Blog

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.

Small item purchases and $30 happiness

With macro uncertainties, it looks like global consumers are focusing more on small items purchases.

In China, happiness starts with $1.

1/ coffee chain Luckin grew 88% in revenue yoy, and has over 10k store now.

For $2 per cup, if people get it every workday, it’s only $40 per month.

There are additional promotions: with $3 membership card/month, coffee can be almost half priced. $1.5 x 20 + $3 => $33 per month.

2/ mobile gaming (mostly micro transactions) market grew most 60% in 2023 summer. See previous post.

for $30-40, you can get a decent new skin on HoK or Justice Mobile. Some skins are as low as $1.

Globally, $20 per month can also buy lot of happiness.

3/ Netflix added 9mn subscribers in 2023q3 and is going to raise price again in the US. After 3 years, 50mn+ more subscribers joined Neflix, vs 2020q3, when the initial covid impact calmed down (2020q3 added 2.2mn vs. 2020H1 added 26mn).

After raising price, it’s only $23 per month.

4/ OpenAI, whose ChaGPT, although a productivity tool, is mostly paid by consumers with $20 per month plan. It has over $100mn revenue per month now, or 4-5mn subscribers, assuming most revenue is from subscriptions.

5/ Midjourney subreddit members grew to almost 1mn. Midjourney entry-level (Basic) plan is $10 per month. It’s reported that its 2023 revenue is over $200mn. Midjourney might have ~2nm subscribers by year end ($24-30mn revenue run rate).

CATL & Tesla growth? What does CATL business look like?

Are they supposed to be growth stocks?

Well, Tesla does have new stories besides FSD -> AI chips (Dojo) + robots (Optimus), which sound to be pretty exciting.

What about CATL’s future strategy?

Is it like Intel? 

Similarities: They are both the supplier for industry-leading products, and has in-house manufacturing (part of the edge is manufacturing. They can expand into other end markets: Intel also serves server market, while CATL serves energy storage market.

Differences: “Wintel” is amazing; however, auto industry hasn’t been winner take all, even for Tesla. Unless auto OEMs are willing to become PC makers and there is one superior EV structure (asset light) and uses CATL exclusively.. << a very unlikely picture. If Tesla had delivered FSD fast enough and good enough, there is a chance to be a layer that takes most market share, but that doesn’t have much to do with the battery layer.

Differences: Although CATL tech is very good and improves every year, it’s not like it created “Moore’s Law” / a long-term road map for the battery industry. It’s not just about visionary or “leadership”, but because the speed of improvements is fast enough – competitors are catching up instead of falling behind.

Is it like Qualcomm?

Similarities: Qualcomm’s SOC enables / provides a platform for smartphones and other IoTs. CATL enables lots of EV brands. Qualcomm faces competition from integrated player like Apple; CATL faces competition from integrated player like BYD.

Differences: similar to Intel, Qualcomm + Android is powerful and ubiquitous. CATL is lack of a powerful OS layer.

Differences: Qualcomm uses foundry for manufacturing and focuses more on design & licensing. CATL produces in-house, but doesn’t do licensing.

Is it like Denso?

Similarities: big player in auto parts.

Differences: Denso’s business lines is more mixed (no synergy), and centered around automakers. CATL focuses on batter-related products, and are supplying to non-auto customers.

Differences: Denso has ~15% gross margin and ~5% net margin. CATL has ~20% gross margin and ~10% net margin.

 

Why books (& reading) are one of the best investments?

You can get 20% return on stocks per year if you are as good as Buffet.

But books and reading have a much higher return!

The cost of a book is basically cost-plus, including those paid to the author. It’s relatively small sum of money – likely less than a meal.

The price is similar for every one. It’s up to each reader to capture all the upside to him/herself. No price discrimination.

The paper and the printing are basically similar across books – different words on a single page don’t mean they cost differently. Therefore, an amazing book costs the same as any other books (well, royalties, length, quality of paper may differ, but you know what I mean).

Put it another way, a good reader can get higher value from reading books, without increasing costs, if he/she can pick “high value” books (for him/herself).

It’s like Nvidia pays similar prices as others for the same wafer to TSMC, but gets higher value/return from it.

Besides money, readers are also investing their time. So, the return of reading a book could be even higher if he/she can read more efficiently, reducing time spent while still get most of the stuff.

It’s like fitting a few more chips onto one wafer.

What’s more impressive? TSMC has a high market share and some bargain power, but printing books seem to be very fragmented and has very good availability.

Enjoy reading great books! That’s likely a higher return investment than Nvidia selling chips.

 

 

Wanda Commercial refinancing YTD

Sources

  • 年初融美元债$700mn,12+% yield
    • 2025 $400mn @11%
    • 2026 $300mn @11%
  • 境内中期票据 15 亿 rmb,6.8%
  • 卖万达电影股权,两笔共40多亿rmb,
    • 一笔中国儒意(背后腾讯出钱)
    • 一笔是东方财富的老婆
  • 卖5座万达广场股权;price? to 大家保险 etc
  • 发行ABS; didn’t work
  • Net income ~$1bn in 23H1

Uses

  • 还40亿 rmb 票据
  • 还$400mn Jul 2023美元债@6.875%
  • Sep 回购 5.58% 38亿puttable境内债 20万达01

Upcoming: Jan 2024 $600mn 7.25% note

Potential: didn’t IPO -> to repurchase 30bn rmb

Is China’s mobile gaming market growing?

Yes.

In 2023 summer (July and Aug), China’s mobile gaming market grew 57% yoy (!).  China’s overall gaming market grew 46.1% yoy in August.

That’s an amazing number, considering it’s quite mature and developed already.

The base effect is obvious though. China’s most recent crackdown on video gaming back to Aug 2021 lasted around a year (or15-month if consider Tencent’ new commercial license as an end).

New game license issuance resumed in Apr 2022, stopped once again in May 2022, and resumed in June 2022. However, big companies like NetEase didn’t get new commercial game licenses until September 2022, and Tencent waited until November 2022 (no new game license issued in Oct 2022).

The crackdown led to yoy decline of ~24% in summer 2022 (July and Aug) for mobile gaming, which is ~70% of the market.

However China mobile gaming is reaching a new high now. It’s 20% higher than 2021 summer.

Comparing with 2020 summer, 3-yr CAGR would be 8.9% for mobile gaming.

BYD export machine running hot; EU’s investigation of Chines EVs; disagreement between Scholz and Macron

BYD started to report overseas sales in July 2022.

So we have the first YoY quarterly number: over 300% yoy growth! (over 4x)

Overseas volume as % of total volume grew from 3% in 2022q3 to 8.6% in 2023q3 – still looks to be a low number.

UBS claimed that BYD has “a sustainable ~25% cost advantage for BYD over legacy competitors”. (Sep 2023)


A few days after UBS report, EU launched investigation into Chinese EV subsidies.

Meanwhile, as Bernstein and WSJ pointed out, not all of those come from Chinese automakers; Tesla, and other European JVs in China exported a ton. Bernstein’s number says Tesla’s export from China to EU is 10x BYD’s.

BMW’s China JV reported that it exported 16,595 units, mainly X3 BEV models, to overseas in the first half of 2023 (didn’t say where, but Bernstein puts BMW China JV EU export number higher than BYD’s EU export).


Unlike French peers, German carmakers still sell a lot of ICE cars in China, and exported EV from China JVs. So Scholz is taking a different view vs. Macron. That’s just one of the issues the two need to reconcile/solve – the other two issues are energy bills and defense, laid out by Politico.