I fully agree with your comment overall, but as someone who lives in Japan I'd like to point out that specifically "people who fill in the holes in roads" is a running joke in Japan since small street constructions always seem way over-staffed.
It's amazing because they do finish a lot faster than in western countries, but you can definitely remove 20% of the people in street-level construction and would see little difference. An example: they hire 2 people on every sidewalk just to tell people to continue walking the way they are walking and make sure they don't fall into a pit while looking at their phones[1].
From what I've learned, this specific job is normally filled by people who have already retired but don't have enough to live, so they are paid peanuts (I believe they might also be paid a bit extra by gvmt to go back to work, but I'm not sure I understood this point properly). They have started to be replaced by "robots" though [2].
Edit: in general everything in Japan seems overstaffed compared to the western counterparts, not just road construction. I believe this is how they can maintain very low levels of unemployment here, but I don't yet know how the incentives work to make it possible.
Sometimes I think what passes for appropriately staffed in the West is actually understaffed. What's called optimal staffing in the West is just at capacity, which is a nice way of saying "with no spare capacity."
Of course, with Japan being the land of just-in-time, one could argue that spare capacity is bad -- but that's an oversimplification. Necessary spare capacity is good. What's bad is not being able to address the reasons the spare capacity is needed. Maybe when it comes to humans doing road work there's no way around the need for spare capacity?
From my experience visiting, there were lots of jobs that were "nice to have". Like the man making sure people don't fall in the hole, or the person standing by the carpark exit making sure the way is clear. You could probably do without them, but it's a nice thing to have.
There's lots of examples of that, people doing things that you would see brutally optimized out of western city and those cities I feel are worse off for that optimization.
One way to look at it is that with their wealth they've decided to employ more people and do more good, versus pocketing the profit and calling it efficiency.
The famous railway networks in Japan are actually privatized rail companies which was very surprising to me. When the rail companies privatized in my country they immediately became worse, as they were suddenly serving a very different purpose. Staff were let go, the trains are cleaned less often, services were cut. That kind of thing.
> From my experience visiting, there were lots of jobs that were "nice to have". Like the man making sure people don't fall in the hole, or the person standing by the carpark exit making sure the way is clear. You could probably do without them, but it's a nice thing to have.
How would liability play out for the construction company in a Japanese court? It's already a low margin competitive industry so passing risk management costs onto customers (the cost of the people monitoring the holes) is likely more practical than risking all profits evaporating and years of litigation, especially if insurance has anything to say about it.
I don't know the answer to that sorry, you're right to point it out though. Perhaps there is actually some prior court case that defined a liability to make those individuals necessary to be compliant.
In Australia we certainly don't have those people, construction companies just use a few signs, some cones and barricades and that's all that is required to be compliant. If someone falls in the hole they shouldn't have walked past the cones!
> In Australia we certainly don't have those people
Not the case in Victoria. Pretty normal to see a couple of workers standing near the site entrance making sure pedestrians don't get run over by a truck. Or standing in front of a closed footpath, directing pedestrians to cross the road.
There's pretty strict rules around keeping the public out of construction sites and excavations. Putting a couple of cones around a hole is definitely not all that's required.
How many people? The more people you put (let's say fixing a pothole) the safer it is. Would you put 1 extra? 10? 100? 1000? At what point do you stop?
There's absolutely a limit where adding another person only adds marginal safety. For instance, children who don't know they should not jump to jump a barrier and then go into a pothole should not be allowed alone on the streets. We are not talking about a random hole in the street, these things are already heavily warded in Japan with barriers and signs AND on top of it there's also workers AND on top of it the mentioned security guards. My argument is, many or all of these security guards are not needed.
In my country the government used to run the railways, but it had been run down for decades and was at the verge of closing all but the busiest commuter lines into the capital. It was privatized, and since then usage has skyrocketed (well pre covid) to over twice the level pre-privatization, so clearly something's gone well.
The privitization is split into different companies, the best having had a 30 year contract to revitilise their line, and they've done wonders with it, while one of the worst was (pre covid) run to the exact orders of the government
My pet conspiracy theory is that Toyota's liberal knowledge sharing with competitors is really a psy-op to keep the competitors perpetually on the edge of ruin while they laugh all the way to the bank.
They don't even have to be deceptive about it. Just in time is obviously about not having undue spare capacity, not about not having any spare capacity at all. But Toyota knows this nuance is going to be lost on buzzword-loving PHBs and greedy management consultant firms. You see the same thing happening with corporate Agile.
Toyota knows it competitive edge is its culture. Culture to change, to be retrospective, to improve. Lean, JIT, Kanban etc were derived from a culture of striving for perfection and to develop the optimal processes for a specific company is something that takes decades and is never ending, not something that can be read in a book, taught on a course in a short period of time.
There's also that story from back when Toyoda was making sewing machines or whatever it was, and a competitor stole copies of the engineering drawings for their latest model. Toyoda shrugged and said something to the effect of "So what? The really valuable knowledge are the mistakes we made coming up with those drawings, and that's not in the drawings. By the time our competitors have managed to get their copy of our machine to the market, we will have innovated away from that for our next model. And that innovation is informed not by what our current model looks like, but by the exploration we did to get there. None of that was stolen, so after this model, they will just keep repeating the mistakes we made."
I have no idea of whether this is true at all, but I really like the story anyway.
----
What I do know is true is that modern day Toyota doesn't mind sharing the solutions they have come up with to various problems, because they think the real value is in the people and processes to (a) identify the problems in the first place, and (b) come up with solutions suited to those specific problems.
Blindly applying Toyota's solutions to Toyota's problems to your organisation, just hoping that you have the same problems as Toyota and that their solutions will work also for you is a recipe for confusion, and not what matters. (Yet virtually every "development methodology" is a specific solution to a specific problem blindly applied to an organisation.)
You see this a lot in IT companies if it works for google, Amazon and Microsoft, It will work for us while we have like a millionth their load or complexity.
> None of that was stolen, so after this model, they will just keep repeating the mistakes we made.
The competitor probably could have purchased a sewing machine and replicated the engineering drawings themselves. Teardowns of competitor products is common today, and I'm sure it was back then as well.
I think Toyota's "helpful" attitude comes from the Western paranoia in the 70-90s that Japan was going to completely dominate worldwide manufacturing. After all, their rise was fast: in the mid 60s, Japanese manufactures were practically begging to sell rudimentary formed metal components in the USA, but by the early 80s, they were a leader in the high-tech manufacturing. That rise caught many people off guard and I've heard it said that the "lost decade" in Japan was the result of American trade policy specifically designed to curtail Japan's growth in manufacturing.
Viewed in that light, it makes sense that Japanese companies would appear "helpful" to American ones. Why else would Toyota co-build a plant with GM in order to teach GM their Kaizen philosophy for manufacturing?
According to the book The Toyota Way: 14 Management Principles from the World's Greatest Manufacturer by Jeffrey Liker, one of the motivations for Toyota was to repay the debt to American manufacturers who participated in the postwar rebuilding of Japan and taught American manufacturing techniques to Toyota and other Japanese manufacturers.
I’ve heard recently (in a video that I don’t know how to find again) that that’s exactly what happened with the automotive industry’s chip shortage: Toyota was the only one that built a stock of chips in advance, and now are the only ones that can keep producing cars without being limited by the shortage.
> Just in time is such a simple principle, but the pursuit of the elimination of waste is now the central mission of any major manufacturer.
> However, most did it wrong. Manufacturers globally saw the headline “elimination of inventory leads to massive efficiency gains” and jumped on that without actually determining what made it work for Toyota.
> They ignored that Japan's small physical size made for short domestic supply chains, less vulnerable to things going wrong.
> They ignored the company's production leveling: finding the average daily demand and producing that regardless of short-term changes and demands.
> They ignored the fact that eliminating excess inventory is different from eliminating all inventory.
> They ignored the principle of growing strong teams of cross-functional workers
predicated on respecting people.
> They ignored the culture of stopping and fixing problems to get things right the first time.
> They ignored huge swaths of the Toyota Way and created a system that’s less effective and less resilient but can impress shareholders through short-term savings.
> How Toyota has effectively implemented this system fills books but many are just reading the covers.
> Even Toyota though is not perfect.
> In 2011 japan was rocked by a 9.0 magnitude earthquake the fourth strongest ever recorded anywhere.
Not only did this cause immense destruction to life and property but it also led Toyota to recognize a flaw in its own system.
> As japan recovered some supply chains were quick to as well. For example, securing plastic resin for door panel production is not difficult: there are plenty of manufacturers globally creating easily substitutable alternatives.
> That’s not the case with, say, semiconductors: the hugely expensive facilities that create these chips require years to construct and after the 2011 earthquake it took many months to mend them back to operating status.
> This surfaced a truth that had never been fully considered: not all supply chains are made equal. Plastic resin can handle supply chain disruption, semiconductors cannot. Therefore Toyota made changes: all along, their mission was not to eliminate inventory full stop; it was to eliminate excess inventory.
> Supply chain disruption is inevitable. It's inevitable in the same way that Titanic's flawed design would eventually encounter an iceberg, or the structural economic vulnerabilities of 2008 would eventually collide with a market panic. Therefore semiconductor inventory is not excess because inevitably, due to the inevitability of disruption, excess semiconductor inventory will eventually become necessary.
> Recognizing this, Toyota in recent years has started to build up a stockpile of two to six months worth of chips and that's why the company is the only major vehicle manufacturer that is unfazed by the semiconductor shortage.
> Toyota followed its own principles. It did not stray from them, and it did not reinvent them. It’s no surprise that Toyota excels at implementing its own system, but it is a surprise that the entire manufacturing world has so wholeheartedly embraced flawed implementation of the system.
I thought I'd rather watch the video than read your long transcript, I clicked it, it's Wendover with his weird speech pattern (some syllables loud, some syllables he runs through, but some vowels he drawls on...).
> Just IN tiime, is such a SIMple principlee, but the purSUIT of the ELIMination of waste...
He probably got some sort of speech coaching because he didn't like the way he sounded. A bad one, I would say. I looked up a video from 2017, he sounded ok, I could hear some long syllables but they weren't as long as they are now. It also sounds like he got a different mic.
What is amazing about this is that it is a super simple game of "what if?". What if our current supply of chips are disrupted? It would mean production halts, there is a non negligible risk and we cannot source new chips in less than 6 months, so stockpile that to keep uninterrupted service. That conclusion probably take quite a while to get to, but the kick off question is really simple.
It's not that simple, though. "What if our supply chain of X is disrupted?" always leads to a problem where one of the most obvious solutions is "Stockpile huge inventories of X!"
That's what manufacturers had done ever since Ford tried to scale up his initial (very Toyota-esque) operation, and scaled it incorrectly but managed to inspire hordes of other manufacturers to repeat his mistakes.
The novelty of Toyota et al. was not that they asked the what if question, but that they answered it unconventionally: they worked on making the supply chains more reliable instead of adding buffers.
That's what makes this next move counter-intuitive to so many people: when Toyota encountered a supply chain that couldn't be made more reliable, they chose the previously-conventional response, apparently in defiance of their whole thing. Except it wasn't.
You're right in that it is simple, but in trying to show that you're making it too simple.
I think that is the easy way to answer that question. Obviously the first "What if .." should be try to source from somewhere else, or perhaps have multiple supply chains. For instance cloth is a commodity, and if they know they can get the quality they need from multiple vendors on all 7 continents - they don't need to do much except perhaps validate the supply chain of some selected to ensure they don't source from the same place. But if they are using a particular mineral that is only available from one location, something else must obviously be done.
Just stockpiling is not really an answer as much as it's trying to just do the least amount of problem solving and thinking.
Yes it's well know in manufacturing management that there is a trade off between utilisation, capacity and cycle time.
Naively it looks like it's cheapest to get close to 100% utilisation by removing all slack from the system.
If you do that though you can't deal with shocks to the system and it costs you much more in the medium term. In the long term your company dies.
This is very well illustrated by John Sterman's Beer game[1]. That's aimed at supply chains in general but works well in any system you can make brittle via cost over optimisation.
This happens a lot because initially over optimisation makes accountants and managers look good to share holders. By the time the chickens have come home to roost they've moved on to another job.
> Just in time is obviously about not having undue spare capacity,
Also about making spare capacity undue. The lazy solution to unreliable delivery is inventory. The efficient solution is working with the supplier to make delivery more reliable.
The West seems to target a "skeleton" staff for nearly everything these days. Management theory piles it as a best practice under "just in time delivery", but the reality is that there no longer is the distinction between fully staffed and skeleton.
I spent some time in Belgium a few years ago and was amazed to see that the two people on road projects who turn the "slow/stop" signs to control traffic on single-lanes were replaced with robots. So the EU has managed to figure out how to go under skeleton at least.
Meanwhile, when I was in Japan, there were indeed lots of seemingly "useless" people doing "useless" jobs on similar sized projects.
The minute we got off the plane in Tokyo we were met by a greeter pointing people to the escalator. Not that you could miss it, that was the only exit.
In the city you could see people guarding potholes. Bus stations (not particularly busy) would have a couple men at each station that would help arrange the passenger luggage: something usually helped by the driver in the West.
Once we took a night stroll in the city, meeting a roadwork on a deserted street. There were 3 men along the fenced dig basically showing us to walk around it.
Can't see any of that sustainable if you pay people full wages.
Does it make sense from an economics standpoint? Perhaps not. But there are benefits to society as a whole. Quite often the people doing those sorts of jobs are getting on in life, if having that job means the person can stay active that could represent a few additional years where they won't be a burden on the health system.
Honestly, giving people these unproductive jobs seems to me like a way of giving out UBI without people even realising it. Because at the end of the day, these jobs are just UBI. They don't add much to the economy.
You seem to be misunderstanding what UBI is. This sounds more like a job guarantee, which is the opposite of UBI. The main point of UBI is to make sure nobody can ever be pressured into accepting such a job.
Those jobs hardly add any value to society (let alone economy). If you just hand out money instead, some people might find something more useful to do, or at least be better off for themselves. See also David Graeber's essay "On the Phenomenon of Bullshit Jobs".
The overhead of employing properly compensated people instead of signage and traffic cones is fairly high, while the work they do is not necessary in any meaningful sense. Which is moot anyway as apparently these folks are paid a pittance.
The US provides EU-level social safety net benefits (healthcare, housing) to the lower classes—via employment in the military, which is in part a stealth jobs program that both major political parties support paying for.
We provide a jobs program for white-collar folks through our excessively-large and exceptionally expensive healthcare system. We pay for whole categories of jobs that don't need to exist and are only adding costs. A double-digit percentage of our healthcare costs are a white-collar jobs program paying people to do nothing useful whatsoever, and we spend a lot on healthcare so even 10% makes it a fairly big program.
Yeah, I'm aware. It's a really sick way to give working-poor families access to universal healthcare, housing, and free education, which other OECD states mostly manage to do without that. But, it's what we've got.
There are similar measures in many EU countries. For example hiring unemployed people to help kids pass particularly busy streets near primary schools.
That's called welfare, which has an advantage of not masking unemployment. (but to be fair am unsure if people doing traffic cone jobs in Japan are considered employed)
Anyhow my point is it has little to do with system redundancy.
Welfare / UBI ignores that one of the realities of the human condition is that we are social creatures that have a desire for purpose and utility to those around us, and having that purposes in a society that recognizes us helps give us dignity and a place in society.
It's not important that all jobs are high value, it is important that we as a society have a way to create and recognize purpose for individuals to support their mental and personal wellbeing. While welfare / UBI may ensure they have basic needs met to sustain life, it does nothing to support their mental and personal wellbeing, which is why these systems often results in the creation of delinquent behaviors that are mostly absent in a system like that in Japan.
It may be the the same economically, but socially it is much better to have someone doing an "unnecessary" job that is recognize as valued by society vs simply collecting a check.
There's definitely a cultural element. I made the point that these jobs must be recognized as valuable by society for a reason.
You say they're standing there in place of a traffic cone, Japanese society instead says that they're there to provide a friendly face ensuring safety of passersby around the dangers of construction work. A traffic cone cannot assist someone who has trouble walking to traverse rough ground because of the work being done. A traffic cone doesn't smile or acknowledge your presence.
That sociocultural element is what has value, and as long as it has value, the job is valuable.
It's what the market demands in the end of the day and why I don't understand Japan too much (yet). If a company hires 4 people just for road signaling, and another 0, I'd hire the second one since it's going to be cheaper.
That's why I believe these market forces push to at-capacity, because if someone has spare capacity or a lot of profit another company is going to come in at a lower margin and win the market. Yes this leads to breaking at some situations that happen once every 20 years, but otherwise you die from competition so that's what we are stuck with.
> I'd hire the second one since it's going to be cheaper.
You would. But the guy doing the hiring is just giving the job to his drinking buddy, sure that the favor will come back in some way or the other.
There’s also the social factor to consider. These extra people could be superfluous, but if they’re not there everyone walking past will condemn that construction site as being unsafe.
Definitely not (except for the most egregious cases), long-term construction sites in Japan are already fully covered by white panels that don't even let you see inside and short-term ones have at least two barriers and lots of signals. While them not being there does not mark "dangerous", I do think them being there could mark them as "extra-safe".
I would do some nitpicking here: The market _rewards_ low prices. But the market _punishes_ low quality and untimeliness.
Why do I phrase it this way? Because the only market mechanism is a sale, you either buy or you don't. Low price is visible before a sale, so a new sale is reward for the low price offered. Quality and timeliness are only apparent after a sale, so can only be punished retroactively by not buying again.
The punishment signal is also far weaker than the reward signal, lots of goods are not bought too often and problems get forgotten by customers, alternatives may have their own problems, etc. And prices can be compared objectively for standardized goods, and at least easily for other goods. Quality and timeliness are often in the eye of the beholder, subject to variation. Comparisons in those areas are also systematically prevented by most vendors.
> Why do I phrase it this way? Because the only market mechanism is a sale, you either buy or you don't.
That's not always true. When a rail network gets privatized, the argument is that the market will make sure it's a good service still. But they are usually paid by the government, not the patrons. The commuters vote becomes negligible, and the rail network has every incentive to optimize for a minimum viable network that meets their government set targets while the service gets worse and worse in all other metrics. Seats are shittier, service is less clean, graffiti stacks up, railstock gets rickety etc.
That railway example is just a plain market failure, and an uncorrectable one at that. Markets only work under certain conditions, very important among those: absence of monopoly, monopsony and limitations that work to the same effects. A privatized rail network is still the only rail network in the country or region, so still a monopoly. If you hack it to bits and split it up too much, it will be useless, connectivity is paramount in that business. If you leave it as a large network (or several large networks) it will be a national or regional monopoly. Having only the government pay is a monopsony. Limiting building new stations and tracks (which the government will have to do at some point, if real estate prices don't) will also prevent competition. So having a functioning rail market is impossible, therefore any argument about market mechanisms involving railways (or roads, water, gas and electricity distribution) is nonsense: Those can never be proper free markets.
I think you're right, I went through a E40m construction project once and the leaders kept saying how they have to chose the lowest bidder. But that was actually a convenient lie they did to avoid having to do due diligence on the bidders.
Mid and long term cost actually turned out a lot higher than if they had chosen the better slightly more expensive bids.
I've seen this turn out the other way on a road contract near my house. (rural Ireland)
Some Italian construction company underbid, and slightly later went bankrupt after finishing some of the work. They just about missed the contract milestones for payment too, and got nothing out of it.
Then, the contract went to someone else who bid, with expected costs adjusted. Think there was an 8-ish month delay.
The way I see the successful low bids happen is usually the following:
1. Submit a low bid for EXACTLY what's written in the RFP
2. Submit contract delay notifications with contract amendment offers, explaining that the original bid was incorrect and therefore if the amendment is not accepted the project can't continue.
3. rinse and repeat
No. 2 is basically legal speak to make sure you get paid(or win a lawsuit in the event that you don't) even though you can't continue construction.
It's basically saying, you made a mistake in your original RFP, therefore nobody could have continued the construction anyway and therefore you have to accept my amendment.
Good bidders will actually tell you that your RFP was wrong. As an example in the construction project I was in the main architect requested custom spliced decades outdated Fiber wiring, when I finally got a hold of the wiring company and asked them why didn't just do MTP, they mentioned that a) they mentioned to the architect that the request wiring makes no sense and b) they were not allowed to talk to us(the engineering department) directly. Same happened with the fire protection rules from anything like stair handles to the garage.
That's also why I don't think the Airport delays in Berlin were engineering mistakes but rather gross mismanagement.
Also, while in this whole project there is a lot of physical work that you can't optimize away, a lot of the physical work delays and do-overs were also due to gross mismanagement, which is something the toplevel comment completely ignores in his assessment of the 4 day workweek.
If you can provide a product with those features at lower cost the market would still reward you.
But yes any description of economics is necessarily a simplification. For example market segments where people intentionally pay more for the exclusivity that comes with being charged more.
Manufacturing and construction are pretty different segments. When manufacturing, it's very easy to optimize for cost because you know exactly how much material is required and how much time, probably to the second, it takes to process such material into the final form. Even shipping prices are pretty well known and can be optimized for through effective packaging.
Construction isn't like that. Companies aren't cranking out a million buildings a day, so they can't remove every ounce of unnecessary materials; they don't control the environment, making delays a fact of life; they don't control all of the companies and people involved; etc. There's just so much unknown when it comes to construction. Like software, it can be difficult to figure out how much padding to put into estimates. It's a good bit easier when it comes to things like roads, but companies still need to contend with unexpected delays a lot.
> It's amazing because they do finish a lot faster than in western countries, but you can definitely remove 20% of the people in street-level construction and would see little difference.
Maybe the 20% extra people is the reason they always finish a lot faster.
Or the reason everything in Japan is so clean and actually efficient and reliable. You won't see the casual litter and degraded infrastructure that is commonplace in US and Western Europe - everything is kept in tip-top shape, because why not? When you have the wealth and workforce to spare, in practical terms it's just better to allocate capacity for this sort of tasks in quantities that might look sub-optimal on a spreadsheet.
So there's normally 2x-3x people working in construction compared to western countries. I'm saying 20% of those workhours are not needed, but that'd still leave quite a lot more workers there compared to western countries and why they finish early. An example is that they normally work around the clock (24/7, but fairly quietly at night).
Japan on paper is the wealthiest (non-microstate) Asian country GDP per capita, but if you compare by PPP it's actually not great because of meh economic productivity/import tariffs, and slips behind South Korea and Taiwan.
Um, actually construction appears to lay people as "overstaffed" due to safety, breaks, inspections and synchronization of people, tools and materials.
Yes, I understand some people that seem to not be doing much are actually safety people, or taking a normal break, but in the case of Japan it's quite extreme and that's why I pointed it out.
Living in Japan also, the rumor I heard was that these additional workers actually work as a shadow force for political parties during elections.
To circumvent strict donation laws, the idea is that these people will be "lended" as administrative staff and door-knockers during elections, in exchange of juicy contracts.
And let's be real, that linked in [2] does not point to a job that requires automation, a blinking panel is enough. Often I have seen the sad spectacle of one mannequin handling a signal on the end and a human on the other end. It is totally a bullshit job.
Some notably low-staffed places in Japan are Yoshinoya and Sukiyaki (beef bowel fast food, often 24/7). It seems there usually just one person working there at a time.
Not sure about Yoshinoya, but I've heard that Sukiya is considered a "black" company. Means they run over people, understaff and, in general, it's not a place you want to work in unless you're desperate. That was few years ago, though, so not sure how is it, now.
I went once at about 3-4PM and also came away feeling like it was a bit understaffed. One worker and about 8 people in there eating. I didn't have to wait a crazy amount of time so maybe it's fine though, gyudon doesn't take a long time to prepare if some ingredients are prepped
By low-staffed, do you mean that you had to wait long time to get your order? These places are highly automated so they don't need as many staff, sure there's few people working there but usually more than enough for serving the orders that come through in just few minutes.
You forget that there's an etos of work in Japan. Besides of getting some small amount of money, those people get a chance for being useful. That is a mental effect as well and should not be overlooked. Not to mention it reduces unemployment.
If you look around there's plenty of such a "pointless" jobs. The same with people steering the traffic at the parking lot entrances/exits at any bigger entertainment or shopping center. Sometimes with almost no pedestrians around and yet those people are standing there in sun or rain alike. Or people who's only job is to smile and open doors for you, less common, but happens as well.
To be honest, though, as a "consumer" of such jobs (pedestrian in this case) I do feel they took an extra mile to make it as safe and comfortable for me as possible. And I do feel better about it.
It may look pointless at the surface, but I think those jobs are extremely important for society at large.
I don't know if Japanese construction projects are overstaffed or not, but if they are, do you think that would change? My guess is that they would continue the same staffing arrangement and get less construction work done.
They even have signs that look like people that day watch out for the sidewalk that stand next to the guy that says watch out for the sidewalk. But everyone who wants to work should have a job.
It's amazing because they do finish a lot faster than in western countries, but you can definitely remove 20% of the people in street-level construction and would see little difference. An example: they hire 2 people on every sidewalk just to tell people to continue walking the way they are walking and make sure they don't fall into a pit while looking at their phones[1].
From what I've learned, this specific job is normally filled by people who have already retired but don't have enough to live, so they are paid peanuts (I believe they might also be paid a bit extra by gvmt to go back to work, but I'm not sure I understood this point properly). They have started to be replaced by "robots" though [2].
Edit: in general everything in Japan seems overstaffed compared to the western counterparts, not just road construction. I believe this is how they can maintain very low levels of unemployment here, but I don't yet know how the incentives work to make it possible.
[1] The guys with the blue shirt, notice one on each side https://c8.alamy.com/comp/HBJ2K1/japanese-people-in-construc...
[2] https://japangasm.files.wordpress.com/2010/07/rwintro-swing0...