End of the Industrial Super-Cycle

There’s an issue, or an unstated assumption, which underlies much of the current worry about the state and direction of the world economy. One way of appreciating it is asking the question: ‘How can the US economy still be expanding vigorously when its industrial sector is in such trouble?’

Consider the evidence. US industrial output has shown consistently negative momentum since 4Q14 and this shows few signs of reversing: in August output fell 0.4% mom, and the regional industrial surveys for September have been grim, with the Philadelphia Fed survey shocking at minus 6 , the Empire State manufacturing survey grim at minus 14.7, the Richmond Fed survey showing minus 5 (worst since January 2013), and the Kansas City Fed manufacturing survey showing minus 6.  At the same time, inventory/shipment ratios have risen to the highest levels since 2009, so far without improvement;  capacity utilization rates have tumbled from a high of 79% in November to 77.6% in August. Exports, meanwhile, have endured consistently negative 6m momentum trends since July 2014, and by July 2015 were falling 7% yoy.

Despite this, domestic demand indicators have remained on balance positive, and, in particular, labour markets have remained robust. In the face of the most dramatic trade/industry downturn since the great recession, 2Q GDP growth came in at 3.9% annualized, largely on the back of a 3.6% annualized rise in personal consumption.

The contradiction between what is happening in the industrial sector and the trajectory of the broader US economy shows up clearly in my momentum indicators.

This divergence between the industrial sector and the overall economy is fairly obvious in the US. What is less obvious is that something very similar is characteristic of the entire global economy.  

As the chart shows (*see below for details on these indicators), on a global basis, the industrial sector is clearly in trouble, but despite that, domestic demand momentum is not merely being maintained, it continues to accelerate slowly, as it has for much of the time since early 2014. 

It seems to me that the most important thing about this aspect of the global economy is to acknowledge that it is really happening, and has been happening for nearly a year now.  There is a deeply entrenched expectation that where the industrial sector leads, the rest of the economy must inevitably follow, from which it follows that a description of the industrial sector cycle is adequate to locate an economy’s current trajectory and potential.

Part of the reason for that is that economists (and everyone else) feel far more comfortable analysing the industrial economy than the services sector.  At the  most basic level, industrial output is far easier to count, movements in industrial prices are far easier to observe, balance sheets of industrial companies easier to take apart, all of which has allowed us a very good idea of how business cycles affect industrial companies.  Similar analysis of the services sector fails at the first hurdle - even counting the output is so uncertain that we rely on hard-fought and contestable conventions and inferences, rather than direct observation. As for pricing, inventories, capital involved. . . .

Historically, there have been good historical reasons for the expectation that where industry leads, the rest of the cycle will follow, and what’s more, over the last 20 years new life has been breathed into those reasons by China's rise.  Nevertheless, the expectation is, in philosophical terms, not necessary but only contingent - and it may be that as China has got richer, the contingency is passing. 

Consider how spending patterns in the US have changed. In 1950, spending on goods accounted for 60.7% of all personal consumption spending, with services accounting for only 39.3%. However, as incomes rose, so the proportion spent on services rose, until in the 12m to June 2015, the proportions were almost exactly reversed, with 67.2% of personal spending going on services, vs only 32.8% on goods. But even that exaggerates the importance of industrial sector supply, since two thirds of spending on goods is on ‘non-durables’ such as food and gasoline. In fact, spending on durable goods accounts for only 10.8% of US personal consumption spending.  This helps explain why a loss of momentum in the US industrial sector need not necessarily be pointing to a wider economic slump. 

Such a pattern should surprise no-one: as a society grows wealthier, so marginal demand shifts from the acquisition of goods to the consumption of services. 

One should expect to find this shifting pattern of demand not just in the US and Europe but, of course, in Asia too.  And given the extraordinary rise in material comfort in China over the last 20 years, one should expect a similar pattern of shifting demand there too.  Although we do not have the data to show this directly, the changing composition of China’s GDP makes it clear that this process is underway.

In 1995, secondary industries (ie, principally manufacturing) accounted for 46.7% of GDP output, whilst tertiary industries (ie, principally services) accounted for 33.6%.  By 2015, the ratios had changed so that tertiary industries accounted for 48.6% of output, whilst secondary industries accounted for 42.2%. But note that as far as domestic demand is concerned, China was also running at trade surplus of approximately 5.5% of GDP, suggesting that domestic demand for secondary industry products had probably sunk to around 36.7% of GDP. 

The shift in underlying demand is even clearer when one looks at marginal contributions to GDP: in 1995 tertiary industries accounted for about 29% of marginal GDP growth; in 2015 that had risen to 68%. 

This shift in Chinese marginal demand away from goods to services, although predictable, is nevertheless the signal that the forces which drove the commodities ‘super-cycle’ - the sudden emergence of demand for goods from a Chinese population transitioning from poverty to material decency -  are no longer the primary forces driving either the commodity cycle, the global industrial sector, or indeed, the world economy.  When China first started emerging from grinding poverty to mass material decency, it was predictable that the first priority for China's population was to acquire more 'stuff'. So it made sense to buy the stuff that made 'stuff' - hence the commodities super-cycle.

But as China's population has grown richer, its marginal demand has shifted from 'more stuff' to 'better services'. Crudely put, its the shift from a new shirt to a sharper haircut.  Unless industrial companies have factored in this slight slowdown in the rate of growth of marginal demand from China, the industrial sector will discover it cannot win the return on capital from its capex that it originally expected.  On the other hand, demand for services will continue to grow relatively unimpaired.

In these circumstances, disappointing industrial demand and all that goes with it can easily co-exist with continued growth of employment, and overall demand in the world economy. 

* (My global momentum indicators for the industrial sector and for monetary conditions take in data from the US, Eurozone, Japan and China; the global domestic demand indicator also includes data from the UK, S Korea and Taiwan. In each case, the indicator measures standard deviation movements from historic seasonal trends for key data. For the industrial indicator, this includes output, exports (local currency value and volume) and where possible indicators for inventory ratios and capacity utilization rates. For domestic demand, I include retail sales, vehicle sales, employment, wages, and selected other indicators where possible. For monetary conditions I include growth of monetary aggregates, movements of the currency vs the SDR, movements in real interest rates and changes in the shape of the yield curve. In each case, for global aggregates, countries are weighted according to 5yr average of US$ denominated GDP.   No single-figure indicator will be perfect, but I am confident that these are sufficiently information-rich to not be completely wrong.)

Cargo-Cult Companies - Japan's 2Q Duponts

The MOF's 2Q survey of private sector balance sheets and p&ls reveals this: more than ever before, corporate Japan's ROE depends only on the ability to source supplies cheaply, and there is little sign it wishes to change this business model or expand its reach. It is a cargo-cult approach, in which all depends on the vale of what washes up on Japan's shores.

On the downside, this confirms that Abenomics' hoped-for rejuvenation of the Japanese economy is nowhere to be seen. On the upside, in the short term, a devaluation of the Rmb will probably aid Japanese profits, rather than erode them as I initially thought.

The 2Q private survey presents a picture of extremes:

• the highest operating margins since my data starts in 1980; 
• the lowest asset turns since my data starts in 1980; 
• the lowest financial leverage since my data starts in 1980. 

At the moment, the gains in operating margins trump all else, raising ROE to 10.4% (just below the post-200 average) and ROA to 3.9% (1SD above the post-2000 average). So it is probably no surprise that as the key ratios which determine return on equity scale off in both directions to to previously unseen extremes, there is no sign of any change whatsoever in corporate behaviour.

And what does it all add up to? Operating profits growth running at just 7% yoy on a 12m basis, and investment in plant and equipment up just 5.5%, only just enough to cover the depreciation allowances claimed.

In 2Q sales rose 1.1% yoy (1.4% 12ma) whilst operating profits jumped 20.5% yoy (7.4% 12ma), and as a result, margins rose to 4.81% (vs 4.52% in 1Q), and 4.3% on 12m. These are the fattest operating margins for Japan since my data begins in 1980.

The reason for the rise in margins is simply an improvement in corporate terms of trade, with the cost of goods sold ratio falling 0.9pps qoq to 76.4%, the lowest since at least 1980 (although on a 12m basis, the 77.3% ratio was matched in 2Q11). There is no further improvement in SG&A /Sales, with the ratio rising slightly to 18.8% (vs 18.2% in 1Q and 18.7% in 2Q14). And there was practically no further improvement in the sales/expenses per employee ratio, as sales per employee fell 3.1% yoy whilst expenses per employee fell 3.4% yoy. 

Both asset turns and financial leverage continue to decline to new lows. Total assets rose 5.2% yoy and 5.9% 12ma whilst sales rose 1.1% yoy and 1.4% on a 12ma, so annualized asset turns fell to 0.87, from 0.95 in 1Q and 0.91 in 2Q14. On a 12m basis, asset turns fall to 0.906x, the lowest since 1980s.

Whilst total assets rose 5.2% yoy in 2Q, shareholders’ net worth rose 6.5% yoy, so financial leverage fell to 2.63, or 2.66x on a 12m basis: again, the lowest since 1980 at least. The cash portion of that net worth continues to rise, up 8.6% yoy in 2Q, equivalent to 11.4% of total assets, or 11.3% on a 12 basis - the highest proportion since 1992 in the immediate aftermath of the zaiteku financing bubble years.  Those cash holdings strip 4.5 percentage points from return on equity, cutting it to 10.4% from the ex-cash ratio of 14.9%.