This thesis concerns econometric time series modelling of cointegrated multivariate systems using a Bayesian approach. The Bayesian approach has become increasingly attractive among researchers in the fields such as biology, though still only a relatively few econometricians use these techniques. Rather than theoretical aspects of Bayesian statistics or computational techniques, we illustrate how the Bayesian methods can be useful in analysing non-linear cointegration models. In the last ten years, non-linear time series models, such as regime switching models, have become popular among applied econometricians to analyse the business cycles, policy evaluation in specific macroeconomic issues and forecasting. Cointegration analysis has been influenced by the non-linearity so that cointegration models that allow regime switching or structural breaks have been analysed by many econometricians. Unfortunately, these nonlinear cointegration models tend to be complicated both in terms of estimation and testing. We consider in this thesis a Bayesian approach to (i) a linear cointegration model, (ii) a cointegration model with Markov regime switching, and (iii) a cointegration model with multiple structural breaks, and show how easily we can analyse these models without any substantial modification. Chapter 2 proposes a simple method for detecting cointegration rank using the Bayese factors, computed by the harmonic mean of the likelihood or Schwarz' Bayesian information criterion. Then we perform Monte Carlo simulations to compare three Bayesian methods (Phillips posterior information criterion, Kleibergen and Paap method, and one proposed method) for the cointegration rank. Provided we have enough large sample size, the Phillips' posterior information criterion gives consistent results, while the results by Kleibergen and Paap method depends on the prior hyperparameters that we specify. In Chapter 3, we develop the cointegration model that allows cointegration relationships to be switched on and off depending on the regime. Unlike the classical method that requires a two-step estimation, the Bayesian method provide a straightforward estimation and testing procedure. In Chapter 4, we consider cointegration model with multiple structural breaks in the level, trend and error covariance. The more general model with breaks in both the adjustment term and the cointegrating vectors are also presented. To date, there is no research that deals with a cointegration model with unknown multiple structural breaks in any subset of the parameters.

The thesis consists of three chapters of self-contained empirical studies. In Chapter 1, we examine long-run and short-run dynamics of US real trade balance with Canada. In addition to the linear error-correction model, the Markov-switching error-correction model is employed, using quarterly data from 1985 to 2008. We find that real exchange rate, real oil price and real new housing price index have statistically significant effects on US real trade balance with Canada in the long run. We acquire evidence of short-run J-curve. Results show that short-run dynamic effects of real oil price are not so fearful, with statistically insignificant effect on real trade balance following an increase in real oil price. House prices could be argued as being strongly relevant for settlement and adjustment of US trade balance in the long run through the wealth effects. However, the immediate (next-quarter) effect of a change in housing wealth is insignificant, consistent with existing literature. US real trade balance with Canada forecasts from our non-linear VAR model outperform ones from the linear VAR in first difference (DVAR) model and ones from the random walk model. The long-term out-of-sample forecastability is not much improved by the oil price and house price variables, which, nonetheless, actively explain in-sample movement of US real trade balance with Canada in the long run. In Chapter 2, we examine the effect of monetary policy and exchange rate on stock price movements in Asia. We employ a Bayesian structural vector autoregression model and impose sign restrictions to identify simultaneously and uniquely contractionary monetary policy shocks and exchange rate depreciation shocks in an integrated framework. This study covers the stock markets of Thailand, Malaysia and South Korea, over the period 1989-2008. Our main results acquired using sign restrictions show that monetary policy shocks result in a strongly persistent effect on market index real stock prices whereas the impact of exchange rate shocks is short-lived over the short run. The variance decomposition suggests that the exchange rate is as important as monetary policy for explaining the dynamics of market and financial sector index real stock prices. More precisely, for all the countries examined, real exchange rate developments have been more important in the short run. In Chapter 3, within the context of a time-varying transition probabilities Markov-switching model of the uncovered interest parity (UIP) condition, we examine if variables measuring fear and volatility have an effect on the probability of switching between the regime where the UIP condition holds and the regime where it does not. The state transition probability depends nonlinearly upon the variables examined. These are the exchange rate volatility, the VIX equity option implied volatility index and the TED spread. Applying this to both US dollar exchange rates and cross (exchange) rates from January 4, 1990 to September 11, 2008, we find that those three variables increase the probability of remaining in the regime where the UIP condition holds. In addition, the probability of switching from the regime where the UIP condition does not hold to the regime where the exchange rate follows the UIP condition decreases as these variables measuring fear and volatility fall, especially the VIX equity option implied volatility index. The smoothed probabilities show that exchange rates examined essentially do not follow the UIP condition except during periods in which the fear and risk variables are increasing, as in the recent global financial crisis in particular.

In most economic models used for theoretical exploration or policy analysis, there is a crucial role for agents' expectations about future outcomes. Generally, it is assumed that economic agents take their decisions according to rationality principles and that they have a fairly accurate knowledge about the economic environment. In other words, they are assumed to know the model of the economy (Rational Expectations Hypothesis). The latter assumption is somewhat extreme, given the evident lack of agreement, even among professionals, about the correct model of the economy. In this thesis I maintain the hypothesis that agents take their decisions rationally, i. e. in order to maximize their utilities given their budget constraints, but I assume that each agent has to learn about the economic environment. More specifically, I consider economic models for monetary policy analysis. The goal is to study how the introduction of learning in these models can affect the design of monetary policy. Policy recommendations that might be sound under Rational Expectations, might lead to disastrous results under learning. I also use learning as a selection device. Some economic models fail to predict a unique Rational Expectations Equilibrium. Nevertheless, a REE is a sensible prediction of the model only if it can be shown that it is the result of some learning process of the economic agents. REE that are unstable under learning are not plausible equilibria.

The first part of this thesis is concerned with tax competition when the tax receipts fund an anti-crime measure. Both the capital and criminals are mobile between two jurisdictions. The resulting pure strategy Nash equilibrium tax rates are distorted from the optimal tax by the equilibrium migration response of the rich; if positive at the equilibrium then tax competition will result in taxes that are too high whilst if it is negative taxes will be too low compared to the optimum. The best response functions of the model are tested using data from England and Wales. The possibility that they engage in tax competition cannot be ruled out. It is possible for a central government to devolve tax raising powers without the distortion occurring if they can impose an optimal sanction. This, though, is independent of the harm caused by the crime and could be politically difficult to introduce. The second part looks at the Ministry of Defence’s procurement policy since 1985. The role of competition has increased but scant attention was played to the trade-off between maximising the benefits of current competition and obtaining future competition. The Ministry of Defence always chose to take the benefits in the short term arguing any loss of competition merely eliminated excess capacity which the Ministry of Defence would no longer have to pay for. Whilst the empirics suggest this is true during the 1990s, the problems encountered on the Type 45 project at the start of the millennium demonstrate the difficulties they have in procuring given the limited number of domestic firms they can contract with. An alternative mechanism of directed buys, with recourse to a competitive market off the equilibrium path, is suggested as a way in which the Ministry of Defence can preserve competition into the future.

Chapter 2 of this thesis employs a dynamic general equilibrium with Taylor wage contracts to show that the use of strict inflation targeting as a disinflation policy may result in a slump in production and a considerable increase in macroeconomic volatility. Important determinants of the magnitude of the macroeconomic oscillations in the post-disinflation state of the economy turn out to be the size of the reduction in the inflation rate and the degree of returns to labor in the production function. In the special case of constant returns, the oscillations are large and permanent. Chapter 3 of this thesis extends the above analysis to an open-economy setting demonstrating that the exchange rate can act as a stabilizer by effectively relieving wages from part of the burden of reducing the inflation rate. The more the economy is open, the smaller the magnitude of the macroeconomic oscillations will be after the disinflation policy is applied. The policy is shown to be infeasible for all practical purposes in a closed economy with constant returns to scale when the full nonlinear model is considered. Chapter 4 of this thesis employs a Markov switching framework to allow for an interesting alternative characterization of macroeconomic news effects on the foreign exchange market. The chapter finds strong evidence for the presence of nonlinear regime switching between a high-volatility and a low volatility state driven by monetary policy announcements that come as a surprise to the market. It also uncovers significant market positioning prior to the announcements, indicating a limiting of risk exposure by market participants who are unsure about the precise outcome of the policy decisions. Chapter 5 of this thesis investigates the impact of monetary shocks on the direction and the composition of international capital flows. It identifies monetary policy shocks in a structural VAR via the pure sign restrictions approach. There are two key findings. First, a US monetary easing causes net capital inflows and a worsening of the US trade balance. Second, monetary policy shocks induce a negative conditional correlation between capital flows in bonds and equity securities. Intriguingly, they cause a negative conditional correlation between equity flows and equity returns but a positive conditional correlation between bond flows and bond returns.

It is well known that the market price of options are inconsistent with option pricing models that assume the innovation of the underlying price follows Geometric Brownian Motion. What is not clear is why this occurs. The testing of option pricing models requires a joint hypothesis to be tested that the options pricing models are correct and that the options markets are efficient. To test the option pricing models, we will examine the relationship between the objective and risk neutral dispersion processes for twelve financial futures markets. These markets have been selected so that we can investigate the dynamics of equity, fixed income and foreign exchange asset classes. Our analysis of the objective dispersion processes allows us to reject the hypothesis that the prices of these twelve markets follow Geometric Brownian motion. For all twelve markets and for various sub-periods of analysis, we find that the optimal models for capturing the dynamics of the objective dispersion process include jump diffusion and stochastic volatility. For the risk neutral dispersion process, we chose to examine the implied volatility surfaces from the closing prices of options (on these same futures markets). It appears that both within and between markets similar dynamics determine the shapes of the implied volatility surfaces. By employing an Analysis of Covariance approach, we found that important consistencies exist within asset classes and between markets. The first order strike price effect (skewness) differs widely among markets but is fairly consistent within the same asset class. The second order strike price effect (kurtosis) is consistent among all markets. The dependency of both strike price effects on the time to expiration is also similar across all markets and suggests that both jump diffusion and stochastic volatility play a role. A comparison of option prices, which are consistent with the objective dispersion process, with actual options prices suggests that significant divergences exist. The actual smile patterns display greater variation in the amplitude compared to those from the objective function. The least degree of discrepancy exists for the foreign exchange markets. Both the stock index and fixed income options dispersion processes display behaviours that diverge considerably. This divergence is primarily due to the existence of negative skews that are not justified by the objective dispersion processes. This suggests that other mechanisms are at work for the risk neutral dispersion processes for these asset classes. The most likely explanations are the existence of risk premia associated with stochastic volatility and non-diversifiable jumps or that transaction costs are relevant.

The thesis aims to explain the determination of employment at the industry level - in particular, the S.I.C. Orders of the U.K. engineering industry. The traditional demand-orientated approach is examined theoretically and empirically. Many developments are made to the models, but implausible and unstable estimates are generally found. More major developments are attempted, modelling desired output and the relationship between investment and employment with some success, but without a generally acceptable model of aggregate employment emerging. The view is taken that a major reason for this is likely to be the neglect of supply factors. Initial attempts to allow for the tightness of the labour market indicate some effect though incorrectly specified. The second half of the thesis undertakes a more rigorous and original analysis, involving the specification of an industry labour supply function, to be analysed in conjunction with the demand function. The appropriate methods of analysis and estimation depend upon assumptions about the interaction of demand and supply and the role of wages. Three stages of development are considered with increasing realism of assumptions, but increasing complexity of analysis and difficulty of estimation. The first assumes flexibility of wages, equilibrating sectors of the labour market and enabling simultaneous estimation of aggregate demand and supply. The second assumes a degree of inflexibility of wages, but homogeneity of the sectors, so that aggregate demand or supply is observed and 'regime' estimation is possible. The third stage allows for non-homogeneity-of sectors so that neither aggregate demand nor supply may be observed. Constrained estimation, via programming methods, results. Exogenous data is used to assess the extent of excess demand and supply in the labour market. Whilst the empirical results are limited, they do indicate the need for supply factors to be modelled and included in the analysis of employment.

This thesis focuses on the liquidity risk and its impact on bond prices of the international markets and comprises three self-contained research papers. In the first research paper, we examine the role of the liquidity in the pricing of sovereign U.S. dollar bonds in emerging markets. We extend Acharya and Pedersen’s (2005) liquidity-adjusted capital asset pricing model to the bond market and find that both liquidity level and multiple liquidity risks are priced factors for the expected excess return of U.S. dollar bonds issued by developing countries. The combined effects of liquidity risk and liquidity level can explain as much as 1% per annum extra yield spread for the countries that have higher liquidity betas. Countries, which have a high correlation with the global market or U.S. stock market, have higher required bond returns than low correlation countries. The liquidity factor helps explain the credit spread puzzle of high yields. Our empirical results also support a flight to liquidity across the studied countries and are robust after controlling for bond characteristics and the U.S. risk factors. The second research paper finds that both liquidity level and liquidity risk are important in explaining the cross-section of domestic government bond returns in 39 countries (both emerging and developed) around the world. After controlling for other market factors and bond characteristics, liquidity level and liquidity risk together can explain as much as 0.41% per annum of extra yield for the highest versus the lowest liquidity risk countries, which are China and Argentina respectively. There is also an evidence of liquidity spillovers from the U.S. equity market to domestic bond markets around the world. Employing a conditional model, which allows both time-series and crosssectional variations in liquidity betas, we find that the impact of liquidity risk is time varying across two different regimes: it increases in times of high uncertainty and is always larger in emerging than in developed countries. Nevertheless, the price of risk or premium required by investors for holding this time-varying risk is relatively modest. The third research paper examines whether liquidity spillovers between sovereign bonds are systematic or idiosyncratic in character. A theoretical model is developed, which demonstrates that idiosyncratic spillovers require returns to be correlated, whereas systematic spillovers require volatilities to be correlated. We apply the model to sovereign bonds in 35 emerging markets, aggregated for some analyses into Asian, European and Latin American regions. We find liquidity spillovers mainly from Latin America to the other regions and they are both systematic and idiosyncratic in character. Further cross-sectional analysis (by country) and time-series analysis (by region) show that systematic spillovers are more important than idiosyncratic spillovers. The conclusion is that most liquidity risk across emerging bond markets is systematic and therefore cannot easily be hedged away. This has important implications for portfolio selection by fund managers and for the regulation of systemic risk.

Under the pressure of the population explosion amconstitutional changes, social and economic conditions in Mauritius went through a process of rapid change after the Second World War. The introduction of universal suffrage unseated the white plantocracy and brought to power the representatives of the urban and rural working classes. This led to fundamen~ changes in the Government social and economic policy. In 1963, Government fixed the wages and conditions of employment of agricultural and non-agricultural workers in the sugar industry. The first chapter describes the structure of the industry and the organisation of cane production on sugar estates. We study the pattern of land owPership characterised by the high concentration of cane land on millers' estates. The distribution of land will influence the rate of production and the allocation of resources inside the industry and the demand for and supply of labour on estates. In the second chapter we study the Island's social system and its influence on Government labour policy. The forces which led to the Government intervention in the labour market originated in the social structure of the Mauritius plantation economy with its rigid divisions. This accounts for the underdevelopment of industrial relations which prompted direct Government intervention in labour matters. The third chapter concentrates on the impact of Government regulations on the level of wages in the economy. The first part focusses on the importance of the sugar industry as a wage leader. We then study the effect of Government policy on the structure of wages and finally its influence on the distribution of employment. The analysis of the macro-economic structure forms an essential part of the work. The fourth chapter comprises a detailed quantitative description of the structure of the Island economy, which provides a convenient framework for the study of the effects of wage increases on the Government finances and the Balance of Payments. The fifth chapter is divided into two parts. In the first part we analyse the labour conditions on millers' estates through an investigation of the conditions of demand for and supply of labour. The second part provides an estimate of the marginal productivity of field labour during the crop and intercrop by means of production functions fitted to 1965 and 1973 data, in order to assess the effect of Government labour policy on the productivity of labour on estates.

Chapter 1 introduces the thesis providing an overview of the common themes and methods underlying this research. Chapter 2 reports an experiment that examines the characteristics of effective leaders in a leader-follower voluntary contributions game. We focus on two factors: leaders’ cooperativeness and their beliefs about followers’cooperativeness. We find that groups perform best when led by cooperatively inclined leaders, partly because they are intrinsically motivated to contribute more than non-cooperative leaders, partly because they are more optimistic about followers’ cooperativeness. Chapter 3 reports an experiment comparing sequential and simultaneous contributions to a public good in a quasi-linear two-person setting. As predicted, we find that overall provision may be lower under sequential than simultaneous contributions. However, we also find that the distribution of contributions is more equitable than predicted when the first-mover is predicted to free-ride, but not when the second-mover is predicted to free-ride. These results can be explained by second-movers’ willingness to punish free-riders, and unwillingness to reward first-movers who contribute. Chapter 4 investigates the impact of social comparisons on reciprocal relationships. Using a three-person gift- exchange game we study how employees’ reciprocity towards an employer is affected by pay comparison information (information about what co-workers earn) and effort comparison information (information about how co-workers perform). We find that pay comparison information does not affect reciprocity, while effort comparison information can influence reciprocal relationships in important ways. Chapter 5 also examines the impact of pay comparisons on effort behaviour. We compare effort in a treatment where co-workers’ wages are secret with effort in two ‘public wages’ treatments differing in whether co-workers’ wages are chosen by an employer, or are fixed exogenously by the experimenter. We find that pay comparisons are detrimental for effort, particularly when coworkers’ wages are exogenous. Chapter 6 summarises the findings of this research and concludes.