9 ms·

The "religious reasons" he's talking about are "zero inventory" or "JIT inventory" pioneered by Japanese auto makers in the 80s and 90s. By optimizing inventory

The "religious reasons" he's talking about are "zero inventory" or "JIT inventory" pioneered by Japanese auto makers in the 80s and 90s. By optimizing inventory management and understanding demand for your product, you can ensure you don't have excess inventory on-site (meaning you don't have money invested in product parts you can't sell and have to pay taxes on). It's the system that virtually every manufacturer (auto or otherwise) uses now. It's why major supply chain issues cause catastrophic manufacturing problems.

j-bos 5y ago
My understanding is that JIT, as performed by Toyota, takes into account the differences in difficulty of sourcing different inventories. If this is true, one would expect that Toyota faced shortages later than other manufacturers. Does anyone have a source proving or disproving this?
- hef19898 5y ago
  Toyota was among the last to be hit. Also, JIT is not having zero inventory. It is about having the right amount of inventory.
  - worldsayshi 5y ago
    Pedantic comment: is that not what lean is about? Or is "lean" and jit essentially the same thing?
    - hef19898 5y ago
      JIT, as Kanban, is part of the Toyota Production System. It is also called Lean nowadays, or simply TPS. The basic principles go back to the US defence industry in WW2, on one hand, and some ingenious ideas from the old Toyota company manufacturing looms. I also see some of these ideas, tacted assembly lines for example, as far back the Venetian Arsenal cranking out Galleons and Galleasses in unprecedented numbers in the 16th (?) century.
  - tuatoru 5y ago
    The conventional wisdom is that Toyota (and Honda, Nissan, etc.) learned the value of keeping buffer stocks after the 2011 Tōhoku earthquake.
    - hef19898 5y ago
      JIT always has some buffer, just smaller ones than in other models. True that Japanese manufacturers increased those after 2011.
- da_chicken 5y ago
  It can take into account predicted difficulties. The pandemic was difficult to predict, basically everyone predicted incorrectly, and chip supply is functionally inelastic over the short term.
  - lenzm 5y ago
    You don't need to know what the pandemic is specifically to plan for a global market shock. Knowing that you have a critical component with a long inelastic supply chain is enough.
    - SahAssar 5y ago
      I'm guessing the problem is to decide what is a reasonable difficulty. Even if you plan for a global market shock, how long do you plan for that to continue? How many companies failing in your supply chain do you plan for?
    - hef19898 5y ago
      And then what? Just stockpile a year of everything? At every company along the supply chain? It is the first time in my life, professional in SCM and otherwise, that something like Covid and a global trade disruption happened together with a global shortage of one class of products. No way you for see or prepare for something like that. Hindsight in 20/20, just imagine the automotive OEMs would have gotten rid of the JIT for chips and Covid (and all the other things causing the shortages) wouldn't have affected chips but, say, certain raw materials like Aluminium.
      - bsder 5y ago
        > And then what? Just stockpile a year of everything? Or run a fab. Fabless isn't a God-given right. Vertical integration is a thing. Vertical integration got a bad reputation because financial engineering dipshits like Carl Icahn would come in and strip mine your company for the single super profitable division and destroy the rest (see: Mesta Machinery).
- forbiddenlake 5y ago
  Yes: https://news.ycombinator.com/item?id=28232469 https://news.ycombinator.com/item?id=28232469
- imeron 5y ago
  https://www.bloomberg.com/news/articles/2021-04-07/how-toyota-s-supply-chain-helped-it-weather-the-chip-shortage https://www.bloomberg.com/news/articles/2021-04-07/how-toyot... It wasn't enough it seems.
xondono 5y ago
What I don’t see discussed is how JIT has become the “securitization” of manufacturing. Everyone had an incentive to do JIT, and it improved efficiency in each individual company, but it did so by inflating a systemic risk that could be easily triggered by a “black swan”.
- wnissen 5y ago
  The thing is that it also reduces the risk of certain systemic risks. Recessions (and even worse, deflation) happen when production is too high relative to demand, but no one has noticed. In the 80s if demand for cars dropped, everyone from the dealers to the manufacturers to the parts suppliers were left with a ton of excess production. In the best case, you merely transfer some production from the end of a boom to the beginning of the end of the bust, but in the worst case all that inventory just sits around and deflates.
hef19898 5y ago
And major manufacturing problems cause catastrophic supply issues some times. manufacturing is part of the supply chain, and it is simply impossible to protect against all eventualities. JIT is not to blame here.
- da_chicken 5y ago
  > it is simply impossible to protect against all eventualities. Of course. No amount of stockpiling will account for a moon-sized asteroid collision. > JIT is not to blame here. Kind of. JIT isn't to blame, but the pressures that reward making every level of industry maximize efficiency, such as by using JIT, is. That is the reason for the supply chain crisis. The market and tax laws pressure companies to optimize for efficiency as much as possible with their supply chains. Another alternative would be to optimize for robustness, but that would naturally be inefficient and would still be unlikely to endure a 100-year catastrophe unless it were completely unable to survive in a competitive market. That doesn't mean that we need to change anything, necessarily, but it does mean that we have structured our economy to be less resilient than it could be with potentially very little actual upside. Even so, some supply chains, like food and clean water, really should be robust rather than efficient. Emergency medical supplies, too, should probably be robust rather than efficient. That's why the Obama pandemic plan included warehouses of medical supplies [0] and FEMA stockpiles food and water. Utility systems, too, should probably favor robustness. Texas has had a horrific year for it's power industry, and it has revealed how a hypercompetitive marketplace doesn't really serve the public best interest. Consider civil engineering. We could optimize that for efficiency. That would result in a lot more bridge collapses, structure failures, and so on. Instead, we generally try to overbuild when failure would cause injury or death. The same is true in most engineering contexts. In many cases, industry simply doesn't design for robustness until they are held liable for failures. Notice the rise in corporate insurance rates after computer security audits due to the risk of data loss, and the sudden panic of executives to suddenly have secure systems where before they were more than happy to have databases of plain text passwords and so on. There are alternatives to profit-motivated business efficiency. We have simply chosen not to do that sort of by accident. [0]: https://www.npr.org/sections/health-shots/2016/06/27/483069862/inside-a-secret-government-warehouse-prepped-for-health-catastrophes https://www.npr.org/sections/health-shots/2016/06/27/4830698...

j-bos 5y ago

My understanding is that JIT, as performed by Toyota, takes into account the differences in difficulty of sourcing different inventories. If this is true, one would expect that Toyota faced shortages later than other manufacturers. Does anyone have a source proving or disproving this?

hef19898 5y ago
Toyota was among the last to be hit. Also, JIT is not having zero inventory. It is about having the right amount of inventory.
- worldsayshi 5y ago
  Pedantic comment: is that not what lean is about? Or is "lean" and jit essentially the same thing?
  - hef19898 5y ago
    JIT, as Kanban, is part of the Toyota Production System. It is also called Lean nowadays, or simply TPS. The basic principles go back to the US defence industry in WW2, on one hand, and some ingenious ideas from the old Toyota company manufacturing looms. I also see some of these ideas, tacted assembly lines for example, as far back the Venetian Arsenal cranking out Galleons and Galleasses in unprecedented numbers in the 16th (?) century.
- tuatoru 5y ago
  The conventional wisdom is that Toyota (and Honda, Nissan, etc.) learned the value of keeping buffer stocks after the 2011 Tōhoku earthquake.
  - hef19898 5y ago
    JIT always has some buffer, just smaller ones than in other models. True that Japanese manufacturers increased those after 2011.
da_chicken 5y ago
It can take into account predicted difficulties. The pandemic was difficult to predict, basically everyone predicted incorrectly, and chip supply is functionally inelastic over the short term.
- lenzm 5y ago
  You don't need to know what the pandemic is specifically to plan for a global market shock. Knowing that you have a critical component with a long inelastic supply chain is enough.
  - SahAssar 5y ago
    I'm guessing the problem is to decide what is a reasonable difficulty. Even if you plan for a global market shock, how long do you plan for that to continue? How many companies failing in your supply chain do you plan for?
  - hef19898 5y ago
    And then what? Just stockpile a year of everything? At every company along the supply chain? It is the first time in my life, professional in SCM and otherwise, that something like Covid and a global trade disruption happened together with a global shortage of one class of products. No way you for see or prepare for something like that. Hindsight in 20/20, just imagine the automotive OEMs would have gotten rid of the JIT for chips and Covid (and all the other things causing the shortages) wouldn't have affected chips but, say, certain raw materials like Aluminium.
    - bsder 5y ago
      > And then what? Just stockpile a year of everything? Or run a fab. Fabless isn't a God-given right. Vertical integration is a thing. Vertical integration got a bad reputation because financial engineering dipshits like Carl Icahn would come in and strip mine your company for the single super profitable division and destroy the rest (see: Mesta Machinery).
forbiddenlake 5y ago
Yes: https://news.ycombinator.com/item?id=28232469 https://news.ycombinator.com/item?id=28232469
imeron 5y ago
https://www.bloomberg.com/news/articles/2021-04-07/how-toyota-s-supply-chain-helped-it-weather-the-chip-shortage https://www.bloomberg.com/news/articles/2021-04-07/how-toyot... It wasn't enough it seems.

xondono 5y ago

What I don’t see discussed is how JIT has become the “securitization” of manufacturing. Everyone had an incentive to do JIT, and it improved efficiency in each individual company, but it did so by inflating a systemic risk that could be easily triggered by a “black swan”.

wnissen 5y ago
The thing is that it also reduces the risk of certain systemic risks. Recessions (and even worse, deflation) happen when production is too high relative to demand, but no one has noticed. In the 80s if demand for cars dropped, everyone from the dealers to the manufacturers to the parts suppliers were left with a ton of excess production. In the best case, you merely transfer some production from the end of a boom to the beginning of the end of the bust, but in the worst case all that inventory just sits around and deflates.

hef19898 5y ago

And major manufacturing problems cause catastrophic supply issues some times. manufacturing is part of the supply chain, and it is simply impossible to protect against all eventualities. JIT is not to blame here.

da_chicken 5y ago
> it is simply impossible to protect against all eventualities. Of course. No amount of stockpiling will account for a moon-sized asteroid collision. > JIT is not to blame here. Kind of. JIT isn't to blame, but the pressures that reward making every level of industry maximize efficiency, such as by using JIT, is. That is the reason for the supply chain crisis. The market and tax laws pressure companies to optimize for efficiency as much as possible with their supply chains. Another alternative would be to optimize for robustness, but that would naturally be inefficient and would still be unlikely to endure a 100-year catastrophe unless it were completely unable to survive in a competitive market. That doesn't mean that we need to change anything, necessarily, but it does mean that we have structured our economy to be less resilient than it could be with potentially very little actual upside. Even so, some supply chains, like food and clean water, really should be robust rather than efficient. Emergency medical supplies, too, should probably be robust rather than efficient. That's why the Obama pandemic plan included warehouses of medical supplies [0] and FEMA stockpiles food and water. Utility systems, too, should probably favor robustness. Texas has had a horrific year for it's power industry, and it has revealed how a hypercompetitive marketplace doesn't really serve the public best interest. Consider civil engineering. We could optimize that for efficiency. That would result in a lot more bridge collapses, structure failures, and so on. Instead, we generally try to overbuild when failure would cause injury or death. The same is true in most engineering contexts. In many cases, industry simply doesn't design for robustness until they are held liable for failures. Notice the rise in corporate insurance rates after computer security audits due to the risk of data loss, and the sudden panic of executives to suddenly have secure systems where before they were more than happy to have databases of plain text passwords and so on. There are alternatives to profit-motivated business efficiency. We have simply chosen not to do that sort of by accident. [0]: https://www.npr.org/sections/health-shots/2016/06/27/483069862/inside-a-secret-government-warehouse-prepped-for-health-catastrophes https://www.npr.org/sections/health-shots/2016/06/27/4830698...