About The Authors

Sunday, September 17, 2017

Autos can't live without China

mike smitka

I take part in an online discussion forum on Japan that occasionally strays into economics and business topics. One ongoing thread is the potential impact of erecting a "bamboo curtain" around China. A couple posts assert that realigning global production following the elimination of trade with China would not be a big deal. They seem to forget the havoc caused by 3/11 (the Japanese tsunami/earthquake), where damage to a mere two plants impeded global automotive production. One produced a dye essential for certain black/red paints. Red isn't all that popular in most markets, but surely the firms that used it for black lost sales to rival manufacturers who had a different pigment mix. [You don't substitute a different pigment without lots of testing – it's finicky, and the pigment layer may be only 19 microns deep. A different particle size or stickiness and you get paint that looks bad or worse, doesn't adhere. BMW owners won't tolerate peeling paint!] Then there was the Renasas plant in Sendai, which was already in the process of shutting down. Work on their new plant in Southeast Asia was accelerated, and round-the-clock teams worked as well to restart production in Japan. Fortunately there were pretty big inventories and the processor involved was used in more than one function. It did mean certain option packages weren't available, but by so China however would not be just two plants.

...without a global market, it would make much less sense for European, Japanese and Chinese suppliers to set up shop in Detroit...

The quick mistake is looking at the name on the front of the building, and assuming that what could be done by a company in one place could readily and quickly be done by the same company elsewhere. Foxconn, which assembles a big slice of the world's cell phones in Shenzhen in Southern China may for example be a Taiwanese firm, but local operations are very much Chinese. Foxconn does have factories in many parts of the world, but they are not making the same things. Ditto Corning (cell phone glass) and so on. Take down China, you take down everything.

To dig deeper, I provide two examples. One is of cell phones. The other is an extended discussion of the role of China in the global automotive industry. As what I have below is already long for a blog post, and quite frankly it's time to get back to class preparation, I don't provide any numbers for automotive trade. You can find more data and more detail in the China chapter of my recent book, Smitka & Warrian, The Global Auto Industry: Technology and Dynamics, up on Amazon on January 1st. See the link at the top of the right-hand column: as an eBook it's a mere $9.99.

In the case of cell phone production, there’s no place in the US where you could hire 50,000 workers in short order, much less the 100,000+ that Foxconn employs in Shenzhen. We do not have a thousand-plus experienced production engineers and foremen and quality managers and logistics experts and purchasing managers who could be dropped into one place. No other country does, either. We do not have firms that can supply, modify and repair the specialized capital equipment. There are components made only in China, and others that flit back and forth across borders – chip/sensor packaging isn’t necessarily in the same country as either the “fab” or the circuit board assembler. Of course there are also many management systems involved, including how to keep Samsung and Apple from seeing or even hearing rumors about each other’s prototypes. It’s Foxconn that knows how to tweak Apple’s design for volume production, and that has the "creative destruction" (= prototype testing) facilities. It’s not just a bunch of low-skill workers.

That’s even more true of the 山寨机 guerilla cell phone industry. The market for niche cell phones involves a network of small companies and finance specialists, where one company can come up with an idea for a phone. An example is putting two SIM cards in a phone instead of one, not a big deal in the US, but it's important in many countries where different carriers have different service areas, and for people who cross borders frequently. So in the background it's necessary to have close ties to wholesalers who ask for something of that sort. A design house can handle the case, specialist firms tweak the circuitry, others do the software patches, other source the parts and components, and finally a job shop assembles the phones. Of course there’s finance involved, lending to such firms – except for the assembler, none may have as much as a dozen employees – is again on the basis of relational capital. It took years for this network to evolve. Pull out a piece and you have nothing. Now that’s a China-based industry, but it should convey the level of sophistication on the China end of things.

For the automotive case, the posts on the NBR Forum included lists of several global components manufacturers, with the implicit assumption that they had cookie-cutter plants in several locations. Sometimes that's the case, though if China is a big piece of the global automotive pie, there may not be enough capacity. Such plants, though, are no longer the core of what's done in China. Three different strategies governed the entrance of global suppliers into China. Let's work through them.

First, some were “simple” branch plants, which suppliers started building in numbers in the late 1990s, such as for wire harnesses or aftermarket parts or other relatively unsophisticated components. (Caution: what was unsophisticated in 1990 may be a high-tech part today!) This business strategy sought to save on labor costs and export all production. For such technologies there are often only modest economies of scale, and plants are scattered in a number of countries. For wire harnesses more are in the Philippines than China, but some production remains in Mexico. (None in the US, except for low-volume high-voltage harnesses and small-lot production for pre-production vehicles.) Assuming a location with an adequate labor force could be located – such plants quite often employ thousands, and need a hundred or more supervisors and assorted other skilled managers – production could be shifted in a few months. But in the interim automotive production would drop to zero, and for some time quality would remain low. But hey, who needs windows that roll up and not just down, or a transmission that shifts properly?

Second, another pattern was to serve as a regional supply base, but with excess capacity for what in the early 2000s were the bigger markets of the EU and NAFTA. In one particular case with which I'm personally familiar, the capital equipment and top supervisors were from Germany, but there task was to replicate a plant in Virginia (which happened to be the company’s first location for a new high-tech component). Such production in China might not be much affected by US policy, assuming they could continue to import certain components. If there was a supply crunch, it would initially be production in Thailand, Korea and Japan that would shut down. But that would be awkward for a global supplier, so surely a portion of what remained of their global production capacity would be redirected from the US to keep them running, at the expense of producers in the US.

Now in this case there was initially no particular human capital on the Chinese end, but physically moving the equipment and then getting a plant up and running again would take 9 months to a year (that's what it took to set it up in China in the first instance). Of course depending on the legal framework – there was no legal mandate for joint ventures for automotive components, but many firms chose that route – their joint venture partner might not let them ship out the capital equipment, and the Chinese government would surely discourage such, even if you could get paperwork on the US end to permit transiting the bamboo curtain. In that case, ordering new equipment and tweaking the line to get it running would take much, much longer. My sense is that it would prove impossible to do in under 3 years, because of the backlog that would arise as everyone tried to place orders simultaneously. And it might take longer – much of the underlying tool and die capacity is now in China (both machines and skilled tradesmen). In any case, when there are only two-three global plants supplying a particular component for global gasoline engines programs, there would be massive disruption. And for things such as fuel injectors or valves or bearings, engines are designed around a particular firm’s component – no double-sourcing, so the engine would have to be modified to use a rival’s fuel injector or similar component, and then re-certified. But by and large rivals have similar footprints in China, so there’d be no workaround in going to another European/American/Japanese firm.

The third pattern is now the most important. China is the world’s biggest auto market (25 million units of light vehicles, more than either the EU or NAFTA). Many models launch first in China or are variations specific to China, and 40% of sales inside China are of vehicles engineered and assembled by domestic Chinese firms, not “global cars” from VW, GM (the two biggest car firms in China) and other foreign auto companies. To serve these customers, Chinese and otherwise, global suppliers have significant operations in China. Three very large global suppliers ($8-$20 billion sales) I’ve visited recently now have major engineering centers in Shanghai that are part of their core global R&D operations. Over the past decade all three have shifted towards local specialization. As very large firms, each has a dozen-plus R&D centers, located in at least a half-dozen countries. At one time they'd each do a little bit of lots of things. Now each center focuses on a specific component or technology as a "global center of excellence". As a reflection of that strategy, R&D in Shanghai is integral to global operations. Indeed, one of these firms (quietly) moved divisional HQ operations there, too, planned over a number of years to coincide with the retirement of several key people in the home country. First the senior person worked in Shanghai for a couple years, and then his designated Chinese successor worked at divisional HQ in the home country for a couple years, with others less senior moving back and forth for shorter stays (6 months or more) over a period of years. To repeat, it took years to set up a new R&D center, to build a team who could work with each other and with the rest of the global enterprise.

One driver for this particular firm was the ability to hire engineers in China. Another, though, was that China was the division’s largest and most profitable market. Indeed, today GM’s main engineering center in Asia is at PATAC in Shanghai, with over 2000 engineers and 15-plus years effort at building up teams (subcompacts continued to be done by Daewoo in Korea, but more and more of the next larger platform is done in China, not just the "top hat"). With GM’s sale of their European operations (Opel/Vauxhall) to PSA, Shanghai is now central to GM’s ability to design vehicles (again, platforms, and not just “top hats” for the Chinese market). To my knowledge other firms (eg, VW) are less China-heavy in their engineering, particularly the Japanese and Korean firms, but all have major operations there.

In sum, the auto industry is today tightly integrated on a global basis. You can’t pull out one piece from any of the 3 major centers (NAFTA, EU, China/Japan/Korea), and major suppliers typically have additional pieces of their global R&D footprint in Southeast Asia, India and Brazil. Factories are not mobile, engineering centers less so. Building a bamboo curtain between the US and China would shut down the US industry, and not just for months. It would be particularly ironic in that a couple initial studies suggest that the role of Detroit as a global engineering center is increasing – it’s no longer just a regional NAFTA role. That includes Chinese suppliers locating R&D centers in Michigan (I’ve visited one). Of course I’ve also visited factories in the US that ship a significant part of their output to China. Without a global market, it would make much less sense for European, Japanese and Chinese suppliers to set up shop in Detroit.

But it would be a huge hit to many industries, not just autos. It wouldn’t make America great again, and if the impetus came from the Washington, it would lead global firms to not put anything of value in the US.

Thursday, September 14, 2017

Sugar tariffs

Mike Smitka
reposted from my Econ 102 Macro Principles blog

First, here are data to help you remember that import prices are not everything. Prices effectively triple between the wholesale price sugar farmers such as the Fanjul brothers receive, and the price you pay in a store. A big baker will pay something much closer to the wholesale than the retail price - if you buy by the train car (not the truckload!), delivery costs per pound are very low. The price gap between Brazilian and US sugar is about 40%. So even if the tariff was eliminated, the price would only fall by about 6¢ wholesale, and by about the same retail. How eager would consumers be to fight over 6¢ per pound? Even though I do some baking, it takes me over 1 year to use a 5 lb bag!

Prices1980's Average2013
Brazil Raw Sugar Price - 14 cents
U.S. Raw Sugar Price22.16 cents20.46 cents
US Wholesale Refined Sugar Price27.06 cents27.22 cents
Grocery Store Refined Sugar Price33.59 cents64.32 cents

Source: US Sugar Prices - American Sugar Alliance, IndexMundi for the Brazil price and XE.com for the Brazilian Real / US$ exchange rate.

Then there is the political economy. The first sugar tariff dates to 1789. Protection was strengthened during the Great Depression with the 1934 Sugar Act, including policies to raise farmer's incomes while at the same time using rationing (esp during WWII) to avoid raising consumer prices. That Act expired in 1974, but in view of his pending election campaign President Ford tripled the import tariff. Presidents Reagan, and George HW Bush also implemented protective measures, while George H Bush was able to veto the Farm Bill in 2008 knowing that Congress would (did!) override his veto. Both Carter (a farmer!) and Clinton (who grew up in a farm district) turned down policies that would have increased sugar protection. There were no changes under Obama. See the Coalition for Sugar Reform for details.

In both the EU and the US the sugar that you buy in a store or get in a restaurant sugar packet is beet sugar. Production dates to the 19th century, when new cultivars with higher sugar content made it a profitable crop, first in Europe and then in the US. In Japan sugar beets were cultivated in the 19th century, but then the expansion of the Empire southward to Taiwan (1895) led to more sugar cane. Today 2/3rds of Japan's sugar is imported, and there remains enough near-tropical land that 20% of domestic output is from sugar cane. Sugar beets are still the overwhelming domestic source. Oh, and that's because of tariffs in all 3 regions.

Sugar growers are a powerful lobby. The Fanjul brothers own over 150,000 acres of Palm Beach County, Florida. That's a potential swing state in national elections (do you know the term "hanging chad"?). Both are politically active – one a Republican, the other not by chance a Democrat. In the Midwest corn farmers are a potent lobby, and in a handful of states so are sugar beet growers. The Senate thus has a big block in favor of agricultural protection. This political economy – enough farmers in enough electoral districts that their vote is essential – is true in Japan, the European Union and NAFTA. In the former two, unless I'm mistaken, direct and indirect farm subsidies are greater than aggregate farm income. The CAP (Common Agricultural Policy) is the single biggest item in the EU budget. Through the good fortune of geography agriculture in the US is inherently more productive, so our overall subsidies are less. It is nevertheless the sector where trade is most constrained by a web of quotas, tariffs, subsidies, cropping restrictions, loan programs and tax breaks.

...trade in sugar could have sweetened the Doha Round...

This matters not just because it was a barrier to the continued integration of the European economies – agriculture has been the biggest sticking point in the various EU expansions, and with attempts to create greater policy cohesion among existing members. On the global trade front, the Doha Round was intended to extend the WTO to cover agriculture, which largely left the sector untouched, other than requesting countries adopt tariffs in place of quotas. [See the textbook for why tariffs are far the better means of protection.] But the politics – dispersed consumers but a geographically concentrated industry, big enough to affect a significant minority of electoral districts in every high-income economy – meant no progress was made. Without progress, however, there was little "benefit" for negotiators from developing countries to take home. The talks have effectively collapsed, and there is no near-term ability to renew them.

...comparative advantage implies we benefit from unilaterally making importing easier...

One real challenge is that several large developing countries are themselves facing pressure to subsidize farmers. China may not be a democracy, but the majority of urban residents have close relatives back on the farm. Keeping urban areas quiet requires making life better in the countryside. Most Chinese farmers now receive cash subsidies. Ten years ago they might have gladly offered many "concessions" to the US and Europe and Japan in agriculture. Now that dynamic is changing. We all lose. Comparative advantage implies we benefit from unilaterally making importing easier. That includes agricultural products. If we're concerned with issues of urban poverty, as is the case now in China, then agricultural imports are particularly beneficial. But because of the politics of "reciprocity," the agricultural sector impedes continued global negotiations. That would be fine if we could rest on our laurels. However, economies are not static, and so areas where all would benefit (healthcare-related sectors, under the rubric of "intellectual property") cannot be addressed because the horse-trading, multilateral most-favored-nation process of global trade negotiations has fallen under the weight of agricultural lobbies. The very real fear is that trade deals are like bicycles: unless they keep moving forward, they fall over and retrogress. Trade in sugar could have sweetened the Doha Round. Politics nixed that.

Tuesday, September 5, 2017

Automotive Employment Decomposed: New vs Used Car Dealers

Mike Smitka
...employments is centered in new car dealers and parts manufacturers...

Scott Wood of Carvana asked me about the increase in employment in automotive retail. With a bit of poking, I found that the Current Employment Survey (CES) run by the BLS provides fairly detailed breakdowns. Now I very strongly suspect that in the survey responses, employees at a new car dealer are not subdivided into those who work in service vs finance vs new sales vs used sales. All are likely classed by the main  business category of new car sales. Still, this provides a starting point. As it happens, while used car dealership employees have risen faster, the bulk of the increase in employment stems from the rise in new car dealership employees. While I'm at it, I also am posting (i) car vs parts/accessories/tires and (ii) assembly vs parts manufacturing. As expected – well, at least by me! – people who work in parts plants vastly outnumber those at "the" car companies. The latter get the political attention, but they're not the ones who "make" cars, they just put together the pieces.



Sunday, September 3, 2017

Quick Update: US Labor Force Graphs, including autos

Mike Smitka

Here is an overview of (i) unemployment across the Great Recession and the subsequent recovery, highlighting U-6 "total pain" versus U-3 "headline unemployment." U-6 peaked at 17% of the labor force. It doesn't reflect those who dropped out but weren't "discouraged" or "marginally attached" by the BLS – currently, as per the graph below, that's still about 2% of prime-age workers, and 4% for age 20-24 workers. See the graph on the right for details.

I also calculated a "normal" level of total employment, using the relatively constant age-specific rates in the period prior to 2007, but adjusting the total for demographic changes, particularly "boomer" retirement. That's the graph immediately below. By that measure we're still a year away from "full" employment, assuming no slowdown. We have however added 12.2 million jobs, relative to population growth, since the trough of the Great Recession.

Finally there's the auto industry. On the retail side employment is at a historic high. However on the manufacturing side, despite robust domestic production, the industry employs about 130,000 fewer workers than at the onset of the Great Recession, or about 12% fewer. That is, one in eight jobs vanished. Why? – productivity. This reflects a long-term trend, it simply takes fewer people to turn out a vehicle today than in 2006, primarily due to more efficient parts production, because that's the sector where nearly 3 out of 4 workers in vehicle manufacturing are located.

Sunday, August 27, 2017

Harvey to clear the industry's excess inventory

Mike Smitka

...the Thu 31 Aug Automotive News morning 'cast gives 300K-500K vehicles, but unfortunately the storm lingers...I hope I am wrong on the 1 million figure...

Harvey will cost insurers billions. Part of that will go towards the purchase of replacement vehicles. Macabre, yes, but that's a timely bump for an industry that has been grappling with excess inventory. Houston proper has 2 million people, and extend out to Harris County and you have almost 4.6 million. Immediately adjacent counties add 1.2 million more. At present the cars of a significant slice of that population are underwater, and will remain so for another day or more. That means they're totaled. Across Texas as a whole that certainly means 1.0 million vehicles, and perhaps more. A week ago the industry was awash with excess industry. Now it's not.

...for the auto industry Harvey is not a disaster but a turn of good fortune...

As sales fell over the past 9 months, the US industry built up inventory. A healthy level is around 60 days of sales. But by June 2017 dealers had 74 days worth of cars and light trucks on their lots, and pared that level only modestly to [Ward's] 69 days in July, or 4.2 million units. Of course this hides significant variation, with GM holding 104 days (0.98 million units) and Ford 76, while Subaru had only 40. Now much of the excess is in sedans, as demand shifted towards light trucks – the latter sold at a 10.8 million unit rate in June, while sedans sold at only a 5.8 million rate. The industry is responding: over the past year, NAFTA light truck production was [Ward's] up 4%, sedan production down 11%. That won't be reversed in the short run – suppliers can't spew out extra parts overnight – so sedan inventories were already set to keep falling.

Now Texas is truck country, as CNN's videos of Houston neighborhoods indicate. So Harvey won't be as big a boon for sedan sales, though Texans who find they're underwater on their loans may be forced to downsize to a mere car. The impact remains: 1.0 million units represents 16 days sales, and will help the industry draw down inventories to closer to 50 days. For the auto industry Harvey is thus not a disaster, but a turn of good fortune.

The US Bureau of the Census tracks the dollar value of automotive retail inventories. The most recent data – $68 billion in June 2017 – do not separate detail new versus used vehicles, or parts and tires versus vehicles. In contrast, the biggest purely retail sector, drugs and druggist sundries, comes to only $60 billion.

Saturday, August 19, 2017

Auto Inventory Cycle Adjustment: More Pain Ahead

Mike Smitka
Washington and Lee University

During the extended housing boom of the early 2000s, consumers used home equity lines to go on a consumption binge. Light vehicle sales remained strong, while the model mix richened. Then the housing bubble reached its peak (and gas prices rose). The good times were over, and at the February 2009 trough sales fell to that of the trough of the 1981 recession. In relative terms the situation was far worse, because in the interim the US population grew by almost 80 million. Relative to employment, sales were a full 15% below the previous post-1976 (before which consistent data are unavailable).

Cars are durable goods, and even though the average vehicle now lasts 12 years, they eventually need to be replaced. Meanwhile the US population continues to increase and incomes have recovered (though for most Americans, not risen). From an excess of vehicles per household going into the recession, the number fell below what households wanted, and has now recovered. In February 2007 SAAR was 16.7 million units; that level wasn't hit again until March 2014. Fueled by low interest rates, longer loan maturities, and high used vehicle prices, sales crept back up. Indeed, Paul Traub of the Federal Reserve Bank of Chicago argues that they have been higher than sustainable.

...it's payback time...

It's payback time. Over the past several months the Light Vehicle SAAR (seasonally adjusted annual rate of sales) fell 7.5% from its December 2016 peak of 18.051 million units. This represents a drop of over 1 million units. Now  inventories rise and fall as a normal response to short-term swings in sales. With a steady fall, though, they mushroom, and mushroom they have. (Thanks to Paul and his May 2017 presentation to my W&L auto seminar for the graph on the left.) Now that the drop appears to be more than a transitory blip, it's time to bring inventories under control and pare production.

The supply chain is like a snake....

The supply chain is like a snake. If it's to eat bigger prey, it has to bulk up, and that takes time as suppliers rehire and otherwise add capacity. (This is on top of the normal investment to replace tooling for old models with that for new. Given the hit balance sheets took from the Great Recession, they also had to repair balance sheets to add capacity.) Going into the recession, production peaked in June 2007. It took over 6 years, until November 2013, to reach the previous peak.

....that's consumed too big a meal.

Now it’s like a snake that’s just consumed too big a meal: it's sluggish, because it takes time to get it out of its system. The Production Index hit 131 in October 2016, and bounced around that level through April 2017. It’s since fallen 7.6%. But that at best brings production in line with sales. It doesn’t pare inventories. That will require either an uptick in sales (fueled by assorted incentives) or a further cut in output. GM for example has chosen the latter, with extended summer vacations – not that they aren’t discounting, after all they can’t totally ignore price cuts by competitors. Output will surely fall more; it’s better to overshoot a bit, given uncertainly about whether sales will fall further. And today labor is no longer a fixed cost. As per Econ 101, the flip side of higher marginal costs makes it relatively more profitable to cut production than prices.

That leaves two questions. One: what is the stable level of output? I’ll save that for another post; too many graphs already! The second is employment. Here the bottom line is clear: productivity continues to rise, a trend visible as far back as data are available. Industry employment is certain to fall, and on a permanent basis. As a dismal scientist, I close with that graph, and with that of auto industry manufacturing employment.

I leave out many pieces. One is that while manufacturing productivity increased, that's much less true of automotive retail: employment there is at a historic peak. Then there are the finance, household formation, population data, household formation, depreciation and other pieces of the peak sales story that Paul Traub sets forth. I have not tried to put together the last couple months of inventory data, or compiled the various media reports of reduced production at GM factories, eg Lake Orion where the slow-selling compact Sonic (and the Chevy Bolt EV) are produced. Finally, while I can't provide any details, I do participate when my schedule permits in the monthly sales analysis roundtable that BWG hosts for their clients. My reading of these various sources is that at a more detailed level, some manufacturers are slower in responding than others, while the impact of the sales slowdown is uneven across manufacturers/brands, with for example the shift towards light trucks amplifying the impact on producers that are sedan-heavy, such as some of the Asian brands.

Thursday, August 17, 2017

Real Yield Curve

Mike Smitka

I look at data of various sorts, often out of mere curiosity. One ongoing puzzle is the evolution of interest rates. I've posted graphs of nominal rates, and implied future rates. Below are similar graphs for real rates, as calculated by Treasury using inflation-adjusted bond (TIPS) yields. The first are the real rates at various maturities. (That's the graph on the left – click to expand.) I then use the difference between yields at different maturities to calculate the implied future interest rate. (Duh, that's the graph on the right.)

At one level these look sensible. We see that longer maturities have higher yields. We see craziness in fall 2008. But real rates remain low, around 1% into the far future. They look sensible, but they don't make sense.

...[they] look sensible, but they don't make sense...

One explanation might be global excess savings, what Ben Bernanke termed in 2005 as a global savings glut, driven by countries where individuals and firms are building up financial assets as their populations grow but where they've run out of sensible domestic investment possibilities. That requires financial outflows (and for that to happen, a trade surplus). And on the US end we do indeed see the flip side of trade deficits and net financial inflows. (Again, you can't have one without the other.)

After all, conceptually the return on investment ought to be higher in labor-abundant, poor economies. But I find it hard to believe that story, when we find it going on year after year. OK, many economies have unstable politics, which might make would-be investors cautious. So savings could pile up. But we don't see sharp breaks that might be consistent with that story, that is, with the ebb and flow of politics. Indeed, if that sort of uncertainty is key, we ought to see a huge Trump bump, because politics in the US now looks zany. Policy change is precluded by political infighting and the failure to appoint staff across the Federal government. We have no ability to address a crisis, much less attend to long-run challenges such as putting in place a true healthcare system, improving eduction or setting our fiscal house in order. The bottom line is that I don't see any such effect in the data.

The other story is secular stagnation: that despite the hype and self-promotion of Silicon Valley and the venture capital vultures who circle in search of easy feeds, there just isn't much happening. Cheaper taxi services – Uber, for example – just aren't working out. And from an economy-wide perspective it's hard to see an economic revolution in that, or better dating apps, or in receiving streams of 140 characters. In the $20 trillion dollar US economy, there's room for a lot of successful new $50 million businesses, but again, from a $20 trillion dollar perspective they are chump change. Robots? – I've been visiting factories for decades, automation is already widespread. The low-hanging fruit has been picked, and "hard" goods are only a 20% slice of our consumption. Artificial intelligence? The Executive Director of our small local United Way of Rockbridge worked in the earliest AI initiative at Stanford in the 1970s. Algorithms aren't new, and the cost of computing has long been near zero. Again, the low-hanging fruit has already been picked.

...the other story is secular stagnation...

My preferred explanation then is that (to borrow the title of Marc Levinson's most recent book), the transport, information and energy revolutions that exploded after 1800 represented An Extraordinary Time. That era has now come to a close, and henceforth we will no longer see the productivity growth that underlay the rise of the US. To borrow again, this time from Robert Gordon, this represents the Rise and Fall of American Growth.

However it's not just an American story, something emphasized more by Levinson than by Gordon. It's the story throughout the OECD economies, Europe and Japan and now even China. We can pray that in the next 3 decades South Asia and Africa converge on the developed economies through their own growth miracles. But for now they're too small, and too isolated financially, to offer a solution to the secular stagnation that we see in the US.

Disclosure: I'm using Levinson's book this fall in my 2 sections of Econ 102, Principles of Macroeconomics. I listened to the Audible recorded book version this spring, and am now reading the hard copy one. Gordon is also available as a recorded book, but given his prolific use of data, I can't imagine consuming it without his graphs in front of me – and I do my listening while driving.
For completeness, Bernanke reviewed his original savings glut story 10 years later, in a 2015 Brookings post. Here too is The Demise of U.S. Economic Growth, a modest-length paper that covers the stagnation themes that appear in the latter 100 or so pages of Gordon's 750+ page book.

Thursday, August 10, 2017

Auto margins and disruption

Just a quick cross-reference in line with the previous post a great quote:

“TSLA, for all Musk's gift-wrapping abilities, is primarily a manufacturer, and history has shown us that the economics of the automotive industry are crap. ”

Graham Osborn, Contributor 10 Aug 2017, 09:51 AM

found in the comments to "Some Thoughts About Tesla's Latest Bond Offering" posted by Montana Skeptic at Seeking Alpha on August 10th.

Wednesday, August 2, 2017

No margins, no disruption: the New Mobility Challenge

Mike Smitka
Washington and Lee University

If you want to disrupt an industry, you need to pick one with fat margins. That's the real challenge for the entire family of "new mobility" models, and more generally for "disruptive" technologies in the automotive footprint. It's one thing to be able to arbitrage regulated monopolies, which is what most incumbent cab companies are. But that only works if your strategy provides you with entry barriers, and the monopolies you break into earn piles of money. Local cab companies are however perfectly capable of developing their own cell phone apps, benefitting from the high ratio of "locals" in the customer base. That's true up and down the automotive value chain: all operate with thin margins. They are thus NOT ripe for "disruption."

...why would anyone want to try to disrupt a market with thin margins?...

At base, motor vehicle manufacturing, distribution, repair, and the final market of transportation services have few monopolies. Yes, if you want to develop the next-generation diesel engine there are only two players who are capable of the underlying material science and have the ability to manufacture with the requisite quality in the requisite volumes. (These two are Federal Mogul and Mahle.) There are however multiple firms capable of "mature" piston production. The same is true for every other component that I can think of: there are often a small handful of "leading" firms but there are also commodity producers of older technologies. The only way to preserve margins is to keep innovating.

Furthermore, it's a complex chain. Assembling vehicles, as Tesla is learning, is the easiest part – and they don't yet have their Model 3 assembly line up and running, despite having more employees than the normal volume assembly plant. But that's the tip of the iceberg: they have yet to solve the national distribution challenge in a cost-effective manner. Assuring that component supply lines can meet your production plan, and that you can take tradeins, provide financing and repair vehicles quickly, all that takes a lot of people and a lot of physical assets. Experience helps, too. If you're to grow rapidly, those have to be in place beforehand. Some resources can be borrowed, including the financing, as happens in the "traditional" franchised dealership system (though Tesla has decided not to do so).

It took over a decade from their establishment of a solid footprint for Honda, Toyota and VW to succeed – and as VW has demonstrated, laurels can be lost. After all, in the early 1960s and again around 1968 VW was THE import market in the US. They're now barely a player. The VW vertically integrated, single model mass production strategy worked for a while, as did Henry Ford's monomaniacal focus on the Model T and nothing but the Model T. Neither VW nor Ford were ever able to lower costs sufficiently to develop a sustainable advantage against the evolving products of rivals. In contrast, Tesla is a high-cost producer with a high-cost distribution system. High costs aren't an insuperable barrier if you aim to break into the premium car segment, but even then you have to keep renewing your product. High costs don't work if you aim for the high volume, middle-segment of the market.

Ditto Uber on the downstream transportation service end: taking over the taxi business would be great, but only if the underlying business was unusually profitable. But it's not. There's little room for cost reduction – materials, labor, overhead. It's not as though existing taxis are new and drivers well-paid. There not much room to cut economic costs, even if in the short run costs can be shifted to unsuspecting parties (Uber's owner/driver contractors). The only way Uber can provide reliably better service than incumbent Yellow Cabs is to have higher peak load capacity, with the requisite assets of vehicles and drivers. Superior service at a comparable price lets them grab market share, but they've not expanded the market. And without a bigger market they can't sustain their high-cost strategy. A cell phone app doesn't lower the cost of a car, or lower the minimum wage paid by alternate jobs. To reiterate: creating a big taxi company does not provide a route to healthy long-run margins.

...creating a big taxi company does not provide a route to healthy long-run margins...

So don't be fooled by the apparent ease of entry by disruptors. Yes, Tesla can draw upon the base of automotive suppliers to launch a car, something that would not have been possible in the more vertically integrated world of the 1960s. Numerous Chinese domestic players have done the same, and one or two may even survive if not thrive, the Geely's and Great Walls. But Tesla can't rapidly expand in the mature markets of NAFTA and the EU, absent a revolution in battery costs that decades of leading-edge chemistry research has yet to deliver. They can't compete in costs against the still-improving technology of the internal combustion engine. They can't compete by eroding the fat margins of incumbents, because margins aren't fat.

Autonomy is much the same. The suite of sensors need to be integrated into a vehicle, software integrated into the actuation of steering, braking and so on. Everything then needs to be tested. Car companies are good at that – that's why they're called assemblers. Many of the pieces are already in place, but consumer acceptance is still uncertain, and the only way average transaction price can rise is if sales fall: the average new car purchaser can only finance so much, given stagnant incomes amidst a driving population that is virtually flat. In other words, implementing a costly technology won't help margins. Indeed, it's already gotten the CEO of Ford fired.

...govt policy can disrupt the (auto) market, but not Tesla, not Uber, not Waymo...

There is one exception: government policy can disrupt the market, by enacting direct and indirect subsidies (such as California ZEV credits or safety mandates). But not Tesla, not Uber, and not Waymo. What is amazing is that, given automotive margins, they purport to have "disruption" as their strategy. It may work on the stock market, at least for a while. It won't work in the automotive market.

Sunday, July 9, 2017

Germany will be first in EVs – but don't invest in VW!

Mike Smitka

The first big market for EVs will be the EU, not China and certainly not the US. This is not common wisdom: after all, isn't Beijing pushing EV technology, including forcing firms to buy batteries from "domestic" players? And then there's Tesla in the US. Of course in terms of hype no one can beat Elon Musk (and in semiconductors, NVDIA). But against global production of 90+ million units Tesla's puny 80,000 units in 2016 is on the order of the reporting errors in global sales data. They aren't disrupting anything. Instead the disruptor will be VW, adding to the presence of Renault and Nissan.

This was one of my key takeaways from the 25th GERPISA conference in Paris in June, garnered across 4 days of R&D center visits, presentations and conversations. [For fellow researchers, next year's GERPISA will meet in June in Sao Paulo Brasil.]

On the regulatory side, EU fuel efficiency rules cinch tighter in 2020-21, and thanks in part to VW the test cycle standards – currently the New European Driving Cycle – will be tightened to better reflect actual EU driving patterns. (Currently, for example, back seats can be removed and the alternator disconnected, while accelerations are unrealistically slow and idles too frequent.) That will make it extremely difficult to meet CO2 mandates. Reconfiguring diesel systems that passed due to "cheats" will lower their efficiency and raise their costs. Keeping to a diesel-centric strategy won't work. This will be a particular challenge for VW. At the same time, France was already moving to limit the presence of diesels in urban areas.

Hence BEVs (battery electric vehicles). VW just announced they've frozen the design of their first model, to be launched in 2020, and will thereafter start turning over their fleet to BEV models. For now they are basing their model on the underlying architecture of the Gulf. However, they are working on new architectures that will facilitate flat battery packs – we at GERPISA visited Renault's vehicle competitive teardown facility, and that's one of the things they look for in their analysis of new platforms.

Then there are batteries. VW will not make their own cells, reflecting both a lack of internal capabilities and to maintain the flexibility to shift their sourcing as cell technology advances. But they will make their own modules and packs, to better control weight and safety. Those are also heavy and bulky, so doing this in their assembly plants makes sense. Ditto Daimler.

Now VW is not alone, even if their diesel-heavy strategy and legal issues places them in an awkward position. Renault continues to update the Clio (and Nissan the Leaf), and both are poised to launch additional vehicles in line with demand. Meanwhile, the various EU members are building out base infrastructure. France already has charging capabilities along major highways; Norway (a small market!) is already 30+% electric. The government role is central, because infrastructure is expensive, and needs to be pervasive. Private efforts suffer from chicken-and-egg issues. For-profit charging ventures inevitably focus on dense areas. But for consumers to make a BEV their sole vehicle, national availability helps: there's always that trip to the beach, or a quick weekend getaway to the countryside. Such locations wouldn't generate enough business to make it pay to set up charging, at least early in the rollout of BEVs. But their presence facilitates market expansion.

After hearing pieces of this story from multiple people, I'm now convinced that BEVs will happen sooner than I expected in Europe. Key is that they will now provide a better value proposition relative to diesels, which intrinsically sip rather than gulp fuel. But batteries remain expensive, so that's not good news for VW. To meet efficiency standards they'll need to sell a lot of EVs. To do so on a competitive basis against the standard gasoline vehicles from other manufacturers means VW will lose money on each one it sells, to the tune of billions of euros as volumes rise. So VW has solved the subsidy dilemma: economies of scale and competitive costs can't be achieved without somehow getting BEV volumes up. EU incentives aren't enough; paying for a test fleet is one thing, doing it for millions of cars is another. VW incentives may well make the difference – indeed, Volkswagen is betting the company that that will be the case. In the interim, though, VW will not be particularly profitable. That interim will extent until 2027, at which point costs will fall, or VW will. Don't invest in VW!

...for the next 10 years, don't invest in VW...

For the US, the current administration is hostile to environmental issues, and is promising to roll back fuel efficiency standards (which will benefit luxury car makers, primarily German, and hence is not particularly helpful to domestic manufacturing). At the same time, the current Congress seems unable to pass any legislation, much less focus on such long-run issues as energy policy. That may change over the next couple elections, but for the time being there will be no national infrastructure policy in the US, and without that BEVs will remain a niche product. Yes, the fines VW is paying will be used to build charging infrastructure, particularly in California. But there's no national vision behind it, only an attempt of VW to get some modest indirect benefit out of their fraud settlement.

Then there's China. For the same budgetary reasons as elsewhere, Beijing will rein in central government subsidies by 2020. Yes, they want electric vehicles, they want to have a presence in new technologies. No, they don't want to pay for it. Economic nationalism fails as a policy when it requires spending serious money. Meanwhile Panasonic and others are building plants there, suggesting that a "Chinese companies first" stance will work no better there than it has in the passenger car market, where VW and GM are #1 and #2. In fact, China has already relaxed its stated limitations on foreign firms setting up new joint ventures: new ones are fine if they're for EVs.

...China may be the largest single BEV market, but it lives on subsidies and those are fading...

One presentation at GERPISA made that clear, using insurance data (not registration data, known to be misleading due to false registrations by various corrupt "car companies" that let them pocket government EV subsidies without actually making vehicles). As everywhere, BEVs are a small share of the market. Once the data were disaggregated geographically, it turns out that in many regions there are essentially no sales. In contrast, in some metropolitan areas BEVs appeared quite popular.

Ah, but the details! In such cities BEVs qualified for a license plate at no cost and with no wait, while buying a plate for an ICE (internal combustion engine) car requires entering a lottery with average wait times of 2 years, and paying up to $12,000 in fees. The BEV exemption thus provides a huge implicit subsidy – but the BEV quota is finite. In one major city the quota of 50,000 was quickly filled. Total BEV sales: 50,800. Absent subsidies, there's as yet no market for BEVs in China, and these subsidies are not set to expand. At the national level they are already shrinking, and there's no systematic rollout of charging infrastructure. Instead, we have owners dropping wires from the window of 5th floor apartments to let them charge their cars. [There were a couple news stories last year on this, if I can find the links I'll edit them into this post.] So while China may be the largest single BEV market, that's due to a confluence of idiosyncratic and transient subsidies, not to effective policy or consumer demand.

So in the end the EU will be first. What is not yet know is whether battery prices will fall sufficiently to become competitive with ICEs. After all, ICE costs are also falling – a decade ago, did anyone foresee "real" cars running on 3-cylinder engines? Lower component count aside, those save weight in a manner that batteries don't, and so allow downsizing in suspensions, frames – everywhere! – with attendant cost reductions. No current commercial battery technology can offer those indirect savings. Even if things go well, it will still be well into the 2030s before EVs will comprise half of sales in development markets. By that time biofuels will also have advanced. My own belief is that in 2030 we'll see a variety of drivetrains coexisting in the global market, with variations in dominant power sources from country to country.