AI, shutdowns and shadow banks

As a global leader in alternatives today, PGM is capturing the potential of tomorrow.

So as you look to diversify your portfolio, PGM offers expertise in seeding, developing and managing a broad range of liquid and illiquid strategies.

With over $320 billion in alts across public and private markets, we are helping clients achieve their long-term goals.

PGM, our investments shape tomorrow today.

Pushkin.

We spend quite a bit of time on this show banging on about stuff that investors are not especially worried about when they probably should be.

That arguably makes us miserable old farts, but whatever.

I'm leaning in.

So here's a few things you could potentially worry about if you want to be nice and miserable like us.

One of them is the total dysfunction in the US government and the shutdown that's now underway.

Americans, if you're wondering, yes, this does make you look very silly.

The other is artificial intelligence.

It's getting hella bubblicious out there and if you really want to worry about something kooky then you can always have a look at shadow banks.

This is Unhedged, the markets and finance podcast from the Financial Times and Pushkin.

I'm Katie Martin, a markets economist in slightly nippy London.

And I'm joined down the line from New York City by an essential worker who, unlike hundreds of thousands of federal US employees, is still coming to work.

And knowing him as I do, you can bet your ass he's still getting paid.

It's Robert Armstrong.

I think you just called me a grouchy old fart.

Was that the word you use collectively to describe the two of us?

That was a bit tough, Katie.

That's a tough way to start the show.

Speaking of tough, I have received several inquiries after our last podcast, which we recorded when I was out in New York, checking that I was still alive after we went for a few little drinks last week when I was out there.

And actually,

we were very fine.

He was quite civilized.

Right.

Yeah.

I think we can be sensible grown-ups when we try.

Not like the murder in Madrid, as I think of it.

Never speak of it again.

Yeah.

So let's...

What are we talking about first?

What's first thought about it?

First, we're talking about the shutdown.

So actually,

talking about it is a terrible idea because shutdowns can turn into open-ups very quickly and we can end up out of date.

But YOLO, let's talk about that.

I mean, the most important thing for us is, weren't we supposed to get a jobs report Friday that we're not going to get now?

I mean, that's a disaster.

So the jobs report for the uninitiated, the non-farms payrolls report is like the sort of benchmark U.S.

employment report.

Comes out the first Friday of every month, come hell or high water.

And this is like the number one data release that comes out of anywhere in the world because it's such an important input for U.S.

monetary policy.

So everyone watches payrolls.

And this Friday, oh, terrible scenes.

We're not going to get payrolls because the Bureau of Labor Statistics is going to be.

Yeah, it literally is the most important economic data point in the world.

And it has been shut down by the shutdown at a time when the labor market is the focus of all the attention of the market and the pundits because the labor market has been weakening dramatically at a time when the rest of the economy doesn't seem to be weakening dramatically.

So, we're sort of trying to figure out: is something weird going on in the jobs market involving immigration?

I mean, listeners to this podcast have heard us rattle on ad nauseum about this, but it's an especially important time to have new labor market data, and we're not going to get it.

So this is very bad.

It's extremely bad.

But the shutdown generally, I think like non-Americans kind of don't get it.

But like there's an annual process from what I understand where Congress has to agree on a budget.

And if they can't agree, then until they can agree, a whole bunch of functions just get shut down.

So like national parks are now unstaffed.

Yeah, and the Bureau of Labor Statistics, all of this stuff.

And this kind of game of chicken that the two political parties play almost always, like games of chicken between children on bicycles, almost always end with somebody turning away at the last second.

Yeah.

And it's a very good bet that that will happen this time.

But sometimes these things do drag on.

Yeah.

I can't, what was the last time?

I think it was about seven years ago.

That it dragged on for a month or so.

Yeah.

And it becomes a pain in in the ass.

But we are in a different kind of political environment than the one we were in the last time we had a serious attempt at playing this game.

And so who knows how things

might or might not escalate.

And Trump, we should also

put the asterisk here that

Trump has threatened to fire people permanently, not just furlough them.

It's not clear if he's allowed to do that, but it's not clear that he cares about what he's allowed to do.

But when does that ever stop to me before?

for?

And so that would raise the political stakes and the economic stakes, I would say.

Yeah.

But like for now, you know, markets are like, woo, woo, woo, all-time high, all-time high, USA.

And, you know, no one seems to really care about this.

Markets, by the way, Katie, don't seem to care about anything.

They don't care about the weak labor market either.

So it's not just the politics they're ignoring.

They seem, you know, very, very close to being totally indifferent to the economics too.

Yeah, they're nihilists dude.

But for all that I am happy to sort of play the misery guts role on this podcast and partially in real life too.

I do actually agree that although this is like a total like an unedifying site, it doesn't actually matter for markets just yet.

The only thing that would make it matter is the thing that really matters for markets is when US politicians clash heads about the debt ceiling.

That's a different kind of political process that you go through that can, in theory, lead to people not being made whole on their treasury holdings, on their holdings of U.S.

government bonds.

In theory,

even that is a slim probability, but this is what the current political environment is all about.

It's not that different from political environments of the past.

It's just that

the kind of extreme low probability events at either end of of the distribution of outcomes get slightly nastier and slightly more possible.

So like the absolutely horrible outcome percentage goes from 1 to 1.2%.

Yeah.

How do you price that?

I don't really know.

Maybe you just whistle past the graveyard and pretend the whole thing didn't happen.

Yeah.

That's what markets seem to be doing.

Well, in the meantime, in the absence of the thing that markets really care about, which is, frankly, jobs data that we we were just saying we would normally get on Friday what we do have to sort of plug the gap is there's another regular jobs data release that comes out of the private sector so it's not an official government piece of data but out the private sector from ADP now that came out on Wednesday it's like a payrolls processor yes they are the person who like runs the software that does all the company's payrolls.

So they have access to this stuff.

Yes, so they sort of know.

But they so they had their data out on Wednesday, said that private sector employment decreased by 32,000 last month.

And that's not like a huge miss as such, but economists had expected for it to go up by 50,000.

In an ideal world, no one would be relying on the ADP numbers to tell them what's going on in the jobs market, but we don't live in an ideal world.

So what this looks like again, as you were saying, the jobs market is a little bit more creaky than it was a few months ago.

This is an extra sign of creakiness, right?

Someone pointed out on their blog, and I'm sorry I don't remember this person's name, and I will email them when I figure it out and apologize for this, but they pointed out that ADP actually uses the Bureau of Labor Statistics weightings of the different sectors of the economy in order to come up with its number.

So if the shutdown lasts long enough, the ADP number will become unreliable too.

So that's just something for the rest of us to look forward to.

Alrighty,

move on to another thing that if you are a miserable sod, you could worry about if you wanted to, which is AI.

Now, AI is all sort of shiny and fantastic and everything,

but

there's a large number of people, a growing number of people I find, who are like not allergic to tech, but who are saying, listen, if you take this sort of AI miracle out of the US economy, how much economic growth is left?

And I think the answer might be not very much.

There's a lot of focus falling on this, right?

Yes.

But there is a lot of trickiness about the numbers.

So this was brought to my attention by a tweet by the famous Harvard economist and public intellectual Jason Fuhrman, who basically did the mathematics on the GDP report and said, look, if you take tech hardware and software out of GDP growth, that's like 90% of the GDP growth we saw in the first half of this year.

And that sounds really bad.

That sounds like you are the AI bubble and a bad cold away from being in a recession, basically.

And a lot of people have been sort of singing this dreary tune.

But as your friend and mine, Katie, Dario Perkins at T.S.

Lombard, which is a clever person at a private economics kind of shop, said

you have to back out the fact

that a lot of the equipment that goes into these data centers is imported,

which appears as a minus.

So the numbers are actually hard to read.

But I

so it's actually, you'd have to, you know, add the imports back and figure out the value of the intellectual property, which comes from the United States.

It'd be very complicated.

But I think it is not true that we'd be in a recession without AI, but there is no question that is a big chunk of American growth and we would miss it if it was gone.

Oh, yeah.

I mean,

we've spoken on this pod before about how big of a slice AI gobbles up in U.S.

stock markets.

And the answer is, you know, a lot is basically all you need to know.

But yeah, it is also a really large part of the real economy because there are just people building data centers all over the place because AI only works if you have massive data centers, which, by the way, churn out loads of of carbon, but that's a whole other question.

I got a stat for you, Katie.

I got a stat for you.

I don't mean to interrupt you, although I do it all the time, for which I apologize very much.

Just the capital expenditure budgets of the big AI monstrosities, the alphabets and the oracles and the so on,

is equivalent to 1% of gross domestic product in the United States.

That's just those seven companies or whatever it is.

That is like a lot.

They are just shoveling money out the door.

And, you know, and

to be fair, five years ago, we were all whining about how there was underinvestment in the U.S.

economy, right?

How there wasn't enough capital expenditures from big American companies, how all the money was going to buybacks.

how we were all eating the seed corn.

And now we have an economy where people are investing.

And we're like, not like that.

Not like that.

Because the thing is,

what we haven't, or like not us, but what the world has not yet really figured out is if you build up these data centers to produce the AI, is anyone going to pay for the AI at the end of this process?

That's a crucial question.

What is the return on the investment?

I don't think the worry...

is that the investment will suddenly stop and growth will cease.

What I think is if people realize

that the AI business model is not that profitable for whatever reason, and I don't know that it is or it's not going to be, I'm just speculating.

Should that happen,

then there's all these assets in the world that are going to have to be massively marked down, starting with the share prices of Meta and Amazon and Alphabet and so forth, right?

And going on to the stuff on the balance sheet of those companies, the value, the carrying value of the AI data centers themselves.

And so what this suggests is it's not like a growth recession.

It would be like a balance sheet recession, which, you know, like Japan had back in the day, where you're like prancing around with all this stuff on the national balance sheet at a very high value, and then suddenly you realize, actually,

we have to haircut all this stuff massively.

So one of the people

sounding the alarm this week is a chap called James Anderson, who was like a big cheese at Bailey Gifford, which is a fund based up in Edinburgh that has made some spectacularly successful bets over the years on

tech, great on Japan.

Great on tech, great on Japan.

So he knows his tech.

He has warned of what he calls a disconcerting rise in AI valuations, saying that Nvidia's planned $100 billion investment in open AI has brought uncomfortable echoes of the dot-com bubble, says Ir in the FT.

You know, even, like I say, even the tech fans are saying, look, this just feels a little bit, this feels a bit toppy, doesn't it?

The pass-the-parcel stuff is disconcerting.

Like, we're all the companies in this business and we'll all start doing business with one another.

I'll throw $100 billion to you.

You'll buy my chips.

I'll do this.

This investment will go there.

That does feel a bit

like Japan 1988 or America 1999.

The kind of circular nature of it all.

But you know, I guess it is easy in a boom like this to be skeptical and

in many booms like this you do ultimately see a pretty horrific separation of losers and winners.

All the money is rushing to one place.

Not all the money is going to get good returns.

But ultimately, these things are good for economies, right?

Real estate booms leave behind them housing that people can live in.

Telecom booms leave behind them fiber that people can send electronic messages over.

Do you know what I mean?

There are good effects of these kind of somewhat bubbly,

like the railroad bubble gave us railroad lines.

Yeah.

Anyway, so I'm just trying to be optimistic.

I'm, you know, you called me a depressed depressed old fart or something equally offensive at the top of the show, and I'm just trying to wear my optimist hat here.

Katie,

it doesn't fit very well, but I'm trying to jam it all.

That's because you've got quite a big head, as you've discussed on other

platforms.

One last thing, because you were writing this week about,

and it immediately sounds scary,

shadow banking.

Yes.

Tell me more about that.

Shadow in it.

Okay.

Shadow banking.

So this is interesting if you are as big a nerd as I am.

What a shadow bank is,

is a company that lends money but is not a bank.

And what not being a bank means, it doesn't take deposits, right?

So, which is what a bank does.

What a shadow bank does is it raises money in other ways

and then lends it out.

So you might have

a subprime auto lender like one Tricolor, which we talked about on the podcast the other day.

That is a Tricolor is a sort of shadow bank.

And what is interesting about them is that banks lend to shadow banks.

And

this raises questions like,

if a bank is lending to a shadow bank, Isn't the bank kind of by proxy doing what the shadow bank is doing?

So the shadow bank tricolor is doing these very dodgy consumer auto loans.

What about the bank that's lending to Tricolor?

I mean dodgy in the sense that like it looks like they're not going to get paid back, not in the sense that they're like

correct.

You know, naughty in some sense.

The long and short of it, and the reason I was writing about it this week, is that over the last year or so, the Fed, which among other things regulates how banks operate,

has forced banks to disclose more about their loans to other lenders.

So basically they're lending into the shadow banking system.

And according to the latest numbers, there are $1.7 trillion from U.S.

banks to non-bank financial companies or shadow banks.

Now,

I don't know how big a problem that is.

I do know $1.7 trillion counts as a very large number.

It's a lot.

This part of math class, I remember, that when you get to the trillions,

these are serious numbers.

Yeah.

Yeah.

And it's growing fast.

In fact, this kind of lending accounts for basically all the lending growth in America right now.

So banks are basically falling over themselves lending to private equity companies, private credit companies, subprime lenders, et cetera, et cetera.

The list goes on.

You know, it's kind of like what's going on underneath the hood here.

And it's not always easy to tell.

Aaron Powell, so all the lending growth comes from deep in the shadows.

All the economic growth comes from the AI build out.

And we have a total blackout on official data to tell us how that's going to affect the U.S.

economy.

Sounds great.

Rack me up for another record high in U.S.

stocks.

All makes absolutely perfect sense.

I mean,

I'm increasingly convinced by the view that there's going to be like a big rally towards the end of this year in stocks because, you know, heaven knows the world has thrown enough crap at markets this year.

And if they can get through that, they can get through anything.

What this tells you is that in the short term, when the market has momentum, almost nothing else matters.

Yeah.

That's true today.

And that has always been true since

dinosaurs were trading stones with one another.

With their little arms, Pressing their keyboard.

Little arms.

Go with the flow.

There's a good case for it.

Speaking of going with the flow, let's come back in a little second with Long Short.

Bonds are back.

And so is All the Credit.

P Jim Fixed Incomes Monthly Podcast Series.

From the latest trends to long-term perspectives, you'll get timely fixed income insights from leading economists, research analysts, and investment professionals.

Whether you're new to bonds or a seasoned investor, tune in to all the credit wherever you get your podcasts.

This podcast is intended solely for professional investor use.

Past performance is not a guarantee of future results.

Alrighty, it's time for long, short, that part of the show where we go long, a thing we love, or short, a thing we hate.

Rob, what are you long or short?

I am short

cyber attacks.

And not just for a generic technological reasons, but because of the following lines from the FT written by our colleagues Harry Dempsey and Leo Lewis in Tokyo.

Japan is just a few days away from running out of a Sahi super dry.

as the producer of the nation's most popular beer wrestles with a devastating cyber attack that has shut down its domestic breweries.

Now, I'd like

to say why this touches me so near to my heart is now I'm going to tell you about my secret vice, Katie.

Most days on my way home from work, I purchase a beer and drink it in a paper bag on the subway on the way home.

Most days, not on the day.

Most days on the day.

Just on a Friday.

A subway beer is a totally standard thing for me.

And very often, it is an ASAHI Super Dry that I have in my paper bag on the subway.

So is there a danger that you will run out of a sahi Super Dry?

I mean,

it would be catastrophic.

Or is it?

No, it's just a domestic glory.

So, hopefully, in America, we still have it.

What are you long at short, Katie?

Enough on my bad habits.

What about yours, Katie?

What are you long at short?

Well, I'm short COVID because everyone's getting it again.

A friend of mine who's got it for the 11th time, she's the only person I know who's had it more than me, I think about seven or eight.

It's going around, and this right here is prior warning that i will get it because i always bloody get it and we'll have a sub we're gonna have to get a sub for you for next week you may as well start casting the net now and find some people who are free in the next couple of weeks because i will get it it will wipe me out it will be very unpleasant and i will complain about it a lot but this is just my lot in life unfortunately so listeners unless i have covid again

I will be back on Tuesday.

Listen up then.

Unhedged is produced by Jake Harper and edited by Brian Erstat.

Our executive producer is Jacob Goldstein.

Topha Vorges is the FT's acting co-head of audio.

Special thanks to Laura Clark, Alistair Mackey, Greta Cohn, and Natalie Sadler.

FT Premium subscribers can get the Unhedged newsletter for free.

A 30-day free trial is available to everyone else.

Just go to ft.com/slash unhedged offer.

I'm Katie Martin.

Thanks for listening.