Stochastic Search Variable Selection in Vector Error Correction Models with an Application to a Model of the UK Macroeconomy

This paper develops methods for Stochastic Search Variable Selection (currently popular with regression and Vector Autoregressive models) for Vector Error Correction models where there are many possible restrictions on the cointegration space. We show how this allows the researcher to begin with a single unrestricted model and either do model selection or model averaging in an automatic and computationally efficient manner. We apply our methods to a large UK macroeconomic model.

An investigation of customer order flow in the foreign exchange market

This paper examines the effect that heterogeneous customer orders flows have on exchange rates by using a new, and the largest, proprietary dataset of weekly net order flow segmented by customer type across nine of the most liquid currency pairs. We make several contributions. Firstly, we investigate the extent to which customer order flow can help to explain exchange rate movements over and above the influence of macroeconomic variables. Secondly, we address the issue of whether order flows contain (private) information which explain exchange rates changes. Thirdly, we look at the usefulness of order flow in forecasting exchange rate movements at longer horizons than those generally considered in the microstructure literature. Finally we address the question of whether the out-of-sample exchange rate forecasts generated by order flows can be employed profitably in the foreign exchange markets

Microstructure Order Flow: Statistical and Economic Evaluation of Nonlinear Forecasts

In this paper we propose a novel empirical extension of the standard market microstructure order flow model. The main idea is that heterogeneity of beliefs in the foreign exchange market can cause model instability and such instability has not been fully accounted for in the existing empirical literature. We investigate this issue using two di¤erent data sets and focusing on out- of-sample forecasts. Forecasting power is measured using standard statistical tests and, additionally, using an alternative approach based on measuring the economic value of forecasts after building a portfolio of assets. We nd there is a substantial economic value on conditioning on the proposed models.

Data-based priors for vector autoregressions with drifting coefficients

This paper proposes full-Bayes priors for time-varying parameter vector autoregressions (TVP-VARs) which are more robust and objective than existing choices proposed in the literature. We formulate the priors in a way that they allow for straightforward posterior computation, they require minimal input by the user, and they result in shrinkage posterior representations, thus, making them appropriate for models of large dimensions. A comprehensive forecasting exercise involving TVP-VARs of different dimensions establishes the usefulness of the proposed approach.

Model Uncertainty in Panel Vector Autoregressive Models.

We develop methods for Bayesian model averaging (BMA) or selection (BMS) in Panel Vector Autoregressions (PVARs). Our approach allows us to select between or average over all possible combinations of restricted PVARs where the restrictions involve interdependencies between and heterogeneities across cross-sectional units. The resulting BMA framework can find a parsimonious PVAR specification, thus dealing with overparameterization concerns. We use these methods in an application involving the euro area sovereign debt crisis and show that our methods perform better than alternatives. Our findings contradict a simple view of the sovereign debt crisis which divides the euro zone into groups of core and peripheral countries and worries about financial contagion within the latter group.

Model Uncertainty in Panel Vector Autoregressive Models.

We develop methods for Bayesian model averaging (BMA) or selection (BMS) in Panel Vector Autoregressions (PVARs). Our approach allows us to select between or average over all possible combinations of restricted PVARs where the restrictions involve interdependencies between and heterogeneities across cross-sectional units. The resulting BMA framework can find a parsimonious PVAR specification, thus dealing with overparameterization concerns. We use these methods in an application involving the euro area sovereign debt crisis and show that our methods perform better than alternatives. Our findings contradict a simple view of the sovereign debt crisis which divides the euro zone into groups of core and peripheral countries and worries about financial contagion within the latter group.

Tracking cost savings from competitive tendering in the short and long run

A major initiative of the Thatcher and Major Conservative administrations was that public sector ancillary and professional services provided by incumbent direct service organisations [DSOs] be put out to tender. Analyses of this initiative, in the UK and elsewhere, found costs were often reduced in the short run. However, few if any studies went beyond the first round of tendering. We analyze data collected over successive rounds of tendering for cleaning and catering services of Scottish hospitals in order to assess the long term consequences of this initiative. The experience of the two services was very different. Cost savings for cleaning services tended to increase with each additional round of tendering and became increasingly stable. In accordance with previous results in the literature, DSOs produced smaller cost reductions than private contractors: probably an inevitable consequence of the tendering process at the time. Cost savings from DSOs tended to disappear during the first round of tendering, but they appear to have been more permanent in successive rounds. Cost savings for catering, on the other hand, tended to be much smaller, and these were not sustained.

Prior selection for panel vector autoregressions

There is a vast literature that specifies Bayesian shrinkage priors for vector autoregressions (VARs) of possibly large dimensions. In this paper I argue that many of these priors are not appropriate for multi-country settings, which motivates me to develop priors for panel VARs (PVARs). The parametric and semi-parametric priors I suggest not only perform valuable shrinkage in large dimensions, but also allow for soft clustering of variables or countries which are homogeneous. I discuss the implications of these new priors for modelling interdependencies and heterogeneities among different countries in a panel VAR setting. Monte Carlo evidence and an empirical forecasting exercise show clear and important gains of the new priors compared to existing popular priors for VARs and PVARs.