Onshoring job and manufacturing data point in different directions. Why?

In the latest data available, UK apparel manufacturing jobs grew in 2013 on 2012, but official data seems to show production fell. In the US, precisely the opposite happened: jobs fell, but production grew. Why?

One explanation, of course, is that some of the data’s wrong. We’re dealing with very small numbers here: the rough calculations below imply domestic production in the US accounted for just 0.8% of the market in 2012, and that means many ways of researching the volumes and jobs involved can be very prone to error. But let’s look at what the numbers seem to show, and then at some possible explanations

1. The scale of domestic production: the past quarter century in the US 

The US has much richer data on this than the UK – but the high-level generalisation most observers agree on is that domestic production over the past 25 years has collapsed, while imports have soared. In 2012, US foreign trade data showed America importing $76.8 billion worth of clothes, at factory prices, excluding import duty. It exported  $5.5 billion – and data from its Federal Reserve shows domestic production (again at ex-factory values) was worth $688 million. The conventional way of calculating a market size, at ex-factory values, is to add imports to domestic production, then subtract exports. Statisticians will point out that in this market the numbers for imports, exports and domestic production don’t all describe quite the same thing. But it’s not far from the truth to say that, at ex-factory values, the US apparel market in 2012 was worth $71.96 billion ($76.8 bn plus $688 mn minus $5.5 bn) and that domestic production accounted for just 0.95% of this.

I’ve then taken the same numbers for Jan-Oct of 2013, looked at how they changed on the same period of 2012, and applied those rates of change to 2013. And that implies this year the market will be worth $74.8 bn (in cash, it’s growing 3.9%) and domestic production will account for just 0.84%. Argue if you will about the precise significance of these numbers: but domestic production has been falling forthe past 25 years almost consistently. The chart below shows the apparent  total market value in billions of dollars by the blue line (left hand axis) and domestic production’s apparent percentage share by the red line (right hand axis):

This looks pretty simple. But it also looks as if something remarkable happened this year

2. The past few years in the US

At first sight something remarkable seems to have happened in the middle of 2013: a 25-year trend reversed

But we have been here before. That June 2013 switch from years of decline to growth also happened in June 2010. But it didn’t last:

What happened, I believe – and there’s a lot of anecdotal evidence that much the same thing is happening now – was that some US garment factories were closing in 2010 for a number of different reasons (difficulty in competing with foreign manufacture, difficulty in recruiting and difficulty in raising finance), while at the same time other factories were opening. For six months in the second half of 2010, new production in the expanding group exceeded production lost in the declining group. The pressures constraining factories didn’t go away: but for a few months they were less important than the trends pushing other businesses forward. The 8.1% October 2013 domestic growth in production is only slightly higher than the 7.1% growth recorded in October 2010.

I’m not saying the current US resurgence (and remember: domestic production in October 2013 was just 0.8% of the national market) is another dead-cat bounce. But the possibility can’t be ruled out.

3. Why the current upswing is getting publicised

One reason the 2013 autumnal uplift is getting more publicity than the 2010 one is that the United Nations Industrial Development Organisation (UNIDO) now publishes a quarterly summary of worldwide movements in manufacturing. Quite misleadingly, its Q3 2013 report says “For the first time over the last several years, growth in the manufacture of wearing apparel was observed in industrialized countries”. It may be the first time UNIDO has noticed growth in rich-country garment making: but the national statistics of Germany and the UK (and Ireland, and Denmark and indeed the entire EU-15 group of industrialised European countries) showed rather faster garment making growth in Q3 of 2010 than US data did. UNIDO didn’t publish its data in those days – but the EU and US did.

But apart from a few cheap gibes about international bureaucrats, there’s another explanation. Onshoring, as a concept, hadn’t been invented back in 2010. Now it’s a stock cliche of management journalists. If UNIDO’s report had been around in 2010, the uptick in US and EU data would have looked like a random fluctuation in a quarter-century of decline. Now it’s the one piece of evcidence that a phenomenon widely dimissed might actually exist – so journalists love it.

What’s not being publicised at the same time is the trend in imports – which are still growing. Not for the first time, domestic production is growing slightly faster than import. Since early 21012, production and imports look as if they’ve grown about the same rate.

But they haven’t. Since 2010, domestic production has fallen from about 1.1% of imports to about 0.8%. The shift is seasonal – but each year, the seasonal uplift in the share held by US production keeps getting to a slightly lower level. In 2013 the Q3 fall was imperceptible to the eye on a graph:

But the share was – almost imperceptible however you try to interpret the data – very, very, very teensily lower than in Q3 2012. In 2012, 0.743%. In 2013, 0.741%.

The real story is: it’s not getting smaller any more

4. What these numbers are telling us

Import data and production data aren’t the same thing. Every single garment arriving in, or leaving from, the US or EU is tracked – and with it a lot of other data, usually including its weight, its description, the price the factory charged, the duties imposed and the cost of its insurance and freight. Governments have been doing this for centuries: we know a great deal more, for example, about England’s export trade in wool throughout the past 800 years than we do about its production or domestic use. And that data is notoriously accurate, give or take the odd recording error, or attempt at outright fraud. It’s a lot easier to count fleeces going onto or coming off a ship than to get an honest answer from a farmer if you ask him how much wool he’s produced. Especially if he suspects you’re going to tax him on the answer.

They sometimes tried asking farmers 800 years ago, but do its equivalent more often today with factories. But what you ask them, usually is “how much, in dollars, have you sold?” And then you deflate the answer by inflation. That answer can be pretty misleading if what you really want to know is whether more clothes are being made, for three reasons:

a. Different products. A few years ago, the rich world’s garment factories tended to make relatively costly clothes. Generally the long production lines for suits or cashmere sweaters in Western factories have closed, and what’s partly come in their place have been highly automated knitting plants (making hosiery or other knitted legwear extremely efficiently) or American Apparel-style workshops making Fast Fashion – probably pricier than it would be in Honduras or Romania, but cutting the delay from drawing board to shop rack by vital days. These businesses can seem less productive if you just divide the value of their output by the number of employees: the raw material is a lot cheaper .

b. Different processes. Certainly in Britain, a good number of newer factories operate on a different basis from twenty years ago. Then, many were integrated design and manufacturing businesses – most of whom just don’t exist in the West any more, though lots of design-led businesses still do their customer contact and design in the West, close to their retail customers and the consumers who keep those retail customers in business.

What we’ve noticed springing up are workshops in Britain that really just do contract manufacturing. Where their predecessor would have sold a skirt they’d designed to Top Shop for, say £15, they now sell their manufacturing only to a retailer for a pound or two. They may well get higher net income per skirt: but their apparent turnover is lower. This doesn’t necessarily mean they’r less profitable

c. Different philosophies. The public accounts M&S and Ralph Lauren have given of their attempts to source “English” (M&S) or “American” (Ralph Lauren) looking clothes domestically are similar. They started off, under some pressure from public opinion, to look for suitable suppliers – then found it really difficult to get any with the capacity to meet their likely orders. It’s taking time to build that capacity, and it really isn’t clear how lucrative that niche is ever going to be – to retailer or to manufacturer.

The net result of all this is that  garment making today involves different kinds of facilities in the West from a decade or two back – and different too from factories in Cambodia, Honduras or China. We’ve still not finished the quiet destruction of old-style Western garment factories, and hardly a week goes by without The Source picking up a story about one closing in North America, Europe or Australia, spending a second considering it for publication, then dropping it because where’s the story?

At the same time, hardly a week goes by without a similar story about someone opening up some kind of facility – but we drop that one too, because we can’t see how readers want to know about somewhere making a few hundred garments a week. Add both sets of stories up, and you might get a story going “50 net new jobs in German garment making last month” – but that seems to be followed by “50 German jobs lost” the next.

5. Watching the tide turning

Walk along a beach and it’s often not easy on a rainy day to see if the tide’s coming in or out: you have to watch a spot fixedly for a few minutes to see if it’s getting wetter. Walk along a  beach you’re not familiar with as the tide stars turning and you really have no idea whether it’s turning or not for quite some tome: there’s a lot of wave motion and it often looks dramatic – but it an take an hour in some places to see whether all those currents, ebbs and flows add up to the sea encroaching or receding.

Working out whether there’s abut to be a change of direction is tough enough with tides, which mostly have been coming an going twice a day for millions of years. What’s happening in the garment trade is messier. The mechanics of who does what are changing – and there;s more to it than just a simply pendulum-style swing to and fro

For what it’s worth, I think questions about onshoring are all the wrong ones:

• If you’re a buyer, you want to know what’s the best way of getting clothes made for you, with your customers, your lead times and your budget.
• If you want to be  a supplier, you need to think about what you’re good at
• If you’re concerned with job creation, or with alleviating poverty in your country, you need to think about what kind of industrial activity might achieve your objectives. “Recreating the world as it was in 1950” isn’t a possibility, so trying to do that is pointless.
• If you’re a journalist, you need to ask yourself whether your audience cares. They’re mostly in one of the three categories above: try asking the questions they need answers to, rather than the non-question some consultant asked at the last conference you attended