#20 -The Data Scientist

Tian Yang is an economist and the Chief Executive of boutique macro research firm Variant Perception which uses a combination of man and machine.


Tian Yang trained as an economist and is now the Chief Executive of boutique macro research firm Variant Perception which uses a combination of man and machine, following Gary Kasparov’s maxim that a man with a machine will beat either alone. They have followed this quantamental strategy for some years, long before it became more fashionable. We discuss this, the capital cycle, the long run cycle for commodities and how Tian solves two puzzles – the fundamentals and what he calls playing the game – the shorter market cycles.


Tian studied economics at Cambridge and started as a trader at Bank of America Merrill Lynch. He joined Variant Perception as an analyst and became Chief Executive just 7 years later. He had intended to become an economics professor but was seduced by the lure of finance. He won various competitions set up by large investment banks in his first two years at Cambridge and became hooked. His enthusiasm for markets is clear in our discussions.


VP produced a report looking at the rubber, whaling and fur industries to detect the long run commodity cycle. Steve has researched the rubber and plantation industries and recalls visiting Manaus, the centre of world rubber production in the early 1900s when tire demand was growing like cloud storage today. Manaus has an opera house which would look good in Milan.

The British stole rubber seeds from Brazil and planted in Malaysia. Later, when much of Asia was occupied by Japan in WWII, the US Government backed various programmes to develop synthetic rubber. In the 1970s, oil price supply shocks made synthetic rubber more expensive and natural rubber gained in importance. But the natural product and the synthetic substitute happily co-existed for decades.

These can be highly complicated cycles.

Tian points out that demand is often inelastic – fuel for heating is a good example. Therefore when supply gets disrupted, you end up with extreme price moves and prolonged cycles. In the long run, producers will try to expand the market and technology can enable alternative uses. For fossil fuels in particular, they have high energy density and are such efficient forms of energy that he sees them as hard to replace. He concluded that you often need a major event – a war or similar catalyst – to accelerate the development of alternative technologies.


The commodities discussion leads straight to the Capital Cycle. Tian became interested in the capital cycle theory after reading the book Capital Returns, the collection of Marathon Asset Management’s writings, edited by Edward Chancellor. He sees this as central to almost every sector and fundamental to the way the world works, but perhaps unsurprisingly given the man plus machine theme underlying Variant Perception’s work, takes a quantitative approach.

VP use 3 lenses to examine the capital cycle:

  1. capex to assets
  2. depreciation to assets and
  3. ROIC

The first factor is the level of capex + R&D to assets – this is a measure of to what degree a company or industry is increasing supply and VP aggregate by sector to give an industry outlook.

Depreciation plus amortisation as % assets gives an further indication as to whether the capacity is growing or shrinking and if it’s ageing.

The really clever filter is ROIC. Tian points out that in an industry like shipping, there can be a lot of capacity outside the quoted sector which may not even have a profit motive – a Chinese shipbuilder SEO  may be seeking to provide employment as much as generate an economic return, while Greek ship owning families may take a much longer term view. In order to determine if a supply shortage is worth buying, VP uses the delta in marginal ROIC to detect whether that capacity shortage is resulting in improved returns and is worth buying.


Tian views the world in 3 time buckets:

Structural Indicators: Modelling long term themes like demographics, capital cycles, debt and currency crises. The fundamentals are the strongest driver over horizons of 2-3 years and more and this is where their work on the capital cycle comes to the fore.

The firm views the business cycle over 6-12 months. They use leading indicators to navigate this time frame and pace emphasis on indicators like building permits and temporary labour which Steve has historically used as powerful indicators of the next trend in revenues for the related sectors.

VP also uses tactical indicators for trading and short term positioning over horizons of days and weeks.


The decades of cheap commodities, cheap labour and cheap financing are over. That makes for a very different investing landscape in which leadership clearly moves from tech to energy. The last decade of tech startups’ prevalence in an era of easy money are gone. Input costs are going up and the cost of labour is rising.

In the age of abundance, you can have independent central banks. In the age of scarcity, governmments will be in charge and we shall move to the China model of the government directing the allocation of credit. And these changes again dictate a change of sector leadership.


Tian joined Variant Perception in 2014 as a research analyst after working as a trader at Bank of America Merrill Lynch for 4 years where he helped to build their Delta One Global Index business. He became Head of Research in 2016 and moved up to become CEO in mid 2021 when founder Jonathan Tepper left to set up his asset management business. He went to high school in Nottingham then read Economics at Cambridge where he was runner up in the JM Keynes Essay Competition in 2009 ; the title was What Keynes would advise if he were alive today.

tian yang


We talked about Joel Greenblatt’s books, How to be a Stock Market Genius and The Little Book that Beats the Market. His two main recommendations were an old favourite and one which was new to Steve and has been added to his list.

This book, The Volatility Machine by Michael Pettis, was a game-changer for Tian. It’s a practical guide rather than theoretical. Pettis is best known as a China expert but has a brilliant mind.

Capital Returns, a collection of Marathon Asset Management’s writings should be well-known to this podcast’s audience, as we have had Jeremy Hosking as a guest.


Variant Perception has a website and a Substack.



Variant Perception was founded by Jonathan Tepper, who is a friend of Steve’s and he suggested Tian as a guest. Thanks Jonathan!