consumer sentiment vs. AMAT stock price:
Thanks for that link, lots of good data there, I've bookmarked it. economagic.com
The last recession we had was in 1990. Consumer sentiment (CS) troughed in 10/90, and AMAT troughed in 11/90. Nice correlation there. Of course, you wouldn't have known the sentiment had troughed until the next one came out higher, so this is not a leading indicator.
Between the 10/90 trough in CS, and the 10/98 trough in CS, there is no correlation between CS and AMAT. If anything, it looks like an inverse relationship (CS troughs coinciding with AMAT peaks, and vice versa).
The 10/98 trough in CS (at 97.4) coincided nicely with the 10/98 trough in AMAT. Again, though, the AMAT price is known to investors in real-time, while a trough in CS is only clear several months later (once the data is out, and 1 or 2 months of rising CS shows that the bottom is in).
The 1/00 peak in CS (at 112, a 50-year peak!) led the 4/00 AMAT peak. This actually looks like it might have been a useful leading indicator. I remember, all through summer and fall 1990, everyone (including me) was saying over and over, "why is AMAT going down?". Maybe if we'd been watching the steadily eroding consumer sentiment, we would have not been so confused.
From 11/98 to 11/00, CS was in a high plateau, between 102 and 112. All the CS numbers during that time are far above the usual historical range of the last 50 years. Beginning in 11/00, CS has been in a steep decline. From 11/00 on, the month-on-month fall in the CS is almost without precedent, in the last 50 years. The latest reading is in 2/01, at 90.7.
Conclusions:
1. Sorry, folks, this ain't the Holy Grail we've been looking for. The last data point (1/00 peak in CS) looks like it is a leading indicator for AMAT peak price. But one data point does not make a trend. By the 2019 cyclical trough, I may have enough data points to be certain of the pattern (very small grin).
2. It's not entirely useless. The CS trough coincided with AMAT trough, more or less, at the last recession in 1990, and at the last semi-equip trough in late 1998. So, it is one more data point to consider. That is, when CS has been climbing for several months off a trough, it is more likely that we've seen the trough semi-equip prices.
3. the CS sentiment numbers and rate of change, since the 10/98 trough, have been outside the 50-year pattern. The last two years have been different than anything in the last 50 years.
4. previous troughs in CS were: 57.6, 51.7, 63.9 (in 1975, 1980, 1990). The present level of CS (90.7) is still far above those levels. So, if we are looking at a recession in 2001-2002, CS probably has a lot further to fall. As a guess, the CS will fall to 58 (the mid-point of the 3 data points above). Again, this guess is based on very few data points, and assumes a recession this year.
5. If CS falls to 58 in 2001-2002, then consumers are going to cut back on spending a lot. End-consumer demand for most things with chips in them will also fall a lot, from current levels. Semi-equip bookings will also continue falling, as will semi-equip stock prices (in the usual stair-step fashion).
6. Whether or not we get a recession soon, or CS goes to 58, it is probably a good idea to not take LT long positions in semi-equips, as long as CS is in free-fall. |
