Tin Men

This past weekend David Kocieniewski wrote a lengthy New York Times article on Goldman Sachs’ influence over 27 industrial warehouses that store aluminium. Slate’s Matthew Yglesias queried the alpha return drivers on Goldman Sachs’ decision.

 

My gut response on reading Kocieniewski’s piece was that it had something to do with trading: probably proprietary desk strategies that used swing trading or that manipulated spreads. I had experienced that in late 2011 during my first trades, and pieced the evidence together from regulatory filings and visits to Mitsubishi UFJ Bank in Tokyo, Japan. Yglesias later confirmed my instincts when he linked to Izabella Kaminska’s FT piece that explained the Goldman Sachs trade.

 

I learned a couple of things from this exchange. Three years of hard work means I am starting to slowly understand how financial markets really work. I saw from a series of past trades how similar dynamics could be repeated today. My gut instincts about the Goldman Sachs trade were correct. But I didn’t have the precise, conceptual understanding of the commodities markets and the relevant central bank regulatory loopholes — despite knowing what contango and backwardation were — to nail the Goldman Sachs commodities trade in the way that Kaminska did.

 

This is a humbling intellectual experience.

2nd February 2012: Gina Rinehart’s FXJ Move

On 31st January 2012, mining magnate Gina Rinehart bought nearly 8% of Fairfax (FXJ) through Morgan Stanley.

 

New Matilda‘s Ben Eltham observed:

 

Precisely why Gina Rinehart is buying a stake in Fairfax remains a mystery. Neither Rinehart nor her company, Hancock Prospecting, have issued any comment on the move. Rinehart simply issued instructions to her broker, and bought up stock worth about $180 million. . . . It’s simply not necessary to buy a stake in a media company to get your message across.

 

Eltham and Jason Wilson each suggest Rinehart’s bid is to gain control of Fairfax and to promote her political views.

 

Examining the pattern of FXJ trading suggests other possibilities. FXJ jumped from $0.74 close on 2nd February to open at $0.82 on 3rd February. There were major sell-offs that day: during a 14-minute rally period from the market open 10:04am (4.27 million shares), 10:16am (1977.69k shares), to 10:18am (1523.01k shares); during the trader lunch period at 1:22pm (2779.19k shares); at 2:24pm (2185.33k shares); and an end of day sell-off (5.25 million shares) which ensured FXJ shares would open lower the following day.

 

The sell-offs fit a well-known strategy used amongst institutional trading desks: the ‘market squeeze’ trade. In late 2011, I watched J.P. Morgan and Japan’s Mitsubishi UFJ bank use this strategy with several other Australian shares. I found out the details from two sources: ASX regulatory filings made on behalf of offshore hedge funds, and from ThomsonReuters’ SIRCA database which has tick data of individual trades.

 

Here’s one way how the ‘market squeeze’ trade works:

 

1. The trading desk buys up huge amounts of a target share: enough to move the share price. This creates a volume spike that will initially move the share upwards: a stochastic market dynamic used in jump diffusion models of mathematics and option pricing.

 

2. The volume spike creates a red alert which attracts other, different traders. Day traders who use technical analysis, charting, or momentum/rally signals now focus on the share. Exchange trade funds and institutional money managers who must rebalance their portfolios are now also interested. This creates a market for the trading desk to sell to. The share also shows up on the daily volume indicators of the major share trading platforms. The financial media becomes interested.

 

3. The trading desk then dumps a large volume of the share at strategic times during the day. This locks-in a short-term or daily profit for the trading desk. It also influences the upper and lower bounds of the share price. Monte Carlo Markov Chain simulation can predict the share price pathways. The trading desk can then adjust its order book and its market execution costs.

 

4. Meanwhile, the trading desk sells off smaller blocks of shares over a 3-4 week period. This tactic influences high-frequency trading systems. It usually means that the share price trends downward as the trading desk has market-maker control — or several trading desks at different firms create a market equilibrium.

 

5. The trading desk can make profits in several ways. It can dump a large amount of shares at the market open which usually means the share price will fall during the day. This tactic will create cyclical and volume-based effects. It can force other traders to sell once their stop-loss levels are breached. Finally, it can sell shares at the peak of a volatility spike and then buy them back at a much cheaper price when the market trends lower. The trading desk’s volume, its order size, and its lower execution and transaction costs means that it can make a profit from spreads of several cents.

 

FXJ’s trading after Rinehart fits the ‘market squeeze’ pattern. It’s possible that Rinehart will pursue the ownership agenda that Eltham and Wilson emphasise: Michael Milken financed Sir James Goldsmith and others to do so in the 1980s era of leveraged buyout deals. But it’s also possible that Rinehart is using value investment criteria to make a quick profit from market volatility. Or, that Rinehart’s announcement enabled Morgan Stanley and/or other trading desks to use a combination of long/short, paired, and event arbitrage strategies. Someone made a killing on trading FXJ on 1st February 2012.

 

Eltham is right: you don’t need an ownership stake for a media company . . . if you pursue other agendas.