The Monday before SEMICON West starts, there are two events that run in parallel in the Marriott Hotel. One is imec's US technology forum. Since I already attended the two-day version in Brussels, and wrote about it over the last couple of months, I attended the other event, the SEMI/Gartner Market Symposium. This is largely focused on the semiconductor equipment industry. But the health of the equipment industry is a sign of the future health of the semiconductor industry. And in Mike Santorini's memorable words, "The semiconductor industry is the speedboat towing the waterskier of EDA." If the boat is going fast then waterskiing is easy, but if it slows then EDA gradually sinks into the water. The symposium started with a presentation by Duncan Meldrum of Hilltop Economics titled Economic Uncertainty, the Global Economy and Semiconductors . He cautioned everyone that the data he was presenting did not yet take account of Brexit. Like almost everyone, he had assumed that the Brexit vote would come down as "stay" not "leave" and so he hadn't really spent any time on it. The overview of the world economy is: The effects of the financial crisis have been dragging on, keeping economic growth on a low-trend path. Uncertainty over economic policy has flared up with Brexit, suppressing key drivers for semiconductor (at least temporarily). The near-term demand-driven outlook for semiconductors has weakened. Global growth from 1993-2007 averaged 3.2%. Then from 2008 to 2010, we went through the financial crisis. Since then, however, from 2011 to 2016, it has only been averaging only 2.4%. Duncan though that there were several reasons for this low growth: There has been a major transfer of debt from the private sector to the public. Private sector balance sheets have gradually improved, although it has taken several years, but government balance sheets have dramatically deteriorated. Consumers and businesses are still risk averse about new spending/investment. There have been several recessions during this period in Japan, Europe, Russia, South America, and a slowing of growth in China. Forecasters have done a terrible job of predicting any of this, and their forecasts tend to start out predicting a return to 3+% growth a year out and then gradually having to reduce them as the reality becomes clearer as time passes. There is a lot of regulatory uncertainty and changing policies (TARP, QE, Dodd-Frank, TPP, Abenomics, Affordable Care Act, etc). Negative interest rates: "I'm so scared I'm willing to give you my money and get back less in 10 years". If you feel like that, there is no way you are going to be comfortable investing and perhaps taking a big loss. Ultimately the only thing that matters is consumer spending, since that drives everything. Companies buy semiconductor equipment only to make semiconductors that go into products that consumers buy. But real consumer spending trends are weak. In the developed world, where most semiconductors end up, growth is below the pre-crisis trends. Cells shaded in pink in the above table are below the 2000-07 trend. China and India are the only areas with positive trends. Europe is not in pink only because it was doing so badly in the earlier period, and Duncan reckons you can take off 0.1% to 0.3% due to Brexit. Semiconductor is a large and mature industry. SEMI tracks production in its MSI reports. MSI stands for "million square inches" and so is the total area of all the wafers manufactured. Duncan's company Hilltop Economics forecasts the future MSI for the next couple of years. However, there are wide error bands around the forecast as you can see in the above graph. The blue line are the actual numbers from SEMI, and the dotted blue line is Duncan's forecast, showing a decline in 2016 with a strong rebound in 2017. Putting it all together, Duncan assumes that economic growth stays on a low-trend equilibrium path. SEMI's MSI dips in 2016 and comes back in 2017 as in the above graph. But there is additional uncertainty from Brexit (which, remember, is not included in any of the graphs or tables here). This has more effect on investment than on consumers but it means that MSI will probably dip deeper in 2016 than in the above graph (shown by the red arrows) but that 2017 level is likely to be lower than forecast in the above graphs. Duncan's final numbers for real GDP growth and MSI are in the table below (still pre-Brexit). This forecast was done in June (and you can see the comparison to the forecast at the start of the year). Previous: Linley Mobile Conference...and That ARM Deal![]()
![]()
↧