Skip to Content

Department of Economics

Applied and Theoretical Econometrics Research

Our Econometrics Research Group conducts research into a range of theoretical and applied issues in econometrics. Our main areas of interest are:

  • Applied Econometrics
  • Time-series Analysis
  • Financial Econometrics

Our work has been published in top rated academic journals such as: The Econometrics Journal, The International Economic Review, Journal of Financial Econometrics, Journal of Banking and Finance, Journal of Money, Credit and Banking, Journal of Time Series Econometrics, The Journal of Multivariate Analysis, Oxford Bulletin od Economics and Statistics, Journal of Time Series Analysis, Economic Modelling, Empirical Economics, Journal of Financial Research, Economics Letters, Applied Economics, The Economic Record, The Energy Journal, Journal of Productivity Analysis, Canadian Journal of Economics.

Econometrics Research Group Members

Our group consists of the following members:

  • Lance Fisher
  • Chris Heaton
  • Roselyne Joyeux
  • George Milunovich
  • Daehoon Nahm
  • Natalia Ponomareva
  • Jeffrey Sheen
  • Shuping Shi
  • Ben Wang

Our Research Projects

We currently run a range of research projects including the following:

Issues related to the identification and analysis of Structural Vector Autoregressions (SVAR) models 

George Milunovich, Lance Fisher, Chris Heaton and Jeffrey Sheen

This work is about providing conditions and developing tests for the identification of SVAR models via non-traditional strategies, such as the use of heteroskedasticity for the purpose of identification and sign restrictions. This has important implications in the areas of finance, macro and microeconomics where traditionally non-testable economic hypothesis are rendered overidentifying, and hence testable by hereby developed methods.

Financial Contagion and Systemic Risk 

George Milunovich, Jeffrey Sheen and Ben Wang

Financial contagion broadly describes the observation that links between financial markets strengthen during periods of financial crises, relative to benchmark relations measured during periods of normalcy. This phenomenon is of concern to both investors and policy makers alike because it implies that financial crises can spread to markets, and asset classes, with no underlying exposure to the source of risk. Contagion in financial systems may be estimated for example using a copula function that describes the dependence properties of the overall financial system. Contagion can amplify systemic risk, which in turn affects and may be caused by macroeconomic and financial factors. State space and hurdle-type models can be used to shed light on these interactions. They provide valuable early warning information that helps our understanding of the drivers of financial risk in a systemic context, which became a major issue in the recent global financial crisis.

Modelling high dimensional time series 

Chris Heaton

The classical approaches to time series analysis generally assume that the number of variables considered is small relative to the number of observations. This is at odds with many modern datasets and consequently the development of statistical techniques that are valid for applications in which the number of variables is of a similar magnitude, or larger than the number of observations is of interest. Potential applications include macroeconomic forecasting, structural macroeconomic analysis, and the analysis of returns in large asset markets.

Testing for multiple-period predictability between serially correlated time series 

Chris Heaton

Tests of one-period-ahead predictability between time series (e.g. the Granger-causality test) are well-established in econometrics and are known to perform well. When these tests are naively extended to cases of multiple-period predictability, the induced autocorrelation results in a severe size distortion that results in false findings of predictability. This project is concerned with the development of test statistics that have good size and power properties in applications for which the prediction horizon is large.

Estimating measures of risk using high-frequency data 

Chris Heaton

The estimation and analysis of parameters that measure market risk (e.g. value at risk, expected shortfall, etc) is essential for the sound management of financial institutions. The growing availability of high-frequency financial data allows for the daily estimation of these parameters without the need for strong assumptions about the intraday stationarity of the distribution of returns. The objectives of this project are to develop techniques for estimating and forecasting measures of market risk, and to develop techniques to assess competing forecasting models and estimation techniques.

Econometric detection and dating of financial bubbles 

Shuping Shi

Recent work on econometric detection mechanisms has shown the effectiveness of recursive procedures in identifying and dating financial bubbles in real time. These procedures are useful as warning alerts in surveillance strategies conducted by central banks and fiscal regulators with real time data. Use of these methods over long historical periods presents a more serious econometric challenge due to the complexity of the nonlinear structure and break mechanisms that are inherent in multiple bubble phenomena within the same sample period. This project aims to develop a new econometric method that is better suited for practical implementation with long historical time series.

Change detection in causal relationships 

Shuping Shi

It is well known that the empirical causal relationships between economic and financial variables are often not stable and systematic. This project aims to implement a procedure to identify changes in causal relationships and the timing of these changes. Measures of interconnectedness amongst financial institutions based on Granger causality are currently used in detecting systemic risk. However, researchers need to assume exogenously defined samples of calm and stress periods to implement this approach. The new test will allow real time identification of changes in the direction of causal relationships, and thus provide important advances in identifying emerging turmoil in financial markets.

Bubbles or market fundamental? 

Shuping Shi

An obvious question is whether the seemingly ever-increasing house prices are driven by market fundamentals or by speculation? Inferences based on housing affordability ratios can be misleading, because they do not explicitly incorporate the effects of other fundamental factors such as interest rate and population growth. A wrong signal of bubble existence may lead to damaging consequences such as false policy action, overbuilding, or even to the acceleration of the bubble process. This project aims to propose an innovative approach to distinguish between a rapid rise in home prices induced by changes in fundamentals and a housing price bubble.

A general theorem for the measurement of scale efficiency 

Daehoon Nahm

Scale efficiency is an important characteristic of a production technology. While economists/econometricians have little difficulty measuring scale efficiency in a non-parametric analysis, they run into difficulties in measuring scale efficiency from a parametric model. This project aims to develop a general theorem that can be utilised in general cases by extending the earlier finding by Nahm and Vu (2013), which is specific to a hyperbolic distance function.

Decomposition of impulse response functions into direct and indirect effects of a structural shock 

Daehoon Nahm

Researchers employing a structural VAR model for their analysis typically assume orthogonality for structural shocks so that their variance-covariance matrix is diagonal. In many cases, however, one cannot simply rule out the contemporaneous correlations between structural shocks, and allowing them can lead to a much more flexible and richer model setting and interesting analyses that are not possible under the usual assumption of orthogonal errors. One of the benefits of allowing contemporaneous correlations between structural shocks is that the impulse responses can be decomposed into direct and indirect effects of a shock. This is different from the usual decomposition of forecasting-error variance in two important aspects: i) it decomposes the impulse itself rather than the forecasting-error variance, and ii) it breaks down the effect of a shock to a single variable (typically a policy variable) while the variance decomposition is based on the assumption that all the variable in the system change by the same amount. The objective of the project is to develop a method that can be employed for the decomposition of impulse responses in a general SVAR or SVARMA setting.

Mixed frequency state space model 

Jeffrey Sheen, Stefan Trueck, Ben Wang

State-space models in economics are typically used to estimate unobserved (economic) states that are important to economic agents. These agents, ranging from ordinary consumers to government policy makers, form their decisions based on whatever economic data is available to them. These data are typically released at different dates and available at different frequencies. The advantage of a mix-frequency state-space model is that it blends all relevant information arriving at different frequency into state measures. One outcome of this project is: Sheen, Trueck and Wang (2015) A daily business condition index for the Australian economy, forthcoming in the Economic Record. In this paper we estimate daily domestic and external business condition indices for the Australian economy.

Modelling exchange rate using principal component analysis 

Natalia Ponomareva, Jeffrey Sheen, Ben Wang

Principal component analysis is a dimension reduction method for analysing data. This method orders the estimated factors according to their power to explain the co-movements of a panel of data. We apply this method to analyse bilateral exchange rates for all major currencies. We ask whether these statistical estimates of the principal components have an economic interpretation. We also test whether we can forecast bilateral exchange rate using these principal components. One outcome of this project is Ponomareva, Sheen, Wang (2015) `The U.S. factor in explaining and forecasting bilateral U.S. exchange rates', unpublished working paper. In this paper we show the first principal component explains more than 50% of co-movements of the bilateral U.S. exchange rates, and is indeed related to U.S. macroeconomic conditions. We also show the principal component forecasts are better than those generated by a random walk for monthly and weekly frequencies.

Asian Bond Market 

Roselyne Joyeux

The recent development of an Asian bond market has reduced the region's dependence on short term bank financing. It is now well established and is expected to reach $US1 trillion in three years and issuance of $US150 billions in 2014. Half of the issuance in the last year has been from Chinese companies and real estate property companies, such as Central China, have figured prominently. With the tightening of credit worldwide Chinese developers' bond prices volatility will increase. Possible fallouts for the Australian economy are very likely. For example, there is some evidence that the latest increases in real estate prices in Sydney and Melbourne are linked to the exit of investors from the Chinese real estate markets investing in Australia. However, so far no existing study has investigated the determinants of Asian bond prices volatility. This project aims at filling that gap.

Chinese real estate market 

Roselyne Joyeux

The real estate industry contributes greatly to Chinese economic growth. In 2009 the real estate industry directly accounted for over one tenth of total GDP growth at the national level, and therefore played a substantial part to the quick recovery of the Chinese economy during the global financial crisis (Deng, Morck, Wu and Yeung, 2011, Wu, Gyourko and Deng, 2012, and Wu, Deng, Huang, Morck and Yeung, 2013). The overall importance is even larger if indirect influences from construction material industries are taken into account. Besides, as reported by the People's Bank of China, at the end of 2010 the outstanding balance of developer plus residential mortgage loans represented around 6.28 trillion yuan RMB, over 18% of total loans. However, over the last decade, residential land prices have displayed remarkable volatility: prices increased on average by approximately 25% per annum in Beijing (annual appreciation in Beijing during 2007, 2009, 2010, and 2013 was even in excess of 50%), while in 2011 land prices in Beijing dropped by 30%. A similar pattern of residential land market volatilities has been observed in other major housing markets, such as Shanghai and Hangzhou (see Deng, Gyourko and Wu, 2012, for more discussion). Some observers and researchers argue that fundamentals of the housing sector, both sector specific and macroeconomic, may have been the driving forces behind housing price volatility at the expense of speculative factors. In order to shed new light on the influence of economic fundamentals and speculation on the volatility of this financial market, we follow the methodology introduced by Engle et al. (2013) and Engle and Rangel (2008).

Commodity prices 

Roselyne Joyeux

Commodity prices and their volatility have increased markedly over the past ten years prompting the G20 leaders, at their summit in June 2012 to state that "excessive commodity price volatility has significant implications for all countries, increasing uncertainty for actors in the economy and potentially hampering stability of the budgets, and predictability for economic planning" (G20 2012 Report). As a result there has been considerable interest in the factors influencing commodity prices and the extent to which they are driven by fundamentals or by the financialisation of commodity markets. Recently two major factors have been presented as driving such commodity price movements: high growth in China and speculation. Indeed on the one hand Chinese growth has been strongly dependent on infrastructure, real estate investment, automobile production for the domestic market and consumer electronics for exports; all activities highly metal-intensive, absorbing an increasing share of the world output of metal commodities (more than one-third for copper). On the other hand, commodity price movements have been frequently assigned to speculative pressures, possibly fueled by low interest rates and generous provision of liquidity by central banks. The relative role of these two series of factors in commodity price volatility has been hotly debated in a recurrent way, but the extraordinary rise in Chinese bank credit in 2009 (as well as the two rounds of quantitative easing by the Fed) has offered this controversy a new lease of life. Whether the boom and bust periods of the commodity markets have matched the movements of growth or inflation in China remains to date an open question due to lack of empirical work. Existing work has focused on the explanation of commodity return movements. However, less attention has been granted to the relationships between commodity price volatility, macro fundamentals and speculation. This is especially the case for Chinese non-ferrous metal futures market and the recent period.

Income and Energy Consumption in Asia 

Roselyne Joyeux

This research studies the relationship between energy consumption and income for a panel of Asian economies. Panel data methodologies are employed to gain the advantage of increased explanatory power of the econometric analysis results from pooling time series and cross section observations. In addition, the analyses incorporate common factors as a means of accounting for variables beyond the bivariate relationship between income and either total final energy consumption or electricity consumption. In the cross section dimension the panel includes 20 economies across Asia, and in the time series dimension it ranges over 41 years from 1971 – 2012.

Finding of the Determinants of the Sacrifice Ratio: using  the BMA Approach 

Natalia Ponomareva

There are a growing number of theoretical and empirical studies on determinants of the sacrifice ratio that measures the cost of disinflation in terms of output loss. Ball (1994) finds the sacrifice ratio to be lower for quicker disinflation and more flexible wage-setting. Other studies consider additional explanatory variables, including institutional settings, economic conditions and the political climate. Some results are robust across different studies, while others not. The large number of potential explanatory variables and contradictory results in the existing literature indicate the presence of model uncertainty. This study deals with this uncertainty using the Bayesian model averaging method to identify the significant determinants of the sacrifice ratio without relying on ad hoc model selection. Our results show that the length of the disinflation is the most important for the cost of disinflation. This supports the "cold turkey'' argument for faster disinflation.

Modelling dependence between commodity currencies and commodity prices 

Natalia Ponomareva

This project employs copula approach to study the relationship between exchange rates and commodity prices for large commodity exporters. Using data for the nominal exchange rates of the four commodity currencies (Australian, Canadian, and New Zealand dollars, and Norwegian krone) against the U.S. dollar and the relevant country-specific commodity price indices, constructed on a daily basis, we _find (i) a positive dependence between the values of commodity currencies and commodity indices, that is, commodity index increases when respective currency appreciates; and provide several explanations this finding (ii) a pronounced increase in (a time-varying) dependence starting from the beginning of the global financial crisis; (iii) no major asymmetries in the tail dependence for most pairs of the exchange rates and commodity indices.