Jason retweeted
TIPS are back in the news and that inevitably means confusion about how investors should view them.
Here's a fun thing to think about: A boat tied to a floating dock rises and falls with every tide. Measure it against the seabed and it never stops moving. Measure it against the dock from where you’ll step aboard, and it barely moves at all.
Long-term TIPS are like that boat: volatile in today’s dollars, but remarkably steady measured against the real spending they’ll support. Seen this way, most of the confusion about TIPS disappears. They do exactly what they were designed to do. And as the safest asset going, their real yield is the natural benchmark against which to judge the stock market’s expected real return when you decide how much to invest in stocks.
Jason retweeted
Seoul this week.
XRPFi Night on Oct 2. XRP Seoul on Oct 3.
The rooms that matter for XRP are the rooms we continue showing up in.
See you there → flare.network/events
Jason retweeted
This may be a useful gauge for market reaction over the weekend. The Saudi Stock market (Tadawul All Shares Index) today underperformed its biggest constituent (Saudi Aramco) by the most in over 5 years.
Major risk off move + spike in oil prices
Jason retweeted
At the same time, no one wants to sell if Trump is ultimately going to TACO, and there is real reluctance to go underweight the US b/c that means going under weight the US tech bros (I have lost faith in their underlying competence tho, thanks to Elon's mismanagement here)
Jason retweeted
I just published a new "Next Economy" post
I hope you enjoy the read!
nexteconomy.co/p/an-outlook-…
Jason retweeted
IT IS TIME!!
PLACE YOUR VOTE FOR FINANCE CHARLATAN OF THE YEAR HERE -->
#FCOTY
forms.gle/KJ4Hm1Gewbd3oiEq6
I really enjoyed reading this when I first started out trading funding products, highly recommend.
The Repo Market, Shorts, Shortages, And Squeezes continues to do well. This week it's #15 on Amazon in Financial Risk Management
Jason retweeted
There is an ever growing chorus of voices insisting that the United States has fallen so far behind in manufacturing that catching up is impossible and that efforts in onshoring/nearshoring will end in catastrophe.
These American Fatalists rest their arguments on the assumption that the future will look exactly like the past. They compare today’s industrial landscape to the old one, see an erosion of factories and supply chains, and conclude the game is over.
But this view assumes that manufacturing in the decades ahead will rely on the same methods, the same cost structures, and the same production logic that defined the twentieth century.
History has never worked that way.
Every time a major shift in technology or process has arrived, the competitive map has been redrawn. New materials, new software driven workflows, new automation, and new forms of additive, modular, and distributed production are already changing what it means to make things.
The question is not whether the United States can rebuild the manufacturing system of the past. The question is whether it can lead the manufacturing system of the future.
And that is where the equation changes.
No country on earth has a deeper culture of innovation than the United States. From advanced robotics to AI integrated design pipelines to novel fabrication techniques, the frontier of industrial reinvention is taking shape here more than anywhere else.
The belief that America cannot possibly catch up, let alone lead, assumes a world that is static.
I don't know for sure how this plays out. China is undoubtedly a formidable adversary. And to assume otherwise would be imprudent. But that goes the same way for those assuming the US can no longer compete.
Place your bets...
Looks like the committee has reached their conclusion and the vote is settled.
wsj.com/economy/central-bank…
Jason retweeted
1/6
Emmanuel Macron: "Today, we are caught between the US and China and it is a matter of life or death for the European industry. We have become the adjustment market and this is the worst-case scenario."
He is absolutely right.
reuters.com/world/china/fran…
Jason retweeted
I spent some time this week researching how to teach LLMs to think Macro. I leave you with 2 thoughts from what I learned this week.
*** Tactical / implementation observation
One can take a RAG approach to inform/educate an LLM about domains specific content (in this case Macro).
Or one can take a fine-tuning approach to generate structural understanding of Macro - which becomes a natural platform to more complex Macro analysis via Agentic workflow.
The RAG approach is probably mainstream across the HF industry - i.e. offers no competitive edge. Furthermore, most RAG algorithms are mediocre for Macro thinking because they optimize for compression while sacrificing the text's surface variation, which is the opposite you want to do when analyzing Macro.
To put it in terms Macro people understand. To make money on a Macro theme, "reading the room" is as valuable as the fundamental understanding of the theme.
*** Theoretical observation
To teach a machine how to think Macro, you need to go into the rabbit hole of semantic meaning. The challenge I identified from current "meaning extraction" architectures, is that they operate at a single level of abstraction - typically the surface level. They extract "What the text says" rather than "What the text means" or "Why it matters".
I approached this challenge with a hierarchical abstraction. Here is a mental map that can help to think about this.
Jason retweeted
It’s fine as long as you are running many signals in the same portfolio.
Your portfolio’s ability to handle noise increases as the number of signals increases.
Think about it.
Imagine you have a game where you need people to guess how tall you are, and you take an average of everyones guesses.
If the person is wrong by more than 5cm, you chop off your finger.
If you only have one guy, he better be on point.
If you have 1000 guys, you can have some guys that are still atrociously wrong, as long as your audience are generally smart, you will get a fairly robust good answer that is very robust to bad guesses.
Then we stop talking about overfitting a single signal and start talking about reducing family wise error rates.
—
Jason retweeted
Passed Pawn's Perspective
1. The markets got through yet another scare in November. Everyone was too all-in on a Santa Rally so they had to be shaken out. I still don't see a reason for a sustained selloff in risk assets, but also don't feel the need to press here. I'm still running lower risk than usual.
2. Commodities are hot again. Copper looks to be breaking out if you ignore the tariff nonsense over the summer. The world is capacity constrained wrt AI buildouts and we need a lot more energy and metal. Focus on the big picture.
3. Eventually we will overbuild (short term capacity) but I don't think that day has come yet. Everyone wants to point to the dotcom bubble. Ok, but we ultimately needed way more internet and compute capacity. The question is when short term expectations need a reality check, and that timing is very hard to predict.
4. We're waiting for the Fed chair announcement but it looks like Hassett. Interestingly STIR markets haven't moved much lately, so the idea of a Fed puppet taking rates to zero has been subdued. I suspect it may have to do with affordability becoming a much more prominent electoral issue recently.
5. Bitcoin had a proper washout since I last wrote. But that's nothing if you track the altcoin markets. It's a graveyard this year. As I've written, the entire space needed a cleanse. Too much easy money, launching coins at absurd market caps leaving horrific risk/reward for new money. For next year to be better, the grifting has to stop and expectations need to come back to Earth.
6. The labor market continues to look weak, but the question remains if it's more of a labor supply issue. AI capex is dominating although that's not a reason to be bearish overall because it's not going to stop soon. K-shaped economy persists and I don't see an off-ramp.
7. After the recent election results, the republicans must be feeling pretty concerned about the midterms. Polling is awful, affordability is poor and consumer confidence is rock bottom. The only thing going well economically is the stock market.
8. Based on that, one could argue they might stop focusing on the markets and try to improve the other stuff. I actually expect the opposite and think they will support the markets at all costs next year, so it seems to me there is a policy put somewhere not too far below.
Have a great week.
-PP
Jason retweeted
The pressure on household expenses is evident in the flow of funds. Each dot represents the sum of household savings vs total debt expense each month (labeled quarterly), colored by year.
-Dots moving right and left indicate more/less savings.
-Dots moving up and down indicate more/less debt costs.
In 🔴2022 we see savings flows plummet (dots move left) then debt costs rising later in the year (dots rise), representing the early 2022 inflation shock paid for by dissaving, and debt service increasing as the Fed begins to hike.
Note how 2023-2025 (🔵🟢🟡) have seen dots rising year over year, but staying relatively flat on the X-axis. This represents rising all-in cost of debt servicing for households *without* much change in the flow to savings.
What this means is that the marginal new dollar earned by increased wages is being eaten up by a split between consumption+borrowing, and NOT becoming an increased flow into savings.
I wager this dynamic could pose an issue if the historic low job growth eventually cracks into outright labor contraction (as suggested by direction of travel in private data since the blackout).
