Will mining investment fall off a cliff?
This article was featured as part of Equilibrium 2013, our annual print publication.
Soft landing for mining?
Many commentators have expressed a relatively sanguine view about the likely course of mining investment over 2014. For example, in its most recent Statement on Monetary Policy, the Reserve Bank wrote that “… given the magnitude of resource projects already committed, particularly liquefied natural gas (LNG) projects, mining investment is expected to remain at an elevated level for a couple of years.”
However, current investment levels look unsustainable. Commencements have been trending down, albeit from extremely high levels since 2009-10 (see Chart 3).
The timing and size of the inevitable decline is important for the domestic economy. First, it is unlikely that non-mining investment will quickly fill the gap created by any sudden decline in mining investment. Private investment outside the mining and housing sectors has been seriously squeezed by mining sector investment for some years and many industries have been under severe pressure (manufacturing, retail, wholesale). Public investment is likely to remain constrained by efforts to achieve a public sector surplus. The timing of the slowdown in mining investment will be an important influence on GDP forecasts. Second, mining construction is relatively labour-intensive and a mining investment decline will be important for aggregate employment. As mining investment translates into export capacity, there will be offsetting effects on GDP but not necessarily on employment because the operational phase of most mining projects is much less labour-intensive than the construction phase.
With this in mind, the capex release in March 2013 provided the first long-term expectations reading for mining investment in 2013-14. This was clearly down from the expectation for 2012-13 made a year ago which, in turn, is not likely to be achieved (see Chart 1). After adjustment for average “realisation ratios” between actual and expected, the expectations suggest that mining investment may rise, but these ratios need to be treated with caution: there have been some spectacular errors around turning points. Neither the sharp upsurge in mining investment in 2006 nor the 1982 recession were anticipated by the survey.
Also that month the ABS published its latest estimates of mineral and petroleum exploration including half-yearly expectations data, and these appear to have moderated (Chart 2).
Commencements and investment
Another way of predicting mining investment is to use the information contained in past commencements data. Mining investment, as measured in the national accounts as gross fixed capital formation, comprises buildings, machinery & equipment and intellectual property products (R&D, exploration and software). The buildings consist of engineering construction (essentially below ground level) and structures (above ground). Data on commencements of engineering construction work for “oil, gas, coal and other minerals” are available from the ABS and may be a useful forward indicator of mining investment activity because (1) engineering construction is one of the earliest activities in most mining projects (after exploration), (2) commencements can be expected to influence activity for long periods of time, especially for large projects, and (3) engineering construction represents as much as 45% of total mining investment.
Commencements appear sensitive to commodity prices and costs through their impact on mining profitability. A simple measure of the overall rate of return on mining capital, net operating surplus relative to the capital stock at replacement cost, suggests that the outlook for commencements is currently less favourable than at any time in the past seven years outside the GFC period (Chart 3).
Econometric estimation (see Appendix and Chart 4 below) indicates that commencements continue to feed into engineering construction for at least three years. In fact, more than two-thirds of engineering work in the current quarter (that is, over one-third of total mining investment) can be attributed to commencements that occurred more than six quarters earlier. Consequently, we can predict engineering construction in the current quarter from commencements data for the past three years using the weighting pattern in Chart 4.
To the extent that much building construction and machinery & equipment investment cannot occur on a mining site until some of the engineering work has been done, it is likely that the potential information in the commencements data is even more important than these figures suggest. Mining industry building construction can be related to recent engineering construction and mining industry machinery & equipment investment can be predicted using past building construction (Chart 7).
A word of caution should be sounded about this analysis. The length of time over which commencements continue to affect investment depends on the gestation period of the projects that are commenced. The gestation period implied by Chart 4 is a maximum of three years. There are several mega projects at the feasibility stage according to BREE’s last report on major projects in October:
These generally appear to have gestation periods of four years or longer, compared with our estimate of three years for the average project. The commencement of mega projects with above-average gestation periods may mean that the predictions from our method overstate near-term investment and understate long-term investment. From an alternative perspective, of five mega LNG projects under way at present, only two are due to startup after 2015. It is important to study the predicted path of the pipeline of projects that have been started but not yet completed to ensure that the projections of work done are plausible.
Assuming commencements remain at their recent average level and using the method described above the projected paths of engineering work done and the pipeline are shown in Chart 5 (see the dashed lines labelled “unadjusted”). Because work done declines rapidly below commencements (note that Chart 5 uses different scales), the pipeline begins to lengthen in late 2014, suggesting the equation is shifting too much work done into the earlier quarters of the projection. To adjust for this, we have added around 5% to the level of work done predicted by the equation through 2013 and 2014 and subtracted a similar amount over the next five years (see the solid lines labelled “adjusted”). These adjustments generate a projected path for the pipeline that continues to decline for much longer than the unadjusted path.
Applying this pattern of future engineering construction yields an aggregate picture in which overall mining investment is maintained at close to current levels until the end of 2013 after which it declines quite rapidly. Translating these forecasts into financial year numbers implies a year-average increase of 20% in mining investment in 2012-13 (mostly reflecting the run-up that has already occurred) but a decline of 6% in 2013-14 and a more serious fall of 21% in 2014-15. This is sufficient to reduce mining investment from an estimated peak of 8.5% of GDP in 2012Q4 to 8.1% in 2013Q4 and then to 6.0% in 2014Q4 (Chart 6). The decline through 2014 would be equivalent to removing more than 2 percentage points from GDP growth in that year.
The commencement of another mega LNG project would alter these atmospheric: for example, a mega project arbitrarily commencing in 2013Q1 with an indicative construction cost of $40 billion, would be sufficient to maintain mining investment into 2014, possibly delaying the decline until late 2015, that is, by 18 months. The RBA prediction for mining investment looks to be somewhere between these two scenarios (see Chart 6).
Finally, we include a worst case scenario that has no more mega projects and a decline in commencements by 50% over the next three years. Because of the lags between commencements and work done (Chart 4), this would not seriously worsen the outlook for mining investment until late 2014 (Chart 6).
What is interesting about this analysis is the potential for mining investment to decline precipitously once existing projects have been completed. This “mining cliff” could appear as early as the first quarter of 2014. The commencement of new mega projects could delay the cliff for 18 months or longer. On the other hand, a decline in commencements in response to declining commodity prices and rising costs is unlikely to affect mining investment until well into 2015 in view of the typical gestation period for projects.
Any sharp decline in mining investment in 2014 would be highly detrimental to growth without an offset. Some of that will come from increased mining exports as new projects continue to come on stream. However, employment is likely to struggle unless non-mining investment fills the breach.
Image: ‘Placer Mining – Columbia, Tuolumne Co., Mining out a ranch’ by Thomas Houseworth & Co., Publisher, https://nypl.getarchive.net/media/placer-mining-columbia-tuolumne-co-mining-out-a-ranch-9f7365. Licence at https://creativecommons.org/licenses/publicdomain/.