Blog

Spring is at the door, and before we know it, work in the field will begin. We will see below that we can expect lower fertilizer prices next summer. This is good news to farmers who are facing tight margins this crop year from lower commodity prices and input prices declining at a slower rate.

In this post, I also look at the past performance of my forecasts.

Fertilizer prices

Figure 1 shows data from Statistics and Data Development Section, Intergovernmental and Trade Relations Branch, Alberta Agriculture and Irrigation and USDA NASS. The Alberta fertilizer products are Anhydrous ammonia (82-0-0) full service with applicator, Urea (46-0-0) and Fertilizer (11-51-0). The US fertilizer price is an index (scale x 10) for the whole United States.

Prices for the three types of fertilizers reported by Alberta Agriculture and Irrigation are lower than a year ago. The price for Anhydrous ammonia (82-0-0) has been significantly trending down since peaking in early 2022. For Urea (46-0-0) and Fertilizer (11-51-0), prices have generally trended down but a slower pace. The US price index has been more stable but has nonetheless lost 10% over the last year.

Drivers of fertilizer prices

Besides fertilizer prices in the previous month, the models find that the three main drivers of fertilizer prices are barge rates on the Mississippi River, the Chicago Mercantile Exchange (CME) corn price and the Henry Hub natural gas price.

I use the downbound grain barge rates from St. Louis as a proxy for up-bound barge rates on the Mississippi River. Figure 2 shows that barge rates peaked in September from low water levels on the Mississippi. The CME corn price has sharply declined over the last year. It even dropped below USD 4/bu at the end of February. The price for natural gas peaked in January but has dropped 50% since.

Figure 2 also shows forecasts for barge rates, corn price and natural gas prices that I will use below to forecast fertilizer prices. According to these forecasts, prices for the main drivers of fertilizer prices will stay low. As fertilizer prices are positively correlated with these drivers, we can expect lower fertilizer prices for the first half of 2024 than for the same period in 2023.

Forecasts

My statistical models summarize information from the data into forecasts for fertilizer prices. The forecasts require scenarios for the models’ predictor variables. The forecasts use the following scenario:

• Barge rates and rail rates are forecasted using time-series models,

• The exchange rate, oil prices, corn price and natural gas price follow their Futures Forward Curves (FFC).

As Figure 2 shows, the FFC for corn currently slopes up, but corn prices are nonetheless expected to stay low. Likewise, the FFC for natural gas slopes up but prices are expected to stay low and to follow their typical seasonal pattern. Barge rates should also follow their typical seasonal pattern.

Figure 3 shows the forecasts. The present economic conditions support lower prices for the three types of fertilizers. Unless there is a significant market shock, we can expect fertilizer prices to continue trending down.

A look back at previous forecasts

This is the fifth blog post in which I forecast fertilizer prices. I thought it would be a good idea to look back at how my previous forecasts to verify to check how close they were to reality.

Any forecast is conditional on assumptions made by the analyst. This means that someone should not judge a model too harshly when looking back at previous forecasts because inputs into the model by the analyst matter just as much. I’m not going to look back at the assumptions I made about predictor variables and whether they correspond to what actually happened. When examining my previous forecasts, I’ll take the position that the model does well when it captures well the general trends taken by fertilizer prices. The idea is to check whether the models catch trends in fertilizer prices well even when values for the predictor variables are not necessarily right.

Figure 4 shows the data.

• November 2022: This was my first forecast and I was hesitant to publish it given it showed sharp increases in already elevated fertilizer prices. At that time, barge rates on the Mississippi were at a record high. The forecasts were totally off. Since then, barge rates have declined and the model has recalibrated to be less sensitive to large swings in barge rates.
• February 2023: For Anhydrous Ammonia (82-0-0), the model forecasted the price to sharply increase because high barge rates months before, but then to drop. In actuality, the price increased slightly and then declined. In the blog post, I had commented that I expected the price to increase but only modestly. In the case of Fertilizer (11-51-0) and Urea (46-0-0), the models were correct in forecasting lower prices.
• June 2023: Other than a bump in the price of Fertilizer (11-51-0) in the summer of 2023, the model correctly predicted trends.
• October 2023: The models forecasted lower prices, which was not accurate in the case of Fertilizer (11-51-0). Prices for Anhydrous Ammonia (82-0-0) and Urea (46-0-0) increased and then declined toward the values predicted by the models. I do not know exactly what shock caused fertilizer prices to increase momentarily in late summer and early Fall. It is possibly a combination of higher barge rates and higher natural gas prices.

Overall, other than the terrible forecasts in the first iteration of the models, I’m happy with their performances. I’ll continue to seek improvements but I expect them to be relatively minor going forward.