How Leveraged Money Becomes Less Leveraged (and less money)

One of the seminal moments in my career was when—during a long rant full of words like “structural, “endogenous” and “secular” in which I tried to defend the reason why I did not want to make an investment—I was interrupted by an older and wiser colleague who said to me, “kid, kid, shut-up. Just shut-up. You don’t get paid to have opinions; you get paid to have positions.” A fitting complement for Sensei KTF’s favorite, “It’s all about the positions.”

What I am referring to, of course, is that prices are set on the margin and that knowing the sensitivity of the marginal actor can and does give you an insight into how price will change. As Kevin hinted earlier this weekend in “A Poorly Run Casino,” we are now facing the economic consequences of the early aughts’ “commodities as an asset class” meme which was further aggravated with the popularity of momentum and/or “trend following” strategies. Money has flowed into commodities, initially from top-down allocators armed with spreadsheets and sharpe-ratios[1], followed my mom and pop’s registered rep who has now accepted as dogma that 15% of a portfolio should go into commodities funds, all but evaporating whatever dubious risk-premium (n)ever existed. Unfortunately for me and you and everyone we know, the size of the flows means that the open interest in many markets is now dominated by speculators—not just any speculator, but “managed money” aka, people like me, people paid to take positions, people who are by and large insensitive to current conditions and must take positions. For those in the cheap seats, let me spell it out for you: the marginal actor has become increasingly price insensitive. A poorly run casino, indeed.

Dr Copper needs a nurse

Despite CFTC reporting parties only making-up 15% of the global copper futures markets, we have strong evidence suggesting that, since 2006, managed-money (the marginal actor) has been dominated by non-economic actors with low price-elasticity, i.e. shameless price-takers. In other words, the volatility and slope of both the supply and demand curves (and by extension the marginal transaction and therefore price) is under the control of a very fickle and non-economically motivated herd that has exploded rapidly in size.

The chart below contains 4 time series,

  1. Weekly spot price of LME copper (bars) 20 week moving average (gray) and  2 standard-deviation Bollinger Bands (red and green)[2]
  2. Money Manager Net Positioning (# of  net-long contracts) / Open interest in high grade copper
  3. % BB of Chart 2 (Price - lower band) / (upper band - lower band)
  4. % BB of Chart 1 (Price - lower band) / (upper band - lower band)

copper bands

What is primarily of note here is the extremely high coincidence of the lower two oscillators, which tell us that money manager sentiment is highly coincident with over-bought and over-sold conditions. Second, it is of note that the money manager net-position can be as high as 20% of total open interest (currently, managed money constitutes 38.4% of the short and 13.9% of the long contracts). This strongly suggests that the overall sentiment of non-economic buyers is the responsible for practically all of the short-term (20 weeks, in this case) volatility in price. In fact, as the scatter-plot below illustrates, the correlation between the two oscillators is a high 0.84, and regressing money manager sentiment to predict price deviation results in an R-squared of 0.90, with a coefficient of 0.954 when electing to use no intercept. In other words, managed money is always longest before the drop and shortest before the rally.

copper pct BB scatter

The scatter-plot above shows the values of both oscillators and their very clear linear relationship, with observations less than one standard deviation from the model’s prediction in grey, those between 1 and 2 in blue, between 2 and 3 in green and greater than 3 in red. The latest observation is circled in red.

Those that would like to play devil’s advocate will counter that “correlation does not imply causation” and regular readers will reproduce Kevin’s “nothing changes people’s opinion like price.” It is our official opinion (and our position) that managed-money is the marginal price-setter and managed-money acts based on past price movements. This creates and perpetuates momentum-sparked self-feedback loops that all but guarantee that, as we approach inflection points, leveraged managed money (whether long or short) will become both less leveraged and less money. As very much economically-motivated buyers and sellers with real positions, real P&L and real margin clerks, the last thing we want to see is a rapidly increasing crowd running at increasingly faster speeds from one side of the boat to the other, which is, as you can see below, exactly what has been happening for the last seven years.

copper mm OI

[1] Inker, Ben. February 2013. We Have Met the Enemy and He Is Us

[2] LMCADY Index. LME spot copper price London settle

4 thoughts on “How Leveraged Money Becomes Less Leveraged (and less money)

  1. Pingback: A Few Good Reads | Derek Hernquist

  2. Kevin

    Copper will have to carry its economic baggage with it. I think in other products the elasticity may get even wilder. Grains, for instance, are still consolidating near prior cycle highs.


Leave a Reply

Your email address will not be published. Required fields are marked *