© 2017 Nicolaus Copernicus University. All rights reserved. http://www.dem.umk.pl/dem D Y N A M I C E C O N O M E T R I C M O D E L S DOI: http://dx.doi.org/10.12775/DEM.2017.009 Vol. 17 (2017) 147−159 Submitted November 28, 2017 ISSN (online) 2450-7067 Accepted December 28, 2017 ISSN (print) 1234-3862 Dorota Witkowska , Krzysztof Kompa * How the Change of Governing Party Influences the Efficiency of Financial Market in Poland A b s t r a c t. Financial market seems to be sensitive to political changes, especially when the change of governing party is connected with essential changes of the economic development concepts. Such situation took place in Poland in 2015, as a result of the presidential and par- liamentary elections. The aim of our research is to investigate the changes occurred on the market, represented by some stable growth open mutual funds, and stock indexes: WIG and TBSP. Analysis is provided applying single index and CAPM models, classical investment performance measures, and statistical interference. K e y w o r d s: stable growth open mutual funds, investment efficiency, Sharpe model, CAPM, Sharpe, Treynor and Jensen ratios. J E L Classification: G11; C12. Introduction Financial market seems to be sensitive to political changes, especially when the change of governing party is connected with essential changes of the concepts concerning economic development. Such situation took place in Poland at the end of 2015, as a result of the presidential and parliamentary elections that were won by the Law and Justice party (PiS) which was in an opposition to the government during two last terms. Now PiS is the largest * Correspondence to: Krzysztof Kompa, Warsaw University of Life Sciences, Department of Econometrics and Statistics, 166 Nowoursynowska Street, 02-787 Warsaw, Poland, e-mail: krzysztof_kompa@sggw.pl; Dorota Witkowska, University of Lodz, Faculty of Management, Department of Finance and Strategic Management, 22/26 Matejki Street, 90-237 Łódź, Po- land, e-mail: dorota.witkowska@uni.lodz.pl. https://orcid.org/0000-0001-9538-9589 http://orcid.org/0000-0002-2810-6654 Dorota Witkowska, Krzysztof Kompa DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 148 party in the Polish parliament having majority in both chambers of parlia- ment. The governing party introduced program called “good change” con- sisting in populist movements such as decreasing of the retirement age or the 500+ Familly Programme, etc. which burden the economy and may affect the financial market. Therefore, here a question arises how the change of ruling party and their economic program influence the situation of Polish financial market. The answer is not easy especially that both sides i.e. governing and opposite par- ties presented completely different arguments which had rather political than economic character. PiS was emphasizing social benefits of proposed pro- grams while the opposition was highlighting the economic consequences and threats for the budget. Some economists even forecasted that financial mar- ket in Poland may collapse since investors do not trust markets with high political risk which comes from social programs together with controversial economic proposals such as increasing taxes from the banking sector or su- permarkets. Therefore, the aim of our research is to find out how the change on the political scene affected the performance of equity and bond markets, togeth- er with stable growth mutual funds (FIO), applying single index and capital assets pricing models together with classical investment performance measures and statistical interference. 1. Data and Methodology Our investigation is carried out using daily logarithmic rates of returns from selected financial instruments:  Warsaw Stock Exchange Index – WIG, representing equity market,  Poland’s Official Treasury Bonds Index – TBSP.Index, representing bond market,  participation units of stable growth open mutual funds (FIO): Credit Agricole Stabilnego Wzrostu (denoted as CA), KBC Fundusz Stabilny (as KBC), Nationale-Nederlanden Stabilnego Wzrostu (as NN), Pioneer Stabilnego Wzrostu (as PIO) and PZU Stabilnego Wzrostu MAZUREK (as PZU). The analysis is provided for the period from 10.10.2013 to 9.12.2016 (the whole period including 826 observations). This time span is divided into five pairs of sub-periods according to the selected events which we take into consideration:  the presidential elections: the first round of election – 10.05.2015, the second round of election – 24.05.2015, How the Change of Governing Party Influences the Efficiency of Financial Market… DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 149  the parliamentary election – 25.10.2015,  the new government appointment – 16.11.2015,  the entry into force 500+ Familly Program – 1.04.2016. The subperiods are defined assuming that the last observation comes from 9.12.2017, and sharing date is a day when the distinguished event took place. It was our concern to obtain subsamples with similar number of ob- servations (detailed information about sample sizes are in Table 2). Investigation of returns and risk generated by the investment portfolios constructed by selected funds is conducted in several steps, beginning from the analysis of the basic parameters and applying statistical interference 1 (assuming the significance level 0.05). Denoting by: E(R) – expected returns, D 2 (R) – variance of returns, RWIG, RTBSP, RFIO – returns from WIG, TBSP and FIO respectively, ,  – parameters of Sharpe model or CAPM, Rbefore, Rafter, before, after – returns from the portfolio and beta coefficients before and after the considered event, respectively, we verify the null hypotheses concerning: 1. rates of return levels, i.e.: E(RFIO) = 0; E(RWIG) = 0; E(RTBSP) = 0, 2. parameters of Sharpe and CAPM models, i.e.:  = 0;  = 0, 3. comparisons of parameters values in two considered sub-periods i.e.: E(Rbefore) = E(Rafter), D 2 (Rbefore) = D 2 (Rafter), before = after, before = after. We apply the classical tests for verification hypothesis of returns equity:  using the Cochran-Cox test statistics: (1)  and using the following statistics: (2) where for the k-th period, – average logarithmic rate of return from the selected instrument, – variance of return, – number of observations, B – benchmark: B=0, B=R1 or B=R2. The comparison of returns in both period is provided using statistics (1) and (2). In the letter case benchmark B is de- fined as an average value of returns obtained in the second considered period (k=1, 2) and the test is provided as two-way test. Comparison of variances is provided using the test with Fisher statistics: (3) 1 All formulas (1)-(5) are discussed in (Witkowska 2016, p. 29-55). Dorota Witkowska, Krzysztof Kompa DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 150 where, – maximal and minimal variance obtained for the both compared samples. The shape of the probability distribution of logarithmic rates of return is examined on the basis of parametric tests verifying hypothesis that sym- metry and kurtosis equal to zero, applying the following statistics: (4) (5) where, At, Kt – the third and the fourth central statistical moments of loga- rithmic rates of returns. The next step of our research is estimation of Sharpe and capital assets pricing models on the basis of daily logarithmic rates of return 2 : (6) (7) where for the t-th period, Rit – rate of return from participation units of the i-th stable growth open mutual fund; Rrt – rate of return from the market index (WIG); Rrt – rate of return from the risk-free instrument (TBSP); i, i – model parameters, it – random component; t – number of observation (t=1, 2, …, T). Parameters of both models are estimated using OLS method. Analysis of parameter significance in the models is provided using the test statistics 3 : and (8) where, – parameter estimates, – standard estimation er- rors from the models (6) and (7). Comparison of parameters obtained in both comparable periods is made applying the test statistics: and (9a) and (9b) 2 Models are widely discussed in literature see for instance: (Zamojska 2012, p. 57-60), (Witkowska 2016, p. 41-48) 3 These measures are discussed in (Witkowska 2016, p. 49-50). How the Change of Governing Party Influences the Efficiency of Financial Market… DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 151 The performance of mutual funds is provided using classical measures i.e. Sharpe, Treynor and Jensen ratios 4 . The first two measures estimate the obtained risk premium evaluated for the unit of risk and the decision about the efficiency is made by comparing the values of the ratios obtained by the mutual fund and the market index. Jensen ratio is the parameter estimate from CAPM, and investment is efficient if is positive. The comparison of the funds’ efficiency is provided applying measures mentioned above which are evaluated for all considered mutual funds in all analyzed time spans, assuming that WIG represents market index, and TBSP is the risk-free instrument. To recognize if the Sharpe ratios are equal, we apply Jobson-Korkie test (Jobson, Korkie, 1981) with Memmel correction (Memmel, 2003), using test statistics – see (Blitz, van Vliet, 2007), (Kurach, Papla, 2014): (10) where for the k-th period of analysis (k=1,2), WSi – Sharpe ratio, ρ12 – corre- lation coefficient, (i=1, 2). 2. Changes of the Equity and Bond Markets In the first step of our analysis we investigate daily rates of return of indexes WIG and TBSP. Tables 1–3 contain basic characteristics of logarithmic rates of return. Bold letters denote rejection of null hypotheses. Table 1. Basic characteristics of daily logarithmic rates of return from the both benchmarks evaluated for the whole period of analysis Basic parameters WIG TBSP Basic parameters WIG TBSP min –5.8250% –0.6366% range 8.8302% 1.2498% max 3.0052% 0.6133% standard deviation 0.9022% 0.1752% arithmetic mean –0.0026% 0.0148% coefficient of variability 343.12 11.80 median 0.0000% 0.0091% interquartile deviation 0.9530% 0.1958% quartile I –0.4495% –0.0774% asymmetry –0.7278 –0.2519 quartile III 0.5036% 0.1184% kurtosis 4.2876 1.0511 Note: Bold letters denote rejection of null hypothesis. It is visible (Table 1) that expected rate of returns from Treasury Bonds is significantly positive while average returns from equity market is negative 4 The description of the efficiency ratios, their application and discussion can be found in many publications. Good examples could be (Borkowski 2014), (Perez 2012), (Zamojska 2012) and Witkowska (2016). Dorota Witkowska, Krzysztof Kompa DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 152 although the null hypothesis cannot be rejected. TBSP can be treated as risk- free instrument because its variability is very low. Time series of daily rates of return from WIG and TBSP are asymmetric with positive kurtosis thus they are not normally distributed. However according to the huge number of observation we assume that probability distribution is asymptotical normal. Table 2. Test statistics verifying the hypothesis about expected returns from both indexes in the considered periods H0: E(RWIG) = 0 and H0: E(RTBSP) = 0 Number of observations Rates of returns from Number of observations Rates of returns from WIG TBSP TBSP TBSP Whole period for the presidential elections: 10.10.2013–9.12.2016 Whole period for the parliamentary elections: 8.09.2014–9.12.2016 826 –0.0597 1.7347 590 –0.3203 0.9267 Period before the first round of presidential elec- tions: 8.10.2013–10.05.2015 Period after the first round of presidential elec- tions: 11.05.2015–9.12.2016 409 0.6326 3.0922 417 –0.6079 0.4903 Period before the second round of presidential elections: 8.10.2013–22.05.2015 Period after the second round of presidential elections: 25.05.2015–9.12.2016 419 0.5652 2.9436 407 –0.5767 0.5054 Period before the parliamentary elections: 8.09.2014–23.10.2015 Period after the parliamentary elections: 26.10.2015–9.12.2016 295 –0.4297 1.3199 295 –0.0541 –0.0376 Period before the new government appointment: 20.10.2014–13.11.2015 Period after the new government appointment: 16.11.2015–9.12.2016 280 –0.5547 0.9040 280 0.2423 –0.0931 Period before entry into force 500+ Familly Pro- gram: 23.07.2015–31.03.2016 Period after entry into force 500+Familly Program: 1.04.2016–9.12.2016 181 –0.3644 1.0793 181 0.2689 –0.5718 Note: Bold letters denote rejection of null hypothesis. Analyzing returns from both markets in distinguished 12 periods of con- sideration, one may notice (Table 2) that only Treasury Bonds generated significantly positive rates of return in the periods before both rounds of presidential election and in the whole analyzed period. It is visible that change on the politic scene caused decline in the both markets. However, the performance of the equity market shows slight improvement in the periods after new government appointment and when 500+ Family Program entered into force. Such situation may be a result of investors emotions and expecta- tions which revealed earlier and anticipated both events. The obtained results are also validated by the tests which are used for comparison of returns in considered sub-periods (Table 3). Higher returns are observed only for TBSP before both rounds of presidential election. Eq- uity market was characterized by the significant increase of risk after the government appointment, presidential and parliamentary elections however How the Change of Governing Party Influences the Efficiency of Financial Market… DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 153 the risk significantly decreased after the 500+ Family Program went into effect. Table 3. Test statistics verifying the hypothesis about expected returns and risk of equity and bond markets in considered periods E(Rbefore) = E(Rafter), D 2 (Rbefore) = D 2 (Rafter) Index Test statistics evaluated due to formulas for returns: for risk for returns: for risk (2) (2) (1) (3) (2) (2) (1) (3) presidential elections, the 1st round presidential elections, the 2nd round WIG 1.3649 1.1186 0.8652 1.5179 1.2906 1.0112 0.7960 1.5823 TBSP 2.5649 2.5592 1.8116 1.0241 2.4287 2.3494 1.6886 1.0381 Parliamentary elections Government appointment WIG –0.3663 –0.3125 –0.0574 1.3068 –0.8394 –0.7142 –0.1330 1.3810 TBSP 1.3560 1.4143 0.2362 1.0040 0.9925 1.0440 0.1758 1.1066 Introduction of 500+ Program WIG –0.5945 –0.6945 –0.1231 1.3649 TBSP 1.5723 1.8236 0.3247 1.3452 Note: positive values of test statistics denote that returns are bigger before the considered event than after. Italic letters denote that risk was smaller after the event than before, bold – rejection of null hypothesis. 3. Mutual Funds Market The mutual fund market is represented by five selected stability growth mu- tual funds. All these funds started their functioning in Poland in years 1999– –2003, the “oldest” is FIO PZU, and the “youngest” FIO Credit Agricole. Analysis is provided for the sub-periods constructed around both rounds of the presidential election because in other sub-periods no essential changes was observed. Table 4. Values of test statistics (2) verifying the hypothesis about expected returns from mutual funds in considered periods H0: E(RFIO) = 0 Period CA KBC NN PIO PZU Whole sample 0.8114 0.4773 0.2734 –0.3305 –0.1463 Before the presidential elections 1st round 2.2363 1.4043 1.1328 0.4378 0.4042 After the presidential elections 1st round –0.3498 –0.5631 –0.4387 –1.0512 –0.6572 Before the presidential elections 2nd round 2.0612 1.2635 1.0595 0.3959 0.3531 After the presidential elections 2nd round –0.2664 –0.4694 –0.4092 –1.0385 –0.6432 Note: Bold letters denote rejection of null hypothesis. Analyzing expected returns form participation units of mutual funds, it is visible that before both rounds of the presidential election mutual funds generated positive returns while after – the negative ones Dorota Witkowska, Krzysztof Kompa DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 154 (Table 4). However, the null hypothesis is rejected only for FIO Credit Agricole (significantly positive returns are observed before the election). Tabele 5. Values of statistics (1)–(3) verifying the hypothesis about expected returns and risk in considered periods E(Rbefore) = E(Rafter), D 2 (Rbefore) = D 2 (Rafter) Period Parameter Formulas no. CA KBC NN PIO PZU presidential elections, the 1st round returns (2) 2.6162 1.8575 1.6271 1.5188 1.0643 (2) 2.2442 2.2198 1.4018 1.4663 1.0493 (1) 1.7034 1.4246 1.0620 1.0549 0.7472 risk (3) 1.3856 1.4007 1.3736 1.0939 1.0489 presidential elections, the 2nd round returns (2) 2.3801 1.6672 1.5480 1.5040 1.0250 (2) 1.9594 1.9103 1.2780 1.3892 0.9671 (1) 1.5127 1.2561 0.9856 1.0205 0.7034 risk (3) 1.4333 1.3515 1.4252 1.1386 1.0912 Note: Bold letters denote rejection of null hypothesis. Table 6. Parameter estimates and determination coefficients of Sharpe models Before After Before After The whole period presidential elections, the 1st round presidential elections, the 2nd round beta alfa beta alfa beta alfa beta alfa beta alfa FIO Credit Agricole 0.2445 0.0002 0.2405 0.0000 0.2441 0.0002 0.2409 0.0000 0.2425 0.0001 R2 0.6230 R2 0.6604 R2 0.6176 R2 0.6641 R2 0.6446 FIO PZU 0.4285 0.0000 0.3404 0.0000 0.4274 0.0000 0.3410 0.0000 0.3752 0.0000 R2 0.8493 R2 0.7756 R2 0.8444 R2 0.7792 R2 0.8010 FIO Pioneer 0.3892 0.0000 0.3347 –0.0001 0.3887 0.0000 0.3349 –0.0001 0.3563 0.0000 R2 0.8843 R2 0.9078 R2 0.8821 R2 0.9098 R2 0.8917 FIO Nationale-Nederlanden 0.3323 0.0001 0.3240 0.0000 0.3324 0.0001 0.3240 0.0000 0.3274 0.0001 R2 0.8672 R2 0.9107 R2 0.8651 R2 0.9125 R2 0.8924 FIO KBC 0.4163 0.0002 0.2730 0.0000 0.4159 0.0001 0.2732 0.0000 0.3300 0.0001 R2 0.7786 R2 0.7118 R2 0.7748 R2 0.7153 R2 0.7181 Note: Bold letters denote statistically significant. The better performance of analyzed funds before the election is also proved by the results presented in Table 5. As one can see, better perfor- mance before the election was visible for FIO Credit Agricole and FIO KBC. FIO Credit Agricole and FIO Nationale-Nederlanden were characterized by significantly smaller risk before election however FIO KBC generated re- How the Change of Governing Party Influences the Efficiency of Financial Market… DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 155 turns with smaller volatility after the election. Null hypotheses are not re- jected for FIO Pioneer, FIO PZU and FIO Nationale-Nederlanden. Beta parameters in the single index models and CAPM are significantly positive however the values of  parameter estimates are rather small (Ta- bles 6–7). That is connected with the fact that portfolios of stable growth mutual funds contain a great share of bonds. At the end of January 2017, Credit Agricole stable growth fund’s portfolio contains only 25% of equity, FIO PZU – 30.3%, FIO Pioneer – 29.3%, FIO Nationale-Nederlanden – 35.3% and FIO KBC –38.2%. The structure of the investment funds’ portfo- lios is also visible when beta parameter estimates are analyzed since the biggest value is observed for FIO PZU and KBC while the smallest for FIO Credit Agricole. Table 7. Parameter estimates and determination coefficients of CAPM Before After Before After The whole period presidential elections, the 1st round presidential elections, the 2nd round beta alfa beta alfa beta alfa beta alfa beta alfa FIO Credit Agricole 0.2171 0.0000 0.2072 0.0000 0.2168 0.0000 0.2074 0.0000 0.2112 0.0000 R2 0.8023 R2 0.7528 R2 0.7998 R2 0.7543 R2 0.7727 FIO PZU 0.4054 –0.0002 0.3103 –0.0001 0.4042 –0.0002 0.3109 –0.0001 0.3485 –0.0001 R2 0.9109 R2 0.8216 R2 0.9079 R2 0.8232 R2 0.8514 FIO Pioneer 0.3754 –0.0002 0.3121 –0.0001 0.3752 –0.0002 0.3121 –0.0001 0.3375 –0.0001 R2 0.9083 R2 0.9258 R2 0.9069 R2 0.9262 R2 0.9088 FIO Nationale-Nederlanden 0.3166 –0.0001 0.3012 0.0000 0.3169 –0.0001 0.3010 0.0000 0.3073 –0.0001 R2 0.9173 R2 0.9302 R2 0.9161 R2 0.9305 R2 0.9235 FIO KBC 0.3885 0.0000 0.2474 0.0000 0.3879 0.0000 0.2475 0.0000 0.3042 0.0000 R2 0.8464 R2 0.7054 R2 0.8445 R2 0.7068 R2 0.7486 Note: Bold letters denote statistically significant parameters. Value and significance of the alpha parameter is important when capital assets pricing models are taken into consideration since this parameter is an efficiency measure i.e. Jensen ratio. In estimated models, none of alphas is significantly positive (Table 7). The best portfolio management can be no- ticed for FIO Credit Agricole and KBC since alphas equaled zero. For the rest of funds Jensen ratios were significantly negative at least in the periods before election and for the whole period, it means that the mutual fund man- Dorota Witkowska, Krzysztof Kompa DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 156 agers did not earned enough return given the amount of risk they were tak- ing. In the next step, we compare betas from the models estimated before and after both rounds of the presidential election (Table 8). Since a positive value of the test statistics means that before the election the parameter was bigger than after the election, it is visible that the risk, measured by beta, signifi- cantly lowered after the election for all mutual funds but FIO Credit Agricole. Taking into account the quality of management, we notice that it was significantly improved after the election by FIO PZU and FIO Nationale-Nederlanden, however one must realize that Jensen alphas re- mained negative. Table 8. Value of statistics comparing betas estimated in both periods H0: before = after Funds Period Sharpe beta CAPM beta CAPM alpha t1 t2 t1 t2 t1 t2 Credit Agricole I round 0.4255 0.4706 1.8703 1.7048 0.8368 0.6278 II round 0.3411 0.3750 1.7730 1.5932 0.4965 0.3601 PZU I round 9.8989 9.7889 15.1606 13.3466 –2.4622 –1.7842 II round 9.6437 9.5620 14.8252 12.9878 –2.2242 –1.5727 Pioneer I round 7.7857 10.2830 10.7358 14.5523 –1.5012 –1.6749 II round 7.7518 10.2474 10.7422 14.3758 –1.1776 –1.2720 NN I round 1.2969 1.6600 3.2745 3.7870 –2.3184 –2.2070 II round 1.3137 1.6850 3.3972 3.8855 –2.1582 –1.9923 KBC I round 13.0273 16.6628 17.2482 17.9215 0.7775 0.6650 II round 13.0234 16.6094 17.2693 17.6845 0.4928 0.4073 Note: Bold letters denote rejection of null hypothesis. The last stage of our investigation consists in evaluation the classical efficiency measures, which are given in Tables 9 and 10. Treynor ratio uses beta as a measure of risk but some Authors apply beta estimated from the single index model – see (Domański 2011, p. 64), (Perez 2011, p.155), and other – take beta from CAPM – see (Borowski 2014, p. 20), (Białek 2009, p. 34). Therefore, we use both approaches in our analysis. The main conclu- sion from our research is that the majority of portfolios were ineffective, except FIO Credit Agricole. Inefficiency appeared more often after than before the presidential election. After the election Sharpe and Treynor ratios show negative risk premium, and they usually decreased in comparison to the first analyzed time span, although Jensen alphas for FIO PZU and Nationale-Nederlanden increased in the samples containing observations after the presidential election. How the Change of Governing Party Influences the Efficiency of Financial Market… DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 157 Table 9. Values of the efficiency measures evaluated for mutual funds before and after both rounds of presidential election Ratio: Sharpe Treynor (β Sharpe) Treynor (β CAPM) Jensen alpha Fund or index Periods before rounds of presidential election 1st round 2nd round 1st round 2nd round 1st round 2nd round 1st round 2nd round CA 0.00392 –0.00015 0.00004 0.00000 0.00004 0.00000 0.00001 0.00001 PZU –0.05050 –0.05011 –0.00044 –0.00043 –0.00047 –0.00046 –0.00018 –0.00017 PIO –0.05753 –0.05635 –0.00049 –0.00048 –0.00051 –0.00049 –0.00019 –0.00017 NN –0.03614 –0.03591 –0.00031 –0.00031 –0.00033 –0.00032 –0.00010 –0.00009 KBC –0.00056 –0.00459 –0.00001 –0.00004 –0.00001 –0.00004 0.00000 –0.00001 WIG –0.00173 –0.00372 –0.00001 –0.00003 –0.00001 –0.00003 x x Periods after rounds of presidential election CA –0.03148 –0.02775 –0.00038 –0.00034 –0.00045 –0.00040 –0.00002 –0.00001 PZU –0.04315 –0.04270 –0.00049 –0.00048 –0.00053 –0.00053 –0.00006 –0.00006 PIO –0.06353 –0.06314 –0.00066 –0.00066 –0.00071 –0.00071 –0.00037 –0.00012 NN –0.03399 –0.03284 –0.00035 –0.00034 –0.00038 –0.00037 –0.00004 –0.00001 KBC –0.04069 –0.03642 –0.00048 –0.00043 –0.00053 –0.00048 –0.00019 –0.00005 WIG –0.03399 –0.02817 –0.00034 –0.00028 –0.00034 –0.00028 x x Note: Bold letters denote that Sharpe and Treynor ratios evaluated for mutual funds are bigger than the ones calculated for WIG and Jensen ratios are statistically significant. Table 10. Values of the efficiency measures evaluated for whole period Sharpe Treynor (β Sharpe) Treynor (β CAPM) Jensen alpha CA –0.01485 –0.00017 –0.00019 0.00000 PZU –0.04640 –0.00047 –0.00050 –0.00011 PIO –0.05976 –0.00057 –0.00060 –0.00014 NN –0.03413 –0.00033 –0.00035 –0.00005 KBC –0.01894 –0.00020 –0.00022 –0.00001 WIG –0.01938 –0.00017 –0.00017 x Note: Bold letters denote that Sharpe and Treynor ratios evaluated for mutual funds are bigger than the ones calculated for WIG and Jensen ratios are statistically significant. Here the question arises if changes of efficiency measures observed in comparable periods are statistically significant. To verify such hypothesis the test with statistics (10) is used for Sharpe ratio, together with the test statistics (1) and (2), applied for average values of Sharpe and Treynor rati- os, evaluated for all analyzed funds. As one may notice in Table 11, neither differences of Sharpe ratios between periods nor differences between mutual fund and market index are significant. However, if average values of Treynors ratios are compared the better investment performance before the election is proved (Table 12) since the higher risk premium was obtained. Dorota Witkowska, Krzysztof Kompa DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 158 Table 11. Values of test statistics in Jobson – Korkie test Funds or index Comparison of Sharpe ratios in two sub-periods between FIO Credit Agricole and WIG 1st round 2nd round CA –0.03148 –0.02775 1st round of presidential election 0.17623 PZU –0.04315 –0.04270 PIO –0.06353 –0.06314 2nd round of presidential election 0.11181 NN –0.03399 –0.03284 KBC –0.04069 –0.03642 in the whole period of analysis 0.20728 WIG –0.03399 –0.02817 Table 12. Values of test statistics for average ratios No. of formula Sharpe Treynor (β Sharpe) Treynor (β CAPM) I round II round I round II round I round II round (2) 1.6922 1.4133 2.7795 2.5495 2.9429 2.8190 (2) 3.9387 2.6980 5.4554 4.0513 5.5685 4.2752 (1) 1.5547 1.2520 2.4765 2.1578 2.6019 2.3534 Note: Bold letters denote rejection of null hypothesis. Conclusion Our research show that in the whole period of analysis (covering more than three years), statistically significant and positive returns were generated only by the bond market, represented by index TBSP. The mutual stabile growth fund FIO Credit Agricole also obtained positive rates of return but only in the periods before both rounds of the presidential election. Taking into account other considered instruments we cannot reject the hypotheses about zero returns. Comparison of the return and risk level let us conclude that: 1. returns from Treasury Bonds significantly decreased after both rounds of the presidential election, 2. risk of the bond market significantly decreased after entry into force 500+ Family Program, 3. returns from equity market did not significantly change but rates of re- turn were lower after both rounds of the presidential election, 4. equity market risk increased after both rounds of the presidential elec- tion, the parliamentary election and appointment of the government, 5. equity market risk decreased after the program 500+ started. How the Change of Governing Party Influences the Efficiency of Financial Market… DYNAMIC ECONOMETRIC MODELS 17 (2017) 147–159 159 References Białek, J. (2009), Konstrukcja miar efektywności Otwartych Funduszy Emerytalnych (Con- struction of efficiency measures for open pension funds), Wydawnictwo UŁ, Łódź. Blitz, D., van Vliet, P. (2007), The Volatility Effect, Journal of Portfolio Management, 34(1), 102–113, DOI: https://doi.org/10.3905/jpm.2007.698039. Borowski, K. (2014), Miary efektywności zarządzania na rynkach finansowych (Management efficiency measures on financial markets), Difin, Warszawa. Domański, C. (2011) (red.), Nieklasyczne metody oceny efektywności i ryzyka. Otwarte fundusze emerytalne (Non-classic methods of risk and performance evaluation. Open pension funds), PWE, Warszawa. Jobson, J. D., Korkie, B. M. (1981), Performance hypothesis testing with the Sharpe and Treynor measures, Journal of Finance, 36(4), 889–908, DOI: https://doi.org/10.1111/j.1540-6261.1981.tb04891.x. Kurach, R., Papla, D. (2014), Inwestycje alternatywne w portfelach otwartych funduszy emerytalnych (Alternative investments in open pension funds’ portfolios), Optimum. Studia Ekonomiczne, 1(67), 71–81, DOI: https://doi.org/10.15290/ose.2014.01.67.06. Memmel, C. (2003), Performance hypothesis testing with the Sharpe ratio, Finance Letters, 1(1), 21–23. Perez, K. (2012), Efektywność funduszy inwestycyjnych (Efficiency of mutual funds), Difin, Warszawa. Witkowska, D. (2016), Zmiana warunków funkcjonowania a efektywność inwestycyjna otwar- tych funduszy emerytalnych (The change of functioning conditions vs investment effi- ciency of open pension funds), Wydawnictwo UŁ, Łódź. Zamojska, A. (2012), Efektywność funduszy inwestycyjnych w Polsce. Studium teoretyczno- empiryczne (Efficiency of mutual funds in Poland. Theoretical-empirical study), C.H. Beck, Warszawa. Jak wpływa zmiana partii rządzącej na efektywność rynku finansowego w Polsce Z a r y s t r e ś c i. Rynek finansowy wydaje się być wrażliwym na zmiany polityczne, zwłaszcza gdy zmiana partii rządzącej wiąże się z zasadniczymi zmianami koncepcji rozwoju gospodarczego. Taka sytuacja miała miejsce w Polsce w 2015 roku w wyniku wyborów prezydenckich i parlamentarnych. Celem naszych badań jest analiza zmian zachodzących na rynku, reprezentowanym przez niektóre stabilne, otwarte fundusze inwestycyjne oraz indeksy giełdowe: WIG i TBSP. Przeprowadzono analizę, stosując modele jednoczynnikowe i CAPM, klasyczne miary efektywności inwestycji i wnioskowanie statystyczne. S ł o w a k l u c z o w e: otwarte fundusze inwestycyjne stabilnego wzrostu; efektywność inwestycji; model Sharpe; CAPM; Sharpe, Treynor i Jensen.