USDA Predicts a 41% Increase in Egg Prices This Year

Yikes!

This week, the Economic Research Service (ERS) of the US Department of Agriculture released its February consumer price forecasts. They predicted that food would be between 1.1 and 5.9 percent more expensive in 2025 than 2024.

But the headline grabber in the report was something else — it predicted a 41% increase in the price of eggs. Eggs doubled in price during 2024, mostly because of avian flu. Are eggs really going to keep going up in 2025?

Yes, probably. But before rushing to the store to buy all the eggs, there a couple of things to understand.

First, the actual ERS forecast is that there’s a 95% chance that egg prices will be between 15 and 75 percent more expensive in 2025 than 2024, on average throughout the year. A 41% increase is the most likely value in that range, but the price will almost certainly change by some amount other than 41%. The real information in the forecast is that there’s a wide range of possibilities.

Second, the prediction refers to the annual average price. The average price of a dozen eggs in 2024 was $3.17. The ERS forecast is that 2025 egg prices will average between $3.65 and $5.55. By the end of January, the price was $4.95, so we’re already 56% higher than last year’s average.

How good is ERS at forecasting food prices?

“Prediction is hard, especially about the future” is a truism attributed to luminaries such as Niels Bohr, Yogi Berra, and Mark Twain (though it’s probably just an old Danish proverb). Until recently, ERS price outlook forecasts were not performing well, with only 16% of forecast ranges including the actual price changes. An ideal forecast would have the actual value inside the forecast range 95% of the time.

In 2022, ERS economists Matthew MacLachlan, Carolyn Chelius, and Gianna Short launched an ambitious project to modernize and improve the food price outlook forecasts. They replaced the outdated legacy approach with a modern time series method. Their new method dramatically improved forecast accuracy, as indicated in their figure below: the orange legacy forecast range is often far from the actual value, whereas the blue time series forecast range often contains the actual value.

The new time series method generates a forecast range from a model based on historical patterns. The model knows how long prices tend to stay up after they increase, and it knows the extent to which months with large price increases tend to follow months with large price increases.

Importantly, the model uses no information other than the history of prices. Just as ChatGPT predicts the next words in a sentence based on past words, the time series method predicts the next egg prices based on past prices. The forecast is not based on the number of layers that have been culled during the avian flu epidemic or estimates of how quickly the epidemic will end.

A dirty little secret of forecasting is that time series methods tend to perform better than methods that use more information or judgment (that link goes to another sterling ERS report). You can learn a lot from careful modeling of past patterns in the data, and the relevant other information is often very uncertain.

The ERS has been an integral part of the USDA for more than a century. In addition to monthly food price outlook forecasts, its staff produce oodles of awesome data products and excellent research publications. The contributions of their staff extend far beyond informing the public about food prices.

The price of eggs will come down when the avian flu epidemic ends. When will that be? The correct answer is that nobody knows, but ERS forecasts will be useful for keeping us informed. I think their $93 million annual budget (28c per American) is well worth it.