Otomobil sahipliğinin modellenmesi

Abstract

Bu çalışmada ulaşım planlamasının anahtar elemanlarından biri olan otomobil sahipliğinin öngörülmesi sorunu incelenmiştir. Türkiye genelinde ve İstanbul'da otomobil sahipliğinin gelişimi için bir model ortaya konmuştur. İlk bölümde, otomobil sahipliğinin tanımı ve önemi vurgulanmış, ulaşım planlamasındaki rolüne değinilmiş, gelişmiş ve gelişmekte olan ülkelerdeki farklılıklar ele alınmış ve bir ulaşım politikası aracı olarak otomobil sahipliği sorunu ortaya konmuştur. İkinci bölümde otomobil sahipliğinin modellenmesinde kullanılan başlıca yöntemler incelenmiştir. Üçüncü bölümde Türkiye gibi gelişmekte olan ülkelerde modellemede karşılaşılan sorunlar incelenmiştir. Ayrıca otomotiv sektörünün dünyada ve Türkiye'deki durumu da incelenmiştir. Daha sonra bu çalışmada kullanılan veriler ve geliştirilen model açıklanmıştır. Dördüncü bölümde geliştirilen model kullanılarak 2005 yılında Türkiye ve İstanbul'daki otomobil sahipliği değerleri öngörülmüştür. Modelden elde edilen sonuçlar diğer çalışmalardaki öngörülerle karşılaştırılmış ve otomobil sahipliğinin modellenmesine ilişkin yapılması gereken çalışmalar konusunda öneriler getirmiştir.

The main objective of this studH is to develop a model for the prediction of car ownership in Turken an<* *n Istanbul Car ownership forecasting plays an important role in both traditional and innovative transport planning. It is often recognised that the level of car ownership in a city or a country is one of the keg factors influencing/ the level of demand for transport facilities including roads, junctions and parking spaces. An increase in car travel contributes more to congestion» and, hence to demand for additional road space, than the same increase in demand for puplic transport. Transport planning is often seen as the processof developing transport policies and invesment programmes to cope with future demand as estimated from computer models. Essential inputs to these models are the expected values in the future of key planning veriables like population, employment and car ownership. These are usually undertaken by the planning team. In anycase, these planning variables are considered as externally defined inputs to the plann ing process. It is not diffucult to see why in developed countries, car ownership is seen as a planning variable that must be forecast before the traditional stages of transport modell ing are undertaken. It is accepted that as incomes increases so will car ownership, as anybody who is able to purchose a car will buy one. The issue is one of forecasting rather than controlling car ownership growth. Most developedcountries have their own car industry and restricting car ownership is seen as detrimental to its development or survival. Moreover given the levels of car ownership in these countries, restricting the purchase of new car's would be seen as unequitable to those just reaching the income levels required to purchase and run a motor vehicle. One of the main transport VII planning issues is to satisfa the demand for more mobility, marshalling resources to cater for additional car travel and implementing policies supporting public transport in the hope of providing a better service and reducing the needfor extra road space. The situation is rather different in developing countries as recognised, for example, bg the World Bank. In developing countries higher - income people can afford and use a car which is not onla a mode of transport, but also a status snnibol. In mana developing countries like Turkey car ownership is growing vera fast, albeit from a vera l°w level. However, this growth cannot be explained ba increasing affluence alone. There are social and cultural factors contributing to it and in mana cases the alternative modes of transport offer a vera poor service. In mana developing countries public transport operators offer a low level of service with old, unsafe, overcrowded buses. This forces mana people to bua a car even if often thea can hardla afford it. This motorisation is alreada creating serious congestion problems in mana large cities, thus generating a demand for addi tionalf acili ties and roadspace. Accidents and enviromental problems are onla some of the side effects of thisgrowth.buttheaare becoming critical in several major cities like Istanbul. Although some developing countries have a local car industra» this is often small and relies on the import of parts and components. Therefore, car ownership growth often makes heavy demands to the balance of payments of a count ra- Scarce foreign currency which could otherwise be spent in the import of mechanising agriculture or in improving health services thus devoted to the import of cars and their parts. Some of the main issues inthechoiceof atransport modelling approach which is appropriate to developing countries are as follows: a) The cost of conventional planning methods: Traditional transport planning techniques are vera demanding in terms of data, technical skills, computer resources and time. All of these are more expensive in developing countries than in Western Europe. An appropriate planning tool should make good use of avaible resources and make modest demands to already over stretched VTII government funds. b) Limitations of data: Good, reliable data are a scarce resource in developing countries. Sometimes even, basxc data like population and income are missing, unreliable or out-of date. Although lowerlabourcostsmake data cheaper, there are higher costs of training and supervision. Moreover, even if good data are collected the high rate of change in manH developing countries would make them obsolete much quicker than in developed ones. c) The nature of the problems is so different from those in the developed countries that new models are required. This is diffucult to achieve. d) Scarcity of technical resources: The personnel with the right qualifications to use conventional methods are relatively scarce in most developing countries, and they may well be better employed in other areas like maintenance, operational and financial management. e) Communication problems: All transport models are a partial and simplified representation of reality and their appropriate use requires an understanding of their power and limitations on the part of the decision-maker, normally a politician. The complex and heavy nature of conventional modelling techiques may alienate decisions -makers. All these issues have been used to suggest that a more appropriate planning approach would de-emphasise large-scale transport modelling and adopt simpler techniques better adapted to local conditions and easier to match with a continuous planning process. It has been a major problem during this study to find reliable data. This fact played an important role in determining parameters needed in this car ownership modelling The parameters which have been used here are as follows: 1 ) GNP/per capi ta 2) Domestic Car Production 3) Domestic Car Prices 4> Car Imports B) Average fuel prices IX The first four parameters of this model for Turken have been taken from Tofaş Data Bank. They had used fixed prices of 1968 to obtain these parameters. Therfore, prices have been transformed to 1990 fixed prices as follows: Deflator = Current Prices / Fixed Prices (Index)i - (Deflator)i / (Deflator)90 (Fixed Prices )90i = Current Prices / (Index) i The index which was obtained in that waH was used to transform the domestic car prices and the average fuel prices into 1990 fixed prices. Fuel prices have been found by supposing that the fuel consumption consisted of "A 60 Super and %40 Regular gas. Same parameters have been used in the model for Istanbul. The value of GNP/Per Capita has been taken from a study (during 1980-86) undertaken by the Chamber of Commerce of Istanbul The missing values between 1987-90 have been obtained the fact that the GNP/Per Capita of Istanbul is about %23 of the Total GNP/Per Capita Turkey. Prices of domestic cars are assumed to be same in Istanbul and in Turkey. The number of imports and domestic cars represent the number of cars registered as private cars in a particular year. Car Ownership per 1000 population c, is modelled by using a regression equation as follows: b. C = a rç X1 i where X is i th variable used to predict the car ownership, and a and b are calibration coefficients. Log-Linear equations are used to calibrate the model. Ln C = Ln a + *? i b; Ln X; Althought all other values started from 1948» reliable data about the fuel prices could onla be found starting from 1970. Because of the insignif acang of the regression coefficients (for example: the price of domestic cars is positive (+) which means as the prices get higher the number car ownership rises too) a model has been developed bg a new regression bg using 1980-1990 data Pinal results have been compared with the real car ownership values and the percentage of error has been calculated. The surveys about the car owership in Turkeg forecosts a 100- 145 thosand percent value. This ratio is for below the global average. The surveg also predicts the fact that global value could not be reoched even in 2005. The survea puts the maximum car ownership ratio for Istanbul as 65 thosand percent.