Semiconductor Industry Cycle Analysis

I confess that I’ve long been fascinated by the semiconductor industry cycle. It seems to be so early cyclical that it is almost counter cyclical! This year has proved rather frustrating for the industry. Many forecasters were predicting an upturn that hasn’t quite happened yet. I want to take a look and see what the chances are of a recovery. I’ll do this by focusing on industry data, metrics from a key bellwether, and anecdotal evidence from companies exposed to the early stages of the semiconductor cycle.

It should be noted that whilst the semiconductor industry has its own dynamics it is still closely tied to the global economy so the sector is only really interesting if you believe that growth will do ok going forward. That said it’s time to look at some industry specific dynamics.


Industry Data

The Semiconductor Industry Association (SIA) has generally been quite positive on the outlook for the industry. While pointing out near term weakness the prognosis given in its press release in May was as follows

“We look for seasonal moderate growth to continue in the second quarter and build momentum as 2012 progresses. However, while forecasts for global economic growth are improving, macroeconomic and geopolitical uncertainties remain.”

However the latest worldwide sales numbers in June reported a .1% sequential decline and a year on year fall of 2%. In order to put this in context here is a long term chart of the SIA three month average billings. Even with the smoothing effect of the average, we can see how cyclical this industry is.

I’ve circled ‘troughs after recoveries’ to represent what bulls are hoping is the situation now. In other words, we have had the usual sharp move upwards that occurs post recession and then we get a mini trough as growth natural slows before picking up again. I think the evidence here is rather compelling and suggests that as long as global growth does ok, the semiconductor industry is set for better times ahead.


Intel Signalling Better Days?

Only Samsung and Taiwan Semiconductor rival Intel $INTC as the industry bellwether and here is how gross margins and quarterly inventory/sales (I/S) ratios have developed since 2008.

Note that the sharp rise in I/S in Q4 of 2008 came as a bit of a surprise but we need to recall that the recession was mainly about the implosion of the banking sector. It was not caused by industry general industrial over-investment or a consumer led boom. Therefore it represents somewhat of an anomaly, however, I think current conditions are more normalized.

The recent data suggest that gross margins are on a downward trend (bad) and I/S is on an upward trend (bad). Whilst this is nothing like 2008-09 it does suggest that growth is weakening. We may well see a ‘snapback’ in the Q3 results due in October. If so, Intel will be off to the races. Well worth following.


Semiconductor Industry Cycle Turning

I think it’s a good idea to look at what the capital machinery manufacturers are saying. If the manufacturers and foundries are feeling confident or they are running at close to capacity (indicating expansion) it will be seen directly in the top and bottom line of these companies. As ever, discussion starts with Applied Materials $AMAT whose recent results and outlook where not particularly good. Management talked of macroeconomic uncertainties causing customers to reduce spending until things pick up. Overall net sales were forecast to decline 25-40% in Q4 and margins continued their decline.

It’s not just AMAT. Its rival KLA-Tencor $KLAC reported gross bookings slightly down in the previous quarter and forecast that new orders for the upcoming quarter would be notably lower. KLA is somewhat more exposed to mobile demand and there has been some weakness there. In a sense KLA is a difficult company to analyze within an industry remit because its revenues are extremely lumpy and management admits that their forecasting record is ‘not very good’.
Despite the opacity in its results caused by its purchase of Novellus, Lam Research $LRCX guided towards weak results in the upcoming quarter and it also announced that it expected wafer fabrication equipment spending to be at the low end of its previous guidance.

Indeed even the positive news of investment by Taiwan Semiconductor and Intel into ASML Holding $ASML is more about an adjustment to the necessity to try and reduce future chip costs as they become ever more complex. In other words, the outperformance of the stock is due to some company specific issues rather than indication of industry strength.


Putting It All Together

I think the anecdotal evidence is supporting what Intel’s key metrics are telling us. However it is always the darkest hour before the dawn. Current conditions are weak and investors should look out for more short term disappointments. However, I think much of this is due to an industry malaise caused by macro-economic weakness and specific issues with solar power and DRAM chip weakness from sluggish pc sales.

The stocks in the sector have largely priced this in and if you buy the economic growth story than the sector is well worth a look down here. The billings data is tentatively suggesting we have passed an inflexion point so if investors can tolerate near term weakness the longer term picture might look better. Expect worse to come near term though.