Not very, judging by the recent bankers' balls-up. And if we look at the solar panel industry - oversupplying its market by 100% this year but still ramping production - you wonder how sane capitalists are.
The DRAM makers have been on a suicide mission called Glut for the past several years leading to government bail-outs in
If ten DRAM companies each invested for 10% market share, the result would be a perfect market.
Instead ten producers invest for 15% market share and the result is bankruptcy and bail-outs.
Yet the solar people led by market leaders Sharp and Suntech are still expanding production. Will governments have to bail out the solar guys?
The banks are the grossest examples of capitalism's inefficiency. Apparently the J P Morgan people, who in the 1990s invented the credit default swap, one of the main instruments of the financial devastation, and who understood its risks, were appalled to find, a decade or so later, how extensively and ignorantly it had been used by other banks which did not understand the risks.
But the copycat nature of capitalism, where each company blindly follows the others, led it to disaster.
Both Chairmen and both CEOs of both HBOS and RBS had no banking qualifications. They were businessmen who followed a trend they did not understand to disaster.
In the microprocessor market we are told that competition works well. But why does an Intel microprocessor cost around $100 while an ARM processor, the world's best-selling 32-bit microprocessor, sell for a dollar or two?
Answer: because anyone can make an ARM processor. whereas only two people make x86 processors.
If there were ten suppliers of x86 microprocessors, each of whom were investing to get 15% of the PC processor market, think how low the cost of an x86 microprocessor would be.
And think how low the cost of a PC would be if you could take out the $100 cost of the processor.
The x86 duopoly of Intel and AMD came about largely by accident. In a normally working chip market there would a lot more than two suppliers of any essential component.
So the x86 market is another market where capitalism is not very efficient.
If efficiency means making quality goods at the lowest cost.
Well the thing is, incentives matter.
Where does the incentive to ramp up production in solar cells come from? Surely not from the Government, with it's massive taxpayer subsidy of the renewable sector based on (in engineering terms) very dubious evididence that this is an efficient way to spend the money.
So with all that cheap taxpayer money being squeezed into the sector, it has to leak out somehow. In this case, in the form of structural overcapacity.
I guess you know more about the DRAM industry that I, but the point I would make is that a situation where each company was investing the same amount in capacity would require information sharing between the manufacturers.
This is known as a Cartel.
So it may be messy and appear to be inefficient, but it's the competitive nature of Capitalism that serves to make us all richer in the end.
But, DWL, in the sense that we'll be paying for inefficient capitalism via lower public spending, reduced public services and higher taxes in the years ahead, to pay back the debts incurred to bail out the inefficent bankers, we are all poorer.
How can lots of companies going bust be efficient? Wasted plant, machinery, Admin costs, employment costs, social costs in terms of the time taken for redundant workers to find new work as well as the cost to suppliers who have cancelled contracts to deal with.
Wouldn't the optimum efficiency incorporate both competition and some sort of oversight by govt?
Also the banking sector has just been bailed out on a global level in a way that is curiously reminiscent of - dare I say it - socialism (keynesian economics)?
Well John, I reckon you have to be right. There's a massive waste of plant, energy, resources and social expense when businesses go under. And no one, anywhere in the world has figured out a way to handle it. China's workers rioted in the credit crunch when export markets dried up, and factories were closed, and many got shot and imprisoned for doing so. So even the Communists haven't figured out an answer to this one. But an answer is very clearly needed, as you say.
PS. John - have just noticed a story how an NXP fab will be sold to a solar panel manufacturer. It raises the possibility of legislating that, when a company needs to close a plant, it is obliged to consult the public authorities about alternative uses, and putting an obligation on the authorities to investigate the possiblities of a transfer of use and to fund the transfer where necessary.
"It raises the possibility of legislating that, when a company needs to close a plant, it is obliged to consult the public authorities about alternative uses"
This does exist in some countries.
"If there were ten suppliers of x86 microprocessors, each of whom were investing to get 15% of the PC processor market, think how low the cost of an x86 microprocessor would be."
Actually if there were 10 supplies each investing to get 15% of a much smaller market we'd still be on single core 2GHz processors. One supplier isn't enough but 3 is probably one too many.
Also comparing an Intel processor with ARM, if the ARM is $2 then on a performance basis the Intel should cost from about $20 to $250, depending on the model, and this isn't too far off reality.
No-one said capitalism was perfectly efficient: Keynes and Schumpeter ('Creative Destruction') both wrote eloquently about just this
In some markets the system, in the long run, on average should end up at some optimum. In others you always get over-capacity cycles with boom'n'bust.
But, rather like democracy, it is the worst system until you see all the others...
Wise words, El Rupester, but it's interesting that Shanghai, in a communist dictatorship like China, can have better networks than San Francisco, in a capitalist democracy like the USA.
Possibly analagous to "... he made the trains run on time"
Although I think networks in Tokyo or Seoul are even better than those in Shanghai...
Aren't there more millionaires in Shanghai than in San Francisco as well :-)
Well Yes, El Rupester, I've heard distinguished American techies argue that exact same point. But it always begs the question: 'Do the people runnning Western wireless networks feel obliged to offer as good a service as is technically possible?'
Mike, in the 'which came first the chicken or the egg' scenario you have to ask how soon BT flagged up the IC requirement. Perhaps they felt the IC companies should develop chip-sets ahead of the networks' deployment schedule. But I seem to remember IC companies launching ADSL chip-sets several years before BT offered ADSL services.
I was there David and BT and all the other major telcos flagged up the IC requirements in 1993-5 with work on the ANSI T1.413 specification for DMT based ADSL.
I tested every one of these chipsets as they were launched and all were nowhere near the required spec until about 1999 when rollouts began. Even then there were some problems and it was accepted that older line cards and terminal equipment would have to be binned but things couldn't wait any longer.
I'm not saying IC companies are always behind the game but in ADSL they were the problem.
That's extremely interesting Mike. Riffling through our archives I find that, in the summer of '94, BT announced they could deliver 2Mbits/sec ADSL using CAP modulation with hardware boxes from Westell and Atlantech, and ICs from ADI.. DMT modulation, though superior, was seen as too power hungry. Anyway BT trialled ADSL extensively in 1994. BUT - despite ADI having the chips, and Westell having the boxes, ADSL was't extensively rolled out by BT in the UK until 2002/4. So the question remains: Why not?
I was with Westell at the time and as you said we were shipping CAP based boxes which were used for the BT and other trials worldwide.
However the standards committees had decided that DMT was the way forward, a decision that was correct in that it allowed for the systems we now see where almost every person has broadband, but was difficult to implement at the time for the IC makers. Indeed Westell was forced to use Globespan CAP modulation ICs for some later trials in 1997 and 98 and replace them with DMT units later.
ADI certainly never had DMT chips that worked back them - it was Alcatel Mietec who cracked DMT first. First rollouts of these occured in 1999 and from 2000 onwards BT rolled out the system as fast as its suppliers could ship it the equipment to install. Over the years I met with hundreds of BT staff and the determination to solve the problems and get ADSL rolled out was definitely there and they were often as frustrated as their customers with the delays on an area where BT was seen to be a world leader.
That is most interesting Mike, a bit of history.
My view of the history is somewhat different from Mike's -- but then I'm biassed ;) ADI & Alcatel were pretty much neck & neck, and the ADI DMT chipset was in BT trials in '97.
Certainly I had ADSL at home in early 2000 which is pretty quick.
That said, I agree with his broader point: there is a big difference between having a chip; having that integrated into a stable product -- and all of that and a whole network.
For example, you mention WiMAX. You could buy a WiMAX dongle two years ago -- but what is the point if the spectrum isn't allocated, or the basestations aren't there yet. Everything needs to come together.
That is different to a new CPU which runs the same old software, and doesn't have any other dependcies.
And, to go back to your original point, I'm still not persuaded that capitalist economies are any worse than other systems in that respect.
What amazes me about ADSL is that it took over 10 years to perfect the technology, BUT less than 8 years for it to become a worthless commodity socket.
Today the prices for a CPE ADSL2 chipset is below $1USD (the whole semiconductor function). Alcatel is long gone and the market is owned by BRCM but is slowly being taken from them by Realtek and Trendchip.
Truly sad....
Well, El Rupester, I suppose for most of its existence telecommunications in the West wasn't subject to capitalistic forces - it was a government owned monopoly. So it's probably a bit unfair to blame capitalism for the shoddy networks. With AT&T and BT not being de-regulated until the 1980s, that old mind-set of a government monopoly - a kind of branch of the civil service - has lingered on to this day. And the BT privatisation was poorly done - the network remained under BT's control, which hasn't really made for a level, competitive, playing field. Maybe the answer is that capitalism hasn't yet been given a fair crack of the whip in providing telecommunications services.
Surely Robert this is the natural result of innovation ending. if the chip companies couldn't find ways to improve the chips, then the chips that were made were doomed to commoditisation. Whatever happened to VDSL?
VDSL still exists but it is only used in Japan (and maybe Korea I'm not sure), however Ikanos is much healthier than most of the ADSL players. IKAN also acquired Conexant's DSL business (which was based on the Globespan chipsets), they got the business for pennies....
Thanks for that, Robert
VDSL is currently used in many European countries for last 100m delivery from FTTC installations but generally it is kept for business use where the cost of the fibre backbone installation can be justified.
There have been countless discussions on domestic rollouts but the business model generally doesn't work.
Thanks for that Mike. Strange they can't make money at it. Virgin is constantly offering faster broadband speeds for more money on the monthly sub.
That's because Virgin picked up the cable networks for a knock down price after they had spent a fortune from their investors destroying the UK road network, oops I mean installing their cable ducts. BT often don't have accessable ducts to the street cabinets so would need to get digging.