Econometric Analysis of Bulk Shipping: implications for investment strategies and financial decision-making

This thesis provides an econometric analysis of the bulk shipping markets and the implications for shipping investment and financial decision making. Chapter 1 sets the scene by providing a historic analysis of bulk shipping markets over the last 55 years. From this analysis, four shipping markets (freight, newbuilding, second-hand and demolition) are distinguished as well as a fifth one (ship finance) that acts as a facilitator to the other four. Also, with the help of correlation analysis, the factors influencing these markets are identified. The chapter then considers five critical interdependent forces (economic structure, ship supply and demand capital flows expressed by investor preferences and investment performance) that comprise the shipping market and move in cyclical patterns. This way, the chapter explains the role of the shipping cycle in devising investment strategies. Based on this analysis, Chapter 1 ends by defining the thesis aim and objectives. Chapter 2 presents the thesis methodology. It critically analyses the methods used in the collection of data and the interpretation of it, as well as the problems experienced while collecting it. The four subsequent chapters present the results from the analysis of the four shipping markets (freight, newbuilding, second-hand and demolition). Based on theory, Error Correction Models describing and quantifying the relationships between the variables are developed for all four markets. This way the thesis fills a gap in maritime economics literature by estimating models where none of the CLRM assumptions are violated. Consequently, statistical inferences from these models can be made safely. Furthermore, by disaggregating into the different ship types according to size, the thesis finds that different variables have different effects on each type, thus proving that each ship type has its own distinctive characteristics. Finally, chapters 4 to 7 compare different econometric methods, the theoretical Error Correction and the atheoretical family of Auto Regressive Moving Average (ARMA) models. It is found that theoretical models, are still to be preferred if one wants to achieve the classical objectives of Econometric Business Cycle Research simultaneously (to describe and forecast cycles and to evaluate policies and test economic theories). However, if not all goals have to be met with a single vehicle, other methods might serve the purpose equally well or even better as is the case with the Auto Regressive Moving Average method whose forecasts outperform those of the ECM method on many occasions. With respect to the period or time charter market, the thesis finds that spot rates are the major determinant for period ones. This indicates the validity of the pure expectation hypothesis of the term structure relationship between spot and period rates. However, this hypothesis is not always valid since on two occasions (Panamax bulk carriers and Aframax tankers), fleet changes, a variable incorporated in the model to depict market changes and risks, is found to be statistically significant. This leads to inconclusive evidence regarding the validity of the pure expectation hypothesis of the term structure relationship and needs further investigation. Finally, it is found that the forecasting ability of the models for the timecharter market is superior to that for the freight market. This can be attributed to the dominance of the stochastic component of the spot rates over the deterministic one, which makes their accurate forecasting a very difficult task. As far as the econometric analysis of new vessel prices is concerned, shipbuilding costs are found to have the most significant effect on the determination of newbuilding prices for all ship types. Time charter rates have an effect only on few ship segments. This is in line with theory that newbuilding prices are cost driven rather than market driven as second-hand ship prices are. It is also found that actual exchange rates do not affect shipbuilding prices but cost variations due to exchange rate fluctuations do. Orderbook as a percentage of the fleet, used as a proxy for shipyard capacity due to data discrepancies and lack of long enough time series for the later, is found significant only for tankers indicating that shipyards' expansion policy is aimed at high value ships like tankers rather than bulk carriers. Finally, newbuilding prices for some ship types may be driven to a certain extent by asset pricing and speculation. Newbuilding and timecharter rates have the greatest effect of all variables on the determination of second-hand prices, in most cases both in the short and the long run. The cost of capital is only significant for bulk carrier owners. The only exception is the Suezmax segment. Suezmax prices have been closely pegged to VLCCs with a discount but not proportional to the size. In other words Suezmaxes are rarely bargain vessels. Second-hand prices of Suezmaxes could therefore be more closely tied to newbuilding prices than for other tanker sizes. Finally, orderbook as a percentage of the fleet has a negative effect on the prices of second-hand vessels only in the long run and only in large and Panamax tankers. The thesis also finds that demolition prices are primarily driven by market conditions and expectations. In addition, the price of scrap steel is also found to have a significant effect on VLCCs due to increasing demand for scrap steel that makes demolition traders eager to offer higher prices for larger tankers to satisfy demand. Finally, it is found that the volume of scrapped ships has a negative effect on the demolition price of medium and large tankers. This is due to new legislation usually stemming from an accident or environmental campaigns that may force shipowners to scrap their ships earlier than they had originally anticipated. Shipowners invest in different ship markets and vessel sizes in the expectation of achieving a reduction in risk via the resulting diversification in their income. However, it is frequently observed that shipping companies focusing on a particular ship type achieve equally good or even better risk levels than those investing in various ship types and size. A question therefore arises whether such diversification strategies really reduce the investor's risk. This thesis investigates the potential of risk reduction benefits for a bulk shipping investor through diversification. It first analyses the traditional risk reduction approach of calculating the variance and the standard deviation of a portfolio. Results show that risk reduction benefits are achieved through diversification. Then, the thesis builds upon the shortcomings of correlation, namely the fact that while markets may tend to diverge considerably in the short-run, like periods of up to a year, they may actually be integrated over longer periods. If the income, expressed in time charter equivalent rates, of the various ship types or sizes is very strongly correlated in the long run, diversification will be less effective than if the ship markets or segments operated independently of one another. An important indication of the degree to which long run diversification is available to shipping investors is given by determining whether the markets are cointegrated. The thesis employs the Johansen method on 247 different combinations of investment in the dry and wet bulk markets. It finds that investing in more than one type of bulk carrier nullifies any risk reduction benefits. Furthermore, risk reduction benefits decrease as diversification increases with no risk reduction benefits obtained when investment involves more than five different ship types/sizes. These results initially seem to be in contrast with portfolio theory which claims that risk reduction benefits increase with higher diversification. However, the reason behind this difference lies with time horizon and the results actually supplement the theory of risk reduction through diversification by showing that the benefits of diversification in most cases may exist in the short run but disappear in the long run. Finally, finding such long run relationships supports the existence of inefficiencies in the shipping freight market, a finding in line with previous research. This thesis also develops an analytical tool for shipmanagers to measure the possible losses, within a specified time horizon and confidence interval, of their portfolios by calculating Value at Risk with the variance-covariance, historical and Monte Carlo simulation methods. This tool allows managers to identify the extent to which each asset contributes to these possible losses, as well as get an idea of the maximum possible losses should the worst case scenario occur. Based on this framework, shipping firms can then make informed decisions about maintaining or expanding lines of business, or whether to hedge financial risks at the firm level. In our example both VaR and CVaR figures are not high enough for the managers to seek hedging of their positions. The thesis argues that even if managers decide to hedge, it makes more sense to enter some of the ships into period charters, if they can, rather than use freight derivatives due to the high transaction and brokerage costs associated with the latter. However, the purpose of the example is neither to show that shipping is low risk nor disregard the use of freight derivatives as hedging tools but rather to promote VaR and CVaR as essential tools for risk measurement and hedging strategy determination in day-to-day shipping operations. In other words, the thesis argues that shipowners shall first try to determine their level of risk-exposure, then decide whether or not this exposure is acceptable to them and then employ different risk management tools to minimise such exposures. Nevertheless, the thesis argues that VaR-CVaR calculations should only be considered as a first order approximation and users should not be lulled into a state of complacency but rather recognise its limitations. Finally, this thesis introduces Real Option Analysis and exotic options in particular as an alternative to the traditional capital budgeting technique for evaluating a series of shipping projects. The thesis considers the option to expand, timing and defer options, the option to choose the best of two assets and the option to vary the firm's production methods. Compound option are used to value the expansion option through ordering an additional number of ships at a predetermined price, showing that such options may increase the shareholders' value substantially. Also a framework to critically assess asset play opportunities is developed. Furthermore, by evaluating investment opportunities using American Exchange Options, substantial differences are found compared to the NPV method in both the value of the investment opportunities and the timing of when the project is undertaken. Chooser options are employed to evaluate the various options open to a shipowner in order to optimise strategic decision making. Finally, Exchange options are used to value the decision to invest in a new ship type. Overall, Real Options are useful tools