Have you heard of the recession? OK, now have you heard of the relationship between supply and demand?
Whoa, Chai! Answer your spam and order some chill pills.
First, DRAM pricing has not always followed the classic price elasticity of demand model, partly because of illegal price fixing:
http://forums.anandtech.com/me...=2336994&enterthread=y
And, the DRAM market is strange anyway--not that I minded picking up some top of the line memory earlier this year for about $10 a GB--possible less than it cost to make it, after figuring in capital investment in the DRAM factory, finishing, shipping, retailing costs, etc.
An older analysis, but, still relevant:
http://www.crn.com/it-channel/...MEZQE1GHRSKH4ATMY32JVN
MarketWatch: DRAM Market Defies Conventional Wisdom
Recently, San Jose-based researcher Dataquest Inc. put another long-standing relationship to the test and found that, contrary to conventional wisdom, theory and facts don't jibe. In 'Killing the DRAM Elasticity Myth,' Dataquest analysts investigate the widely held assumption that the market for memory modules exhibits a high degree of price elasticity. In economic theory, demand for goods or services is said to be elastic if the quantity demanded is sensitive to changes in price.
A simple plot of DRAM shipments and price against time certainly supports the notion that demand is highly elastic. Indeed, throughout this decade, megabyte shipments accelerated even as price per megabyte declined, albeit haphazardly during some periods. Moreover, a simple statistical test establishes a tight correlation between quantity demanded and price. While professional bean counters know that correlation does not demonstrate causality, lay observers assumed that there was indeed a functional relationship. In other words, folks jumped to the conclusion that falling prices were driving sales of memory chips.
That ain't so.
Dataquest took a hard look at the numbers and showed that the DRAM market is far less responsive to price changes than believed. Coincidentally, semiconductor researcher Advanced Forecasting Inc. (
www.adv-forecast.com), Cupertino, Calif., last week announced a new forecasting model of DRAM average selling price.
The model is based on a proprietary set of 30 monthly economic factors that do not include any DRAM data. Just as the Dataquest analysis indicates that quantity is not a function of price,
Advanced Forecasting's research shows that price, in turn, is determined largely by variables outside the DRAM market.
There's also currency fluctuation (I recall seeing on this forum a user from Australia surprised that prices there didn't reflect the discounts we were getting in the US, and someone pointed out that their currency had fallen--something I am afraid we have to look forward to here, since we printed, oh, a few trillion US-Zimbabwe dollars over the past few years.)
And, there is this, which makes sense, if you recall when DDR3 RAM first came out:
JUNE 11, 2003
Samsung's "Sashimi Theory" of Success
CEO Yun Jong Yong says getting hot products to market before they become "dried fish" lets the electronics giant fetch high prices
Yun Jong Yong has a vivid way of describing his survival strategy in today's hypercompetitive consumer-electronics market, where today's hot-selling novelty can become a cheap commodity within months. The vice-chairman and CEO of Samsung Electronics calls it the "sashimi theory." When prime fish is first caught, it's very expensive at a top-notch Japanese restaurant. If some of the fish is left over, it sells the next day for half the price at a second-tier restaurant. By the third day, it goes for one-fourth the price. "After that," Yun says, "it's dried fish."
So the secret to success in consumer electronics is getting the most advanced products onto retail shelves ahead of the competition. That way, you can charge premium prices -- until the rest of the pack catches up and the product is no longer fresh. "In our business, there's money to be made if you can reduce the lead time to customers," Yun says. "If I can reduce that time by one week, it makes a big difference. If you're two months late to the market, the game is finished."
[/b] Yes, there's a small supply at the introduction of a product, but prices can go crazy--like when certain hot cars come out and are sold for way over sticker price. See reference to "popular delusions" below.
Also, not to get boring, there's the large up front costs of investing in fabs, and training workers--huge up front costs. But, once a fab is up and running, the marginal cost of making more RAM (let's see, some underpaid workers, and some raw sand--silicon) is very low--
But, farmers have been doing this for a long time: the price of a crop X is high, so everyone goes out and plants crop X--and the next year, the price is low, since there is so much X--yes, price elasticity, but with the added production/investment factor--which can make prices cycle despite simple supply/demand factors. Note: a few years ago there was an article on lemon pricing--prices go up, lots of trees get planted, prices go down, trees get pulled out and replaced with other citrus--I think is was an 8 or 9 year cycle....
"Popular delusions and the madness of crowds": great book on pricing, and bubbles....remember when investing in DRAM manufacturing was like printing money? (Micron, in Idaho....) Now it's the fast track to printing IOUs.... Housing prices in the US? Partly supply and demand, partly employment, cost of money/mortgages, partly pricing bubble.
Anyway, there is price elasticity, but other factors can put some strange oscillations in the pricing.
Oh, think airlines: high capital costs, high operating costs, demand extremely variable--and then they make it painful with their weird pricing with saturday layovers and early buy and last minute buy, with powerful algorithms to maximize the per seat per mile revenue--
and still they lose boatloads (not planeloads) of money.
Supply and demand...wish it were so simple.
NX
Edit: horrific spelling errors, bad cut and paste corrected....
Facebook
Twitter