Citi's Dirk Willer on Why You Buy the Bubble | #654

Dirk Willer · Citi

The bond selloff reflects higher real rates from strong growth, not Fed credibility or fiscal worries.

Decomposing the back-end move into real rates, break-evens and asset swap spreads shows real rates did all the work. Inflation expectations stayed well behaved, so the market did not say the Fed is losing credibility. Asset swap spreads, which usually show fiscal stress, did not move, so the market did not price a big fiscal problem. The market is saying growth is strong, so real rates and r-star must rise. That is a good problem to have and not that bad, which may explain why equities can rise alongside higher rates.

“And it was all about real rates of the market saying, listen, you have a growth problem, you know, which is a good problem to have”at 15:11

Dirk Willer · Citi

Treasury can cut coupon issuance and sell more T-bills at the November refunding to ease yields.

The Treasury has firepower. Its buybacks have been a 'pea shooter' so far, and it is not close to running out of bullets. At the November refunding it could issue far fewer coupons than expected and sell mostly T-bills. If bill issuance creates front-end stress, the Fed would have to buy them, and it can buy bills if the problem is in repo. Also, when the debt ceiling debate draws down the Treasury General Account, bank reserves rise, so bills can settle without problems, as in 2023. The earlier mistake was intervening in the back end without a panic.

“So he could really issue a lot fewer coupons than expected, and really just sell a lot of t bills.”at 15:59

Dirk Willer · Citi

Treasuries will protect portfolios in the next deflationary bust, but not while inflation is the problem.

Right now the problem is inflationary, so Treasuries do not help. The market moved from assuming every shock is deflationary to assuming every shock is inflationary, which is wrong. In a deflationary bust, the Fed would cut and buy the back end, so bonds would work. Candidate deflationary shocks include a bursting AI bubble or AI working too well and pushing unemployment above 10%. The 30-year real yield of about 3.3% is attractive for long-term investors who expect an equity setback. Buy bonds after the Fed has hiked once, and hold for a long horizon.

“Right now we are in an inflationary problem, and of course they don't help you... when you get the next deflationary shock, bonds will be your friend again.”at 19:50

Dirk Willer · Citi

TIPS offer a higher Sharpe ratio than nominal bonds, so lever them for better returns.

TIPS generally offer a higher Sharpe ratio than nominal bonds. Because they are less volatile, you can lever them up a little and earn a higher return than nominal Treasuries. The argument rests on the lower volatility of TIPS and their risk-adjusted return. It is offered as a portfolio construction point rather than a market call.

“tips are in general a higher Sharpe ratio proposition than nominal bonds.”at 36:18

Dirk Willer · Citi

TIPS only work in stagflation if the Fed doesn't hike; hikes hurt them

TIPS have a higher Sharpe ratio than nominal bonds and are less volatile. In a stagflation scenario, TIPS only work if the Fed does not act. If inflation rises but the Fed does not hike, TIPS are great. If inflation rises and the Fed hikes, TIPS are not great, because rising front-end rates sell off all bonds. Very short-dated TIPS may still work, since the inflation coupon more than offsets the rate rise, but that is hard for retail investors to trade.

“if you're in a stagflation scenario, they only work if the Fed doesn't act. So if there's inflation but the Fed doesn't hike, then tips are great. If there's inflation but the Fed hikes, then tips are not great”at 37:12

Dirk Willer · Citi

Expects a good year-end equity rally; a pullback into the midterms should recover after the vote.

November and December are seasonally strong, and they are even stronger in good years, because investors pre-position for a strong next year. Tax-loss selling is also lighter. Usually there is a pullback into the midterms, averaging around 5%, which is quickly recovered after the vote. The main risk to the year-end rally is an Iran war re-escalation after the midterms.

“But in principle we think year end will be good in equities.”at 8:05

Dirk Willer · Citi

S&P can keep ignoring rising rates unless Treasury volatility (MOVE index) rises too much.

His team found the level of rates barely matters for equities. What matters is the volatility of the sell-off, measured by the MOVE index (implied volatility of Treasuries). Equities start to care when MOVE rises too much. Recently MOVE has started to move, yet the S&P was roughly sideways and only slightly down, so price action has been impressive. He thinks the S&P can probably keep ignoring interest rates.

“We found the level of rates almost doesn't matter for equities. What matters is the volatility of the sell off.”at 8:16

Dirk Willer · Citi

Tech leadership persists until the bubble peaks; the big sector rotation only follows the peak.

During bubbles, concentration is high and tech keeps outperforming. Broadening episodes are short-lived. Tech outperformance usually ends when the bubble peaks, not before, and the large sector rotation comes after. In 2000 the Nasdaq peaked in March but the total-return S&P peaked in September, because money rotated into rate-sensitive sectors like utilities and housing, defensives like consumer staples, and even finance. If tech stops working, the whole story falls apart. If you don't like tech, it's unclear what you should like other than bonds and consumer staples.

“So the outperformance of tech ends when the bubble peaks, not before, typically.”at 24:22

Dirk Willer · Citi

Overoptimism indicator flags short-term 5% pullbacks and has worked well out of sample

Citi's POLS indicator is a short-term risk gauge that combines positioning, overoptimism and leverage measures. It is hard to find an indicator that has worked as reliably as this one. It is tactical and calls for about 5% pullbacks on average, and it has worked extremely well out of sample, which Dirk says he likes.

“It's a tactical one. It just calls for 5% pullbacks on average. But it has been working extremely well out of sample”at 41:43

Dirk Willer · Citi

Bubble peak likely tied to compute overcapacity and liquidity withdrawal; the Fed is not there yet.

Bubbles tend to pop when liquidity is withdrawn, as in 2000 when the Fed hiked and then drained liquidity added for Y2K. His team's approach is to date compute overcapacity, since stocks might peak six to nine months before fundamentals do. Human behaviour tends to overcapitalise large opportunities, so overcapacity in compute is likely eventually even if fundamental analysts don't see it close yet. The first Fed hike is usually not a problem, and the last one is. It is unlikely the Fed is already far enough along its hiking path to really hit the bubble.

“I would say human behaviour is to overcapitalize the opportunity, and the bigger the opportunity, the more you overcapitalize it. And so in the end, there will be overcapacity of compute.”at 4:38

Dirk Willer · Citi

US equities entered bubble territory mid-last year; bursting could bring a bigger recession than 2000.

Using GMO's price-based definition (an asset rising more than two standard deviations above its long-term real trend), US equities entered bubble territory around mid-last year. Valuation measures such as Shiller's CAPE also point to a bubble. Bubbles are not a reason to be bearish: investors make money riding them but give back a fair chunk later. Equity market cap is about 240% of GDP versus just over 100% in 2000, when the bursting bubble caused a recession. So the next bust could be bigger, the Fed would likely cut aggressively, and rates could return to zero. Valuation tells you nothing about the next year, but he expects a meaningful equity setback over roughly the next five years.

“partly because I believe that the equity market at some stage over the next five years or so will have a very meaningful setback.”at 18:32

Dirk Willer · Citi

Breakdown of the top seven S&P stocks signals the bubble's end; an earlier exit than fundamentals

Bubbles get highly concentrated in market leaders, so when the leaders break down it is a bad sign for the bubble's longevity. Citi's bubble indicator takes the top seven S&P 500 names each January 1. It switches on and signals an exit when four of those seven fall below their 200-day moving average. Dirk says this saves a lot of money in bubbles and gets him out earlier than turning fundamentals, which typically come six to nine months later. He hopes to use it over the next year or two.

“whenever you, our bubble indicator is switched on, you get out of the market when four of those seven break down, that saves you a lot of money in these bubbles”at 40:48

Dirk Willer · Citi

Warsh Fed is hawkish; his shadow-rate indicator sits at highs, echoing Greenspan-era reaction.

Warsh gave no formal reaction function, but he listed the indicators he cares about. His team built a z-score of those indicators and summed them into a 'Warsh shadow rate', which tracked Fed funds almost perfectly under Greenspan in the 1990s and then acted as a leading indicator. It is now at highs, which they read as hawkish. The last meeting was unanimous, suggesting he has enough sway on the committee. The bias is hawkish but not irrationally so: if the data says no hike, the Fed won't hike.

“right now it's at the highs. We think that means he's hawkish.”at 11:44

Dirk Willer · Citi

Fed is not close to overtightening; buy bonds only once one hike is left.

Stocks usually have a small pullback around the first hike, then stop falling until the last hike, when the Fed overtightens. The Fed is not close to overtightening. Bonds should be sold about six months before the first hike and bought back when one hike is left. The market keeps pricing more hikes until the Fed sounds close to done, so buying early is premature. If the cycle looks like 1999, it is too early to buy bonds. Only if the Fed just takes back three cuts does the front end become attractive.

“And so we would argue they are not close to overtightening.”at 12:43

Dirk Willer · Citi

The Fed's hiking cycle will be over next year, so bonds can be added back

Dirk expects the Fed's hiking cycle to be behind us next year, almost certainly. That would mean bonds can be added back to portfolios, perhaps quite aggressively depending on how things play out. The view depends on oil falling as the Iran situation resolves, since lower oil would help central banks that target oil or headline inflation end their hiking cycles.

“Then next year, almost certainly the Fed hike and cycle will be behind us, which means you can add bonds again to your portfolio, maybe quite aggressively depending how it goes.”at 43:24

Dirk Willer · Citi

Brazil is a favourite: a market-friendly government, big rate-cut room, and real currency upside.

Brazil is an idiosyncratic play in a market where most trades reduce to oil and AI. The new government will likely be market-friendly, restoring fiscal order and deregulating. Real rates of 9-10% leave room for aggressive cuts, because the economy was never weak enough to justify them while fiscal policy was strong. Now that fiscal is being repaired, cuts can come. This is a commodity story that benefits from lower rates and business-friendly policy. The real should appreciate as investors bring money back, and the carry trade will last a long time.

“So now the fiscal gets paid back, you can get very aggressive rate cuts. It's a commodity story that benefits from lower rates and the new business friendly government.”at 26:43

Dirk Willer · Citi

Global forces drive about 70% of emerging market outcomes, more than local stories

Emerging markets are the tail that gets buffeted by global markets, so global forces are probably about 70% of the game for them. Dirk illustrates this with Brazil in 2002: the real rallied strongly the day Lula came in, but it had probably been driven by the S&P, which was in free fall before the election and turned around at the same time as the real. In his view, when the global cycle turns, investors should follow the global cycle rather than the idiosyncratic story.

“emerging markets are the tail that are getting racked by global markets, so the global forces are probably 70% of the game for these emerging markets.”at 46:42

Dirk Willer · Citi

Korea is set to rejoin the AI story once levered retail positioning is cleaned out.

Korea's story is a DRAM story, and fundamental analysts expect DRAM to stay in short supply for a long time. But Korea rose 80% in one quarter, driven by highly levered retail investors, and such rallies usually end badly. Positioning is still not clean. Once it has been cleaned out after the ramp and the sharp fall, Korea would be easier to buy. Since he is positive on AI in the US, emerging Asia including Korea should be on the positive side for consistency.

“It's just having its positioning cleaned out a bit more after this huge ramp up, followed by a very significant fall. It would be easier to buy than if it doesn't happen.”at 29:27

Dirk Willer · Citi

Be cautious on European equities: no AI story, hit hardest by energy shock, plus ECB and political headwinds.

Europe lacks the AI story and gets hurt the most by an energy shock. The ECB may be over-tightening. French equities usually sell off into presidential elections that the market sees as unfriendly, which is a headwind. The euro is likely to weaken. Any global market is highly correlated with the S&P, and Europe is unlikely to decouple, so it would not be a source of outperformance. He is not extremely bearish, but Europe would not be a pick for him.

“be more cautious in Europe just because, yeah, it doesn't have the AI story, but it also has... It really gets hurt the hardest from an energy shock.”at 29:47

Dirk Willer · Citi

Clear long-term bull case for gold from central bank reserve diversification; avoid while the Fed is hawkish.

In the short run, gold tracks US two-year real rates, so a hawkish Fed argues against buying now. In the bigger picture, central banks are seeking reserve assets outside the dollar. When Putin diversified, he bought euros, renminbi and gold. Euros were a bad bet because the EU sanctioned Russia, and the renminbi cannot store wealth for China, which leaves gold. Gold bottomed when the Bank of England stopped selling and may peak when the PBOC stops buying. China could still buy a lot of gold if it wants to cut dollar holdings.

“So in the very big picture, there is a clear bull case for gold. I just wouldn't pick it up while people are concerned about a hawkish Fed”at 34:43

Dirk Willer · Citi

Gold would make a comeback once oil falls and central banks ease; could surge if equities peak

If oil falls back after the Iran situation resolves, hiking cycles end and some central banks move to easing. Gold would then make a comeback, and that could drive a big rotation across commodities and bonds. In a scenario where equity markets peak, the Fed would change direction, rates could fall to very low levels, and the Fed would probably have to buy the back end of the curve, so gold would go straight up. Dirk stresses that this is not the base case for next year.

“and when that happens, then gold would make a comeback too. So there really could be a big rotation”at 44:13

Dirk Willer · Citi

Sentiment is strangely subdued, unlike the aggressive sentiment typical of late-stage bubbles.

Bubbles usually show very aggressive sentiment, including among retail investors, who often pile in before a bubble ends. Current sentiment is strangely subdued: the AAII bearish indicator is quite high. Breadth is also narrowing, which is typical of bubbles as everyone chases the same thing. Muted sentiment suggests the bubble is not yet in its exuberant final phase.

“Sentiment is also strangely subdued, and then of course usually bubbles have very aggressive sentiment. Even retail investors”at 5:56

Dirk Willer · Citi

Midterms unlikely to matter much for markets; a Democratic sweep wouldn't change fundamentals much.

He does not expect much market relevance from the midterms. A Democratic sweep, as prediction markets suggest, would not necessarily change economic fundamentals much. The main midterm effect is a usual pullback of about 5% into the vote, which is typically recovered afterwards. The Iran war risk after the midterms is a bigger threat to the year-end rally.

“we don't expect much of market relevance out of the midterms, I don't think a democratic's fee, as the prediction markets predict, would necessarily do that much in terms of economic fundamentals.”at 7:40

Dirk Willer · Citi

AI working too well could push unemployment above 10%, creating a deflationary shock.

He lists AI working too well as one candidate for the next deflationary shock. In that scenario unemployment would rise to 10% or more, which would push yields down and make Treasuries work as a hedge. It is offered as a plausible scenario rather than a forecast, and he gives no detailed mechanism for the labour impact.

“one deflationary shock could be AI working too well, and unemployment going to 10% plus or whatever it is”at 20:25

Dirk Willer · Citi

ECB is hiking without an AI boom and is closer to a policy mistake than the Fed.

The ECB is hiking because of oil prices, even though second-round effects from oil into inflation are not yet visible. The European economy is more sensitive to high energy prices than the US economy, and it lacks an AI-driven investment boom to offset the hit. The first signs of trouble have appeared in French spreads, which show the cost of being too hawkish. That makes the ECB far more exposed to a policy error than the Fed.

“the ECB is a lot closer to a policy mistake than the Fed, because the ECB essentially is hiking without an AI boom”at 23:09

Dirk Willer · Citi

Dollar bullish bias: equities, rates and oil terms of trade favour the dollar over the euro.

The team's dollar model has four parts: relative equity performance, relative rates, terms of trade, and macro data momentum. Currently it is about 3:1 for the dollar on equities, rates and terms of trade. Data momentum still favours Europe. Relative equity performance depends on how long the AI bubble lasts, and US equities should keep outperforming Europe. Oil, via the Iran endgame, keeps the dollar bullish for now. Rates favour the dollar because the ECB is risking a policy mistake and French spreads widened. The euro is very oversold short term, so it may not be an ideal entry point.

“So we have a dollar bullish bias. Now in the short term, the Euro is of course extremely oversold, so maybe it's not the ideal entry point”at 23:44

Dirk Willer · Citi

China's GDP growth doesn't show up in equity returns; not convinced buying the China index pays off.

Across markets, equity earnings track GDP growth, but in China there is essentially no correlation. Huge GDP growth did little for equities, unlike India, where GDP gains show up in the market. Possible reasons include stakeholders other than shareholders getting the spoils, or hyper-competitive local industry. China has competitive, cheap AI, but he is not convinced owning the index would capture that. Exports rise while domestic demand stays weak, and the economy is still not back to where it was.

“they have very competitive AI and it's cheap. But you know what, I'm just not convinced that if I buy the China Index I will benefit from those factors.”at 31:49

Dirk Willer · Citi

Copper looks good now as global data momentum rises; stronger dollar caps the upside

Citi's global data change index tracks macro momentum in the data. Copper has a much higher Sharpe ratio when that index is rising, and it is rising now, with US, European and global data all improving. Copper does best when global data momentum rises alongside a weaker dollar, which is not the case now: growth momentum is strong but the dollar is probably somewhat stronger, so copper can still do well but is not in a hyper-bullish phase. Structural drivers such as AI and the energy transition support demand, though China was the last big copper cycle and is hard to compare with today's.

“And whenever that is moving higher, copper has just a much higher Sharpe ratio than if it's not. And right now it is moving higher.”at 38:34

Dirk Willer · Citi

Credit spreads are too tight to tighten further; credit would underperform in a recession

Dirk does not like credit. If the equity bubble bursts and the economy slides into recession, credit will significantly underperform. Holding credit as a hedge would not cost much because spreads are very tight and supply is endless, so spreads have little room to tighten from here. He therefore sees credit as a probably sensible hedge for an investor mainly worried about a bubble bursting.

“when the equity go bursts and you get a you into recession scenario, credit will really underperform, and it will just not cost you very much because spreads are really so tight and supply is endless as far as the eye can see... credit spreads will just not tighten from here.”at 42:30

Dirk Willer · Citi

Oil should fall back to around $70 once the Iran situation resolves next year

Dirk expects the Iran situation to be resolved almost certainly next year. The Iran war began with a large oil oversupply, so once it resolves oil should return to about $70, possibly lower. Lower oil would feed into central banks that target oil or headline inflation, bringing hiking cycles to an end and leading to easing in some places. He also notes that buying oil as a hedge against an Iran-driven spike is expensive, because implied volatility, though down from March, is still very high.

“Almost certainly the Iran situation will be resolved, which means oil will be back to $70 and, I mean, it could be even lower, but probably something like that, because we started the whole Iran war with a big oversupply of oil.”at 43:36