Kub's Den

dtn-img

By Elaine Kub
DTN Contributing Analyst

Hat tip for this week's column topic to DTN Senior Analyst Darin Newsom and his army of Twitter followers, especially Chuck Penner from LeftField Commodity Research in Winnipeg (@LeftFieldCR) who put forth a theory that grain markets seem more likely to pull back, or reverse direction, in reaction on the days after "a big move." Chuck is far from alone here. I think we've all, at some point in our lifetimes of watching grain prices, developed some kind of gut instinct that the market will pare back gains the next day after they were made, or bounce back a little the day after exploring fresh lows. We have probably all uttered the words "Turnaround Tuesday" when the market switches its mood the day after starting the week in what seemed like a confident direction.

Is this gut instinct actually supported by the data? If we see a big move in one direction on one day, does that actually help us predict what direction the grain markets will move on the next day -- up, down, or sideways?

First, let's think about a roulette wheel. The dealer spins it, and the ball lands on either a red number, a black number, or a green space. There are 18 red numbers, 18 black numbers, and 2 green spaces, so 38 possibilities overall. Thus, there is a little over 47% chance the ball will land on red, a little over 47% chance the ball will land on black, and a little over 5% chance the ball will land on green. If you just bet $1 on "black," over and over for a thousand spins, at the end you would expect to still have $473 (i.e. to have lost $527).

Well, the corn market's result on any given day can either be up, down, or sideways. I did the math, and over the past 10 years (2,521 trading sessions), there was a little over 49% chance that any given day would end with a higher close, a little over 48% chance that any given day would end with a lower close, and about a 2% chance that any given day would end up exactly where it closed the day before.

However, we believers in behavioral economics might be tempted to say the corn market is not a roulette table. Its daily result isn't determined by some random spin of a wheel. Rather, its price behavior from one day to the next depends on a lot of complex market indications and traders' sentiment.

In statistics we call roulette a "Markovian" process, which means it is truly random and memory-less. The probability of the next spin's result is always the same, no matter what happened on the previous spin or on the last 10 previous spins. Even if there was a rare series of 15 red numbers landed in a row, on the sixteenth spin, the probability of another red number being hit would still be just a little over 47%. But it's hard to say whether or not the corn market's day-to-day price behavior is Markovian or not, until we look at the data.

If our Turnaround Tuesday instincts are correct, then the grain markets do have a memory, and they do care what happened the day before. Maybe they really are more likely to switch direction after a "big" day.

To test this idea, I had to pick some arbitrary definition for "big" and I settled on any move of 3% or more. Daily moves of even 8 cents felt "big" in 2006, but in the wild market of 2008, when prices were higher, moves had to be more like 20 cents to feel "big." Today, my 3% rule selects mostly double-digit daily moves.

Over the past 10 years, there have been 307 "big days" in the corn market. They occur less frequently now -- only 4% of trading sessions in 2015 and 2016 -- than they have over the course of the whole decade -- 12% of the past 2,521 days. Here's how the market performance turned out on the days after those big days:

No change in price on day after a big upward move: 1 session

No change in price on day after a big downward move: 4 sessions

Closed higher on day after a big downward move: 74 sessions

Closed lower on day after a big upward move: 75 sessions

Total turnarounds after big moves: 149 (48% of the time)

Closed higher again on day after a big upward move: 82 sessions

Closed lower again on day after a big downward move: 74 sessions

Total instances of continued direction after big moves: 156 (51% of the time)

Well, that's looking more and more like a roulette wheel, now isn't it? I realize that just looking at historical daily changes doesn't fully capture the psychological adventure of watching the corn market sail 7 cents upward in the middle of a session, past the $4 mark for the first time in nine months, then ultimately close more than 10 cents lower, as it did on April 21. But examining these dry probabilities after the fact with cool heads can prevent us from getting too tempted to start a 'Corn Croupier Hedge Fund' that invests in one-day, up-or-down futures positions whenever triggered by a 3% move in prices.

It actually looks like the corn market is less likely to react to "big days" with a turnaround in direction, and more likely to follow "big days" with a small continuation of the original direction. The average move on the continuation days was 6.9 cents. The average size of the turnarounds was 9.4 cents, but remember, those were less likely to occur, and it's virtually impossible to predict which course the market will choose to pursue. Even after "big days," behavioral economics aside, we are stuck with the same possibilities as before: up, down, or sideways.

And that, of course, brings us to this week. The days when the USDA releases its World Agricultural Supply and Demand Estimates (the monthly WASDE reports) frequently tend to be "big days." Tuesday of this week was no exception -- a 10 3/4-cent upward move in corn prices, in harmony with a 57 1/4-cent upward move (!!!) in soybean prices. Even after all this data exploration I have done, writing now on Tuesday afternoon, I still have no way to statistically predict with any confidence whether the market will follow that performance with a higher continuation or a lower turnaround at Wednesday's close.

One final note. I looked to see if directional switches were especially likely on Tuesdays, you know, like "Turnaround Tuesdays." And to the surprise of absolutely no one who has made it this far through this column, they were not. On any given day, the corn market has a roughly 50% probability it will switch directions from the previous day's close. On Tuesdays, lovely Tuesdays, it's actually notably LESS likely to switch direction: only a 45% probability. Maybe we should start calling them Tenacity Tuesdays.

Elaine Kub is the author of "Mastering the Grain Markets: How Profits Are Really Made" and can be reached at elaine@masteringthegrainmarkets.com or on Twitter @elainekub.

(CZ/BAS)