Out of Sample Data and Quant Quake #2

September 2019 Investment Update

Dear investors and well wishers,

Portfolio Update

The fund returned 10.9% in August, ahead of the MSCI world total return at 0.2% and the ASX200 total return at -2.4%, bringing us to +25% net for calendar year-to-date.

Out-of-sample Data

This was a good reporting season, not because the numbers were this or that, but because our core investment theses stood up to new, out-of-sample data.

In particular:

Carvana punched through their mid-term profit target of over $3,000 a car, reporting $3,175/car, up from ~$2,100, and maintained their organic growth rate at over 100%.

At 0.4% market share, with a cheaper and better offering than competitors, we expect the firm to reach 4% market share in the mid-term, which would be a 10x increase in size. Given that we bought at a ~1.5x forward sales, there’s every chance of a sustained increase in multiple too. 

Pinduoduo showed they’re going to take the fight to Alibaba, and are now taking something like 31% of e-commerce growth in China, vs Alibaba at 43%.

Margins also increased significantly, suggesting that, like Carvana, Pinduoduo is under-pricing their offering and only scratching the surface of potential demand. It’s rare indeed to find two companies over valued at over US$10 billion growing at >100% organically while improving their economics.

In the past few weeks, Pinduoduo’s enterprise value increased from sub $20b to over $35b. Since the firm is growing at about two thirds the rate of Alibaba in dollar terms, this may be only a fraction of the firm’s future value. For a sense check, two thirds of Alibaba’s market cap is over $300 billion.

We remain astonished that a newcomer could capture 10% of the Chinese e-commerce in just four years, with most of that in the last two. There remains a striking opportunity for someone to pull this off in the West.

Afterpay is highly topical these days and I’m reluctant to discuss it, as most readers have likely passed it on at lower prices, or sold their shareholdings at substantial profits.

But their latest update demonstrated the product can scale in the UK, as uptake has been faster than it was in the US, which was itself faster than Australia, where the firm is near ubiquitous in major segments of retail. Adding 200,000 British customers in a few weeks is quite an achievement.

Repeat customers are now spending over 20x a year on average. When you incorporate two forms of rapid expansion - in this case an increase in the number of customers and an increase in spend per customer - you get the kind of growth that can lead to extraordinary changes in value.

We suggested years ago that loss rates might fall as poor borrowers were weeded out of the customer pool. We can now support this thesis with out-of-sample data, as loss rates have fallen from 1.5% to 1.1%.

A significant majority of our portfolio is now invested in mid/large cap global stocks posting organic revenue growth of over 100%, with improving economics.

Quant Quake - Under the Surface

One of my favourite stories from the GFC was the quant quake of August 2007, when a large number of market neutral long/short funds lost an inordinate amount of money in a brief period of time.

These funds were using computers to trade pairs of stocks. An Australian equivalent might be to buy ANZ when it’s cheap and sell CBA when it’s more expensive, for no net equity risk, then wait for the values to move back in line.

All these strategies have an Achilles heel: everyone is using the same data, so comes to the same conclusions.

Magnifying the problem, as more capital is put to work, risk-adjusted returns actually improve. These strategies show their highest returns and lowest volatility right before they’re about to blow up.

Never trust a straight line in finance!

For the reasons above, these funds are largest right before they crash, so when they fall, they can quickly lose more than they ever made.

The quirky part is that because the funds were market neutral, longs and shorts were the same size. When they liquidated, there was no net change to equities. All the chaos was all going on below the surface.

Something like that happened last week, as momentum leaders sold off and highly shorted companies ripped upwards.

Our largest positions were largely unscathed, with Afterpay, Carvana and Pinduoduo still well ahead of where they reported a few weeks ago, perhaps because they're short-seller favourites too. Ofcourse, we're investing for long-term outcomes, so are largely unfazed by these kinds of movements when they strike.

However the sell-off last week has hit some of the highest quality companies in the world, many of which lost 20-30% in a couple of days. [Some of these advanced 200-300% over the past year or so, so best to keep this all in perspective. Long term holders are still the real winners.]  

It’s generally wise to buy market leading companies during indiscriminate liquidations by quant funds, so we plan to slowly and incrementally add to the highest quality firms if and as the sell-off continues.

The companies we like, such as Shopify, Twilio and Alteryx, will shortly report revenues 5-10% bigger than those of only two months ago. As we move forward in time, and prices move down, the valuation picture can change quite dramatically. 


Only a week ago, there seemed a consensus that recession was imminent, US bond yields were going to zero, and at least one person told me he’s quitting his buy-side job after making a bundle in gold stocks and Government bonds year-to-date.

I have the good fortune to speak to sophisticated and significant investors as part of my day-to-day, and anecdotally I can confirm there is a strong bias to reducing equity exposure, with a common plan being to wait until the coast is clear to reallocate to equities.

This sounds very sensible, but in practice means buying back in when prices are higher. Equities are always most dangerous when the data is positive and the coast looks clear.

I thought I'd share some interesting charts that show a different perspective.

1. Central banks have now been easing for some time, and this typically leads to an improvement in manufacturing data, which is where the current slowdown has hit hardest. 

2. Sentiment is quite negative, which as mentioned, I've observed first-hand.

3. Hedge funds have very low net equity exposure. These charts are somewhat amusing, as they show the most sophisticated market participants are consistently selling at the lows and buying at the highs.

Hedge fund exposures are quite relevant to equity returns, as most equity trades are rotations from one stock to another.

Hedge fund trades can be different, particularly if they borrow when they buy, and repay debt when they sell. This results in incremental changes to the supply and demand of stocks, in a way that long only fund managers trading with each other do not. 

4. Finally, this chart shows the extremity of last week’s value/momentum reversal (it was largely momentum that reversed, not growth). 

For what it’s worth, previous momentum sell-offs like this marked good times to buy, not to sell.

Hope you’ve had a good week,