Capex survey points to further falls in investment – Dec qtr 2015

The Australian capex survey results for the December quarter are disappointing on all levels. Capex remains firmly lower than last year, forecasts are for further declines in the next two quarters and the first view of 2017 looks like capex will continue to fall. There is one small bright point – actual December capex increased over September. This may be positive for the December quarter National Accounts. The reason for the increase can be traced back to activity in NSW and unfortunately, the forecasts by business suggest that it is not likely to continue.

Estimate 5 for capex in the 2015/16 financial year highlights that total capex is still expected to fall by -17% or $26.6b below last year

Expected capex for the full financial year to June 2016 is forecast to come in at $124b, which is 17% below last year. So far, the Sept plus Dec 15 capex actuals are tracking at -17.8% below the same two quarters last year in real terms and -15% in nominal terms.

The benefit of estimate 5 data is that it is made up of 50% actuals and 50% forecast. The forecast period now looks out over a shorter time frame (next 6 months) so is becoming a more accurate view of capex plans:-

Source: ABS

The decline in capex for this financial year, at estimate 5, is looking like it might reach levels not seen since the last recession in 1991/2.

It doesn’t matter whether you look at capex expectations by industry or asset type – ALL areas are contributing to the decline. But the decline is being led by Mining and Buildings and Structures:-

Source: ABS

The only positive in this is that expected capex in Manufacturing doesn’t look like it will decline as much as it did last year. Apart from that, capex in Other Selected Industries and Mining will fall faster than last year as well.

But are the actuals in the GDP results following the capex survey? The data is only up to the Sept qtr and for the moment, yes, private capital investment is following the lead of the capex survey. But we haven’t seen the larger falls in capex that the survey is forecasting. That said, capex has been slowing, then declining since 2012:-

Source: ABS

At estimate 5, we have two quarters of actuals and two quarters of forecasts for capex. The Dec quarter actuals came in higher than the Sept qtr actuals and this could be positive for GDP in the Dec qtr. I’ll come back to this in more detail shortly. But this means that the bigger falls in capex are forecast for the last two quarters of the 2015/16 financial year (which I’m calling ‘Half 2’ – H2 in the table below). From ABS 5625.01(a) and 5625.01(b):-

Actual Capex for H1 v Forecast Capex for H2 2015/16 Financial Year

Source: ABS, $ M, original, current prices

Across every industry, businesses are forecasting capex in the March and June 16 quarters to be between 15 and 20% below capex for the first half.

These forecasts for the March and June 16 quarters do not compare well with actual capex for the same quarters last year either:-

Actual Capex for Mar & Jun 2015 v Forecast Capex for Mar & Jun 2016

Source: ABS, $ M, original, current prices

I’ve outlined before that the Capex Survey (what this post is based on: ABS 5625) doesn’t cover all industries and tends to overstate the size of Mining. The industries that are excluded are: Agriculture, Forestry and Fishing (Division A), Public Administration and Safety (Division O), Education and Training (Division P), Health Care and Social Assistance (Division Q), Superannuation Funds (Class 6330).

The size of the capex change tends to be overstated in the survey compared to the actuals in the National Accounts, but are directionally in line. The Capex survey data is used in the development of the quarterly National Accounts.

The first look at expected capex for 2017 looks just as poor

“Non-mining business investment is forecast to pick up in the second half of the forecast period, reflecting the improvement in domestic demand”, RBA, Statement on Monetary Policy, February 2016

Estimate 1 for the next financial year 2016/17 was also released with the December data. This is a 100% forecast and is looking further out to the full financial year 2016/17, so its accuracy is lower than later estimates.

The view at estimate 1 tends to understate final capex. So adjusting the estimate using the 2016 realisation ratio for estimate 1 (which bumps up the estimate by 20%), we get our first capex estimate for 2017 of $99b. Let’s assume that 2015/16 capex spend comes in at the estimate 5 forecast of $124b. That means that estimate 1 is pointing to a further 20% fall in capex for 2016/17.

The falls in capex at estimate 1 are across the board as well:-

• Mining – estimate 1 adjusted upward based on a +3% realisation ratio = $35b, which represents a further 33% decline on current forecast for mining capex of $53b in this 2015/16 financial year
• Manufacturing – estimate 1 adjusted upward based on a +27% realisation ratio = $8.3b, which represents a further 1.1% decline on the current forecast for this financial year
• Other Selected Industries – estimate 1 adjusted upward based on a +49% realisation ratio = $62b, which still represents a further 1.5% decline on the current forecast for this financial year

(The realisation ratios that I’ve used here are the larger of the last 5 years)

But according to the RBA in the latest Statement on Monetary Policy Feb 2016:-

“The ABS capital expenditure (Capex) survey of investment intentions and the Bank’s liaison point to a sharp fall in mining investment in 2015/16. The subtraction from GDP from lower mining investment is expected to peak this financial year.” RBA, Statement on Monetary Policy February 2016

Here’s hoping that the later estimates of the 2016/17 financial year capex improve, because so far, the falls for next year are looking just as bad as this year.

But actual capex in the survey INCREASED in the Dec quarter

It looked like there was a spot of good news in the capex survey data. Actual capex in the December quarter grew by +0.8% in real terms. A turnaround possibly? It unfortunately didn’t take long to see that this is all driven by one state – NSW, which contributed +1.97% points to the +0.8% quarterly increase in actual capex. The only other states to make a positive contribution to growth in the latest quarter were VIC and SA – and contribution to growth was small at best.

Source: ABS

But its nots looking like NSW is on a new investment trajectory either – the latest quarter of higher growth seems like a function of a lower result in the Sept qtr.

Source: ABS

Capex in NSW is still 8% below its 2011 peak in real terms and 4% below the Dec 14 quarter (same time last year).

Reverting to current prices, capex in NSW for the next two quarters is also forecast to be 18% lower than in the actuals in the Sept15/Dec15 quarters. NSW capex in Mar 16 + Jun 16 is forecast to be $11,051m versus the Sept 15/Dec 15 actuals of $13,544m. In other words capex in NSW is not likely to continue increasing in this financial year.

With capex falling and business continuing to forecast lower capex for 2017, the signs for the Australian economy are not good. While mining has mostly been leading the falls, the declines in capex are now evident many across the other industries – which suggests that the ‘transition’ has so far been tepid at best. Business remains wary of investing to increase capacity in light of overall lower growth.

Internet job vacancies point to low employment growth – January 2016

The monthly internet job vacancies data is another indicator of future employment growth. This month internet job vacancies continued to grow at a steady pace of +0.4% trend and 0% seasonally adjusted. While there has been a trend of growing vacancies, it does appear to have slowed over the last few months. This does not suggest high, or accelerating, growth in employment over the next few months. But the internet vacancy index has been tracking along like this for a while now – and employment growth continued to outperform.

Source: Dept of Employment 24 Feb 2016 (left hand axis = # of vacancies index)

The Dept of Employment gathers the data on job ads from 3 main job search websites – CareerOne, Seek and Australian Job Search (AJS). It suggests that the index in early 2015 was understating the growth in vacancies:-

“Over the five months to August 2015, there was a reduction in the number of new job advertisements listed on the Australian JobSearch (AJS) website. This may have been as a result of the transition of employment services from Job Services Australia to jobactive. The number of job advertisements listed on AJS began to recover in September; however, the data in this report should be used with caution.” Source: Dept of Employment, 24 Feb 2016

However, the current level of the internet vacancy index (IVI) is lower than the two previous peaks of 2006-08 (45% lower) and 2009-11 (20% lower). Comparing this to employment growth at those same peaks shows that while the periods of employment growth line up with the periods of increase in the IVI, the relative size of the growth during those periods does not correlate at all:-

Source: ABS

For example, the latest peak in annual employment growth exceeded the previous peak in late 2010 by 4% – yet current vacancies are running at 20% below that corresponding peak in late 2010. It’s difficult to reconcile the relative growth of the two data points.

An interesting aspect of the IVI report is the regional internet vacancy index. It provides some insight into the distribution of internet job vacancies and it seems directionally in line with the state-based analysis of employment growth in my previous post – Is the performance of the labour market really that good?

Change in internet vacancies over the year to January 2016

Source: Dept of Employment

According to the report, the largest % increases in vacancies over the year were recorded in:

“Southern Highlands & Snowy NSW (up by 34.4%), Canberra & ACT (26.1%), Bendigo & High Country VIC (23.4%), NSW North Coast (21.8%), and Ballarat & Central Highlands (21.6%)”

The largest % falls in vacancies over the year were recorded in:

“Bathurst & Central West NSW (down by 44.8%, albeit from a low base), Port Augusta & Eyre Peninsula SA (18.6%), Perth (15.9%), North West Tasmania (13.7%) and Central Queensland (12.8%)”

VIC is the only state where all regions saw an in increase in the IVI over the last year – yet employment growth had been subdued in the last year compared to previous peaks. NSW and QLD were mostly all positive. ACT was positive. But in most regions in WA the growth in the IVI was negative. In all regions in SA, TAS and NT the IVI was negative for the year.

Is the performance of the labour market really that good? January 2016

The labour force data continues to raise more questions about its own methodology than it answers about the performance of the labour force. There remains a very wide 95% confidence interval around the ‘true’ change in employed persons between January and December, which is somewhere between a -65k decline in employed persons and a +50k increase in employed persons. But the 95% confidence interval around the ‘true’ unemployment change figure is now just skirting zero at the lower bound, between -9.6k and +70k. So while the interval is wide, it’s starting to be more likely that unemployment increased in January.

There is a different way to present the data to get an idea of how the market is performing. In this post, I will focus on the difference between the monthly change in employed persons and the monthly change in the labour force during the last year. The reason for looking at the labour market in this way is to gain a different perspective on the drivers behind changes in unemployment and to understand the ‘health’ of employment growth over time – is it accelerating or slowing? This provides a different view on the performance of the labour market by state. All does not seem well in NSW, our main employment growth engine. The labour markets in VIC, WA, TAS and NT are also showing some signs of deterioration. The most positive performance is in QLD and ACT, with SA still mixed.

Overall annual employment growth remains over 300k persons

At this point in time, the labour market still looks like it is performing above average on an annual basis, with employment growth well above the 10 year average for January (+186k). Growth in full-time (FT) employed persons continues to exceed growth in part-time (PT) employed persons and total unemployed persons continues to decline across all time periods.

Source: ABS

What you can’t see from this chart is that the annual growth in employment is skewed to one state, with NSW accounting for 56.3% of the National annual growth in employed persons. NSW represents 32% of all employed persons in Australia.

Unemployment continues to fall, but not in all states

The good news is a bit more limited when it comes to the change in unemployed persons, especially when you break it down on a state by state basis and when you start to look at the performance over more recent time periods. It’s really only NSW, SA and QLD where unemployed persons has continued to fall in any large and meaningful way throughout the year (looking at time periods of six months or less).

It looks like unemployment has started falling in WA in the last quarter as well, but this is a function of the labour force growth slowing faster than employment growth. I’ll come back to this point in more detail shortly.

Source: ABS

The mixed performance of total unemployed persons across the states raises some questions of just how widespread this strong labour market performance really is.

When employment grows faster than the labour force, unemployment falls

This is true even when employment growth is slowing faster than labour force growth, which is what has been happening at a National level over the last 5 months:-

Source: ABS

Employment growth has exceeded labour force growth throughout most of 2015. The positive gap between the two measures since August 2015 equals the decline in total unemployed persons during that time of -28.9k persons.

The trend shows that both the growth in employment and the growth in the labour force has been slowing during that time. Whether this is an enduring trend or not remains to be established. But the gap between the two has been narrowing since October 2015 – which means employment growth is slowing faster than the labour force. If this trend continues over the next few months, unemployment could start to grow again on a National basis.

As an aside, why is labour force growth slowing?

The analysis above looks at the labour force as the sum of all employed persons plus all unemployed persons that are looking for work. Another way to view the labour force is by 1) contribution from population growth and 2) contribution from changes in participation. It seems we are seeing less contribution from participation growth over the last six months:-

Source: ABS

The contribution from growth in participation has more than halved since peaking in August 2015. In the last six months, participation has fallen in VIC, SA, WA, TAS and NT.

I’m less concerned with the apparent drop off in population growth over the last two months. This seems to be a regular feature of the data. But there has been a slow-down in what underlying population growth has added to the labour force since its peak in Feb 2008.

But we live in a country where economic fortunes have differed greatly among the states and is in the process of shifting from the mining-led states back towards to the eastern seaboard. It’s worthwhile looking at each state in a bit more detail to understand this transition.

NSW – the main engine driving the so-called post-mining transition is sputtering

The state of NSW has been the strongest performing state in terms of employment growth during 2015. As mentioned above, NSW alone accounted for over 56% of the annual National employment growth in 2015. The current level of annual growth in employment in NSW is extremely strong in historical terms as well – it’s the highest level of annual employment growth on record for NSW. Previous peaks were between +110k and +120k growth in annual employment. The current level of annual employment growth in NSW is +170k employed persons (Jan 2016). This is almost on par with the National 10 year average in January.

The monthly change in employment growth in NSW has now halved over the last six months from the peak of +17.3k growth in employed persons in May 2015 to +7.9k growth in employed person in January 2016.

Growth in FT employed persons had far exceed growth in PT employment over the last year, with FT employment growth of +142k versus PT employment growth of 28k persons. This has reversed as of December, with PT employment now growing faster than FT on a monthly basis. As of January, FT employment growth has slowed to a very low +2.6k persons – down from the peak of 15.9k as recently as July.

What has ‘saved’ unemployment from growing in NSW is that the labour force growth has also slowed – faster than the slow-down in employment growth.

Source: ABS

Since the peak in labour force growth in June 2015, total unemployed persons has actually declined by 20k in NSW. While this is great news, the gap between the measures is narrowing. This state will be important to watch.

VIC – labour market is deteriorating

There are two problems in VIC. The first is that employment growth during 2015 has been subdued. It’s a big state, representing 25% of employed persons, yet only accounted for 14% of employment growth in 2015. Most of the reason for this seems to be that PT employment has been slowing and is now declining (last 4 months). There are early signs that FT employment may have also peaked in this current cycle. For a short time, unemployment was declining. Which brings me to the second problem. Employment growth has now slowed faster than labour force growth and unemployment has started to increase again in the latest quarter.

Source: ABS

It’s a negative pattern – employment growth is slowing from a low peak and unemployment has started to increase again over the last 3 months.

QLD – labour market performance is good

During 2015, QLD had made a bigger contribution to employment growth than VIC, accounting for 25% of the annual National employment growth. For most of 2015, employment has grown faster than the labour force in QLD resulting in lower unemployment. The trend of the employment growth looks positive too – it has been accelerating. The only thing slowing down employment growth in QLD has been lower PT employment growth. Growth in FT employment is still accelerating.

Source: ABS

Unemployment in QLD continues to fall, but this has slowed to zero in the last two months.

SA – performance has been mixed

Employment in SA has continued to grow, with recent growth on par with previous peaks in growth. But most of the current employment growth has been in PT employed persons, whereas FT employment growth was only positive for the last 4 months of 2015. But at least the growth in PT employment was still higher than the growth in labour force – and unemployment has fallen as a result.

Source: ABS

WA – unemployment falling, but it’s hardly a robust market

This is the frontline of the slowing mining investment engine. While employment has started growing again in the last half of 2015, it is well below recent growth levels.

All of the growth has been PT in nature as well. FT employed persons has declined by 24k persons in the last year, whilst PT employed persons has increased by 28k. Cost cutting continues in earnest, but at least there are PT jobs available.

The decline in the level of unemployed persons is mostly the result of slowing growth in the labour force, rather than strong growth in employment.

Source: ABS

TAS – the labour market has deteriorated

The annual rate of employment growth in TAS has turned negative over the last 3 months. Only for a short period during 2015 was employment growth positive and above the level of growth in the labour force. FT employment continues to decline (-2k FT employed persons), but at least in the last few months there has there been some small level of growth in PT employed persons (not enough to offset the falls in FT employment though).

Source: ABS

Unemployment has started to increase again in the last few months of 2015. The current level of unemployed persons, 17k persons, whilst still elevated, remains below the 2013 peak of 20k persons.

NT – employment is declining and unemployment is miraculously falling

The mining transition continues to hurt NT, with employment declining in the latter half of 2015. Even though employment has been declining on a monthly basis, it hasn’t been declining as fast as the labour force. As a result, the level of unemployment has actually declined over the last six months.

Source: ABS

Just looking at a declining rate of unemployment wouldn’t give you the full story about the labour market in NT.

ACT – labour market is improving

The level of employment growth in ACT had been consistently low throughout 2012-2014. But since the latter half of 2014, and in the latter half of 2015, employment growth has started to accelerate. The current level of employment growth is not high by historical standards, but it’s a good sign that it is accelerating and this is different to many other states. Unemployment has grown on an annual basis because the labour force size has grown faster than employment.

Source: ABS

Another positive sign is that FT employment has overtaken PT employment growth as of October 2015.

It doesn’t take much scratching below the surface to see that in most states, the labour market is no longer as strong as the annual figures suggest. It also highlights the need to review a range of different measures, rather than just relying one figure at a point in time as a gauge of labour market strength. The big watch out at the moment is slowing employment growth relative to the labour force across the bigger population states, especially NSW.

Australian balance of trade worsens in the December 15 quarter

Last week, the ABS released the latest Balance of Trade figures for the month of December 2015. It is the first view of what the nominal trade balance for the December quarter is likely to show. This early view shows that December quarter net exports is likely to detract from nominal GDP growth in the fourth quarter. The main driver of this result was the continued fall in the total value of exports of -3.2% for the quarter. Further falls in commodity prices and a slightly stronger $A during the quarter seem to be the main drivers. The 4^th quarter GDP/Balance of Payments figures in real terms (adjusted for price movements) will be the test to see if export volumes have continued to hold up, given concerns over slowing Chinese growth.

The size of the trade deficit was partially ‘saved’ by a -0.7% decline in the value of imports versus the previous quarter. If imports had continued to grow, then the nominal trade deficit would have been higher. But in the detail of this decline in imports, there appears to be a rapid slowdown in the import of consumption goods over the December quarter. This is only one quarter where the trend has changed, but is something to watch, especially as an indicator of a change in household spending patterns.

The value of the trade deficit widened again in the December quarter.

This first view of the Dec quarter trade deficit shows a large increase versus the Sept quarter. At this rate, the nominal value of net exports is likely to detract approx. -0.51% points of growth from nominal GDP in the Dec quarter GDP.

Source: ABS

For the year, this has been the largest trade deficit recorded in Australia (in the available data from the ABS). The main reason behind the deterioration in our trade balance has been the lower value of our exports.

The total value of exports fell by 3.2% in the Dec quarter

The December quarter results show a weaker performance across many of the export categories, but not all.

Source: ABS, The Macroeconomic Project

The export of rural goods was the only area where contribution to growth improved in the December v. September quarter. Given that in USD terms, rural commodity prices were fairly stable over the quarter, but fell in $AUD terms (due to a slightly stronger AUD over the quarter), it’s likely that volumes were stronger in the quarter. While export of services also made a positive contribution to growth, the contribution slowed to half in the December quarter.

Of all the export categories, the main contributor to the fall in export value was non-rural goods (commodities, metals, machinery, other manufactures etc.) contributing -3.5%pts to the overall decline. Non-rural goods represented 58% of the total value of exports in the December quarter, so has a fairly large impact on the overall result.

The big part of the fall in value was related to the continued falls in commodity prices throughout the last quarter of 2015. Priced in USD, commodity prices fell by 9% over the quarter.

Source: RBA

Priced in $A terms, commodity prices fell by more than 11% and this was the result of a strengthening in the AUD over the quarter (on avg +4.2% versus Sept quarter – Source: RBA). At the very least, we know that export volumes didn’t grow enough to offset the negative impacts of both price falls and a stronger currency on the total value of non-rural exports.

This result is likely to be income-negative in the December quarter. Falling revenue for commodity producers/exporters will continue to place pressure on their operating environment (lower income, profit and more pressure on cost reduction) and will also see continued falls in tax receipts for the government.

The 4^th quarter data will be important to see how volumes have held up – as one indicator of underlying demand. The question of the Chinese slow down and its potential impact on Australian exports remains unclear. As of the Sept quarter, export volumes of our largest export (metal ore & minerals) continued to grow – albeit at a slower pace.

Source: ABS

The latest China Resources Quarterly for Summer 2015/6, shows that Australian volumes of iron ore exports to China were slightly lower in the Dec quarter versus the Sept quarter.

Source: Westpac

But for profit, income and tax receipts, its nominal terms (volume*price) that still matters the most. The chart above clearly shows the value of iron ore exports falling.

What ‘saved’ a worse trade deficit was a fall in the nominal value of imports

The value of imports fell by -0.7% in the December quarter and growth reversed across most import categories.

Source: ABS

The best performing categories were the import of intermediate goods (which didn’t decline further in the latest quarter) and the import of services (very little change from last quarter).

Whilst the declines across the bigger import categories aren’t large, they still represent relatively large reversals from the previous quarter.

As I highlighted in a previous post, there have been some unprecedented trends with regard to imports recently. The following chart from Sept 15 Balance of Payments takes a historical view of chain volumes (import values in real terms) for goods & services imports:-

Source: ABS

What is unprecedented is the change in the trend of imports of goods and services over the last several years.

Services imports clearly peaked back in the June 2013 quarter and quarterly volumes have been declining ever since. The Sept 2015 quarter volume (real $’s) measure of services imports is now 16% below the 2013 peak. The implicit price deflator for services (services imports price index) over the same period of time has increased by over 24%.

Goods imports peaked in June 2012 – since then, the quarterly volumes have fallen by only -2.4%. In other words, remained fairly flat – but this is still a big shift in the historical trend. The implicit price deflator for all goods imports has increased by only 8%. This low increase is being influenced by a 40% fall in Fuels & Lubricants import prices during that time (source: ABS 6457.03).

Part of what has driven this decline in goods imports since June 2012 has been the decline in the import of capital goods as the investment stage of the mining boom started to unwind. This has subtracted -7.53% points from the growth of goods imports in real terms over that time. But on the other side of the ledger, the import of consumption goods has continued to grow, contributing +4.12% points to the growth of goods imports (also since June 2012). In nominal terms over that same time, the contribution from growth in consumption goods imports far outstripped the decline in capital goods imports and contributed +9% pts to import growth versus capital goods which detracted -3.2% pts from import growth.

The nominal value of consumption goods imports has declined in the December quarter

So the import of consumption goods (such as food, household electrical, textiles, clothing, footwear etc.) has been one of the stronger performers in terms of import growth – in both real and nominal terms. Its also the second biggest import category in nominal terms. Given the categories it covers, it could be used as another barometer of household spending and demand in the economy – spending which has remained fairly stable. It’s important to note that the import of consumption goods hasn’t ‘peaked’ and reversed in the same way that services or capital goods imports have peaked. The red flag in this latest release of the balance of trade figures is that the value of consumption goods imports appears to have declined in the December quarter:-

Source: ABS

This isn’t unusual in the data – clearly there have been periods when the value of consumption goods imports has declined quarter on quarter. But the November & December rolling quarters are reversing at least 12 months where the import of consumption goods has consistently grown at +4%.

It’s possible that the fall could be price-related (as the currency has continued to fall since 2011). But the data also shows that since the AUD currency depreciation commenced, consumption goods import prices and consumption goods import volumes have been growing in tandem. In fact, during 2015, growth in both price and volume of imported consumption goods (at least up to Sept 2015) has been accelerating:-

Source: ABS

What we know so far about December is that import prices for consumption goods still grew by +1.4% (quarter on quarter), yet the nominal value of consumption goods imports declined by -1.1% which detracted -0.3%pts from the growth in imports. We’ll have to wait until the Q4 data to see the impact on volume. It’s too early to say what this means, but it’s certainly something to watch.