A report from Reuters (hat tip: Felix Salmon) attributes the wild commodity price moves last week to algorithmic stop-loss trading.
Jim Brown offered these details on the oil trading:
Those funds interviewed said the massive amount of stop losses that were triggered was beyond comprehension…. When the crash finally came the number of positions liquidated was staggering. As each technical level was broken it triggered more stop losses and more short selling to capture the drop….
Credit Suisse analysts said the high frequency and algorithmic trading accounted for about half of all the volume in the oil markets.
Every day, futures prices can and do move in response to how many people want to buy or sell the contracts. As I explained in a recent study in Brookings Papers on Economic Activity, inventory arbitrage forces the spot price to move along with the futures price. But as I also explained there, this does not mean that sentiment or speculation alone can put the price of oil at any arbitrary value. Ultimately, the critical question is whether the spot price is one at which the physical quantity produced is equal to the physical quantity consumed. Whether today’s price indeed accomplishes this was the focus of my discussion of these events last weekend.
But clearly, there are lots of big traders out there who are thinking not along these lines but rather in terms of Keynes’ beauty contest, supposing that all that matters is guessing what other people think the price will be, and fooling themselves into believing that by programming their computers to buy high and sell low they are going to outsmart the other players.
Count me among those who maintain that buying high and selling low is unlikely to be a successful trading strategy. But if enough people believe otherwise, they can wreak a bit of havoc on the rest of us before they themselves go belly up.
“Fatih Birol, chief economist of the International Energy Agency, says world oil production has already peaked.”
http://www.futurepundit.com/archives/008065.html
I really don’t believe oil is that kind of natural market, the kind where producers are always and already pushing for full price discovery.
We know, for instance, that the Saudis have a long term strategy and seek “moderate” prices in the short term.
Why wouldn’t speculators have the power to do price discovery in place of aggressive producers? And if their attention is inconstant, why wouldn’t their influence be chaotic?
Professor,
Good post.
It would seem to me that so many stops, to cause this kind of decline in the market, should signal a huge insecurity in the investment community that the price of oil will be sustained. Those who sold on the higher stops probably jumped back in at the lower price and that is why the price has risen the past few days, but how comfortable will they be even at the lower price?
Thanks for stimulating my thoughts.
I don’t think your scepticism about algo traders is right. There are features of markets — trending and mean-reversion — that enable both short-term algo traders and systematic trend followers to make money over time. “Buy high, sell higher” is the mantra of the trend-follower; “buy low, sell high” may work for value investing, where there is a fundamental value to hold on to and probably a yield (e.g. in equities) but is a dangerous game in commodity markets — especially if you are leveraged or suffering a negative roll yield.
I agree that the sell-off had a technical cause (the trigger was the hike in COMEX silver margins). So did the flash crash of May 2010 (somebody selling S&P futures in an illiquid market) but that didn’t stop markets being depressed until August on concerns about a double-dip and deflation — these crashes to tend to occur against a backdrop of fundamental concerns.
Anyway, I hope that is right, as I shorted equities last week. Sell in May and go away; short in May and spend the summer worrying.
As always, insightful remarks.
From the title of the post, I was surprised not to see mention of the CME’s substantial change in performance bond requirements for crude. For example, the Brent Crude contract (BZ) requirement was raised from $5250 to $6500. This contract is for a 1000 barrels or roughly $114,000 in crude oil.
A technical detail to be sure, but one that has triggered margin calls throughout a number of leveraged speculators. Rumors indicate this move was to “shake out” the small investors from a volatile market (WSJ and Forbes).
As these investors “de-leverage”, we experience supply driven price declines coupled with lower demand.
Silver was also impacted with increased requirements.
So in the short term prices can fluctuate due to speculation, but ultimately the prices have to reflect demand for the underlying commodity? Is that about it?
It looks like the time t=0 observed excess supply is trumping the time (t+n) expected excess demand.
Steve: Yes, I would agree with that.
Does this not suggest that some regulation is needed in futures markets?
To trade oil, mineral and grain futures would it not be sensible requirement that traders either produce, store, or consume these raw materials?
Since these commodities are input costs in our economy does it not seem wrong that they are used as investment vehicles to gamble upon – ultimately leading to way higher trading volumes and by consequences higher prices than would Logically happen if if trading was limited to physical product?
And I’m still waiting for gasoline price to drop even a penny! that market sure follows oil when it’s going up. “Up by dimes, down by pennys.”
As I have written before here, I think the market is pretty disconnected from any understanding of what the true marginal cost of production is. Until someone can show me data from the Saudis, the simple truth is that no one knows what the MC of oil is. If the markets really understood it, we wouldn’t have gone from 157 to 30 to 110 to 97.
There is no other explanation for market behavior. We don’t know what we don’t know, and the attempts to speculate on the MC based on the current market price are about as trustworthy as a 2008 estimate of real estate prices from Zillow. There is no real data here, so all we have is speculation, and then algo’s on top of the speculation that love volatility.
I think people don’t want to admit that as important a variable as the MC of oil is so unknown – and so they fight the conclusion.
And yet when you push the data you get back never directly answers the question.
If we are talking about puts, then buy high and sell low may be just the thing (for the price of the underlying asset that is).
It seems dangerous that we allow such speculation in the market.
I can understand the idea that these huge price swings are temporary and ultimately the price will reflect underlying demand, but a big enough temporary price swing is all it would take to cause a recession. And even if it doesn’t cause a recession, such wild swings surely make it difficult for users of the commodities to plan their businesses.
Jeremy: If I as a producer want to hedge by selling long, and only those who are physically involved in the product take positions, who is supposed to take the short side to sell me insurance for the hedge?
Speculators play a potentially vital role of both information discovery and providing a counterparty for desired hedges by producers or consumers. They perform this function well if speculators try to buy low and sell high. They perform this function poorly if they try to buy high and sell low.
The goal of an improved regulatory structure should be to eliminate some of the destabilizing speculation we have seen but not to throw out the potential benefits in the process.
My thought for the day:
When flying out of JFK Airport, leave an extra hour more than you budgeted to allow for the awful traffic.
From JFK… SK
HFT should be prohibited by law and the perpetrators then encouraged to set up in Atlantic City, “Lost Wages” (a.k.a. Las Vegas), Morocco, Macao, Singapore, Tel Aviv, and elsewhere to allow algos to “simulate” financial markets trading and let “gamblers” engage in all manner of unregulated activity 24/7. The can flash crash the gambling venues all they want.
But the now publicly traded exchanges would object because the trades occurring at trillions of times a picosecond would reduce the fees the exchanges receive from HFT.
Shut it down!!!
Many institutional investors have implemented “real return” asset classes in the last few years. One of the components is typically a long, index-oriented commodities fund (typically based on the GSCI). 15 years ago it was controversial for a pension plan or endowment to invest in commodities. Now it it is not controversial; in fact, we are beginning to see them in defined contribution plans Thus, I think that at the margin there has been a steadily increasing demand for the long only position that can easily get pushed up into sustained momentum trades when negative supply shocks hits or demand takes off (emerging markets). It takes awhile, but the prices eventually have to reflect actual demand and supply conditions. I think the result is greater volatility in the commodities market, but I’m not sure we really have the data to test this.
Thank you for an excellent post John Butters. For those somewhat unfamiliar with how real-time traders and investors actually work, please read or re-read the post. Despite the legitimate quibbles raised by JB, JB’s description and JDH’s view of strategies and time horizons are fully compatible.
Jim, you may be right that futures market liquidity fluctuations can cause short run volatility in crude prices, but I doubt that such shocks get passed through to retail gasoline prices. A couple of years ago I wrote a short piece making the point that retail gas prices don’t incorporate all the wiggles we see in crude. If you’re interested, the paper is here.
The theory that speculators can’t make money except when they play a stabalizing role is correct only if there is a unique equilibrium state with random fluctuations around it. If there are multiple possible equilibriums, speculators can profit by destabalizing the market.
Hedge funds and banks became “real” commodity hedgers when they recently bought commodity operations such as grain elevators and oil storage facilities. They made these purchases to get around the legal limits on the number of commodity contracts speculators are allowed to hold.
The legal limits on speculator positions were in place to protect the markets from the influence of speculators; who’s role is to provide market liquidity.
The spike in grain prices a few years ago was speculator driven. The hedge funds became aware that many grain elevators are required by State law to remain 100% hedged. Speculators ran the price of wheat and other grains to record high prices; the farmers and elevators operators were bled to death by ever increasing margin calls until the cost of interest became too expensive for the real hedgers to hold their positions. A classic short squeeze manipulated by speculators.
Once the speculators bridged the legal limits on commodity contracts the volatility in the futures markets exploded. The narrative of peak oil and food is the smoke screen for these market abuses.
For speculators to make sure money (bordering syllogism)
Necessary conditions
Federate the real assets holders (better profits publication for equities,hold the assets for commodities)
Conditions sufficient
Use of derivatives on the top of the assets prices.
Create volatility (volatility is the daily income of the traders)
Set up prices convexity,reflexivity
Drag the retail investors.
Make good use of the charts.
Make good use of your in house chartist publications.
Whisper to the press.
Exit strategy wait for the accumulation point (crowding out)
Commodities are not “assets”; they are inputs to production.
Speculators use synthetic instruments to “create volatility” (and increase tail risk, ultimately) in order to make money.
The argument that HFT and the like increases liquidity and efficiency of price discovery is horse dung.
Now producers and intermediaries are forced to engage in non-productive commodities speculation to hedge increasingly volatile commodities prices or lock in supplies at affordable prices for their businesses, which exacerbates the forward effects on price volatility further into the future.
But nothing will be done until the HFTers, specs, et al., crash the system.
This was not a Flash Crash unlike what we had in Equities.
Where does Credit Suisse get the following information
“Credit Suisse analysts said the high frequency and algorithmic trading accounted for about half of all the volume in the oil markets.”
They really have no way of knowing this…Not even a decent guess.
As someone who trades in these markets,I believe that the media is exaggerating the effects of High Freq Trading (HFT, Algo…). While there are a lot of speculators in the energy markets there are only a handful of true algorithmic traders in this space, the rest are not so high tech.
Something like 80% of options on crude don’t even trade electronically…