic agents. Therefore, all these variables lead to increase on consumption and a decrease on the saving tendency. Hence it is expected that the increasing of real exchange rate has a negative effect on current account (Brissimis et al., 2010). Similar results obtained for European countries (Aristovnik, 2006; Gehringer, 2015; Bollano and Ibrahimaj, 2015). Exports indicate demand for a local product and imports reflect supplies from foreign countries to meet local production requirements. Shortly export can be regarded as a credit to local economy whereas import implies a debit for a local economy. From this point it could be expected that export has a positive and import has a negative effect on current account. In the literature, foreign direct have a positive spill-over effects on host countries' current account by means of bringing technology and know-how, contributing to development of companies, integration into the global economy and increasing (Mencinger, 2008). Generally, it is suggested that government budget deficits leads to current account deficits via redistributing income from future generations to present generations. In this respect of twin deficit hypothesis, government expenditure could be seen as an important factor for budget deficits (Zorzi et al., 2009). Therefore model has been established in the respect of literature as an equation (1). (1) In this study we used balanced panel data set in the panel data analysis process. Balanced panel data implies that the all year’s data has been obtained for each country and there has not any deficient data. Panel data set in includes of 5 horizontal section units. i symbolizes country and t symbolizes time; i=1-5 (5 countries) and t=1997-2015 (19 years). The total number of observations in data set (i×t = 95) is 95.

4. Analysis Process In the panel data analysis pooled OLS model can be used if all observations are homogenous. When observations include unit and/or time effects, it can be suitable to use fixed effects or random effects models (Yerdelen Tatoğlu, 2012: 163-164). Likelihood ratio (LR) test was used in the model in the respect of to determine whether there are unit and time effects. In LR test, it is

80 European Journal of Economic Studies, 2017, 6(2)

examined whether standard error of unit effects is equal to zero (H0: σµ=0). Otherwise, LR test is used to examine whether standard error of time effects is equal to zero (H0: σλ=0) (Yerdelen Tatoğlu, 2012: 170). Pooled OLS model can be used, if unit and time effects are not determined in LR test. In spite of this condition, if unit and/or time effects are determined in test results, it can be said that the model is one sided or two sided.

Table 1. LR Test

Unit Effect Time Effect χ2 47.26 0.62 prob. 0.0000 0.2151

The results of LR test exhibit that there is an only unit effect in the model. Consequently, the model is one sided. Hausman specification test is used to specify whether unit effect is fixed or random. Hausman test infers that if there is no correlation between error components (ui) and explanatory variables (xkit), both fixed effects and random effects estimators are appropriate. In any case, if there is a correlation between error components and explanatory variables, random effects estimator is inappropriate. In Hausman test, null hypothesis implies that there is no correlation between error components and explanatory variables (Hill et al., 2011: 559). It can be said that random effects are appropriate when there is not a correlation between ui and xkit, and fixed effects are appropriate when there is a correlation between ui and xkit (Gujarati, 2003: 650).

Table 2. Hausman Test

χ2 46.25 prob. 0.0000

Hausman test results show that unit effects are fixed. Therefore, analysis is made in accordance with one sided fixed effects model. After these findings, model was examined in the scope of variation from basic assumptions. One of these assumptions is constant variance (homoscedasticity) assumption. Constant variance assumption implies that while unit values of explanatory variables change, variance of error term remains fixed. If this assumption does not valid, model includes heteroscedasticity (Wooldridge, 2012: 93). Modified Wald Test was used to examine this assumption.

Table 3. Test for Heteroscedasticity

Modified Wald Test 2 X 5.01 prob. 0.4152

Heteroscedasticity results imply that there is no heteroscedasticity. Constant variance assumption is valid. Other basic assumption is autocorrelation assumption; there is no correlation between error terms of independent variables (Wooldridge, 2012: 353). If this assumption does not occur, it implies that there is correlation between error terms of independent variables. Durbin- Watson test of Bhargava, Franzini and Narendranthan test and Baltagi-Wu LBI test were used to examine this assumption. In the respect of values obtained for both tests are less than 2, it can be said that there has been auto-correlation in the model of fixed effects.

81 European Journal of Economic Studies, 2017, 6(2)

Table 4. Test for Autocorrelation

Modified Bhargava et al. Durbin-Watson Test Baltagi-Wu LBI Test

1.1070284 1.2147892

Another assumption is about correlation between units. In studies such as domestic and regional economies, neighborhood effects can show spill-over in themselves. In such cases, correlations have spatial view rather than temporal view (Greene, 2012: 389). This assumption is tested through Friedman’s Test. According to the Friedman’s test of cross sectional independence test statistics and probability values, there is a correlation between units.

Table 5. Test for Correlation between Units

Friedman’s Test of Cross Sectional Independence 2 X 15.802 prob. 0.0033

According to the results of analysis, there have been autocorrelation and correlation between units problems in the model. In order to solve these problems, standard errors which are resistant to deviations from assumptions were produced by using method of Driscoll-Kraay estimator.

Table 6. Analysis Results

Explanatory Variables Coef. t-stat. p-value GDP -0.0670 -1.27 0.274 RER -0.0631 -4.53 0.011** EXP 0.9172 14.74 0.000* IMP -0.9372 -11.20 0.000* FDI -0.0509 -4.46 0.011** GOV 0.0862 2.41 0.073*** Cons. 4.3684 8.44 0.001* R2: 0.8658 Prob. 0.0000 Note: (*) significant at %1 level, (**) significant at %5 level, (***) significant at %10 level.

Analysis results show that the GDP variable effects on CA negatively but it is statistically insignificant. RER effects on CA negatively and this result is statistically significant. In this regard one unit appreciation in RER leads to 0.06 % decrease in CA. Coefficient of EXP has a positive and statistically significant impact on CA. It can be described that one unit increase in EXP gives rise to 0.91 % increase in CA. IMP variable has a negative and statistically significant effect on CA. One unit increase in IMP cause 0.93 % decrease in CA. Coefficient of FDI has a negative and statistically significant effect on CA. It implies that one unit increase in FDI leads to 0.05 % decrease in CA. GOV variable has a positive and statistically significant effect on CA. One unit increase GOV gives rise to 0.08 % increase in CA. Effect of RER, EXP and IMP variables on CA are coherent with the theoretical expectations. However FDI and GOV variables effect on CA are not consistent with the theoretical expectations. Also it can be seen that there have been strong effect of EXP and IMP variables on CA.

5. Conclusion Current account balance is an important subject in economy literature since it has been an indicator of economic crisis variable in pre-crisis period of countries. With regard to European countries, current account imbalances in periphery regions of Europe leads to declining of and losses on foreign asset holdings in these regions. By this way, European countries with current account surplus are affected negatively from these events. From this point,

82 European Journal of Economic Studies, 2017, 6(2) the control of current account imbalances has been considered as a primary goal for European countries. In this study we analyze current account dynamics for five Central European countries in the period of 1997-2015 by using panel data methodology. Real exchange rate has a negative and statistically significant effect on current account in Central European countries. This result is coherent with the theoretical expectations, and it can be said that appropriation in the real exchange rate brings to increase purchasing power of household and rising demand for imported goods. Furthermore, increasing value of the property assets of domestic agents also affect current account negatively. Exportation has a positive effect on current account and this result implies that increasing of foreign demand for local products gives rise to foreign currency access to the economy. Conversely the demands for foreign goods have an impact on current account negatively in Central European countries. This is because that importation stated the debt conditions for countries. The results of exportation and importation are also convenient with theoretical perspective. However foreign direct investment has a negative effect on current account in Central European countries and this result contradicts with the theoretical expectations. But it can be said that in the long run the FDI’s positive effects on current account could be turned negative by the way of repatriation of profits to investor country and this negative effect could be extended if the investment funds gain from the host country through credits channel (Moura, Forte, 2010). Government expenditure has positive effects on current account in Central European countries. It can be said that this result is also adverse with theoretical expectations. Theoretically, twin deficit hypothesis implies that if the government expenditure financed by the government incomes, it leads a current account deficit in the economy. However, Finn (1998) asserts that government expenditure on final goods has a positive effect on private sector’s investment and domestic output. In this respect, government expenditure could be financed without government income and, government expenditure could be financed by the increasing private investments. Therefore it can be said that government expenditure impacts on current account positively in Central European countries. Current account balance is sensitive to international trade movements as respect to import and export. Improving of the policies to increase export and decrease import are important agenda for Central European countries. It can be used regulations on foreign investors to limiting the repatriation of profits to host country. The impact of government expenditure on private investments is positive. Therefore government expenditure does not generate current account deficit in Central European countries.

References Hassan et al., 2015 – Hassan, K., Hoque, A., Rao, A. (2015). Sustainability of Current Account Balance in Asean Countries: Evidence from a Panel Error Correction Model. The Journal of Developing Areas, vol. 49, no. 6, pp. 189-204. Brissimis et al., 2010 – Brissimis, S. N., Hondroyiannis, G., Papazoglou, C., Tsaveas, N.T., Vasardani, M.A. (2010). Current Account Determinants and External Sustainability in Periods of Structural Change. European Working Paper Series No: 1243, Germany: European Central Bank. Belke and Schnabl, 2013 – Belke, A., Schnabl, G. (2013). Four Generations of Global Imbalances. Review of , vol. 21, no. 1, pp. 1-5. Lane and Pels, 2012 – Lane, P.R., Pels, B. (2012). Current Account Imbalances in Europe, IIIS Discussion Paper No: 397. [Electronic resourse]. URL: https://www.tcd.ie/iiis/ documents/discussion/pdfs/iiisdp397.pdf [Accessed: 30.01.2017]. Gosse and Serranito, 2014 – Gosse, J., Serranito, F. (2014). Long-run Determinants of Current Accounts in OECD Countries: Lessons for Intra-European Imbalances, Economic Modelling, vol. 38, pp. 451-462. Cavdar, Aydin, 2015 – Cavdar, S.C., Aydin, A.D. (2015). Understanding the Factors Behind Current Account Deficit Problem: A Panel Logit Approach on 16 OECD Member Countries, Procedia Economics and Finance, vol: 30, pp.187-194. Bertola and Prete, 2015 – Bertola, G., Prete, A.L. (2015). Reforms, Finance and Current Accounts, Review of International Economics, vol. 23, no. 3, pp. 469–488.

83 European Journal of Economic Studies, 2017, 6(2)

Karras 2016 – Karras, G. (2016). Macroeconomic Volatility and the Current Account: Evidence from a Panel of OECD Countries, International Economic Journal, vol. 30, no. 3, pp. 322–338. Erauskin, 2015 – Erauskin, I. (2015). Savings, the Size of the Net Foreign Asset Position and the Dynamics of Current Accounts. International Review of Economics and Finance, vol. 39, pp. 353-370. Kim, 2015 – Kim, S. (2015). Country Characteristics and the Effects of Government Consumption Shocks on the Current Account and Real Exchange Rate. Journal of International Economics, vol. 97, pp. 436-447. Martin, 2016 – Martin, F.E. (2016). Exchange Rate Regimes and Current Account Adjustment: An Empirical Investigation. Journal of International Money and Finance, vol. 65, pp. 69–93. Tan et al. 2015 – Tan, Z., Yao, Y., Wei, S. (2015). Financial Structure, Corporate Savings and Current Account Imbalances, Journal of International Money and Finance, vol. 54, pp. 142-167. Moral-Benito, Roehn, 2016 – Moral-Benito, E., Roehn, O. (2016). The Impact of Financial Regulation on Current Accunt Balances. European Economic Review, vol. 81, pp. 148-166. Aristovnik, 2006 – Aristovnik, A. (2006). The Determinants and Excessiveness of Current Account Deficits in Eastern Europe and the Former Soviet Union. William Davidson Institute Working Paper No:827, USA: University of Michigan. Gehringer, 2015 – Gehringer, A. (2015). New Evidence on the Determinants of Current Accounts in the EU, Empirica, vol. 42, pp.769-793. Bollano, Ibrahimaj, 2015 – Bollano, J., Ibrahimaj, D. (2015). Current Account Determinants in Central Eastern European Countries, Graduate Institute of International and Development Studies International Economics Department Working Paper Series, No: HEIDWP0022-2015, Switzerland. Zorzi et al., 2009 – Zorzi, M.C., Chudik, A., Dieppe, A. (2009). Current Account Benchmarks for Central and Eastern Europe a Desperate Search? European Central Bank Working Paper Series No: 995, https://www.ecb.europa.eu/pub/pdf/scpwps/ecbwp995.pdf? 9da81a457e30259e36a6df428edf2a51 [Accessed: 30.01.2017]. Kang, Shambaugh, 2016 – Kang, J.S., Shambaugh, J.C. (2016). Current Account Deficits, , vol. 31, no. 85, pp. 153-199. Kollmann et al., 2015 – Kollmann, R., Ratto, M., Roeger, W., Veld, J., Vogel, L. (2015). What Drives the German Current Account? And How Does It Affect Other EU Member States? Economic Policy, vol. 30, no. 81, pp. 47-93. World Bank – World Bank (2016). World Development Indicators. [Electronic resourse]. URL: http://data.worldbank.org/data-catalog/world-development-indicators [Accessed: 15.11.2016]. Gervais et al., 2016 – Gervais, O., Schembri, L., Suchanek, L. (2016). Current Account Dynamics, Real Exchange Rate Adjustment, and the Exchange Rate Regime in Emerging-Market Economies. Journal of , vol. 119, pp. 86–99. Mencinger, 2008 – Mencinger, J. (2008). The ‘Addiction’ with FDI and Current Account Balance, International Center for Economic Research Working Paper Series No: 16/2008, [Electronic resourse]. URL: http://www.biblioecon.unito.it/biblioservizi/RePEc/icr/ wp2008/ICERwp16-08.pdf [Accessed: 30.01.2017]. Yerdelen Tatoğlu, 2012 – Yerdelen Tatoğlu, F. (2012). Panel Veri Ekonometrisi, Stata Uygulamalı, Beta Yayınları, İstanbul. Hill et al., 2011 – Hill, R.C., Griffiths E.W., Guay. C.L. (2011). Principles of , Fourth Edition, Wiley, USA. Gujarati, 2003 – Gujarati, D.N. (2003). Basic Econometrics, Fourth Edition McGraw-Hill, New York. Wooldridge, 2012 – Wooldridge, J. (2012). Introductory Econometrics: A Modern Approach, Fifth Edition, South-Western, Cengage Learning, USA. Greene, 2012 – Greene, W.H. (2012). Econometric Analysis, Seventh Edition, Prentice Hall, USA. Moura, Forte, 2010 – Moura, R., Forte, R. (2010). The Effects of Foreign Direct Investment on the Host Country Economic Growth-Theory and Empirical Evidence, FEP Working Papers No.390, pp. 1-32. Finn, 1998 – Finn, M.G. (1998). Cyclical Effects of Government’s Employment and Goods Purchases. International Economic Review, vol. 39, no. 3, pp. 635-657.

84 European Journal of Economic Studies, 2017, 6(2)

Copyright © 2017 by Academic Publishing House Researcher s.r.o.

Published in Slovak Republic European Journal of Economic Studies Has been issued since 2012. ISSN: 2304-9669 E-ISSN: 2305-6282 2017, 6(2): 85-95

DOI: 10.13187/es.2017.6.85 www.ejournal2.com

Does Government Size Affect Economic Growth in Developing Countries? Evidence from Non-stationary Panel Data

Murat Cetin a , * a Namik Kemal University, Turkey

Abstract The Armey curve suggests that there is an inverted U relationship between government size and economic growth. In order to investigate this relationship for 12 developing countries from 1990 to 2012, this study uses panel data methodology including panel unit root, cointegration and causality tests. The results show that i) the series are integrated at order of I(1), ii) there exists a long run equilibrium relationship between the variables, iii) economic growth is positively correlated with the government consumption expenditure, iv) economic growth is negatively correlated with the squares of government consumption expenditure, v) there exists a causality running from the explanotary variables to economic growth in the long run and short run. The study provides an evidence that there exists an inverted U relationship between government consumption expenditure and economic growth implying the validity of Armey curve in these countries. The study may also provide some policy implications. Keywords: government size, economic growth, Armey curve, panel data analysis.

1. Introduction Economic growth and its determinants has been one of the main topics investigated by theorists and politicians. According to growth literature, there are two fundamental kinds of growth theory. The first is the neoclassical growth theory. It is well known as the exogenous growth model presented by Solow (1956), Swan (1956), and Koopmans (1965). The second is the new growth theory developed by Romer (1986; 1990), Lucas (1988), Barro (1990), Rebelo (1991), Grossman and Helpman (1991), Aghion and Howitt (1992), and Jones (1996). This theory is also known as the endogenous growth model. The neoclassical theory of growth generally focuse on capital accumulation and its relation to savings and population growth. It suggests that in the long run economy will reach a steady state where per capita output is constant. It also suggests that there is a linear relationship between a number of variables and economic growth in the long-run. According to this theory, government policy cannot influence the steady-state growth rates. As a result, the impact of government policy on the long run growth has not been investigated in this model. The new growth theory suggests that both transition and steady state growth rates are endogenous and there are several determinants of long run growth. Here, long run growth rates can differ across countries and convergence in income per capita cannot occur. However, according

* Corresponding author E-mail addresses: [email protected] (M. Cetin)

85 European Journal of Economic Studies, 2017, 6(2) to this theory government policy can affect economic growth either directly or indirectly. In this model there are three basic fiscal instrument affecting the long run growth rates: expenditure, taxation and the aggregate budgetary balance. Firstly, these instruments affect the efficiency of resource use and the rate of factor accumulation. These developments influences a country’s long- run growth performance (Barro, 1989; 1990; Brons et al. 1999). A part of the new growth theory focuses on the relationship between government size and economic growth. The literature on public expenditure and economic growth stresses on the presence of a historical relationship between government size and GDP growth. This is called as the Armey curve (Armey, 1995), Rahn curve (Rahn and Fox, 1996) or BARS curve (Barro, 1989; Armey, 1995; Rahn and Fox, 1996; Scully, 1994). This literature uses the form of an inverted U-shaped curve. The Armey curve is based on the law of diminishing factor returns and implies the idea that there is a positive correlation between public expenditure and GDP up to a certain point. After that the correlation becomes negative. In other words, after this point an increase in public expenditure leads to a decrease in GDP. So, Armey curve exhibits a relationship similar to that of Kuznets’ curve. According to Armey curve, the government size and economic growth may be modelled by using a quadratic function (Vedder and Gallaway, 1998). Barro (1990) investigates the impact of different sizes of government on economic growth. According to Barro, an increase in taxes decreases economic growth, while an increase in government expenditure raises marginal productivity of capital. So, economic growth increases. If the government is small, the second force dominates. If the government is large, the first force dominates. The study’s main finding reveals that the relation between government expenditure and economic growth is non-monotonic. The Armey curve can be formulized in different shapes in order to test whether an “inverted U” relationship exists between public expenditure and economic growth. The empirical research on this topic aims to test the presence of this relationship in different countries by using several econometric techniques. Examples are given by Miller and Russek (1997), Vedder and Gallaway (1998), Kneller et al. (1999), Folster and Henrekson (2001), Pevcin (2004), Chen and Lee (2005), Angelopoulos et al. (2008), Herath (2010), Magazzino and Forte (2010), Afonso and Furceri (2010), Wu et al. (2010), Ijeoma and O’Neal (2012), Roy (2012), and Altunc and Aydın (2013). But, the empirical literature provides inconclusive findings regarding the relationship between government expenditure and economic growth. Miller and Russek (1997) investigate the link between government expenditure and economic growth in both developed and developing countries. The results indicate that debt-financed increases in government expenditure slow economic growth and tax-financed increases enhance economic growth for developing countries. The results also indicate that there is no relation between debt-financed increases in government expenditure and economic growth and there is negative link between tax-financed increases and economic growth for developed countries. Vedder and Gallaway (1998) test the validity of the Armey curve in the cases of United States, Sweden, Denmark, Canada, Britain and Italy over the period 1947-1997. The results show that there is empirical evidence supporting the validity of the Armey curve for all these countries. Employing panel data for 22 OECD countries, Kneller et al. (1999) show that productive government expenditure increases economic growth, while non-productive government expenditure does not. Folster and Henrekson (2001) investigate the impacts of expenditure and fiscal measures on economic growth for rich countries over the period 1970-1995. The study finds a strong negative relationship between public expenditure and economic growth. Using panel data regression analysis based on five-year arithmetic averages, Pevcin (2004) examines the relationship between government expenditure and economic growth for European countries. The empirical findings support the presence of the Armey curve over the period. Using a threshold regression approach, Chen and Lee (2005) analyse the non-linear relationship between government expenditure and economic growth in Taiwan. Applying the two- sector production function, the study provides evidence that government size has a threshold effect and that there is a non-linear relationship between the variables implying the presence of Armey curve in Taiwan.

86 European Journal of Economic Studies, 2017, 6(2)

Angelopoulos et al. (2008) analyze the relation between government spending and economic growth in developed and developing countries. Using a panel OLS and 2SLS, they find evidence that there is a nonlinear link between government expenditure and economic growth. The results show that an efficient public sector has a positive impact on economic growth. Herath (2010) investigates the relationship between government expenditure and economic growth in the case of Sri Lanka by using second degree polynomial regressions. The findings show that there is a positive relation between the variables. The findings also support the Armey’s idea of aquadratic curve for Sri Lanka. Magazzino and Forte (2010) investigate the existence of Armey curve for the EU countries in the period 1970-2009 by using time-series and panel data techniques. The study provides empirical evidences generally supporting the presence of Armey curve. Afonso and Furceri (2010) analyze the impacts of size and volatility of government revenue and spending on economic growth in OECD and EU countries by applying panel regression analyses. The findings suggest that both variables are harmful to economic growth. In particular, the results show that government consumption and investments have a negative effect on economic growth. Wu et al. (2010) examine the causal relation between government spending and economic growth by using the panel Granger causality method presented by Hurlin (2004) and panel data set from 1950 to 2004. The study finds evidence of a positive relation between government spending and economic growth. The sudy also finds bi-directional causality between the variables for the different sub samples of countries. Ijeoma and O’Neal (2012) examine the impact of government expenditure on economic growth for Nigerian economy from 1980 to 2011. Using ARDL bounds testing approach, the results indicate that government recurrent and capital expenditures are positively correlated with economic growth in the short-run. In the long run there is a positive relation between government recurrent expenditure and economic growth, while government capital expenditure is negatively linked to economic growth in Nigeria. Using time-series data covering the period 1950-2007, Roy (2012) analyses the relationship between government size and economic growth in the United States. The study particularly investigates the impacts of government consumption and government investment expenditures on US economic growth. Based on the results of a simultaneous-equation model, government consumption expenditure decreases economic growth, while government investment expenditure increases economic growth in the United States. So, the study shows that the overall impact of total government spending on economic growth is uncertain. Altunc and Aydın (2013) examine the presence of Armey curve for Turkey, Romania and Bulgaria by using ARDL bounds testing approach to cointegration from 1995 to 2011. This study finds an empirical evidence that the Armey curve is valid for Turkey, Romania and Bularia. Following the empirical lietrature, this study’s main aim is to investigate wether the Armey curve (the inverted U relationship between government size and economic growth) exists in developing countries over the period 1990-2012. In this purpose, we employ panel unit root tests developed by Maddala and Wu (1999), Hadri (2000), and Im et al. (2003). We also employ the cointegration methods developed by Kao (1999) and Maddala and Wu (1999) to examine the long- run relationship between the variables. Long-run estimation is conducted by panel OLS method. Finally, the long run and short run causality between the variables is investigated by panel vector error correction model (PVECM). The remainder of this study is organized as follows. Section 2 describes the model and data of the empirical analysis. Section 3 presents the empirical methodology. Empirical results are reported in Section 4. Section 5 concludes the study with some policy implications.

2. Model and Data In this study, we investigate the relationship between government size and economic growth in selected developing countries. We use panel data covering the period 1990-2012 gathered from the World Development Indicators (WDI) online database (2014). The countries examined in this study are Brazil, Gabon, Colombia, Costa Rica, Peru, Bostwana, China, Malaysia, Mexico, South Africa, Thailand and Turkey. This sample is selected on the bases of upper-middle income country and data availability. In order to test the existence of the inverted-U shaped relation between

87 European Journal of Economic Studies, 2017, 6(2) government size and economic growth (Armey curve), the following quadratic function presented by Vedder and Gallaway (1998) can be used

LNGDP    LNGOV  LNGOV 2  it 0 1 it 2 it it (1)

where GDP, GOV and GOV2 represent per capita real income, government consumption expenditure as a percentage of real GDP and square of government consumption expenditure as a percentage of annual real GDP, respectively. So, government consumption expenditure is used as an indicator of government size. The data are transformed to natural logarithm because log-linear form provides a better result. α1 and α2 are the slope coefficients and the sign of the coefficients is expected to be positive and negative, repectively (Vedder and Gallaway, 1998; Herath, 2010; Altunc and Aydın, 2013). εt is the error term assumed to be normally distributed with zero mean and constant variance. Table 1 presents the descriptive statistics of the variables employed in the analysis. Figure 1 shows the plots of the series.

Table 1. Descriptive statistics

Balanced panel: N=12, T=23, Observations=276 Variable Unit Mean Median Std. Dev. Min. Max. LNGDP GDP per capita, 2005=100, $ 8.308 8.410 0.535 6.137 9.053 LNGOV Goverment consumption 2.616 2.579 0.297 2.080 3.177 expenditure/GDP, 2005=100, $ LNGOV2 Square of LNGOV 6.935 6.651 1.573 4.328 10.094

LNGDP LNGOV 10 3.2

3.0 9 2.8

8 2.6

2.4 7 2.2

6 2.0

3 3 0 0 0 7 7

3 3 0 0 0 7 7

9 0 9 0 1 9 0

9 0 9 0 1 9 0

------

------

0 0 1 1 1 2 2

0 0 1 1 1 2 2

1- 90 1- 00 1- 10 2- 97 2- 07 3- 94 3- 04 4- 91 4- 01 4- 11 5- 98 5- 08 6- 95 6- 05 7- 92 7- 02 7- 12 8- 99 8- 09 9- 96 9- 06

1 1 1 1 1 1 1

1- 90 1- 00 1- 10 2- 97 2- 07 3- 94 3- 04 4- 91 4- 01 4- 11 5- 98 5- 08 6- 95 6- 05 7- 92 7- 02 7- 12 8- 99 8- 09 9- 96 9- 06

1 1 1 1 1 1 1

LNGOV2 12

10

8

6

4

3 3 0 0 0 7 7

9 0 9 0 1 9 0

------

0 0 1 1 1 2 2

1- 90 1- 00 1- 10 2- 97 2- 07 3- 94 3- 04 4- 91 4- 01 4- 11 5- 98 5- 08 6- 95 6- 05 7- 92 7- 02 7- 12 8- 99 8- 09 9- 96 9- 06

1 1 1 1 1 1 1

Fig. 1. The plots of LNGDP, LNGOV and LNGOV2 series

3. Econometric Methodology As the main aim of this study is to examine the cointegration and causality relationship between goverment size and economic growth over the period, our econometric strategy consists of three steps. In the first step, we investigate the order of integration in the variables by using panel unit root tests presented by Maddala and Wu (1999), Hadri (2000) and Im et al. (2003). Using cointegration methods developed by Kao (1999) and Maddala and Wu (1999), the second step tests the cointegration relationship between the variables. In the third step, the long-run parameters are estimated and final step investigates the Granger causality between the variables by applying PVECM.

88 European Journal of Economic Studies, 2017, 6(2)

3.1 Panel Unit Root Tests Im et al. (2003) provides a very simple panel unit root test which is well known as IPS test. They employ a separate ADF regression as follows:

pi

yit   i  i yi,t1  ∑ij yi,t j   it (2) j1

where i = 1, . . .,N and t = 1, . . .,T

The test allows for a heterogeneous coefficient of yit-1 and bases on averaging individual unit root test statistics. In this test, the null and alternative hypotheses are as follows:

H 0 : i  0 for all i (3)

H1 : i 0 for i = 1, 2, ….. N1

(4) H1 : i  0 for i = N1+1, ….. N (5)

The IPS t-bar statistic indicates an average of the individual ADF statistics and is estimated as follows: 1 N t NT  ti (6) N i1

where tힺi is the individual t-statistic for testing H0 hypothesis. In case the lag order is always zero, IPS provides simulated critical values related with t-bar for different number of cross-sections N and series lenght T. IPS reveals that standardized t-bar statistic exhibits an asymptotic N(0,1) . The unit root test developed by Maddala and Wu (1999) uses the Fisher (p) test. Under cross- sectional independence of the error terms εit, the joint test statistic can be expressed as follows: N p  2ln( i ) (7) i1

In this procedure, the null and alternative hypotheses are similar to IPS’s hypotheses. Using the ADF estimation equation in each cross-section, this test computes the ADF t-statistic for each individual series. So, the Fisher-test statistics are calculated and are compared with the appropriate χ2 critical value. Hadri (2000) presents a panel version of the Kwiatkowski et al. (1992) test. In this procedure, the null hypothesis implies that there exists stationarity in all units. The null hypothesis is tested against the alternative of a unit root in all units. The test is based on Langrange test and the residuals are obtained from the following regression:

yit   mi dmt   it , m = 2, 3 for i = 1, …… N. (8)

The test statistic is then given by

1 N T S 2 H  it (9) LM 2   2 NT i1 t1 ˆ ei

89 European Journal of Economic Studies, 2017, 6(2)

1 T with ˆ 2  eˆ 2 . ei T t1 it

3.2 Panel Cointegration Tests Kao (1999) suggests several residual-based panel tests and they have parametric properties. In these tests, the null hypothesis implies that there exsists no cointegration. In this procedure, the DF and ADF unit root tests are added to panel cointegration analyses. The main feature of these tests is that they base on the spurious least squares dummy variable panel regression equation as follows:

yit   i   it   eit , i = 1,……N; t = 1,…….T (10)

t t in which y  u and x   are restricted to be atmost I(1) with u ∼ (0, 2 ) i.i.d. it s1 is it s1 is it u  2 and εit∼ (0,  ) i.i.d.. The ADF type panel statistic developed by Kao bases on the following AR (p) regression

eˆit  peˆi,t1  1eˆi,t1 ......  p eˆi,t p  vitp (11)

Kao (1999) formulates the ADF panel test statistic as follows:

N (e'Q v ) 6Nˆ i1 i i i  v N ' ˆ sv (e Q e ) 2 0v i1 i i i  ADF  (12) 2 2 ˆ 0v 3ˆ v 2  2 2ˆ v 10ˆ 0v

' 1 ' where Qi  I  X ip (X ip X ip ) X ip , and X ip indicates a matrix of observations on the

(eˆ ,eˆ ,...... , eˆ ). regressors i,t1 i,t2 i,t p

N T vˆ 2 2 i1 t1 itp s  (13) v NT

where vˆitp implies the estimate of vitp . The panel ADF test has a asymptotically N(0,1) distribution. Hence, in addition to the Kao test, we also employ Fisher’s test to aggregate the p-values of the individual Johansen maximum likelihood cointegration test statistics. In the Fisher procedure which is a non-parametric test the homogeneity in the coefficients are not assumed (Maddala and Kim, 1998; Maddala and Wu, 1999).

3.3 Panel Granger Causality Test If there is a cointegration relationship between the variables, this implies a causal relation between the variables. However, this does not show the direction of causality. To test the causal relations between the series, we can use a two-step process. In the first step the residuals are estimated from the long-run model. In the second step the estimated residuals are included to error correction model as an error correction term (ECT). This model is known as dynamic error correction model. The model is expressed as follows

90 European Journal of Economic Studies, 2017, 6(2)

q q q 2 LNGDPit  1   1ikLNGDPitk   2ik LNGOVitk  3ik LNGOVitk  ECTit1  it (14) k 1 k 1 k 1

where  and q represent the first difference operator and the lag length, respectively. ECT denotes the error-correction term which contains estimated residuals from the cointegration regression (Eq. 1). μ is the serially uncorrelated error term. γ reflects the long-run equilibrium relationship among the variables. If θ2 or θ3 is not equal to zero, it is determined to be a short run causal relationship. If γ is not equal to zero, it is determined to be a long run causal relationship. If γ and θ2 or θ3 are not equal to zero, it is determined to be a joint causal relationship.

4. Empirical Findings Table 2 reports panel unit root test results. The findings indicate that the series are not stationary in level. After taking the first difference, the series are stationary. So, it is concluded that all variables are integrated at order of I(1). These results enable us to apply the cointegration tests.

Table 2. Panel unit root test results

Variables IPS ADF-Fisher PP-Fisher Hadri test test test statistics test statistics statistics statistics Panel A: Level LNGDP 4.576 5.851 5.293 11.799a0.000 LNGOV -0.805 29.220 21.627 8.704a0.000 LNGOV2 -0.727 28.713 21.981 8.784a0.000 Panel B: First difference ΔLNGDP - 122.139a0.000 133.135a0.000 0.645 9.257a0.000 ΔLNGOV - 128.351a0.000 145.549a0.000 0.092 9.481a0.000 ΔLNGOV2 - 124.941a0.000 145.046a0.000 0.045 9.199a0.000 Notes: The optimal lag lengths are selected automatically using Akaike information criteria (AIC). The LLC test uses Newey-West bandwidth selection with Bartlett kernel. a denotes significance at the 1 % level. p-values are given in parentheses.

Table 3 presents the results of Johasen-Fisher and Kao cointegration tests. Fisher statistics estimated from trace and maximum eigen tests indicate that there are two cointegration vectors implying the presence of a long-run relationship between the variables at the %1 level. Kao test results indicate the existence of a long-run relationship between te variables. All the findings provide an evidence that there is a cointegration relationship between per capita real income, government consumption expenditure and square of government consumption expenditure over the period.

91 European Journal of Economic Studies, 2017, 6(2)

Table 3. Panel cointegration test results

Cointegration tests Fisher statistics Fisher statistics (from trace test) (from max. eigen test) Panel A: Johansen-Fisher None 255.6a0.000 195.8a0.000 At most 1 121.0a0.000 120.0a0.000 At most 2 32.880.106 32.880.106 Panel B: Kao ADF statistics 1.876b0.030 Notes: The optimal lag length is selected using AIC. a and b denote significance at the 1% and 5% level, respectively. The values in parenthesis are p-values.

The estimations of long-run parameters are conducted by using panel pooled OLS method. The results are presented in Table 4. Diagnostic tests show that there are the problems of serial correlation and heteroscedasticity in the model. We apply the processes of AR(1) and White cross- section to resolve these problems. The results show that economic growth is positively correlated with the government consumption expenditure. This indicate that an increase in government size can enhance economic growth. The results also show that economic growth is negatively correlated with the square of government consumption expenditure. These findings provide an evidence supporting the presence of an inverted U shaped relationship between government size and economic growth.

Table 4. Panel regression estimation results (Dependent variable: LNGDP, Method: Pooled panel OLS)

Variables Coefficients t-statistics Standart errors

LNGOV 1.084 3.252 a0.001 0.333 LNGOV2 -0.267 -4.114 a0.000 0.065 Constant 8.158 18.202 a0.000 0.448 AR(1) 0.968 251.622 a0.000 0.003 Diagnostic tests R2 0.996 Adjusted-R2 0.996 F-statistic 22628.85 a0.000 Durbin-Watson statistic 1.420 2 a LMh ( ) statistic 262.852 0.000 Baltagi-Lee (2) statistic 385.831 a0.000 Notes: a denotes significance at the 1% level. The values in parentheses are p-values

Table 5 reports the results of the long-run, short-run and joint Granger causality. The results suggest that the lagged error correction term is negative and statistically significant at 5 % level as expected. This implies a causality running from government consumption expenditure and the squares of government consumption expenditure to economic growth in the long run. It is found that there exists a causal relation running from government consumption expenditure and the squares of government consumption expenditure to economic growth in the short run. It is also found that there exists a joint causal relation running from the explanatory variables to economic growth. The Granger causality findings provide an evidence that government consumption expenditure (government size) causes economic growth in developing countries over the period.

92 European Journal of Economic Studies, 2017, 6(2)

Table 5. Panel Granger causality test results (Dependent variable: LNGDP)

Series Short run Long run Joint (Short run and Long run) F-statistic ECT(-1) F-statistic LNGOV 2.621b0.051 LNGOV2 2.628b0.051 ECTit-1 -0.020 a0.000 LNGOV/ECT 6.061 a0.000 LNGOV2/ECT 6.116 a0.000 Notes: The optimal lag length is selected using AIC. a and b denote significance at the 1 % and 5 % level, respectively. The values in parentheses are p-values.

5. Conclusion and Policy Implication The determinants of economic growth have been discussed by theorists and econometricians for a long time. Growth literature presents two fundemental models: exogenous growth model and endogenous growth model. The first model suggests that there is a linear relationship between a number of variables and economic growth in the long-run. In this model, government policy cannot influence the steady-state growth rates. The second model is well known as new growth theory. In this model government policy can affect economic growth either directly or indirectly. In this contex, a fundemental strand of the new growth theory concentrates on the inverted U relationship between government size and economic growth. This is generally called as Armey curve. The study investigates the cointegration and causal relationship between the government consumption expenditure and economic growth in the context of Armey curve. We employ panel data covering 1990-2012 for 12 developing countries. Panel unit root tests indicate that the series are integrated at order of I(1) implying that we can apply the cointegration tests. Panel cointegration tests reveal that there exists a lon run relationship between the variables. Panel pooled OLS estimations suggest that the coefficients of government consumption expenditure and the squares of government consumption expenditure are positive and negative, respectively as expected. Granger causality test based on VECM shows that there exists a causal relation running from government consumption expenditure and the squares of government consumption expenditure to economic growt in the long run and short run. All the empirical findings reveal that there exists an inverted U-shaped relationship between government consumption expenditure and economic growth. So, the study provides an empirical evidence that the Armey curve is valid for developing countries over the period. The empirical results also imply that there is an optimal level of government consumption expenditure. Therefore, governments should avoid excessive consumption expenditure. Otherwise, these excessive expenditure hamper to economic growth. On the other hand, this study can be repeated by considering different kinds of government spending. This empirical study may also bring about new empirical studies. In this respect, a further empirical research may include the individual countries or the sub groups of the panel.

References Afonso, Furceri, 2010 – Afonso, A., Furceri, D. (2010). Government size, composition, volatility and economic growth. European Journal of , 26, 517-532. Altunc, Aydın, 2013 – Altunc, O. F., Aydın, C. (2013). The relationship between optimal size of government and economic growth: Empirical evidence from Turkey, Romania and Bulgaria. Procedia-Social and Behavioral Sciences, 92, 66-75. Angelopoulos, Philippopoulos, Tsionas, 2008 – Angelopoulos, K., Philippopoulos, A., Tsionas, E. (2008). Does public sector efficiency matter? Revisiting the relation between fiscal size and economic growth in a world sample. , 137, 245-278. Armey, 1995 – Armey, D. (1995). The Freedom Revolution, Regnery Publishing Co., Washington, D.C. Barro, 1989 – Barro, R. J. (1989). A cross country study of growth, saving, and government, NBER Working Paper No. 2855.

93 European Journal of Economic Studies, 2017, 6(2)

Barro, 1990 – Barro, R.J. (1990). Government spending in a simple model of endogenous growth. Journal of Political Economy, 98(5), 103-125. Brons, de Groot, Nijkamp, 1999 – Brons, M., de Groot, H.L.F., Nijkamp, P. (1999). Growth effects of fiscal policies. Tinbergen Discussion Paper, Amsterdam: Vrije Universiteit. Chen, Lee, 2005 – Chen, S.T., Lee, C.C. (2005). Government size and economic growth in Taiwan: A threshold regression approach. Journal of Policy Modeling, 27, 1051-1066. Folster, Henrekson, 2001 – Folster, S., Henrekson, F. (2001). Growth effects of government expenditure and taxation in rich countries. European Economic Review, 45(8), 1501-1520. Grossman, Helpman, 1991 – Grossman, G.M., Helpman, E. (1991). Quality ladders in the theory of growth. Review of Economic Studies, 58(1), 43-61. Hadri, 2000 – Hadri, K. (2000). Testing for stationarity in heterogeneous panel data. Econometric Journal, 3, 148-161. Herath, 2010 – Herath, S. (2010). The size of the government and economic growth: An empirical study of Sri Lanka. SRE-Discussion Papers, 2010/05. WU Vienna University of Economics and Business, Vienna. Ijeoma, O’Neal, 2012 – Ijeoma, E.K., O’Neal, E.J. (2012). Government expenditure and economic growth in Nigeria, 1980-2011. International Journal of Academic Research, 4(6), 205- 209. Im, Pesaran, Shin, 2003 – Im, K.S., Pesaran, M.H., Shin, Y. (2003). Testing for unit roots in heterogeneous panels. Journal of Econometrics, 115, 53-74. Jones, 1995 – Jones, C.I. (1995). Time series tests of endogenous growth models. Quarterly Journal of Economics, 110(2), 495-525. Kao, 1999 – Kao, C. (1999). Spurious regression and residual-based tests for cointegration in panel data. Journal of Econometrics, 90, 1-44. Kneller, Bleaney, Gemmell, 1998 – Kneller, R., Bleaney, M.F., Gemmell, N. (1998). Growth, public policy and the government budget constraint: Evidence from OECD countries. Discussion Paper No.98/14, School of Economics, University of Nottingham. Koopmans, 1965 – Koopmans, T.C. (1965). On the Concept of Optimal Economic Growth. The Econometric Approach to Development Planning, North-Holland, Amsterdam. Kwiatkowski, Phillips, Schmidt, Shin, 1992 – Kwiatkowski, D., Phillips, P. C. B., Schmidt, P., Shin, Y. (1992). Testing for the null of stationarity against the alternative of a unit root. Journal of Econometrics, 54, 159-178. Lucas, 1988 – Lucas, R. (1988). On the mechanics of economic development. Journal of , 22, 2-42. Maddala, Wu, 1999 – Maddala, G.S., Wu, S. (1999). A comparative study of unit root tests with panel data and a new simple test. Oxford Bulletin of Economics and Statistics, 61, 631-52. Maddala, Kim, 1998 – Maddala, G.S., Kim, I.M. (1998). Unit Roots, Cointegration, and Structural Change. Cambridge University Press, Cambridge. Magazzino, Forte, 2010 – Magazzino, C., Forte, F. (2010). Optimal size of government and economic growth in EU-27. MPRA Paper, No. 26669. Miller, Russek, 1997 – Miller, S.M., Russek, F.S. (1997). Fiscal structures and economic growth: International evidence. Economic Inquiry, 35(3), 603-613. Pevcin, 2004 - Pevcin, P. (2004). Does Optimal Spending Size of Government Exist?. EGPA- European Group of Public Administration, Katholieke Universiteit Leuven, Belgium, Retrieved August 18, 2008, from http://soc.kuleuven.be/io/egpa/fin/paper/slov2004 /pevcin.pdf Rahn, Fox, 1996 – Rahn, R., Fox, H. (1996). What Is the Optimum Size of Government?. Vernon K. Krieble Foundation. Rebelo, 1991 – Rebelo, S. (1991). Long-run policy analysis and long-run growth. Journal of Political Economy, 99(June), 500-521. Romer, 1986 – Romer, P.M. (1986). Increasing returns and long-run growth. Journal of Political Economy, 94(5), 1002-1037. Romer, 1990 – Romer, P.M. (1990). Endogenous technological change. Journal of Political Economy, 98(5), 71-102. Roy, 2012 – Roy, A.G. (2012). Revisiting the relationship between economic growth and government size. Economics Research International, 1-8.

94 European Journal of Economic Studies, 2017, 6(2)

Scully, 1994 – Scully, G.W. (1994). What is the Optimal Size of Government in the US?. National Center for Policy Analysis, Policy Report 188. Solow, 1956 – Solow, R. (1956). A contribution to the theory of economic growth. Quarterly Journal of Economics, 70, 65-94. Swan, 1956 – Swan, T. (1956). Economic growth and capital accumulation. Economic Record, 32(Nov), 334-361. Vedder, Gallaway, 1998 – Vedder, R.K., Gallaway, L.E. (1998). Government size and economic growth. paper prepared for the Joint Economic Committee, Retrieved September 11, 2008, from http://www.house.gov/jec/growth/function/function.pdf World Bank, 2016 – World Bank, (2016). World Development Indicators. http://data.worldbank.org/data-catalog/world-development-indicators Wu, Tang, Lin, 2010 – Wu, S.Y., Tang, J.H., Lin, E.S. (2010). The impact of government expenditure on economic growth: How sensitive to the level of development?. Journal of Policy Modeling, 32, 804-817.

95 European Journal of Economic Studies, 2017, 6(2)

Copyright © 2017 by Academic Publishing House Researcher s.r.o.

Published in Slovak Republic European Journal of Economic Studies Has been issued since 2012. ISSN: 2304-9669 E-ISSN: 2305-6282 2017, 6(2): 96-103

DOI: 10.13187/es.2017.6.96 www.ejournal2.com

UDC 338.532.64; 336.711

The Role of Inflation and its Targeting for Low-Income Countries (Lessons from Post-Communist Georgia)

Vakhtang Charaia a , Vladimer Papava b , * a Business and Technology University, Georgia b Ivane Javakhishvili Tbilisi State University, Georgia

Abstract The inflation index regrettably fails to fully reflect the expectation of the population in developing and, especially, poorer countries. Some of the commodity groups (e.g. electronics, new and used cars, furniture, hotel and restaurant services, etc.) fail to reflect the problem of the low- income population. Under these conditions, a logical question arises concerning the kinds of problems which might occur when the main goal for a central bank’s is only to retain price stability, which is known as inflation targeting. For countries where import exceeds export by several times, it should be clear that calculations must be made not only by the traditional inflation index but also according to their consumer basket made up exclusively of imported goods and services (imflation). Agrarian inflation, that is agflation, becomes more and more popular in economics. The agflation index use area is restricted because it fails to reflect the change in on such substantial spheres as medication and . In the paper we propose a new statistical indicator, munflation, which reflect price fluctuations on the medication, utilities and nutrition. Keywords: inflation, inflation targeting, low-income population, imflation, agflation, munflation.

1. Introduction Inflation is an important macroeconomic indicator for the analysis of an established economic situation as well as forecasting the economic development for any country. Post-Communist Georgia regrettably belongs to the list of countries which have experienced a hyperinflationary spiral (Gurgenidze et al., 1994). In particular, at the initial stage of the restoration of independent statehood, mistakes made by the Georgian Government and the National Bank of Georgia (NBG) have led to in the country (Khaduri, 2005: 20-24; Papava, 1995). To overcome the problem of hyperinflation, complex economic reform was carried out in Georgia (Kakulia, 2008; Papava, 2011) as a result of which macroeconomic stability has been achieved (Papava, 1996; Wang, 1998; Wellisz, 1996). Based on this particular experience as well as

* Corresponding author E-mail addresses: [email protected] (V. Papava)

96 European Journal of Economic Studies, 2017, 6(2) international practice, price stability preservation has been established, and quite logically so, as the primary goal of the NBG’s monetary policy (Kakulia, Gigineishvili, 2005). The purpose of this research is to study the aspect of the traditional inflation indicator which acquires a great importance in a country’s development, especially in poorer countries, based on the Georgian example.

2. Why Inflation Index is not Understandable for Low-Income Population The National Statistics Office of Georgia (Geostat) has been engaged in the inflation monitoring of six Georgian cities at over 1,700 retail and service outlets since 1992. The “consumer basket,” which helps to indicate the consumer price index (CPI) or the average inflation rate, incorporates 12 commodity groups and given the correspondent weights encompasses 305 different goods and services (Geostat, 2017b). The basket is rather balanced and covers all of the products required by an ordinary citizen. However, the majority of these commodity groups fails to reflect the problem of the Georgian population as well as for other developing and, especially, poorer countries. In particular, for Georgia, where poverty represents the most urgent problem for 30 % of the population (NDI, 2017), the price dynamics; that is, their decrease or increase on expensive alcoholic beverages, furniture, recreation and entertainment, and hotel and restaurant services are of no importance. Consequently, we can conclude that the abovementioned basket for those poverty affected fails to adequately reflect the structure of expenditure for the country’s average (that is, poorer) consumer. It is a fact that the inflation index and its internationally recognized and approved calculation practice regrettably fails to fully reflect the expectation of the population in developing and, especially, poorer countries as conditioned by the perception of the average . Moreover, this can also possibly cause the rise of negative emotions based on distrust among society. In particular, when the official low inflation rate is characterized by a significant price increase on essential products for low-income , these negative emotions occur when the low-income group sees a price reduction for only lesser important products. Given the fact that in 2013 around 746 million people (of which 383 million are in Africa and 327 million are in Asia) lived in extreme poverty (Roser, Ospina, 2017) and that in 2014-2016 around 10.9 % of the world population was starving (WHES, 2016) (that is, as poverty remains a global problem (e.g. Sachs, 2005)), it is obvious that the average price level as assessed by the traditional inflation index is meaningless at its best. The price dynamics on , basic medicine and elementary utilities is what concerns people most of all. Ultimately, using the inflation index alone cannot guarantee successful decision-making. Under these conditions, a logical question arises concerning the kinds of problems which might occur when the main goal for a central bank’s monetary policy is only to retain price stability. This is known as inflation targeting.

3. On the Inflation Targeting Beginning from 1967, New Zealand experienced a stretch of high inflation lasting for more than two decades (the average annual inflation was 15 % and it peaked at 20 %) (e.g. Sherwin, 1997: 261). In 1984, the Reserve Bank of New Zealand issued an act under which the desired maximum inflation level was set for the monetary policy which paved the way for the so-called inflation targeting. By doing this, New Zealand was the first country in the world to renounce the internationally recognized priority of the monetary aggregates and exchange rate (e.g. Archer, 2000; Bernanke et al., 1999; Brash, 2002; Fischer, Orr, 1994; Spencer et al., 2006). New Zealand’s example was soon adopted by several other countries – Canada, the UK, Finland, Sweden, Australia and Spain (Debelle et al., 1998). By 2006, there were 25 inflation targeting countries (Mishkin, Schmidt-Hebbel, 2007: 1) with the number growing to 62 by 2017 (CBN, 2017). It should be mentioned that the NBG (starting from 2009), like the central banks of Georgia’s main trade partner countries (Turkey, Russia, Azerbaijan, Belarus, etc.), has already been exercising inflation targeting for years. Central banks employing inflation targeting frequently justify their decisions to do so and state that they have reached not only their desired target (price stability) but have also contributed to stable economic growth such as, for example, was announced by Canada’s central bank (Bank of Canada, 2006: 3). The fact that the average inflation level was comparatively low in both

97 European Journal of Economic Studies, 2017, 6(2) developing and developed countries utilizing inflation targeting is proven by research studies (Mishkin, Schmidt-Hebbel, 2007; Vega, Winkelried, 2005). This said, however, inflation targeting does have serious opposition (e.g. Plushchevskaya, 2012; Snooks, 2008). For example, , the Nobel Prize winner in Economics, is almost confident that this system will be changed because the central banks of developing economies are incapable of managing their inflation which is frequently imported (Stiglitz, 2008). In the opinion of Jeffrey Frankel, a Professor at Harvard University and a member of President Bill Clinton’s Council of Economic Advisers: “One reason that inflation targeting gained such wide acceptance as the monetary-policy anchor of choice was the demise of its predecessor, exchange-rate targeting, in the currency crises of the 1990s; pegged exchange rates had come under fatal speculative attack in many of these countries whose authorities thus needed something new to anchor the public’s expectations concerning monetary policy. Inflation targeting was in the right place at the right time” (Frankel, 2012). However, subsequently inflation targeting died and central banks have not yet decided what new commitment monetary policy should be given in order to become a new hope for stability (Frankel, 2012). The Bank for International Settlements resists inflation targeting which, in most cases, runs counter to financial stability (BIS, 2010; Jones, 2016). Inflation targeting does not take into account the financial cycle and thus produces an excessively expansionary and asymmetric monetary policy (Weber, 2015). It is noteworthy that inflation targeting creates a great deal of questions. A major argument in favor of inflation targeting – that it has contributed to a decline in inflation since the early 1990s – is questionable at best. From the 1980s on, the inflation trend was already on the decline where globalization and China’s integration into the world economy – and not inflation targeting – have probably been the most important reasons for the drop (Weber, 2015). It is a fact that Georgia, like other developing and, especially, poorer and import-dependent countries, is experiencing problems that cannot be fully explained by the inflation index alone. In particular, such examples for Georgia are: a) 80 % of the consumer basket comprises imported products; therefore, inflation is also to a certain degree imported; b) A high level of dollarization takes place (over 70 % by the end of 2016 (NBG, 2017b)*); c) Over half of the employed (in 2015, the number of self-employed in the total number of employed was more than 57 % (Geostat, 2017a)) do not receive any income/ to a bank account and a considerable part of the population lives on money transfers from abroad. The tightening or mitigation monetary policy by the NBG, therefore, may be accompanied by a completely unplanned revaluation or devaluation of the national currency based on the regional and world political and/or economic actions which can completely “absorb” the decision taken by the NBG. This problem is clearly fixed by the NBG itself which indicates the difficulty in forecasting a purposeful inflation rate: “The forecast is largely dependent on exogenous factors affecting the market and contains risks in both upward and downward directions. The main risks continue to stem from the external sector; in particular, from the economic conditions of trading partner countries and the global strength of the US dollar as well as international commodity prices. The current forecast does not expect any significant changes in these factors” (NBG, 2017a: 7). For the lower income part of the Georgian population, like in any other country, it is difficult to provide for their satisfaction when the government is keeping the inflation rate at a certain level (targeted inflation by NBG – 5 % in 2016, 4 % in 2017 and 3 % in 2018). The country’s relatively low level of inflation in the past resulted from the drop in oil prices on international markets (Nasdaq, 2017) and the devaluation of national in the neighboring (main trade partner) countries (IMF, 2016: 12-13) (correspondingly, by a reduction in the prices for their products on the Georgian market). Ignorance of the importance of the national currency exchange rate rather painfully affects a wider group of the population and business which ultimately increases both fear and uncertainty (Anguridze et al., 2015).

* For Georgia, this problem has been studied in detail in the work: (Kakulia, Aslamazishvili, 2000)

98 European Journal of Economic Studies, 2017, 6(2)

4. Modifications of the Imflation In order not to mislead a country’s population, its central bank, its government and business as well as for an adequate reflection of the reality in developing and mostly poorer countries, other indices must also be used together with the inflation index. For countries where import exceeds export by several times, it should be clear that calculations must be made not only by the traditional inflation index but also according to their consumer basket made up exclusively of imported goods and services. Such an index can be called imflation which is a combination of two terms – “import” and “inflation” (Charaia, Papava, 2017). It is noteworthy that if targeting parameters also include imflation together with ordinary inflation, then central banks will need to adequately respond to the issue of national currency devaluation in order to prevent price increases of imported goods on the domestic market owing to the particularly large volume of import. It should be mentioned that the NBG has been engaged in the monitoring of “imported” and “local” commodity prices (e.g. NBG, 2017a: 18) which, to be sure, is very important for Georgia’s economy and its negative trade balance. Given that the Georgian consumer basket contains many goods and services (both imported and domestic) which are generally not purchased by lower-income citizens, neither inflation nor the imflation index will meet the goal of assessing price dynamics in those spheres important for the poorer population, especially in developing countries. As is well known, agrarian inflation (or the growth of average prices for agricultural products) or the agflation index, becomes more and more popular in economics. The term “agflation” is relatively new and its introduction is associated with the substantial increase in the prices for fruit, eggs, grain and other commodities in 2006-2007 (Chorafas, 2016: 139). The agflation measurement is very important in developing and, especially, poorer countries which are characterized by permanent increases in foodstuff prices (for example, in India*). Many studies have proven that the agflation index is higher and rather more important in developing poorer countries where food products constitute about half of the total consumer basket (Georgia – 31 % (Geostat, 2017b), Russia – 50 % (Mashirova, Stepashova, 2015), Azerbaijan – 50 %, Armenia – 50 %, Turkmenistan – 60 % and Tajikistan – 57 % (EPRC, 2012: 12, 32)) in contrast to developed Western countries (USA – 15 % (EPRC, 2012: 32), Eurozone countries – 18 % (Eurostat, 2016) and Turkey – 24 % (Daily News, 2016). It is noteworthy that agflation is not only a problem for developing economies. This is evidenced by the challenges experienced by the new EU Member States (the ten Eastern European countries which joined in 2004) (IMF, 2008). Prices for a whole range of products (dairy, vegetables and sugar) are going up not only in the EU but all over the world whereas the prices for some individual products (e.g., pork) in the EU are higher on average than in the rest of the world (EC, 2016). As is also well known, food inflation is not only higher, more instable, shows great volatility and lasts longer than non-food inflation, it also needs more time to adapt to new reduced prices which is unlike the process of price increases (e.g. EPRC, 2012: 31). The agflation index use area is restricted because it fails to reflect the change in prices on such substantial spheres as medication and utilities. For example, the average inflation index in Georgia in 2016 exceeded the average index for agflation. However, the annual average price increases for products such as medication and utilities, which are a major concern for the wider population, has exceeded the average annual inflation rate (see Table 1). Given that the population in poorer countries gives special attention to how prices of food products, medication and utilities (mainly water, electricity, sewage, gas and other fuels) fluctuate, the statistical indicator adequately reflecting these prices should be calculated. Hence, we propose a new statistical indicator, munflation. This new term comes from the first letters of the English words – medication, utilities and nutrition (Charaia, Papava, 2017). The respective parameters for medication, utilities and food products from the consumer basket should be used for a munflation calculation. Food products prices are also used for the agflation calculation as mentioned above.

* For example, the agflation index acquires a special importance for India (see, Suryanarayana, 2008)

99 European Journal of Economic Studies, 2017, 6(2)

Table 1. Consumer Price Indices in Georgia in 2016 (the respective month of the previous year=100)

2016 Groups I II III IV V VI VII VIII IX X XI XII Average Total 105.6 105.6 104.1 103.2 102.1 101.1 101.5 100.9 100.1 99.8 100.2 101.8 102.2

Food and non- alcoholic 105.1 104.6 102.3 101.9 101.4 100.7 101.1 100.6 100.3 100.2 99.9 101.6 101.7 beverages Housing, water, electricity, gas 108.4 106.0 105.6 105.4 105.5 106.1 106.1 101.5 99.9 99.3 99.0 100.1 103.6 and other fuels Healthcare 111.0 110.5 110.2 107.5 105.7 104.3 104.1 103.3 101.2 101.3 101.5 102.1 105.2 Source: (Geostat, 2016).

According to Geostat, the “food and non-alcoholic beverages” commodity group is comprised of 92 items (Geostat, 2017b) which comply with the products included in the agflation basket. It is noteworthy that this basket can differ from country to country given the more or less dependence of the local population on individual products. Further according to Geostat, the “healthcare” commodity group currently comprises 19 goods and services. The “utilities” commodity group comprises 17 products (Geostat, 2017b). To calculate munflation, all three groups of products included in the consumer basket need some adjustment. For example, from if we look at the “food and non-alcoholic beverages” group, soft drinks such as cola and other similar drinks can be withdrawn because they are generally not consumed by the poorer population. Further, we can remove clinical thermometers from the “healthcare” group (which, notwithstanding poverty, can be found in every family) and maternity services (which are government-funded in Georgia). Additionally, building materials (which are not generally used by lower-income segments of the population) can be taken out of the “utilities” commodity group. The issue of the possibility of extending the existing inflation targeting practice and studying the indicators of imflation, agflation and munflation in developing and relatively poorer countries, together with the inflation index, is the subject for a separate study.

5. Conclusions and Recommendations There is no reason to doubt the significance of the inflation index for economic development or the fact that the inflation indicator is used as a central bank’s target. For more than a quarter of a century, a rather rich experience of inflation targeting has been collected based on the experiences of many countries as well as serious research. Two groups of – opponents and supporters of inflation targeting – have been outlined. The critics of inflation targeting believe that the association of a central bank’s monetary policy with the planned inflation target alone; that is, for countries mostly dependent on import and where inflation is also imported together with the incoming goods, is one of its weak points (as evidenced by many studies) and fails to produce adequate results. Hence, for those countries whose economies are largely import-dependent, the importance of imported inflation, or imflation targeting, should be studied together with the inflation targeting. In its corresponding publications, the NBG also clearly confirms the difficulty in forecasting an inflation target because of the country’s so-called imported inflation. For over a decade, the agflation index has been calculated in world practice and this is of special importance for poorer countries where the food problem is especially acute. For developing poorer countries, the dynamics of average prices and not only for food but also for medication and utilities is of importance. This purpose is met by the munflation index

100 European Journal of Economic Studies, 2017, 6(2) which reflects the dynamics of the average prices of all three commodity groups deemed the most important for poorer groups of the population. The statistics office of a developing poorer country must calculate the imflation, agflation and munflation indices, together with the inflation index, which requires the development and practical implementation of a special methodology (especially for imflation and munflation). The calculation of these indices at more or less perfect levels creates an objective possibility for the central bank of a developing poorer country to diversify the inflation targeting system with the imflation, agflation and munflation targeting components. This should be a subject of an individual study.

References Anguridze et al., 2015 – Anguridze, O., Charaia, V., Doghonadze, I. (2015). Security Problems & Modern Challenges of the Georgian National Currency. Tbilisi, Ivane Javakhishvili Tbilisi State University. [Electronic resource]. URL: http://analytica.tsu.ge/www/files/2015/07/Security-Problems-Modern-Challenges-of-the- Georgian-National-Currency2.pdf Archer, 2000 – Archer, D.J. (2000). Inflation Targeting in New Zealand, International Monetary Fund, March 20-21. [Electronic resource]. URL: https://www.imf.org/external /pubs/ft/seminar/2000/targets/archer.htm Bank of Canada, 2006 – Bank of Canada (2006). Renewal of the Inflation-Control Target: Background Information – November 2006. Ottawa, Bank of Canada. [Electronic resource]. URL: http://www.bankofcanada.ca/wp-content/uploads/2010/06/background_nov06.pdf Bernanke et al., 1999 – Bernanke, B.S., Laubach, T., Mishkin, F.S., Posen, A.S. (1999). Inflation Targeting. Lessons from the International Experience. Princeton, Princeton University Press. BIS, 2010 – BIS (2010). Perspectives on Inflation Targeting, Financial Stability and the Global Crisis, Bank for International Settlements Papers, No 51, March. [Electronic resource]. URL: http://www.bis.org/publ/bppdf/bispap51.pdf Brash, 2002 – Brash, D. (2002). Inflation Targeting 14 Years On, Reserve Bank of New Zealand, January 5, [Electronic resource]. URL: http://www.rbnz.govt.nz/research-and- publications/speeches/2002/speech2002-01-05 CBN, 2017 – CBN (2017). Inflation Targets Table for 2017, Central Bank News,

101 European Journal of Economic Studies, 2017, 6(2)

Eurostat, 2016 – Eurostat (2016). Comparative Price Levels of Consumer Goods and Services. Eurostat, Statistics Explained, December. [Electronic resource]. URL: http://ec.europa.eu/eurostat/statistics-explained/index.php/Comparative_price_levels_of_ consumer_goods_and_services Fischer, Orr, 1994 – Fischer, A.M., Orr, A.B. (1994). Monetary Policy Credibility and Price Uncertainty: The New Zealand Experience of Inflation Targeting, OECD Economic Studies, No. 22, Spring. [Electronic resource]. URL: http://www.oecd.org/newzealand/33940940.pdf Frankel, 2012 – Frankel, J. (2012). The Death of Inflation Targeting, Project Syndicate, May 12, [Electronic resource]. URL: https://www.project-syndicate.org/commentary/the-death-of- inflation-targeting?barrier=true Geostat, 2016 – Geostat (2016). Consumer Price Indices in Georgia, National Statistics Office of Georgia, [Electronic resource]. URL: http://geostat.ge/cms/site_images/_files/ english/price/Consumer%20Price%20Index%20to%20the%20same%20month%20of%20previous %20year%20-%20eng.xls>. Geostat, 2017a – Geostat (2017). Distribution of Population Aged 15 and Older by Economic Status, 1998-2015. Employment and Unemployment, National Statistics Office of Georgia, [Electronic resource]. URL: http://geostat.ge/cms/site_images/_files/english/labour/new/ 01%20Distribution%20of%20population%20age%20of%2015%20and%20older%20by%20econo mic%20status.xls Geostat, 2017b – Geostat (2017). Inflation Calculation Methodology Note, National Statistics Office of Georgia, [Electronic resource]. URL: http://geostat.ge/cms/site_images/_ files/english/methodology/Inflation%20Calculation%20Methodology%20Note%202017.pdf Gurgenidze et al., 1994 – Gurgenidze, L., Lobzhanidze, M., Onoprishvili, D. (1994). Georgia: From Planning to Hyperinflation, Communist Economies & , vol. 6, no. 2. IMF, 2008 – IMF (2008). Agflation in the New Member States—Some Stylized Facts, IMF, [Electronic resource]. URL: https://www.imf.org/external/CEE/2008/120107.pdf IMF, 2016 – IMF (2016). Regional Economic Outlook: Caucasus and Central Asia, International Monetary Fund, October 21. [Electronic resource]. URL: http://www.imf.org/external/pubs/ft/reo/2016/mcd/eng/pdf/cca1016p.pdf Jones, 2016 – Jones, C. (2016). BIS Warns of too Much Focus on Inflation, The Financial Times, June 26. [Electronic resource]. URL: https://www.ft.com/content/5e4d0560-3a13-11e6- 9a05-82a9b15a8ee7 Kakulia, 2008 – Kakulia, M. (2008). Before and After the Introduction of the Lari: Georgian National Currency in Retrospect. In E. M. Ismailov, ed. Central Eurasia: National Currencies. Stockholm, CA&CC Press. Kakulia, Aslamazishvili, 2000 – Kakulia, M., Aslamazishvili, N. (2000). Dollarization in Georgia: Size of the Problem, Factors and the Ways of Solution. Tbilisi, Paata Gugushvili Institute of Economics, Georgian Academy of Sciences. (in Georgian.) Kakulia, Gigineishvili, 2005 – Kakulia, M., Gigineishvili, N. (2005). The Primary Objectives and Priorities of Monetary Policy in Georgia, Problems of Economic Transition, vol. 48, no. 4. Khaduri, 2005 – Khaduri N. (2005). Mistakes Made in Conducting Economic Reforms in Post-Communist Georgia, Problems of Economic Transition, Vol. 48, no. 4. Mashirova, Stepashova, 2015 – Mashirova, S.P., Stepashova, А.Yu. (2015). Analiz stoimosti potrebitel’skoi korziny” [Consumer Basket Cost Analysis], Molodoy Uchenyi, no. 1. [Electronic resource]. URL: http://moluch.ru/archive/81/14639/ Mishkin, Schmidt-Hebbel, 2007 – Mishkin, F., Schmidt-Hebbel, K. (2007). Monetary Policy under Inflation Targeting: An Introduction, Banco Central de Chile Working Papers, No. 396, December. [Electronic resource]. URL: http://si2.bcentral.cl/public/pdf/documentos- trabajo/pdf/dtbc396.pdf Nasdaq, 2017 – Nasdaq (2017). Crude Oil. WTI (NYMEX) Price, Nasdaq, [Electronic resource]. URL: http://www.nasdaq.com/markets/crude-oil.aspx?timeframe=3y NBG, 2017a – NBG (2017). Monetary Policy Report, February 2017. Tbilisi, National Bank of Georgia, [Electronic resource]. URL: https://www.nbg.gov.ge/uploads/publications/ inflationreport/2017/150217publish_MPR_February.pdf

102 European Journal of Economic Studies, 2017, 6(2)

NBG, 2017b – NBG (2017b). Money Aggregates and Monetary Ratios. Statistical Data, National Bank of Georgia, [Electronic resource]. URL: https://www.nbg.gov.ge/uploads/ depozitaricorporeisheniinglisurad/money_aggregates_and_monetary_ratioseng.xlsx NDI, 2017 – NDI (2017). NDI Poll: Economy Still Top Concern for Georgians; Support for NATO and EU Stable, National Democratic Institute, January 17, [Electronic resource]. URL: https://www.ndi.org/publications/ndi-poll-economy-still-top-concern-georgians-support-nato- and-eu-stable Papava, 1995 – Papava, V. (1995). The Georgian Economy: Problems of Reform, Eurasian Studies, vol., 2, no. 2. Papava, 1996 – Papava, V. (1996). The Georgian Economy: From ‘Shock Therapy’ to ‘Social Promotion,’ Communist Economies & Economic Transformation, vol. 8, no. 2. Papava, 2011 – Papava, V. (2011). On the First-Generation Post-Communist Reforms of Georgia's Economy (A Retrospective Analysis), The Caucasus & Globalization, vol. 5, issue 3-4. Plushchevskaya, 2012 – Plushchevskaya, (2012). O sostoyatel’nosti teoreticheskogo fundamenta targetirovaniya inflyatsii i novokeinsianskikh modelei obshchego ravnovesiya [On the Validity of the Theoretical Bases of Inflation Targeting and New Keynesian General Equilibrium Models], Voprosy ekonomiki, no.5. (in Russian.) Roser, Ospina, 2017 – Roser, M., Ospina, E. O. (2017). Global Extreme Poverty. Our World in Data. [Electronic resource]. URL: https://ourworldindata.org/extreme-poverty/ Sachs, 2005 – Sachs J.D. (2005). The End of Poverty. Economic Possibilities for Our Time. New York, Penguin Press. Sherwin, 1997 – Sherwin, M. (1997). Inflation Targeting—The New Zealand Experience. In Price Stability, Inflation Targets and Monetary Policy. Bank of Canada, May 1. [Electronic resource]. URL: http://www.bankofcanada.ca/1997/05/price-stability-inflation-targets-monetary- policy/, http://www.bankofcanada.ca/wp-content/uploads/2010/07/cn97-14.pdf Snooks, 2008 – Snooks, G.D. (2008). The Irrational “War on Inflation”: Why Inflation Targeting is Both Socially Unacceptable and Economically Untenable, The Australian National University Global Dynamic Systems Centre Working Papers, No. 1, March. [Electronic resource]. URL: https://0a85f5e1-a-62cb3a1a-s-sites.googlegroups.com/site/institutegds/home/w orkingpapers/WP001.pdf?attachauth=ANoY7cpy5w_EenI0XhuOlCm0jgEJXR4LDX3HCCXfX7Z7 JGYvDgnfSPdSztRpxIrI1mmHwCEgOSQTWWjquluaySUkycpsZ4LKHsLqXo28R7d1J-fgBc3 RIc0F3MqMRiN6nJXl6Gvu1YvgsiZJ0MEuwLEXDB1lo5LusA2jVc8Nzdup_kSbvpgBnVzLk7LnD4I 484y7gLlNN-5WgPhopLtVJP3WI co5mjlhpWWJodzM3uD3Gj42zfm_RX0%3D&attre directs=1 Spencer et al., 2006 – Spencer, G., Reddell, M., Hodgetts, B., Hunt, C., Bollard, A. (2006). Inflation Targeting: The New Zealand Experience and Some Lessons, Reserve Bank of New Zealand, January 18. [Electronic resource]. URL: http://www.rbnz.govt.nz/research-and- publications/speeches/2006/speech2006-01-18 Stiglitz, 2008 – Stiglitz, J. (2008). The Failure of Inflation Targeting, Project Syndicate, May 6. [Electronic resource]. URL: http://www.project-syndicate.org/print_commentary/stiglitz99/ Suryanarayana, 2008 – Suryanarayana, M.H. (2008). Agflation and the PDS: Some Issues, Indira Gandhi Institute of Development Research, WP-2008-009, April. [Electronic resource]. URL: http://www.igidr.ac.in/pdf/publication/WP-2008-009.pdf Vega, Winkelried, 2005 – Vega, M., Winkelried, D. (2005). Inflation Targeting and Inflation Behavior: A Successful Story? International Journal of Central Banking, vol. 1, no. 3. [Electronic resource]. URL: http://www.ijcb.org/journal/ijcb05q4a5.pdf Wang, 1998 – Wang, J.-Y. (1998). From Coupon to Lari: Hyperinflation and Stabilization in Georgia, Caucasica. The Journal of Caucasian Studies, vol. 1. Weber, 2015 – Weber, A. (2015). Rethinking Inflation Targeting, Project Syndicate, June 8. [Electronic resource]. URL: https://www.project-syndicate.org/commentary/rethinking-inflation- targeting-price-stability-by-axel-weber-1-2015-06?barrier=true Wellisz, 1996 – Wellisz, S. (1996). Georgia: A Brief Survey of Macroeconomic Problems and Policies, Studies & Analyses, Working Papers, No. 87. Warsaw, CASE. WHES, 2016 – WHES (2016). 2016 World Hunger and Poverty Facts and Statistics, World Hunger Education Service, September. [Electronic resource]. URL: http://www.worldhunger.org/2015-world-hunger-and-poverty-facts-and-statistics/#hunger- number

103 European Journal of Economic Studies, 2017, 6(2)

Copyright © 2017 by Academic Publishing House Researcher s.r.o.

Published in Slovak Republic European Journal of Economic Studies Has been issued since 2012. ISSN: 2304-9669 E-ISSN: 2305-6282 2017, 6(2): 104-114

DOI: 10.13187/es.2017.6.104 www.ejournal2.com

Study on Client-Satisfaction Factors in Construction Industry

Meliha Duljevic a , *, Mersid Poturak a a International Burch University, Bosnia and Herzegovina

Abstract Client satisfaction represents a crucial factor in the development and management of the construction process, as well in the creation of professional company-client relationships. Moreover, it is one of the major determinants of project success and therefore is a fundamental issue for construction managers who must constantly seek to improve their performance in order to survive in the marketplace. Providing superior quality and keeping customers satisfied are rapidly becoming the ways that companies use to differentiate themselves from competitors. The main objective of this study is to establish a comprehensive list of factors used for measuring client satisfaction and to study their influence on client satisfaction in the construction industry. The purpose of this study is to identify main client satisfaction factors and to advance both theoretical and practical understanding of their satisfaction in construction industry. A survey conducted in this study was focused on clients of different companies from construction industry in Bosnia and Herzegovina (B&H). The perceptions of clients with respect to the performance of contractors were measured using five factors including timeliness, cost, quality, client orientation and safety. Through the analysis of data generated by the survey, it is concluded that all the factors identified in the client-satisfaction model do not possess the same significance when it comes to satisfying clients. The approach of this research is useful to construction firms, not only in B&H, but also in other places, for identifying and improving their weak areas and improving the service quality for their clients. Keywords: client satisfaction, satisfaction factors, construction industry, Bosnia & Herzegovina.

1. Introduction Satisfaction is difficult to define and therefore there is little consensus of the definition of satisfaction. Locke (1970) asserted that satisfaction is a function of comparison between an individual’s perception of an outcome and its expectation for that outcome. Levels of satisfaction achieved hence are dependent on an individual’s perceptive thinking and is subjective in nature in the context of satisfaction measurement. For organizations, customer satisfaction is an effective approach to differentiate themselves from competitors and gain (Woodruff, 1997), but it is also one of the key issues in their efforts towards improving quality (Fornell et al., 1996). Companies use different forms of customer/client satisfaction approaches in developing and monitoring product/service offerings in order to manage and improve customer relationships (Burns & Bush, 2006). Likewise, measuring

* Corresponding author E-mail addresses: [email protected] (M. Duljevic), [email protected] (M. Poturak)

104 European Journal of Economic Studies, 2017, 6(2) customer satisfaction has numerous benefits for companies, such as improvement in communication between parties, enabling mutual agreement, assessment of progress toward the goals, and monitoring accomplished results and changes (Burns, Bush, 2006; Naumann, Giel, 1995). In the construction domain, satisfaction and client satisfaction in particular, plays a fundamental role in determining the perceived success of a project (Ashley et al., 1987; Bresnen, Haslam, 1991). In the construction industry, client satisfaction has remained an elusive and challenging issue for some considerable time (Banwell, 1964; Latham, 1994; Egan, 2002). However, the importance of customer satisfaction and orientation has increased due to the tightened competition and more demand from customers as a response to the industry’s poor performance. Client satisfaction in the construction industry can be defined as how well a contractor meets the client’s expectation. Satisfaction can be described in terms of a process of “expectancy disconfirmation” that is, the confirmation or disconfirmation of an expectation where satisfaction is founded mostly on meeting or exceeding expectations (Maloney, 2002). Customer satisfaction has been identified as a quality dimension in construction (Yasamis et al., 2002) and as an important factor representing success of a project (Chan, Chan, 2004; Delgado, Aspinwall, 2005). Customer satisfaction additionally can be considered as a method for expanding the construction process (Egan, 1998; Liu, Walker, 1998) and a tool for mutual learning (Love et al., 2000; Bertelsen, 2004). The main task of the construction industry is to provide clients with facilities that meet their needs and expectations. Assuring operational quality at every stage of the construction process should make sure that the quality of the final product will satisfy the final client (Jang et al., 2003). The subject matter of client satisfaction in the construction sector could be trace back to the 1980’s. According to research conducted by Ashley et al., (1987) on the determinants of the success of construction projects, six criteria intended for measuring success were highlighted. These include budget, schedule, client satisfaction, functionality, contractor satisfaction, and project-manager/team satisfaction. Therefore, the creation of a common client satisfaction measurement or approach is crucial in the construction industry. In spite of the fact that the construction industry has become aware of the importance of client satisfaction, it is equally important to know how well the industry is meeting client expectations. The main objective of this study is to identify a comprehensive list of factors for measuring client satisfaction and to study their influence on client satisfaction in the construction industry. Firstly, the literature is examined with the focus on the concept of satisfaction and its assessment and the survey was prepared. Then based on the results of the survey, the major factors influencing client satisfaction are identified. This study would contribute knowledge area by identifying the factors of clients’ satisfaction, criteria for measurement, and actual levels of satisfaction, as perceived by clients. For managers particularly this research offers information on groups of factors leading to high levels of customer satisfaction in construction business. They can use the information in allocating recourses and making better decisions on which factors to focus.

2. Literature review 2.1 Satisfaction Concept in Construction Industry Satisfaction is regarded as a function of comparison between an individual’s perception of an outcome and its expectation for that outcome (Locke, 1970), or a comparison of pre-purchase expectations and post-purchase product or service performance (Churchill, Serprenant, 1982). The construction industry is that part of the economy, which deals with the design, construction, maintenance, and utilization as well as with the modulation, modification and demolition, or deconstruction of constructs (Rußig et al., 1996). It is a major sector in most national economies and a major contributor to environmental changes, both in terms of designing the built environment as well as in terms of anthropogenic effects on the environment. In the construction industry, the measurement of client satisfaction is often associated with performance and quality assessment in the context of products or services received by the client (Parasuraman et al., 1985; Soetanto, Proverbs, 2004). A customer’s background and experiences play important roles in providing the relevant standards of comparison, or frame of reference (Smith et al., 1969). The comparison involves what the customer believes will happen with what is actually provided (Parasuraman et al., 1985). Different customers are likely to have different standards/expectations, which are pertinent to their judgment to the products or services.

105 European Journal of Economic Studies, 2017, 6(2)

Customer services literature suggests that a customer’s expectations and perceptions of performance have a direct effect on their satisfaction (Locke, 1970; Oliver, 1981; Parasuraman et al., 1985). A number of models have been developed to ease the measurement of satisfaction including SERVQUAL (Parasuraman et al., 1985; Siu et al., 2001), performance assessment (Soetanto, Proverbs, 2004) and business excellence models (EFQM, 2005).

2.2 Construction Industry in Bosnia & Herzegovina Construction can be defined as ‘the mobilization and utilization of capital and specialized employees, materials, and equipment to assemble materials and equipment on a particular site in accordance with drawings, specifications, and contract documents arranged to serve the purposes of the client’ (Merrit et al., 1996, p. 4.1). Construction sector in B&H is for many years in a very difficult state. Although in 1990’s this sector was employing about 100,000 people, and comprised between 10 % and 15 % of total B&H employment, today there are only 33,149 employees in the construction sector, making just 4.7 % of all employees in B&H (Statistic Agency, 2014). It is a compelling notion that in 2013, the value of completed construction work in B&H was merely 1,403,605,000 KM (Šehanović, 2008). When it is considered that there are 4.263 domestic construction companies in B&H (cumulative data taken from the data of Republic Agency for Statistics of RS and Federal Agency for Statistics of FB&H), it is obvious that the fight for the one’s market stake is tough. Because of that, it is necessary to use all the potential from all business strategies that can result with a positive outcome of that fight. The construction sector in Bosnia and Herzegovina has a long history and great potentials, especially when it comes to human resources. Most building materials are readily available in B&H. Virtually all of the components needed to repair existing structures and to build new ones, starting from motorways and airports to commercial premises and housing, are produced by the B&H businesses. This sector of the industry offers considerable scope for export development (FIPA, 2014). Currently, this sector amounts to 5,3 % of B&H exports in the total realized sale of the B&H products at the foreign market. Important note to emphasize for this sector that in 2011, according to FIPA (2014) it achieved the strongest nominal growth comparing to other sectors (4,8 %). 2.3 Client-Satisfaction Factors 2.3.1 Timeliness Clients, contractors and consultants often see timely completion of a construction project as a key criterion of project success. In the construction industry, the goal of project control is to ensure the projects finish on time, within budget and achieving other project objectives. It is a difficult task undertaken by project managers in practice, which involves continuously measuring progress; evaluating plans; and taking corrective actions when needed (Kerzner, 2003). NEDO (1983) states that a disciplined management effort is needed to complete a construction project on time, and that this determined management effort would help to control both costs and quality. This is equivalent to saying that the client’s objectives can be achieved through a management effort that recognizes the interdependence of time, cost and quality. During the last few decades, numerous project control methods, for instance Gantt Bar Chart, Program Evaluation and Review Technique (PERT) and Critical Path Method (CPM), have been developed (Nicholas, 2001; Lester, 2000). Even with the wide use of these methods and software packages in practice, numerous construction projects still suffer time and cost overruns. 2.3.2 Cost Cost as defined by Stewart (1991) represents the total amount of all the resources necessary to perform the activity. “Cost is among the main considerations during the project management life cycle and can be regarded as one of the most important parameters of a project and the driving force of project success” (Azhar et al., 2008: 7). Gido and Clements (2003) asserted that cost performance is an effective technique in project management and it is widely accepted in the literature and industry. Ashworth (1994) stated that one of the client’s requirements when it comes to construction project is the estimation of its expected cost. Appropriate cost control is important since it is the general trend towards greater cost-effectiveness and guarantees construction costs not only in the context of initial costs, but with regard to life-cycle costs or total cost appraisal. Cost estimation presents the base for project management, business planning, budget preparation as well as cost and schedule control (Marjuki, 2006).

106 European Journal of Economic Studies, 2017, 6(2)

2.3.3 Quality A project might be completed on time and within budget, however unless it achieves the specified quality or performance criterion it will be considered to be a disappointment or even a complete failure. The concept of quality is multidimensional and incorporates aspects that may be appraised subjectively. The Latham Report (1994) distinguished a number of quality aspects that clients may look for in a satisfactory construction project: „pleasant to look at; free from defects on completion; fit for the purpose; supported by worthwhile guarantees; satisfactory durability and customer delight“. Some of these aspects are inherent in the design of the project, while others relate to how successfully the contractor constructs that design on site. From the viewpoint of a construction company, quality management in construction projects should mean maintaining the quality of construction works at the required standard to achieve customers’ satisfaction that would bring long-term competitiveness and business survival for the companies (Tan, Abdul-Rahman, 2005). In addition to the abovementioned, Harris and McCaffer, (2001) stated that quality management practices incorporates all the means employed by managers in an effort to implement their quality policies. These activities comprise quality planning, quality control, quality assurance and quality improvement. 2.3.4 Client Orientation The emergence of the customer as the champion for change has increased the pressure on the construction industry to offer the higher quality along with better service in order to satisfy customer needs and expectations. Rising customer expectations and greater competitiveness has become a key characteristic of the construction industry in developed economies over the last decade (Copare, 1992; Raftery et al., 1998). The term ‘customer/client’ should be comprehended in its wider sense to include all parties and individuals who would influence the character, scope and nature of the product or service that the business needs to provide. Newcombe (1999) introduced a view of construction projects as a coalition of powerful individuals and groups, the stakeholders, who are by definition the clients of that specific project. 2.3.5 Safety Safety is an economic as well as humanitarian concern that requires appropriate management control. Safety and health must be managed in the same way that other aspects of a company are managed (Peterson, 1979). Benefits of safety and health may embrace less injuries, less property damage, less down time, improvement in morale, developed industrial relations, enhanced productivity, reduced cost and improved quality (Promfret, 1997). Additional benefits include less compensation insurance, fewer hidden costs, better supervisor morale, increased efficiency, and improved marketability (Levitt, Samelson, 1995). Hinze and Parker (1978) stated that good safety performance and high productivity are compatible and that safety should not be sacrificed in an attempt to enhance productivity. In identifying factors that influence performance, Chan et al. (2004) identified the implementation of an effective safety programme as a critical success factor of construction projects. Assaf et al. (1996) also identified adherence to safety rules and regulations within such programme as essential.

3. Research methodology Both primary and secondary data have been utilized in this research. Primary data of this research was obtained directly via a survey, while secondary data was based on different books, journals, reports, and the like. Since the study required the collection of data from a wide range of clients who have experience with local contractors, an online survey was designed and conducted. The survey was designed as a research instrument to examine the levels of client satisfaction as perceived by clients based on consultant performance using a series of satisfaction determinants, as developed in earlier satisfaction assessment models (Parasuraman et al., 1985; Soetanto, Proverbs, 2004; EFQM, 2005). The research sample has been drawn from building construction clients, who had one or more than one building projects completed in the last 15 years to make sure that they have the relevant knowledge and experience to accurately answer the survey. The survey-making process lasted for three months. Above mentioned surveys were collected in the period of one month and few days, precisely speaking from the 24th of April until the 30th of May, 2017. The surveys were sent to managers of different companies who were clients of some of B&H construction companies. The total number of collected surveys is 165. The survey was consisted of close-ended questions, in which rating questions were asked. Saunders et al. (2009) stated that

107 European Journal of Economic Studies, 2017, 6(2) this type of questions employ the Likert-style rating scale, where participants are supposed to mark one of existing statements. All statements have five levels that illustrate the participants’ opinion about it. The research model is developed from a detailed overview of literature review in order to identify the factors and their influence on client satisfaction in construction industry. The model proposed in this study is not an exhaustive one it can be further extended by adding other variables we have not made reference to. The key elements of the model are: 1. Independent Variables (IV) a. Timeliness b. Cost c. Quality d. Client Orientation e. Safety

2. Dependable Variable (DV) a. Client Satisfaction

4. Data analysis and results 4.1 Descriptive statistics In the first part of the survey, respondents were asked several demographic questions. The following tables show the summarized results obtained from those questions. Table 4.1 shows the gender distribution for the sample. The majority of the respondents surveyed were male (70,3 %) while female respondents occupied a smaller portion compared to them (29,7 %).

Table 4.1. Gender of the Respondents

Frequency Percent Valid Percent Cumulative Percent Valid Male 116 70.3 70.3 70.3 Female 49 29.7 29.7 100.0 Total 165 100.0 100.0

In terms of age groups, the respondents aged below 25 comprise 15,8 % of the total, those 25 to 34 years 18,8 %, those 35 to 44 years 32,1 %, and 45 and over groups 33,3 % (Table 4.2).

Table 4.2. Age of the Respondents

Frequency Percent Valid Percent Cumulative Percent Valid Below 25 26 15.8 15.8 15.8 25-34 years old 31 18.8 18.8 34.5 35-44 years old 53 32.1 32.1 66.7 45 and above 55 33.3 33.3 100.0 Total 165 100.0 100.0

Table 4.3 shows that about 2,4 % of the respondents were educated up to elementary school and 24,8 % were educated up to high school. The number of respondents attaining higher education included 51,5 % of the respondents were educated up to the undergraduates level. A considerable number (13,3 %) of respondents have completed Master level, and relatively lesser number of them (7,9 %) have a PhD degree.

108 European Journal of Economic Studies, 2017, 6(2)

Table 4.3. Level of Education of the Respondents

Frequency Percent Valid Percent Cumulative Percent Valid Elementary school 4 2.4 2.4 2.4 High school 41 24.8 24.8 27.3 Undergraduate 85 51.5 51.5 78.8 Master 22 13.3 13.3 92.1 PhD 13 7.9 7.9 100.0 Total 165 100.0 100.0

4.2. Validity of survey questions Back Translation technique was used in this research as a validation process, where the survey was first translated into Bosnian language, and then translated back into the original English language, by a different person. The objective was to ensure that the original translation is accurate. The term “back translation” is used in survey research literature and in translation studies to refer to the translation of a translation back into the source language (Harkness, 1996). The basic steps involved in this process were as follows: 1. A source text in English language (Source Language Text One, SLT1) was translated into Bosnian language (Target Language Text or TLT). 2. The TLT was translated back into the language of SLT1 by a second translator, unfamiliar with the SLT1 and uninformed that there is an SLT1. This second translation, the back translation, is SLT2. 3. SLT1 was compared to SLT2. 4. Based on the similarities between SLT1 and SLT2, conclusion was drawn that there was a great equivalence between the TLT and the SLT1.

4.3. Reliability assessment Cronbach’s coefficient alpha analysis is the most widely used formula for assessing the internal consistency of measures in marketing research (Peter, 1979). A low coefficient indicates that the sample items have not been able to capture the construct, while a large alpha coefficient indicates that the given item correlates well with the true scores. Cortina (1993) and Kline (1999) have argued that an acceptable value for Cronbach’s alpha could reach around and above 0.7 (0.65 to 0.84); values significantly lower than 0.7 indicate an unreliable construct. In this study, Cronbach’s alpha is measured for the second and third part of the survey that contains 19 items in total. From the table below we can see that all items are reliable since their Cronbach’s alpha coefficients are higher than standard value of 0.70.

Table 4.4. Reliability Test

Variables Cronbach's Alpha Cronbach's Alpha N of Items Based on Standardized Items Timeliness .837 .840 3 Cost .858 .862 3 Quality .855 .855 3 Client Orientation .820 .824 3 Safety .886 .886 3 Client Satisfaction .874 .875 4

4.4 Hypothesis Testing Hypotheses raised in this study are as follows:

H1: Timeliness positively influences client satisfaction with a construction company. H2: Cost positively influences client satisfaction with a construction company.

109 European Journal of Economic Studies, 2017, 6(2)

H3: Quality positively influences client satisfaction with a construction company. H4: Client orientation positively influences client satisfaction with a construction company. H5: Safety positively influences client satisfaction with a construction company.

Table 4.5. Regression Testing Hypotheses

Model Unstandardized Standardized Coefficients Coefficients B Std. Error Beta t Sig. 1 (Constant) 1.329 .240 5.528 .000 Time .179 .086 .203 2.086 .039 Cost -.020 .094 -.021 -.208 .835 Quality .248 .104 .260 2.378 .019 Client_Orientatio -.069 .084 -.071 -.820 .413 n Safety .373 .078 .445 4.782 .000

In order to test the hypotheses a Liner Regression method was used. Based on the obtained results the following conclusions are drawn: From the Table 4.11 we can see that our first independent variable „Timeliness“ has a level of significance 0.039 which is smaller than 0.05. Therefore, we can state that Hypothesis 1 is accepted. In other words, we claim that timeliness factor is positively related to client satisfaction with a construction company. Second independent variable „Cost“ has a level of significance 0.835 which is higher than 0.05 and we can state that Hypothesis 2 is rejected. In other words, we claim that cost factor is negatively related to client satisfaction with a construction company. Third independent variable „Quality“ has a level of significance 0.019 which is smaller than 0.05. Therefore, we can state that Hypothesis 3 is accepted. In other words, we claim that quality factor is positively related to client satisfaction with a construction company. Fourth independent variable „Client orientation“ has a level of significance 0.413 which is higher than 0.05, therefore we can state that Hypothesis 4 is rejected. In other words, we claim that client orientation factor is negatively related to client satisfaction with a construction company. Fifth independent variable „Safety“ has a level of significance 0.000 which is smaller than 0.05 (<0.05). Therefore, we can state that Hypothesis 5 is accepted. In other words, we claim that safety factor is positively related to client satisfaction with a construction company.

5. Conclusion The main objective of this study was to identify a comprehensive list of factors for measuring client satisfaction and to study their influence on client satisfaction in the construction industry. Based on the previous similar researches in the context of construction, the following five factors were identified to be the focus of the study. Those factors included timeliness, cost, quality, client orientation and safety at work. Data generated through the client-satisfaction survey was analyzed and certain hypotheses were tested. Out of five hypotheses, three of them were accepted, which indicated that three factors are considered important for client satisfaction, and that they actually have a positive impact on client satisfaction. The first factor that proved to have positive effects on the clients' satisfaction is timeliness. In the last two decades, the market is characterized with many projects in the private sector where the deadlines play a very important role to the private clients because this actually represents the realization of their ideas. The results indicated that the goal of each client is the fastest possible completion of works. Likewise, the results of the research suggest that companies that want to have satisfied clients must take into account this phenomenon, because in the following period the effectiveness of the planning and scheduling jobs will continue to play a major role in the eyes of clients. The second factor that showed significant positive effect on clients’ satisfaction was cost. The test results showed that the quality factor in construction is essential for each client in terms of

110 European Journal of Economic Studies, 2017, 6(2) visual appearance, then with regard to fulfilling the clients' needs in the use of the facility, as well as the life safety of people inside the building. Based on the obtained results, contractors are recommended to take into consideration clients' wishes, needs as well as the fears, during the negotiation and the execution of works. Contractors should convince the client that he knows the materials used for the present work, help him in the selection of materials, and share his own experiences related to materials. In addition, the contractor needs to convince the client that the continual supervision is maintained over work execution, and to be aware of quality as an essential dimension of the overall client satisfaction. The third factor, in addition to timeliness and quality that was found to have an extreme importance to clients’ satisfaction is safety. Safety in construction can be accomplished by working in accordance with certain measures and standards. All results from this study suggest that when it comes to health and safety measures and their application, they not only reassure the client, but also significantly improve the image and reputation of the contractor. Contractors are advocated that in order to protect their workers, their own as well as clients' interests to perform regular training of employees on work safety, and to introduce a system that guarantees the application of the regulations related to safety. Also, contractors should pay special attention to supply the workplaces or building sites with the necessary equipment and plan on applying safety measures. Contractor, who maintains a safe, clean and organized work environment, reassures the client, is regarded as a reliable partner, and is preferable in the range of clients. The analysis showed that by failing to accept two hypotheses, the conclusion drawn is that not all client-satisfaction factors are perceived to be equally important by the clients. Specifically, two of the client-satisfaction factors, cost and client orientation, showed significant differences. Regarding cost, we assume that this factor is not considered unimportant, but that the respondents evaluated it in this way because many of them during the planning phase, preparation of project documentation and contracting of works phase, do not have a clear vision of all the works and materials that will be needed. Therefore, clients partially change their ideas and decisions in terms of material selection during the execution of works. It can be concluded that the majority of clients, when it comes to the selection of materials, give more importance to their wishes, than to keep it within the agreed budget. Lastly, results showed that the relationship between the contractors and the client does not belong to the forefront of the decisive factors. One of the reasons why respondents consider this factor important but not crucial, is that clients maybe consult engineering companies for professional advising when it comes to gathering the necessary documentation, rather than consulting construction companies. Although the findings of the study contribute to the body of knowledge, there are certain limitations of the study. Firstly, the study has focused on the clients of the construction industry in B&H and this restricts the generalization of the study. Also, a model of the clients' satisfaction developed in the thesis contains five factors taken for the purpose of the study. There may be several other factors that may have impact on clients' satisfaction, and hence this lead to another limitation of the study. If there are future researches to come, the research researcher can use qualitative methods to have a better understanding customer satisfaction, since this research was only based on quantitative method. Similarly, it would be interesting if others dimensions are also added in the future researches except for the ones used in this research, or if different situational and control variables are used in future research. Even though the above mentioned findings are based on input from the B&H construction market, we believe that the contractor firms in other countries, who specialize in building works for private sector may also benefit from the findings or at least the approach of this research as well. Though it needs to be remembered that the configurations that are the outcome of this research can only serve as one aspect in enhancing the understanding of the factors related to clients’ satisfaction.

References Ahmed, Kangari, 1995 – Ahmed, S.M., Kangari R. (1995). Analysis of Client-Satisfaction Factors in Construction Industry. Journal of Management in Engineering, Vol. 11, No. 2, pp. 36-42. Ashley, 1987 – Ashley, D.B., Luirie, C.S. and Jaselskis, E.J. (1987). Determinants of construction project success, Project Management Journal, Vol. XVII No. 2, pp. 69-79.

111 European Journal of Economic Studies, 2017, 6(2)

Ashworth, 1994 – Ashworth, A. (1994). Cost studies of buildings. Essex: Longman Group Limited. Assaf et al., 1996 – Assaf, S.A., Al-Hammad, A.M., Ubaid, A. (1996) Factors effecting construction contractors' performance, Building Research and Information, 24 (3), 159-163. Azhar et al., 2008 – Azhar, N., Rizwan U. Farooqui, Ahmed, S.M. (2008). Cost overrun factors in construction industry of Pakistan. Advancing and Integrating Construction Education, Reseach & Practice, 499-508. Banwell, 1964 – Banwell, G.H. (1964). The Placing and Management of Contracts for Building and Civil Engineering Work, HMSO, London. Bertelsen, 2004 – Bertelsen, S. (2004). Lean Construction: Where are we and how to proceed? Lean Construction Journal. Vol.1, pp. 46-69. Bresnen, Haslam, 1991 – Bresnen, M.J., Haslam, C.O. (1991). Construction industry clients: a survey of their attributes and project management practices. Construction Management and Economics, Vol. 9, pp. 327-42. Burns, Bush, 2006 – Burns, A., Bush, R. (2006). Marketing Research, 5th ed., Pearson Prentice-Hall, New York, NY. Chan, Chan, 2004 – Chan A. P. C., Chan, A. P. L. (2004). Key performance indicators for measuring construction success. Benchmarking: An International Journal, 11(2), 203-221. Chan et al., 2004 – Chan, A. P. C., Scott, D., Chan, A. P. L. (2004). Factors Affecting the Success of a Construction Project. Journal of Construction Engineering and Management, 130 (1), 153-155. Cheng et al., 2006 – Cheng, J., Proverbs, D. and Oduoza, C. (2006). The satisfaction levels of UK construction clients based on the performance of consultants: results of a case study. Engineering Construction and Architectural Management, Vol 13, No.6, pp567-583. ISSN 0969- 9988. Churchill, Serprenant, 1982 – Churchill, G.A., Serprenant, C. (1982). An investigation into the determinants of consumer satisfaction. The Journal of Marketing Research, Vol. 19, pp. 491-504. Copare, 1992 – Copare, P.B. (1992). Planning your construction services. Transactions of the American Association of Cost Engineers, American Association of Cost Engineers, Morgantown. Cortina, 1993 – Cortina J. (1993). What is coefficient alpha? An examination of theory and applications. Journal ofApplied Psychology, 78, 98–104. Delgado, Aspinwall, 2005 – Delgado D. J., Aspinwall A.M. (2005). Improvement tools in the UK construction industry. Construction Management and Economics, Vol. 23 Issue 9, pp. 965. De Vaus, D.A. (1994), Surveys in Social Research, 3rd ed., UCL Press, Guildford. Dysert, 2006 – Dysert, L. R. (2006). Is "estimate accuracy" an oxymoron? AACE International Transactions, s.l.: EST01.1 - 01.5 EFQM – EFQM (2005). European Forum for Quality Management, available at: http://www.efqm.org/ (accessed 17 January 2017). Egan, 2002 – Egan, J. (2002), Accelerating Change, Construction Industry Council, London. Egan, 1998 – Egan, J. (1998). Rethinking Construction, Department of the Environment, Transport and the Regions, London. Fornell et al., 1996 – Fornell, C., Johnson, M.D., Anderson, E.W., Cha, J. and Bryant, B.E. (1996). The American customer satisfaction index: nature, purpose, and findings. Journal of Marketing, Vol. 60, pp. 7-18. Gido, Clements, 2003 – Gido, J., Clements, J.P. (2003). Successful project management. New York: South-Western. Harkness, 1996 – Harkness, J.A. (1996). Thinking Aloud about Survey Translation. Paper presented at the International Sociological Association Conference on Social Science Methodology, Colchester. Harris, McCaffer, 2001 – Harris, F., McCaffer, R. (2001). Modern Construction Management. 5th ed. London: Blackwell Publishing Hinze, Parker, 1978 – Hinze, H., Parker, H. (1978). Safety: productivity and job pressures. Journal of construction Division, ASCE, Vol. 104, No. 2. pp. 241-249. Kärnä, 2004 – Kärnä, S., Junnonen J.M., Kankainen J. (2004). Customer Satisfaction in Construction, 12 p.

112 European Journal of Economic Studies, 2017, 6(2)

Kerzner, 2003 – Kerzner, H. (2003). Project Management A Systems Approach to Planning, Scheduling, and Controlling. John Wiley and Sons InC., New Jersey. Kline, 1999 – Kline P. (1999). The Handbook of Psychological Testing. Routledge, London. Latham, 1994 – Latham, M. (1994). Constructing the Team, HMSO, Department of Environment, London. Latham, 1994 – Latham, M (1994). Constructing the Team: Final Report of the Government/Industry Review of Procurement and Contractual Arrangements in the UK Construction Industry, HMSO, London. Lester, 2000 – Lester, A. (2000). Project Planning and Control. Butterworth Heinemann, Oxford. Levitt, Samelson, 1995 – Levitt, R.E., Samelson, N.M. (1995) Construction safety management, John Wiley & Sons Inc. New York. Locke, 1970 – Locke, E.A. (1970). Job satisfaction and job performance: a theoretical analysis. Organizational Behavior and Human Performance, Vol. 5, pp. 484-500. Love et al., 2000 – Love, P., Smith, J., Treloar, G. and Li, H. (2000). Some empirical observations of service quality in construction. Engineering, Construction and Architectural Management, Vol. 7 No. 2, pp. 191-201. Maloney, 2002 – Maloney, W.F. (2002). Construction product/service and customer satisfaction. Journal of Construction Engineering and Management, November/December, pp. 522-529. Marjuki, 2006 – Marjuki, M. (2006). Computerized building cost estimating system, Malaysia: University Teknologi Malaysia. Merrit et al., 1996 – Merrit, F.S., Loftin, M.K., & Ricketts, J.T. (1996). Standard handbook for civil engineers (4th ed.). New York: McGraw Hill. National Economic..., 1983 – National Economic Development Office (N.E.D.O.) (1983). Faster Building for Industry. Her Majesty’s Stationery Office, London. Naumann, Giel, 1995 – Naumann, E., Giel, K. (1995). Customer satisfaction measurement and management: using the voice of the customer. International Thomson Publishing, USA. Newcombe, 1999 – Newcombe, R. (1999). From client to project stakeholders: a stakeholder mapping approach. Proceedings of the CIB W55 and W65 Joint Triennial Symposium, pp. 57–67, Cape Town, Republic of South Africa, 5–10 September. Nicholas, 2001 – Nicholas, J. (2001). Project Management for Business and Technology. Prentice Hall, New Jersey. Oliver, 1981 – Oliver, R.L. (1981). Measurement and evaluation of satisfaction process in retail setting. Journal of Retailing, Vol. 57, pp. 25-48. Parasuraman et al., 1985 – Parasuraman, A., Zeithaml, V.A. and Berry, L.L. (1985). A conceptual model of service quality and its implications for future research. Journal of Marketing, Vol. 49 No. 4, pp. 41-50. Peter, 1979 – Peter, J. Paul (1979), "Reliability: A Review of Psychometric Basics and Recent Marketing Practices," Journal of Marketing Research, 16 (February), 6-17. Peterson, 1979 – Peterson, D. (1979). The OSHA compliance manual, McGraw Hill Book Co., Inc., New York. Promfret, 1997 – Promfret, B. (1997) Psychology of safety, National safety. pp. 14-15. Raftery et al., 1998 – Raftery, J., Pasadilla, B., Chang, Y.H., Hui, E.C.M. and Tang, B.S. (1998). Globalization and construction industry development: implications for recent developments in the construction sector in Asia. Construction Management and Economics, 16, 729–37. Rußig et al., 1996 – Rußig, V., Deutsch, S., Spillner, A. (1996). Branchenbild Bauwirtschaft: Entwicklung und Lage des Baugewerbes sowie Einflussgrößen und Perspektiven der Bautätigkeit in Deutschland, Ducker&Humboldt: Berlin-München. Siu et al., 2001 – Siu, K.W., Bridge, A. and Skitmore, M. (2001). Assessing the service quality of building maintenance providers: mechanical and engineering services. Construction Management and Economics, Vol. 19, pp. 719-26. Smith et al., 1969 – Smith, P.C., Kendall, L.M. and Hulin, C.L. (1969). The Measurement of Satisfaction in Work and Retirement: A Strategy for the Study of Attitudes, Rand McNally & Company, Chicago.

113 European Journal of Economic Studies, 2017, 6(2)

Soetanto, Proverbs, 2004 – Soetanto, R., Proverbs, D.G. (2004), Intelligent models for predicting levels of client satisfaction. Journal of Construction Research, Vol. 5 No. 2, pp. 233-55. Statistic Agency B&H, 2014 – Statistic Agency B&H (2014). Construction works in 2013, November. Stewart, 1991 – Stewart, R. D. (1991). Cost Estimating. 2nd ed. New York: John Wiley & Sons, Inc. Šehanović, 2008 – Šehanović S. (2008). Uticaj imidža na poslovanje građevinskih preduzeća u Bosni i Hercegovini, Master thesis, Univerzitet u Tuzli, Filozofski fakultet. Tan, Abdul-Rahman, 2005 – Tan, C. K., & Abdul-Rahman, H. (2005). Preliminary Research into Overcoming Implementation Problems in Construction Projects.Proceeding of the 4th Micra Conference.Faculty of the Built Environment, University Malaya, 08.15-08.28 Willis et al., 2012 – Willis, P.G., Brown, K.A. & Prussia, G.E. (2012). Does employee safety influence customer satisfaction? Evidence from the electric industry. Journal of Safety Research, Vol. 43, No. 5-6, pp. 389-396. Woodruff, 1997 – Woodruff, R.B. (1997). Customer value: the next source for competitive advantage”, Journal of the Academy of Marketing Sciences, Vol. 25 No. 2, pp. 139-53. Yasamis et al., 2002 – Yasamis, F., D. Arditi and J. Mohammadi (2002). Assessing contractor quality performance. Construction Management and Economics 20, pp. 211-223.

114 European Journal of Economic Studies, 2017, 6(2)

Copyright © 2017 by Academic Publishing House Researcher s.r.o.

Published in Slovak Republic European Journal of Economic Studies Has been issued since 2012. ISSN: 2304-9669 E-ISSN: 2305-6282 2017, 6(2): 115-123

DOI: 10.13187/es.2017.6.115 www.ejournal2.com

Enablers of Successful Knowledge Sharing Behavior: KMS, Environment and Motivation

Alparslan Özlen a , * a Gebze Technical University, Turkey

Abstract Knowledge sharing is suggested as a key element for Knowledge Management in sustaining organizational competitiveness. This work investigates the relationships proposed by a knowledge sharing model implying that knowledge sharing practices contribute to organizational and individual performance as a result of (a) qualified Knowledge Management Systems, (b) suitable knowledge sharing environment and (c) organizational knowledge sharing motivation. The proposed model is tested by using the data obtained from surveying various private and public Bosnian enterprises. At the end of data collection period, 207 usable surveys are achieved. According to the results, Knowledge Management in Bosnia is still developing yet. Model test suggests advanced Knowledge Management Systems, suitable knowledge sharing environment and high organizational knowledge sharing motivation influence knowledge sharing and successful knowledge sharing increases the performance of both individuals and the organization. The results suggest that successful knowledge sharing can be achieved through considering technical (KMS), social (environment) and the individual (motivation). Keywords: knowledge management implementation, knowledge sharing, individual performance, organizational performance

1. Introduction Dynamic and competitive Knowledge-based economy requires ability to transform knowledge resources to organizational survival and competitiveness. This encourages researchers and practitioners to analyze the organizational ability in identifying, capturing, creating, sharing and accumulating knowledge (Nonaka, Takeuchi, 1995) as knowledge management processes. Knowledge Management (KM) fundamentally aims to maximize the flow of existing knowledge through individuals and organizations which are strongly dependent upon individuals’ knowledge sharing (KS) behavior (Bock et al., 2005). Successful knowledge sharing is supposed to contribute to the organizational performance (Argote et al., 2000) and organizational effectiveness (Alavi, Leidner, 2001). The literature reports few Knowledge Management studies for Bosnia and Herzegovina concentrated more on the implementation level of Knowledge Management and its adoption. They report weak knowledge management understandings of Bosnian organizations (Handzic, Lagumdzija, Celjo, 2007; Biloslavo, Kljajic-Dervic, 2011; Bartlett et al., 2012; Ozlen et al., 2012)

* Corresponding author E-mail addresses: [email protected] (A. Özlen)

115 European Journal of Economic Studies, 2017, 6(2) and suggest more to enhance Knowledge Management for those organizations in terms of measurement and technology (Handzic et al., 2007), Knowledge Management strategies (Ozlen et al., 2012). However, the literature is weak in providing about knowledge sharing behavior of Bosnian enterprises. This study aims to strengthen existing KM literature by evaluating some enablers of successful knowledge sharing behavior for Bosnian enterprises. The proposed knowledge sharing model suggest that successful knowledge sharing leverages organizational and individual performance as a result of (a) qualified Knowledge Management Systems (KMS), (b) appropriate knowledge sharing environment and (c) high organizational knowledge sharing motivation. In order to test the model, the data is collected by surveying Bosnian public and private enterprises. Further sections provide the relevant literature, the research model, the research methodology, the findings and the discussion of the findings.

2. Discussion Knowledge Sharing Environment Knowledge sharing practices are extremely important in keeping and enhancing gained valuable intellectual capital and therefore organizational success. Hence, the identification of influencing factors and the outcomes of these practices is necessary. The literature suggests culture, structures, and technology as the environmental antecedents of knowledge sharing (Alavi et al., 2006). KM is determined by social (Ribiere, Sitar, 2003) and/or technical (Tsui, 2003) elements in enhancing knowledge processes and therefore working knowledge and finally advanced performance. Social factors are identified to have greater importance than technical factors to enhance organizational knowledge management (Handzic, 2011). The literature suggests organizational culture as one of the main determinants of knowledge sharing (Alavi, Leidner, 1999). Modern technologies for open communication and knowledge acquisition require networked structures (Handzic, 2011). Moreover, individualistic cultures are suggested for knowledge acquisition, while cooperative cultures are necessary for high knowledge sharing (Alavi, Leidner, 1999). Effective organizational management creates an enabling environment for knowledge generation and supports collaboration and knowledge sharing (Fink, 2000). A variety of measures such as rewards and incentives, and ensured management commitment are necessary in developing a knowledge sharing culture (Handzic, 2011). Development of a knowledge sharing culture as the best strategy for KM program are encouraged through (1) leading by example; (2) branding KM through incentives such as kind messaging, formal communications, and rewards and recognition and (3) making KM fun (O'Dell, Hubert, 2011). Therefore, a supportive organizational culture as a knowledge sharing facilitator is required to be satisfied in leveraging the interactions among knowledge workers. In this research, we use the term knowledge sharing environment instead of supportive organizational culture. The literature also suggests information technology as an important factor for establishing a knowledge sharing platform (Hahn, Subramani, 2000). Supportive technical environment increases the collaboration among the people (O'Dell, Hubert, 2011). Knowledge Management Systems (KMS) (a type of information systems) are supportive technologic knowledge sharing instruments. A flexible corporate infrastructure is necessary for enterprise- based knowledge management systems for instant, ad hoc and intensive collaborations (Liu et al, 2005). Furthermore, KMS is recommended as an enabler for KMS use (Jennex, Olfman, 2004, 2005, 2006; Jennex, 2008) in increasing knowledge sharing. Final antecedent variable is the motivation for sharing knowledge which needs to be evaluated for successful knowledge sharing (Gu, Gu, 2011). Motivators and demotivators are influential for organizational knowledge sharing (Oye et al, 2011). Task routineness and open communication improve only mandatory sharing behaviors and solidarity sharing is enhanced by voluntary sharing behaviors (Teng, Song, 2011). Teh and Yong (2011) observe that (a) sense of self-worth and in-role behavior are positively related to the attitude toward knowledge sharing; (b) both subjective norm and organizational citizenship behavior are independent and positively related to intention to share knowledge; (c) but attitude toward knowledge sharing is negatively related with intention to share knowledge; and

116 European Journal of Economic Studies, 2017, 6(2)

(d) individuals’ knowledge sharing behavior is influenced by intention to share knowledge. Intrinsic motivation and joint relationships and interpersonal interactions among employees are suggested to facilitate successful knowledge sharing (The, Yong, 2011). Consequently, this study employs knowledge sharing environment, Knowledge Management Systems and knowledge sharing motivation as the determinants of superior Knowledge Sharing.

Knowledge sharing Vygotsky’s (1978) socio-cultural theory of learning suggests knowledge sharing and social interaction by the social/individual and the public/private mechanisms for knowledge acquisition and representation. Learning is supposed to be started on the social environment through the interactions between learners and the expert knowledge holders (Vygotsky, 1978). Individual learners take the concepts and strategies to other contexts and meanings and interpret them with social interactions. Consequently, learning starts in the public through the use of knowledge. Therefore, individuals understand, adjust, and implement the learned knowledge in their private domain. As a result of Polanyi’s (1966) conceptualization, SECI model (Socialization, Externalization, Combination, and Internalization) is proposed by Nonaka and Takeuchi (1995) in order to explain tacit and explicit knowledge sharing in creating knowledge. After knowledge sharing processes, organizational knowledge is transformed into individual or group knowledge through internalization and socialization while individual and group knowledge are transformed into organizational knowledge through externalization and combination. The literature reports that knowledge sharing is suggested as a fundamental knowledge- centered activity through which employees can mutually exchange their knowledge and contribute to knowledge application and ultimately the competitive advantage of the organization (Wang, Noe, 2010). This research uses knowledge sharing as the central variable for the proposed research model.

Performance Variables Knowledge sharing has been extensively evaluated for the organizational KM including KS effectiveness in knowledge networks (Hansen, 2002), on individual performance (Teigland, Isko, 2003) and on organizational performance (Argote et al., 2000). Wang and Wang (2012) suggest that both explicit and tacit knowledge sharing practices enhance performance and innovation by contributing to firm performance. According to their results, while explicit knowledge sharing is found to be more significantly influencing innovation speed and financial performance, tacit knowledge sharing is observed to have more significant effect on innovation quality and operational performance. KMS use (knowledge sharing with knowledge sharing instruments) is considered as the influencing factor of success in KMS success literature (Jennex, Olfman, 2004, 2005, 2006; Jennex, 2008). According to Wang and Wang (2012), there are few studies measuring the direct relationship between knowledge sharing and firm performance. This study evaluates success variables (individual performance and organizational performance) as an outcome of knowledge sharing (KMS use).

Proposed Knowledge Sharing Model In the proposed knowledge sharing model, knowledge sharing environment, KMS and sharing motivation are included as the possible drivers of Knowledge Sharing. Knowledge sharing dimension is considered as the use of KMS for knowledge sharing purpose. Finally, individual performance and organizational performance are proposed for the ultimate outcomes of the model as success measurements (DeLone, McLean, 1992, 2003) (Figure 1).

117 European Journal of Economic Studies, 2017, 6(2)

Knowledge

Sharing Environment

Individual Performance

Knowledge Knowledge Management Sharing Systems

Organizational Performance

Knowledge Sharing Motivation

Fig. 1. Knowledge Sharing Model

Therefore, the following hypotheses can be asserted for the research model in Figure 2.

H1. “Social Environment” is positively influential on “Knowledge Sharing”. H2. “Knowledge Management Systems” is supposed to enhance “Knowledge Sharing”. H3. “Knowledge Sharing Motivation” has a positive contribution on “Knowledge Sharing”. H4. “Knowledge Sharing” positively affects “Individual Performance”. H5. “Knowledge Sharing” increases “Organizational Performance”.

3. Data and Methodology In order to collect the data to empirically analyze the proposed research questions and to verify the constructed research model, a survey study is employed. While constructing the questionnaire a seven-point Likert scale (1 for the negative end point as “strongly disagree” and 7 for the positive end point as “strongly agree”) is used and the survey is distributed both on English and Bosnian language. For data collection, convenience sampling is preferred by considering the availability of the respondents. The survey targets individual knowledge workers (respondents) in Bosnian public and private enterprises. Peter Drucker popularizes the term ‘knowledge worker’ in 1968 (Drucker, 1968). He suggests knowledge workers as in the main focus point where they produce ideas, concepts, and information rather than a manual skill or muscle. He suggests knowledge as today’s main cost, investment, and product. According to Drucker, knowledge increasingly becomes the main exchange matter on knowledge based economy. Mainly high status employees in organizational chart such as supervisors, presidents, executive committee members, auditors and CEOs are targeted. A response rate of 69 % (207/300) is achieved from distributed surveys. Performed analyses include (1) descriptive statistics for the strength of the factors and the demographic information and (2) factor analysis, reliability test and regression analyses for testing the relationships and the reliability and the validity of the constructs. SPSS 18 software program is used for all tests.

118 European Journal of Economic Studies, 2017, 6(2)

4. Results Demographic Information The respondents are mainly from operational (35,3 %), administrative (26,6 %) and educational (17,4 %) departments. Their positions within the organizations are as follows: clerical workers (42 %), managers (28,5 %), university lecturers (21,7 %) and so on. Males and females are almost equally represented (52,7 % vs. 47,3 % respectively).

Table 1. Respondents’ Departments

Respondents According to Frequency Percent Their Departments Administration 55 26,6 Auditing 7 3,4 Education 36 17,4 Finance 11 5,3 Human Resources 1 ,5 Law 10 4,8 Marketing and Sales 9 4,3 Operations 73 35,3 Research and Development 5 2,4 Total 207 100,0

KM Implementation The respondents are also asked to evaluate their organizations by considering the implementation levels of knowledge management. According to the results, very few organizations (30/207) have no KM strategy. Most organizations have as at least a KM strategy (82/207). 62 respondents stated that their organizations have an implemented KM strategy. Moreover, 50 respondents rated their organizations as successful in knowledge sharing. 27 reflect that KM practices are a part of their organizational culture. 35 considered their organizational internal environment is approvable for emerging of KM. And finally, 25 respondents measured their organizational external environment is approvable for emerging of KM (Figure 2).

Fig. 2. KM Implementation

119 European Journal of Economic Studies, 2017, 6(2)

Research Model Test Factor Analysis (FA) is employed using Varimax rotation in SPSS in order to evaluate whether the used items point the proposed factors. It is identified that all items show the proposed dimensions as seen in Table 4. The sample size is found to be suitable for FA, since KMO results change between 0,713 and 0,867. The factors with those items are observed to be reliable (Cronbach's Alpha values are between 0,767 and 0,884). In terms of construct validity, item loadings are identified to be quite high. The mean values for all dimensions are also calculated and it is observed that they are all just above average (change between 4,405 and 4,853).

Table 3. FA and Reliability Results

Item-Factor Sampling Reliability N of Correlations Adequacy Statistics Factor Mean Items Interval for Item Cronbach's KMO Measure Loadings Alpha Knowledge 4 4,636 0,732-0,835 0,764 0,767 Sharing Knowledge Management 4 4,607 0,773-0,875 0,713 0,844 Systems Organizational Knowledge 7 4,405 0,543-0,803 0,860 0,869 Sharing Motivation Knowledge Sharing 4 4,555 0,736-0,814 0,780 0,772 Environment Individual 6 4,853 0,698-0,829 0,867 0,850 Performance Organizational 6 4,691 0,741-0,857 0,862 0,884 Performance

After measuring the strengths of dimensions, proposed model is tested and the results are provided in Table 5. According to the table, all relationships are found to be significant. While the influencing factors of Knowledge Sharing are explained well ( =0,649) by the model, performance variables are weakly ( =0,231 and =0,191) explained.

Table 4. Regression Results

Relations Adjusted Standardized Sig. Dependent Independent R Square Coefficients Knowledge Sharing ,212 *** Environment Knowledge Management Knowledge ,649 ,406 *** Systems Sharing Knowledge Sharing ,313 *** Motivation Individual ,231 ,485 *** Performance Knowledge Sharing Organizational ,191 ,441 *** Performance ***: p<0.001

120 European Journal of Economic Studies, 2017, 6(2)

According to the results (Figure 2), Knowledge Sharing is influenced by KMS moderately (0,406); by Organizational Motivation well (0,313); and by Knowledge Sharing Environment weakly (0,212). When the performance variables are considered, knowledge sharing has moderate influence on both individual performance (0,485) and organizational performance (0,441).

=0,231

Knowledge 0,406*** Individual Management Performance Systems 0,485*** =0,649

Knowledge 0,212*** Knowledge Sharing Sharing Environment =0,191

0,441*** Organizational

Performance 0,313*** Organizational Knowledge Sharing Motivation

Fig. 3. Model Test Results (***: p<0.001)

5. Conclusion This study empirically test a knowledge sharing model which implies knowledge sharing practices enhance both organizational and individual performance through (a) qualified KMS, (b) suitable knowledge sharing environment and (c) organizational knowledge sharing motivation. According to the results, it is identified that KM implementation in BiH is poor. Some of the surveyed organizations are found to have no KM strategy. While some others have in initial stages, very few of them has implemented KM practices as a part of their organizational strategy. According to the results, all the assumed relationships are verified meaning that advanced KMS, suitable knowledge sharing environment and high organizational knowledge sharing motivation leverage knowledge sharing and successful knowledge sharing increases individual performance and the organizational performance. The mean values of dimensions advise that Bosnian enterprises have simple KMS, weak knowledge sharing environment and organizational knowledge sharing motivation which are the proposed enablers of knowledge sharing. Knowledge sharing dimension is also detected to be weak. Moreover, both individuals and organization do not seem to have satisfactory performance as a result of knowledge sharing. Low KM implementation level may be the reason for these weak values. Given that all hypotheses are significantly supported by the collected data, the structured models and the relationships are not so strong. Low KM implementation and therefore low mean results may be the reason for these consequences. According to the results, only 50 (out of 207) respondents assume that their organizations are successful in knowledge sharing. Therefore, it may be expected that the respondents are not well aware of knowledge sharing enablers and possible outcomes of successful knowledge sharing in individual and organizational levels. The identified relationships among knowledge sharing, knowledge sharing environment and performance can assist the companies in order to get better performance through knowledge sharing. Therefore, future research may reflect the strategies and programs to leverage firm performance. In this study, the concentration is on Bosnian managerial practices. The study is quite unique in that the knowledge sharing literature is weak in Bosnian environment. Therefore, the study provides valuable theoretical and practical insights since the collected data represents the leading companies in BiH.

121 European Journal of Economic Studies, 2017, 6(2)

Future research may evaluate different types of KS (such as solicited and voluntary) and include additional antecedent variables and characteristics of learning organizations to further explain knowledge sharing behavior in KM.

References Alavi, Leidner, 1999 – Alavi, M., Leidner, D. (1999). Knowledge Management System: Issues, Challenges and Benefits. Communications of the Association for Information System, 1(7), 2-41. Alavi, Leidner, 2001 – Alavi, M., Leidner, D. (2001). Knowledge Management and Knowledge Management Systems: Conceptual Foundations and Research Issues. MIS Quarterly, 25(6), 95-116. Alavi et al., 2006 – Alavi, M., Kayworth, T. R., Leidner, D. E. (2006). An empirical examination of the influence of organizational culture on knowledge management practices. Journal of Management Information System, 22(3), 191-224. Argote et al., 2000 – Argote, L., Ingram, P., Levine, J. M., Moreland, R. L. (2000). Knowledge Transfer in Organizations. Organisational Behavior and Human Decision Processes, 82(1), 1-8. Bartlett, Čičić, Ćulahović, 2012 – Bartlett, W., Čičić, M., Ćulahović, B. (2012). A Survey of Knowledge Management Adoption in Public Administration. Bilgi Ekonomisi ve Yönetimi Dergisi, 7(1), 199-224. Biloslavo, Kljajic-Dervic, 2011 – Biloslavo, R., & Kljajic-Dervic, M. (2011). Survey of Possibilities of Applying the Model of Knowledge Management in Enterprises with Regard to the Bosnia and Herzegovina Market. MIC 2011: Managing Sustainability? Proceedings of the 12th International Conference (s. 1143-1163). Portoroz: University of Primorska, Faculty of Management Koper. Bock, Zmud, Kim, Lee, 2005 – Bock, G.-W., Zmud, R. W., Kim, Y.-G., & Lee, J.-N. (2005). Behavioral intention formation in knowledge sharing: Examining the roles of extrinsic motivators, social-psychological forces, and organizational climate. MIS Quarterly, 29(1), 87-111. DeLone, McLean, 1992 – DeLone, W., McLean, E. (1992). Information Systems Success: The Quest for the Dependent Variable, Journal of Information System Research, 3(1), 60-95. DeLone, McLean, 2003 – DeLone, W., & McLean, E. (2003). The DeLone and McLean Model of Information Systems Success: A Ten-Year Update, Journal of Management Information Systems, 19(4), 9-30. Drucker, 1968 – Drucker, P. F. (1968). The age of discontinuity: Guidelines to our changing society. London: Transaction Publishers. Fink, 2000 – Fink, S. (2000). Crisis Management Planning for the Inevitable. AMACOM, New York. Gu, Gu, 2011 – Gu, Q., Gu, Y. (2011). A Factorial Validation of Knowledge-Sharing Motivation Construct. Journal of Service Science and Management, 4 (1), 59-65. doi: 10.4236/jssm. 2011.41009 Hahn, Subramani, 2000 – Hahn, J., Subramani, M.R. (2000). A Framework of Knowledge Management Systems: Issues and Challenges for Theory and Practice. Proceedings of the Twenty- first International Conference on Information Systems, Brisbane, Australia, December 10-13, pp. 302-312, Paper 28, available at http://aisel.aisnet.org/icis2000/28. Handzic et al., 2007 – Handzic, M., Lagumdzija, A., Celjo, A. (2007). A Survey of Knowledge Management Adoption in Public Administration. Journal of Information & Knowledge Management (JIKM), 6(3), 219-230. Handzic, 2011 – Handzic, M. (2011). Integrated Socio-Technical Knowledge Management Model: An Empirical Evaluation. Journal of Knowledge Management, 15(2), 198-211. Hansen, 2002 – Hansen, M. T. (2002). Knowledge networks: Explaining effective knowledge sharing in multiunit companies. Organization Science, 13: 232–248. Jennex, Olfman, 2004 – Jennex, M., Olfman, L. (2004). Modeling knowledge management success. Proceedings of the Conference on Information Science and Technology Management, (CISTM). Jennex, Olfman, 2005 – Jennex, M., Olfman L. (2005). Assessing Knowledge Management Success. International Journal of Knowledge Management, 1(2), 33-49.

122 European Journal of Economic Studies, 2017, 6(2)

Jennex, Olfman, 2006 – Jennex, M., Olfman, L. (2006). A Model of Knowledge Management Success, International Journal of Knowledge Management, 3(2), 51-68. Jennex, 2008 – Jennex, M. (2008). Exploring System Use as a Measure of Knowledge Management Success. Journal of Organizational and End User Computing (JOEUC), 20(1): 50-63. Liu, Olfman, Ryan, 2005 – Liu, S.C., Olfman, L., & Ryan, T. (2005). Knowledge Management System Success: Empirical Assessment of a Theoretical Model, International Journal of Knowledge Management, 1(2), 68-87. Nonaka, Takeuchi, 1995 – Nonaka, I., Takeuchi, H. (1995). The knowledge creating company: How Japanese Companies create the dynamics of innovation. Oxford University Press Inc., New York. O’Dell, Hubert, 2011 – O’Dell, C., Hubert, C. (2011). The New Edge in Knowledge: How Knowledge Management is Changing the Way We Do Business (1st ed.). Wiley. Oye et al., 2011 – Oye, N.D., Mazleena, S., Noorminshah, A. (2011). Knowledge Sharing in Workplace: Motivators and Demotivators. International Journal of Managing Information Technology, 3(4), 71-84. Ozlen et al., 2012 – Ozlen, K., Mahmutović, Z., Mekić, E., Herić, E. (2012). KM Applications in Bosnian Managerial Practices. ISSD 2012, Sarajevo, BIH. Polanyi, 1966 – Polanyi, M. (1966). The Tacit Dimension, London, UK: Routledge and Kegan Paul. Ribiere, Sitar, 2003 – Ribiere, V.M., Sitar, A.S. (2003). Critical role of leadership in nurturing a knowledge-supporting culture. Knowledge Management Research & Practice, 1, 39-48. Teh, Yong, 2011 – Teh, P.L., Yong, C.C. (2011). Knowledge sharing in is personnel: organizational behavior's perspective. Journal of Computer Information Systems, 51 (4). pp. 11-21. ISSN 0887-4417. Teigland, Wasko, 2003 – Teigland, R., Wasko M. (2003). Integrating Knowledge through Information Trading: Examining the Relationship between Boundary Spanning Communication and Individual Performance. Decision Sciences, 34(2), 261-286. Teng, Song, 2011 – Teng, J.T.C., Song, S. (2011). An exploratory examination of knowledge- sharing behaviors: solicited and voluntary. Journal of Knowledge Management, 15 (1), 104-117. Tsui, 2003 – Tsui, A. B. (2003). Understanding expertise in teaching: Case studies of ESL teachers. New York: Cambridge University Press. Vygotsky, 1978 – Vygotsky, L. (1978). Mind in society: The development of higher psychological processes. Cambridge: Harvard University Press. Wang, Noe, 2010 – Wang, S., Noe, R.A. (2010). Knowledge Sharing: A Review and Directions for Future Research, Human Resource Management Review, 20, 115-131. Wang, Wang, 2012 – Wang, Z., Wang, N. (2012). Knowledge sharing, innovation and firm performance. Expert Systems with Application, 39 (10), 8899-9808. doi: 10.1016/j.eswa.2012.02.017

123 European Journal of Economic Studies, 2017, 6(2)

Copyright © 2017 by Academic Publishing House Researcher s.r.o.

Published in Slovak Republic European Journal of Economic Studies Has been issued since 2012. ISSN: 2304-9669 E-ISSN: 2305-6282 2017, 6(2): 124-143

DOI: 10.13187/es.2017.6.124 www.ejournal2.com

The Relationship between Short-Run Interest Rate and its Economic Determinants: Consumer Price Index, Industrial Production Index, Household Consumption and Exchange Rate. An Empirical Research for the Four Most Developed Countries

John Rigas a , George Theodosiou b , * , George Rigas b , Apostolos Goulas c a J.P. Morgan Chase Bank, UK b Technological Educational Institution of Thessaly, Greece c University of Thessaly, Greece

Abstract This study investigate the relationship between the real money market rate (RMMR) and its economic determinants (Consumer price index (CPI), industrial production index (IPI), household consumption expenditure (HCE) and exchange rate (EXRAT) by using a multivariate VAR model, and examine the existence of a causal relationships between the model variables based on a vector error correction model (VECM) in the four developed countries. The results suggest the existence of a long-run relationship between the real money market rate (RMMR) and its determinants for the four developed countries, in which none of the four determinants have a significant effect on RMMR. The results of causality analysis showed that there exists a bidirectional causality: 1. between change of RMMR and rate of change of CPI for four countries, 2. between rate of change of CPI and rate of change of IPI for one country, 3. between rate of change of IPI and change of EXRAT for one country, and 4. between rate of change of CPI and change of the ratio real HCE/real GDP (gross domestic product) for two countries. Moreover, there is unidirectional causality from the changes of RMMR determinants to the change of RMMR for many countries. Keywords: real money market rate, consumer price index, industrial production index, household consumption expenditure, exchange rate, VAR model, VEC model, cointegration, causality.

1. Introduction The interest rate is distinguished in nominal and real. Nominal interest rate is the unadjusted interest rate for inflation. The real interest rate is given by the Fisher equation: r = ((1+i)/(1+p)) – 1 Where r is the real interest rate, i the nominal one and p is the inflation rate over the year (Interest rate, Wikipedia, n.d.). However for low levels of inflation the linear approximation r ≈ i – p is widely used to calculate the real interest rate. The economic theory states and the monetary authorities of countries (e.g. central banks) confirm that the control of in an economy is achieved with the management of

* Corresponding author E-mail addresses: [email protected] (G. Theodosiou), [email protected] (J. Rigas), [email protected] (G. Rigas), [email protected] (A. Goulas)

124 European Journal of Economic Studies, 2017, 6(2) interest rates (Monetary policy, Wikipedia, n.d.). The key interest rates are those at which central banks lend money to commercial banks. The monetary policy by the central banks through changes in interest rates and the fiscal policy by the governments affect the money supply and the decisions of households and companies for spending, investing and saving (, Wikipedia, n.d.). It is important for the researchers to analyze and interpret the path movements of interest rates and their determinants as interest rates paths influence the whole economy. Recent studies refer extensively to the relationship between interest rate (IR) and separately with each one of its key determinants: consumer price index (CPI), industrial production index (IPI), household consumption expenditure (HCE) and exchange rate (EXRAT). The basic difference between these studies and this one lies in the fact that these investigated the existence of a long-run equilibrium among two or three or four variables, while this study is expanded to include all five variables for a number of countries. Moreover, it is investigated the causal relationship among the model variables for each one of the selected developed countries. In this research, the real money market rate is employed as a proxy of the short-run interest rates because it gives a general depiction of each country’s economy and provides liquidity funding for the global financial system. The data that are used are extracted from the four developed countries across the globe. To begin with, the inflation in these countries is more controlled as these countries have good monetary policies and political stability. They are also highly industrialized which means that the industrial production index will have a crucial role. Moreover, the economic safety that is offered, results in exchange rates with no big fluctuations and it makes the trading and investments with other countries to flourish. The 4 developed countries that were selected for our research are: France, Germany, Japan and USA. The purpose of the study is to examine the relationship between the real money market rate (RMMR) and its economic determinants (CPI, IPI, HCE and EXRAT), and additionally, to analyze the causal relationships between the model variables by using a multivariate autoregressive VAR model for each one of the four developed countries which were chosen for the research.

Consumer price index (CPI)

Industrial production

index (IPI)

Household consumption Interest rates (IR) expenditure (HCE)

Exchange rates (EXCR)

Fig. 1. Model Specification

Three econometric techniques were used to analyze the time series data on real money market rate (RMMR) and its economic determinants, namely: 1. the Augmented Dickey–Fuller (ADF) unit root tests, for the existence of unit root (stationarity tests) 2. the Johansen cointegration test for the existence of a long-run cointegrating relationship among the model variables, and the determination of an error correction model (ECM), and 3. the Granger causality tests, that determine whether one variable is useful in predicting another.

2. Literature review on the relationship among interest rates and its determinants The effect of consumer price index (CPI) on Short-run Real Interest Rate (SRIR) and the existence of casual relationship between them are examined in this study, in order to shed more light on the empirical knowledge derived from the findings of various research works in the last

125 European Journal of Economic Studies, 2017, 6(2) decades. As stated by Kane and Rosenthal (1982) the short term interest rates are important determinants of inflation rate, which is defined as the percentage change of CPI from period to period. Namely, short term interest rates are efficient in predicting inflation, and that has implications in monetary policy. Diba and Oh (1991) found a very high negative relationship of inflation and real interest rate, while the nominal interest rate is weakly correlated with inflation. A later study conducted by Booth and Ciner (2001) contradicts the findings of Diba and Oh (1991) regarding the correlation of inflation and nominal interest rate. Kandil (2005), in his study of fifteen developed countries with strong industry, concluded that both interest rate and money supply are underlying factors for the formation of price levels and they are strong correlated with each other. On the other hand, Cologni and Manera (2008) took into account the significant effect of the large rises in oil prices the recent years to the and assessed the performance of the G-7 countries’ economies. Al-Khazali (1999) examined the relationship between interest rates and inflation in nine countries of the Pacific-Basin. His results did not give any evidence of a relationship between the variables. Nagayasu (2002) found that the impact of interest rates in the inflation evolution in Japan for the period 1980 to 2000 is very strong, especially when using short term interest rates. Allen and Mapfumba (2006) concluded that the gap of the neutral and real interest rate is the major determinant of the inflation growth. Some years earlier, Gjerde and Saettem (1999) employed VAR model methodology for Norway with model variables interest rate, inflation, industrial production index, exchange rate and oil prices with purpose to examine how each variable affects and in what degree the rest variables. The most crucial conclusion was that interest rate affects significantly the inflation. It was also showed that there is no direct relation between interest rate and industrial production. The joint effect of industrial production index IPI on SRIR and the existence of causality between IPI and SRIR are also examined in this study. The empirical research findings about the link between IPI and SRIR are mainly referred to the effect of interest rates on IPI. Wei (2008) examined the interest rates’ impact on the industrial production and the stock market in China. He concluded that the interest rates policies affect IPI and have a short-run effect in the industrial production which is not persistent in the long-run. Tunali (2010) found out that IPI is negatively related with interest rate. Bianchi et al. (2010) found that the long - short term interest rate spread and the long term real interest rate are marginally significant for industrial output only for Italy. Papapetrou (2001) found that interest rate is negatively related with industrial production output in Greece. Later, Gogas & Pragidis (2010) analysed the forecasting power of the yield curve and argued that the yield curve in combination with stock index has significant predicting power on the Greek Industrial Production Index (IPI). Household consumption is closely linked to savings and to disposable income. Carlino (1982) argues that a survey on the findings of various studies revealed inconsistencies as to the significance and the sign of the estimated intertemporal relationship between consumption and interest rate. Baum (1988) derived a relationship among real rate of interest, consumption, and the personal wealth and he showed that the real interest rate has a small and not statistically significant effect on consumption-savings decisions. One year later, Campbell and Mankiw (1989) estimated the relation between consumption, income and interest-rates. They considered that the data are generated by two types of consumers, one consuming their permanent income (expected long term income) and the other their current income. Sullivan and Lombra (1992) examined analytically and empirically the role of interest rates and of non-interest terms on loan using yearly US data. The main results of this work were that non-interest terms on loan were a determinant of household spending on housing and durables goods and become less important as deregulation and innovation have proceeded. In contrast to the previous researchers, Hahm (1998) argued that there is a significant positive relationship between consumption growth and changes in expected real interest rates. He argued that the failure of previous studies to find a significant link between consumption and real interest rates can be attributed either to the use of inappropriate instruments or using an inadequate measure of consumption (e.g. including the problematic housing services). Nakagawa and Oshima (2000) reported that for the treatment of the low consumption problem that the Japanese economy faced, Krugman (1998) suggested that a decrease in real interest rates caused by inflation expectations could increase household consumption. The question for them was whether the decline in real

126 European Journal of Economic Studies, 2017, 6(2) interest rates will lead to stimulating household consumption in Japan. To answer this question and to verify the assertion of Professor Krugman (1998), they analyzed the relationship between real interest rates and household consumption using data from Japan, U.S.A., Britain and France. They found that the relationship real interest rate – household consumption is supported for the USA and UK, where consumers are more willing to borrow, but it does not work in Japan, since Japanese like to save and are unwilling to use consumer credit even if real interest rates decline. Zhang and Wan (2002) examined the effect of interest rate changes on household consumption in China and ascertained that nominal interest rates are less relevant to household consumption decisions than inflation rates do. Cromb and Corugedo (2004) investigated the sensitivity of consumption level to interest rates in a simple model of consumption under conditions of certainty. For most examined parameter values, they found that higher interest rates are linked with lower level of consumption. Finally, Neumeyer and Perri (2005) studied the relationship between real interest rates and business cycles for five small open developing economies and for five small open developed economies. They found that real interest rates are countercyclical and lead the cycle and the consumption. Consumption is procyclical and present higher volatility than output. Moreover, they documented that developing countries business cycles are more volatile than in developed ones. The relationship between interest rate and its fourth determinant, the exchange rate, was investigated by many researchers, which led to interesting conclusions. Edison and Path (1993) assessed the relationship between real interest rates differentials and real exchange rates, using the exchange rates of the U.S. dollar against the other G-10 currencies. Their results indicated the no- existence of a long-run relationship between real exchange rates and real interest rates differentials. Baxter (1993) re-examined the relation real exchange rates - real interest differentials using the same data. She found that there exists a relationship between the two variables with the strongest link at trend and business-cycle frequencies and that there is no relationship between them at high frequencies. This explains why prior studies, focused on high-frequency components of the data, found no statistical relationship. Ogaki and Santaellar (2000) studied the relationship between exchange rate and interest rate for Mexico, and found that one-month and three-month interest rate differentials have opposite results on the exchange rate. Namely, increases in the one-month interest rate differential, all other things constant, tend to appreciate the exchange rate, while increases in the three-months interest rate differential tends to depreciate the exchange rate. Nakagawa (2002) tried to presents evidence in the relationship between real exchange rates and real interest rates differentials for the U.S. dollar against British pound, German mark, Canadian dollar and Japanese yen. He concluded that the confusion between theory, that supports a relationship between these variables, and the empirical findings, that are not clear, is due to the fact that the nonlinearity is not recognized in adjustment of real exchange rate. When threshold nonlinearity was introduced into a traditional model he had results that supported the link real exchange rates – real interest rate differentials. Zettelmeyer (2004) studied the impact of monetary policy on the exchange rates in Australia, Canada, and New Zealand using 3-month market interest rates. The main finding is that 2-3 percent appreciation of the exchange rate would require an interest rate change of about 100 basis point. This is consistent with the prevailing view about the impact of interest rates on exchange rates. Kanas (2005) findings support the existence of a relation between the US/UK real exchange rate and the real interest differential, which is justified by the theoretical knowledge about the real exchange rate determination. Gochoco-Bautista and Bautista (2005) found that contracting domestic credit expansion and increasing the interest rate differential, they both contribute to reduce the exchange market pressure. Afterwards, Bautista (2006) examined the relationship between the real exchange rate and the real interest differential in six East Asian economies and showed that the relation changes with the nominal regime. During fixed exchange rate regimes the relation were characterized by positive time-varying correlations, while during free fall regimes correlations were negative. Chen (2006) using weekly data from six developing countries tried to answer the question of whether a higher interest rate steadies exchange rates. The empirical results indicated that higher nominal interest rates leads to increased probability of switching to a crisis regime, and that a high interest rate policy cannot defend the exchange rate. The empirical results of Choi and Park (2008) showed

127 European Journal of Economic Studies, 2017, 6(2) that the tight monetary policy and the consequent rise in interest rates were not effective in stabilizing the exchange rate during the Asian currency crisis and after the end of it. In conclusion, interest rates have been widely investigated for their relationship with the key macroeconomic variables of inflation, industrial production, household consumption and exchange rate. As discussed above, it has high importance in knowing the path movements of interest rates as they are used for the form of government policies. What is important to mention here is that there has been discussion within the literature about the unidirectional relation of interest rate to these four variables, but as mentioned above, most of the studies do not discuss the opposite direction of the relationship. Moreover, none of the above researches had as an objective the evaluation of the joint effect of the four macroeconomic variables in the interest rate and therefore a different approach like this is required. Something else that is important to mention here is that, most of the researches that have been mentioned in the literature review have used more or less the same methodological tools which are employed in the current study. Finally, a wide range of countries have been used in the literature review, but they gave different results for the same relationships. This is probably due to the special characteristics of each country. For that reason, the current study employs a sample of the eighteen most developed countries which is anticipated to give a robust answer in the problem.

3. Research hypotheses The objective of the study is to investigate the effect of various key economic indicators (variables) on short-run real interest rate in the case of developed countries. Namely, it is investigated the joint effect of some key economic indicators on short-run real interest rate for a number of the most developed countries. In recent decades the relationship between interest rate and various key economic indicators has become a subject of extensive research. The research hypothesis was created after the literature survey took place and was found that each of the four economic indicators consumer price index (CPI), industrial production index (IPI), household consumption expenditure (HCE) and exchange rate (EXRAT) affects individually either with its current values or its lags the interest rates or they are affected by changes in the interest rates. These indicators indicate how well an economy is and how well it will be in the future. After the ascertainment that there exist a link between the interest- rate and the four selected indicators in some countries, we formulated the research hypothesis. The research hypotheses of the study is to investigate the existence of a long run relationship between the real money market rate (RMMR) and its economic determinants (CPI, IPI, HCE and EXRAT), and to examine the causal relationships between the model variables for each one of the 4 developed countries which are chosen for this study. The data used in the study are quarterly and are obtained from International Financial Statistics (International Monetary Fund, IMF, 2010). The range periods of the time series data for the 4 developed countries are appeared in Table 1 (Appendix). The methodology of VAR models is used for this empirical research and it is based on the following reasons. The economic analysis suggests that there are long-run relationships between various economic variables included in the explanation of relevant economic phenomena (Brooks, 2008: 336). The estimation of these relations with the classical OLS method assumes that the variables are stationary, otherwise it raises the problem of "spurious regression” (Brooks, 2008: 319). The problem is treated by the stationarity and cointegration tests of variables (Cointegration, Wikipedia, n.d.). The cointegration analysis is used to estimate the long-run parameters and for its application, when there are more than two variables, VAR models are required. The usefulness of the VAR model can also be seen in the estimation of short-run parameters or imbalanced parameters. The estimation of these parameters makes use of long-run parameters estimated with the integration method. These procedures are performed with the vectors error correction models (VECM) and are based on VAR models technique. Moreover, VAR models are used to capture the evolution of multiple time series, and the interdependencies between them (Brooks, 2008: 291). From an economic point of view the joint dynamics of the VAR model variables are a depiction of the underlying economic relationships among the model variables (, Wikipedia, n.d.). The study is limited to the developed countries on the grounds that they have specific common characteristics (Developed country, Wikipedia, n.d.) and meet certain standards, as they are

128 European Journal of Economic Studies, 2017, 6(2) appeared in Table 2 (Appendix), which guarantee the reliability of data used and enhance the results of the statistical tests. Developed countries, however, differ among themselves on the temporal evolution of economic indicators (First World, Wikipedia, n.d.), which is mainly related to: 1. the production’s structure of each country; 2. the contribution of each sector in the GDP formulation; 3. the propensity to consumption that the people of each country show; 4. the economic robustness of each country; 5. the degree of new technologies implementation; 6. the degree of industrial development; 7. the innovative activity and entrepreneurship; 8. the competitiveness of products and services; 9. the competitive advantages of each country; 10. the maturity of the society; 11. the scientific and professional training of citizens; 12. the applied social policy; 13. the applied fiscal and monetary policy; 14. the joining of a country in a union of countries like the European Union and the euro- area; 15. the climatic conditions, the historical evolution of the country, the mentality of population, the geographical position and a lot of other factors of smaller importance. This differentiation is the one that can justify different values in the estimated coefficients of the variables in long-term equilibrium relationship. The estimated coefficients essentially are the long-run estimated elasticities of the short-run real interest rate as for the other four variables: consumer price index (CPI), industrial production index (IPI), household consumption expenditure (HCE) and exchange rate (EXRAT). In this research, a total number 4 from the developed countries (see Table in introduction), as the most developed and for which data were available, was selected across the globe, most of them from the European Union which has the largest number of developed countries. A considerable number from these countries comes from euro-area. Except from the fact that the economies of developed countries meet certain standards, they also implement effective economic policies and provide reliable data and information to various organizations and databases.

4. Methodology 4.1. Model structure A fifthvariate VAR model was used in order to test the existence of long-run relationships, to determine the interest rate function and to analyse the causal relationship. The expected long-run equilibrium relationship (cointegration equation) is specified as follows: RMMRt = b0 + b1LNCPIt + b2LNIPIt + b3RRHCGDt +b4EXRATt+et (1) where: RMMRt = Real money market rate = Money market rate (MMRt) - Inflation (INFt) =MMRt- INFt INFt = ( LNCPIt-LNCPIt-1 )*100*4=INFt(Q)*4 INFt = Annual inflation corresponding to quarter t INFt(Q)= ( LNCPIt-LNCPIt-1 )*100=The quarterly inflation LNCPIt = The natural logarithm of CPIt LNIPIt = The natural logarithm of IPIt RRHCGDt = Real household consumption expenditures (RHCEt)/ Real gross domestic product (RGDPt) = RHCEt / RGDPt RHCEt = Household consumption expenditures / GDPt Deflator (2005=100) RGDPt= Real GDPt = GDPt / GDPt Deflator (2005=100) EXRATt = The current exchange rate. The time series data of CPI and IPI are expressed in natural logarithms in order to obtain stationarity in their variance and also to capture multiplicative time series effects (Granger, Newbold, 1986), cited by Dritsakis & Adamopoulos (2004). The ratio RRHCGD = RHCE/RGDP is defined to show the real size of the household consumption.

129 European Journal of Economic Studies, 2017, 6(2)

In modelling the real money market rate equation, the methodology of unit root (Augmented Dickey-Fuller (ADF) unit root test) was used to determine whether each time series entering the model is stationary and moreover its order of integration. The long-run cointegrating relationship among the variables was examined by using the Johansen co-integrating test. The vector error correction model (VECM) was employed to determine the short-run dynamics of the variables in the model. Finally, Granger causality tests based on VECM were applied to explore the directions of causality between the model variables.

4.2. Augmented Dickey–Fuller (ADF) Test In the first part of our analysis, we have to check for the existence of unit root test or otherwise if each one of our variables is stationary. According to Brooks (2008, p. 319), two main problems may arise from the use of non-stationary data. The first is that in case of an unpredicted change in a specific moment, the effect of this change will exist to the infinite and in the same degree of significance. The other is that the regressions will not be true as it results in high R2 even if the variables show no sign of correlation to each other. After checking that the time span of the variables is big enough, we made sure that ADF tests can be performed for the existence of unit root of model variables. The Augmented Dickey–Fuller (ADF) (Dickey, Fuller, 1979) regression tests refer to the t-statistic of δ2 coefficient on the following three regression equations: k ΔΧt = δ2Χt-1 + ∑βiΔΧt-i + et (2) i=1 k ΔΧt = δ0 + δ2Χt-1 + ∑βiΔΧt-i + et (3) i=1 k ΔΧt = δ0 + δ1t + δ2Χt-1 + ∑βiΔΧt-i + et (4) i=1 Where: i =1,2,3,…,k the number of time lags δ0, δ1, δ2 and βi i=1,2,3,…k are the parameters and t is the time trend. The null and the alternative hypothesis that are testing in the three models (2, 3 and 4) for the existence of unit root in variable Xt are as follows: H0 : δ2=0 (The series Xt contains a unit-root, hence it is non-stationary). Ha : δ2 < 0 (the H0 is not valid). The hypotheses were tested by t-statistic of δ2 using the critical values of MacKinnon (1991). The econometric package EViews 5.1 (2005), that was used for the ADF tests, gave the critical values of MacKinnon at 1 %, 5 % and 10 % level. Dickey-Fuller (1979) showed that the asymptotic distribution of t-statistic is independent of the number of lags of the dependent variable’s first differences. What affects the values of t-distribution is the presence or absence of deterministic terms such as the intercept and time trend. The minimum values of Akaike criterion (AIC) and of Schwartz criterion (SCH) determined the optimal specification of ADF equations and the appropriate number of lags. Regarding the test of autocorrelation in disturbance terms (residuals) the Breusch-Godfrey test or otherwise the Lagrange Multiplier (LM) statistical criterion was used. The number of time-lags should be such that there are no auto-correlated residuals.

4.3. Cointegration Test Before the explanation of the Johansen cointegration methodology that it is based on the VAR models, a presentation of the reasons that drove to the specific choice is made. The VAR models make possible to examine whether fork variables that are not stationary in levels there exists a long-run relationship, and how these k variables relate to each other. Moreover, a Vector Error Correction Model is defined that links the short-run dynamics with the long-run relationship. Apart from that, Brooks (2008, p. 291) states that a VAR model offers the flexibility to each variable to be depended not only in its own lags but to the lags of the other variables, and that the VAR forecasts are much better than those of the traditional models.

130 European Journal of Economic Studies, 2017, 6(2)

The Augmented Dickey-Fuller (ADF) unit root tests determine whether each series entering the VAR model is stationary and also its order of integration. Given the results of unit root tests, the Johansen cointegrating test examines whether there is a long-run cointegrating relationship among the model variables. These variables can be cointegrated if there is one or more linear combinations among the variables that are stationary (I(0) integrated). If the variables are cointegrated, then there is a stable long-run linear relationship between them. In the case of k variables can be up to k-1 linearly independent cointegrating vectors (cointegrating equations). The number of linearly independent cointegrated vectors is called "order of cointegration" and may range from 1 to k-1. The Johansen cointegration tests are based on the methodology of VAR models which enable the researcher to determine the maximum number of cointegrated vectors (cointegrating equations The VAR models constitute a system of equations where all variables are endogenous and each one is determined as a function of the past values (lags) of all variables of the model. The selection criteria of likelihood ratio (LR), of Akaike, of Schwartz and of HQ are used to determine the VAR lag order and the number of lags required in the cointegration test. For testing the number of cointegrated vectors in the VAR model are used the trace (Tr) test and the maximum eigenvalue (max-eigen) test, proposed by Johansen & Juselius (1990). The null hypothesis in the trace test is that there are at most k cointegrated vectors or there are at most k linear combinations among the model variables that are stationary. Namely, the number of cointegrating equations r is less than or equal to k, where k=0,1, 2...,m-1, and m the number of model variables. The hypotheses that are sequentially tested by trace test are: H0 : r ≤ k against the alternative Ha : r ≥ k+1, k=0,1, 2...,m-1. In the max-eigen test the null hypothesis that there are at most k cointegrated vectors (H0 : r ≤ k) is tested against the alternative hypothesis of k+1 cointegrated vectors ( Ha : r = k+1), k=0,1, 2...,m-1. Thus, the hypotheses that are sequentially tested by max-eigen test are: H0 : r ≤ k against the alternative Ha : r = k+1, k=0,1, 2...,m-1. Based on the two criteria, the number of the cointegrated vectors is determined at 5 % and 1 % levels for each one of the 18 developed countries. Johansen and Juselius (1990) suggest the use of trace test when there are different results from the two tests.

4.4. The Error Correction Model (ECM) The error correction model that is defined from the long-run cointegration relationship for the equation of RMMR can be expressed as follows: ΔRMMRt =Lagged( ΔLNCPIt, ΔLNIPIt, ΔRRHCGDt, ΔEXRATt) + λut-1 + Vt (5) where: Δ refers to first differences of the variables, ut-1 are the estimated residuals from the long-run relationship (cointegrating equation) and represents the deviation from it in time t, -1 <λ <0 is the short-run adjustment coefficient, Vt is the white noise error term. The error correction model analyses the short-run dynamics and links the short-run and the long-term behaviour of the model variables. The selection criteria of Akaike and of Schwartz are used to determine the number of lags required in the VECM.

4.5. Granger causality tests The Granger causality tests, like the Johansen cointegration tests, are based on the methodology of VAR models. The Granger causality test determines whether one variable is useful in predicting another. Namely, for each pair of model variables X and Y it is said X Granger cause Y if and only if the prediction of Y is better by using the lag values of X together with the lag values of all other model variables (Y including). The VEC Granger Causality/Block Exogeneity Wald tests and the χ2 (Wald) statistics are used to examine the Granger causality among the model variables. Granger causality is distinguished in unidirectional and bidirectional. Unidirectional causality exists from X to Y if X Granger causes Y but Y does not Granger causes X, and bidirectional if X Granger causes Y and Y Granger causes X. The reliability of this test depends on the order of the VAR model and on the stationarity of the variables. The reliability of this test is reduced if the variables are not stationary. For each one of the18 developed countries, the estimated vector error

131 European Journal of Economic Studies, 2017, 6(2) correction model (VECM) was used to test the existence of causal relationships among the model variables.

5. Data The data that are used in this study were obtained from International Financial Statistics, (International Monetary Fund, IMF, 2010). The time series data are quarterly covering for each one of the 4 developed countries the range period appeared in Table 1 (Appendix). Household consumption expenditures and Gross domestic product GDP time series data are converted from nominal to real values in national currency. The national GDP Deflator (2005=100) for each country was used to adjust nominal values to real values ( base year 2005). The base year for calculation of the Indices CPI and IPI is 2005 (Index Numbers (2005=100): Period Averages, IMF (2010)). The endogenous variables of VAR model include the real money market rate (RMMR), the natural logarithm of consumer price index (LNCPI), the natural logarithm of industrial production index (LNIPI), the ratio (RRHCGD) of real household consumption expenditures (RHCE) to real gross domestic product (RGDP) and the exchange rate (EXRAT) that refers to the current exchange rate. Afterwards, some explanations for the economic indicators, by which are defined the model variables, are given below: Money market rate (MMR) is short-term interest rate such as the three-month EURIBOR rate (Euro area countries), the rate on three-month commercial paper (USA) and the interbank offer rate for overnight deposits (UK). Analytic presentation of MMR for each one of the 4 developed countries is given in Table 3 (Appendix). Real money market rate (RMMR) is the nominal interest rate adjusted for inflation and measures the purchasing power of interest income. The linear approximation r ≈ i – p is used to this study to calculate the real interest rate, where r is the real, i is the nominal interest rate and p is the inflation rate over the year. The consumer price index (CPI) is today in UK the official measure of inflation (Consumer Price Index (UK), Wikipedia, n.d.). An increasing trend in CPI can raise interest rates and bond yields and cause a fall to bond prices. Likewise, a decreasing trend in CPI can cause a fall to interest rates and bond yields . The inflation rate (INF) is the percentage change in CPI from period to period, and can be defined as: INFt = ((CPIt - CPIt-1)/ CPIt-1)*100 ≈ (LNCPIt - LNCPIt-1)*100) (Sweidan, 2004; Katos, 2004: 992). The Industrial Production Index (IPI) moves at the same time as economic activity (business cycle) and can be considered an accurate measure of industrial production and of manufacturing employment. High levels of industrial production can lead to high levels of consumption, to rapid rise of inflation and increase of interest rates. As such, IPI becomes a leading indicator of interest rates. Household consumption expenditure (HCE) covers all domestic expenditures (from residents and non-residents) for individual needs. This includes expenditure on goods and services, rent for owner-occupied residences and the consumption of garden produce. Gross domestic product (GDP) is a very important measuring the economic activity of countries and is defined as the monetary value of all goods and services produced within a country in a specific time period. The exchange rate of each one of the 4 developed countries is determined by national currency units per US Dollar, except from USA that is determined by US Dollar per Special Drawing Rights (SDR). The Euro Area member countries exchange rates, until the participation of national currencies within the Eurosystem, are expressed as national currency units per US Dollar. After the participation of national currencies within the Eurosystem the countries exchange rates are presented as Euros per US Dollar. The exchange rates of the Euro Area member countries were converted from national currency exchange rates into Euro exchange rates at official conversion rates (Euros per national currency).

6. Estimation Results The results of the Augmented Dickey-Fuller (ADF) tests are presented in Table 4 (Appendix). The Akaike information criterion (AIC) and the Schwartz criterion (SC) determined the best specification of ADF equations and the corresponding number of lags. Regarding the autocorrelation test in error terms, the Lagrange Multiplier (LM (1)) test was applied. The two statistical test trace test and Max-eigen test showed that there exists one cointegrating equation for each one of the 18 country. A unique long run relationship between real

132 European Journal of Economic Studies, 2017, 6(2) money market rate (RMMR) and its determinants (LNCPI, LNIPI, RRHCGD and EXRAT) is accepted for each country, which includes all five variables. The Johansen maximum likelihood cointegration test results are appeared in Table 5 (Appendix) and the normalized cointegrating coefficients of cointegrating equation for each country in Table 6 (Appendix). Deviations from long-run equilibrium relationship could happen in the short-run due to profound changes to one or more variables of the model. The short-run dynamics were analysed by applying an error correction model (ECM). The estimated long run relationship for each country was used to include an error correction mechanism in a VAR model. The derived error-correction model has then the following form: ΔRMMRt =Lagged( ΔLNCPIt, ΔLNIPIt, ΔRRHCGDt, ΔEXCRt) + λut-1 + Vt (6) The results of Granger causality tests are appeared in Table 6. Since the reliability of Granger causality tests depends on the order (k) of the VAR model and on the stationarity of the variables, the Granger tests were applied using the VEC model, which uses the first differences (stationary variables). That enables us to see and the economic meaning of the Granger causality relations. The selection criteria of Akaike and Schwartz were used to select the order of the VEC model for each country (Table 7 (Appendix)). The detailed results of statistical tests for each one of the four developed countries are presented below: 1. France. The ADF tests showed that LNCPI and LNIPI are stationary at levels, while RMMR, RRHCGD and EXRAT are stationary at first differences. The results of cointegration tests suggest that there exist a long-run relationship that presents the following form: RMMR=378.534LNCPI+234.101LNIPI+4745.830RRHCGD-223.851EXRAT–5145.047(17) s.e. (132.291) (535.384) (1781.650) (116.214) t [2.861] [0.437] [2.664] [1.926] In the long-run, LNCPI and RRHCGD have a significant positive effect on RMMR. LNIPI affects positive RMMR and EXRAT negative but both not significantly. The error-correction models present the following form: D(RMMRt)=Lagged(D(LNCPIt), D(LNIPIt), D(RRHCGDt), D(EXRATt)) - 0.005ut-1 + Vt (18) s.e. (0.001) t [-3.260] Adj. R2 = 0.440, ECM Lags (in first differences) = 4 2 The Adj. R = 0.440 is quite large and the coefficient of ut-1= -0,005 has a negative sign and is statistically significant (t=-3.260). The results of Granger causality tests denote that there exists a unidirectional causality from D(LNCPI) to D(LNIPI), from D(LNIPI) to D(RMMR) and from D(EXRAT) to D(LNCPI), and a bidirectional causality between D(RMMR) and D(LNCPI) at the 5% level. 2. Germany. The ADF tests showed that RMMR, RRHCGD and EXRAT are stationary at levels, while LNCPI and LNIPI are stationary at first differences. The results of cointegration tests suggest that there exist a long-run relationship that presents the following form: RMMR=71.864LNCPI – 57.320LNIPI - 115.723RRHCGD + 25.178EXRAT–14.919 (19) s.e. (16.110) (16.419) (134.863) (8.343) t [4.461] [-3.491] [-0.858] [3.018] In the long-run, LNCPI and EXRAT have a significant positive while LNIPI has a significant negative effect on RMMR. RRHCGD affects negative RMMR but not significantly. The error-correction model presents the following form: D(RMMRt)=Lagged(D(LNCPIt), D(LNIPIt), D(RRHCGDt), D(EXRATt))-0.083ut-1+Vt (20) s.e. (0.026) t [-3.153] Adj. R2 = 0.471, ECM Lags (in first differences) = 5 2 The Adj. R = 0.471 is quite large and the coefficient of ut-1= -0,083 has a negative sign and is statistically significant (t=-3.153). The results of Granger causality tests denote that there exists a unidirectional causality from D(LNCPI) to D(LNIPI), from D(LNIPI) to D(RMMR) and to D(RRHCGD) and from D(EXRAT) to D(RMMR) and to D(LNCPI), and a bidirectional causality between D(RMMR) and D(LNCPI) at the 5 % level.

133 European Journal of Economic Studies, 2017, 6(2)

3. Japan. The ADF tests showed that RMMR and LNIPI are stationary at levels, while LNCPI, RRHCGD and EXRAT are stationary at first differences. The results of cointegration tests suggest that there exist a long-run relationship that presents the following form: RMMR= 9.739LNCPI – 9.920LNIPI – 83.726RRHCGD + 0.024EXRAT + 43.185 (23) s.e. (1.886) (1.815) (14.318) (0.006) t [5.163] [ -5.466] [-5.847] [3.860] In the long-run, LNCPI and EXRAT have a significant positive while LNIPI and RRHCGD have a significant negative effect on RMMR. The error-correction model presents the following form: D(RMMRt)=Lagged(D(LNCPIt), D(LNIPIt), D(RRHCGDt), D(EXRATt))-0.420ut-1+Vt (24) s.e. (0.187) t [-2.247] Adj. R2 = 0.458, ECM Lags (in first differences) = 2 2 The Adj. R = 0.458 is quite large and the coefficient of ut-1= -0,420 has a negative sign and is statistically significant (t=-2.247). The results of Granger causality tests denote that there exists a unidirectional causality from D(RMMR) to D(LNCPI) and from D(LNIPI) to D(LNCPI) and to D(RRHCGD), and a bidirectional causality between D(RMMR) and D(LNIPI) at the 5% level. 4. USA. The ADF tests showed that RMMR, LNCPI and RRHCGD are stationary at levels, while LNIPI and EXRAT are stationary at first differences. The results of cointegration tests suggest that there exist a long-run relationship that presents the following form: RMMR=5.972LNCPI + 3.675LNIPI – 224.306RRHCGD – 6.379EXRAT + 122.872 (33) s.e. (1.649) (3.900) (42.830) (3.063) t [3.622] [0.942] [-5.237] [-2.083] In the long-run, LNCPI has a significant positive effect on RMMR while RRHCGD and EXRAT have a significant negative effect on RMMR. LNIPI affects positive RMMR but not significantly. The error-correction model present the following form: D(RMMRt)=Lagged(D(LNCPIt), D(LNIPIt), D(RRHCGDt), D(EXRATt))–0.137ut-1+Vt (34) s.e. (0.052) t [-2.659] Adj. R2 = 0.377, ECM Lags (in first differences) = 5 2 The Adj. R = 0.377 is quite large and the coefficient of ut-1= -0,137 has a negative sign and is statistically significant (t=-2.659). The results of Granger causality tests denote that there exists a unidirectional causality from D(RMMR) to D(EXRAT) and from D(LNIPI) to D(RRHCGD), and a bidirectional causality between D(RMMR) and D(LNCPI) at the 5 % level.

7. Conclusions In this study we investigate the relationship between the real money market rate (RMMR) and its economic determinants (LNCPI, LNIPI, RRHCGD and EXRAT), and additionally, examine the causal relationships between the model variables for each one of the four selected developed countries based on VECM. The results support the existence of a long-run relationship between the real money market rate (RMMR) and its determinants for the four developed countries. Specifically, the results indicated that: 1. LNCPI has a significant positive effect on RMMR for France, Germany, Japan, and USA; 2. LNIPI has a significant positive effect on RMMR for UK, while LNIPI has a significant negative effect on RMMR for, Germany and Japan; 3. RRHCGD has a significant positive effect on RMMR for, France, Japan and a significant negative effect on RMMR for the USA; 4. EXRAT has a significant positive effect on RMMR for, Germany, and a significant negative effect on RMMR for, Japan, and the USA. The causality analysis results showed that there exists: 1. A bidirectional causal relationship between change of RMMR (D(RMMRt)) and rate of change of CPIt (D(LNCPIt)=quarterly inflation /100) for France, Germany, and the USA;

134 European Journal of Economic Studies, 2017, 6(2)

2. A unidirectional causality from rate of change of IPI to change of RMMR for France and Germany; 3. A unidirectional causality from change of exchange rate to change of RMMR for Germany; 4. A unidirectional causality from D(RMMRt): 1. to rate of change of CPI for Japan and 4. to change of exchange rate for the USA.

References Al-Khazali, 1999 – Al-Khazali, O. (1999). Nominal interest rates and inflation in the Pacific- Basin countries. Management Decision, 37(6), pp. 491-498. Allen, Mapfumba, 2006 – Allen, D.E., Mapfumba, T. (2006). Real Interest Rates and Inflation in Norway. School of Accounting, Finance and Economics & FIMARC Working Paper Series, Edith Cowan University, Working Paper 06.02. Jun 2006 00:00 No. 06.02 – Real Interest Rates and Inflation in Norway archived 01 Jun 2006 00:00. Baum, 1988 – Baum, D. (1988). Consumption, Wealth, and the Real Rate of Interest. A Reexamination Journal of Macroeconomics, Winter 1988, 10(1), pp. 83-102. [Electronic resource]. URL: https://docslide.com.br/documents/consumption-wealth-and-the-real-rate-of- interest-a-reexamination.html Bautista, 2006 – Bautista, C.C. (2006). The exchange rate–interest differential relationship in six East Asian countries. Economics Letters, 92. https://doi.org/10.1016/j.econlet.2006.01.016 pp. 137–142. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/S016 517 65 06000395 Baxter, 1993 – Baxter, M. (1994). Real exchange rates and real interest differentials. Journal of Monetary Economics, 33(1), pp. 5-37. Bianchi et al., 2010 – Bianchi, C., Carta, A., Fantazzini, D., De Giuli. M.E., Maggi, M. (2010). ACopula-VAR-X approach for Industrial Production Modelling. , 42(25), pp. 3267-3277. Booth, Ciner, 2001 – Booth, G.G., Ciner, C. (2001). The relationship between nominal interest rates and inflation: international evidence. Journal of Multinational Financial Management, 11, pp. 269–280. [Electronic resource]. URL: https://www.researchgate.net/ publication/222297529_The_relationship_between_nominal_interest_rates_and_inflation_Inte rnational_evidence Brooks, 2008 – Brooks, Ch. (2008). Introductory Econometrics for Finance, Cambridge University Press, p. 336, 2008. [Electronic resource]. URL: https://books.google.gr/books?id= CrPRZexTZLkC&printsec=frontcover&hl=el&source=gbs_ge_summary_r&cad=0#v=onepage&q& f=false Brooks, 2008 – Brooks, Ch. (2008). Introductory Econometrics for Finance, Cambridge University Press, p 319, 2008. [Electronic resource]. URL: https://books.google.gr/books?id= CrPRZexTZLkC&printsec=frontcover&hl=el&source=gbs_ge_summary_r&cad=0#v=onepage&q& f=false Brooks, 2008 – Brooks, Ch. (2008). Introductory Econometrics for Finance, Cambridge University Press, p 291, 2008. [Electronic resource]. URL: https://books.google.gr/books?id=CrPRZexTZLkC&printsec=frontcover&hl=el&source=gbs_ge_s ummary_r&cad=0#v=onepage&q&f=false Campbell, Mankiw, 1989 – Campbell, J. Y., Mankiw, N. G. (1989). Consumption, Income and Interest Rates: Reinterpreting the Time Series Evidence, NBER Macroeconomics Annual 4, pp.185- 216. www.nber.org/chapters/c10965.pdf Carlino, 1982 – Carlino, G. (1982). Interest rate effects and intertemporal consumption, Journal of Monetary Economics, 9, pp. 223-234. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/0304393282900435 Chen, 2006 – Chen, S.S. (2006). Revisiting the interest rate–exchange rate nexus: a Markov- switching approach. Journal of Development Economics, 79, pp. 208– 224. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/S0304-3878(05)00003-9 Choi, Park, 2008 – Choi, I., Park, D. (2008). Causal relation between interest and exchange rates in the Asian currency crisis. Japan and the World Economy,

135 European Journal of Economic Studies, 2017, 6(2) https://doi.org/10.1016/j.japwor.2007.01.003 20, pp. 435–452. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/S0922142507000114 Cologni, Manera, 2008 – Cologni, A., Manera, M. (2008). Oil prices, inflation and interest rates in a structural cointegrated VAR model for the G-7 countries. Energy Economics, 30, pp. 856–888. [Electronic resource]. URL: http://www.sciencedirect. com/science/article/pii/S0140988306001393 Cromb, Corugedo, 2004 – Cromb, R., Corugedo, E. F. (2004). Long-term interest rates, wealth and consumption, Publications Group, Bank of England, Working Paper no. 243. Diba, Oh, 1991 – Diba, B.T., Oh, S. (1991). Bounds for the correlations of expected inflation with real and nominal interest rates. Economics Letters, 36(4), pp. 385-389. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/016517659190203W Dickey, Fuller, 1979 – Dickey, D., Fuller, W. (1979). Distributions of the estimators for autoregressive time series with a unit root, Journal of American Statistical Association, http://dx.doi.org/10.1080/01621459.1979.10482531 74, pp. 427–431 [Electronic resource]. URL: http://www.tandfonline.com/doi/abs/10.1080/01621459.1979.10482531 Dritsakis, Adamopoulos, 2004 – Dritsakis, N., Adamopoulos, A. (2004). Financial Development and Economic Growth in Greece: An Empirical Investigation with Granger Causality Analysis. International Economic Journal, http://dx.doi.org/10.1080/1016873042000299981 18(4), pp. 547–559. [Electronic resource]. URL: http://www.tandfonline.com/doi/abs/ 10.1080/1016873042000299981. Econometric package EViews 5.1, 2005 – Econometric package EViews 5.1 (2005). Edison, Path, 1993 – Edison, H.J., Path, B.D. (1993). A re-assessment of the relationship between real exchange rates and real interest rates:1974-1990. Journal of Monetary Economics, 31, pp. 165-187. [Electronic resource]. URL: https://docslide.net/documents/a-re-assessment-of- the-relationship-between-real-exchange-rates-and-real-interest.html Gjerde, Saettem, 1999 – Gjerde, O., Saettem, F. (1999). Causal relations among stock returns and macroeconomic variables in a small, open economy. Journal of International Financial Markets, Institutions and Money, 9, pp. 61–74 [Electronic resource]. URL: http://www.science direct.com/science/article/pii/S1042443198000365 Gochoco-Bautista, Bautista, 2005 – Gochoco-Bautista, M.S., Bautista, C.C. (2005). Monetary policy and exchange market pressure: The case of the Philippines, Journal of Macroeconomics, 27, pp. 153–168. [Electronic resource]. URL: http://www.sciencedirect. com/science/article/ pii/S0164070404000771 Gogas, Pragidis, 2010 – Gogas, P., Pragidis, I. (2010). The interest rate spread as a forecasting tool of Greek industrial production. International Journal of Business and Economics, 3(1), pp. 37-46. Granger, Newbold, 1986 – Granger, C., Newbold P. (1986). Forecasting Economic Time Series, New York: Academic Press. [Electronic resource]. URL: https://books.google.gr /books?hl=el&lr=&id=oDWjBQAAQBAJ&oi=fnd&pg=PP1&dq=GRANGER,+C.+%26+NEWBOLD,+ P.,+1986.+Forecasting+Economic+Time+Series,+New+York:+Academic+Press.&ots=jZ8y6Knqmx &sig=YGPerzJVE4i_S4QUFLgxIPFYWZ8&redir_esc=y#v=onepage&q&f=false Hahm, 1998 – Hahm, J.H. (1998). Consumption adjustment to real interest rates: Intertemporal substitution revisited, Journal of Economic Dynamics and Control, 22, pp. 293- 320. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/S0165- 1889(97)00053-5 International Monetary Fund, IMF, 2010 – International Monetary Fund, IMF 2010. International Financial Statistics Yearbook, 2010, Washington DC. Johansen, Juselius, 1990 – Johansen, S., Juselius, K. (1992). Testing Structural Hypotheses in a Multivariate Cointegration Analysis at the and the Uncovered Interest Parity for the UK. Journal of Econometrics, 53, pp. 211-244. Kanas, 2005 – Kanas, A. (2005). Regime linkages in the US/UK real exchange rate–real interest differential relation, Journal of International Money and Finance, 24, pp. 257-274 Kandil, 2005 – Kandil M. (2005). Money, interest, and prices: Some international evidence. International Review of Economics and Finance, 14, pp. 129–147. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/S1059056004000176 .

136 European Journal of Economic Studies, 2017, 6(2)

Kane, Rosenthal, 1982 – Kane A., Rosenthal, L. (1982). International interest rates and inflationary expectations, Journal of international money and finance, 1, pp. 97-110. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/0261560682900079 Katos, 2004 – Katos, A. (2004). Econometrics: Theory and Practice, 1st Edn., Zygos, Thessaloniki, Greece. Krugman, 1998 – Krugman. P. (1998). Japan’s Trap. [Electronic resource]. URL: http://web.mit.edu/krugman/www/japtrap.html [Accessed 26 January 2011]. MacKinnon, 2010 – James G. MacKinnon (2010). Critical Values for Cointegration Queen’s Economics Department Working Paper No. 1227 Department of Economics Queen’s University 94 University Avenue Kingston, Ontario, Canada K7L3N6 1-2010Tests qed_wp_1227.pdf 02-Jan- 2010 14:27 191K Nagayasu, 2002 – Nagayasu, J. (2002). On the term structure of interest rates and inflation in Japan. Journal of Economics and Business, 54, pp. 505–523 [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/S0148619502001042. Nakagawa, 2002 – Nakagawa, H. (2002). Real exchange rates and real interest differentials: implications of nonlinear adjustment in real exchange rates. Journal of Monetary Economics, 49, pp. 629–649. Nakagawa, Oshima, 2000 – Nakagawa, S., Oshima, K. (2000). Does a Decrease in the Real Interest Rate Actually Stimulate Personal Consumption? An Empirical Study, Research and Statistics Department, Bank of Japan, Working Paper Series, Working Paper 00-2. Neumeyer, Perri, 2005 – Neumeyer, P.A., Perri F. (2005). Business cycles in emerging economies: the role of interest rates, Journal of Monetary Economics, 52, pp. 345–380. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/S0304- 3932(05)00003-6 Ogaki, Santaellar, 2000 – Ogaki M., Santaellar, J.A. (2000). The exchange rate and the term structure of interest rates in Mexico. Journal of Development Economics, 63, pp. 135–155. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/S0304 387800001036 Papapetrou, 2001 – Papapetrou, E. (2001). Oil price shocks, stock market, economic activity and employment in Greece, Energy Economics, 23(5), pp. 511-532. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/S0140-9883(01)00078-0 Sullivan, Lombra, 1992 – Sullivan, E., Lombra, R. (1992). Interest Rate Effects on Intertemporal Consumption: Do Financial Regulation and Deregulation Matter? Journal of Economics and Business, 44, pp. 115-125. [Electronic resource]. URL: http://www.sciencedirect. com/science/article/pii/0148-6195(92)90010-8 Sweidan, 2004 – Sweidan, O. (2004). Does inflation harm economic growth in Jordan? An econometric analysis for the period 1970-2000. International Journal of Applied Econometrics and Quantitative Studies, 1-2, pp. 41-66. [Electronic resource]. URL: http://www.usc.es/economet /reviews/ijaeqs123.pdf Tunali, 2010 – Tunali, H. (2010). The Analysis of Relationships between Macroeconomic Factors and Stock Returns: Evidence from Turkey Using VAR Model. International Research Journal of Finance and Economics, Issue 57. [Electronic resource]. URL: http://www. internationalresearchjournaloffinanceandeconomics.com/ISSUES/IRJFE%20issue%2057.htm [Accessed 19 January 2011]. Wei, 2008 – Wei, C.-C. (2008). An Empirical Analysis of the Effect of China Interest Rate to Industrial Production and Stock Markets Index by Using the DCC-MEGARCH Model. International Research Journal of Finance and Economics, Issue 18. Available from: [Electronic resource]. URL: http://www.eurojournals.com/irjfe_18_01_Ching.pdf [Accessed 18 January 2011]. Wikipedia Interest rate, n.d. – Wikipedia, n.d. Interest rate. In: Wikipedia: the free encyclopaedia [online]. St Petersburg, Florida: Wikimedia Foundation. [Electronic resource]. URL: http://en.wikipedia.org/wiki/Interest_rate [Accessed 11 March 2011]. Wikipedia n.d. Consumer Price Index (UK)) – Wikipedia, n.d. Consumer Price Index (UK). In: Wikipedia: the free encyclopaedia [online]. St Petersburg, Florida: Wikimedia Foundation. [Electronic resource]. URL: http://en.wikipedia.org/wiki/Consumer_Price_Index [Accessed 4 March 2011].

137 European Journal of Economic Studies, 2017, 6(2)

Wikipedia, n.d cointegration – Wikipedia, n.d. Cointegration. In: Wikipedia: the free encyclopaedia [online]. St Petersburg, Florida: Wikimedia Foundation. [Electronic resource]. URL: http://en.wikipedia.org/wiki/Consumer_Price_Index [Accessed 11 March 2011]. Wikipedia, n.d Developed country – Wikipedia, n.d. Developed country. In: Wikipedia: the free encyclopaedia [online]. St Petersburg, Florida: Wikimedia Foundation. [Electronic resource]. URL: http://en.wikipedia.org/wiki/Developed_country [Accessed 2 March 2011]. Wikipedia, n.d. First World – Wikipedia, n.d. First world. In: Wikipedia: the free encyclopaedia [online]. St Petersburg, Florida: Wikimedia Foundation. [Electronic resource]. URL: http://en.wikipedia.org/wiki/First_World [Accessed 4 March 2011]. Wikipedia, n.d. Keynesian economics – Wikipedia, n.d. Keynesian economics. In: Wikipedia: the free encyclopaedia [online]. St Petersburg, Florida: Wikimedia Foundation. [Electronic resource]. URL: http://en.wikipedia.org/wiki/Keynesian_economics [Accessed 11 March 2011]. Wikipedia, n.d. monetary policy – Wikipedia, n.d. Monetary policy. In: Wikipedia: the free encyclopaedia [online]. St Petersburg, Florida: Wikimedia Foundation. [Electronic resource]. URL: http://en.wikipedia.org/wiki/Monetary_policy [Accessed 11 March 2011]. Wikipedia, n.d. Vector autoregression – Wikipedia, n.d. Vector autoregression. In: Wikipedia: the free encyclopaedia [online]. St Petersburg, Florida: Wikimedia Foundation. [Electronic resource]. URL: http://en.wikipedia.org/wiki/Vector_autoregression [Accessed 2 March 2011]. Zettelmeyer, 2004 – Zettelmeyer, J. (2004). The impact of monetary policy on the exchange rate: evidence from three small open economies. Journal of Monetary Economics, 51, pp. 635– 652. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/S0304- 3932(04)00019-4 Zhang, Wan, 2002 – Zhang, Y., Wan, G.H. (2002). Household consumption and monetary policy in China. China Economic Review, 13, pp. 27–52. [Electronic resource]. URL: http://www.sciencedirect.com/science/article/pii/S1043951X01000554

Appendix

Table 1. Data time period for the 18 developed countries

Data time range for the 18 developed countries* Countries Starting date Ending Date France 1970Q1 2009Q4 Germany 1960Q1** 2009Q4 Japan 1960Q1 2009Q4 USA 1971Q1 2009Q4 * The data starting point of the data range differs from country to country according the available time series data for each country ** The data before 1990Q4 refer to Western Germany, as it was impossible to extract data for the eastern part of the country

138 European Journal of Economic Studies, 2017, 6(2)

Table 2. The common characteristics of developed countries*

The common features and standards of developed countries refer to: 1. the real per capita GDP which is greater than a certain threshold 2. the real per capita income which is higher than the minimum allowable subsistence level 3. the policies for the reallocation of country wealth 4. the policies for the education and health of the population 5. the social policy 6. the human Development Index that includes: 6.1. life expectancy at birth 6.2. mean years of schooling 6.3. expected years of schooling and 6.4. gross national income (GNI) per capita 7. the reliability of the provided and published statistical data 8. the democratic governing for decades 9. the acceptance of economic development as a combination of economic growth and human development, and many other common features and standard of smaller importance. *Sources: 1. First World, wikipedia, n.d. 2. Human Development Index, wikipedia, n.d

Table 3. Money market rates (MMR) for the 4 developed countries

Countries Money market rates (MMR)* France The three-month EURIBOR rate, which is a three-month interbank rate Germany Period averages of ten daily average quotations for overnight credit. Japan The lending rate for overnight loans USA The rate on three-month commercial paper * Source: International Monetary Fund, IMF, 2010. International Financial Statistics, Country Notes 2010

Table 4. Augmented Dickey–Fuller (ADF) unit root tests for the 4 countries

At levels At first differences

Test Test Order statistic Country Variable OSRE* Lag statistic LM(1)** OSRE Lag LM(1) of (DF/AD (DF/ADF) Integr. F)

0.065 0.000 RMMR 1 5 -1.239 1 4 -7.869 [0.799 I(1) [1.000] ] 0.296 LNCPI 2 10 -3.732 - - - - I(0) [0.587] France 0.052 LNIPI 2 1 -3.000 - - - - I(0) [0.819] 0.740 2.064 RRHCGD 2 4 -2.443 1 3 -5.300 [0.390 I(1) [0.151] ] 0.000 0.113 EXCR 2 1 -2.163 1 0 -8.566 [1.000 I(1) [0.737] ] The critical values for OSRE=1 at 1%, 5% and 10% are -2.580, -1.943 and -1.615 resp/ly The critical values for OSRE=2 at 1%, 5% and 10% are -3.473, -2.880 and -2.577 resp/ly The critical values for OSRE=3 at 1%, 5% and 10% are -4.019, -3.439 and -3.144 resp/ly RMMR 2 3 -3.987 0.000 - - - - I(0)

139 European Journal of Economic Studies, 2017, 6(2)

[1.000] 2.515 2.467 LNCPI 1 3 -0.871 2 2 -4.505 I(1) [0.113] [0.116] 0.315 0.013 LNIPI 3 10 -2.615 3 9 -5.605 I(1) Germany [0.574] [0.908] 0.000 RRHCGD 3 1 -3.879 - - - - I(0) [0.999] 0.514 EXCR 1 1 -2.001 - - - - I(0) [0.473] The critical values for OSRE=1 at 1%, 5% and 10% are -2.577, -1.943 and -1.616 resp/ly The critical values for OSRE=2 at 1%, 5% and 10% are -3.465, -2.877 and -2.575 resp/ly The critical values for OSRE=3 at 1%, 5% and 10% are -4.008, -3.434 and -3.141 resp/ly 2.727 RMMR 1 7 -2.367 - - - - I(0) [0.099] 1.144 0.000 LNCPI 2 4 -2.809 1 3 2.113 I(1) [0.285] [1.000] 0.520 LNIPI 2 5 -4.725 - - - - I(0) [0.471] Japan 1.421 2,142 RRHCGD 2 2 -2.352 1 1 -9.185 I(1) [0.233] [0.143] 1.260 0.000 EXCR 1 1 -1.913 1 0 -9.495 I(1) [0.261] [1.000] The critical values for OSRE=1 at 1%, 5% and 10% are -2.577, -1.942 and -1.616 resp/ly The critical values for OSRE=2 at 1%, 5% and 10% are -3.464, -2.876 and -2.575 resp/ly The critical values for OSRE=3 at 1%, 5% and 10% are -4.041, -3.450 and -3.150 resp/ly 0.713 RMMR 3 5 -4.233 - - - - I(0) [0.398] 1.191 LNCPI 2 4 -3.460 - - - - I(0) [0.275] 0.00 US 0.244 0 LNIPI 1 1 2.208 1 0 -6.627 I(1) [0.621] [1.00 0] 0.034 RRHCGD 3 1 -3962 - - - - I(0) [0.854] 0.00 1.531 0 EXCR 3 3 -2.849 1 2 -5.741 I(1) [0.216] [1.00 A 0]

The critical values for OSRE=1 at 1%, 5% and 10% are -2.580, -1.943 and -1.615 resp/ly The critical values for OSRE=2 at 1%, 5% and 10% are -3.474, -2.881 and -2.577 resp/ly The critical values for OSRE=3 at 1%, 5% and 10% are -4.022, -3.441 and -3.145 resp/ly * OSRE = optimal specification of regression equation OSRE= 1. No intersept no trend, =2. Intersept no trend, =3. Intersept and trend ** LM(1)= Lagrange multiplier for first order autocorrelation test *** I(0)= Integrated zero order, I(1)= Integrated first order

140 European Journal of Economic Studies, 2017, 6(2)

141 European Journal of Economic Studies, 2017, 6(2)

Table 6. Normalized cointegrating coefficients of cointegrating equations

Country Variables RMMR LNCPI LNIPI RRHCGD EXRAT C 1.000000 -378.5341 -234.1014 -4745.830 223.8514 5145.047 France (132.291) (535.384) (1781.65) (116.214)

* Standard error in parentheses

RMMR LNCPI LNIPI RRHCGD EXRAT C 1.000000 -71.86437 57.31992 115.7227 -25.17812 14.91902 Germany (16.1101) (16.4194) (134.863) (8.34343)

* Standard error in parentheses

RMMR LNCPI LNIPI RRHCGD EXRAT C 1.000000 -9.739248 9.920346 83.72603 -0.024418 -43.18530 Japan (1.88635) (1.81482) (14.3185) (0.00633)

* Standard error in parentheses

RMMR LNCPI LNIPI RRHCGD EXRAT C

1.000000 -5.972046 -3.675459 224.3058 6.379426 -122.8722

USA (1.64870) (3.90011) (42.8303) (3.06310)

* Standard error in parentheses

Table 7. Granger causality test results

Dependent X2 Wald statistic – significance level Country Variable Excluded Variable D(RMMR) D(LNCPI) D(LNIPI) D(RRHCGD) D(EXRAT) 12.807* 19.733* 7.175 6.512 D(RMMR) - (0.012) (0.001) (0.127) (0.164) 11.734* 3.653 2.891 11.753* D(LNCPI) - (0.019) (0.455) (0.576) (0.019) 1.637 10.585* 4.883 2.837 France D(LNIPI) - (0.802) (0.032) (0.300) (0.586) 2.277 4.158 8.932 3.214 D(RRHCGD) - (0.685) (0.385) (0.063) (0.523) 5.115 5.725 1.732 1.613 D(EXRAT) - (0.276) (0.221) (0.785) (0.806) Lag lengths = 4, df = 4 D(RMMR) D(LNCPI) D(LNIPI) D(RRHCGD) D(EXRAT) 24.123* 18.381* 3.717 22.850* D(RMMR) - (0.000) (0.003) (0.591) (0.000) 17.407* 0.826 3.836 19.328* Germany D(LNCPI) - (0.004) (0.975) (0.573) (0.019) 7.907 16.443* 7.956 3.529 D(LNIPI) - (0.161) (0.006) (0.159) (0.619) D(RRHCGD) 6.562 4.001 13.229* - 7.938

142 European Journal of Economic Studies, 2017, 6(2)

(0.255) (0.549) (0.021) (0.160) 4.370 3.719 5.662 3.106 D(EXRAT) - (0.498) (0.591) (0.341) (0.681) Lag lengths = 5, df = 5 Lag lengths = , df = 4 D(RMMR) D(LNCPI) D(LNIPI) D(RRHCGD) D(EXRAT) 4.523 11.587* 5.163 1.845 D(RMMR) - (0.104) (0.003) (0.076) (0.398) 6.792* 9.310* 2.755 2.364 D(LNCPI) - (0.034) (0.001) (0.252) (0.307) 13.479* 5.629 0.538 4.561 Japan D(LNIPI) - (0.001) (0.060) (0.764) (0.102) 3.583 4.216 22.987* 2.488 D(RRHCGD) - (0.167) (0.122) (0.000) (0.288) 0.694 1.055 1.075 4.258 D(EXRAT) - (0.707) (0.590) (0.584) (0.119) Lag lengths = 2, df = 2 D(RMMR) D(LNCPI) D(LNIPI) D(RRHCGD) D(EXRAT) 3.045* 7.311 7.099 7.802 D(RMMR) - (0.693) (0.319) (0.213) (0.168) 20.138* 2.145 10.120 4.855 D(LNCPI) - (0.001) (0.829) (0.072) (0.4.34) 9.875 8.041 6.143 0.482 USA D(LNIPI) - (0.079) (0.154) (0.293) (0.993) 3.661 3.799 12.760* 3.795 D(RRHCGD) - (0.599) (0.579) (0.026) (0.579) 13.732* 8.842 6.647 6.674 D(EXRAT) - (0.017) (0.116) (0.248) (0.246) Lag lengths =7, df = 7 * Indicate significance at the 5 % level

143