Pork producers' ability to deliver a proper and consistent supply of hogs to meet packer-processor needs remains an ongoing discussion.
The discussion ranges from the need to meet the 1.5% steady annual growth in consumer demand, to the accuracy of USDA Hogs & Pigs reports and industry analysts' ability to predict future supply.
As we consider ways to measure and manage supply in the pork chain, the concept of “process capability” emerges. Collectively, the industry needs a better understanding of process capability and how to assess it. The overriding question is: Can producers do a better job of matching the market hog supply to their customer's needs — the packers and processors?
The following manufacturing example will help illustrate the concept of process capability:
If we are in the business of supplying raw materials to North American Processing, Inc. (NAP), NAP would tell us how many units to deliver each week, plus set maximum and minimum specifications for the units supplied.
For example, NAP's purchasing department might tell us it needs 1.87 million units weekly, but it would accept up to 1.95 million units. Their absolute minimum is set at 1.65 million units.
Further, NAP might specify that each unit must have certain quality characteristics, such as a target weight of 185 lb., again with upper and lower limits for weight. NAP's purchasing agent might set other quality-based standards for our product. For example, he could target a specific characteristic we'll call “BF,” again setting upper and lower specification limits.
As a long-time supplier of raw materials to NAP, we have addressed several quality-related issues with our products over the years. When NAP became concerned about our BF measurements, we responded by lowering BF measurement to meet their specifications.
As in almost every vendor-customer relationship, there have been rocky moments and threats of using competitive products from overseas, but overall, we have a good relationship.
Last week, a NAP purchasing agent called with concerns about the number of units we deliver each week. They sent charts and a few calculations (see Chart 1 and 2).
Deliveries per Week
There are 375 weeks of deliveries represented in the charts, beginning with the first week of 1997 and continuing through the last week of March 2004. The note accompanying the charts indicated that the mean number of units we delivered each week is 1.873 million units, with standard deviations of 147,151 units above or below that level. They went on to point out that, based on current specifications, the process capability index (Cp) for the number of units delivered per week was 0.34.
We determined that during the 375-week period, we delivered a number outside of their specifications 130 times. In other words, we missed their specifications for the number delivered 34.7% of the time.
As painful as it is to face, our system is not capable of delivering to our customer's specifications. We must determine how we become capable, or face the inevitable.
The Real World
It is apparent that the fictitious company supplying “units” to NAP represents U.S. market hog production collectively supplying our customers in the processing side.
While the story may be fictional, the facts are all too real. The chart represents the actual federally inspected hog slaughter since the beginning of 1997. The target and specifications are based on real market signals.
To help establish specification limits, we studied the “below cost” price paid for our animals when supply exceeded 1.95 million head. Conversely, delivery of less than 1.65 million head reflects an inadequate hog supply, which may result in reducing shifts, or possibly plant closings.
Unfortunately, there is no written specification for the collective number we need to deliver each week. However, the marketplace clearly signals that the specifications used in our example are close to correct.
If Chart 1 represented an input you or I needed to manufacture a product, how long would it take any of us to find a new supplier that could deliver the right number we requested every week?
Consider the effect on market price received when deliveries are above or below the customer's specifications. For this analysis we used the same dataset for federally inspected slaughter (FIS). We removed the holiday weeks with less than five full days of kill.
We used the weekly weighted average of the Iowa-Southern Minnesota direct market price as reported by USDA. All prices are reported as dollars per hundredweight of carcass ($/cwt., carcass). Again, the dataset covers the period from the beginning of 1997 and runs through March of 2004.
Chart 2 provides the federally inspected slaughter and the average weekly carcass price. The correlation coefficient of deliveries and price is -0.7544. This means that increases in deliveries are well correlated with decreases in price.
The chart plots all of the deliveries and the corresponding average price for the week, and includes a linear regression trend line. The average penalty our customer imposes for deliveries above the upper specification limit (USL) is $21.59/head. Delivery outside of specification 34.7% of the time is never acceptable and the penalties are severe.
Where should producers focus first to provide consistent deliveries within specification? What should producers do to reduce variation and improve customer satisfaction? To solve these problems, we must use a multidisciplinary approach.
More Global Information
Livestock producers need more accurate information about inventories. Since 1960, the broiler industry has used a weekly USDA report on egg “sets” to predict the number of birds that will be delivered at a specific time in the future.
One important reason the chicken industry appears able to adhere more closely to the demand trends is because they have better global metrics to predict future supply. One could make a strong case that a broadly available global metric to predict future supplies would be in the best interest of livestock producers.
The long-term pork cycle is but one reason for the lack of process capability of livestock producers. Seasonal variation is the second important reason.
Chart 3 (p. 18) represents a weekly measure of pigs weaned/mated female/ year (P/MF/Y) in 2003. The data set includes approximately 350 farms and 550,000 animals. The mean P/MF/Y is 20.01 with a standard deviation of 0.48. We have added a moving average trend line to this data to more closely reflect the real impact of seasonal variation.
Industry research has focused on the “seasonal infertility” of late summer and early fall. But another, equally important variable is “seasonal hyper-fertility,” and it is one of the major reasons live animal production struggles to meet the quantity specifications of the marketplace. This hyper-fertility period is responsible for the majority of deliveries above the upper specification limit set by our customers. The average penalty is $21.59/animal delivered.
Clearly, seasonal hyper-fertility creates serious economic damage. This begs the question — are we spending our research dollars and implementing management interventions correctly?
All variation needs to be addressed and managed if live animal production is ever going to become process-capable.
The global metric and management focus on reducing variation will result in improvement in the process capability index for live animal production. This process is absolutely necessary if we are to have a healthy, viable production industry.
If we choose to ignore the problem, our domestic market will be taken over by foreign competitors who have made process capability a priority. No amount of protectionist legislation or duties will ultimately change that fact.