“When we learn about hedging in school, we learn that as the futures market moves, the cash market should move with it, and ideally in a 1:1 relationship,” says Brett Crosby. He's a Wyoming rancher who's also founder and president of Custom Ag Solutions, which designs and manages agricultural research projects.
“Back home, you know that ideal world doesn't exist. The relationship between futures and cash prices, and the degree of that relationship, varies depending on a variety of factors, including cattle weight, cattle price, cost of gain and carcass value,” Crosby says.
That's where basis — the difference between cash and futures prices at any given time, in a particular location — comes into play.
By way of basic review, basis is equal to a local cash cattle price, minus the near-term futures price. If you kept your peepers propped open during algebra class, you know that also means that basis added to a futures price is equal to a predicted cash price.
“Research has shown that the best price forecast often is to use the futures market adjusted for expected basis,” explains Kevin Dhuyvetter, a Kansas State University (KSU) agricultural economist. “That means if I want a price forecast for the specific weight of my cattle and in my specific geographic location I need a corresponding basis forecast.”
That's been the weakness of historic basis information commonly used in the industry. The average historical basis — usually for 3-5 years — is typically based on weight ranges of 50-100 lbs., and usually broad geography.
Plus, Dhuyvetter points out, “All of a sudden you have a year like 2006. By last October, corn was trading at over $3, roughly $1 more than any of the previous 3-5 years. So a basis forecast based on the historical average then included a year that didn't remotely reflect the year of interest.”
Forecasting price close to home
As a producer seeking to more accurately predict feeder-cattle prices for budgeting and marketing decisions, Crosby was frustrated with these limitations. He and Dhuyvetter saw the value in developing basis forecasts with more weight, geographic and price specificity. Working together and with KSU peers James Mintert and Terry Kastens, and partnering with USDA's Risk Management Agency (RMA), the answer became beefbasis.com.
Virginia Guzman, RMA Chief of Non-Insurance Programs, explains RMA funded development of beefbasis.com to improve risk management tools available to cow-calf producers.
It's a free, Web-based, risk-management tool that allows producers to forecast basis for specific cattle weights, lot sizes and classes, within a narrow in-state geography. Information is currently available for 15 states; 10 more are coming.
Beefbasis.com is built upon a minimum of 10 years of auction-transaction data from the Agricultural Marketing Service for each of the states. The database currently has more than 10 million transaction records — including cattle price, sex, lot size and location — and grows each month. The number of transaction records for each state ranges from 50,000 to more than 800,000. Technically, Dhuyvetter explains data for two of the states were modeled on other states in the same region, adding more accuracy, because fewer than 10 years of actual data existed.
“We know many cow-calf producers don't use futures or options contracts, but they still have to make decisions about marketing and retained ownership. They still have to calculate prices to use in their annual budgets,” explains Mintert, KSU livestock marketing economist. “If you can more accurately project prices for a budget or more accurately assess the value of your cattle for the current market, you can make better decisions.”
Suppose you've got some heavy five-weight steers to market in northeast Kansas. Specifically, these steers are Medium and Large #1 steers that average 584 lbs.
You can call an order buyer to see what's being offered, but that doesn't tell you what their value is relative to the market. You can look at last week's report from the closest auction, but that doesn't tell you about today.
You could even try to calculate an estimated price adjusted for historical basis. In your location, you might look up state data and use the October feeder-cattle futures contract. As of Oct. 8, the nearby feeder-cattle futures price was $113.95, and the October Chicago Mercantile Exchange (CME) steer calf basis for 2004-2006 for Dodge City (for 600-lb. steers) averaged about $5 to just over $8 across the month. So, you arrive at a forecast cash price of $118.95 to $121.95.
Now, try the same thing with beefbasis.com. Based on the Feeder Cattle model (FC) at the site (more later), estimated basis is $11.10 for an expected cash price of $125.05. For the Live Cattle and Corn-based model (LCC — also more later), the estimated basis is $8.93, for an expected cash price of $122.88. That's based on the specific weight, muscle and frame classifications of the steers, and lot size sold at the closest market included in the data set, which is at Salina, much closer than Dodge City.
In this scenario, historical basis undervalues the calves, relative to the beefbasis.com models. As to which method provides more accuracy, Dhuyvetter explains, “Remember that both the historical average and the beefbasis.com models are forecasts, so neither are perfect. However, based on extensive, statistical out-of-sample testing, beefbasis.com models outperformed historical averages.”
“The lighter the calves and the further away from the futures contract, the more important these forecasts become,” says Kole Swanser, a partner in Custom Ag Solutions. He points out users receive a confidence level for each basis/price forecast. A hedge analysis can also be had at the touch of a key.
“We want feedback from users so we can continually improve it,” Swanser says. He points out the tool was developed with input from cow-calf producers.
Like other forecasting tools, though, this isn't some sort of crystal ball. Like historic basis, this has limitations, too.
For one thing, with 10 years of data, only three years reflect historically high corn prices. That's one reason beefbasis.com enables producers to use two different models to calculate basis.
According to Mintert, the FC model calculates feeder basis as a function of feeder-cattle futures. The LCC model calculates it as a function of live-cattle futures and corn futures. “In today's environment, we wanted to include a model that allows us to directly capture the effect of the corn market,” Mintert says.
Moreover, video-auction data is currently not used within the models. These folks know how important video auctions are and are working to incorporate them. There are no easy solutions, though.
Aside from access to proprietary video auction market data, and the relatively slim number of years from which to draw, basis is calculated using futures prices at delivery date. Since video contracts usually include a fairly broad marketing window, there's no telling the specific delivery date. Keep in mind these models are based on actual market prices, as opposed to prices that buyers could have paid.
“As a producer, I want to see how prices actually behaved in the past, not how a breakeven budgeting approach says prices should have behaved,” Dhuyvetter says. “This model explains what actually happened in the past, so I would argue it's a better model for predicting how things might be in the future.” He emphasizes beefbasis.com forecasts allow producers to consider specific cattle weights rather than ranges.
Plus, producers can query the price database at beefbasis.com to compare average weights and prices between locations or weights and prices across years at a given location.
Forecasts are just that
Keep in mind the resulting basis forecasts are not some absolute value. Like Expected Progeny Differences used to estimate the genetic merit of cattle, these estimates represent a statistical range. For instance, in the example cited earlier, the basis in the FC model is $11.10, ranging from $6.71 to $15.48. For the LCC model it is $8.93, raging from $4.61 to $13.24.
Such broad ranges are tough for some to accept, compared to the single number offered by an historic basis. Imagine, though, how much range is associated with a historic-basis price forecast, which fails to account for lot size, specific weight, specific class and more localized geography.
Crosby emphasizes that the tool allows him to more accurately forecast prices than previously possible.
“I used the basis tool a lot this past summer to make decisions about whether to contract my calves and when. Before all of this came along, the best I could do was use a complicated cost-based model where I'd back out the cost of gain past 600 lbs., and a different cost of gain for the weight up to 600 lbs,” Crosby says. “At the very least, I want to be able to forecast the position I'll be in given different market scenarios. This tool allows me to do that.”