136 Abstract The importance of borrowing is fundamental for central public administration and it consists in sources of fi nancing budget defi cit and refi nanc- ing government debt. In the last years, a lot of countries had diffi culties regarding the payment of public loans at their maturity due to the burden of government debt to GDP ratio. In this situa- tion, investors lose their confi dence not only in the country that is facing problems, but also in other states that pay their debt at maturity. For this reason, they are careful at any change that affects sovereign rating. From our investigation we found that sover- eign rating has a negative infl uence on bonds’ interest rate. As such, decision makers from central public administration should focus on improving sovereign ratings in order to decrease interest rates. Keywords: central public administration, sovereign rating, interest rate, fi nancial market. THE CORRELATION OF SOVEREIGN RATING AND BONDS’ INTEREST RATE IN EU MEMBER STATES*1 Emilian-Constantin MIRICESCU Emilian-Constantin MIRICESCU Associate Professor, Department of Finance, Faculty of Finance, Insurance, Banking and Stock Exchange, Center of Financial and Monetary Research, Bucharest University of Economic Studies, Bucharest, Romania Tel.: 0040-213-191.900 E-mail: emilian.miricescu@fi n.ase.ro * Acknowledgements: This work was supported from the European Social Fund through Sectorial Operational Programme Human Resources Development 2007 – 2013, project number POSDRU/159/1.5/S/134197, project title “Performance and Excellence in Postdoctoral Research in Romanian Economics Science Domain”. Transylvanian Review of Administrative Sciences, No. 45 E/2015, pp. 136-148 137 1. Introduction In contemporary days, the need of the governments to borrow money is gener- ated by the increased demands from citizens to cover the collective needs as most countries had to spend more public money than the revenues collected through man- datory taxes, capital income, external grants and other fi nancial resources (Mirices- cu, 2011). Văcărel et al. (2003) highlighted the need for borrowing as many countries around the world are confronted with the problem of public budget defi cit. Whole loans borrowed by the government, local public administration authorities and by other public institutions, along with related interest and commissions which were not paid, represent the public debt at a certain time. In terms of public borrowing importance, Stroe and Armeanu (2004) emphasized that the fi nancing of temporary cash-fl ow problems and the budget defi cit through loans instead of taxes shows some benefi ts, such as: effi ciency, avoidance of social discontent, and relative decrease of fi scal eff ort over time. If interest rates are too high the government debt will increase rapidly. In this context, Văcărel et al. (2003) highlighted that the elevated level of interest rates collected on foreign loans contrib- uted to the external debt crisis of developing countries. But Moşteanu et al. (2008) explained that because of the serious problems that faced some debtor countries, the Paris Club was founded within the international fi nancial system that can restructure and even cancel public debt. For Romania, sovereign rating has a particular signifi cance, as in June 2014 from the total public administration debt 54.14% was borrowed from the external markets. We consider that an opportunity for central public administration to decrease interest rates is sovereign ratings improving with the purpose of reducing the public debt burden. Figure 1 shows that government debt to GDP ratio in the Euro Area had an in- creasing trend, starting from 69.2% in 2000 and reaching 93.9% in 2014 Q1. Gov- ernment debt to GDP ratio in the European Union started from 61.9% in 2000 and reached to 88% in 2014 Q1. Compared with 2000, the Euro Area member states in- creased with 24.7 percentage points (pp) in their debt to GDP ratio at the end of 2014 Q1. Compared with 2000, the European Union member states increased with 26.1 pp in their debt to GDP ratio at the end of 2014 Q1. In our opinion, central public authorities should have a suitable management of public debt portfolio in order to bett er fulfi l the citizens needs on long term. Figure 2 shows that for 16 member states, government debt to GDP ratio exceeds the ceiling specifi ed by Euro convergence Maastricht criteria (maximum 60%), and for 12 member states the index complies with Maastricht criteria. We fi nd in Figure 2 that government debt to GDP ratio in the European Union member states at the end of 2014 Q1 varies from 174.1% in Greece to 10% in Estonia. In our view, large govern- ment debts lead to large interest expenses in the public budgets, and we study the infl uence of sovereign rating on bonds’ interest rate. 138 69.2% 68.0% 67.8% 68.9% 69.4% 70.1% 68.3% 66.3% 70.1% 79.9% 85.3% 87.3% 90.6% 93.9%92.7% 88 .0 %87.2%85.2%82.5% 80.0% 74.8% 62.5%59.0%61.4% 62.6%62.1%61.9%60.5%61.1%61.9% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Euro Area European Union Figure 1: Government debt to GDP ratio in the Euro Area and in the European Union Source: Our results based on data provided by Eurostat 174.1% 123.7% 112.2% 84.3% 49.5% 44.3% 40.3% 38.2% 22.8% 10.0% 20.3% 39.0%40.4%45.6% 58.4% 58.6% 68.0%75.1% 73.5%75.3% 77.3% 78.7% 91.1% 96.6% 96.8% 105.1% 135.6% 132.9% G re ec e It al y Po rt ug al Ir el an d C yp ru s B el gi um Sp ai n Fr an ce U ni te d K in gd om H un ga ry Sl ov en ia G er m an y M al ta A us tr ia N et he rl an ds C ro at ia Fi nl an d Sl ov ak ia Po la nd C ze ch R ep ub lic D en m ar k Sw ed en L ith ua ni a R om an ia L at vi a L ux em bo ur g B ul ga ri a E st on ia Figure 2: Government debt to GDP ratio in the EU member states at the end of 2014 Q1 Source: Our results based on data provided by Eurostat Figure 3 shows that government debt to GDP ratio in Romania had an increasing trend, starting from 12.6% in 2007 and reaching to 38.4% in 2013. Debt interest to GDP ratio in Romania also had an increasing trend, starting from 0.7% in 2007 and reaching to 1.7% in 2013. So, government debt to GDP ratio increased faster than debt interest to GDP ratio, as a consequence of sovereign bonds’ interest rate diminution. Specialized global rating agencies are Standard & Poor’s, Moody’s Investor Ser- vices and Fitch Ratings. However, sovereign ratings are also provided by local or regional rating agencies, specialized services from banks, foreign trade agencies and fi nance journals (Miricescu, 2011). Sovereign ratings split countries in two clusters: (i) investment grade with rat- ings equal to or above BBB- (for Standard & Poor’s and Fitch Ratings) and Baa3 (for Moody’s Investor Services); (ii) speculative grade with ratings equal to or below BB+ (for Standard & Poor’s and Fitch Ratings) and Ba1 (for Moody’s Investor Services). 139 12.6% 13.4% 23.6% 30.5% 34.7% 37.9% 38.4% 1.7%1.8%1.6%1.4%1.2%0.7%0.7% 2007 2008 2009 2010 2011 2012 2013 Government debt to GDP ratio Debt interest to GDP ratio Figure 3: Government debt to GDP ratio and debt interest to GDP ratio in Romania Source: Our results based on data provided by Eurostat Table 1: Rating scale Standard & Poor’s AAA AA+ AA AA- A+ A A- BBB+ BBB BBB- BB+ BB Fitch AAA AA+ AA AA- A+ A A- BBB+ BBB BBB- BB+ BB Moody’s Aaa Aa1 Aa2 Aa3 A1 A2 A3 Baa1 Baa2 Baa3 Ba1 Ba2 Standard & Poor’s BB- B+ B B- CCC+ CCC CCC- CC - SD D - Fitch BB- B+ B B- CCC+ CCC CCC- CC C DDD DD D Moody’s Ba3 B1 B2 B3 Caa1 Caa2 Caa3 Ca - C C C Source: Our results based on data provided by Bran and Costică (2003) and rating agencies An earlier version of this paper (Miricescu, 2012) was presented and published in the Proceedings of the 6th International Conference on Globalization and Higher Education in Economics and Business Administration. The paper is organized as fol- lows: Section 2 reviews the existing fi nancial literature regarding the correlation be- tween sovereign ratings and the interest rate on sovereign bonds, Section 3 presents the database and research methodology used in our study, Section 4 analyzes the correlation between sovereign ratings and the interest rate on sovereign bonds for 25 developed and emerging EU member states and Section 5 concludes. 2. Literature review In recent years, the correlation between sovereign rating and interest rates on bonds was studied in several quantitative papers; almost all the articles emphasize systematic analysis of sovereign ratings (Cantor and Packer, 1996). Sovereign ratings were established by the rating agencies Moody’s and Standard & Poor’s. Cantor and Packer’s (1996) study was carried out on 49 developed and emerging countries by using a regression analysis having data for the 1987-1994 period. The authors stated that investors in sovereign bonds are excessively pessimistic, when countries with low credit ratings intend to issue government bonds on the international fi nancial 140 market; investors also require high yields on such securities. In order to validate this conclusion, we mention that in July 2014 Romania obtained from Standard & Poor’s an investment grade rating (BBB–) and borrowed money at 4,16% interest rate com- pared with Germany which obtained from Standard & Poor’s an investment grade rating (AAA) and borrowed money at 1,11% interest rate. Min (1998) analyzed determinants of diff erence between government bond yields in emerging economies and the US government bonds yields. The main factors iden- tifi ed as signifi cant in economic terms were: (i) public debt to GDP ratio – positive infl uence, (ii) public debt service to exports – positive infl uence, (iii) net foreign as- sets to GDP ratio – negative infl uence, and (iv) international reserves to GDP ratio – negative infl uence. According to the author, the impact of sovereign ratings on yield spread of government bonds was high. Reisen and von Maltz an (1999) examined the impact of changes in sovereign rat- ings on diff erence between government bond yields having 10 years maturity for 29 emerging economies and the US government bond. The sovereign ratings were es- tablished by the rating agencies Moody’s, Standard & Poor’s and Fitch. The article was carried out by using a sample panel data over the period 1989-1997. The authors concluded that the spreads of government bond yields had already varied before the change of sovereign ratings, namely the fi nancial market anticipated in advance sov- ereign ratings’ changes. In our opinion, the changes in ratings are performed usually once a year (if it is necessary), but investors generally include gradually positive and negative percep- tions on the issuers into bonds’ price. There are relatively few examples of more than one rating change during a year. According to Reisen and von Maltz an (1999), when sovereign rating decreased, the negative infl uence on bond yield spreads continued for about 20 days after the an- nouncement of change. We believe that investors have confi dence in opinions expressed by rating agencies, which are brought into the interest rates of sovereign issuers. Beck (2001) used a panel data and showed that determinants of government bonds spreads are long-term variables, such as determinants of sovereign ratings and fore- casts indicators for economic growth – negative infl uence, domestic infl ation – posi- tive infl uence and international interest rates – positive infl uence. The study was per- formed on nine emerging countries, including data over the period December 1998 – August 2000. Kaminsky and Schmukler (2002) examined 16 emerging countries by using a panel data analysis with data over the period 1990-2000. Sovereign ratings were established by the rating agency Standard & Poor’s. The authors concluded that sovereign ratings and outlooks of emerging countries infl uence government bond yields and also the bond yields of nearby countries – which are vulnerable on macroeconomic indica- tors. The infl uence of sovereign ratings and outlooks on government bond yields are stronger in the period of crisis. The reason for this correspondence is that economic trade increased between nearby countries, and the problems are almost the same. 141 Gonzalez et al. (2004) revealed that sovereign ratings have a moderate correlation with interest rate spreads and this correspondence is more evident for downgrades than for upgrades, and may be more signifi cant for equity prices than for bond prices. Also, liquidity, taxation and historical volatility are determinants for spreads. The authors explain that the basic rationale for using ratings is to achieve information economies of scale and solve principal-agent problems. Powell and Martinez (2008) revealed, by taking into consideration sovereign rat- ings which were established by rating agencies Moody’s and Standard & Poor’s, that a rating model should include: (i) variables of the real sector (economic growth lev- el, GDP per capita, infl ation and unemployment rate); (ii) external sector indicators (external debt, real exchange rate volatility and international reserves), (iii) variables that highlight the political and institutional environment (especially the eff ectiveness of government), and (iv) indicators of public debt and consolidated budget defi cit. The scholars used a regression based on a structured data panel, which showed that sovereign ratings of emerging economies such as Brazil, Argentina, Mexico and Chile have a greater impact on the spread of Credit Default Swap than other fi nancial fac- tors such as: the yield of US T-bills and the yield index of US corporate bonds. Hartelius, Kashiwase and Kodres (2008) carried out a study on 33 emerging econ- omies and the US economy with data collected over the period January 1991 – Febru- ary 2007. The authors used sovereign ratings established by the rating agency Stan- dard & Poor’s. The scholars found that the main determinants for emerging market spreads are both sovereign rating and the US interest rates. Greenidge, Drakes and Craigwell (2010) explored empirically the causality direc- tion between external debt and sovereign rating by using panel causality tests for 32 developed and emerging countries having data over the period 1998-2008. The au- thors used sovereign ratings established by the rating agency Standard & Poor’s, and they concluded that there is a bidirectional causality relationship between external public debt and sovereign rating. The scholars highlighted that sovereign rating is one key element that determine if the lender provides the credit and the interest rate. Jaramillo and Tejada (2011) emphasized that an investment grade sovereign rating is related to low interest rate spreads on fi nancial markets. Based on a panel data anal- ysis over the period 1997-2010 for 35 emerging economies, the researchers showed that sovereign investment grade rating reduces the interest rate spreads against a benchmark with almost 36% from the interest rate spreads involved by the macro- economic fundamentals of these countries. Sahinoz and Gonenc (2011) examined for the period 1998-2008 18 emerging market economies in a panel data. The authors con- cluded that these countries should enhance the economic growth, reduce the public debt to GDP ratio and sustain the institutions’ strength to facilitate recovering the sovereign ratings and to have lower interest rates. Canuto, Santos and Porto (2012) analyzed for the period 1998-2002, 66 countries by using a panel data analysis. The authors used sovereign ratings established by Moody’s, Standard & Poor’s and Fitch. The authors emphasized that a high sovereign 142 rating (which involves a low sovereign risk) is the result of high per capita income, high economic growth and a low government debt. Thus, a low sovereign risk deter- mines a low interest rate on government bonds. Afonso, Gomes and Taamouti (2014) examined 21 EU countries from January 1995 until October 2011 in a panel data. The authors used sovereign ratings established by the rating agencies Standard & Poor’s and Moody’s and Fitch. The authors concluded that rating upgrades do not have any signifi cant eff ect on volatility, but sovereign downgrades increase bonds volatility after two lags. From the fi nancial literature we conclude that sovereign rating has a major impact on interest rates of government bonds, especially in emerging countries which have frequently speculative ratings. In many cases, rating changes performed by rating agencies are already contained in sovereign interest rates. In this case, the interest rate change is small after the press announcement. 3. Database and methodology Our investigation will be conducted on annual data (end of December of each year), over the period 2001-2013 for 25 countries members of European Union by ap- plying a regression analysis. Of the 28 European Union member states we removed Germany, Estonia and Croatia. We excluded Germany because for this country the dependent variable is zero in all cases. According to European Central Bank, there are no Estonian sovereign debt securities that comply with the defi nition of long-term in- terest rates for convergence purposes, and no suitable proxy indicator has been iden- tifi ed; so we decided to eliminate Estonia from our analysis. We excluded Croatia from our analysis because for this country there are less data for long-term interest rates. Independent variable is sovereign rating (RATING) established by the rating agency Standard & Poor’s as a measure of risks associated to the borrower countries. Dependent variable (SP) is the spread between each member state long-term interest rate and Germany’s long-term interest rate. Long-term interest rate represents a con- vergence criteria corresponding to sovereign bonds close to 10 years maturity. Afon- so, Gomes and Taamouti (2014) used 10-year government bonds taken from Reuters for interest rate. We consider Germany’s long-term interest rate to be a proxy for risk- free interest rate, as Germany has the highest GDP from the EU member states and Germany had in many investigated periods the lowest interest rates for sovereign bonds. Pungulescu (2013) also used German government bonds as benchmark rate, and she omitt ed Germany from the regressions. Afonso, Gomes and Rother (2007) used a linear scale and they grouped the ratings in 17 categories, by putt ing together in the same bucket the few observations below B- (as presented in Table 2). We analyzed 25 countries that had ratings ranging from AAA to CC, which leads to the following linear transformation, from qualitative variables in quantitative vari- ables (Table 3). 143 Table 2: Linear transformation of ratings RATING AAA AA+ AA AA- A+ A Transformation (i) 17 16 15 14 13 12 RATING A- BBB+ BBB BBB- BB+ BB Transformation (i) 11 10 9 8 7 6 RATING BB- B+ B B- below B- Transformation (i) 5 4 3 2 1 Source: Afonso, Gomes and Rother (2007) Table 3: Linear transformation of ratings RATING AAA AA+ AA AA- A+ A A- Transformation (i) 20 19 18 17 16 15 14 RATING BBB+ BBB BBB- BB+ BB BB- B+ Transformation (i) 13 12 11 10 9 8 7 RATING B B- CCC+ CCC CCC- CC Transformation (i) 6 5 4 3 2 1 Source: Our results based on data provided by Standard & Poor’s Figure 4 shows that sovereign rating correlates with interest rate spread in 2012. Figure 4: Analysis along with sovereign rating and interest rate spread in 2012 Source: Our results based on data provided by European Central Bank and Standard & Poor’s Afonso, Gomes and Rother (2007) emphasized that one alternative way to over- come the criticism of assuming that the distance between two notches is equal for ev- ery combination of sequential notches is to apply alternative transformations besides the usual linear one. The reason why they applied the logistic transformation is that at the middle of the scale, ratings can rise rather quickly, as the sovereigns deliver some improvements. Both at the bott om and the top end of the rating scale, the in- crease of an additional notch is slower, since the improvements are more demanding. 144 We adjusted and applied the logistic transformation described previously, as follows: SR=ln[Ratingi/(1- Ratingi)], where, Ratingi = (2 * i – 1)/(2 * n); in our study n = 20 Table 4: Logistic transformation of ratings RATING AAA AA+ AA AA- A+ A A- Transformation (i) 20 19 18 17 16 15 14 SR 3.66 2.51 1.95 1.55 1.24 0.97 0.73 RATING BBB+ BBB BBB- BB+ BB BB- B+ Transformation (i) 13 12 11 10 9 8 7 SR 0.51 0.30 0.10 -0.10 -0.30 -0.51 -0.73 RATING B B- CCC+ CCC CCC- CC Transformation (i) 6 5 4 3 2 1 SR -0.97 -1.24 -1.55 -1.95 -2.51 -3.66 Source: Our results based on data provided by Standard & Poor’s Over the period 2001-2013, as a result of ratings logistic transformation, the rating ranges between -3.66 and 3.66. The independent variable (SR) in our model is related in 2012 to the following countries: (i) 3.66 to Germany, Luxembourg, Netherlands, Finland, Denmark, Sweden and the United Kingdom; (ii) 2.51 to France and Austria; (iii) 1.95 to Belgium and Czech Republic; (iv) 1.24 to Slovenia; (v) 0.97 to Slovakia and Poland; (vi) 0.73 to Malta; (vii) 0.51 to Ireland, Spain and Italy; (viii) 0.30 to Bulgaria and Lithuania; (ix) 0.10 to Latvia; (x) -0.10 to Cyprus, Hungary and Romania; (xi) -0.30 to Portugal, and (xii) -1.95 to Greece. Figure 5 shows that in 2012 the transformed sovereign rating (SR) correlates with interest rate spread (SP), which validated the transformation and the use of SR as in- dependent variable. Figure 5: Analysis along with transformed sovereign rating and interest rate spread in 2012 Source: Our results based on data provided by European Central Bank and Standard & Poor’s 145 4. Empirical results On short run we studied for each year the intensity by which the transformed sovereign rating infl uences the interest rate spread, by using cross-section regression at the same moment in time. Cross-section regression is used to highlight the static re- lation between variables for each year studied separately. From the regression equa- tions for each year, we found that as the transformed sovereign rating (SR) increases, the interest rate spread (SP) decreases and vice versa. However the intensity relation between two variables is diff erent from year to year depending on investors’ risk per- ception: (i) in 2003, 2004, 2005, 2006 and 2007, sovereign rating had a small impact on interest rate, estimated coeffi cients ranging from –0.32 to –0.54, and the models explanatory power ranging from 24.93% to 40.09%; and (ii) in 2001, 2002, 2008, 2009, 2010, 2011, 2012 and 2013, sovereign rating had a huge impact on interest rate, esti- mated coeffi cients ranging from –0.43 to –1.95, and the models explanatory power ranging from 50.34% to 73.77%. Table 5: Cross-section regression results Year Regression equation Adjusted R Square Signifi cance F 2001 SP = 2.63 – 0.77 * SR (0.25)§ (0.11)§ 66.25% 0.00 2002 SP = 1.48 – 0.43 * SR (0.20)§ (0.08)§ 50.34% 0.00 2003 SP = 1.76 – 0.53 * SR (0.31)§ (0.13)§ 39.09% 0.00 2004 SP = 1.92 – 0.54 * SR (0.33)§ (0.13)§ 40.09% 0.00 2005 SP = 1.37 – 0.40 * SR (0.33)§ (0.13)§ 24.93% 0.00 2006 SP = 1.30 – 0.37 * SR (0.28)§ (0.11)§ 28.90% 0.00 2007 SP = 1.28 – 0.32 * SR (0.22)§ (0.09)§ 34.96% 0.00 2008 SP = 4.02 – 1.01 * SR (0.43)§ (0.17)§ 58.56% 0.00 2009 SP = 4.25 – 1.24 * SR (0.57)§ (0.25)§ 50.36% 0.00 2010 SP = 4.15 – 1.12 * SR (0.46)§ (0.20)§ 55.57% 0.00 2011 SP = 6.85 – 1.95 * SR (0.55)§ (0.24)§ 73.77% 0.00 2012 SP = 4.51 – 1.38 * SR (0.40)§ (0.19)§ 67.82% 0.00 2013 SP = 2.98 – 0.89 * SR (0.25)§ (0.13)§ 66.23% 0.00 §Standard Error of OLS estimators, all estimators show signifi cance at 1% level. Over the period 2001-2013, F-values show signifi cance at 1% level. Source: Our results based on data provided by European Central Bank and Standard & Poor’s On the long run, for the entire period 2001-2013 we studied the intensity by which the transformed sovereign rating infl uences the interest rate spread, by using data 146 panel regression (data panel regression is used to highlight the dynamic relation be- tween variables for the entire period of time). According to the data panel regression methodology, the fi rst stage is represented by the stationary analysis, and we chose to use the test Im, Pesaran, Shin both for dependent variable and also for independent variable. Im, Pesaran and Shin (2003) argued that they have developed a simple com- putational procedure for testing the unit root hypothesis in heterogeneous panels. Table 6: Im, Pesaran, Shin panel unit root test Variable IPS statistic Signifi cance level SP - level -5.01 0.00 SR - level -6.24 0.00 Source: Our results based on data provided by European Central Bank and Standard & Poor’s The SP variable is stationary because the signifi cance level is lower than the threshold of 0.01. The SR variable is stationary because the signifi cance level is lower than the threshold of 0.01. One possible explanation for this situation is the fact that in order to defi ne the quantitative variable SR we carried out the logistic transformation of qualitative variable sovereign rating, after we performed the linear transformation. In our opinion, another possible explanation resides in minor fl uctuations of sover- eign ratings for each country, excepting the last periods, when sovereign ratings had rapid changes. Table 7: Data panel regression results Regression equation Adjusted R Square Signifi cance level SPt = 3.26 – 0.96 * SRt (0.14)§ (0.06)§ 44.41% 0.00 §Standard Error of OLS estimators, all estimators show signifi cance at 1% level. For the entire period 2001-2013, F-values show signifi cance at 1% level. Source: Our results based on data provided by European Central Bank and Standard & Poor’s From the regression equation for the entire period 2001-2013, we found that as the transformed sovereign rating (SR) increases by 1% it determines the interest rate spread (SP) to decrease by 0.96%. The intensity relation between the variables consid- ered as panel data is moderate as the model explanatory power is 44.41%. For the entire period 2001-2013 we analyzed the direction of presumed causality from transformed sovereign rating (SR) and interest rate spread (SP), on the Grang- er causality test. We found a bi-directional causality among transformed sovereign rating (SR) and interest rate spread (SP), as the signifi cance level is lower than the threshold of 0.05. Table 8: Granger causality test Null Hypothesis: F-Statistic Signifi cance level SP does not Granger Cause SR 5.86 0.02 SR does not Granger Cause SP 11.16 0.00 Source: Our results based on data provided by European Central Bank and Standard & Poor’s 147 5. Conclusions and recommendations Comparing the results of the cross-section regression with the results of the panel data regression we found that, in both analyses, the sovereign rating has a strong negative infl uence on bonds’ interest rate, as previous studies found. In our opinion, public decision makers from EU member states must improve sov- ereign ratings in order to decrease interest rates. Thus, central public administration will have money to satisfy bett er the collective needs of the citizens living in their territory. From the regression analysis we found that the correlation between the two indexes is stronger during fi nancial or economic crises, and it is weaker during the periods of economic growth. However an important element of the interest rate is the risk-free interest rate, which increases or decreases independently of sovereign rating. In order to improve sovereign ratings and decrease interest rates, we recommend that public decision makers from the EU member states to increase, for example: (i) GDP per capita, (ii) real GDP growth, (iii) foreign exchange reserves, (iv) political stability, (v) government eff ectiveness, and so on, and to decrease, for example: (i) government debt to GDP ratio, (ii) public budget defi cit, (iii) infl ation, (iv) unemploy- ment, and so on. In the last three decades, the importance of sovereign rating for international in- vestors became fundamental as there are states having problems with the payment of principal loans and interests. References: 1. Afonso, A., Gomes, P. and Rother, P., ‘What “Hides” Behind Sovereign Debt Ratings?’, 2007, European Central Bank Working Paper Series, no. 711, pp. 1-65. 2. Afonso, A., Gomes, P. and Taamouti, A., ‘Sovereign Credit Ratings, Market Volatility, and Financial Gains’, 2014, European Central Bank Working Paper Series, no. 1654, pp. 1-52. 3. Beck, R., ‘Do Country Fundamentals Explain Emerging Market Bond Spreads?’, 2001, Center for Financial Studies Working Paper, no. 2001/02, pp. 1-32. 4. Bran, P. and Costică, I., Economica activităţii fi nanciare şi monetare internaţionale (The Eco- nomics of International Financial and Monetary Activity), Bucharest: Economica Publish- ing House, 2003. 5. Cantor, R. and Packer, F., ‘Determinants and Impact of Sovereign Credit Ratings’, 1996, FRBNY Economic Policy Review, no. 2, pp. 37-53. 6. Canuto, O., Santos, P.F.P. and Porto, P.C.S., ‘Macroeconomics and Sovereign Risk Rat- ings’, 2012, Journal of International Commerce, Economics and Policy, vol. 3, no. 2, pp. 1-36. 7. Gonzalez, F., Haas, F., Johannes, R., Persson, M., Toledo, L., Violi, R., Wieland, M. and Zins, C., ‘Market Dynamics Associated with Credit Ratings – A Literature Review’, 2004, European Central Bank Occasional Paper Series, no. 16, pp. 1-36. 8. Greenidge, K., Drakes, L. and Craigwell, R., ‘Debt and Sovereign Credit Ratings: Ev- idence from Panel Causality Tests’, 2010, [Online] available at htt p://www.ccmfuwi. 148 org/fi les/publications/conference/2010/9_4-Craigwell_Drakes_Greenidge-p.pdf, acces- sed on September 24, 2014. 9. Hartelius, K., Kashiwase, K. and Kodres, L.E., ‘Emerging Market Spread Compression: Is it Real or Is it Liquidity?’, 2008, IMF Working Paper, no. WP/08/10, pp. 1-36. 10. Im, K.S., Pesaran, M.H. and Shin, Y., ‘Testing For Unit Roots in Heterogeneous Panels’, 2003, Journal of Econometrics, vol. 115, no. 1, pp. 53-74. 11. Jaramillo, L. and Tejada, C.M., ‘Sovereign Credit Ratings and Spreads in Emerging Markets: Does Investment Grade Matt er?’, 2011, IMF Working Paper, no. WP/11/44, pp. 1-17. 12. Kaminsky, G. and Schmukler, S.L., ‘Emerging Markets Instability: Do Sovereign Rat- ings Aff ect Country Risk and Stock Returns?’, 2002, World Bank Economic Review, vol. 16, no. 2, pp. 171-195. 13. Min, H.G., ‘Determinants of Emerging Market Bond Spread: Do Economic Fundamen- tals Matt er?’, 1998, The World Bank Policy Research Working Paper, no. 1899, pp. 1-33. 14. Miricescu, E.C., The Sovereign Rating, Bucharest: ASE Publishing House, 2011. 15. Miricescu, E.C., ‘The Correlation of Sovereign Rating and Bonds’ Interest Rate in the EU Member States’, in The Proceedings of the 6th International Conference on Globalization and Higher Education in Economics and Business Administration, 2012, pp. 798-803. 16. Moşteanu, T., Câmpeanu, E., Cataramă, D. and Miricescu, E.C., Gestiunea datoriei pub- lice (Public Debt Management), Bucharest: Universitară Publishing House, 2008. 17. Powell, A. and Martínez, S.J.F., ‘On Emerging Economy Sovereign Spreads and Rat- ings’, 2008, Inter-American Development Bank Working Paper Series, no. 629, pp. 1-42. 18. Pungulescu, C., ‘Measuring Financial Market Integration in the European Union: EU15 vs. New Member States’, 2013, Emerging Markets Review, vol. 17, pp. 106-124. 19. Reinsen, H. and von Maltz an, J., ‘Boom and Bust and Sovereign Ratings’, 1999, Interna- tional Finance, vol. 2, no. 2, pp. 273-293. 20. Sahinoz, S. and Gonenc, R., ‘Determinants of Credit Ratings, Sovereign Bond Spreads and Real Interest Rates in Emerging Markets’, 2011, Iktisat Isletme Ve Finans, vol. 26, no. 305, pp. 9-35. 21. Stroe, R. and Armeanu, D., Finanţe (Finances), Bucharest: ASE Publishing House, 2004. 22. Văcărel, I., Bistriceanu, G., Bercea, F., Anghelache, G., Moşteanu, T., Bodnar, M. and Georgescu, F., Finanţe publice (Public fi nances), Bucharest: Didactică şi Pedagogică Pub- lishing House, 2003.