Why Macro is “Pretty Risk-On” for Equities | Tian Yang of Variant Perception

**Jack Farley** (0:00)
I'm joined once again by Tian Yang, co-founder and head of research at Variant Perception. Tian, what do you think about the stock market right now? The S&P, the NASDAQ, the semiconductor stocks, they have been trading quite weekly for the past month.
Do you think that this is a bottom?

**Tian Yang** (0:19)
So I think time horizon matters.
So on our macro risk indicators, that generally looking forward more three to six months, we're still in a risk on macro regime, where it sees the mix of growth, inflation, policy liquidity is broadly still supportive risk assets. So in general, I think the environment is still good.
Now, clearly with the kind of hype around AI, the kind of mania in Korea, there's obviously signs of access linked specifically to AI. And that's clearly why we've obviously had that bubble like price action. You had a bunch of LPPL bubble exhaustions go off in early June. We've had the correction. But from here, I would say that this is kind of like what the gold and Bitcoin tops felt like. If you go back to like some of 2025, it's kind of like the fundamental story hasn't really changed, right? Like, you know, Trump's going to be a great crypto president. You know, they're going to pass a genius act or that. But, you know, despite that, the price action just exhausts similarly for gold at the beginning of the year. And I think you're a little bit at that stage with Semi's where there's enough, like, underlying things on bottlenecks and, you know, the future, that's all still intact. The fundamentals haven't changed, but the price action is kind of telling you, you know, it's become like a meme, too much money's got in. And typically when that tops and rolls over, it's quite hard for it to regain the highs. Obviously, we're well, well below the highs. So certainly we're recording this just as we have the situational awareness news. So certainly, yeah, sure, could you get like a very sharp squeeze, a tradable bounce? Absolutely. But I think the medium term picture is that we need like a new narrative, right? The narrative around Agenic AI, the narrative around bottlenecks. That's what fueled this leg of the rally from April on. That narrative's clearly done. And we're waiting around for the next narrative. Maybe it's going to be like, you know, real world, world modeling, things like that. You need something else ready to come along with like a wild moment, I think, for us to really get back to the highs. So, yeah, I think we're more in the kind of phase of the market where the equity market is broadening out. The money's not necessarily leaving the market, it's just rotating to laggards and other areas, which generally I think reflects the risk on. So, yeah, I think broadly, equity outlook is fine, but on the semi-specifically, I think we, in terms of out position, we're keeping some of the exposure, but we're really been rotating to more of the value laggards that's been popping up the portfolio in July. So only energy, only financials, only health care, those have been pretty important as more portfolio balanced.

**Jack Farley** (2:52)
Tell me about the LPPL exhaustion signals. Remind me what that stands for.
What were the signals you got a month, two months ago about there's too much risk, too much leverage in Korea, exhaustion? And also do the signals work the other way? Are you getting bullish LPPL signals, or do they only work on the downside?

**Tian Yang** (3:12)
But they should work both ways. So LPPL stands for Log Periodic Power Law. So it's our bubble and crash detection system. So it's kind of designed to kind of get you that final OF leg of the move that typically is not obvious, right? It's more like after something has been moving exponentially for a while, it looks like it's stable. But then because inherently of like, just the way markets were forced, forced price action, there's always that final leg. Like recently we had it on China, for example, right? China sold off, sold off, sold off, just kept going lower. In May, April, it looked like it was bottoming, and you have a final OF leg in June. And it was in June that all the LPPL buy signals went off. Similarly on the Semi, when it was rallying up in April, in May, mid-May, it kind of was like, wow, this is a bit much, but then it has that final leg up, and then it triggers the model. So it's kind of designed to catch these log periodic power law waves. So you can think of these as just essentially a pattern matching model. But it's really looking for not just things to be exponential, but that the kind of wave pattern, the market speeds up as it's going exponential. And that's typically what you see at the end of these exhaustion moves. So you kind of saw a bunch of these from the beginning of June as we were kind of making the making the new highs. And so typically these are more tactical models. You know, they typically are good for the next month. So they're not necessarily long-term models, but they give you a sense of when things are extreme to face, right? So the most recent has been obviously all the semi-AI related cells, and then the China, you know, Indonesia, a lot of these laggard, you know, biasing we've been writing about.