TTK Links: Home Page | TTK Moodle | ÕIS
TTK/UAS Repository For Digital Teaching and Learning Resources

Aktsiate tootluse prognoosimine regressioonanalüüsi abil

Rosenfeldt, Märt (2020) Aktsiate tootluse prognoosimine regressioonanalüüsi abil. [thesis] [en] Forecasting stock market returns using regression analysis.

[img] PDF - Published Version
Restricted to Registered users only

Download (720kB)
[img] Microsoft Word - Published Version
Restricted to Registered users only

Download (373kB)

Abstract

Lõputöö oli koostatud teemal „Aktsiate tootluse prognoosimine regressioonanalüüsi abil“. Lõputöö eesmärk on anda ülevaade investeeringutest ning selgitada aktsiahinna tegureid ja koostada regressioonanalüüsi põhjal mudel aktsiaturu tootluse prognoosimiseks Investeerimist mõistetakse kui rahaliste vahendite paigutamist mingisugusesse varaklassi kasu saamiseks või rikkuse hoidmise jaoks. Investeerimise eesmärk on teenida kasu kapitalilt ehk investeeringu väärtuse tõusust ja/või teenida jooksvat intressitulu. Investeerimisvahendeid võib jaotada kaheks: kõrgema tootluse ja riskiga varaklassid ja madala tootluse ja riskiga investeerimisvahendid. Lõputöös keskendus autor aktsiatele, mis on kõrgema tootlusega. Teoreetilisest analüüsist selgus, et pikaajaliselt rikkuse kasvatamiseks on parim viis aktsiatesse investeerimine. Tehnoloogia arengu tagajärjel on maailm jõudnud selleni, et igal ühel on võimalik investeerida oma raha endale sobivasse varaklassi. Üheks populaarseimaks investeerimisinstrumendiks igapäeva investorile on olnud aktsiad, sest nende soetamine ei nõua rohkelt kapitali, samuti on aktsiad üks likviidsemaid investeerimisvahendeid ning tegemist on pigem passiivse varaklassiga. Uuringu käigus jõudis autor järeldusele, et enamikul muutujatest on aastase vaatlusperioodi tootluse prognoosimise jaoks tagasihoidlik jõud kogu vaatlusperioodi jooksul. See on kooskõlas varasemalt koostatud ja töös kajastatud Goyal ja Welchi (2008) uuringuga. Lõputöös pakub autor välja alternatiivse meetodi aktsiaturu tulude prognoosimiseks. Selleks on osade summa meetod, nagu on välja toonud Ferreira ja Santa-Clara (2008). Modelleerides eraldi aktsiaturu tootluse komponente, on valimivälise regressioonanalüüsi tulemused oluliselt paremad. Märkimisväärne osa paranemisest tuleneb pelgalt dividenditootluse ja kasumi kasvu komponentidest. Lõputöö eesmärk on täidetud ja on selgitatud investeerimisalaseid mõisteid ning erineva tootluse ja riskiga varaklasse. Samuti on lõputöös selgitatud aktsia hinna konsensuslik hinna kujunemine ning kirjeldatud aktsia hinda mõjutavaid tegureid. Rakendusliku tööna on koostatud regressioonanalüüsi põhjal eksperimentaalmudel aktsiaturu hinna prognoosimiseks 16 teguri põhjal. Empiirilise uuringu tulemusti analüüsides võib öelda, et enamike autori poolt valitud tunnuste alustel on raske genereerida usaldusväärseid valimisiseseid prognoose aktsiapreemia kohta. 16 muutuja hulgas on neli tegurit, mis andsid positiivse valimivälise R-ruudu väärtuse. Tulemustest võib järeldada, et loobudes traditsioonilisest regressioonist aktsiate tootluse prognoosimisel ja keskendudes osade summa meetodile, mis hõlmab dividenditootlust, S&P 500 aktsiaindeksi ettevõtete kasumi kasvutempot ja aktsia turuhinna-tulu suhtarvu kasvutempot, toob see investoritele statistiliselt ja majanduslikult olulist kasu.

Abstract [en]

The final thesis was written on the topic “Forecasting stock market returns using regression analysis”. The aim of this thesis is to provide an overview of investments and to explain the factors of stock price and to create a model for stock market performance forecasting based on regression analysis. Investment is defined as placing funds in a class of assets for the purpose of gaining wealth or maintaining wealth. The purpose of the investment is to earn a capital gain, ie an increase in the value of the investment, and / or to earn current interest income. Investment instruments can be divided into two categories: asset classes with a higher return and risk and investment vehicles with a low rate of return and risk. In the thesis, the author focused on stocks that have higher returns. Theoretical analysis shows that investing in equities is the best way to grow wealth in the long run. As a result of technological advancements, the world has reached the point where anyone can invest their money in a suitable asset class. One of the most popular investment instruments for everyday investors has been equities, as their acquisition does not require a lot of capital, equities are also one of the most liquid means of investment and are rather a passive asset class. Theoretical analysis shows that investing in equities is the best way to grow wealth in the long run. As a result of technological advancements, the world has reached the point where anyone can invest their money in a suitable asset class. One of the most popular investment instruments for everyday investors has been equities, as their acquisition does not require a lot of capital, equities are also one of the most liquid means of investment and are rather a passive asset class. During the study, the author concluded that most variables have modest power over the entire observation period for predicting annual performance over the observation period. This is in line with a previously compiled and documented study by Goyal and Welch (2008). In the thesis, the author proposes an alternative method for forecasting stock market returns. This is the sum of parts method, as outlined by Ferreira and Santa-Clara (2008). By modeling separate components of stock market performance, the results of the non-sample regression analysis are significantly better. A significant part of the improvement is due solely to the components of dividend yield and earnings growth. The objective of the thesis is fulfilled and the concepts of investment and asset classes with different rates of return and risk are explained. The thesis also explains the consensus price formation of the share price and describes the factors influencing the share price. An experimental model for estimating stock market price based on 16 factors has been developed based on regression analysis. Analyzing the results of empirical research, it can be said that it is difficult to generate reliable in-election predictions for stock premium based on most of the traits selected by the author. Among the 16 variables, there are four factors that yielded a positive out-of-sample R-squared value. The results suggest that by abandoning the traditional regression in stock return forecasting and focusing on the share sum method, which includes dividend yield, S&P 500 stock index corporate earnings growth rate and stock market price to earnings ratio, it brings about statistically and economically significant benefits for investors.

Item Type: thesis
Advisor: Diana Tandru
Subjects: Economy and Management > Micro and Macroeconomics
Economy and Management > Economic Analysis
Divisions: Service Economy Institute > Business management
Depositing User: Märt Rosenfeldt
Date Deposited: 31 Jan 2020 09:02
Last Modified: 31 Jan 2020 09:02
URI: http://eprints.tktk.ee/id/eprint/4751

Actions (login required)

View Item View Item