Abstract
A recent Australian Productivity Commission Staff Working Paper has estimated a 25 % decline in value added total factor productivity in Australian coal mining over the period 2000/2001–2006/2007. The decline coincides with a period of rapidly increasing global demand for, and price of, coal, which has made it economically feasible to mine coal that was previously uneconomic. This “digging deeper” hypothesis is plausible, although our initial surprise at the magnitude of the estimated decline has motivated our investigation of productivity in Australian coal mining, its financial implications and its likely causes. We begin by applying the Staff Working Paper methodology to revised data to estimate the magnitude of the decline. This exercise shrinks the estimated magnitude of the decline to 21 %. We continue by applying an alternate methodology to the revised data to re-estimate the magnitude of the decline. This exercise generates no further change to the estimated rate of decline. We then compare estimates of the decline obtained from value added and gross output frameworks, and we find a considerable shrinkage in the estimated productivity decline that highlights the importance of the Domar factor in productivity measurement. We continue by exploring the financial consequences of the productivity decline. We conclude by joining the search for sources of the productivity decline.
Similar content being viewed by others
Notes
This is not Australia’s first mining boom. Battellino (2010), then Deputy Governor of the RBA, and Banks (2012), Chairman of the PC, recount several previous booms, some demand-driven and others initiated by new discoveries, dating back as far as the 1850s gold rush. The principal differences between the current demand-driven boom and previous booms are (1) this boom has been stronger and has lasted longer than previous booms, and (2) this boom is the first to occur under a floating exchange rate regime, under which the RBA has allowed the AUD to appreciate and act as an automatic stabiliser that has kept domestic inflation relatively low.
ABS cat. no. 8415 reports survey information on employment, total income, selected labour costs, purchases and selected expenses, total expenses and operating profit before tax (which, apart from inventory change, equals the difference between total income and total expenses). Subtracting purchases and selected expenses from total income creates value added, or factor income, and subtracting selected labour costs from factor income generates capital income. Again apart from inventory change, the ABS splits capital income into six expense items associated with capital (one negative) and operating profit before tax. Alternatively, adding the value of inventory change to total income and subtracting total expenses yields operating profit before tax. The Data Appendix provides further details.
The ABS (2007) has conducted experimental comparisons of average annual growth of a capital services index based on wK and rK. In mining they find a difference of less than a tenth of a per cent in capital services growth from 1985/1986 to 2005/2006, despite a one per cent difference between rK and wK. They find similar results across a range of market sector industries. This suggests that a capital services index may be relatively insensitive to the choice between wK and rK, although it does not follow that estimated productivity growth is also relatively insensitive to the choice. Nonetheless, it is with our coal mining data because of the pattern of capital deepening.
We have also constructed Fisher gross output and value added productivity indexes. Both Törnqvist and Fisher indexes are superlative indexes, and productivity trends obtained from Fisher and Törnqvist indexes are virtually identical.
Outsourcing consists of “contract mining expenses” and “other contract, subcontract and commission expenses;” the two components were merged prior to 1996/1997.
For the mining sector the ABS (2011) reports 11 % value added productivity growth over 1994/1995–2000/2001 and 19 % productivity decline over 2000/2001–2006/2007. It also reports 6 % gross output productivity growth over 1994/1995–2000/2001 and 12 % productivity decline over 2000/2001–2006/2007.
Georgescu-Roegen (1951) introduced profitability as a financial performance indicator, and noted its desirable property of being independent of the size of production units, a useful property since real gross output has increased nearly fourfold during the sample period.
In Fig. 4 Π, Y/X and P/W are indexes set to unity in 2000/01. The actual value of Π is 1.31 in 2000/2001, which corresponds to an AUD1.26 billion nominal profit.
Δ(Y/X) = Δ(W/P), making the latter a dual, price-based, productivity index, only if ΔΠ = 1, a sufficient but not necessary condition for which is π = 0. Thus, if profit is expensed, rK replaces wK in the productivity and price recovery indexes, and Δ(W/P) is dual to (i.e., equivalent to) Δ(Y/X). This relationship constitutes a potential advantage of the use of the return to capital; if prices are measured more accurately than quantities, productivity measurement can be based on Δ(W/P). This is the approach Hsieh (2002) applies to revisit productivity growth in East Asian economies.
The longstanding claim that open cut mines are more productive than underground mines appears frequently, but empirical evidence is scarce. Barger and Schurr (1944) found open pit mines to have higher labour productivity than underground mines in the US in the 1930s, and attributed the difference to the difficulties of introducing mechanisation and exploiting scale economies in underground mines. James (1952) found labour productivity to be three times as high in open cut mining as in underground mining in Pennsylvania. The New South Wales Department of Primary Industries (nd) reports output (both raw and saleable) per employee for open cut and underground NSW mines, and finds the former roughly double the latter. The Queensland Department of Employment, Economic Development and Innovation (nd) reports much smaller ratios in Queensland mines, with a productivity advantage in open cut mines ranging from 10 to 40 %. These are labour productivities, and we know of no total (or total factor) productivity study that addresses this issue.
Mudd (2009) provides data on open cut and underground coal mining in Australia and each of its states from 1939/40, and he provides data on overburden removal from 1942/1943, with gaps during the 1973/1974–1990/1991 period. He also warns that post-1990/91 overburden data provide lower bounds to actual amounts of overburden. The sizes of the estimated decline in YR/B and estimated increase in B/X may seem suspiciously large, but ABARES (2011) proclaims substantial “technological advancements” in overburden removal.
Coal mining also has well-documented difficulties in attracting skilled employees to remote locations. Unlike the capital input, however, labour is expensed when it is employed, not when it is needed.
To illustrate the importance of the weather, the 2011 Queensland floods cost the industry more than AUD 2 billion in foregone revenue. As for rail and port infrastructure bottlenecks, it is not uncommon to see scores of ships off the Queensland coast waiting to be loaded, a delay that costs companies hundreds of millions annually to compensate for delays. As for chronic labour shortages, a study prepared for the Queensland Resources Council predicts that if currently planned investment projects go through, the mining sector in Queensland alone will require an additional 40,000 skilled workers. Consultancy Deloitte (2011) includes these three factors, together with government regulation and tax policy, and stakeholder activism, among the top ten issues confronting mining companies in 2012. Discussions with a Rio Tinto executive revealed the company’s belief that the primary drag on productivity is neither resource depletion nor internal capital lag, but the inadequate external coal supply chain, improvements to which occur discretely so that the supply chain goes through alternate stages of capacity shortage and excess capacity.
References
ABARES (2006) Australian coal exports: outlook to 2025 and the role of infrastructure. www.abare.gov.au/publications_html/energy/energy_06/coal_exports.pdf
ABARES (2011) Australian energy resource assessment 2011. Chapter 5 coal. http://adl.brs.gov.au/data/warehouse/pe_aera_d9aae_002/aeraCh_05.pdf
ABS (2007) Experimental estimates of industry multifactor productivity, Australia: information paper. Cat. No. 5260.0.55.001. www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/5260.0.55.0012007
ABS (2011) Experimental estimates of industry multifactor productivity, Australia: detailed productivity estimates. Cat. No. 5260.0.55.002. www.abs.gov.au/AUSSTATS/abs@.nsf/ProductsbyCatalogue/A834424832179D03CA25739B00158F99
Balk BM (2009) “On the relation between gross output—and value added—based productivity measures: the importance of the Domar factor”, macroeconomic dynamics. 13 Supplement S2:241–267
Balk BM (2010) An assumption-free framework for measuring productivity change. Rev Income Wealth 56 Special Issue 1 (June), S224–S256
Banks G (2012) Australia’s mining boom: what’s the problem? In: Banks G (ed) Advancing the reform agenda: selected speeches. Productivity Commission, Canberra. http://www.pc.gov.au/speeches/reform-agenda
Barger H, Schurr SH (1944) The mining industries, 1899–1939: A study of output, employment and productivity. National Bureau of Economic Research, New York. http://www.nber.org/books/barg44-1
Battellino R (2010) Mining booms and the Australian economy. Address to the Sydney Institute, 23 February 2010. www.bis.org/review/r100224d.pdf
Davis HS (1955) Productivity Accounting. University of Pennsylvania Press, Philadelphia
Deloitte (2011) Tracking the trends 2012: the top 10 trends mining companies may face in the coming year. www.deloitte.com/mining
Domar ED (1961) On the measurement of technological change. Econ J 71:284 (December), 709–729
Georgescu-Roegen N (1951) The aggregate linear production function and its applications to von neumann’s economic model, Chapter IV. In: Koopmans TC (ed) Activity analysis of production and allocation. Wiley, New York
Hilmola O-P (2006) As currency changes matter: improving the control of profitability and productivity in manufacturing companies. Int J Prod Qual Manage 1(4):321–338
Hsieh C-T (2002) What explains the industrial revolution in East Asia? Evidence from the factor markets. Am Econ Rev 92:3 (June), 502–26
James CM (1952) Measuring productivity in coal mining. Research report no. 13, Industrial Research Department, Wharton School of Finance and Commerce, University of Pennsylvania, Philadelphia, PA
Kendrick JW (1961) Productivity trends in the united states. Princeton University Press, Princeton, NJ
Kendrick JW, Creamer D (1961) Measuring company productivity: handbook with case studies. Studies in Business Economics 74. The Conference Board, New York
Mudd GM (2009) The sustainability of mining in Australia: key production trends and their environmental implications for the future. Research report no. RR5. Department of Civil Engineering, Monash University and Mineral Policy Institute (revised April 2009). http://users.monash.edu.au/~gmudd/sustymining.html
New South Wales Department of Primary Industries (nd). www.dpi.nsw.gov.au/minerals/resources/coal/coal
OECD (2001) OECD productivity manual: a guide to the measurement of industry-level and aggregate productivity growth. www.oecd.org/dataoecd/59/29/2352458.pdf
Rio Tinto (2011) Annual report www.riotinto.com/annualreport2011/
Topp V, Bloch H, Parham D (2007) Mining productivity: the case of the missing input? www.pc.gov.au/research/conferences/productivityperspectives/pp2007
Topp V, Soames L, Parham D, Bloch H (2008) Productivity in the mining industry: measurement and interpretation, Productivity commission Staff Working Paper. www.pc.gov.au/research/staffworkingpaper/mining-productivity
Queensland Department of Employment, Economic Development and Innovation (nd) www.dme.qld.gov.au/mines/coal.cfm
Zheng S, Bloch H (2013) Australia’s mining productivity decline: implications for MFP measurement. J Prod Anal. doi: 10.1007/s11123-012-0329-4
Acknowledgments
We thank the Productivity Commission and the Australian Bureau of Statistics for their assistance. We have received helpful comments from Bert Balk and Harry Bloch, and from audiences at the University of Queensland, the Asia-Pacific Productivity Conference at Academia Sinica, Taipei, the North American Productivity Workshop at Rice University, Houston, and an International Workshop on Efficiency and Productivity at Universidad Miguel Hernández, Elche, Spain. We are grateful to three referees for their constructive criticism.
Author information
Authors and Affiliations
Corresponding author
Data Appendix
Data Appendix
TSPB kindly provided our capital quantity index, and additional indexes they used in their study. Our remaining data were obtained online from ABS and ABARES.
The ABS reports nominal values of total income and total expense, its materials expense, labour expense and capital expense components, and employment (cat. no. 8415 since 1996/1997, cat. no. 8414 from 1992/1993 through 1995/1996, and cat. no. 8402 for 1991/1992). We derive a nominal gross output series from the ABS nominal total income series by deducting income not directly related to mining production activities (interest income, government funding and other income), and by adding the value of inventory change. Our nominal materials expense is the same as the ABS purchases and selected expenses item. Our nominal labour expense is the same as the ABS selected labour costs item. Our nominal capital expense is the sum of the ABS depreciation and amortisation, interest expenses, insurance premiums, natural resource royalties expenses and bad and doubtful debt items, less capitalised purchases and capitalised wages and salaries. The sum of our nominal materials, labour and capital expense items equals the ABS total expenses item. Our nominal value added series is the difference between our nominal gross output series and our nominal materials expense series. We define S CK as capital expense divided by the sum of capital expense and labour expense, with capital expense and labour expense defined above. We define S RK as the sum of capital expense and profit (defined as gross output less total expenses) divided by value added, the sum of capital expense, profit and selected labour costs. We convert all series, apart from S CK and S RK , to indexes based at 2000/2001.
We define real gross output as the ratio of nominal gross output to an implicit price deflator obtained by dividing nominal gross output by tonnes of saleable coal, as reported by ABARES (Australian Mineral Statistics). We obtain real materials expense by dividing nominal materials expense by an ABS index of mining input prices (cat. no. 6427). We define real value added as the difference between real gross output and real materials expense. Our measure of employment coincides with that of the ABS, the number of persons employed at the end of June, by management unit, and we define the price of labour as the ratio of nominal labour expense to employment. We derive a capital price index as the ratio of nominal capital expense to the TSPB capital quantity index. The USD/AUD exchange rate is obtained from the Reserve Bank of Australia.
Simple correlations between our nominal values of gross output, materials expense, and value added series and those of TSPB are all above 0.99. However our nominal gross output series has a 2 % wider range, and our nominal materials expense series has a 7 % narrower range, than those of TSPB. Consequently our nominal value added series has a 10 % wider range than that of TSPB. However virtually all of the differences occur before 2000/2001; the nominal value added range post-2000/2001 is less than 1 %. Our employment series has marginally narrower range than that of TSPB, with all differences occurring before 2000/2001. We have identified two possible sources of these differences. One is that we follow the post-2000/2001 ABS procedure of excluding capitalised purchases and capitalised wages and salaries from nominal capital (and therefore total) expense, whereas TSPB appear to follow the pre-2000/2001 ABS procedure of including capitalised items in nominal gross output. A second concerns employment data, which we collect at the management unit level from ABS sources. However prior to 2000/2001, the TSPB employment series appears to comingle employment data at the management unit level with employment data at the establishment level. ABS ceased reporting establishment level information in 2000/2001. The consequence of these differences is that our output per unit of input series exhibits faster growth than that of TSPB prior to 2000/2001, and marginally slower decline post-2000/2001, whether output is measured as gross output or value added.
Rights and permissions
About this article
Cite this article
Lovell, C.A.K., Lovell, J.E. Productivity decline in Australian coal mining. J Prod Anal 40, 443–455 (2013). https://doi.org/10.1007/s11123-013-0340-4
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11123-013-0340-4