Structural Evolution of Economy and Environment in Australia

Richard Wood

r.wood@physics.usyd.edu.au

 

This research was conducted as part of a PhD by the author, Richard Wood, at ISA, University of Sydney, with the goal being to construct a time series of input output tables in current and constant prices alongisde integrated environmental extensions in order to help understand the key relationships in an evolving economic structure that are driving resource use and greenhouse gas emissions in Australia.

The approach involved looking at the factors and relationships that underpin economic growth in Australia. This research sought to understand the changes in these factors by taking a historical perspective to the determinants of environmental impact through an investigation of structural changes over a period of 30 years. A detailed model is developed using the macro-economic tool of input-output analysis. This model makes it possible to investigate inter-relationships and intra-relationships between sectors of the environment, the economy and the population at disparate scales.

The research is composed of two major components - the estimation of a time series of a detailed environmental-economic model; and the analysis of strucutral change and environmental impact.

 

An estimation of an input-output time series for Australia in current and constant prices

 

The goal of this work was to create a time series of input-output tables for Australia in both current and constant price (nominal and real value) so as to facilitate the analysis of structural change in the Australian economy. Creating such a database is not without significant challenge due to the lack of data consistently available over a time series, and due to methodological changes within the available data. The approach involves use of all available macroeconomic data, and the implementation of large scale mathematical programming techniques to allow for data-reconciliation under partial information.

 

The end result is a database that contains a time series of annual input-output tables for the Australian economy. The database covers the years 1975-2005 and is constructed based on 17 input-output tables published by the ABS and 10 additional Supply-Use tables.

 

Changes in classification have been dealt with by reconciling the 100 or so industry level data with product level data on approximately 1000 commodities for 11 points of time. Confidential and/or missing data has been estimated using both economic trends and aggregate data. Tables have been estimated to be consistent with the current series of input-output tables published by the ABS. Numerous methodological changes have occurred over time. These have been dealt with more recently (from 1995) by estimating relative rather than absolute change based on a consistent set of supply-use tables. Prior to 1995, most major changes occurred with the classification, and these were dealt with as above. Other methodological changes were considered relatively minor in terms of overall economic evolution and subject to mathematical balancing. Within the annual time series, all years with no input-output or supply use data are interpolated based on relative changes in value added components (by industry where available) between consecutive published input-output tables.

 

Finally, the complete set of tables were balanced to published National Accounts data which covers all components of value added, imports and final demand, whilst maintaining all economic balances (total supply = total use, etc). All balancing has occurred using optimisation techniques with estimates on relative standard error of the source data.

 

The constant price time series is based on the current price input-output time series for Australia. The input-output tables are estimated in real values based on a two-step process. Firstly, price indices are constructed for all sectors for the productive, consumptive, import and export components of the economy. Available price indices cover final consumption (61 products); producer - manufacturing (186 products), construction (7 products, only since 1996), transport (33 products, only since 1996), services (37 products, only since 1996). Where producer price indices were not available, consumer price indices were assumed equivalent. These price indices are applied to each product sold within the tables. Constant price value added and final demand aggregates are sourced from the National Accounts. The tables are rebalanced satisfying row/column balances using the available National Accounts data and the estimated deflated tables.

 

It is not currently considered possible to reconstruct each cell of an input-output table over the last 30 years under perfectly consistent methodology and classification. However, this database uses all best available data to provide the most consistent measure of economic evolution that to the Author’s knowledge is currently available. Final demand and value added components are balanced to consistent aggregates, and intermediate usages are based on consistent statistics since 1995, and based on detailed classifications prior.

 

Whilst large changes in small values have occurred over time in these tables, in the author’s opinion economically important activities have generally been well represented in classification and methodological terms. As such, the overall or holistic accuracy of the model for analysing key parts of economic evolution is considered valid. Analysing trends in changes in values is the key focus of the database, but analysing a year-to-year (particularly in earlier years) change would not necessarily reflect real data (noting also that all input-output tables are only updated with real data sporadically). Some individual sectors have been particularly vulnerable to classification and methodological changes, and it would be unadvisable to base results individually on these sectors.

 

For more details about this work, please contact Richard Wood r.wood@physics.usyd.edu.au

 

 

This work has been or is being published in international peer-reviewed journals in the following papers. Please contact the author for copies.

Wood, R. (2009). "Construction, stability and predictability of an input-output time series for Australia." In revision.

A time series of Australian input-output tables has been developed. This paper gives an overview of the construction techniques employed, including the major issues encountered. In order to delineate environmentally important processes, a range of fine detail commodity data was utilised in order to expand the system from roughly 100 sectors into a temporally consistent 344 sectors. Confidential and inconsistent data was overcome using the KRAS method, a recent modification of RAS (Lenzen et al. 2006, 2008; Lenzen et al. 2007). The article then analyses the stability of input-output coefficients. Similar to the work in Dietzenbacher and Hoen (2006), a time series of input-output tables is analysed. The issue of stability of coefficients and multipliers was investigated under the Leontief and Ghosh models of supply/demand. The predictability of the models was then examined under updated final demand or primary input components. Finally, the predictability of the input-output model over varying time scales and under different updating techniques is also investigated.

 

Lenzen, M., R. Wood, et al. (2007). "Some comments on the GRAS method." Economic Systems Research 19(4): 461-465.

Lenzen, M., B. Gallego, and R. Wood, Matrix balancing under conflicting information. Economic Systems Research, 2009. 21(1): p. 23-44.

Lenzen, M., B. Gallego, and R. Wood, A flexible approach to matrix balancing under partial information. Journal of Applied Input-Output Analysis, 2006. 11 & 12.

 

Wood, R. and M. Lenzen (2009). "Aggregate measures of complex economic structure and evolution a review and case study." Journal of Industrial Ecology 13(2): 264-283.
It is perhaps in the nature of complex systems that they call for aggregate measures that enable analysts to grasp their structure and evolution without being overwhelmed by their very complexity. Complex interindustry theory and models are a typical case, where the underlying database - an input-output table - routinely contains thousands of data points for a single year. Within input-output analysis, quantitative measures have been developed that describe and characterize interindustry interactions and that have been used to compare economies, both in a static taxonomy and through their evolution over time. First, we review and critically discuss a number of concepts that have been proposed and applied to interindustry systems, such as interconnectedness, interrelatedness, linkages, and economic landscapes. Second, we apply these concepts to a case study of the Australian economy between 1975 and 1999 in terms of environmental headline indicators. Our results enable the reader to judge the usefulness and ability of the measures in capturing the key structural elements and evolutionary processes governing the interaction between the economy and the environment. For the Australian case study, the measures showed a diversifying economy occurring together with a specialization of environmental flows. © 2009 by Yale University.

 

Wood, R. and M. Lenzen (2009). "Structural path decomposition." Energy Economics 31(3): 335-341.
We combine Structural Decomposition Analysis (SDA) and Structural Path Analysis (SPA) in order to examine the temporal changes within a full production chain perspective. To our knowledge this work constitutes the first formulation of what we call Structural Path Decomposition (SPD). SPD provides noteworthy insight in two instances: first it extracts and ranks those interactions within an economy that are most important in driving change; second it provides a temporal perspective to standard input-output-based Life-Cycle Assessment. In this paper, we develop the mathematical model of SPD and provide two case studies of the most important changes in structural paths in Australia from 1995 to 2005. © 2008.

Wood, R. and C. Dey (2009). "Australia's Carbon Footprint." Economic Systems Research: In press.

Wood, R. (2009). "Structural decomposition of Australia's greenhouse gas emissions." Energy Policy: In press.

A complex system of production links our greenhouse gas emissions to our consumer demands. Whilst progress may be made in improving efficiency, other changes in the production structure may easily annul global improvements. Utilising a structural decomposition analysis, a comparative-static technique of input-output analysis, over a time period of around 30 years, net greenhouse emissions are decomposed in this study into the effects, due to changes in industrial efficiency, forward linkages, inter-industry structure, backward linkages, type of final demand, cause of final demand, population affluence, population size, and mix and level of exports. Historically, significant competing forces at both the whole of economy and industrial scale have been mitigating potential improvements. Key sectors and structural influences are identified that have historically shown the greatest potential for change, and would likely have the greatest net impact. Results clearly reinforce that the current dichotomy of growth and exports are the key problems in need of address. © 2009 Elsevier Ltd. All rights reserved.

 

Wood, R., M. Lenzen, and B. Foran, A material history of Australia:Evolution of material intensity and drivers of change. Journal of Industrial Ecology, 2009.

 

Wood, R. and M. Lenzen (2009). Principal methodological approaches to studying sustainable consumption: Scenario Analysis, Ecological Footprints and Structural Decomposition Analysis. In: Suh S, Handbook on Input-Output Economics for Industrial Ecology, Elsevier.

 

Wood, R. and M. Lenzen (2006). "Zero-value problems of the logarithmic mean divisia index decomposition method." Energy Policy 34(12): 1326-1331.
Recently, the Logarithmic Mean Divisia Index (LMDI) approach to energy decomposition has been espoused as the preferred indexing method. Whilst the LMDI method provides perfect decomposition, and is time-reversal invariant, its strategy to handle zero-values is not necessarily robust. In order to overcome this problem, it has been recommended to substitute a small value [delta]=10-10-10-20 for any zero values in the underlying data set, and allow the calculation to proceed as usual. The decomposition results are said to converge as [delta] approaches zero. However, we show that under this recommended procedure the LMDI can produce significant errors if applied in the decomposition of a data set containing a large number of zeroes and/or small values. To overcome this problem, we recommend using the analytical limits of LMDI terms in cases of zero values. These limits can be substituted for entire computational loops, so that in addition to providing the correct decomposition result, this improved procedure also drastically reduces computation times.