Geoff Vanden Heuvel
Director of Regulatory and Economic Affairs
USDA requires all U.S. dairy manufacturing plants who produce more than 1 million pounds per year of butter, nonfat dry milk, block cheddar cheese, barrel cheddar cheese or dry whey to report their sales quantity, moisture, and prices to USDA every week. Currently reporting are 16 entities reporting 40-pound cheddar blocks, 11 entities reporting 500-pound cheddar barrels, 19 entities reporting butter, 26 entities reporting nonfat dry milk, and 14 entities reporting dry whey. Every Wednesday USDA releases the prior week’s prices and sales volumes. This is the National Dairy Product Sales Report (NDPSR) and you can see the latest one here. The Federal Milk Marketing Order (FMMO) system then uses the months’ worth of NDPSR prices to establish the Class III and Class IV milk price. The FMMO Class III and IV prices are announced by the fifth of the month for the previous month.
I have always heard there was a correlation between Chicago Mercantile Exchange (CME) cash prices for dairy commodities and prices used by USDA’s NDPSR to set FMMO Class III and Class IV prices. This week I was able to get access to some actual correlation data charts that show those relationships. These charts were produced by Bozic, LLC and I thank them for the permission to use them here.
Butter shows the highest correlation between a simple average of a week’s worth of daily CME butter prices and the weekly NDPSR butter price lagged one week.
These charts represent about 10 years’ worth of price data – more than 500 data points. The vertical axis is the weekly NDPSR price and the horizontal axis is the average of a week’s worth of daily CME prices. In this chart, the NDPSR is compared to the previous week’s average of the CME price. You can see the correlation for butter is very tight. Nonfat dry milk correlation is slightly less than butter. Running the numbers, the closest correlation occurs when the NDPSR is compared to the average of the previous three trailing weeks instead of just the previous week. Even then, the correlation is .9359 which is pretty high, but you can see the times that there are outliers.
For block cheese the correlation is pretty high when you use a one-to-two-week lag in the comparison. So, a week’s block cheese NDPSR price correlates best to the average weekly CME block price from 1-2 weeks prior.
For Barrel cheese the best correlation fit is for a one-week lag between the weekly average of the daily CME barrel prices and the following week’s NDPSR barrel price, but here too you see outliers.
For dry whey there is a correlation, but it is the weakest of the five dairy commodities that USDA uses to establish FMMO Class III and IV prices. The closest correlation is to compare the weekly NDPSR with the previous months trailing weekly average. On the other hand, the price range for dry whey
between the lowest recorded price of about 26 cents a pound and the highest price of about 67 cents is much less than the historical price ranges for the other commodities.
So why does this matter? What the data tells us is that CME prices are a rather good predicter of Class III and IV prices in future periods, but they are not perfect, and they are not consistent for all the different commodities.