Advanced economies are in the throes of what can, without exaggeration, be called a global supply chain crisis. Car makers unable to meet demand and even in many cases produce full featured cars due to chip shortages. Shipping rates spiking then collapsing due to a lack of containers and truckers at ports. Petrol and food shortages in the UK. Warnings even in the US to buy Christmas gifts now because they might not be available later. And there are also increasing reports of shortfalls of drugs and medical supplies.
The original sin is optimizing businesses for efficiency at the expense of having slack to contend with unexpected developments….anything from the factory of a key supplier blowing up to the macro-level disruption of Covid. More buffers means more resilience. Having them doesn’t mean all bad outcomes would have prevented. But we would have had fewer deficits and many would have resolved faster. In other words, companies around the world set out to increase collective tail risk in pursuit of profit.
How did this result come about? A lot of commentators want to point to mechanisms like globalization or consolidation, as opposed to why did executives decide to pursue those fads?
I went to Harvard Business School from 1979 to 1981. Managing global supply chains, outsourcing and offshoring were not yet A Thing. About 40% of the class had worked in manufacturing before coming to HBS and at least that many had undergraduate degrees in engineering. We had no Wall Street analysts (those two-year programs were just getting started). In my nearly 90 person section, there were only two bankers. Both had applied to HBS not long after completing their commercial banks’ two year credit officer program.
So having come in to the world of commerce at the tail end of the old regime, here are the management practices that produced companies with sprawling supply chains as important parts of their operations, as opposed to old industrial model of internal integration. I’ll try to lay these developments in rough time order, but they obviously overlap.
Just in time manufacturing. In the later 1970s, the business press regularly described how Japanese and German manufacturers were eating the lunch of their sclerotic American competitors. Carmakers were Exhibit 1. The experts conceded that the foreign insurgents did have the good fortune to be operating out of newer plants, but a big part of the problem was outdated practices and bad relations with workers 1 (the success of the NUMMI, where Toyota took over the management of a plant with a reputed-to-be-terrible workforce and got it performing at over the median for quality standards for Toyota operations proved the problem was much more “bad management” than “bad unions.”
Japanese “just in time” manufacturing amounted to turning a problem into a virtue. The Japanese car-makers of the 1960s couldn’t afford to carry big inventories, nor was it prudent for them to invest in larger facilities just to have inventory on site. The Japanese and Germans made deep inroads into the US market in the strong dollar period of the early 1980s, after Volcker relented in 1892 on this high interest rates, but they never gave it back when the dollar normalized.2 Japanese methods, most of all “just in time” management, were widely adopted in the West.
Outsourcing. Electronic Data Systems pioneered in outsourcing before it was even called outsourcing, acting as a “facilities manager”for big corporate and government IT departments. EDS would often buy the data center, taking on the employees, and run the operation for a fee. Even though companies had been selectively outsourcing before (for instance, the legal department hiring outside law firms for litigation or other specialized or one-off work), the EDS approach was more radical, in essence taking over an entire, usually pretty large, and usually mission critical corporate function and entrusting it to outsiders. The fact that EDS was competent and much better at optimizing the use of then very price IBM computers than their customers were meant the deals were win-wins. I don’t recall ever reading of an EDS deal going bad; one assumes they either didn’t or EDS would do what it took to make amends.
GE envy/imitation. During the Jack Welch era, the business press and management gurus held up GE a paragon of American manufacturing.3 One of the much-touted GE principles was not to be in a market unless they could be number one or number two. This rule became the fodder for many management consultant studies urging either divestitures or acquisitions to achieved the supposedly magical market position (never mind that market definitions are very fuzzy). Even though consolidation plays are hard to execute and seldom succeed (every study ever done has found that between 2/3 and 3/4 of all acquisitions fail, as in do not deliver value to the buyer), unless a deal was obvious misconceived, Wall Street would generally give a prospective buyer’s stock a pop.
In other words, the fad for consolidations can’t be attributed to LBO artists, or at least not directly. They were making financial plays in the 1980s, taking over undervalued, overly diversified conglomerates and selling the parts for more than the price of the former whole.
However, weak management sought to better defend their castles by building higher walls and deeper moats in the form of poison pills and “golden parachutes” for the executives, in theory to make taking over a company prohibitively expensive. In practice, they served to grease more deals, since the management of the seller got big payoffs, and the management of the buyer could expect raises for now running a bigger and more complicated operation.
In banking, the regulators drove consolidation. Conventional wisdom was that 16,000 banks (the number as of 1988) was way too many, and Canada needed only 5, so the US could and should shrink the number a ton. Somehow they didn’t read any of the studies that showed that banking has a negative cost curve, that banks show higher costs per dollar of assets once they pass a not very high threshold. In other words, if a bank bought another bank and cut costs in the consolidation, they could have reduced costs to the same degree if not more as separate entities.
Concentrated supply networks. By the late 1980s, if not earlier, another managerial fad for manufacturers was to greatly thin their supplier networks. The claim was that having fewer suppliers would allow the buyer to invest in their supplier and forge deeper relationships. That never seemed to be the operative truth. Someone who becomes a large, as in overly large, customer to a business can push them around. I always assumed this fashion was primarily about extracting better prices or valuable non-price terms.
Offshoring. For global manufacturers, this was hardly a new idea. Automakers had plants in foreign markets largely to serve customers in those markets. However, Mexican maquiladoras quickly became popular among American auto and parts makers and Japanese selling into the US market. From Wikipedia:
Maquiladoras date back to 1964, when the Mexican government introduced the Programa de Industrialización Fronteriza (‘Border Industrialization Program’)….
In 1989, the federal government put in place specific procedures and requirements for maquilas under the “Decree for Development and Operation of the Maquiladora Industry”.[8] After the Mexican debt crisis of 1980 (see Latin American debt crisis), the economy liberalized and foreign investment increased. Factory jobs began to leave central Mexico, and workers followed the jobs from central Mexico to the maquilas in the north and on the border.[9] In 1985, maquiladoras overtook tourism as the largest source of foreign exchange, and since 1996 they have been the second largest industry in Mexico behind the petroleum industry.
Although the plural of anecdote is not data, in 1984 I ran the numbers on a potential acquisition of a Mexican business by a US aircraft maker. The short version is the deal made no sense (buyer and seller were 10X apart and when I modeled their difference in positions, tax, union relation, and FX/country risk, the difference made sense). Needless to say, this transaction would not have benefitted from maquiladora treatment. But I did visit Monterey (the capital of maquiladora-land) and at least on a drive-through, it consisted of impressively big heaps of raw materials and awfully rough-shod, small footprint plants. It looked like the manufacture as of then was on the order of metal-bending and assembly.
Back to Wikipedia:
With the introduction of NAFTA in 1994, Northern Mexico became an export processing zone. This allowed multinational corporations from the US to produce products cheaply. Corporations could use a maquila to import materials and produce a good more cheaply than in the US by paying Mexican laborers to lower wages and paying less money in duties. Mexicans work for approximately one-sixth of the U.S. hourly rate. During the five years before NAFTA, maquila employment had grown at a rate of 47%; this figure increased to 86% in the next five years. The number of factories also increased dramatically. Between 1989 and 1994, 564 new plants opened; in the five years following, 1460 plants opened.
Wikipedia contends:
However, the maquiladora growth is largely attributable to growth in US demand and devaluation of the peso, not NAFTA itself.
Help me. The “US demand” cannot be tidily picked apart from NAFTA. However, Mexico did have a currency crisis in early 1995.
Nevertheless, the motivations for outsourcing IMHO are not properly understood. In the auto business, which is typical of a lot of US industry, direct factory labor cost is 11% to 13% of product cost. The offsets against that are greater supervisory and coordination costs (longer shipping times and financing costs, and with that, greater risk of being stuck with inventories related to products that aren’t selling well) and just plain old screw ups due to having more moving parts.
In other words, outsourcing is better understood as a transfer from factory labor to managers and executives, at the cost of greater operational risk.
This cynical take is confirmed over the years by many executives telling me privately that the case in their company for outsourcing was weak but management went ahead because reasons, the most important often being they knew Wall Street would bump the stock price up.4
Misrule by Wall Street. In the Stone Age when I attended business school, even that bastion of capitalism hewed to the belief that companies served many constituencies, including (gasp) communities, and one of the big jobs of management was to navigate sometimes conflicting demands.
By the end of the 1980s, the notion that public companies have a sole duty to serve shareholders, a theory with no legal foundation but was successfully promoted by Milton Friedman and his followers, had become well enough accepted so as to restructure executive pay around stock price performance, and not business fundamentals.
Wall Street was remarkably forgiving of executive efforts to ‘splain away risk blowups, including the sort that resulted from over-optimizing for efficiency. It became almost a joke to seem companies every five years or so engage in a writeoff purge, pretend that all that Bad Stuff was part of Discontinued Operations and thus had nothing to do with the company as of now. I can’t recall anyone serious ever suggesting that the implication of the loss dump was the prior year earnings had been overstated, and where were the fines and sanctions for accounting/securities fraud?
Planned obsolescence. Goods companies learned Silicon Valley’s evil lesson: that software can be deployed to shorten product lives. I happily drive a 2003 Buick with <57,000 miles on it. Readers believe and I have no reason to doubt them, that it is unlikely that current cars will last more than 10 years because the manufactures will not support all the chippy components in them beyond that horizon.
So to achieve that end, cars (and other products) have chip-created feature bloat. But pretty much all the chips come from China or Taiwan. Oopsie! What if they have problems, like China's severe power shortage, or they decide to withhold some chips to make a point?
The general point is extended supply chains make the producer hostage to Bad Shit Happening to their suppliers. But quite a few companies made this vulnerability worse by integrating chips to add little to no value (what is wrong with turning on a car with a key, fer Chrissakes?) when the more likely justification was to make sure the product wouldn't last all that long either due to chip failure or software obsolescence.
Readers may be able to cite additional management fads practices to the miss, but the overarching point is hopefully evident: executive practices created the fragile supply chain mess. Deregulation and favorable trade deals greased the wheels, but they came about because businessmen, pundits, the press and pols were all pumping for them to advance supposedly desirable ends, and not as ends of their own.
_________
1 The success of the NUMMI joint venture, where Toyota took over the management of a plant with a reputed-to-be-terrible workforce and got it performing at over the median for quality standards for Toyota operations proved the American problem was much more “bad management” than “bad unions.”
2 Many accounts back in the day described how Detroit was in denial as well as incapable. Executives drove company cars that were babied every day by company mechanics, so they had no idea how they actually performed. They also failed to embrace small cars, both due to projecting their own preferences (in snowy areas, a heavy car is a plus), and profit (they thought they’d be cannibalizing sales of bigger, pricier cars, but better they cannibalize those sales than a competitor).
3 A friend who turned around and still runs a successful mid-sized US manufacturer who worked at GE under Reg Jones and then Welch hotly disputes the popular myth. She says Jones was the architect of GE’s success, that Welch mainly but slowly ran it into the ground.
4 This is not to say that outsourcing never made sense but that it was often adopted when there wasn’t much in the way of good reasons.
73 comments
Comments are closed.
It’s a mathematical and physical truth that the more optimised a system is, the less robust it is when perturbed. JIT is highly optimised and has just had a major perturbation.
Interestingly, the effects of the 2011 Great East Japan Earthquake led many Japanese manufacturers to adjust their JIT philosophy, by increasing the diversification of suppliers, so that they would less vulnerable to one being temporarily out of action.
At a former employer, we were a beneficiary, getting a contract we’d had long been fighting for. It was only on the basis that the customer wanted to make sure that there were if possible at least two suppliers for every vital part, but we weren’t complaining.
Agree 100%, what we’ve seen is an increasing brittleness in business.
My overwhelming feeling by about halfway through this essay, before reading the comments, is that a lot of this can be explained as overoptimization. Jack Welch focusing on the corporation as a financial instrument really seems like optimizing to Graham & Dodd methods, for example: they outlined the metrics that they like to follow, so corporations optimize to those metrics. Similarly “By the end of the 1980s, the notion that public companies have a sole duty to serve shareholders” – by optimizing for the markets, corporations could also move white collar compensation to the stock market through options and RSUs.
Re: Jack Welch — my late brother was a GE company man. He worked in research and adored his work for decades, until Welch came along. A few years into the Welch reign, my brother retired in disgust. They lured him back to work as a consultant, paying him far more than his former salary and benefit package, and he was free of office politics. He did that for a few years, then happily quit it to spend more time on his little sailboat.
Re cars being babied by mechanics, my dad worked on an assembly line for one of the big U.S. automakers as a summer job in the late 1940s. He said the workers always knew which cars were going to be checked by the engineers and would spend extra time making sure everything was right with those.
Monoculture is a big risk factor and the ongoing market concentration due to failure to enforce antitrust (primarily due to Robert Bork crippling it by imposing his Chicago School theories as out-of-thin-air jurisprudence) must bear a share of the responsibility.
We have stress tests for financial institutions, but nothing of the sort for critical infrastructure like the electric grid or telecommunications networks, let alone the more complex but a little more forgiving industrial supply chains.
It would be very interesting to see what happens with recovery of supply. It looks to me like the disruptions are akin to the physical process of wave interference (with the attendant local amplification). I am seeing a number of conditions worsening rather than gradually improving. Energy shortfalls popping up everywhere across generation methods, number of container ships at anchor in Los Angeles increasing, etc etc
I work for a company which does a lot of business with automotive manufacturers globally. We now do not believe that the supply side will recover until 2023. Which really means, we now fully expect 2022 to suck and we hope that 2023 would be better.
I know – “the Jackpot wasn’t any one thing” … yet it sucks.
Very interesting. I feel this in the sales force. I’m hunting around for a new car, but don’t absolutely need one. The sales team stresses to just buy what they have in stock that’s acceptable and can’t promise when or if they’ll get new stock in. They also give up pestering very quickly. With so little product to move, they’re only interested in serious buyers and not trying to nurture warm leads. That very much jibes with no return to “normal” until at least 2023.
Yes I’m expecting something like this. For instance, electronic component shortages in 2021, thus the downstream process (pcb assembly) is starved of input. When the component shortage gets relieved in 2022, the pcb assembly houses facing the backlog released on them, may be swamped. The concept can be applied anywhere there’s a chain of events that had all the slack taken out of it.
Incidentally, JIT isn’t necessarily about taking out all the slack; it is about knowing where it is so that you can make intelligent decisions about it. Lower the level of the water so that you can see where the rocks are, rearrange them for smoother flow. It is the smoother flow that is the real power move. Focusing on an end goal to save money on your water bill is a superficial interpretation.
Superficial as it may be, I think there are/were a lot of business leaders who see JIT as a way to push inventory management /expense way from them and further up the supply chain. They then get to pat themselves on the back and earn a big bonus for reducing cost and increasing efficiency…allegedly.
A lot of new car dealerships are going to go under by the time 2023 or 2024 comes around.
One of our local dealerships is building what looks like a faux port (a great waste of water in these trying times of drought, I feel) resembling San Pedro to lure the giant ships that once brought every make and model to our shores from Japan, and just how long the Car Go Cult can hang on, is all about hope.
The big three have been trying to get rid of their marginal dealers for decades now. This would be benefit to them if this is true.
Did a drive-by of the 10 or so new car dealerships in the area yesterday.
Some are down to putting used cars in the front as a phalanx and others not even bothering to provide left and right flanks to hide the big empty lot behind the ‘spoketemkin village’.
A few had handfuls of new cars mixed with the used, but that was it.
If I knew I wasn’t going to get any new inventory for a year or 2, i’d get busy filling that lot with used cars, but that isn’t happening, as it wasn’t their forte heretofore.
Certainly, a big part of the many current issues is managements philosophical shift from more resilient vertical integration with many possible local backup suppliers to the current offshored just in time single (or very few) suppliers.
How much of the current stresses to the “global supply chain” are covid19 related, and how much are new things that won’t change anytime soon?
I think weather related disasters are a new stressor that will be increasing. A couple big droughts (or even floods at the wrong times) in global “breadbasket” regions would lead to shortages. Or how about a plant disease that spreads more easily in weather stressed monocultures? Abnormal freezes like we saw in TX shutting down everything for a time. Are there likely to be other issues like shortages of rare earth metals, etc, not connected to weather but to resource depletion and politics?
Is the maintenance of the US tarrifs on Chinese goods a realization that increasing domestic manufacturing is a good idea or just coincidence?
The freeze in Texas was not abnormal or unexpected. Power plant operators choose not to spend any money weatherizing their plants and equipment. A decision they have again made for this winter.
There were more causes than that. The gas operators, from well to processing and pipelines, also did not weatherize or register as essential industry with the grid operator to avoid having their electricity cut when the grid needed to reduce load.
I knew someone who bought a fancy Audi convertible the year of the big tsunami in Taiwan. She ended up getting a car with no map/gps/back-up camera/info touch screen thingy because the parts for it were all from Taiwan. The wait was more than half a year IIRC. Not life threatening I know but at the time I was taken aback by the idea of only one source for those parts and the inability to source something from somewhere else in a reasonable time frame.
On the specific shortage of automotive semicon, I found this article interesting:
https://www.supplychaindive.com/news/nxp-semiconductors-automakers-just-in-time-shortage-chip/604424/
There is hint in there why the shortage seems to last so long, even after the original causes are gone. Everyone now wants to move away from JIT in semicon, and build up several months of reserves. As result, there is a one-time extra demand peak to full those reserve buffers.But suppliers (like NXP in the article) will only do so much to expand production to cover that one-time peak, because that would give them idle factories when the peak is over.
That would explain why the pressure stays on – every time there is some slack capacity, someone buys that capacity to expand their reserve buffer, which makes hiccups more likely as production capacity stays strained. And every time there is a new hiccup, the desire to hold a buffer grows more.
It might be the slo-mo high-tech version of the stereotypical toilet paper problems (or the UK petrol crisis of last week). Once there are signs of a shortage, demand goes up instead of down.
Yes. Another mass behavior is that, during a shortage, buyers will place orders in multiple places, expecting them to all be late, and then when the first one arrives, cancel the rest. This drives the distributors crazy, but there’s not much they can do, because times like these are when the distributors gain and lose customers in a big way, depending on whether they are able to save their customers in a crunch.
> Yes. Another mass behavior is that, during a shortage, buyers will place orders in multiple places, expecting them to all be late, and then when the first one arrives, cancel the rest.
This is an important but ignored factor to what is happening. A totaly false demand impulse is being sent down the line in a scramble to obtain supplies. One level of business has no insight into what the next level is doing.and that implies there is going to be a massive wreck, even worse than what is already happening, in store.
Greed rules the business person’s mind and whenever demand increases for whatever reason, expanding their business is of utmost importance and now it is based on fake demand.
It is not credible that the reasons for the power shortage in China are what we are being led to believe. It sounds like horseshit to me. The insight that the Chinese government has into the different levels of production within China and what is being shipped out are greater than the blind stuff orderers and I view the electricity shortage as a way to neuter the greed of their own. There is ruin to expanding production in the face of fake demand.
Automotive chips also have a much much higher storage and operating temp spec than commercial. You can’t just substitute an identical commercial spec chip into automotive.
Imagine all the temps the chips have to withstand—From underhood shutoff soak in Death Valley all the way to cold start in Alaska. It’s a tough ask.
The extended temp range is a complication at circuit-design level, but in terms of the technology in the individual components, all the electronics parts have to withstand 260C+ at least briefly when the circuit board assemblies are reflow-soldered. And components whose function is thermally limited, the market wants to max that out regardless of automotive or not. There are other things too about automotive qualified parts, like a more formalities in the QC process, lot-traceability and so forth.
How did the corporistas escape the clutches of the deadly cartels in the maquiladora zone?
They don’t – protection is paid.
One of my suppliers lost an entire semi truck load of industrial metals on the road between the Mexican mill and the US plant in Houston.
(for some reason, there is no US plant making nickel silver–I can’t imagine why… /s)
The same supplier tells me he has expected 5000 pounds of this material for over a month now. I asked him if the truck is missing again? Doesn’t know.
I’m going to add one more dubious practice to the list: Zero cash on hand. I actually remember an executive talking about this back in 2005. He said, “we want ALL of our money working for us; if it’s in the bank, it’s not working for us.” I remember wondering how we’d deal with cash flow variations that might be larger than expected.
The answer, of course, was short-term loans (a.k.a., “commercial paper”). And this worked reasonably well during steady times. But then in 2008 credit markets seized up, and the entire company got caught flat-footed. We couldn’t pay our suppliers. Even payrolls were even at risk. They finally worked through it, but they’ve always kept more cash on hand since then.
I was in U.K manufacturing from 1970 – 2000 and with J.I.T Manufacturing the Japanese creamed the
U.K industries.On buying our first Japanese C.N.C. machine in 1979 the Japanese were able to put an engineer from Paris into our factory within 24 hours,while Herbert &; Ward a world famous Birmingham Lathe manufacturer 60 miles up the road took 4 weeks! No wonder they went out of business.
Very quickly old truths and verities changed.Seeing a shop floor stacked up to the ceiling with W.I.P . which was associated with manufacturing health.”My they must be doing well here!” became a heresy.
All of these techniques JIT Cell Manufacture. TPM, TQM had as their goal the reduction of time.Imagine,the best manufacturer is one that would take the order (and the money) – with the materials instantly appearing and being manufactured into the finished unit.exactly when the customer order is to be delivered.
Zero lead time, therefore, is the goal of every process including R&D.
Many, many years ago i remember seeing a BBC documentary on the death of the UK car industry, and it highlighted how the old British Leyland tried to adopt Just in Time. But the British fundamentally misinterpreted its use – they thought that it was about efficiency (removing excess warehousing), whereas for the Japanese it was all about increasing quality (ensuring that it was easier to track faults).
Interesting Japanese style planning. Making their manufacturing quality more resilient, not less.
Part of the problem is that to show that A is a problem, B has to happen. If B doesnt’ happen for long enough, then A will seem not to be a problem.
As a matter of fact,we have many many As, but most of them never materialise, or at least “never on my watch”, which for most terms and purposes is the same.
A world where long, sensitive supply chains was pretty much an impossibility before end of 1980s, for a variety reasons, one of them being strategic. But for the last 30-40 years, it never really was a global wide-ranged issue (when it was global, it was specific – like oil for a while; when it was wide, it was not global in general), until CV19 – which shown that silently the world has shifted.
When I started in IT in the mid 1990s, it was on the IBM AS/400 platform, and EDS was the software vendor that I relied on for support.
EDS support was incredibly competent and their employees provided not only solutions to my problems, but a fare amount of my early education on the system.
The AS/400 had a built in self-service, education module that I could log into, and use when I wasn’t busy. It was a high quality tool that together with hands on experience, could teach a motivated user to become a competent systems administrator.
For the first ten years my experience was of emersion in a sort of guild system of skilled craftspeople who passed on their knowledge in users groups and informally in their day to day interactions.
As the demand for IT expertise increased this community was over taken by a system of credential-granting programs run by software vendors, manufacturers and IT management service providers.
My experience with support gradually went from fast, competent and friendly support by people with deep knowledge of their subject, to slow, frustrating waits on phone queues, only to be connected a person who handled my problem by adhering to a script. If my problem was routine, the script probably led to a solution, but it many times had nothing to do with the knowledge of the person I was talking to.
There are many times at the moment, when getting support is a matter of sending a request via email, filling out an online service ticket or opting for online “chat”, often, I suspect, with an “expert system”, IOW “chatting” with a computer that acts as a buffer between the customer and human support personnel.
The environment I work in went from few choices among high quality vendors, to many choices among mediocre vendors, and at the same time, the human side of things went from highly-skilled and competent “experts” to people with shallow skill sets, but lots of “credentials”.
The bad choices I see “management” making routinely, lead me to believe in a future littered with a wide variety of IT apocalypses, even the possibility that I walk into work one day to find a situation where nothing works and there is nothing I can do to help.
It doesn’t help that the public is under the impression the “new” means better, and that ageism is rampant in the IT world, leading to refusal to listen to anyone older than 30.
Meanwhile, the biggest Chevrolet dealer in my region, that usually has 300 new pick-up trucks to choose from, now has none, and is paying insane prices for used trucks at auction.
And consider that selling a vehicle in 1994 didn’t involve the internet at all, and today every sale involves the ‘services‘ of 6 or 7 web-based systems each one costing a small % of the deal.
IMHO, management is not sure to what extent these ‘services‘ actually make a positive contribution to the business, but there are promises that seem logical, and everyone else is doing it.
Worse, if the retailer does bite on the sales pitch, the vendor will sell their manufacturer on their wiz-bang web service, who will make its use a “brand-standard” and force the retailer to buy it.
BTW, every one of these vendors requires that your business shares all its data, including customer information as a prerequisite to delivering their services.
I’ve never seen management refuse that request, and once data-sharing is established, I’ve never had management order a disconnection.
The PMC is marching in short-sighted lock-step towards an abyss, that’s how we got here.
Everything is Enron, GM, Calpers, and the DNC, all the way down.
Apple TV has brought Asimov’s Foundation to television but in the real world there doesnt seem to be anything that will shorten the collapse. I for one would like to read about people doing things to lead us back from the precipice.
After reading about all these shortages across the world from the US, UK , China, etc. I began to wonder if this was a part of a larger pattern. For decades we have been using a supply system with little slack and was as tight as a drum. In our future supply system we will be using one more robust, more flexible and certainly more local. But to get from here to there will mean a steady increase of such disruptions as factors like climate change, resource depletion, etc. all make their presence felt and we are forced to change accordingly. So what I am saying is that we may be on the slow up-slope of a very steep bell curve and so we had better make our own preparations with this in mind.
I do wonder if the equilibrium state for American consumption is at a level far lower than the present one. And the resolution to our supply chain crisis won’t bring us back to the world of “plenty” the US consumer once knew.
The rotten fruit on our tree of debt and leverage is ability to bring future consumption into the present. But it really does seem that tree won’t bear fruit anymore. We’ve literally been throwing “money” at our problems with diminishing effect for the past two decades.
At this point reversing the “efficiencies” we introduced into our supply chain via “on-shoring” or creating redundancies in some process won’t do much to help maintain our consumption levels. If anything reversing these industrially engineered and operations researched, consultant implemented, “efficiencies” would reduce corporate profits enough to make them unpalatable to the powers that be. This reduced rate of profit wouldn’t be enough for the debt-servicing required to maintain our leverage merry-go-around. As a society we’d have to de-lever and consume less(far less).
This “crisis” seems to me to be an inflection point in our piece-wise consumption function.
Mining landfills may become big much sooner than I expected.
Of course there’s a computer inside your Buick that controls the engine and presumably is holding up nicely. I don’t think the turn to electronics really is about planned obsolescence since there are many legitimate reasons for their use, not the least of which is greater reliability. I’ve read that one reason for the chip shortage is that automakers want older well proven chips whereas the chip foundries want to make the latest high tech designs for the performance competitive consumer electronics business. Indeed I believe I got this link here.
https://jalopnik.com/i-asked-experts-why-carmakers-cant-just-transition-to-n-1847739665
Perhaps the solution is for these companies to have their own specialized foundries. Next year Hyundai plans to do just that. It may be easier in Korea than in today’s US.
Another issue for car electronics is the LOW volume, compared to computer usages of chips.
Automakers only need a small run of semicon for an entire product year. They had been ordering ahead, getting in the queue at the semi fab. But with the lockdowns, they cancelled their orders because they expected sales to crater. So when sales rebounded and they wanted to order semi, the automakers were at the back of the line for new semis.
Interestingly, the whole issue of low volume but mission critical semicon is being addressed by a Japanese consortium called Minimal Fab.
They use wafers that are 0.5 inches in diameter, and they do not need a clean room. The work in progress is transferred between machines in an airtight container. They claim to reduce the cost of a fab from billions of dollars to about ten million. Of course, the tech level is about 20 years behind state of the art. Design rules around 0.25 microns, transistor densities around 100k/mm2 (versus today’s density of 100M/mm2).
But, since most auto parts (except Tesla) don’t need massive computing, the old tech level is adequate for the auto industry. This is confirmed by earlier comments in the thread about auto companies wanting to use old, proven tech.
It would be interesting to get in on this, but it seems to be highly localized inside Japan. Very little news coverage outside Japan. Almost no literature in English. The Minimal Fab consortium members provide all the consumables for the process.
Sorry, Carolinian – I missed the hook to your comment.
Of course, Minimal Fab would be just what you suggested – a specialized foundry for auto makers.
Thats fascinating. i knew there was a pretty big market in discontinued chips. Companies that buy and re sell old stock.
This is an entirely different approach, and most of the old stuff doesn’t need or even want modern shrunk geometries.
You can still buy today opamps designed by Bob Widlar in 1970. Still useful today.
Thanks for confirming that “old stuff doesnt need modern geometries”. I meant to add to my apology to Carolinian that, by itself, a car company doesn’t have the volume to run a decent-sized state/art conventional fab – unless they are willing to take an operating loss to guarantee supply. That’s why Minimal Fab is such an interesting concept/product.
Very rapid obsolescence is a big issue in electronics. In addition to the market in used parts, there are places that keep copies of old tape readers just to transfer stuff to modern storage devices.
IIRC, the military has a program of “workalike” transistor circuits that have the same operating parameters as old vacuum tube equipment, so they can be plugged into old military gear. I suspect something like this is going on in the 1950s era air traffic control system.
I don’t know the details to what Hyundai has in mind but they have made such an announcement. They are a worldwide car company with large volume but not even close, I think, to Toyota.
It’s quite likely that these simpler chips are more or less alike and could work in most cars now made so perhaps carmakers could buddy up.
I agree it makes sense for carmakers to “buddy up”. But, I looked at the list of the 150 or so companies in the MinFab consortium, and not one auto company. Some division of Mitsubishi, but not the auto division.
Also, re: Hyundai, they are Korean; and the Japanese and Koreans have this intense rivalry (read hatred). So, I doubt the consortium would accept them as a member. Further, Hyundai pulled out of non-commercial sales in Japan because no one was buying their stuff.
do you have a link to Hyundai’s fab effort?
Now that is something she (or someone) should write a book about!
I am puzzled by the view that a slowdown in the delivery of discretionary purchases is considered a crisis. Modern cars can easily last 15-20 years, as Yves has demonstrated, yet people are howling with displeasure because it is less convenient to buy cars and trucks to make fashion statements. Americans are addicted to the regular purchase and replacement of material goods (mostly unneeded) as a way to medicate their dwindling spiritual health. How sad that the iPhone upgrade has now become an act of renewal of personal worth.
When I was a kid, car turnover was pretty frequent as you had to have the newest model, plus the cars didn’t last that long. My truck has 163k miles and I feel another 163k is not out of the question. Cars would dead by 50-75k miles back in the day.
In a place such as SoCal where you drive all over tarnation to get anywhere and are often subject to stop and go traffic, it doesn’t really matter what kind of car you have in reality, all kinds can go slow enough in a traffic jam, nor could you really ever do more than say 90 mph, negating that spendy Ferrari’s horsepower.
That said, you are what you drive rules used to apply in LA, an instant signifier of status that was seldom wrong.
Engines were expected to be rebuilt after 100k miles. They used to sell them (for Ford and Chevy) in the Sears Catalog. Take that Amazon!
Crisis is a relative word, I guess. There really is something weird going on, with similar problems showing up in very different places and situations. The automotive semicon problems are a highly visible example of a deeper issue.
One problem is, we don’t not have a method restrict the problems to people with undeserving demands. If someone needs to buy a car for a good reason, the cars are for them just as rare and overpriced as for fashion statement people.
And second, a lot of problems are currently kept away from end-consumers, by all kinds of buffers and tricks and efforts. That may not last forever, if they stubbornly keep growing. So even if the current situation is not worthy of the word crisis, it is still good to figure out what’s going on, in case it expands to a worse situation with more impact on essential production
“current cars will last more than 10 years because the manufactures will not support all the chippy components in them beyond that horizon”.
All those very fancy low-volume supercars will be unmaintainable in 20 years when electronics starts failing. No-one will reverse engine an engine management system to determine the engine mapping, for a market of 100 cars. Then you have the body management and infotainment systems too…
The custom tyres will be difficult to source before then. Wound: meet salt.
You missed the words “that it is unlikely” that current cars . . .
Current cars are priceless. I wouldn’t get one even if the price were zero.
Peak car design was decades ago.
Also current cars are as much surveillance machines as cell phones.
From VW’s 2019-10 Privacy Policy for Car-Net Security & Service 2020 models in USA:
“please note that precise GPS location data and your driver behavior data will not be used for VW marketing purposes or shared with unaffiliated third parties to use for their own marketing purposes without your affirmative consent.”
Or course, by definition parties VW contract with are affiliated, so “affirmative consent” is not necessary. Also, any purpose that isn’t “marketing” doesn’t require consent.
Peak car design was decades ago? During the ” unsafe at any speed” era?
If the engineers at GM at that time had taken the anti roll bar that they were going to put in the rear suspension of the Corvair and beaten the GM accountants to death with it, Nader never would have had the material for that book. The friggin bean counters insisted the five bucks it would have cost was too much and that was that.
The rest is history.
Peak car design was during the 90’s, before the total crapification of everything.
In the era of aftermarket Chinese auto parts you’d be amazed at how old a car can get before parts aren’t available.
And the electronics that are being put into cars are not that exotic. A last century processor will do.
Yes, and rewriting the engine management code to run on that processor, might be, as they say, “A spot of bother..”
All cars are a spot of bother. The point is whether it’s possible or not. Think you’d be surprised how much.
Plus of course you’d have to have quite an old car for the parts not to be available from the dealers themselves.
Should we expect the “experts” based on their knowledge and lessons learned from the current situation, to claim that artificial intelligence will be the next fad, I mean cure, to make supply chains more resilient in the future?
EDS has plenty of deals that went bad and plenty they lost money on. I worked with them as an outside vendor at two separate companies. The key to their success was the contract. EDS was extremely skilled at writing and negotiating contracts. Companies on their first foray into outsourcing were at a severe disadvantage. Eventually, companies gained experience and this advantage fell away and in some cases EDS ended up in bad contracts.
There was nothing inherently proprietary about how EDS worked. As I saw first hand, EDS would take over a data center, fire all of the long term and very experienced ( higher waged ) employees with no severance and replace only half with lower paid new hires providing bare minimum contracted services. Instant profit.
Up until around 2000 or so, 50% of EDS’ business came from GM. By 2008 EDS was in very bad shape as offshoring became more popular and they could not compete on price any more and sold themselves to HP.
I have been contemplating this topic for a couple of decades. Even to the point of researching a book on the role of inventories as telling a tory through history. Starting with storing grain centuries ago as a food source between harvests to today. Companies have leaned inventories off their balance sheets for a few reasons in my view. One is to use their leverage to force them on suppliers, creating a daisy chain, with increasing geographic reach, which is inherently weak. The second, is to please Wall Street — who typically know nothing real. The third is a redefining and concentration of power in the C-suite which justifies higher compensation. The problem with all this is it works until it doesn’t. Putting Humpty Dumpty back together for something we evolved into over decades will be a mess.
Interesting sequence of events. Can we suppose that at some point in the future global competition (“efficiency”) might reach an equilibrium – sort of homogenized – despite the frenzy to find an advantage for profit? There will always be an advantage for quality. But should quality equate with profit? Shouldn’t quality be the goal of progress; and profit be more or less static? So just to carry that thought out then, wouldn’t competition for quality be a good thing? Whereas today our idiotic competition for profit/shareholder-value is quite literally wasting and destroying the planet. ?
Corporations could use a maquila to import materials and produce a good more cheaply than in the US by paying Mexican laborers to lower wages and paying less money in duties. Mexicans work for approximately one-sixth of the U.S. hourly rate. —quoted from a Wikipedia article
Trade is based on comparative advantage, not absolute advantage. If it was the case that the US was three times as good as Mexico in growing corn, and twice as good at making cars, it would still be in the interest of the US to grow corn and trade it for cars, for they get more cars that way. The same is true for Mexico, in that they get more corn by trade than by growing it themselves.
Though she [i.e., Portugal] could make the cloth with the labor of 90 men, she would import it from a country where it required the labor of 100 men to produce it, because it would be advantageous to her rather to employ her capital in the production of wine, for which she would obtain more cloth from England, than she could produce by diverting a portion of her capital from the cultivation of vines to the manufacture of cloth. —David Ricardo
The big problem with this premise is that it presumes a static set of comparative advantages …
JM Greer has an interesting take on this–What if the shortages are actually the first symptom of the world hitting the “Limits to Growth” as described by the book of the same name?
The hypothesis was that eventually shortages of raw materials and rising pollution would put limits on industrial production.
“Given the nature of today’s global economy, groaning as it is under the burdens of dysfunctional centralization and excess complexity, a flurry of seemingly unrelated shortfalls and delays is exactly how a contraction in industrial output would show up first, as marginal producers of components and raw materials fail to contribute their quotas to the manufacturing sector.
If this is correct, we have reached the point at which the decline in resource availability and the increase in the total pollution load on the environment and economy have begun to throw monkey wrenches into industrial output.”
https://www.ecosophia.net/a-prayer-for-nonbelievers/
Thanks Yves.
When I read articles like this I often have the feeling of an itch being scratched or an “ah ha, I knew this on some level” and now I actually know how to articulate what I sense or feel. Focus on management strategies and profits and not the people who produce or create has been such a mistake and had such dire consequences for so many for far too long. The Chicago School and the MBAs who ruined manufacturing jobs must have a very special place in hell waiting for them. Seriously, those family blog-ers have a lot of responsibility for the current mess in which find ourselves. Where is the FDR who can convince today’s 1%ers that a few million bucks less in their pockets won’t kill them?
And I agree with Charles upthread, I long for solutions and good news stories. I know this is not manufacturing but somehow a new book called Pastoral Song by James Rebanks comes to mind. I heard a lovely interview with him recently and have the book on order from my local bookseller. He owns a small farm and has turned his back on big ag methods. His farm is healthy and turns a small profit. He encourages people to demand more local products when they’re shopping. Shopping local and lessening my general non life-supporting consumption, keeping my car as long as possible are all things I do so that I feel like I am making some positive contribution to the world – or at least not adding so much to the mess. Solving the problems described in Yves’ article are just way beyond me. Although, I do feel some hope when I hear about people refusing to go back to BS jobs and unions going on strike. Some days I even think it may not be too late.
Excellent piece Yves. Iirc, there was a kerfuffle about the entertainment screens in autos. Didn’t last. But here’s your line I loved:
Unless the Silicon Valley Lords buy a law making it illegal to sell and make cars which turn on with a metal key in a metal lock, there may well remain a niche market of people who would buy a car which turns on with a metal key in a metal lock. Would the niche market be big enough that a niche carmaker could live on the proceeds of serving that market?
Excellent piece Yves. Iirc, there was a kerfuffle about the entertainment screens in autos. Didn’t last. But here’s your line I loved:
“But quite a few companies made this vulnerability worse by integrating chips to add little to no value (what is wrong with turning on a car with a key, fer Chrissakes?)”
I do wonder if the equilibrium state for American consumption is at a level far lower than the present one.
In a complex non-linier system, one witn distortion in it’s feedback, there are many possible quasi sable equlibria. For example being healthy alive and functioning fulls is one state, and being disabled after an accident, another.
The chance from one quasi stable equilibrium to another take an impetus, war, plague, death of a head of states being some of the impeti.
After a state change, there is mostly a new equilibrium, and rarely, very rarely, a return to the previous equilibrium,
As a test: Choose any war, and examine the conditions of a country before and after a war. Or a plague.
There is no “normal. There is much effort to steer the system to the benefit of those who have the most control, or influence over the system. Similarly there are always perturbations in the equilibrium, some caused by those who dislike the current equilibrium (or status quo.)
The main fact is: “There Is No normal”
i remember when nafta billy clinton was holding up dell computers as a ideal that should be emulated, nafta billy was enthralled that dell had 4000+ suppliers scattered all over the world.
we see how dim witted that thinking is today, i saw it as dim witted when nata billy said that on t.v. back in the 1990’s.
read this,
https://www.amazon.com/End-Line-Coming-Global-Corporation/dp/0767915879
Paul Volcker was born in 1927, so he wasn’t relenting on anything in 1892. :)
Sony apparently partnering with TSMC to build a chip factory in Japan
Complex global supply chains.
It seemed like a good idea at the time.
The neoliberals designed a global economy for maximising profit.
The neoliberals had designed a global economy that had no resilience to any sort of shock.
No one thought about anything other than maximising profit.
An ideological triumph for the neoliberals, but a disaster for the global economy.
Companies used to put money aside to tide them over when times got hard.
Private equity firms stripped all the spare cash out of firms and loaded them up with as much debt as they could carry in the good times.
Companies used to keep a lot of stock that would tide them over if there were any problems with supply.
What a waste of money.
Just in time delivery systems are what you need to maximise profit.
Companies used to have spare capacity that they could bring online quickly when demand rose suddenly.
What a waste of money.
That spare capacity was just sitting idle most of the time.
….. etc …..
it was pretty bad when you had a president extolling this crap on t.v. almost daily. nafta billy clinton was ecstatic as he tore apart americas wealth that took 200 plus years to accrue.
I don’t think auto makers today are planning obsolescence in quite the way you imply. For instance, the key has been a chip since the 1990s. Plenty of those cars have now rusted out before the silicon shorted out.
The oily bits are still very well made, though there are new issues (more below). The planned obsolescence comes from the infotainment systems and their integration.
It used to be the stereo was just a stereo and was made in a standard, replaceable size. The oily bits and the bodywork were good for 20 years if you took care of them, and you could upgrade the stereo to latest features quickly and easily. Now the stereo is a computer tied to the engine and body computers and to the manufacturer over the mobile data network. It talks to the steering wheel controls using proprietary protocols, puts messages on the LCD screen in the instrument cluster and shows messages important to the operation of the car, like when to change the oil. It’s also no longer a standard size, and likely integrated into the dash in such a was as to be hellishly difficult to replace physically in addition to the electronic issues.
The pace of change in electronics thus makes the car feel obsolete in use long before the oily bits wear out.
As to the oily bits, the advent of direct gasoline injection and high-boost/low RPM turbocharging has quite a few of us concerned about the real world durability of some modern engines. The systems have been pretty common since ~2010, so we should have some decent data about now, I just haven’t dug deeply enough into it to form a strong opinion.