Your browser does not support JavaScript!

Doctoral theses

Current Record: 5 of 31

Back to Results Previous page
Next page
Add to Basket
[Add to Basket]
Identifier 000455177
Title Modeling Time Series : the case of the Greek new-car sales sector
Alternative Title Υποδειγματοποίηση χρονολογικών σειρών : εμπειρική διερεύνηση της αγοράς αυτοκινήτων στην Ελλάδα
Author Βουλγαράκη, Μαρία
Thesis advisor Τσιώτας Γεώργιος
Reviewer Τσερκέζος Δικαίος
Τζίνιους Μαργαρίτα
Μάρκελλος Ραφαήλ
Κυρίκος Δημήτριος
Φλώρος Χρήστος
Βενέτης Ιωάννης
Abstract The purpose of this thesis is to study the time series modeling and make a comparative empirical analysis of the Greek new car-market, report and discuss the research results, and investigate whether time series methods can successfully be applied in marketing data and give reliable forecasts. Although the practice of forecasting marketing data, like sales, is a widely researched area, it hasn't been extensively applied for the Greek new-car sales sector. Therefore, this research is an attempt to report and discuss the new car sales forecasting practices concerning Greek companies in a turbulent economic time interval and a strictly supervised economic environment. The design of this empirical research application reports on sales forecast practices, using monthly data of the Greek new-car sales sector, available from the Association of motor vehicles importers representatives (AMVIR) of Greece. The monthly new car registration number is assumed, to be equal to the monthly new car sales level in the Greek market. The methodology of this applied study uses an in-sample and an out-of-sample time series modeling and forecasting, in a variety of new-car sales firms in the Greek retail market, testing various data sets in a time frame of the last two decades (1998 till 2016). Despite the difficult economic conditions in the Greek market, the researcher develops a comparative study, by modelling various time series and analysing them. Simple time series models are used, like the Mean or Average, the Naive and the Seasonal Naive models, but also some more sophisticated ones, like for example the Linear Models with Seasonal Dummies (LMSD), the Exponential Smoothing state space model (ETS), the Seasonal Autoregressive Integrated Moving Average (SARIMA), and the family of Seasonal Autoregressive Integrated Moving Average-General Autoregressive Conditional Heteroscedastic Model (SARIMA- GARCH). Furthermore, in addition to the original form of the variables and their log transformation, the use of the family of Box-Cox (1964) data transformation is applied in this research, where the transformation parameter lambda is determined by the method of Guerrero (1993). All these alternative approaches improve the quality and performance of the data, which ultimately proves that transforming variables provides a powerful tool for developing models. Lastly, the approach of combined forecast, as a forecasting tool, of various time series forecasting models is reported and discussed. This technique uses information from various individual forecasting methods, assuming either equal weights or optimally weighted forecasts, depending on the limitations, set by each type of combined forecasts studied. The empirical findings of this study gave evidence of an improvement in forecasting and confidence intervals, using the appropriate data transformation process. In addition, there is the prevailing conclusion that it is extremely difficult to find a single model, that can capture and forecast new car sales for all companies and at all times. Each car firm should be treated separately. However, research suggests that data transformation, the use time series models, and the combination of their predictions in the right way, prove to be beneficial in improving the accuracy of predictions for all cases of this research in the Greek car market. The conclusions of this thesis are very interesting cause the research takes place over a difficult time for the Greek economy. During the research period the country signed three (3) Memorandums of Understanding, on specific economic policy conditions, which lead to an economic supervision of the country, by a decision group referred as TROIKA, which was formed by the International Monetary Fund, the European Central Bank and the European Commission. Greek government was forced to implement austerity measures to its citizens, as a consequence of the supervision. Greek consumers, due to economic uncertainty, liquidity shortage, and bank crises, prolonged buying durable goods, like cars, which resulted in a sharp fall of new car sales in the Greek market.Finally, the originality of this research and the added value in economic science is that it is the first time, to my knowledge, that new car sales sector of the Greek market is treated as time series, with an extensive application of time series modeling, for the first two decades of the 21st century. Additionally, this thesis research contributes in depicting the best way for predicting sales, and helps improve the quality and accuracy of sales forecasts, hedge against risk, and reduce failure, for the Greek new-car decisions makers. It also emphasizes the problems, that evolved from the diminishing new car sales, due to the economic crisis in the Greek new car market.
Language English
Subject Exponential smoothing state space model (ETS)
Forecasting
Greek market
Linear model with seasonal dummies
Naive model
New car retail sector
Seasonal autoregressive integrated moving average (SARIMA)
Seasonal naive model
Time series
Αφελές
Γραμμικό με εποχιακές ψευδομεταβλητές
Εκθετικής εξομάλυνσης χώρου-χρόνου
Ελληνική αγορά
Εποχιακό αυτοπαλίνδρομο κινητού μέσου υπό συνθήκη ετεροσκεδαστικό
Εποχιακό αφελές
Μετασχηματισμός Box-Cox
Προβλέψεις
Πωλήσεις αυτοκινήτων
Συνδυασμός προβλέψεων
Χρονολογικές σειρές
Issue date 2020
Collection   School/Department--School of Social Sciences--Department of Economics--Doctoral theses
  Type of Work--Doctoral theses
Permanent Link https://elocus.lib.uoc.gr//dlib/3/1/5/metadata-dlib-1683182607-720116-28166.tkl Bookmark and Share
Views 282

Digital Documents
No preview available

Download document
View document
Views : 0