WeWork was more black swan than unicorn for SoftBank

[TAIPEI] SoftBank Group Corp's WeWork debacle is unlikely to be repeated. It also means that for the Vision Fund to deliver venture capital-size returns, it'll need to rely on singles and doubles instead of hitting home runs.

That's not because Masayoshi Son has learned his lesson. His defiance at an investor conference on Wednesday, including almost an hour spent justifying the WeWork investment, tells me that SoftBank's chairman hasn't changed.

But simple mathematics indicates that WeWork won't happen again. The Vision Fund has US$97 billion, of which it's already deployed more than US$85 billion and has ceased making new investments. It has 88 portfolio members, which means it has doled out around US$1 billion apiece. Prior to SoftBank's non-bailout-bailout (Mr Son insists it wasn't a bailout. Most everyone else disagrees.),  the Vision Fund and SoftBank combined had invested US$10.5 billion into WeWork with a valuation of US$47 billion.

Only three investments come close: The Vision Fund has put US$5.3 billion into ride-hailing company Didi Chuxing, with a price tag of US$56 billion (The exact amount isn't clear. SoftBank filings show that Didi was the only holding in its Delta Fund as at Sept 30, 2018, with SoftBank Group transferring its stake to that fund for US$3.7 billion. It is then reported to have invested a further $1.6 billion.); US$3.5 billion into Grab Holdings Ltd, with a valuation around US$14.3 billion; and just under US$3 billion into video platform ByteDance Ltd, valued at around US$75 billion.

Rather than plough the remaining US$12 billion or so left for follow-on investments into any single mega-unicorn, the Vision Fund would probably get better returns putting the cash into smaller startups. Theoretically, there's a lot more upside to giving US$1 billion to a nobody than US$10 billion to a somebody.

This means that by both size (the large amount committed), and scale (its almost 30 per cent stake), WeWork and its collapse are more like a black swan than a unicorn representative of the Vision Fund's broader portfolio. In his 2007 book — The Black Swan: The Impact of the Highly Improbable — Nassim Taleb writes about the extreme impact of rare and unpredictable outlier events. While it could be argued that WeWork's decline was predictable, few expected it to implode on this scale.

The uniqueness of the WeWork mess was on display in SoftBank's September-quarter earnings. Its US$4.6 billion writedown on WeWork was a significant contributor to SoftBank's first loss in 14 years. Rare: Yes. Unpredictable: Somewhat. Extreme impact: Absolutely.

Based on its current investments, only a full write-off on Didi could top the havoc caused by the office-rental company. But if we're going to see another huge loss in the portfolio, it'd probably be an encore performance by WeWork.

There are two caveats to this analysis: The Vision Fund could well put that US$12 billion or so back into ByteDance, Didi or Grab, boosting exposure. Or the Vision Fund 2 may come along and invest into existing portfolio companies, and even find new mega-unicorns to prop up. The latter, while possible, seems unlikely because the runway of massive startups is quite limited.

The flip side to such lower exposure is that the rest of the companies in the Vision Fund's portfolio are too small to drive huge returns. I managed to track down the names and valuations of 64 of the fund's portfolio companies and found that only seven have a valuation of US$10 billion or more, while 10 are in the US$5 billion to US$9.9 billion range, according to data from SoftBank statements and CB Insights' unicorn list. More than half have price tags of US$2 billion or less.

I suppose you could argue that this gives the Vision Fund plenty of upside. Yet it's also true that a large portion of those startups will amount to nothing, literally. It's the nature of venture investing that most of the companies you bet on die before you see a dime in returns. And there's no evidence to suggest that Mr Son and his team are better at picking startups than anyone else (quite the contrary, perhaps).

While swinging for the fences on dazzling startups may excite the crowds, a run is a run even if you get there by bunting.