Blog

US-China direct flights recovery

Recovery Tracker

after China reopened in 2023, flights were set to increase from 16 per week to 24 per week, announced in March 2023.

In Aug 2023, two sides agreed to double capacity of 48 per week, ramping up to 36 per week on Sep 1, and 48 per week on Oct 29.

Flights would further increase to 70 per week starting Nov. 9.

Technical difficulties

US need to avoid Russian airspace, which requires longer distance and thus refueling.

Impact on tourism

e.g. SF: visitors from Mainland would be only ~20% of 2019 level.

“In 2019, 518,000 of San Francisco’s 4.3 million international visitors were from China, according to data provided by SF Travel. Though visitors from Mexico outnumbered them by about 100,000, visitors from China spent the most of any group, accounting for $1.2 billion of the $7.7 billion international tourists spent in the city that year.

This year, visitors from China are expected to number only one-fifth of their 2019 total, and expected to spend just under $450 million. That brings the city’s total international visitor spending down from 2019’s $7.7 billion to an expected $5.9 billion in 2023.”

— SF Chronicle (https://www.sfchronicle.com/sf/article/international-tourism-china-recovery-18188305.php)

 

Other sources:

https://www.regulations.gov/document/DOT-OST-2020-0052-0165

 

US new home sales resilient

Looks like the monthly sales is still health. Month to clear inventory is steady and up a bit to ~7.8 months in Oct 2023.

Currently monthly sales pace is better than 2018 and 2022, despite record high interest rate in recent years.

New homes for sales has gone up more. So the number month to clear new home inventory has gone up to 7-8 months recently vs. an average of 6.2 months in 2018. And is much better than the 2020-21 average of 5.1 months.

Better availability should be good for inflation and soft-landing scenario.

New residential sales Oct 2023

See the other post for China new home sales – the inventory stood at over 20 months the last time I checked.

Commercial real estate problems summary

A good summary from Rob Stuckey, head of Carlyle’s U.S. real estate funds, on US office building weakness, from Insights and Indicators podcast by Carlyle:

  1. Already weak before pandemic
    • oversupplied
    • low operating margin
    • high correction to GDP / exposure to macro cyclicality
  2. Secular trend of work-from-home / technology trend

Factors to value real estate

  • demand drivers (macro/GDP, demographics)
  • technology
  • operating margin (high maintenance/recurring capital expenditure)
  • tenant stickiness (demand ever increasing)

 

Meituan’s changing financial reporting

Back to 2021-2022, Meituan’s quarterly results experienced various changes in reported metrics, which looks a bit dubious and problematic – whether it’s due to conflicts when measuring performance internally and to investors, or gov’s implicit requirement, or regulation changes.

Here are the 4 changes:

1/ Food delivery revenue split (2021q4): “Commission” split into “Commission” and “Food delivery services”, not segment changes.

2021q3

2021q4

2/ no more “GTV of food delivery” and “number of domestic hotel room nights” (2022q1)

2021q4

2022q1

3/ big change in 2022q2: new segment reporting of “Core local
commerce”, which combines previous “Food delivery”, “In-store, hotel & travel” & some business previously in “New initiatives and others”, e.g. Meituan Instashopping (美團閃購)

2022q1

2022q2

This segment reporting is used as of today.

Plus, in operating metrics, “Number of food delivery transactions” is now “Number of On-demand Delivery transactions”.

2022q1

2022q2

This is interesting – according to the footnote, “Number of On-demand Delivery transactions” includes number of transactions from food delivery and Meituan Instashopping businesses. While it’s consistent with “Core local commerce” definition, it’s hard to argue why business like Meituan Grocery (美團買
菜), which is under “New initiatives and others”, is not on-demand delivery transaction.

Plus, since “Core local commerce” now includes in-store, hotels etc., which has nothing to do with “delivery”, it’s hard to know the unit economics for delivery.

4/ No more reporting of “Number of Transacting Users”, “Number of Active Merchants” and transaction per user (2023q1)

2022q4

Gone in 2023q1

US-China recent deals round-up

Biotech

Nov – Modern Shanghai plant break ground; the $1bn deal was signed in July

Oct – Junshi’s PD-1 drug, with US & Canada right purchased by partner Coherus was approved by FDA

Agricultural purchases

Oct – signed 11 purchasing agreements/contracts, worth multiple billions in value

Nov – 600k ton soybeans; and then 3mn+ tons; a good summary here

Industrial / Areospace

Nov –  GE Aerospace’s 25 GEnx-1B engines order from China Eastern Airlines to power its Boeing 787 fleet.

Nov – Xiamen Airlines purchases.

Market is speculating more Boeing orders.

Consumer internet / tech

Nov – Meta’s Oculus is coming to China in late 2024 (w/ a lower-end version of Quest 3), with Tencent as partner; WSJ reported the talk between Meta and Tencent back in Feb 2023.

Nov – Nvidia to release 3 new chips for China market (H20, L20 and L2), available as soon as the end of this year.

US car vs. home loans

Read the q3 quarterly report on US household debt and credit (here). One interest takeaway is how divergently different loans perform, vs. the GFC era.

When the GFC hit, both all loans perform badly. Transition into delinquency (30+ days) for credit cards, mortgages, and auto loans reached over 10%. Mortgages delinquency were picking up faster and looks worse than auto loans.

This time around, mortgages looks fine (as of 23q3), and delinquency is going up not only slowly but at the level even lower than 2005-06, while auto and credit cards are deteriorating at a faster rate.

By age groups, for 18-29 and 30-39, the percentage of auto loan balance falling 3-month behind is reaching about the similar level of GFC era. (Another theme: younger generation is under more pressure than the older for the past 2 years)

To reflect back, there was a shortage of cars during the pandemic and used car prices were shooting up. It could cost some money if someone bought a car back then and sold it this year, as 2nd-hand car price has been on a downward path.

The selling and downward pricing trend could be a self-reenforcing process.

Meanwhile, house is a more resilient asset class and current macro is still ok. After 10 years, houses are very likely to worth more but cars very likely won’t.

As shown with the FHFA house price index, which is keeping up.

…which is very different from the GFC era when HPI declined and under pressure for years.

China new home sales 2023 Sep vs. 2020 Sep

Facts for China new home sales

2020 2023
Sep Sep
‘0000 sqm 万，平方米
Supply 供应面积           7,529           2,990
Demand 成交面积           5,228           2,262
Inventory 库存         50,739         51,221
Sales pace (month)             9.71           22.64

Inventory actually didn’t increase much, flat after 3 three years.

But the willingness to purchase (new homes) has decreased, area sold in Sep 2023 is less than half (43%) of 2020 Sep level.

Therefore, the resulting month-to-clear-for-sale-homes is more than doubled from ~10 month to almost 23 month – it will need almost 2 years to clear new house inventory at current sales pace.

The above figure is for 100 cities in China.

To look at the bottom 10 cities: back in Sep 2020, the worst 10 cities needed 21.5 – 35.3 month to clear inventory whereas in Sep 2023, the worst 10 cities will need 57.6 – 93.9 month to clear inventory.

Source:

http://m.fangchan.com/news/320/2023-10-25/7122778841134993672.html

http://news.dichan.sina.com.cn/2020/10/27/1274844.html

 

 

 

Meta’s growth potential?

Meta’s bottom-line looks amazing – diluted EPS almost tripled from a year ago (+168% yoy).

How?

  • Headcount shrunk 25%
  • Revenue grew 25%
  • $20bn+ buyback in the past 4 quarters

Cheers to Susan Li, the new CFO announced back in 2022q2 earnings. Delivering numbers that investors needed.

Efficiency has improved dramatically – quarterly operating income per full-time employee more than tripled from $65k to $208k.

AI story is impressive; and Metaverse is not dead.

What are the concerns?

1) two-year cagr not impressive: at the midpoint of 2023q4 guidance, two year revenue cagr (vs 2021q4) is <7%.

Two year ads revenue carg for US, Canada and Europe is 6.7% in 2023q3.

Remember, most of Meta’s revenue is ads in US, Canada and Europe (2/3 in 2023q3). User growth obviously is not meaningful. It needs ARPU to grow. While ads pricing won’t be strong given macro uncertainties, it will then rely on showing more ads to users, which won’t be something people would enjoy.

2) operating cost would be higher: infrastructure cost would rise due to AI investments. Reality Lab operating cost would be higher. Two large layoffs were done; hard to cut further. More importantly, new revenue streams are less lucrative than ads (which has over 80% gross margin).

2) regulation, fine: Meta was sued – that has hit the headline. Meanwhile, EU’s DMA would take effect next year. Plus, AI is very data-driven. However, can companies easily get data this time around?

Will EPS continue to grow at 15% or above for 2024, 2025 and beyond? I think doable, but is AI an easily profitable business? Let’s see.

Tesla SOTP… $360bn?

Two biggest component would be electric vehicles and AI.

EV: BYD is <$100bn; BYD delivered higher profits than Tesla; BYD also has energy storage business

AI Models (FSD): OpenAI is <$100bn; latest valuation appeared to be $86bn

AI Chips: AMD is ~$160bn. Tesla should be years behind in terms of external revenue profits (as a business) etc. Let’s just use $160bn, as Tesla has some other business.

Then it sums up to be $360bn = $100bn + $100bn + $160bn

Last time I checked (before earnings), it’s more than doubling that number…

Robots? Boston Dynamics was $1.1bn back in Dec 2020.

Charging? ChargePoint ($CHPT) is ~$1bn market cap.

If someone wants to add ride-hailing services I mean Lyft is <$5bn, not significant.

Even additionally add Enphase and First Solar, which were ~$13bn and $16bn, still not big enough to move the needle.

Insurance could be big. However, if it’s not good enough/taking over now, why it should be in the future? It shouldn’t be a futuristic thing; auto insurance has a history of over 100 years.

$400bn or lower sounds about right.