EKONOMETRIA. Zastosowanie matematyki w ekonomii. Redaktor naukowy Janusz Łyko

Podobne dokumenty
EKONOMETRIA. Zastosowanie matematyki w ekonomii. Redaktor naukowy Janusz Łyko

Transakcje insiderów a ceny akcji spółek notowanych na Giełdzie Papierów Wartościowych w Warszawie S.A.

SSW1.1, HFW Fry #20, Zeno #25 Benchmark: Qtr.1. Fry #65, Zeno #67. like

Stargard Szczecinski i okolice (Polish Edition)

EKONOMETRIA. Zastosowanie matematyki w ekonomii. Redaktor naukowy Janusz Łyko

Patients price acceptance SELECTED FINDINGS

Katowice, plan miasta: Skala 1: = City map = Stadtplan (Polish Edition)

Tychy, plan miasta: Skala 1: (Polish Edition)

EFEKT DŹWIGNI NA GPW W WARSZAWIE WPROWADZENIE

Machine Learning for Data Science (CS4786) Lecture11. Random Projections & Canonical Correlation Analysis

Dochody gospodarstw rolnych a ryzyko walutowe

Revenue Maximization. Sept. 25, 2018

/ / * ** ***

EKONOMETRIA ECONOMETRICS 1(39) 2013

Cracow University of Economics Poland. Overview. Sources of Real GDP per Capita Growth: Polish Regional-Macroeconomic Dimensions

Network Services for Spatial Data in European Geo-Portals and their Compliance with ISO and OGC Standards

Zakopane, plan miasta: Skala ok. 1: = City map (Polish Edition)

Karpacz, plan miasta 1:10 000: Panorama Karkonoszy, mapa szlakow turystycznych (Polish Edition)

Hard-Margin Support Vector Machines

Machine Learning for Data Science (CS4786) Lecture 11. Spectral Embedding + Clustering

PRACE NAUKOWE Uniwersytetu Ekonomicznego we Wrocławiu

MaPlan Sp. z O.O. Click here if your download doesn"t start automatically

ARNOLD. EDUKACJA KULTURYSTY (POLSKA WERSJA JEZYKOWA) BY DOUGLAS KENT HALL

Prognozowanie średniego miesięcznego kursu kupna USD

Studia Ekonomiczne. Zeszyty Naukowe Uniwersytetu Ekonomicznego w Katowicach ISSN Nr

OSTC GLOBAL TRADING CHALLENGE MANUAL

Analysis of Movie Profitability STAT 469 IN CLASS ANALYSIS #2

STATYSTYCZNA WERYFIKACJA MODELU CAPM NA PRZYKŁADZIE POLSKIEGO RYNKU KAPITAŁOWEGO WPROWADZENIE METODOLOGIA TESTOWANIA MODELU

Miedzy legenda a historia: Szlakiem piastowskim z Poznania do Gniezna (Biblioteka Kroniki Wielkopolski) (Polish Edition)

Wybrzeze Baltyku, mapa turystyczna 1: (Polish Edition)

Financial support for start-uppres. Where to get money? - Equity. - Credit. - Local Labor Office - Six times the national average wage (22000 zł)

Wojewodztwo Koszalinskie: Obiekty i walory krajoznawcze (Inwentaryzacja krajoznawcza Polski) (Polish Edition)

Rozpoznawanie twarzy metodą PCA Michał Bereta 1. Testowanie statystycznej istotności różnic między jakością klasyfikatorów

Domy inaczej pomyślane A different type of housing CEZARY SANKOWSKI

KRZYSZTOF JAJUGA Katedra Inwestycji Finansowych i Zarządzania Ryzykiem Akademia Ekonomiczna we Wrocławiu 25 LAT EKONOMETRII FINANSOWEJ

Magdalena Sokalska Szkoła Główna Handlowa. Modelowanie zmienności stóp zwrotu danych finansowych o wysokiej częstotliwości

MODELOWANIE FINANSOWYCH SZEREGÓW CZASOWYCH Z WARUNKOWĄ WARIANCJĄ. 1. Wstęp

Multi-Criteria Decision Support using ELECTRE methods

Institutional Determinants of IncomeLevel Convergence in the European. Union: Are Institutions Responsible for Divergence Tendencies of Some

WYKORZYSTANIE TEORII CHAOSU ZDETERMINOWANEGO W PROGNOZOWANIU KROKOWYM ROCZNEGO ZUŻYCIA ENERGII ELEKTRYCZNEJ PRZEZ ODBIORCÓW WIEJSKICH

POLITYKA PRYWATNOŚCI / PRIVACY POLICY

PRZYKŁAD ZASTOSOWANIA DOKŁADNEGO NIEPARAMETRYCZNEGO PRZEDZIAŁU UFNOŚCI DLA VaR. Wojciech Zieliński

Krytyczne czynniki sukcesu w zarządzaniu projektami

Has the heat wave frequency or intensity changed in Poland since 1950?

Miedzy legenda a historia: Szlakiem piastowskim z Poznania do Gniezna (Biblioteka Kroniki Wielkopolski) (Polish Edition)

Updated Action Plan received from the competent authority on 4 May 2017

WYBRANE TESTY NIEOBCIĄŻONOŚCI MIAR RYZYKA NA PRZYKŁADZIE VALUE AT RISK

Proposal of thesis topic for mgr in. (MSE) programme in Telecommunications and Computer Science

EKONOMETRIA. Zastosowanie matematyki w ekonomii. Redaktor naukowy Janusz Łyko

PEX PharmaSequence monthly report - January 2018 Total open market (sell-out report)

Sequential Monte Carlo: Selected topics

Latent Dirichlet Allocation Models and their Evaluation IT for Practice 2016

Pielgrzymka do Ojczyzny: Przemowienia i homilie Ojca Swietego Jana Pawla II (Jan Pawel II-- pierwszy Polak na Stolicy Piotrowej) (Polish Edition)

DYNAMICZNE MODELE EKONOMETRYCZNE

OeconomiA copernicana. Małgorzata Madrak-Grochowska, Mirosława Żurek Uniwersytet Mikołaja Kopernika w Toruniu


Towards Stability Analysis of Data Transport Mechanisms: a Fluid Model and an Application

Eliza Khemissi, doctor of Economics

y = The Chain Rule Show all work. No calculator unless otherwise stated. If asked to Explain your answer, write in complete sentences.

EKONOMETRIA. Zastosowanie matematyki w ekonomii. Redaktor naukowy Janusz Łyko

Linear Classification and Logistic Regression. Pascal Fua IC-CVLab

Wyzwania praktyczne w modelowaniu wielowymiarowych procesów GARCH

Jak zasada Pareto może pomóc Ci w nauce języków obcych?

ABOUT NEW EASTERN EUROPE BESTmQUARTERLYmJOURNAL

DM-ML, DM-FL. Auxiliary Equipment and Accessories. Damper Drives. Dimensions. Descritpion

deep learning for NLP (5 lectures)

Test sprawdzający znajomość języka angielskiego

TESTOWANIE STABILNOŚCI PARAMETRÓW WIELOCZYNNIKOWYCH MODELI MARKET TIMING Z OPÓŹNIONĄ ZMIENNĄ RYNKOWĄ 1

Cracow University of Economics Poland

Rodzaj obliczeń. Data Nazwa klienta Ref. Napędy z pasami klinowymi normalnoprofilowymi i wąskoprofilowymi 4/16/ :53:55 PM

PORTS AS LOGISTICS CENTERS FOR CONSTRUCTION AND OPERATION OF THE OFFSHORE WIND FARMS - CASE OF SASSNITZ

WYKORZYSTANIE MODELI AUTOREGRESJI DO PROGNOZOWANIA SZEREGU CZASOWEGO ZWIĄZANEGO ZE SPRZEDAŻĄ ASORTYMENTU HUTNICZEGO

Jazz EB207S is a slim, compact and outstanding looking SATA to USB 2.0 HDD enclosure. The case is

Zestawienie czasów angielskich

OPTYMALIZACJA PUBLICZNEGO TRANSPORTU ZBIOROWEGO W GMINIE ŚRODA WIELKOPOLSKA

pl. Grunwaldzki 24, Wrocław


Bank i Kredyt. Czasopismo NBP poświęcone ekonomii i finansom. National Bank of Poland s Journal on Economics and Finance. czerwiec 2008 ROK XXXIX

Leading organiza5on represen5ng the Business Services Sector in Poland ABSL. June 2013

Ukryte funkcjonalności w oprogramowaniu i urządzeniach elektronicznych. mgr inż. Paweł Koszut

KORELACJA 1. Wykres rozrzutu ocena związku między zmiennymi X i Y. 2. Współczynnik korelacji Pearsona

MODELE PROGNOSTYCZNE SPRZEDAśY ENERGII ELEKTRYCZNEJ ODBIORCOM WIEJSKIM OPARTE NA WYMIARZE FRAKTALNYM, LOGISTYCZNE I KRZYśOWANIA HEURYSTYCZNEGO

Wykaz linii kolejowych, które są wyposażone w urządzenia systemu ETCS

Demand Analysis L E C T U R E R : E W A K U S I D E Ł, PH. D.,

Wojewodztwo Koszalinskie: Obiekty i walory krajoznawcze (Inwentaryzacja krajoznawcza Polski) (Polish Edition)

Wykaz linii kolejowych, które są wyposażone w urzadzenia systemu ETCS

Sustainable mobility: strategic challenge for Polish cities on the example of city of Gdynia

Presentation of results for GETIN Holding Group Q Presentation for investors and analyst of audited financial results

Wykorzystanie testu t dla pojedynczej próby we wnioskowaniu statystycznym


photo graphic Jan Witkowski Project for exhibition compositions typography colors : : janwi@janwi.com

Wroclaw, plan nowy: Nowe ulice, 1:22500, sygnalizacja swietlna, wysokosc wiaduktow : Debica = City plan (Polish Edition)

Immigration Studying. Studying - University. Stating that you want to enroll. Stating that you want to apply for a course.

TTIC 31210: Advanced Natural Language Processing. Kevin Gimpel Spring Lecture 8: Structured PredicCon 2

AN ATTEMPT OF APPLY THE WEIBULL DISTRIBUTION IN ROAD TRAFFIC LOSSES ANALYSIS PRÓBA ZASTOSOWANIA ROZKŁADU WEIBULLA DO ANALIZ STRAT W RUCHU DROGOWYM

Helena Boguta, klasa 8W, rok szkolny 2018/2019

Journal of Agribusiness and Rural Development CHANGES IN PRODUCTIVITY OF SELECTED BRANCHES OF THE POLISH FOOD SECTOR IN THE YEARS

FORMULARZ APLIKACYJNY CERTYFIKACJI STANDARDU GLOBALG.A.P. CHAIN OF CUSTODY GLOBALG.A.P. CHAIN OF CUSTODY APPLICATION FORM

Instrukcja konfiguracji usługi Wirtualnej Sieci Prywatnej w systemie Mac OSX

Transkrypt:

EKONOMETRIA 26 Zasosowanie maemayki w ekonomii Redakor naukowy Janusz Łyko Wydawnicwo Uniwersyeu Ekonomicznego we Wrocławiu Wrocław 2009

Spis reści Wsęp... 7 Beaa Bal-Domańska, Ekonomeryczna analiza sigma i bea konwergencji regionów Unii Europejskiej... 9 Andrzej Bąk, Anea Rybicka, Marcin Pełka, Modele efeków głównych i modele z inerakcjami w conjoin analysis z zasosowaniem programu R. 25 Kaarzyna Budny, Kuroza wekora losowego... 44 Wikor Ejsmon, Opymalna liczebność grupy sudenów... 55 Kamil Fijorek, Model regresji dla cechy przyjmującej warości z przedziału (0,1) ujęcie bayesowskie... 66 Paweł Hanczar, Wyznaczanie zapasu bezpieczeńswa w sieci logisycznej... 77 Roman Hupas, Meody szacowania wewnąrzdziennej sezonowości w analizie danych finansowych pochodzących z pojedynczych ransakcji... 83 Aleksandra Iwanicka, Wpływ zewnęrznych czynników ryzyka na prawdopodobieńswo ruiny w skończonym horyzoncie czasowym w wieloklasowym modelu ryzyka... 97 Agnieszka Lipiea, Sany równowagi na rynkach warunkowych... 110 Krysyna Melich-Iwanek, Polski rynek pracy w świele eorii hiserezy... 122 Rafał Piszczek, Zasosowanie modelu logi w modelowaniu upadłości... 133 Marcin Salamaga, Próba weryfikacji eorii paryeu siły nabywczej na przykładzie kursów wybranych walu... 149 Anoni Smoluk, O zasadzie dualności w programowaniu liniowym... 160 Małgorzaa Szulc-Janek, Influence of recommendaions announcemens on sock prices of fuel marke... 170 Jacek Welc, Regresja liniowa w szacowaniu fundamenalnych współczynników Bea na przykładzie spółek giełdowych z sekorów: budownicwa, informayki oraz spożywczego... 180 Andrzej Wilkowski, O współczynniku korelacji... 191 Mirosław Wójciak, Klasyfikacja nowych echnologii energeycznych ze względu na deerminany ich rozwoju... 199 Andrzej Wójcik, Wykorzysanie modeli wekorowo-auoregresyjnych do modelowania gospodarki Polski... 209 Kaarzyna Zeug-Żebro, Rekonsrukcja przesrzeni sanów na podsawie wielowymiarowych szeregów czasowych... 219

6 Spis reści Summaries Beaa Bal-Domańska, Economeric analysis of sigma and bea convergence in he European Union regions... 24 Andrzej Bąk, Anea Rybicka, Marcin Pełka, Main effecs models and main and ineracions models in conjoin analysis wih applicaion of R sofware... 43 Kaarzyna Budny, Kurosis of a random vecor... 53 Wikor Ejsmon, Opimal class size of sudens... 65 Kamil Fijorek, Regression model for daa resriced o he inerval (0,1) Bayesian approach... 76 Paweł Hanczar, Safey sock level calculaion in a supply chain nework... 82 Roman Hupas, Esimaion mehods of inraday seasonaliy in ransacion financial daa analysis... 96 Aleksandra Iwanicka, An impac of some ouside risk facors on he finie- -ime ruin probabiliy for a muli-classes risk model... 109 Agnieszka Lipiea, Saes of coningen marke equilibrium... 121 Krysyna Melich-Iwanek, The Polish labour marke in ligh of he hyseresis heory... 132 Rafał Piszczek, Logi model applicaions for bankrupcy modelling... 148 Marcin Salamaga, Aemp o verify he purchasing power pariy heory in he case of some foreign currencies... 159 Anoni Smoluk, On dual principle of linear programming... 168 Małgorzaa Szulc-Janek, Analiza wpływu rekomendacji analiyków na ceny akcji branży paliwowej (Analiza wpływu rekomendacji analiyków na ceny akcji branży paliwowej)... 178 Jacek Welc, A linear regression in esimaing fundamenal beas in he case of he sock marke companies from consrucion, i and food indusries... 190 Andrzej Wilkowski, Abou he coefficien of correlaion... 198 Mirosław Wójciak, Classificaion of new energy relaed echnologies based on he deerminans of heir developmen... 208 Andrzej Wójcik, Using vecor-auoregressive models o modelling economy of Poland... 218 Kaarzyna Zeug-Żebro, Sae space reconsrucion from mulivariae ime series... 227

PRACE NAUKOWE UNIWERSYTETU EKONOMICZNEGO WE WROCŁAWIU nr 76 Ekonomeria 26 2009 Małgorzaa Szulc-Janek Wrocław Universiy of Economics INFLUENCE OF RECOMMENDATIONS ANNOUNCEMENTS ON STOCK PRICES OF FUEL MARKET Summary: In he aricle he auhor answers he quesion wheher in accordance o effeciveness heory of capial marke, recommendaions of analyss have a linear influence on sock prices or maybe he resuls are closer o chaos heory and is combinaion of fracal analysis. Sock prices irrespecive of he capial marke secor are he reflecion of he public informaion available on he marke. Neverheless invesor decisions are commonly driven by individual needs, requiremens and paerns of behaviour, very ofen irraional and no conneced wih public informaion. Chaos heory is offering an alernaive o he classical approach owards he way capial marke is funcioning. From is perspecive capial marke is a nonlinear, dynamic sysem driven by deerminisic chaos. According o E. Peeres here exiss long-erm influence of informaion from he pas and endency so as no o include fundamenal informaion abou he sock prices ino marke valuaion. The ool used in he aricle o measure he impac of marke influx of informaion on he sock prices is even analysis. Key words: even analysis, esimaion window, even window, abnormal rae of reurn, recommendaion, p-value. 1. Inroducion In he aricle he auhor is answering he quesion wheher in accordance wih effeciveness heory of capial marke, recommendaions of analyss have a linear influence on sock prices or maybe he resuls are closer o chaos heory and is combinaion of fracal analysis. Sock prices, irrespecively of he capial marke secor, are he reflecion of he public informaion available on he marke. Neverheless invesors decisions are commonly driven by individual needs, requiremens and paerns of behaviour, very ofen irraional and no being conneced wih public informaion. Chaos heory is offering an alernaive o he classical approach owards he way he capial marke is funcioning. From is perspecive capial marke is a nonlinear, dynamic sysem driven by deerminisic chaos. According o E. Peeres here exiss long-erm influence of informaion from he pas and endency so as no o include fundamenal informaion abou he sock prices ino

Influence of recommendaions announcemens on sock prices... 171 marke valuaion [Szablewski, Pniewski, Barosiewicz 2008, p. 213]. The ool used in he aricle o measure he impac of marke influx of informaion on he sock prices is he even analysis. 2. The even analysis A company is an insiuion paricipaing in he marke and being managed in order o creae income for is owners. I is no an isolaed uni bu is prone o various ouside and inside facors. Even analysis is an alernaive way of esing he resuls of differen facors on financial markes or resuls of decisions made inside he companies on heir marke value. I analyses closer and disan background of he organizaion. Figure 1 depics main facors which cause sock prices variaions. I is worh menioning ha inside facors such as splis, dividends or income forecass may influence he prices sronger han ouside poliical changes and differen ypes of risk such as for insance ineres rae risk. Inside facors are playing he role of hedge insrumens, seps aken by company so as o hedge from ouside independen facors. This relaion depends on a company profile. Recommendaions Poliical changes Changes in indexes Managemen changes Sock prices Income forecas, new conracs Dividends Splis Ineres rae changes Changes in correlaed marke secors Fig. 1. Inside and ouside facors changing value of sock prices Source: own figure on he basis of [Gurgul 2006].

172 Małgorzaa Szulc-Janek An advanage of his mehod is coming from is simpliciy and few requiremens, which makes i adjused o various secors and companies. According o A. McWilliams and D. Siegel even analysis is consising of he following seps [McWilliams, Siegel 1997, pp. 626-657]: 1. Definiion of evens providing new informaion for he company. 2. Economic heories research clarifying he influence of he paricular evens on he sock prices. 3. Creaing he group of companies affeced by he evens. 4. Even window seing (required clarificaions for windows longer han wo days). 5. Eliminaion of companies affeced by evens no aken ino analysis. 6. Saisical measuremens of esimaed and abnormal raes of reurn, and heir significance level. 7. Simulaing he resuls of negaive abnormal raes of reurn in he sample and Wilcoxon es resuls. 8. Boosrap analysis for small samples and is impac on he whole resuls. 9. Economic heories and zero hypohesis verificaion. 10. Aaching he lis of companies aken ino consideraion in he analysis wih daes of evens lised. The approach presened above is very ofen discussed as some poins such as for insance very shor even windows (suggesed no longer han wo days) may be no appropriae for all evens. Even is defined as new public informaion available on he marke which may be conneced wih plans or implemened acions wihin he companies affeced [Gurgul 2006, pp. 33-34]. The informaion should fulfil he following requiremens [Tabak, Dunbar 1999]: 1) i has o be explici and precise, 2) public appearance on he marke mus be easy o idenify, 3) informaion canno be driven by marke paricipan before is appearance, 4) informaion used in he even analysis is he only one for he ime being available for he secor, company. There are some dispuable poins in he even analysis such as he lengh of even window and window of esimaion. The even window is a period in which he impac of informaion is analyzed. I can have fixed or flexible lengh bu he fixed windows may be used only for appropriaely large samples (due o he fac ha only in considerably numerous samples effecs of under- and overesimaed evens can be balanced). In he experimen fixed windows equal five days (wo days before he even, he day of he even occurrence and wo days afer he even) are applied wha will be furher explained. General rule is ha he even window should be as long as required so as o cover he whole even. I is also imporan o apply appropriae esimaion window. As opposed o he even window, his is he window used for esimaion parameers of he model for calculaions esimaed rae of reurn in even window. I is crucial o choose he esimaion windows wihou

Influence of recommendaions announcemens on sock prices... 173 disurbing evens. In realiy i is difficul o choose periods for pre-even windows wihou an influence of oher evens and o decrease heir impac bu no omi hem oally in he aricle he arificial variable z i is used for days of esimaion window: z i } z = {0,1, where i 1, he dae of even =. 0, he res of he days In he aricle esimaed and abnormal raes of reurns are used. Imporan sep of he even analysis is conneced wih choosing he righ dependence beween marke insrumen (in he aricle WIG20 index) and he insrumen analyzed (for insance sock prices). The mos common model is marke model [Gurgul 2006, p. 44]: Rˆ = α + β R, i, i i m, where: R ˆi, rae of reurn on sock prices in period, α i free parameer, β i sensiiviy of marke changes, rae of reurn on marke insrumen in period (e.g. index WIG20). R m, Parameers of his model are esimaed using OLS mehod: T 1 = 1 ( R R )( R R ) ˆ β =, i i, i m, m T 1 2 ( R 1 i, Ri) = 1 T 1 T 1 ˆ α ( ˆ i = R 1 i, β R 1, ) =, = m T 1 where: = 1,..., T 1 belongs o he group of indexes of observaions in pre-even window. T is an index of firs observaion in he even window. The model wih arificial variable Z i menioned above is applied in his analysis: Rˆ = α + β R + z. i, i i m, i Using arificial variable for abnormal raes of reurn esimaion in pre-even window, he evens wih he same direcion as he esimaed one (buy, sell or keep he sock) are aken ino analysis bu do no have he considerable impac on he resuls. Marke model was checked by economerical es for auocorrelaion, heeroscedasciy, normaliy of he residuals.

174 Małgorzaa Szulc-Janek 3. Daa The se of evens for he analysis was prepared on he basis of announcemens on he pages of sock marke newspaper Parkie. The assumpion was made ha recommendaions which appear here are firs recommendaions, his meaning ha hey have no been available for marke paricipans before. Only recommendaions from he bigges rade houses such as DM BZ WBK, DM BH, DB Securiies, ING Securiies, Millennium DM, DI BRE BANK were analyzed. The secor which is analyzed in he aricle is fuel secor due o las ime violaions, considerable number of evens and a share of he bigges fuel companies in marke index WIG20 (for companies creaing he index here is low probabiliy ha prices are he resuls of inside under he couner acions because hese companies are inensively examined by invesors, heir fundamenal informaion is analyzed horoughly). Moreover, he correlaion beween he liquidiy of he companies from his secor and possibiliy o manipulae wih oo opimisic or pessimisic recommendaions is no high. The ime horizon for analyzed socks is from June 2006 ill January 2009. Recommendaions o buy, sell or keep were analyzed separaely. The day of even was he day when he recommendaions appeared for he firs ime. If here were oher onedirecion recommendaions in he even window, hey were ignored. In he preeven window arificial variable for oher one-direcion recommendaions was applied. Wha was analyzed in he marke model applied were daily raes of reurn (logarihmic) calculaed according o he equaion: R = ln( P ) ln( P), where: P sock price in period. + 1 Addiionally in he analysis he following hypoheses were esed [Gurgul 2006, p. 77]: recommendaion BUY is a posiive signal from he marke and i generaes increases of he sock prices; recommendaion KEEP does no have any influence on he sock prices; recommendaion SELL is a negaive signal from he marke and i generaes decreases of he sock prices. The fixed even window is applied wih he lengh of five days (wo days before he even, he day of he even, wo days afer he even). This lengh is suggesed in lieraure of his subjec as wihin five days he even has go considerable impac on he prices. In mos cases wo days before is he period when price changes are conneced wih inside rading, influx of informaion coming from inside of he company. Day of he even is he day when marke verifies he price on he basis of informaion announced and wo days afer is he ime when he impac of he even is decreasing. According o he definiion of he even analysis only he days wih visible even impac can be couned for analysis and ha is why he win-

Influence of recommendaions announcemens on sock prices... 175 dow has go such lengh. I is checked if in he even window he abnormal raes of reurn are significan. Abnormal raes of reurn are calculaed according o he equaion [Gurgul 2006, p. 77]: AR = R E( R Ω ), i, i, i, EW where: Ω EW se of informaion in even window. Significance of esimaed raes of reurn is checked on he basis of saisical Suden es wih (N k 1) degrees of freedom: sa ˆ σ Nk = 1 1 1 0 + T 1 2 ˆ AR = ( AR ) AR = 0 Nk T 1 σ 1 Nk = 1 AR = T AR AR 0 + T 1 = 0 i,, AR, where: N k number of even in k group (for k = 1: buy, for k = 2: sell, for k = 3: keep), T' he lengh of esimaion window. The lengh of esimaion window (pre-even) is equal o 20, 25 and 30 days analysis is conduced for differen lenghs. 4. Empirical resuls Firsly, he possibiliy o apply linear marke model for he analysis was checked. The following ess were used: Durbin Wason auocorrelaion es, Goldfeld Quand heeroscedasciy es, Doornik Hansen normaliy of residuals es. The analysis was prepared wih he usage of GRETL. Resuls are shown in Table 1. Irrespecive of he pre-even window (20, 25 or 30 days of esimaion) he linear model can be used. The grea majoriy of each simulaion in every class in each esimaion window is showing he lack of auocorrelaion, homoscedasciy and normaliy of residuals. In he ables 2, 3 and 4 below resuls of Suden es obained for every day in he even window are shown. Presened AR is he average abnormal rae of reurn used for he whole sample of evens (N k is he number of recommendaions analyzed in each group in differen esimaion windows).,

176 Małgorzaa Szulc-Janek Table 1. Linear marke model verificaion Esimaion window 30 25 20 Linear model esing resuls Tes % of success in simulaion BUY KEEP SELL GQ 84.21 100.00 83.33 DW 84.21 90.00 66.67 DH 78.95 90.00 100.00 GQ 89.47 100.00 71.43 DW 84.21 91.67 71.43 DH 89.47 91.67 85.71 GQ 66.67 92.31 85.71 DW 75.00 92.31 85.71 DH 66.67 84.62 85.71 Source: own calculaions. I is noiceable ha in he firs group BUY here is no significan influence of recommendaions on he sock prices. Abnormal raes of reurn are insignifican. Very low effeciveness of recommendaion can be conneced wih difficulies wih inerpreaion, low level of confidence in analyss, risk aversion of he invesors on fuel marke. The bigges influence of he recommendaion is visible wo days before he even (day 2 ) or, depending on he lenghs of esimaion window a he day of he even (day 0 ). Consrucing models wih he usage of shorer esimaion windows is increasing he significance level of abnormal reurn in he day 0. Table 2. Resuls in he group BUY T' N k T 30 19 e 0.453 0.135 0.202 0.044 0.099 AR 0.008 0.002 0.004 0.001 0.002 p-value 0.656 0.894 0.842 0.965 0.922 T 25 19 e 0.295 0.037 0.369 0.024 0.147 AR 0.005 0.001 0.006 0.000 0.002 p-value 0.771 0.971 0.716 0.981 0.885 T 20 24 e 0.159 0.041 0.359 0.12 0.145 AR 0.003 0.001 0.006 0.002 0.002 p-value 0.875 0.968 0.723 0.906 0.886 Source: own calculaions.

Influence of recommendaions announcemens on sock prices... 177 Table 3. Resuls in he group SELL T' N k T 30 19 e 0.327 0.816 0.526 0.339 0.027 AR 0.006 0.016 0.010 0.007 0.001 p-value 0.757 0.452 0.621 0.749 0.980 T 25 19 e 0.370 0.281 0.552 1.718 1.051 AR 0.021 0.016 0.032 0.100 0.061 p-value 0.724 0.788 0.601 0.137 0.334 T 20 24 e 0.067 0.452 0.169 0.182 0.227 AR 0.001 0.009 0.003 0.003 0.004 p-value 0.949 0.667 0.871 0.862 0.828 Source: own calculaions. In he second group he resuls are also insignifican as above, however, considerably higher influence on sock prices is visible in his group. The significance is he highes on he day afer he recommendaion (p-value = 0.137). Opimal is he window for he pre-even period of 25 days. Table 4. Resuls in he group KEEP T' N k T 30 19 e 0.036 0.186 0.024 0.188 0.046 AR 0.001 0.003 0.000 0.003 0.001 p-value 0.972 0.857 0.981 0.855 0.964 T 25 19 e 0.158 0.009 0.287 0.021 0.303 AR 0.003 0.000 0.005 0.000 0.005 p-value 0.877 0.993 0.779 0.984 0.768 T 20 24 e 0.343 0.327 0.262 0.253 0.253 AR 0.006 0.005 0.004 0.004 0.004 p-value 0.738 0.749 0.798 0.805 0.805 Source: own calculaions. There is also, like in previous groups, insignifican impac of public announcemens on sock prices. Resuls from his class are worse han in BUY and SELL ones. The significance is higher for shorer pre-even windows. Invesors do no pay a lo of aenion o his recommendaion.

178 Małgorzaa Szulc-Janek 5. Conclusions To sum up here is no ground on he basis of saisical es used for saing ha recommendaions of analyss have go a significan impac on changes refleced in sock prices (significance level for his es was 5%). The highes influence was visible in he group SELL. Resuls in his group showed ha impac is visible afer he even, however, i is no considerable impac. There is no influence of insider dealers. Announcemens o BUY socks is no being refleced in he prices. I can be, as saed before, he resul of risk aversion and low confidence o financial insiuions or can indicae ha he marke is so effecive ha every informaion available on he marke has been already discouned and is embedded in he prices. Moreover, shorer esimaion windows resul in beer reacion of recommendaions on he sock prices. The mos difficul analysis was choosing he significan evens. This also can be applied so as o improve he resuls and furher assess he significance of he process more accuraely by applying discree wavele ransform which, on he basis of variaion analysis, is reflecing changes in sock prices. Moreover, no every ime series fixed even windows are correc. When sample is oo small such misakes like over- or underesimaions canno be eliminaed. Also dynamic prices wihin he whole day, no only closing prices may be a good maerial for even analysis. This may also resul in increasing significance of he announcemens. Lieraure Gurgul H., Analiza zdarzeń na rynkach akcji, Oficyna Ekonomiczna, Kraków 2006. McWilliams A., Siegel D., Even sudies in managemen research: Theoreical and empirical issues, Academy of Managemen Journal 1997, Vol. 40. Szablewski A., Pniewski K., Barosiewicz B., Value Based Managemen, Deloie, Polex, Warszawa 2008. Tabak D., Dunbar F., Maerialiy and Magniude: Even Sudies in he Courroom, NERA Working Paper, 1999. ANALIZA WPŁYWU REKOMENDACJI ANALITYKÓW NA CENY AKCJI BRANŻY PALIWOWEJ Sreszczenie: Zbadano, czy zgodnie z eorią efekywności rekomendacje analiyków w branży paliwowej przedsawione w rzech grupach: kupuj, rzymaj, sprzedaj, wpływają na kursy akcji. Ceny i sopy zwrou wypadkowe informacji rynkowych i syuacji makroekonomicznej obarczone są mankamenami psychologicznymi. Inwesorzy, podejmując decyzje, opierają się nie ylko na informacjach fundamenalnych spółek, ale akże na subiekywnych

Influence of recommendaions announcemens on sock prices... 179 przesłankach. Teoria rynku frakalnego zakłada chaoyczność w mechanizmach rynkowych i brak wpływu rekomendacji analiyków na graczy giełdowych. Do analizy wpływu rekomendacji na ceny użyo zwyżkowej sopy zwrou obliczonej za pomocą modelu rynkowego Sharpa oraz esu isoności Sudena. Wyniki świadczą o małym wpływie rekomendacji analiyków na ceny akcji, co jes zgodne z badaniami przedsawionymi w lieraurze przedmiou.