Online information source for semiconductor professionals

Guest blog: A Perspective on the Impact of Japan‚??s Disaster

30 March 2011 | By Mark Osborne | Editor's Blog

Popular articles

Oberai discusses Magma’s move into solar PV yield management space - 29 August 2008

‚??Velocity‚?? the new buzzword in Intel‚??s PQS annual awards - 12 April 2012

Applied Materials adds Jim Rogers to Board of Directors - 29 April 2008

TSMC honors suppliers at annual Supply Chain Management Forum - 03 December 2008

Sematech Litho Forum: Sematech mulling multi-beam mask writer effort - 12 May 2010

This guest blog was provided by Bill McClean, IC Insights, covering new analysis of the impact on the semiconductor industry from the recent disasters in Japan.

One thing we have learned in over 30 years of analyzing the semiconductor market is that when the forecasting environment is at its most confusing state (e.g., the memory market bubble of 1984, the internet bubble of 2001, the financial collapse of 2008/2009, etc.), the best thing to do is go back to basics.

These “basics” are the underlying trends and assumptions that are already in place. By revisiting these trends and assumptions, one can step back from the confusion and gain some perspective on the situation. While a forecast may still need to be changed, putting the situation into perspective helps prevent the emotional “whipsaw” reaction of drastically cutting or increasing a forecast only to find that it needs to be significantly changed again a few weeks or months later.

Bill McClean, IC InsightsThe following is a review of some of the basic assumptions concerning the current global situation following the earthquake and subsequent tsunami that struck northern Japan on March 11, 2011.

In an attempt to assess the impact of the Japanese earthquake and tsunami on the semiconductor industry at this early stage, one can only make some assumptions regarding the most basic of business issues: supply and demand. Taking a top-down approach, the first consideration is worldwide GDP growth and how it might be affected by the current situation in Japan.

Effect on GDP

Presently, Japan’s 2011 economic growth is essentially impossible to predict. Early estimates indicate that the earthquake and tsunami damage will cause Japan’s GDP to decline by 1-3% in 2011. Most economists agree there is little doubt that Japan’s economy will take a significant hit in the first half of this year and enter a recession (i.e., two quarters in a row of negative GDP). Moreover, some forecasts believe that Japan’s 2Q11 GDP could decline up to 12% (in comparison, during the worldwide financial shock, Japan’s 1Q09 GDP shrank by 5.4%)!

However, over the long run, advanced economies facing significant disasters have oftentimes performed better than initially expected as increased investment outlays to rebuild infrastructure help foster strong economic growth. This rebuilding GDP growth spurt is expected to begin in Japan in 4Q11.

As shown in Figure 1, Japan represented 7.5% of worldwide GDP in 2010. In what might be considered a “worst case” scenario, Japan’s GDP declined 5.2% in the severe global recession year of 2009.



If the Japanese economy declined 5.2% in 2011 (not our current forecast), and the GDP growth rates in the other regions of the world were kept the same as shown in Figure 1, then 2011 worldwide GDP growth would drop from our January McClean Report estimate of 3.9% to 3.4%.

A GDP growth rate of 3.4% would result in a US$260 billion shortfall in worldwide GDP as compared to our 3.9% forecast presented in January’s McClean Report.
It should be noted that although Japan’s economy this year may not register a 5.2% decline as shown above, IC Insights believes that the resulting supply chain disruptions from the earthquake could still cause worldwide GDP to drop to 3.4%, and still result in a US$260 billion shortfall in worldwide GDP this year as compared to our original 3.9% forecast. IC Insights’ worldwide GDP forecasts are summarized below.

•    January McClean Report worldwide GDP Forecast — +3.9%
•    IC Insights’ current worldwide GDP Forecast — +3.6%
•    IC Insights’ pessimistic worldwide GDP Forecast — +3.4%

A worldwide GDP growth rate of 3.4% would be 0.2 points below our current forecast and the long-term average worldwide GDP growth rate of 3.6%. However, in our opinion, this level of growth would still be sufficient to support IC Insights’ current semiconductor market growth forecast of 10%.

It should also be noted that in the first few days immediately after the earthquake, the price of oil dropped from over US$100 per barrel to US$97 dollars per barrel. However, since that time, oil prices have moved up to over US$105 per barrel (due mostly to the situation in Libya and speculation).

As shown in Figures 2-8 of the January McClean Report, oil prices have historically had a good correlation with worldwide GDP growth. The lower the oil price, the better the outlook for worldwide GDP, which in turn would serve to offset some of the negative effect on worldwide GDP from slower economic growth from Japan.

However, oil prices to well above US$100 per barrel would serve to further slow worldwide economic growth from our current forecast.

One estimate now making the rounds among the economists is that for every US$10 increase in the price per barrel of oil over US$90, worldwide GDP growth would be reduced by 0.2 points. Thus, if oil averaged US$100 per barrel in 2011, it would drive down our current worldwide GDP forecast from 3.6% to 3.4%.

Effect on Electronic System Sales

Moving one step down from worldwide GDP, electronic system sales were $1,237 billion in 2010, which represented only 2.2% of worldwide GDP (Figure 2).



Taking the pessimistic situation of a 3.4% worldwide GDP growth rate in 2011 (as compared to our current forecast of 3.6% growth), and the associated US$260 billion negative impact on worldwide GDP, and multiplying it by 2.2% yields an electronic systems sales loss (as compared to our January McClean Report forecast) of $5.7 billion. Subtracting US$5.7 billion from IC Insights’ current 2011 electronic system sales forecast of US$1,348 billion would put electronic system sales at about US$1,342 billion for this year, an 8.5% increase over 2010 compared to our current forecast of 9.0%. This 8.5% increase would still be much better than the 6.0% long-term growth rate for electronic system sales.

IC Insights believes that any negative impact on electronic system sales from the current situation in Japan will only delay those sales, not destroy the demand for the systems. Thus, any shortfall in electronic system sales experienced in 2011 due to the Japan earthquake is likely to be gained back in 2012.

IC Insights estimates that the Japanese market (i.e., consumption) represented approximately 9% of the world’s electronic system sales in 2010, which equates to about $111 billion. Thus, if the entire $5.7 billion potential electronic system sales shortfall was due to a decline in demand from the Japanese market, the 2011 Japanese electronic systems market would decline about 5% to $105 billion.

For 2010, JEITA (Japan Electronics and Information Technology Industries Association) shows that electronic system production in Japan was US$80.5 billion (Figure 3). This US$80.5 billion represented only 7% of worldwide electronic system sales in 2010. In contrast, Japan accounted for 16% of the worldwide semiconductor market.

IC Insights believes that this mismatch is due to the fact that many large Japanese electronic system producers purchase their semiconductors from their headquarters in Japan (which are then logged as a semiconductor sale to Japan) but use a great amount of these semiconductors in electronic system production facilities located outside of Japan (which accounts for the relatively low amount of system production in Japan itself).

Not taking into account supply chain interruptions, the Japanese earthquake should not have much of an effect on the rest of the world’s demand for electronic systems. It should be noted that it is likely that the psychology of the Japanese people will be to act in a conservative manner in the current environment, which of course will have a negative effect on electronic system demand from Japanese consumers. However, rebuilding electronic infrastructure in the hard hit areas of Japan is likely to serve to partially offset some of the softness from the Japanese consumer portion of the electronic system market.



Effect on Semiconductor Sales

Assuming a loss $5.7 billion in worldwide electronic system sales this year, and assuming about 25% of electronic system sales are semiconductor content, about $1.4 billion in lost worldwide semiconductor sales value would result. Using the “pessimistic” scenario (i.e., 3.4% worldwide GDP and a loss of $5.7 billion in electronic system sales) and subtracting $1.4 billion from IC Insights’ current 2011 forecast of $346.8 billion for the worldwide semiconductor market, puts the worldwide semiconductor market at $345.4 billion, still a 10% increase over 2010.

As was shown in Figure 3, JEITA states that $43.4 billion worth of semiconductors were produced in Japan last year. That amount represented about 14% of the 2010 worldwide semiconductor market of $314.2 billion—24% of the worldwide O-S-D (Optoelectronics, Sensors, and Discretes) market and 12% of the worldwide IC market.

Thus, it appears that the world’s electronic supply chain is twice as dependent on Japan-based production for O-S-D devices than it is for ICs.

Although it appears that major DRAM and flash memory fabs were not significantly damaged by the earthquake and tsunami, flash memory and DRAM spot prices jumped by 20% in the days immediately following the quake.

As evidenced by this example, supply does not necessarily need to be actually disrupted for prices to surge, just the “potential” of a supply disruption can cause component buyers to step up their purchases for the “just in case” scenario. IC Insights expects that over the next six months, many electronic system producers will attempt to acquire extra IC inventory, especially in anticipation of the seasonally strong second half of the year.

IC Capacity in Japan

In 2010, Japan had the largest base of installed IC capacity of any region in the world. Figure 4 breaks down the world’s installed monthly wafer production capacity by geographic region or country as of mid-2010 (this data was excerpted from IC Insights’ Global Wafer Capacity Report).



To clarify, each regional number in Figure 4 is the total installed monthly capacity of fabs located in that region regardless of the headquarters locations for the companies that own the fabs. For example, the wafer capacity Korea-based Samsung has installed in the U.S. is counted in the Americas capacity total, not the Korea capacity total. The ROW “region” consists primarily of Singapore, Israel, Malaysia, but also includes countries such as Russia, Belarus, India, South Africa, and Australia.

As of mid-2010, Japan still led the world in IC wafer capacity with a 22% share. However, the gap between Japan and Taiwan has closed quickly over the past few years, with Taiwan forecast to surpass Japan as the leading IC capacity region in 2011. Korea is home to 15.2% of all capacity, followed by the Americas region with a 14.7% share, China with 8.4%, and Europe with 8.1%.

Analysis of the wafer size distribution by region at the mid-point of 2010 highlights some interesting differences in the seven separate regions (Figure 5). For wafers 200mm in diameter and smaller, Japan was the top region in terms of the amount of capacity (although Europe has the largest capacity of 100mm wafers specifically). The fabs running small size wafers tend to be older and typically process low-complexity, commodity type products or specialized devices.



At only 18.7%, Japan holds a relatively low share of the world’s 300mm IC fabrication capacity. It appears that as many Japanese companies have moved to the fab-lite model, the desire to install large amounts of 300mm capacity in the country has dissipated.

The supply chain in the semiconductor industry is a long and complex one. Moreover, in many cases, all that needs to happen to disrupt the chain is to pull out one piece of the puzzle (e.g., a specialty gas, molding compound, silicon wafers, a key module for a piece of semiconductor production equipment, or raw material) and the whole process comes to a standstill.

One area of concern is the production of raw 300mm silicon wafers from Shin-Etsu Handotai (SEH). It is estimated that about 20% of the world’s raw 300mm wafers are produced at SEH’s Shirakawa plant. This facility was still closed as of March 23, 2011 and SEH issued a statement that it would have a difficult time bringing this facility back into production with the rolling blackouts still in effect (a raw wafer production facility uses a tremendous amount of electricity to melt the polysilicon and pull the single-crystal boules that are eventually sliced into wafers).

Another area of concern is the resin bismalemide triazine, known simply at BT. BT is used to produce laminate substrates for PBGAs and CSPs, many of which are used in cellphones.

It appears that two of Mitsubishi Gas Chemical’s, which makes over 90% of the world’s BT, facilities that make BT were damaged in the earthquake. However, it is still unclear when Mitsubishi will be able to bring those facilities fully back on-line. In the meantime, there are already several other companies looking at producing and qualifying possible replacement materials for the BT resin.

IC Insights believes that current levels of inventory of wafers and packaging materials will help to avert serious shortages. Also, materials facilities (e.g., raw wafers, plastic resin, etc.) can oftentimes be brought back on-line much quicker than IC fabs. This was evident many years ago when Sumitomo’s plastic resin facility had a major fire.

Although there were some short-term disruptions, serious long-term supply chain problems did not materialize. IC Insights believes that other disruptions to the semiconductor supply chain are likely to emerge over the next few weeks that no one can anticipate today. However, any major disruptions in the semiconductor industry supply chain are currently not expected to last past August of this year.

Hsinchu, Taiwan Study Example

About three years ago, IC Insights was contracted to perform a proprietary market research report for a large insurance company. This company wanted to develop a model that showed how much in electronic system sales would be lost if the IC fabs in Hsinchu, Taiwan were shut down for one, two, or three months due to an earthquake or typhoon. We knew this company was serious when they showed us a plot of the location and severity of every earthquake and the path of every typhoon that had hit Taiwan over the last 50 years!

Currently, about 60% of the world’s IC fabrication capacity is located in seismically active regions (note: South Korea is not considered a seismically active region). Although the IC industry has always had a significant portion of its fabrication capacity in these “dangerous” areas, most buyers of ICs don’t give this a second thought (although they should).

Toshiba stated only a couple of months ago that a significant amount of IC wafers had to be scrapped when it lost power to one of its flash memory fabs. Since so many steps in the IC fabrication process are temperature sensitive, IC fabs have always had difficulty dealing with power outages that cause the temperatures in the process flow to quickly depart from what is required.

If the back-up generators are not able to instantly provide enough power, much of the “work in progress” needs to be scrapped (very expensive). It is possible that the rolling blackouts in Japan (now expected to last through at least the end of April) can be worked around considering the power company is announcing specific times to blackout certain geographic areas (possibly giving the fabs time to enact alternate sources of power).

It should also be noted that when a fab that has been shut down is once again ramped up, it takes 4-6 weeks from starting a raw wafer until a finished processed wafer arrives at the end of the line. Thus, IC devices do not immediately come rolling out of a fab after it is re-started, and a significant lag time must be taken into account.

We still remember being shocked at how far-reaching the impact of an extended shutdown of the IC fabs in Hsinchu, Taiwan would be. Since the two largest foundries (TSMC and UMC) have such a significant presence in Hsinchu, the number of different parts and customers (both IC suppliers and electronic system houses) that would be affected by such a shutdown is staggering.

The bottom line was that for every month the IC fabs in Hsinchu were shut down, the electronic system industry would lose at least $10 billion in sales!

Summary

In the final analysis, there is no doubt that supply will be constrained in numerous areas relating to the electronic system and semiconductor industries due to the earthquake and tsunami in Japan. However, on a worldwide basis, demand for electronic systems and semiconductors is expected to be only slightly lessened due to the disaster in Japan. Moreover, any lessening of system or semiconductor demand in 2011 due to the earthquake is forecast to be delayed and pushed into 2012, but not destroyed.

Figure 6 looks at the typical quarterly seasonal patterns for worldwide electronic system and semiconductor sales as well as the “post-earthquake” quarterly effects on Japan’s GDP and worldwide semiconductor sales. With regard to worldwide electronic system and semiconductor sales, the effects of the earthquake would have been much more severe if it had taken place in the third quarter when the strong seasonal upturn usually begins.



Because the earthquake happened so late in 1Q11, the effect on Japan’s 1Q GDP is not expected to be severe. As shown in Figure 6, Japan’s economy is expected to suffer the biggest negative impact in 2Q11. However, beginning in late 3Q11 or early 4Q11, it is forecast that Japan’s GDP will once again be growing as the rebuilding process, with its attendant spending and investment programs, begins in earnest. IC Insights believes that this rebuilding process will gain momentum in 1H12 and continue to impact Japan’s GDP in a positive way.

Similar to Japan’s GDP, IC Insights believes that worldwide semiconductor sales in 1Q11 will be only slightly impacted by the earthquake and that 2Q11 will be the quarter that is most affected. With a spurt of semiconductor inventory building likely to occur in 2Q11, semiconductor sales are expected to be only moderately negatively impacted instead of severely negatively impacted. Moreover, any “lost” sales in 2Q11 are forecast to eventually be shipped in 3Q11 or 4Q11.

TSMC recently stated that although it is likely to feel some impact from supply disruption because of the earthquake, it still believes that it will hit its full-year 2011 targeted growth rate of 20%.

Overall, IC Insights believes that the slightly negative 1Q11 and slightly positive 4Q11 earthquake effects on worldwide semiconductor sales will essentially cancel each other out. The same is true for the moderately negative 2Q11 and moderately positive 3Q11 effects. This “cancelling” effect assumption is the primary reason that IC Insights has not changed its full-year forecast of 10% semiconductor industry growth in 2011.

If key components of the semiconductor supply chain are constrained for an extended period of time, there is a potential shortfall to IC Insights’ current forecast for IC unit volume shipment growth of 9% in 2011. However, assuming that electronic system and IC demand remains strong, any significant constraint to IC supply will likely lead to increasing IC average selling prices (ASPs), especially during the seasonally strong second half of the year. With IC Insights’ current forecast calling for only a 1% increase in the overall IC ASP in 2011, there is considerable upside potential to this forecast.

Overall, IC Insights still believes that the IC market will grow by at least 10% in 2011, with any shortfall in unit volume shipments being offset by increasing IC ASPs.
One issue that could also boost the IC market and IC ASPs this year, since they are reported in U.S. dollars, is the weakness of the U.S. dollar. On Wednesday March 16, 2011, the U.S. dollar reached an all-time low against the Japanese yen (76.8 yen per dollar) but rebounded to 81.0 yen per dollar by March 23, 2011. Considering that the average yen per U.S. dollar exchange rate in 2010 was 87.9, there could easily be a 5-10% “boost” to the 2011 Japanese IC market caused purely by the strength of the yen when converted into U.S. dollars!

In terms of the supply chain, we should keep in mind that one company’s misfortune is another’s opportunity. Whether it is automobiles, electronic systems, silicon wafers, or flash memory, other suppliers worldwide will be aggressively attempting to make up for any shortfall in supply coming out of Japan. While it is not always easy to switch suppliers of key products, businesses eventually find a way to move forward.

IC INSIGHTS MESSAGE:

Our prayers and thoughts are with the Japanese people, for those lives that were lost as well as to the survivors of the earthquake and tsunami that hit northern Japan on March 11, 2011.

Related articles

IHS iSuppli thinks that high IC inventories in the supply chain limit impact of Japan disaster - 18 March 2011

Silicon producers in Japan suffering power problems - 17 March 2011

Ed Korczynski leaves SST magazine - 18 June 2008

TI‚??s supplier awards revisited - 08 April 2008

Renesas Electronics provides impact assessment update: 300mm fab remains closed - 23 March 2011

Reader comments

No comments yet!

Post your comment

Name:
Email:
Please enter the word you see in the image below: