Here are a couple other phrases the Original Poster should know. The "first mover advantage" is when one of the first (successful) firms takes advantage of economies of scale - mentioned above - to lower costs. This discourages new entrants. The other is "learning by doing" where production teaches the firm better methods, thus improving technology and lowering costs. Which further discourages new competition.

@Daniel One could summarize your point by saying that building the next generation of a semiconductor fab/foundry costs only $20 billion if it is done by TSMC with the expertise and knowledge of the employees in Taiwan who build the current generation. If anyone else tries to do it, it will be way more expensive or more likely impossible.

Even worse - the factory needs a lot of supply infrastructure (other companies) that generally do not exist in the wild, so there is a very strong push to put a factory where you find the specialized firms which is - where other factories are.

I do not remember whether it was Thailand or Taiwan, but some years go a valley got flooded - that had 5 HDD factories in it, which all stopped producing for some time. Not a LONG time ago. You know how many factories worldwide existed making HDD? 5. Specialized suppliers available.

Even those $20 billion (or increase it with an order of magnitude or two if you want) are a tiny fraction of the budgets of some of the biggest countries. Wouldn't it be rational for them to invest even at a loss, as an insurance? If something happened to Taiwan it might lead to the total collapse of our modern civilization.

@TomTom: That was Thailand, and no, that was definitely not the entire world production in that single valley. A sizeable fraction, yes, and it did cause a price hike.

The listed reasons seem rather speculation based on what happens to be in Taiwan -- as opposed to how Taiwan's semi industry outcompeted alternatives in other geographies. (Which would then feed into the steady-state explanation offered at the beginning.)

You could say there was a reason why someone had to get into a strong position, quite automatically. And because of some coincidences, you are asking why Taiwan is in this position, and not for example South Korea, or maybe Croatia.

@TomTom The supply infrastructure might be the dominant element which locks in their success. I remember a senior Intel person writing about their vast supplier network, and how hard that made setting up large-scale production in new places (tried to find link, but Google-bombed). Just as one point, Taiwan's semi industry consumes ~10% of their water (1m tons/wk?); generally, the chemical needs are broad, exotic, and often nasty.

For all the items you mentioned about Taiwan, didn't the US do similar things and have similar traits?

@robotcookies I actually addressed that in my answer. Many places had all those things. South Korea, USA, NL etc. as I mentioned in my answer some things come down to luck. In real life events are not deterministic. For example, in graduate admissions you might end up with 100 excellent equally good candidates to fill 30 spot. Which of those 100 get accepted will ultimately comes down to luck. Or you can always smoke and never get a lung cancer. Similarly if there are 10 places with suitable conditions but world economy can sustain only single big semiconductor company it’s down to sheer luck

Or even better example, if you purchase lottery ticket you have 100% mathematically equal conditions for winning the pot than every other ticket holder. Yet satisfying the preconditions is by no means a guarantee of winning.

One other factor that might have played into Taiwan dominating is the proximity with China, especially Hong Kong and the Shenzhen region. Taiwan is in a unique position to be close enough to the cheap manufacturing labor of China while also having the political landscape to allow the heavy investments needed for semiconductors to be safer. Think Apple, get the chip from TSMC and then move them quickly to Foxconn for cheap assembly. It would definitely be less convenient if they had to move from Thailand or Korea, since those are further away.

@bracco23 it may be one other factor, but I doubt it has had a large influence. Chips are among those high-USD/kg-density commodities that manufacturers may choose to ship three times around the globe between various stages of the assembly process, if there is but a tiny cost advantage to be had in each stage.

Any answer that doesn’t discuss how Morris Chang invented the concept of the pure play foundry when everyone thought it was utter bollocks. He identified it as TSMCs only opportunity given they lacked IP, design and even sales skills. And now TSMC, not Taiwan, dominates world semiconductor manufacturing. TSMC is so dominant it has made more than 100% of all the open fab profits during the entire history of the market.

Here is Morris Chang explaining why the foundry business was uniquely suited to Taiwan. semi.org/en/Oral-History-Interview-Morris-Chang

@SafeFastExpressive can you please provide link to any peer reviewed source mentioning Morris Chang as being responsible for this? In fact in his interview you linked he pretty much summed what I have written in the answer. It does not seem like this is as all just due to effort of one person. Economics stack exchange is science site we do not got here by rumors but by hard scientific evidence

@1muflon1 You want me to find a peer reviewed study of whether Morris Chang was the founder of TSMC and responsible for its major strategic and technical decisions? I wasn't aware that historical fact needs peer-reviewed studies. If you want to ignore Morris as the one most responsible, and just focus on Taiwan's available resources as Morris delineated in his interview and it clearly dictates why the pure play foundry business model took root in Taiwan, the value of the model is why it grew so rapidly.

@SafeFastExpressive sorry but this is fallacy. I asked for peer reviewed source for him being responsible for Taiwan being the worlds major manufacturer of chips not for him doing strategic decision. Look you are clearly not scientist so you don’t understand. Just because Chang was CEO of the company that does not mean he is responsible for company success and when we talk about whole country it’s even less likely. It is completely possible that if another manager would be there Taiwan would do even better and so Chang made everything worse

For example, is in US Biden responsible for all inflation for low economic growth and terrible stock market? I would not say that he is responsible at least for sure not for everything because inflation would likely pick up even if trump would be president as inflation is likely caused by too much increase in money supply during Covid. At best he could have had marginal impact. Yet according to you Biden is president so anything currently happening in US must be his doing. That is profoundly anti science view. The same holds for Chang, just because he was CEO does not mean he did help

@1muflon1 Not sure if your grasp of science is as strong as you think. TSMC is by far the largest semiconductor company in Taiwan. It controls about 50% of the worldwide foundry market and is close to 50% of total Taiwanese semiconductor revenues. And a significant portion of the remaining companies are successful because TSMC created the fab business (UMC) or because they benefitted from providing complementary services to TSMC. That is basic math.

It's factual history that Morris Chang built TSMC into one of the world's largest by pioneering the open Fab business. It required a leader with a deep experience in semiconductors who would pick a radical new business model. It would be like saying Reagan's tax cuts were inevitable when not only his opponent Jimmy Carter was directly opposed to them, but even his leading Republican primary opponent. Only Reagan could have passed his tax cuts, and only Morris Chang could have built TSMC. Burden of proof is on you to show who else would have done it anywhere near as well.

@SafeFastExpressive You do not know what you are talking about. Please have a look at this effect. Again none of it is argument that Chang contributed to the success. What you need to show is that given potential outcomes $Y_i$ (defined as potential semiconductor production in Taiwan), where $i$ is 1 if Chang is CEO of that company and 0 when he isnt, next let C be a dummy for Chang being CEO then what you need to show is that $E[Y, C=1]- E[Y, C=0]=E[Y_1, C=1]- E[Y_1, C=1]+E[Y_0, C=1]- E[Y_0, C=0]$ is statistically significant and

positive in addition to showing that the selection bias $E[Y0,C=1]−E[Y0,C=0]$ is zero

" Burden of proof is on you to show who else would have done it anywhere near as well" <- this is deeply anti-science. Burden of proof is on those who make positive claims not on those who make negative claims. You can't even prove negative empirically, only positives can be empirically proven! Thats like saying prove god does not exist. Burden of proof is on those who say new drug works not on those who say it doesnt. Taking new drugs assuming they work until they are proven harmful is recipe for disaster. It is on you to show that the Chang intervention had significant effect not on

people who say it did not. Where is even any evidence that Chang was inventor of that model. also you say this is simple math. If it is so simple, tell me how much in number of units or if not in percentage terms did Taiwan production increased just due to Chang? Also I expect some proof it was due to his direct intervention not just saying because he was CEO. You could put random person off the street and make them CEO of Microsoft and they would still sell more software amidst period of growing demand.

You need to show that Chang had positive effect, conditional on things like institutions, industrial policies, Taiwanese subsidies, openness, demand changes.

Ok, I got it. Morris Chang founded TSMC and over his 30 years averaged over 15% annualized revenue and profit growth, which is an extreme outlier in corporate performance, but any random person off the street could have started TSMC from scratch and done the same. It's just like taking over Microsoft because anyone can build a massive competitive moat very easily.

And it would be easy to refute my arguments. Show any open fab business that existed before TSMC that had any significant success. That would clearly show that Chang didn't do anything innovative and wasn't necessary to creating that innovation.

@SafeFastExpressive no it is not easy. If there would be such company that does not mean he did not had an effect. Again you have poor understanding of causality as we scientists understand it. The problem is that you need to see counterfactual, how TSMC would do under different manager than Chang, of course that is impossible as we cannot travel back in time and replace him but clever scientific methods can fish out this effect even from combination of observational data and other techniques, but you cannot claim that there is effect just because he was there

@SafeFastExpressive okay so over those 30 year he was responsible for all that 15% annualized revenue and profit growth? Where is your source for that? All TSMC employees showed 0 initiative. All decisions were purely due to him. All other execs were just drones? Also, all those subsidies from Taiwan 0 effect? Since according to you its all just due to him? All that good business environment 0 effect? You are claiming that when it comes to revenue $E[Y,C=1]−E[Y,C=0]=0.15$, while bias is 0. Again its not up to me to disprove you. If someone says they saw aliens its not up to me to prove

aliens do not exist. Again, your argument is equivalent to an argument, children got vaccine and later we saw it developed autism, so it must be vaccine that gave children autism. No, no, no. What you need to figure out is what would happen to the same children if it did not got a vaccine and compare it to what happened when it got a vaccine. If autism would occur in counterfactual where kid did not get a vaccine its not vaccine that caused autism. Also large claims require large amount of evidence if you make claim like success of whole country in one industry is due to single person you need

strong evidence

@1muflon1 Berkshire Hathaway has roughly a 20% annualized return since Warren Buffett took over as CEO. As an economist, how would you measure how much of that outperformance was due to WEB's efforts?

@1muflon1 Getting back to Chang, was he not responsible for the TSMC culture that enabled executives and employees to make their contributions? I've worked at organizations where management pushed down decision-making and authority to gain the contributions of as many employees as possible, and for organizations where decisions were top-down and employees/mid-level managers were only expected to carry out orders like drones. How do we establish that the Chief Executive Officer (Chang) does or does not get credit for fostering an environment that enabled the organization to outperform?

@SafeFastExpressive "As an economist, how would you measure how much of that outperformance was due to WEB's efforts?" For example, by synthetic control method. Collect data some performance measures, here revenue is actually red herring, we care about chip production, so the performance measure should be growth of output. Collect data on output growth of similar technological firms, there aren't that many chip makers so you would need to go little bit beyond just that. Then collect data on all important control variables, size of the company, level of state support, institutional quality

@SafeFastExpressive "Getting back to Chang, was he not responsible for the TSMC culture that enabled executives and employees to make their contributions?"Good question was he? How do you know he set up all the work-culture rules and not his HR manager? Do you have any research showing he was the one who set the work-culture at TSMC?

@SafeFastExpressive "How do we establish that the Chief Executive Officer (Chang) does or does not get credit for fostering an environment that enabled the organization to outperform?" Good start would be to find some evidence that some extremely important things like the fabless business model were his idea. That still does not mean that Chang is the sole factor, but that would show he had significant impact.

Also good counter example is Elon Musk, he was CEO of many companies such as Tesla and paypal, but when you actually do research on what he does in the companies you discover that all the heavy lifting is done by other execs and he is basically just showman/salesman like Saul Goodman. Moreover, his erratic 'showmanship' seems to even hurt tesla since his tweets several times affected the company negatively (which is why board of directors was trying to get rid of him). Yet I am sure that his many fanboys would swear on theirs mothers graves that all these companies are successful due to him

I am not saying Chang is equivalent to Musk, the point is just to show that firms can be successful despite having CEO that does not help much. Next, there is research showing that quality of CEO on average improves profitability, but your claim is that Chang is the X factor that is responsible for all success, that is other high quality CEO would fail where Chang succeeded. Your claim implies if TSMC would be run by another smart MIT educated CEO it would be failure, but Chang was the chosen one who made Taiwan success inevitable, that is extraordinary claim that needs strong support

@1muflon1 Elon Musk is a great example. In the case of Tesla, he provided all initial startup funding. He had been funding JB Straubel's research into battery development, and convinced the founders to hire JB as CTO, becoming the most important Tesla hire. Then elon led every funding round after that, which was key to their success. Finally, when Martin Eberhard's roadster plan would have had a \$250k build cost, Elon saved Tesla from bankruptcy by firing Martin and replacing him, to lower build cost below \$100k. Elon is clearly the person most responsible for Tesla's success.

And the case of SpaceX is another great example, Elon founded it, provided all of its startup funding, and put all of his remaining wealth into it. He recruited one of the greatest rocket engine designers of all time, Tom Meuller, to build the Merlin. And Tom has publically discussed how Elon constantly challenged him and his team on critical decisions, and in several cases forced them to make changes that ended up improving the engine substantially. Tom Meuller himself would tell you that SpaceX would never have succeeded without Elon Musk.

@1muflon1 When you attribute corporate culture to "HR", I think it demonstrates an out of touch academic perspective. HR is just a liability minimization department that establishes rules/procedures/pay scales/benefits that best match the CEO's goals/objectives and preferred culture. Everyone reports to the CEO, either directly or indirectly, and he/she decides what key priorities are. There is no chance that any strategic effort or key technical investment occurred at TSMC without Chang's approval during the 30 years he was in control. He hired the key VPs. He made the important decisions.

@1muflon1 And we don't need a "synthetic control method" to know Warren Buffett is responsible for Berkshire Hathaways outperformance. The outperformance was caused by intelligent resource allocation (acquisitions, stock picking, capital allocation), and as CEO WEB is in charge of resource allocation. As further evidence are 50 years of annual meeting Q&As where he discusses his resource allocation decisions. In the case of Morris Chang, we have numerous interviews where he discusses inventing Fabless semiconductor model, investment decisions, management, etc and no one disputes his account.

@SafeFastExpressive Yes he provided funding which could be done by any VC. Also, if he would be good for the company the board would not constantly try to get him out and even sued him.

@SafeFastExpressive "There is no chance that any strategic effort or key technical investment occurred at TSMC without Chang's approval during the 30 years he was in control." again empty statement unless you prove it

@SafeFastExpressive actually funny you mention that "And we don't need a "synthetic control method" to know Warren Buffett is responsible for Berkshire Hathaways outperformance." <- you do need some method, not necessarily Synthetic control but something for that

as far as we know beating market in long run should be nearly impossible due to EMH and there is lot of statistical evidence for at least the weak EMH version. So for all we know Warren Buffett is just example of survivorship bias en.wikipedia.org/wiki/Survivorship_bias

@SafeFastExpressive "When you attribute corporate culture to "HR", I think it demonstrates an out of touch academic perspective" <- I did not attributed anything to anything here. I am just giving plausible counterexamples. All this demonstrates is that A) laymen have poor understanding of causality B) laymen have poor understanding logical fallacies