New Tests for Cointegration in Heterogeneous Panels
This paper makes the following contributions to the existing literature on panel cointegration. First, two new tests based on the principle of weighted symmetric estimation are proposed for panel cointegration testing. Second, the asymptotic distributions of these new tests are examined, and these are shown to be well defined Weiner processes that are free of nuisance parameters. Third, the size and power properties of the proposed tests are studied with a Monte Carlo simulation, and their properties are found to be superior to those of the existing tests across a range of environments.
Demographics and asset prices in Australia: Do the dynamics of population ageing matter?
The effects of population ageing on asset markets are complex. Recent literature has raised concerns of significant downward pressure on asset prices, housing and financial, due to the rapid demographic transition associated with retiring Baby Boomers. Awareness of this demographic transition and speculation over the consequent effects on asset markets prompted the asset meltdown debate. This thesis contributes to the asset meltdown debate and addresses the question whether demographic transitions, particularly the increasing proportion of the population in the old age cohort due to the retirement of Baby Boomers, will precipitate a dramatic decline in house and stock price in Australia. The structural vector autoregressive model used for the empirical analysis is an important improvement over the reduced-form regression strategies usually employed in the literature. Both the demographic and non-demographic variables used in the empirical analysis are treated as endogenous and reverse causality between the variables is taken into account. The population ageing dynamics are modelled using impulse response functions and, thus, an insight into the potential magnitude of demographic shocks, particularly retirement shocks, is obtained. The analysis quantifies the responses in real house and stock prices to such shocks. The structural shocks are characterised as a sequence of shocks, often with different signs at different points in time, rather than one-off shock. The cumulative effect of such a sequence of shocks on the evolution of real house and stock prices over time is examined using historical decomposition. In addition, the forecast error variance decomposition is used to quantify the percentage contribution of the total variation in real house and stock prices to each structural shock in the models for different forecast periods. The findings support the optimists’ view in the asset meltdown debate. Predictions that population ageing, or more specifically, changes in age structure particularly due to retiring Baby Boomers, will lead to pronounced downward pressure on real house or real stock price in Australia are rejected. The findings suggest that Baby Boomers are unlikely to sell enough housing and financial assets in retirement to precipitate a market meltdown, or a sudden and sharp decline in real house or stock prices. With the benefit of hindsight, we also see that the one fourth of the Baby Boomers are already retired and the Australian housing market and stock market does not show signs of collapse or substantial price decreases. Poterba (2001) provides a possible explanation for these findings, namely, even though changes in age structure affect asset demand, these effects are simply too small to be detected among the other shocks to house and stock prices. Moreover, the anomaly as revealed by asset ownership statistics, that the older population cohort continues to hold or accumulate assets rather than de-accumulate as originally predicted by the life cycle hypothesis sheds light on why population ageing does not exert a pronounced downward pressure on asset prices in Australia.
New Tests For Cointegration in Heterogeneous Panels
In this paper, the performances of panel data unit root tests are considered and various estimation methods under different properties of data are compared. It is shown that weighted symmetric estimation increases the power of the tests without adversely affecting the size, for most data properties and most panels of dimensions N and T. The presence of serial correlation and cross-sectional correlation does not reduce the power of the tests significantly.
Panel Data Unit Roots Tests Using Various Estimation Methods
In this paper, the performances of panel data unit root tests are considered and various estimation methods under different properties of data are compared. It is shown that weighted symmetric estimation increases the power of the tests without adversely affecting the size, for most data properties and most panels of dimensions N and T. The presence of serial correlation and cross-sectional correlation does not reduce the power of the tests significantly.
The Effects of Fiscal Policy on Output, Unemployment and Housing Prices in Australia
The effects of fiscal policy on economic issues are complex. Recent literature has raised concerns about the significant impact of fiscal instruments on GDP, consumption, investment, unemployment and housing due to the financial crisis. This thesis addresses the question how government expenditure and/or tax revenue impacts on GDP, unemployment and housing in the case of Australia. We use three different methods to exam the effects of fiscal policy on the Australian economy.
The structural vector autoregressive model used for the empirical analysis is an important improvement to the reduced-form regression strategies usually employed in the literature. The first method applied was short-run identification method using EVIEWS, which produced no significant impact and results that conflict with current economic theory. The analyses suggest that traditional approaches measuring the effects of fiscal policy on GDP and unemployment rate must be rethought. Findings from this research illustrate the dangers of incorrectly invoking an economic assumption of linking the fiscal instruments to GDP and unemployment rate.
The second method was sign restrictions identification method using RATs program, based on Uhlig's criteria using SVAR. The results from the second method suggest that whenever the government wants to increase GDP or reduce the unemployment rate, they can increase spending or reduce tax revenue. This result appears to be consistent with those of major contributors to the literature (see for example Blanchard and Perotti, 2002; Mountford & Uhlig, 2009). Although the results come from two different identifications methods, pure-sign-restriction and penalty-function, these findings all followed the same directions. Findings on the unemployment rate response to fiscal shocks, confirmed those of studies such as Monacelli et al. (2010) and Holden & Sparrman (2014), but differed with Bruckner & Pappa (2010).
The third method uses sign restrictions identification, in conjunction with a Bayesian method executed in MATLAB, to control the signs of the responses from other variables to fiscal shocks. This third can control the results to one specific model. We imposed restrictions to examine the relationships between fiscal policy shocks and the housing price index. Our results show that government spending has positive effects on the housing price index, but that tax revenue shocks do not. This result holds validity for fiscal and housing market policymakers, in the light of housing bubbles and economic fluctuations.
Introductory Econometrics: A Practical Approach (2nd edn), by Hamid R Seddighi (Routledge, New York, 2012), pp 385
'Introductory Econometrics: A Practical Approach' is most aptly named. It explains step by step the applications of econometric methods in practice. The text also discusses the key econometric issues in a non-mathematical way and provides a summary of key issues in econometrics. The book covers all econometric topics for typical undergraduate courses: the classical regression model, simultaneous equations, qualitative variables and panel data, time series econometrics and the like. One of the attractive features of this book is that its 16 topics are grouped into five teaching units. This gives both lecturers and students guidance in teaching as well as in studying the subject.