Oil price shock

Following the closing of the Strait of Hormuz, oil prices increased by roughly 60%. Accordingly, I consider a 60% increase in the West Texas Intermediate (WTI) oil price and how it ripples through the Canadian economy.² The model assumes that the oil price affects Canada’s GDP, the exchange rate, gasoline prices, and prices for goods, including food. I will not show results for all variables.

Figure 1 shows responses to a 60% increase in WTI lasting either one month or six months. The shock does not persist and the WTI returns to its previous level rapidly after a shock. The Canadian dollar becomes stronger in response to higher oil prices, rising by about 3% for the one-month shock and up to almost 4% for the six-month shock. In both cases, the exchange rate returns rapidly to its previous level after the end of the shock. The gasoline price follows the same pattern as the oil price, rising by 20% for the one-month shock and up to 30% for the six-month shock. The gasoline price declines quickly when the oil shock ends. In practice, I doubt that will be the case, as gasoline prices tend to rise fast but to fall slowly (rocket and feather pattern).

A higher oil price affects prices for all goods. The price index for All-items increases by about 1% for the one-month shock and by as much as 1.5% for the six months shock. This says that we can expect the oil shock to lead to an inflation rate about 1 to 1.5% higher because of the oil shock. This is consistent with Brandon’s results and those reported by the Scotiabank.

Figure 1 shows that the Food price index decline in response to the oil shock. The impact is small: -0.2% for the one-month shock and up to -0.6% for the six-month shock. It is the change in the exchange rate that drives the decline in food prices. Price indices for Fresh fruit and vegetables and Fresh and frozen meat (excluding poultry) decline more than for food in general. For Fresh or frozen poultry, the price index moves up and down around zero. I do not show confidence intervals in Figure 1, but those for the food price indices all include zero. This tells us that the impacts of the oil shock on food prices are not statistically different from zero, however.

This is not too surprising. The cost of producing food domestically increases in response to higher oil prices, but the impact is small. The cost of production for imported food is also higher, but it becomes cheaper to import because of the appreciation in the exchange rate. The net effect appears to be a small decline in food prices, which is not statistically significant.

One impact of the oil shock is to lower the competitiveness of domestically produced food. In addition to an increase in production costs, Canadian food producers must compete with cheaper imported food. We can expect that the profitability of Canadian farms to decline in 2026.

Fertilizer price shock

The Middle East produces about 25% of global urea exports. No surprise then that global nitrogen prices shoot up with the closing of the Strait of Hormuz.³ The Urea FOB US Gulf futures have jumped by about 50% since the start of the conflict.

I trace in this section the impacts of a 50% increase in urea prices on food prices in Canada. The model is similar to the one for the oil shock but adapted to the specificity of the fertilizer market.

Figure 2 shows the impact of a 50% increase in the price of urea in Canada. Again, I consider the impact of one and six months shocks. For the one-month shock, the price of urea increases by 50% in the first month and then declines as rapidly the following month. For the six-month shock, the price of urea increases by 50% and then returns to its previous level rapidly after the end of the shock.

The figure shows movements in the price of corn and soybeans in Canada. But, these changes are not statistically significant (confidence intervals not shown). Nitrogen fertilizers like urea are important for corn production, but the timing and duration of the shock matter. A shock at the end of February comes at a time when planting decisions are already made and most inputs are already purchased.⁴ Thus, we do not expect that a shock on the price of urea to impact prices for corn or soybeans for 2026. Furthermore, note that changes to corn and soybean prices in response to the urea price shock are not statistically significant.

Figure 2 shows small changes for the price indices for All-items and Food. They are not statistically significant. For Fresh fruit and vegetables, Fresh or frozen meat (excluding poultry) and Fresh or frozen poultry, the changes are larger, but, again, they are not statistically significant.

How can a positive price shock to an input as important as fertilizer not cause crop prices and food prices to increase? There are several reasons. First, as previously mentioned, timing matters a lot. The shock from the Iran conflict came when crop planting decisions were made and inputs mostly already purchased. I expect, in response to the fertilizer shock, small changes in plantings for this crop year, and consequently this means marginal impacts on crop prices. Second, the duration of the shock also matters. A permanent shock on fertilizer prices would affect planting decisions in the following crop years and would then affect crop prices. Third, as the USDA data show, farm costs account for a small share of the value of food to consumers. Therefore, even a large shock on fertilizer prices will result into a small shock to prices paid by consumers for most food products. It is for products with less value-added after the farm, like fresh fruits and vegetables, that we will observe the largest shock. Even for those, we expect a certain rigidity in retail prices, meaning that they adjust slowly.

Conclusion

Results from my model do not align with what some media have reported about the consequences of rising oil and fertilizer prices. Reality is sometimes boring, and there is always a bias in reporting real consequences. However, if the oil and fertilizer shocks last, I expect consumer food prices to eventually rise. Hopefully, the situation in the Middle East will normalize soon, for the best of everyone.

There will be consequences to higher oil and fertilizer prices upstream retail. I did not look into those but they can be important. In particular, it might be tough for fresh fruit and vegetable farms. Their production costs will increase and they will face tougher competition from imports because of a stronger Canadian dollar. Likewise for field crop farms, in particular those growing crops that require nitrogen fertilizers (e.g., corn). Let’s hope that input prices decline soon.

Fertilizer prices

Figure 1 shows fertilizer price 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.

The price of Anhydrous ammonia (82-0-0) has generally trended down since the peak in 2022, but the price has slightly increased since bottoming down in summer 2024. The price for Urea (46-0-0) has also declined but it has been more volatile than Anhydrous ammonia. The price for Fertilizer (11-51-0) has been more stable than the other two products, but it has also declined since the peak in 2022. The US fertilizer price index has also generally trended down since peaking in 2022, but has been more stable than the Alberta fertilizer prices.

Figure 1. Alberta fertilizer prices and US fertilizer price index

Forecasts for Alberta fertilizer prices

Figure 2 shows forecasts for fertilizer prices based on Alberta input price data. The economic conditions currently point toward relatively stable fertilizer prices. Crop prices are low, slowing down demand for fertilizers, thus keeping their prices low. The models forecast some seasonality in fertilizer prices but no large price swings.

Figure 2. Forecasts of Alberta fertilizer prices

Forecasts for Statistics Canada fertilizer prices indices

Fertilizer price indices are from the Statistics Canada table 18-10-0266. The indices are for the entire Canada.

Comparing Figures 2 and 3, the indices have moved in sync with Alberta fertilizer prices, although the indices are less volatile. Accordingly, forecasts for the price indices are for more stable prices. The models find that under the current conditions, fertilizer prices should stay relatively stable over the next six months.

Figure 3. Forecasts of Statistics Canada Fertilizer price indices

Conclusion

The economic data do not point toward upward pressures on fertilizer prices. I do not expect the trade wars could cause an increase in fertilizer prices in Canada. In the trade war between Canada and the United States, fertilizers have been mostly exempt, although a 10% US tariff has applied on Canadian potash. That tariff should have a negative impact on Canadian potash prices. I do not expect the Canadian government will put tariffs on fertilizer imports from the United States. If the trade war between the United States and China continues, it will cause lower crop prices in the United States, which will dampen the demand for fertilizers and keep fertilizer prices low.

Model setup

The model is meant to capture the most important realities of supply management in Canada. It is not calibrated to a sector in particular. The import price is lower than the domestic price. This is generally consistent with what we observe for supply management sectors but not always. Lately, with the avian influenza outbreaks in the United States, prices for eggs in Canada are lower than in the United States. Some dairy products may also be cheaper in Canada than in the United States.

A typical economic analysis of production quotas uses supply and demand curves. I take this approach but I specifically model production decisions by individual farms to obtain the market supply. Farms differ in their cost structures, explaining differences in their size and profitability. Assuming non-identical farms allows me to examine how policies impact the number of farms, their sizes, and some issues regarding efficiency.

The model assumes the following:

• Farms produce an identical product.
• There are 1,000 potential farms who differ in their production costs. A farm produces if it can generate a positive profit.
  
• The production quota is 1,500 units.
• Market access (i.e., import quota) is 10% of domestic production. The import quota is filled because the import price is lower than the domestic price. Thus, the total quantity supplied is 1,650 units (1,500x1.10).
• The price paid to farms is fixed and equals $10/unit. The reason for fixing the price is that prices paid to farms in supply-managed sectors are based on cost-plus formulas. The price level is such that it may leave some of the quota rent to processors.

Baseline scenarios

The figure below illustrates the baseline scenarios. The figure shows two supply curves. The one labelled Supply is the normal supply curve. I will use the Supply, inefficient allocation when discussing supply management. I explain how to interpret that curve below.

Table 1 complements Figure 1 by showing the outcomes of the baseline scenarios. The total value of the quota corresponds to the rent that goes to farms from selling at a price above their marginal cost. The producer surplus is an economic measure of producer welfare. It is calculated as the area above the supply curve but below the price, and includes the value of the quota.

Autarky

Autarky is the scenario where there is no trade (imports and exports). This is a situation that Canada probably never experienced, and will likely never experience. The market equilibrium in autarky in Figure 1 is where the supply and the demand cross at point a. The price is $9.44/unit, domestic production is 1,875 units, there are 717 farms and the producer surplus is $2,352.

Free trade

In the free trade scenario, there is no barrier to trade and foreign goods enter Canada freely. The scenario assumes an import price of $8/unit. The price is constant because Canada is a small country, sufficiently small that it has a marginal impact on world prices for most commodities. In practice, the relevant world price for Canadian imports is often the US price. The total quantity demanded is at the intersection between the world price and the demand at point c, and equals 2,503 units. Domestic production is at the intersection between the world price and the supply at point b, and equals 712 units. Imports correspond to the difference between the quantities at points c and b, and equal 1,792 units. There are 282 farms and the producer surplus is $1,558.

Supply management

I consider two cases for supply management. The first case assumes an efficient quota allocation with farms freely buying or selling their production quotas. As a result, the most efficient farms, who value quota the most, have the production rights.

In the second case, I consider there are frictions that cause inefficiencies in the quota allocation. The ceiling on dairy quota prices in P5 provinces (Ontario, Quebec, New Brunswick, Nova Scotia and Prince Edward Island), and in British Columbia, is an example of a friction that leads to an inefficient allocation. The data on quota transactions show large differences in quota values across provinces and small transaction volumes relative to production in provinces with constrained quota prices. Bruno Larue, Alphonse Singbo and I show in a study published in 2017 that constraints on quota transactions make dairy farms less efficient and herd expansion unprofitable. For poultry and eggs, I expect that quotas are more efficiently allocated because frictions are not as significant as in the dairy sector. In Figure 1, the supply with an inefficient allocation lies above the supply with an efficient allocation.¹

In both cases, the domestic production equals the quota set at 1,500 units and the price paid to farms is $10/unit. Market access equals 10% of domestic production, corresponding to 150 units. Consumption equals 1,650 units.

In the case of an efficient quota allocation, the marginal cost is at point d and equals $9.02/units. The value of the quota is calculated as the surface delimited by the price on the demand ($10/unit), the marginal cost ($9.02/unit), zero, and the quota volume (1,500 units): (10.00 - 9.02)x(1,500 - 0) = $1,470. The producer surplus is $3,118 (inclusive of the quota value) and there are 582 farms.

In the case of an inefficient allocation, the marginal cost is at point e and equals $9.50/unit. The total value of the quota is $750, the producer surplus is $2,644 and there are 699 farms.²

Analysis

I do not expect the free trade to happen in the near future. But, in the long term, the likelihood of that scenario increases. Compared to the two supply management cases, free trade leads to fewer farms, the loss of the quota value and a lower producer surplus. However, consumption is higher, which benefits consumers.

An inefficient quota allocation means a larger number of farms. The rules that cause frictions in quota transactions are a choice that provincial marketing boards have made to maintain a larger number of farms. Many farms choose to keep producing because the value they can obtain for their quota is low relative to the return they obtain from farming. The rules make quotas more affordable but less accessible because the volumes traded are lower, and as a result it slows down farm consolidation. In the case with an inefficient allocation, there are 699 farms, while with an efficient allocation there are 582 farms. The larger number of farms comes at the cost of a lower quota value and a lower producer surplus.

The downside of rules restricting quota transactions is that farms operate at a less efficient scale. In the model, the average production is 2.15 units with an inefficient allocation, while it is 2.58 units with an efficient allocation. The average farm size with an efficient quota allocation is similar to the average farm size in the scenarios for free trade and autarky.

Market access expansion

As mentioned at the outset, political parties in the latest elections have vowed to protect supply management but not to keep it in its current state. Market access concessions are a likely outcome of the negotiations. With this in mind, I investigate with my model a scenario where market access increases from 10% to 20%.

Figure 2 shows the outcome of expanding market access to 20%. At a quantity of 1,800 units, the total quantity given 20% market access and a 1,500 quota, that the price on the demand at point f is about $9.60/unit, below the price paid to farms of $10.00/unit. In practice, this is possible if the increase in market access is sufficiently strong. If such a situation occurs, it would be problematic for the economic viability of processors. Below, I consider two adjustments in response to this situation: A) a cut to the price paid to farms and B) a 10% cut to the quota. These are not optimal responses to increased market access. The values are selected with the purpose of providing an example. An optimal response likely includes adjustments to the price and the quota.

These two adjustments may not be necessary in the case of a small increase in market access. If market access is slowly increased over a few years, these adjustments may not be perceived because they are implemented through the absence of modifications in response to a growing demand. For example, without increased market access, it could be the case that the production quota would be increased by 2% in response to a growing demand. However, in the case with increased market access, the production quota may not be increased in response to demand growth.

Figure 3 shows the two responses to increased market access. In panel A), the price paid to farms is cut from $10/unit to $9.56/unit: the production quota stays at 1,500 units, imports are 300 units and total consumption is 1,800 units. In panel B), the production quota is adjusted down by 10%: the production quota is 1,350 units, imports are 270 units, and consumption is 1,620 units.

The first part of Table 2 shows the solutions for the case illustrated in Figure 2. The next two rows show solutions for market access expansion with no change to the quota but a price cut from $10.00/unit to $9.56/unit as in panel A) of Figure 3. The last two rows show solutions for market access expansion with a 10% quota cut but not change to the price as in panel B) of Figure 3.

Supply management, no change to the quota and the price

The price paid to farms and the quota stay do not change in response to market access expansion in this scenario. This is not a viable situation as mentioned above because of the burden it imposes on processors.

Supply management, no change to the quota, and price cut

This scenario considers that the price paid to farms is cut from $10/unit to $9.56/unit. I set the price at $9.56/unit because it is below the price at point f but above the price at point e in Figure 3. This leaves a rent to processors, improving their economic viability.

With an efficient quota allocation, the number of farms stays the same at 582, but the value of the quota drops from $1,470 to $810 and the producer surplus declines from $3,118 to $2,458.

With an inefficient quota allocation, the number of farms drops from 699 to 650, the value of the quota declines from $750 to $90 and the producer surplus decreases from $2,644 to $1,984.

Supply management, and 10% cut in production quota, and no change to the price

Cutting the production quota by 10% means that, at the total quantity of 1,620 units, the price on the demand is above $10.00/unit, leaving a rent to processors.

With a 10% cut, the total value of the quota and the producer surplus increase whether the quota allocation is efficient or not. This does not have to be the case. If the quota cut is smaller, say 5%, the values for the quota and the producer surplus would have declined, but the rent to the processors would have been negative, making this economically unfeasible. This illustrates that it might be difficult to adjust the quota in response to increased market access and satisfy all the actors involved in the supply chain.

Table 2 shows that with a 10% quota cut, the number of farms declines from 582 to 525 in the case of an efficient allocation, and declines from 699 to 677 in the case of an inefficient allocation.

Analysis

Table 2 shows that the response the industry adopts to increased market access matters a great deal. The values in Table 2 depend on the choices I made (10% quota cut or a price cut from $10.00/unit to $9.56/unit). In practice, this shows that choices made by producer boards will be consequential.

The best course of action in response to increased market access likely involves pulling on several levers, including adjustments to the price and the quota. What the model shows is that it will be difficult, if not impossible, not to lose farms in the process. The best outcome in the model regarding the number of farms is when the quota allocation is efficient, the price is cut and the production quota stays the same.

Concluding thoughts

Canada conceded market access in CETA, CPTPP and CUSMA. My thought at the time Canada signed these agreements was that sectors under supply management knew what to expect for many years ahead and could adapt to increased international competition. Given recent developments, this is obviously not the case.

Supply management is weaker now than when CUSMA was signed in 2019. For that reason, the impacts of further market access concessions will likely be different. I do not know how close Canada is from having pseudo supply-management systems for dairy, poultry and eggs. The tipping point from a meaningful production quota to a meaningless quota might not be very far. That is an empirical question that deserves studying.

My model shows that the number of farms is lower with an efficient quota allocation, while an inefficient quota allocation, because of rules that limit quota transactions, supports a larger number of smaller farms. The efficient quota allocation provides more value in terms of quota and producer surplus, and farms are more resilient to increased market access.

My understanding is that quota allocations are quite efficient in poultry, eggs and in dairy in the Prairies (Alberta, Saskatchewan and Manitoba). In the P5 provinces and British Columbia for dairy, various quota rules, including price ceilings, create frictions, lower the number of transactions and yield an inefficient quota allocation. If the trade negotiations lead to a significant and quick increase in market access, the risk is to see the number of dairy farms decline sharply in P5 provinces in particular.

The Canadian government compensated farms after signing CETA, CPTPP, and CUSMA. The payments are substantial. I expect the federal government to provide similar support if additional market access is given to the United States. These payments will slow down farm exit but will not stop it. Economies of scale mean that larger farms are more efficient, and given them, consolidation is an inevitable fact.

Liberalizing dairy quota transactions in P5 provinces would yield a high quota price. This is certainly not be a desirable outcome to some, but it does have upsides. Some farms would choose to sell their quota, giving them significant compensations and the opportunity to exit on their own terms. Farms who wish to expand will have access to quotas and exploit economies of scale. As a result, they will be in the best position to compete with American farms if additional market access is granted in upcoming negotiations.

Canadian exports of cattle and beef

Canada exports feeder cattle (young cattle going to feedlots) and fed cattle (about 18 months old cattle ready for slaughter) almost exclusively to the United States. Figure 1 shows monthly Canadian cattle exports to the United States since 2020. Exports of feeder cattle are small compared to the exports of fed cattle. Alberta exports the most fed cattle, on average, about 12,776 per month in 2024. Quebec is the largest exporter of feeder cattle. Saskatchewan and Manitoba exports of feeder cattle surged in late 2021 and early 2022 following the 2021 drought in the Prairies.

Figure 2 shows monthly Canadian exports of beef to the United States and to other countries. Alberta is by far the largest beef exporter with exports exceeding CAD 4.2 billion in 2024. Ontario follows with CAD 988 million. In 2024, 87% of beef exports were to the United States.

Estimates of the economic impacts

Like my previous post on the impact of US tariffs on the Canadian hog and pork industry, I use an equilibrium displacement model. This type of model has been used before to study livestock and meat trade between Canada and the United States (see for example, Brester, Marsh and Atwood (2004), Pendell, Brester, Schroeder, Dhuyvetter and Tonsor (2010) or Hahn, Sydow and Preston (2019)).

I model the supply chains in Canada and the United States. In the model, Canada exports feeder cattle, fed cattle and beef to the United States, and exports beef to other countries. The United States imports feeder cattle from Canada and Mexico, and fed cattle and beef from Canada. In practice, beef trade between Canada and the United States flows two ways, with Canada being a net beef exporter to the United States. The objective of the model is to capture the net flow of beef to the United States. I do not consider the impacts of tariffs on feed costs and on the costs for other inputs.

The model is calibrated to 2024 data except for beef consumption in Canada, for which the latest data available is for 2023. Calibration to 2024 data means that the model takes the economic conditions observed in 2024 and then examines the impact of 25% tariffs on US imports of feeder cattle, fed cattle and beef, everything else remaining the same. The length of run of the model is a year, meaning that the results are for the impacts of the tariffs a year after they are introduced, hence giving time to markets to adjust.

The results cannot be interpreted as a prediction of what would happen if tariffs were applied in 2025, but rather as a counterfactual exercise of the impact of tariffs for 2024. The economic conditions in 2025 are different than those in 2024; notably, the Canadian dollar is weaker. Nonetheless, the model gives a good idea of the size of the economic impacts that can be expected in 2025.

Table 1 shows the impacts on the Canadian cattle and beef industry of 25% tariffs on all US imports of cattle and beef from Canada and Mexico. The model finds a small impact on export volume of feeder cattle to the United States, although prices for feeder cattle in Canada decline by 16.5%. Total production of fed cattle in Canada declines by 2.5%, exports of fed cattle decrease by 18.6% and the price of fed cattle is 22.6% lower because of the tariffs. The tariffs cause retail beef prices in Canada to decline by 22.7%, which yields a 13.6% increase in beef consumption in Canada. Exports of beef to the United States drop by 46.6% but are partially replaced by a 46% increase in exports to other countries.

The economic impacts of the 25% tariffs on the Canadian cattle and beef sector are substantial. Total farm cash receipts for feeder and fed cattle declines by CAD 4.02 billion. The value of feeder cattle exports declines by CAD 46 million and the value of fed cattle exports drops by CAD 806 million. The value of beef consumption in Canada declines by CAD 1.14 because the increase in consumption volume does not fully compensate for the price drop. The value of beef exports to the US declines by CAD 2.04 billion, whiles value of beef exports to other countries increases by CAD 90 million.

Other considerations

As discussed in my previous post for hogs and pork, my model does not consider the effect of variations in Canada-US exchange rate. Currently, the average Canada-US exchange rate for 2025 is 4.4% lower than in the average for 2024. A lower exchange rate diminishes the impacts of the tariffs. With a 4.4% lower exchange, a 25% tariff becomes equivalent to 20.6% tariff.

The tariff war impacts markets in various ways. The model does not consider the impacts of tariffs on feed and other input costs and the impacts of retaliatory tariffs by Canada and other countries. It is possible that the tariff war causes lower production costs for cattle and beef in Canada and that foreign markets open to Canadian beef because of retaliatory tariffs on US beef, hence lowering the impacts estimated in this blog post.

Canadian exports of hogs and pork

Canada exports feeder pigs and fed hogs to the United States. The trade data allow us to distinguish by classifying feeder pigs as those that weigh less than 50 kg and fed hogs as those that weigh 50 kg or more. Manitoba is the largest exporter of feeder pigs, mostly from the export to the United States of weanlings weighing less than 7 kg. Ontario exports the most fed pigs to the United States.

Figure 1 shows monthly Canadian exports of live pigs to the United States since 2020. Exports of fed pigs are seasonal but have been relatively stable since 2001 on an annual level. Likewise, exports of feeder pigs have been holding fairly steady.

Figure 2 shows monthly Canadian exports of pork to the United States and to other countries. In 2020 and 2021, exports to other countries were boosted by the outbreak of African swine fever in China. China’s hog inventory has since recovered and the global demand has diminished. Pork exports to the United States are about a quarter of Canada’s total, and have been holding steady.

Estimates of the economic impacts

I use an equilibrium displacement model, a type of model that has been used before to study livestock and meat trade between Canada and the United States (see for example, Brester, Marsh and Atwood (2004), Pendell, Brester, Schroeder, Dhuyvetter and Tonsor (2010) or Hahn, Sydow and Preston (2019)).

I modelled the supply chains in Canada and the United States, starting with feeder pigs (weanlings), then fed hogs that are slaughtered to produce pork that is sold at wholesale and then at retail. I specifically consider Manitoba weanlings meant for export to the United States. Canadian fed hogs are either slaughtered domestically or are exported to the United States. Pork produced in Canada is either consumed domestically, exported to the United States or exported to other countries. I do not consider the impacts of tariffs on feed costs and on the costs for other inputs.

I did my best to calibrate the model to 2024 data. This was not always easy. Calibration to 2024 data means that the model takes the economic conditions observed in 2024 and then examines the impact of 25% tariffs on US imports of feeder pigs, fed hogs and pork, everything else remaining the same. The results cannot be interpreted as a prediction of what would happen is 25% tariffs were applied in 2025. The economic conditions in 2025 are different; notably, the Canadian dollar is weaker. More on that below. Nonetheless, the model gives a good idea of the size of the economic impacts that can be expected in 2025.

Lower exchange rate

My model does not consider that the Canada-US exchange rate in early 2025 is about 6% lower than it was on average in 2024. Effectively, a lower exchange rate diminishes the impacts of the tariffs. Indeed, in the past, countries have adopted monetary policies to lower their exchange rate as a strategy to get around punishing tariffs. The value of the Canadian dollar is expected to decline in response to the US tariff.

A 6% lower exchange rate means a 19% effective tariff in 2025 compared to the 2024 baseline. I calculated the impact of 19% tariffs in scenario 1. The impacts on feeder pigs are lower but for the rest of the supply chain, the impacts are the same as those in Table 1. This is because 19% tariffs are sufficient to cause exports of slaughter hogs and pork to the United States to drop to zero.

I found that a 7% drop in the exchange rate compared to 2024 would be sufficient for fed hog exports to the United States to continue, although export volumes would be much lower. For pork exports to the United States to continue, it requires about a 15% drop in the exchange rate given a 25% tariff.

Farm prices

The price data are from Statistics Canada Farm product index series. I use price indexes because they allow comparisons across provinces and groups of commodities. Farm price indexes consider a representative basket of commodities in each province. This means that price indexes vary across province because economic factors specifics to them and because bundles of agricultural products are not exactly the same across provinces.

Total crop price index

For all provinces, except for the Atlantic and British Columbia, the crop price index peaked in 2022 and has trended down since. These trends follow from commodity crop prices. Forecasts do not indicate a significant potential for prices to rebound over the next year.

The size of the 2022 peak depends on each province crop mix. The Prairies are important producers of commodity crops and we observe much more prominent peaks in those provinces. In Quebec and Ontario, the peaks are smaller, as fruit and vegetable productions are relatively more important.

Total livestock and animal products price index

Livestock prices have trended upward in all provinces since 2020, but have been volatile. The forecasts show that prices are likely to peak in the middle of 2025 to then decline following seasonal patterns, most notably for hog prices.

Total price index

When considering total farm price indexes, we observe that in provinces who largely depend on commodity crops, in particular Manitoba and Saskatchewan, prices peaked in 2022 and then have dropped since. The other Prairies Province, Alberta, has a large crop sector but also a large cattle sector. Prices have declined in that province but less so than Manitoba and Saskatchewan. Quebec and Ontario have diversified agriculture production and prices have stagnated since 2022, but have been volatile. For the Atlantic and British Columbia, prices have generally been trending up since 2010.

Farm cash receipts

The quarterly farm cash receipts data are from Statistics Canada Farm cash receipts series. The most recent observations are for the third quarter of 2024.

Total crop cash receipts

Crop receipts are seasonal, in particular in provinces like British Columbia, Ontario and Quebec, where non-storable crops account for a greater share of receipts. Crop receipts in the Prairies provinces are mostly from grain and oilseeds, storable commodities, somewhat smoothing crop receipts as farmers sell them throughout the year.

Crop receipts have trended up between 2010 and 2020 in all provinces. However, since peaking in 2022, they have weakened and flattened in the Prairies provinces. Similarly, they have flattened in Ontario over the last five years. For other provinces, crop receipts have continued to trend up.

Total livestock cash receipts

Growth in livestock receipts was weak in all provinces between 2010 and 2015, except for British Columbia. Growth has picked up since. Livestock cash receipts have trended up in recent years in all provinces and are expected to continue trending up in 2025.

Total farm cash receipts

For total farm cash receipts, the trends depend on the relative sizes of crop and livestock receipts in each province.

Farm cash receipts have flattened and are not expected to grow in Manitoba, Saskatchewan and Ontario. Cash receipts in these provinces more heavily depend on crops. In Alberta, it is the stength of the cattle sector that is pulling cash receipts up. In other provinces, total farm cash receipts are trending up and are expected to continue growing.

In total for Canada, farm cash receipts are stagnating following the trends in Manitoba, Saskatchewan and Ontario, who together account for more than half of the total receipts.

Are there farm recessions in Canada?

There is no short answer to that question, and nuances are necessary in a complete answer.

Given declining prices and revenues, the commodity crop sector is likely in a recession, and will continue to be in 2025 in all likelihood. Prices and revenues are generally strong in the livestock sector and should remain strong in 2025. However, this does not means that all types of livestock operations are doing well.

The farm sectors in Manitoba and Saskatchewan are likely in recession. The Ontario farm sector might also be in a recession but it is less severe because of more diversified farm revenues. The conclusion is similar for Alberta, who is seeing total farm cash receipts growing driven by its large cattle production. The farm economy appears more robust in other provinces

I haven’t mentioned input costs thus far, an important factor in the financial health of farms. Farm input price inflation has slowed down over the last two years and prices for some inputs, like fertilizer and fuel, have declined. Nonetheless, input prices continue to grow and against lower revenues in the commodity crop sector, it means increased financial stress to crop farms. Livestock farms are generally doing better as revenues are growing.

I do not have the data to evaluate the recent financial health of crop farms but the data above suggest it is declining. There are several programs available to Canadian farmers to cover production losses, increased input costs and market conditions. Generally, these programs are designed to cover losses from sudden shocks. They are not adapted to the current situation where commodity prices decline over an extended period of time, while input costs continue growing. If low crop prices continue, governments may wish to step in to keep too many farms from a dire outcome.

Cost of production estimates

The CDC makes available the results of the annual cost of production surveys at this page, along with a publication that describes the process of estimating the cost of production. The process has become a lot more transparent over the last two years. I have been slowly gaining a deeper understanding of the whole process, and accordingly improving my models.

I go over the details I did not get right in my September post.

Consumer price index

According to Statistics Canada, the CPI rose 2.0% on a year-over-year basis in August 2024. I was quite surprised to find that the CDC considers that it increased by 2.89%.

The way Statistics Canada calculated the annual change in the CPI is by taking the value of the CPI in August 2024, 161.8, and the value of the CPI in August 2023, 158.7, giving a 2% inflation rate (161.8/158.7 - 1 = 0.0195). This is the conventional way of calculating the change in the CPI. I figured out that the CDC took the mean of the CPI between September 2023 and August 2024, 159.9, and the mean of the CPI between September 2022 and August 2023 (155.4) to find a 2.89% increase in the CPI (159.9/155.4 - 1 = 0.0289).

I cannot say that the method used by the CDC is wrong but it is certainly unconventional. I’m not able to find good reasons to proceed that way. Seasonality should not be an issue as it is an annual rate that is calculated. Taking averages smooths out monthly shocks but large sudden shocks rarely ever happen on the CPI, in particular when calculating an annual rate.

Looking back at previous announcements, I found that the CDC has used the same method to calculate the change in the IPC for the last three years. I cannot confirm for 2021 and prior as changes in the CPI are not mentioned in the announcements.

If the CDC sticks to the same calculation method in the long run, I consider that it is acceptable, although the method is unconventional. Given enough time, Statistics Canada and CDC calculation methods should yield the same average annual change in the CPI. What the CDC cannot do, for credibility reason, is being inconsistent in the way it calculates the change in the CPI.

Drought maps

Agriculture and Agri-Food Canada (AAFC) publishes monthly a map that shows drought intensity across Canada. This is a great map that I frequently use during the growing season. However, as I’m interested in the impacts of drought on agriculture, I find it difficult to gauge how much farm land is affected by a drought. Below, I show in a map the extent that droughts affect farm land.

Note that monitoring drought over the entire territory it very relevant, especially to evaluate the risk of forest fire. My goal is, however, different as I seek to examine how droughts specifically affect agriculture.

Drought conditions for all Canada

The figure below reproduces the CDM map published by AAFC. I simplified the map by omitting to plot the impact labels and the impact lines that the AAFC map includes.

The map shows that a very significant part of Canada is under droughts of various intensities. The most intense drought region is Northern Alberta and the Northwest Territory. No region of Canada is an exceptional drought. The drought affects mostly northern regions where there is no agriculture. There are drought conditions in southern part of Canada but it is difficult to tell how much farm land is under drought.

Drought conditions for farmland

The next map show CDM map but only for the agricultural ecumene, defined as the areas where the country’s main agricultural activities take place. I use the agricultural ecumene as a proxy for farm land. I added to the map, in green, the parts of the agricultural ecumene that are not under a drought. As pointed out earlier, the data do not tell us whether there is a surplus of moisture in these areas.

The maps give a better idea of how much farm land is under a drought. Southern Saskatchewan and Northern Alberta, and a good part of the agricultural ecumene in British Columbia are in moderate or severe drought conditions. In the East, north-east parts of the agricultural ecumene in Quebec are in moderate drought conditions.

Droughts over time

What motivated me to collect drought data was to observe their frequencies. The following figures show by province the shares of the agricultural ecumene affected by drought conditions. The CDM data are only available since 2002, too short a period to determine whether climate change has made droughts more frequent and more intense.

I was surprised to find that there is a Wikipedia page about Drought in Canada. The page notes that the most serious drought was between 1999 and 2004 in the Prairies. Unfortunately, the CDM data do not allow us to fully observe that event.

Atlantic provinces

The agricultural ecumene in Atlantic provinces is small and meteorological events tend to affect the whole areas all at once. This is particularly obvious in Figure 3 for Prince Edward Island.

Atlantic provinces are not seriously affected by major drought events. The latest serious episode of drought was in 2020. Observe there were practically no drought before 2010.

Central Canada provinces

Quebec and Ontario are affected by episodes of drought, more frequently and more severely in Ontario. The most important drought was in 2007 and 2008 in Ontario. The data do not allow to identify episodes of excess moisture, which could be a more important problem than droughts in Quebec and Ontario.

Western Canada

Droughts are frequent and sometimes severe in Western Canada. Alberta and British Columbia are the driest provinces. The 2021-22 drought was widespread and intense with some areas under exceptional drought conditions. That drought seems to have never ended in Alberta, but Figure 4 does no allow us to observe whether it is the same areas that are continuously affected by drought conditions. British Columbia is under the worst drought conditions since the data begin in 2002. Large squats of the agricultural ecumene in Saskatchewan have stayed under drought conditions since 2001. Drought conditions returned to Manitoba in 2023 and 2024 but have recently disappeared as Figure 2 shows.

Farm gate prices to decrease in February?

The CDC typically releases at the beginning of October the results of its annual cost of production study. The National Pricing Formula determines the annual adjustment in farm gate milk prices by averaging the annual inflation rate for consumer prices and the rate of change in milk production costs.

The inflation rate has significantly declined over the last year. In the most recent data from August, the inflation rate is at 2.0%, compared to 4.0% in August 2023.

Figure 1 shows estimated milk total production cost. I followed the indexation from the CDC described in its cost of production study to estimate and forecast monthly production costs. You can read the 2023 study at this link. You can see from the figure that the total cost has been declining, by about 6% between summer 2023 and summer 2024. This is my estimate but it should not be too far from what the CDC will find. The CDC will use the cost study for 2023, scheduled for release in October 2024, to calculate the change in production costs.

Figure 1. Estimated milk total production cost

From the estimated change in the costs of production and the inflation rate, the farm gate price could decline by 2% (+2% x 0.5 + -6% x 0.5). However, under the exceptional circumstances mechanism, the adjustment in the farm gate milk price might not follow the formula. There are three criteria for the exceptional circumstances mechanism:

1. Stakeholders make the request;
2. Large difference between the inflation rate and the growth of production costs (8% here);
3. Estimated production costs exceed producer revenues.

Meeting the first criterion will depend on stakeholders. The second criterion is met. I do not think the third criterion will be met. Thus, it seems like a real possibility that milk prices will decline in February 2025. At the very least, dairy farmers should not expect the CDC to increase farm gate milk prices in 2025.

Farm gate price forecasts

I use two models to forecast farm gate milk prices by province. Here are brief details for those interested.

• Milk component models: This is the type og model I used in my February post. The farm gate price forecasts build from forecasts of the volumes and prices by milk classes.
• Time-series models: These models are a mix of elastic-net and ARIMA models.

The models are estimated separately for each province. The time-series models tend to produce smoother forecasts than the milk component models.

I assume CDC lowers farm gate milk prices on February 1, 2025, by 2%, consistent with my calculations above. Nonetheless, the models forecast slightly higher prices in 2025 compared to 2024. How is that possible? The CDC only influences prices. Domestic and international market forces play an important role, in particular in influencing the end-use of milk.

Comparison of Statistics Canada and Producteurs de lait du Québec milk prices

In my February post, I used monthly data from Les Producteurs de lait du Québec (PLQ), where the price is measured as the value to farm of one hectolitre at the average composition. My forecasts in this post use data from the Statistics Canada table 32-10-0077. Here, I examine whether the two data sources give comparable values.

Figure 5 compares prices for the data from PLQ and Statistics Canada. The figure shows a gap between the two series but that they move in sync until August 2021. The gap is mostly explained by transportation costs.

The area in light blue is the period on and after August 2021. From that date on, the prices from the two sources do not track as well. What happened in August 2021 is that PLQ introduced a new pricing method that gives greater incentives to producers to increase the relative butterfat composition of their milk. What I do not understand is why that change caused prices to no longer track each other. PLQ reports a value at the average composition. Although components pricing changed, it should not change the fact the value at the average composition can be calculated as the total revenue to farm divided by the total milk volume. My understanding of how Statistics Canada calculates the average price is by surveying farmers and asking for the average price they received, and by using administrative data. Thus the two sources should produce very similar values. It does appear there were a break in the way the average price is reported by PLQ.

Figure 5. Comparison of farmgate prices published by Statistics Canada and the PLQ

Figure 5 gives a hint as to why my model forecasted in February declining prices. The prices reported by Statistics Canada declined, consistent with my forecasts, but those reported by PLQ stayed relatively flat. In my previous post, I estimated the model using data from PLQ and most of the data series were from before August 2001. Thus, my model forecasted prices as if there never had been a changed in the way the average price is reported, hence producing a forecast more consistent with Statistics Canada data. As I use data from Statistics Canada in this post, I do not expect this will be a problem this time.

Fertilizer prices

As usual, let me begin by showing recent prices for fertilizers. 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.

Over the last two years, we see that the price for Anhydrous ammonia (82-0-0) has trended down but that prices for Urea (46-0-0), Fertilizer (11-51-0), and the US price index have sort of flattened. Movements in the short run are more concerning as all prices have increased over the last few months. It seems that downward trends in fertilizer prices are reversing, or perhaps that prices are going toward a new equilibrium around their actual levels.

Figure 1. Alberta fertilizer prices and US fertilizer price index

Drivers of fertilizer prices

As I have discussed in previous posts, among the most important drivers of fertilizer prices in Canada are barge rates on the Mississippi River, the Chicago Mercantile Exchange (CME) corn price and the Henry Hub natural gas price. Let’s have a look at them before examining forecasts for fertilizer prices. In Figure 2, I forecast barge rates using time-series models, and the corn prices and the natural gas price follow their Futures Forward Curves (FFC).

Figure 2 shows that barge rates on the Mississippi have trended down but that my models forecast them to increase in the next few months. That forecast follows from typical seasonal patterns in barge rates which increase when water levels on the Mississippi are low. Unless Mississippi water levels drop, we should observe fairly stable barge rates. Corn prices appear to have bottomed out in February, but they have stayed low since. Low corn prices mean a weakening of the demand for fertilizers. As we learn more about the condition of the new crop and prices for 2024, we will gain insights about the demand for fertilizer for the 2025 crop, and in consequence about fertilizer prices in the coming months. Natural gas prices are climbing, raising the cost of producing certain fertilizers.

Overall, Figure 2 shows there are opposite forces affecting fertilizer prices in Canada. That is where models are particularly useful in summarizing information into forecasts.

Figure 2. Drivers of fertilizer prices

Forecasts for Alberta fertilizer prices

Figure 3 shows forecasts for fertilizer prices based on Alberta input prices data. The economic conditions currently point toward relatively stable fertilizer prices. The models forecast a small increase in the price for Anhydrous Ammonia (82-0-0) following higher natural gas prices, but small declines in the prices for Fertilizer (11-51-0) and Urea (46-0-0).

Figure 3. Forecasts of Alberta fertilizer prices

Forecasts for Statistics Canada fertilizer prices indices

I prepared forecasts for the fertilizer price indices from the Statistics Canada table 18-10-0266 for industrial product price indices. The prices indices are for all of Canada.

Comparing Figures 3 and 4, we observe that fertilizer price indices have moved in sync with Alberta fertilizer prices. Accordingly, it is not surprising that the forecasts for the price indices are like those for Alberta fertilizer prices. The models forecast that under the current conditions, fertilizer prices should stay relatively stable over the next six months.

Figure 4. Forecasts of Statistics Canada Fertilizer prices indices

Farmland values vs. the stock market

Is it better to invest in farmland or in the stock market? I’ll answer this question from a historical perspective by looking at the return of investing $100 in farmland versus investing $100 in the stock market in 1996, the year the available FCC data begin.

Figure 1 shows how $100 invested in farmland or in the stock market would have grown through time. Each gray line shows the return from an investment in farmland for a region. The S&P/TSX composite index is in blue and the S&P 500 is in red. The S&P/TSX composite index is a measure of the expected return from investing in Canadian stocks in Toronto while the S&P 500 is a measure of the expected return from investing in US stocks (New York).¹ The S&P/TSX (Toronto) does not set the bar high in terms of yield, averaging a 4.8% annual return. The S&P 500 (New York) average annual return was 7.1%.

Note that the returns to owning farmland I discuss only consider the direct return from buying and selling land. They do not take into account revenues from exploiting or renting out farmland, and costs such as property taxes. Similarly, the stock indices are not corrected for dividends. In both cases, the data tend to underestimate returns but I do not know to what extent. These considerations could affect the comparisons below.

The figure shows that for almost every region, the return to farmland by far exceeds the return from investing in the stock market. Regions for which the average annual return from owning farmland is below the return from investing in the S&P/TSX (Toronto) are Prince in Prince Edward Island, Truro-Shubenacadie and Pictou-Antigonish in Nova Scotia. Regions where the average annual return beats the S&P/TSX (Toronto) index but is below the S&P 500 (New York) are Kings and Queens in Prince Edward Island, Western in New Brunswick, and Cariboo-Chilcotin and Peace-Northern in British Columbia.

Farmland values increased spectacularly in several regions, in particular in Ontario and Quebec. The highest average annual return is for the Laurentides-Lanaudière region in Quebec at 11.6%. A $100 investment in farmland in Laurentides-Lanaudière in 1996 is worth $1,920 in 2023. Four regions in Ontario follow: Mid Western, Central West, South East and South West. Two regions in Quebec are next: Chaudière-Appalaches and Montérégie.

@tabl-tab1 shows the average return to farmland by province, weighted by area (more on that below). In all provinces, the average return is above the S&P/TSX (Toronto). The average return in only two provinces is below the S&P 500 (New York): Prince Edward Island and Nova Scotia. The table includes the standard deviation as a measure of risk. Based on that measure, the risk is lower from investing in farmland than investing in the S&P 500 (New York) or in the S&P/TSX (Toronto).

Overall, these data tell us that farmland was an excellent investment between 1996 and 2023. In most regions of Canada, the return on farmland exceeded the average return from the stock market, at a lower risk!

Mapping farmland values in Canada

Let us now have a look at farmland values across Canada on maps to gain a better understanding of the geography of farmland values. It is not easy to map farmland values through time. To simplify, I’ll approach the task in two steps. First, I’ll map farmland values for selected years. Then I’ll show an animation of how farmland values have changed through time.

FCC shows farmland values on maps where regions are coloured according to farmland values, regardless of the surface of land that is actually farmed. I’ll take a different approach by colouring only the surface of land corresponding to the agricultural ecumene to give a more accurate visual.² Statistics Canada defines the agricultural ecumene as the areas where the main agricultural activities take place.

Figure 2 shows maps of farmland values for 1996, 2005, 2014 and 2023. As farmland values increase, the map darkens. This is true for all regions as farmland values have increased everywhere. There is consistency through time with the highest-valued land in 1996 being also the highest valued in 2023. The most valuable land is in the South Coast region near Vancouver in British Columbia. The agricultural land in that region cover a small area and is barely visible in the maps. Elsewhere, the most valuable land is along the Saint Lawrence River, by lake Erie and lake Ontario, where most of Canada’s population is located.

Notice how much more farmland there is in the Prairies provinces compared to the rest of Canada. Farmland in the Prairies tends to be more affordable than farmland in Ontario, Quebec and British Columbia. Also note that there is a net difference in the value of land in Alberta and Saskatchewan.

The animation below offers a more dynamic view. Observe that only in rare cases that farmland values decline year to year. The animation shows from the map darkening rapidly that farmland values increased fast in the 2010s, contrasting with the prior years when farmland values were more stable.

Farmland values in Canada and provinces

I show below average farmland values across Canada. This is not something that FCC shows in its annual farmland value report, although FCC gives the annual average percentage change in farmland values for Canada. I calculated average farmland values by weighting regions by the area covered by the agricultural ecumene for the 2021 Census of Agriculture. Table 2 shows the results, and also contains values by province and the annual percentage change.

The annual percentage changes I calculated are different than those reported by FCC. My guess is that it mostly comes from different weighting, and from only considering cultivated non-irrigated farmland. I do not know what weighting FCC uses when aggregating values. I tried as an alternative to weight the data according to the land use data from the Census of agriculture but I could not replicate FCC’s results based on the annual percentage changes of individual provinces.

Concluding remarks

The data presented above show that for most regions of Canada, investing in farmland in 1996 would have been a great investment. This conclusion holds true for the period considered and thus someone should not conclude that the return from investing in farmland always exceeds the return from investing in the stock market.

Table 3 shows the average annual return by decade from investing in farmland in Canada and the average annual return in the stock markets in Toronto and New York. The farmland value data are from Statistics Canada, and thus also include the value of buildings, unlike the FCC data I used above. These data allow me to go furhter back in time than the FCC data.

From table 3, an investment in the S&P/TSX (Toronto) or the S&P 500 (New York) during the 1980-1990 or the 1990-2000 decades would have done better than an investment in farmland for most provinces. Afterward, the return from investing in farmland exceeds the return from investing in the stock market. The 1980s and 1990s were not kind to agriculture with stagnating or declining agricultural commodity prices and a wave of bankruptcies in the 1980s. In the 2000s and 2010s, growth in biofuel production caused agricultural commodity prices to rise, and demand for farmland to grow, causing farmland values to increase.

What to expect for the next decade? Farmland values will follow commodity prices and profitability in agriculture. If there is new demand for agricultural commodities, for example from increased biofuel production for airplanes, I expect farmland values to continue their strong growth. If agricultural commodity prices stagnate or decline, then farmland values may decline. Climate change will also play a role in future farmland values, but it is difficult to predict what that role will be.

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.

Figure 1. Alberta fertilizer prices and US fertilizer price index

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.

Figure 2. Drivers of fertilizer prices

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.

Figure 3. Forecasts of Alberta fertilizer prices

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.

Figure 4. A look a previous forecasts of Alberta fertilizer 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.

Agri-food trade

Published in 2018, the Report of Canada’s Economic Strategy Tables: Agri-food sets a target of CAD 85 billion for agri-food exports by 2025. Canada’s exports surpassed the target early with agri-food exports totalling CAD 93.9 billion in 2022.¹ Agri-food exports were CAD 99.9 billion in 2023. A major factor explaining that success is price inflation for agri-food products.

Figure 2 shows Canada’s monthly agri-food trade between 2010 and 2023. Exports, imports for both agriculture and food have been trending up and are highly seasonal. Exports of agricultural products largely surpass imports. In 2023, Canada exported on average CAD 3.9 billion and imported CAD 1.8 billion per month of agricultural products. The agricultural trade surplus was CAD 25.6 billion in 2023.

Food exports and imports are comparable in value. Food imports exceeded exports until 2019 when Canada became of net food exporter. Canada returned to a negative food trade balance in 2020 but since 2021 Canada has been a net food exporter. In 2023, Canada exported on average CAD 4.4 billion and imported CAD 3.9 billion per month of food. The food trade surplus was CAD 6.1 billion in 2023.

Trade by country

Let’s now examine who are Canada’s main agri-food trading partners. I will use bump charts to show Canada’s top ten trading partners over time. The charts show at the top Canada’s number one partner and at the bottom Canada’s tenth trading partner. The charts do not show trade values but the size of the circles show the relative importance of the trade relationship, a larger circle indicating a greater value.

Agriculture

The United States is by far Canada’s main export destination for agricultural products. Japan, China and Mexico round the top four. From positions five to ten, there is no consistency with countries jumping up or down several positions from one year to the next. India, Bangladesh, and Indonesia jump in and out of the top five depending on their imports of pulses, cereals and canola.

The ranking of Canada’s agricultural imports by origin is quite consistent. The United States is number one, Mexico number two, China number three and Chile number four. After that, it varies from year to year. Note that Thailand was ranked number five in 2010 and that it has dropped out of the top ten in 2023. Peru was out of the top ten until 2014 and it is now number four. CPTPP likely helped Peru’s exports to Canada. The main imports from Chile and Peru are edible fruits and nuts.

Food

Again, it’s no surprise that the United States is the main destination, and by far, for Canada’s food exports. China, Japan, Mexico and South Korea, follow in that order since 2016. The Philippines, the Netherlands and the United Kingdom have ranked sixth, seventh and eighth in the last three years.

The United States is again the main trading partner when examining food imports. Then the list of countries is quite different from the previous charts. Italy has been number two since 2013. Brazil stole in 2023 the number three position from France, who dropped to number four. Mexico is in fifth position followed by China and then by the United Kingdom who has not been lower than in seventh position except in 2018 when it fell to number eight.

Trade by commodity

I also show rankings for agri-food trade by commodities using bump charts. The top traded commodity is at the top and the size of the circles shows the relative value of trade.

In the figures below, I include the HS4 code in the shortened product descriptions to avoid confusion. You can verify product classification at this page.

Agriculture

Prairies crops dominate the ranking of Canada’s exports of agricultural commodities. Wheat is number one, followed by canola. In third are legumes, who often trade position with crustaceans. Soya is in fifth position since 2019. Live bovine is in sixth but occupied a higher spot before 2010. Live bovine exports to the United States have declined and have stayed lower since the United States adopted mandatory Country of Origin Labelling (m-COOL) for red meat in 2008, even though the dispute was resolved in 2015.

Not too surprisingly for a northern country, Canada’s two main agricultural imports are fresh fruits and vegetables. The ranking of Canada’s agricultural imports tends to vary significantly from year to year depending on commodity prices. The top ten includes several products that are not grown, or grown in small quantities, in Canada like bananas, citrus, rice and grapes. Observe the recent rise of corn imports in the ranking following the drought in the Prairies in 2021.

Food

The products occupying the top four of Canada’s food exports tend to trade positions: bread and pastries, canola oil, beef, and pork. In the case of bread and pastries, it’s a bit of mystery to me as I do not know the comparative advantage that explains strong exports of bread and pastries by Canada (wheat production?). Moreover, Canada is also a strong importer of bread and pastries, although exports largely exceed imports. In the case of chocolate products, who are consistently among the top ten food exports, although Canada does not produce cocoa, lower sugar prices in Canada compared to the United States makes Canadian firms competitive on the US market.

Recent lower pork prices have caused pork to fall to number four. On the opposite, with higher prices from droughts in the US and Canada, beef has climbed to the third spot in 2023.

Food imports include several specialty products like wine, coffee, and certain alcohols. Bread and pastries occupy the top position, just like in the case of exports. Two-way trade also occurs for preparations for animal feeds and for chocolate. Quite obviously, Canadians like imported alcoolique beverages.

Are we going to see increased protectionism and in what form?

Whether we will see more protectionism depends on a few factors, including:

• Political situation: we will see more protectionism depending on who gets elected in the United States in November and likewise in the next rounds of elections elsewhere in the world. The new caucus in the US House of Representatives could also lead to increased protectionism.
• Agri-food prices: When farming and producing food is profitable, the impetus for protectionism is much lower. The USDA recently forecasted net farm income to decline by 25.5% in 2024. If this forecast materializes, this could lead to increased pressures for protectionism in the United States.
• Protests in Europe: We expect changes to come after the protests that rocked several European countries since January. That could include increased farm support but also increased protectionism.

The figures above show the obvious fact that the United States is Canada’s main agri-food trade partner. This depency on the US market has its risks. What could we expect from the United States if it moves to increase protection of its agri-food sector?

• Tariffs: Former President Trump stated that he would impose tariffs on all imported goods if re-elected. This, it appears, would not exclude Canada and agri-food products and would essentially mean the end of free trade with the United States.
• Return of m-COOL for red meat: Since the US lost at the WTO against Canada and Mexico regarding m-COOL for red meat, steps are occasionally taken for its return. Currently, profitability is low in the hog sector but is high in the cattle sector. I would not be surprised if we see increased pressure to reinstate m-COOL once profitability in the cattle sector declines.
• Measures regarding milk: In the second round of the dairy dispute between Canada and the United States, the CUSMA panel sided with Canada. In reaction to the decision, Ambassador Katherine Tai stated We will continue to work to address this issue with Canada, and we will not hesitate to use all available tools to enforce our trade agreements and ensure that U.S. workers, farmers, manufacturers, and exporters receive the full benefits of the USMCA. This quote leaves open the possibility that the United States could take measures against Canada’s dairy sector even though the panel sided with Canada.

Canada’s agri-food export portfolio comprises a large number of countries and products. If any country importing Canada’s agri-food restricts imports, this may result in non-negligible economic impacts to Canadian agri-food sectors. It’s hard for me to guess how import restrictions by countries other than the United States could manifest. Based on recent history, I cannot rule out that China could impose restrictions on the import of certain commodities as it had done for canola and pork a few years ago. However, I cannot predict if, when, and how such restrictions could be imposed. India imposes tariffs on pulses depending on domestic prices. In December, it temporarily removed a 50% tariff on the import of pulses. The tariff is set to come back on March 31st 2024.

Canada’s trading partners will continue to put pressure for market access in sectors under supply management. Recent trade agreements have maintained supply management systems but weakened them by increasing access to Canada’s domestic market. Country members of these agreements may challenge Canada’s TRQ administration, event though economic factors explain why they are unfilled. Recent examples include challenges for dairy TRQs by the US under CUSMA and by New Zealand under CPTPP. Although these challenges are not protectionnism measures, there can be a cause for concern. We will see if the door is now open for more challenges.

Where do farm gate milk prices come from?

Before discussing the forecasts, let me offer discuss briefly how farm gate milk prices are derived. I’ll simplify the description of the system, perhaps some will say oversimplify, and I’ll focus on farms and processors.

Farmers sell raw milk to processors who manufacture from it consumer products or ingredients for further processing. Milk is assigned a class depending on its end use:

• Class 1: Fluid milk and beverages;
• Class 2: Yogurts, sports milk drinks, ice cream, etc.;
• Class 3: Cheese;
• Class 4: Butter, butteroil and dairy product for the food industry (e.g., milk powders);
• Class 5: Dairy products used as an ingredient in the further processing sector.

Processors value milk components: butterfat, proteins, and other solids. It is the quantities of components in kilograms that are billed rather than the volume of milk in litres. Prices for the milk components across classes are set by the Canadian Dairy Commission (CDC): see the prices here and here. For class 4A, following agreed upon rule in the Canada-United States-Mexico Agreement (CUSMA), prices for proteins and other solids are set monthly based on the USDA nonfat dry milk price. The CDC buys and sells butter at a support price which stabilizes seasonal variations in the production and the demand for butterfat.

Revenues from the sales of raw milk are pooled and provincial boards derive from them farm prices for butterfat, proteins and other solids. Farm prices for the components do not equal their average values when sold to processors. Farm prices reward the production of butterfat relative to the sum of proteins and other solids. The table below shows the Quebec farm gate prices for December 2023. The volume of milk going to level 1 or level 2 depends on the ratio of the sum of proteins and other solids over butterfat and as such it is not possible to calculate the average value based on the average composition. Below, my objective will be to forecast the average value, which was $98.28/hl in December 2023.

The forecasts

I wrote more than a year ago about the trends affecting the Canadian dairy industry. These trends are still relevant today and the models I developed account for them.

For my forecasts of Quebec farm gate milk price, I assume the following:

• No adjustment to milk prices and to the support price for butter in 2024 other than what will take into effect on May 1st;
• GDP grows by 1% in 2024 as forecasted by the Bank of Canada in its latest Monetary Policy Report.
• The exchange rate stays constant at the latest value observed at the time of the forecasts.
• The USDA nonfat dry milk price follows the futures forward curve.
• Total milk volumes and component concentrations follow their recent trends and their typical seasonal patterns.
• The butter stock-to-use ratio follows typical seasonal patterns.

Regarding the total raw milk volumes, the model predicts a 1.6% increase for the P5 (Ontario, Quebec, New Brunswick, Nova Scotia and Prince Edward Island) provinces in 2024.

Figure 1 shows the USDA nonfat dry milk price converted into CAD per kilogram. From the futures forward curve, we can expect a higher price for USDA nonfat dry milk in 2024 compared to 2023 but lower than in 2022. This will support a slightly higher price for class 4A, which includes skim milk powder.

Figure 2 shows raw milk volume shares for selected milk classes. The model does not forecast large changes in the distribution of milk across classes in 2024 compared to 2023 except for a smaller share of milk going into class 1A in the last quarter of 2024. This will have a negative impact on farm gate prices as we will see below.

Figure 3 shows Quebec average farm gate prices for 2022, 2023, and forecasts for 2024. For 2024, farm gate milk prices in Quebec should increase by 0.8% compared 2023, representing $0.007/litre. The model finds higher prices for the first 8 months of 2024 compared to 2023, but slightly lower prices in the months afterwards because of a smaller share milk forecasted to flow into class 1A in the last quarter of 2024 (see Figure 2).

The main takeaway is that economic conditions do not support a significant increase of farm gate milk prices in Quebec in 2024. The May 2024 adjustment to the farm gate price of milk will contribute to slightly higher prices but not enough to counter the impact of smaller volumes going into class 1A in the fourth quarter.

What else?

I did not touch on several issues that impact farm gate milk prices in Quebec, including trade and revenue pooling between P5 and Western Milk Pool (WMP: British Columbia, Alberta, Saskatchewan and Manitoba) member provinces. Dairy imports from the United States, Europe and New Zealand impact Canada’s domestic milk production and can also affect farm gate prices. Regarding revenue pooling, I expect that it has a marginal impact on farm gate prices although it should help smooth them out. My models implicitly account for trade and revenue pooling, and several other issues not discussed.

I plan on further refine the models and to update the forecasts occasionally. I hope eventually to extend the models to forecast farm gate dairy prices in other provinces.

Where are Canadian farms?

Before showing maps of infrastructure in Canada, I think it is appropriate to recall where farms are located. In Figure 1, each red dot represents one farm randomly located within the census consolidated subdivision it belongs to because I do not have the exact location of farms. I explain how I produce this type of map in the first post of this series.

The map shows greater concentrations of farms in Southern Ontario and in Quebec along the Saint Lawrence River. There are farms all across the agricultural ecumene. Observe the white spots bounded by the agricultural ecumene. These are large cities. The locations of Montréal, Ottawa, Toronto, Regina, Edmonton and Calgary are easy identified.

Rail roads

The map below shows railroads in Canada based on data from Open Government for the National Railroad Network. The history of the development of the Canadian railroad system is fascinating and shows how crucial railroads were to economic development and settlement in the Western parts of the country.

The Canadian National (CN) and Canadian Pacific (CP) are the two main rail companies in Canada. The Grain Growers of Canada observes that 94 percent of Canadian grain is shipped by rail. It is not surprising to find that the railroad system covers well the agricultural ecumene in the Prairies and allows for the export of grain through Eastern or Western ports. In the East, railroads are less crucial to agriculture because of the commodities produced, the proximity of the Saint-Lawrence Seaway and the existing road network.

Grain elevators

Grain elevators are often located along railroads and therefore it seems natural to map them while also showing the railroad network. I obtained data on the location of grain elevators from Agriculture Canada through Open Government. The data for 2021 include the location of 441 elevators, two more than the 439 listed by the Canadian Grain Commission. I am aware that not all elevators are shown as the data seem only to include elevators situated by railroads.

The three types of elevators shown in the map below differ in the class of their grain company licences. Primary elevators receive grain directly from producers for storage and forwarding. Process elevators receive and store grain for manufacturing and processing. Terminal elevators receive grain from other elevators for cleaning, storing and treating before it is shipped elsewhere. The map shows that most primary and process elevators are in the Prairies. Elsewhere, the elevators mapped are terminal.

I must note that Statistics Canada offers the Grain Supply Chain Dashboard: Real-time Grain Movement by Rail. In addition to show the railroad network and the location of elevators, the dashboard shows number of cars by corridor and dwelling time.

Roads

The data for the roads mapped below are for the National Road Network made available through Open Government. The data appear to contain all numbered roads in Canada and winter roads.

The road system is extensive and for the most part is built on the agricultural ecumene. Outside of the agricultural ecumene, there are roads mainly for logging and mining, and to reach remote communities. The road system is denser on the agricultural ecumene than outside of it, especially in Quebec and Ontario. The agricultural ecumene in the Prairies is gridded with roads.

Pipelines

No farmer wishes to see a pipeline crossing over his/her farm. Pipelines provide an essential service to the economy but no direct service to agriculture. The building of the Keystone pipeline exemplifies how the development of such infrastructure is controversial for several reasons, including the seizure of farmland by eminent domain.

The figure below maps pipelines in Canada using data from Natural Resources Canada made available through Open Government. There are some gaps in the data as there are discontinuities in the pipe system.

Most pipelines are in oil-producing provinces, in the Prairies, and on the agricultural ecumene. Elsewhere in Canada, there are fewer pipelines, with some pipelines on the agricultural ecumene in British Columbia, Ontario and Quebec.

Power lines

The construction and exploitation of power lines is perhaps less controversial than that of pipelines as power lines are associated with a lower environmental risk. Nonetheless, power lines may not be welcomed on farmland as they can disturb farming activities and degrade the landscape.

The data for the power-line network mapped below are from Natural Resources Canada and made available through Open Government. The data seem to include only large transmission lines and not include lower voltage distribution lines. There are some gaps in the data as there are discontinuities in some lines.

One particularity of Canada is that there are several hydropower stations located north, far away from population centres. This is particularly obvious in Manitoba, in Ontario and Quebec. The greatest concentrations of power lines are in southern Quebec and southern Ontario, over the agricultural ecumene. Elsewhere, the concentration of power lines is much lower.

Farmland, population and infrastructure

Given the size of Canada, it is hard to appreciate distances in the maps above, and, in order to show well infrastructure, its size is disproportionate in the maps. Nonetheless, the maps show that the agricultural ecumene represents a small share of land in Canada. Farms share land with the population, industries and the infrastructure that support the economy. Inevitably, this leads to disagreements, compromises over the best use of land, and sometimes to legal outcomes.

It is important to protect farmland but not at all cost. It might be better not to exploit less productive marginal farmland or to use it for other purposes. Sometimes, the economic importance of certain projects may justify the loss of farmland. In any case, it is worth approaching the conversion of farmland holistically with a long-term perspective, in particular for projects that lead to the irreversible loss of farmland.

Yield curve inversion is meaningless

The yield curve shows the relationship between the yield and the maturity of securities. In a normal market, yields increase with maturity as investors ask for higher returns on riskier long-term investments. The yield curve is inverted when yields decline with maturity. This is a situation where investors consider that the near term is riskier than the long term. Inversion of the yield curve is considered an early signal of a recession.

The animation below shows the yield curve between January 2018 and November 2023 in two panels. The panel on the left shows the yield curve: on the horizontal axis is maturity increasing from left to right and on the vertical axis is the yield. The panel on the left compares the yield on the 3-Month treasury bill and the 10-Year bond to determine when there is an inversion. The area in blue shows the spread when the yield on the 10-Year bond exceeds the yield on the 3-Month treasury bill. The area in orange shows the spread when the yield on the 10-Year bond is lower than the yield on the 3-Month treasury bill, that is, when the yield curve is inverted.

The animation shows that the yield curve has been inverted for nearly a year and a half. Yet, Canada has not been in recession. Such a continuous false positive signal makes inversion of the yield curve a meaningless indicator that a recession is coming. We cannot conclude from the yield curve that a recession is coming.

The next animation displays the effective interest rates paid by households and businesses on the right panel. It shows how the Bank of Canada quantitative tightening measures and increases of its policy rate tightened credit conditions for households and businesses. Interestingly, the household average interest rate has flipped back above the average business interest rate recently.

Sahm’s rule recession indicator weakly indicates a recession

The Sahm’s rule recession indicator is an early recession indicator that compares the average unemployment rate in the last three months and the minimum monthly unemployment rate over the last 12 months. If the difference between these two values exceeds 0.5, it is a signal that a recession is beginning.

The figure below shows the Sahm’s rule recession indicator for Canada. The latest employment data are for November 2023. The figure shows that the indicator has been rising and has been above 0.5 for the last three months. This tells us that Canada is on the brink of a recession, but the signal is still weak. In previous recessions, the indicator sharply rose.

Labour market strength: the vacancy to unemployment ratio

The vacancy to unemployment ratio is not a recession indicator but it offers a good measure of the strength of the labour market. The ratio tells us how many jobs there are available for each unemployed worker. A ratio above 1 means there are more jobs available than there are workers looking for a job.

The figure below shows the ratio peaked in the middle of 2022. The labour market was particularly tight in Quebec and British Columbia, but has softened over the last year in part because of the influx of immigrants. Nonetheless, it is still strong compared to before the pandemic.

The labour market appears in a strong position such that if a recession arises, it should not affect workers like previous recessions. That is, if Canada experiences a mild recession, we can expect the unemployment rate to stay relatively low.

Is a recession coming?

The Sahm’s rule recession indicator says that a recession is approaching but the signal is still weak. If the indicator sharply increases in the December unemployment data, then it will be clear that we are heading into a recession. For now, the signal is still too weak to make a definitive statement from it.

The signs point toward a weaker Canadian economy in 2024. Whether this means a recession, I do not know. I agree with what appears to be the consensus among economists that the Bank of Canada will begin lowering its policy rate in 2024 and that the economy will strengthen from then on.

Recent data

Forecasts

I forecast fertilizer prices from scenarios for the models’ variables:

• Exchange rates, oil prices, barge rates and rail rates stay at their October 2023 averages thus far,

• The corn price follows its Futures Forward Curve (FFC),

• Three scenarios for natural gas prices (Henry Hub):

  • FFC: natural gas price follows the FFC.

  • FFC -25%: natural gas price declines by 25% from its FFC.

  • FFC +25%: natural gas price increases by 25% from its FFC.

The FCC for corn shows that markets expect corn prices to rise slowly throughout the crop year, as expected under normal market conditions. For natural gas, prices are expected to increase until February and then to decline.

Figure 5 shows the forecasts for the three scenarios. The forecasts are for stable fertilizer prices, and even a small decline in the case of anhydrous ammonia. This is not too surprising. Corn prices are one of the main predictors of fertilizer prices. Lower corn prices appear to overcome increases in natural gas prices and barge rates, resulting in lower expected fertilizer prices.

Figure 5. Forecasts of Alberta fertilizer prices

Fertilizer inventories

The St. Lawrence Seaway strike could cause higher fertilizer prices in Eastern Canada. Strong fertilizer inventories should help ease the impact of the lack of supplies in the short term. Data for fertilizer inventories for September in Figure 6 show that Eastern provinces were not in a bad position before the strike began. Eastern inventories of UAN were low compared to the previous two years. For MAP and Urea, they were similar to the previous two years. Let’s hope the strike will not last and that given decent inventories, impacts on fertilizer prices will be minimal.

In Western Canada, inventories compare to previous years. This reinforces the forecasts above that conditions are there for stable fertilizer prices in the next six months.

Figure 6. March fertilizer inventories in Canada (Statistics Canada table 32-10-0036

Corn, soybean, hog and poultry & eggs

I show in Figure 1 maps corn and soybean farms because it is common practice to rotate these two crops. Most corn and soybean farms are in Ontario and Quebec. There is a concentration of soybean farms in Manitoba but few corn farms. In that region, farmers rotate soybeans with crops like peas, dry beans, canola and oats. Across Canada, there are almost twice as many soybean farms than corn farms. Soybeans require fewer growing degree days than corn so it’s not all that surprising to see more of it in Canada.

Figure 1 also shows the locations of hog & pig farms, and poultry & egg farms. Both use corn as the main ingredient in feeds with soybean meal added to increase protein content. Thus, we expect hog & pig farms, and poultry & egg farms to be located near corn and soybean farms. That is indeed the case for hog & pig farms. Poultry & egg farms, however, are located across Canada, with greater concentrations in Ontario and Quebec. This reflects the supply management system in Canada where production quotas are managed at the provincial level.

Wheat, canola, dry pea & bean and other grain

Figure 2 maps wheat, canola, dry pea & bean, and other grain farms. There is no NAICS code specifically for canola but the NAICS code for oilseed (except soybean) is almost entirely comprised of canola farms. The classification for other grain farming comprises grain farms that do not enter in the other categories or who primarily engage in growing a combination of oilseeds and grains.

The crops in Figure 2 are most important in the Prairies and are often planted in rotation. There are wheat farms all across the Prairies and in Southern Ontario. Canola is dominant in the Prairies, especially in the north of the agricultural ecumene. There are more than twice as many canola farms than there are wheat farms. Dry pea & bean farms concentrate in Saskatchewan. The other grain category is the largest. It encompasses farms who tend not to specialize in a single crop and shows that farms across the country tend to diversify into several grain and oilseed crops.

Hay, dairy, beef and sheep & goat

Hay feeds cattle. It is therefore not surprising to find in Figure 3 that hay farms, dairy cattle & milk farms, and beef cattle farms are located in the same areas. Supply management for milk, administered by provincial boards, favours a distribution of dairy cattle & milk farms across Canada. Beef cattle farms are remarkably well distributed across Canada.

It would have been interesting to map barley farms as barley is used as a feed especially in the Prairies. Unfortunately, the data do not include a category specifically for barley. To complete the figure, I added sheep & goat in Figure 3. They are not very common. Comparatively, there are eleven beef cattle farms for every sheep & goat farm. Most sheep & goat farms are Ontario, Quebec, and British Columbia.

Potato and tobacco

Figure 4 shows the geographic distributions of potato and tobacco farms. I chose to plot potato farms because people tend to think the Prince Edward Island and New Brunswick are the largest producers of potato in Canada. In 2022, Alberta was the largest potato producer in Canada, just ahead of Prince Edward Island. Manitoba was third, right behind, and then there is quite a drop to New Brunswick in fourth place and Quebec in fifth place.

I mapped tobacco farms because I was curious to know whether there were still many of them and where they are located. According to the 2021 Census of Agriculture, the vast majority were in southern Ontario, there were a few in Alberta, one was in Quebec and one was in Prince Edward Island. The map does not show every farm who grew tobacco as it shows farm who primarily produced tobacco. Thus, there are likely more farms who grew tobacco but for which it was not the main source of income.

The growth of soybean production

When I left Quebec in 2004 for my PhD for California in 2004, I would see very few soybean fields in my home region. As I came to visit every summer in the following years, I would see more and more of them. Higher prices, better genetics, hotter summers and know-how made them suitable for Quebec’s climate.

Figure 5 shows how the number of soybean farms has changed over the last three censuses. First, observe the growth in the number of soybean farms. It went from 6,423 farms in 2011, to 7,859 farms in 2016 and to 11,741 farms in 2021. In Ontario and Quebec, we can see the distribution of farms getting denser and expanding north. In Manitoba, there is a clear growth in the number of soybean farms between 2011 and 2016 and then the number of soybean farms declined between 2016 and 2021. Farmers switched away from soybean after disappointing yields in dryer years.

Figure 6 shows the average yields and seeded areas for soybean by province. Ontario and Quebec have the highest average yields. Yields are lower and more volatile in Manitoba. Thus, it’s not too surprising that farmers in Manitoba switched to other crops given yield risk with soybeans. On the opposite, with higher and more predictable yields, Ontario and Quebec farmers have slowly increased soybean acres.

What is a farm?

That is the question. Definitions vary across countries, over time and purpose. Statistics Canada changed the definition of a farm in the 2021 Census of Agriculture. Before 2021, a farm was defined as an agricultural operation that produced at least one product intended for sale. The new definition from the 2021 Census of Agriculture says:

A farm is a unit that produces agricultural products and reports revenues or expenses for tax purposes to the Canada Revenue Agency.

The definition of a farm has changed several times in the Census of Agriculture and for that reason comparing the number of farms over time does not offer a consistent picture. It is nonetheless informative of long-term trends.

Below I will map farms according to whether they are classified as crop or livestock farms. The classification follows the North American Industry Classification System (NAICS). According to NAICS, a crop farm is an establishment for which 50% or more of its production consists of crops. Correspondingly, a livestock farm is an establishment for which 50% or more of its production consists of animal production or aquaculture. I take that production is measured on the basis of revenue.

Maps

I obtained 2021 Census of agriculture data from Statistics Canada table 32-10-0231. The table gives the number of farms by NAICS up to 6 digits. I will use only the first three digits to determine whether farms are classified as crop or livestock farms. I plan on mapping farms using finer definitions in futures posts.

The Census data give the number farms by standard geographic areas. The smallest area available, used to make the maps below, is the census consolidated subdivision which is defined as a group of adjacent census subdivisions within a census division. The area covered by a census consolidated subdivision is relatively small in populated areas.

The data do not give the exact locations of farms. For that reason, the figures below are dot distribution maps where each dot represents one farm randomly located within the census consolidated subdivision it belongs to. Given that the census consolidated subdivisions where farms are located are small for most of the maps below, it will not matter that I do not have their exact locations to show geographic concentration. All farms, big or small, are represented within a map by a dot of the same size. Dot size varies across map according to the scale. The dots are partially transparent such that superposed dots become darker, showing a greater concentration of farms.

I locate farms only within the agricultural ecumene. Statistics Canada defines the agricultural ecumene as the areas where the country’s main agricultural activities take place. Limiting the random placement of farms at the intersection of their census consolidated subdivisions and the agricultural ecumene means that farms will be located on the maps in areas where we can expect to find them.

Note that the scales on the maps below will vary depending on the province. This can give of incorrect perception of the concentration of farms across provinces. For that reason, the maps should not be used to compare farm density across provinces. The map for all Canada at the end can be used to compare farm density across Canada.

Newfoundland and Labrador

According to the Census, there were 344 farms in Newfoundland and Labrador in 2021. Farms are located on the coast with the largest concentration around Saint John in the East. There are farms in Labrador but I suspect that most are greenhouses.

Prince Edward Island

Canada’s smallest province had 1,105 farms according to the 2021 Census. The agricultural ecumene covers most of Prince Edward Island. The distribution of farms is quite uniform and similar for crop and livestock farms.

Nova Scotia

The 2,741 farms in Nova Scotia are spread across the province. Most farms are on the western part of the peninsula covered by the agricultural ecumene.

New Brunswick

The distribution of the 1,851 farms in New Brunswick is spotty, with several small agricultural ecumene areas across the province.

Quebec

The Census counted 29,380 farms in Quebec in 2021. Most of them are in the south, in the Saint Lawrence River Valley. There are concentrations of farms in other regions like Lake Saint-Jean and Abitibi.

Ontario

Ontario has the most farms in Canada with a total 48,346. Most of them are in the southern part of the province along Lake Huron, Lake Erie and Lake Ontario. There are farms further north in a few small areas. The map shows the greater concentration of crop farms compared to livestock farms.

Manitoba

Most of the 14,543 farms in Manitoba are in the south of the province that is part of the Canadian Prairies. There are significantly more crop farms than livestock farms.

Saskatchewan

Saskatchewan had 34,128 farms in 2021 according to the Census of Agriculture. The southern third of the province is part of the Prairies, with farms across the area. There are two and a half more crop farms than livestock farms. Two thirds of Saskatchewan farms are oilseed and grain farms, mostly producing canola and wheat.

Alberta

Alberta is the province with the second-largest number of farms with 41,505 farms according to the 2021 Census of agriculture. There is almost an equal number of crop and livestock farms. There is a greater concentration of livestock farms in the Western part of the province. Agriculture goes surprisingly far north with several farms along the Peace River Valley.

British Columbia

The 15,841 farms in British Columbia are distributed across several small regions. The greater concentrations of farms are in the Okanagan Valley, at the mouth of the Fraser River, and on the eastern coast of Vancouver Island. There is also a concentration of farms in the northeast part of the province, in the Peace River Valley.

Canada

Finally, here is the complete picture for the 189,874 farms in Canada. The maps show the largest concentration of farms is along the Saint-Lawrence River valley and by the Great Lakes in Southern Ontario.

Farmland in Canada covered 153.7 million acres in 2021, which might appear large. However, it represents about 6.2% of Canada’s total land and less than 20% of the United States farmland.

Conclusion

I plan on making more maps in the future. I think there are interesting analyses to perform with these data. If you have any suggestion of what you’d like to see mapped, please reach out to me.

The next blog post will likely be in September.

A look at the data

Drivers of Alberta fertilizer prices

In my November 2022 post, I showed that the current price is the best predictor of fertilizer price next month. The importance of corn futures, natural gas prices and barge rate on the Mississippi River was much smaller. With natural gas prices and barge rates spiking in the Fall, I thought it was worth re-assessing the importance of variables to check whether the models still find that these variables matter.

Figure 4 shows the main drivers of fertilizer prices. The best predictor of fertilizer price is the price last month. Other variables are much less important and include the corn price, the natural gas price and lags for the barge rate on the Mississippi River. Compared the November 2022 post, the sets of variables selected by the models are similar but barge rates have lower importance.

Forecasts

I forecast fertilizer prices from plausible scenarios for the models’ variables. The three scenarios are as follows:

• Exchange rates, oil prices, barge rates and rail rates stay at their June 2023 averages thus far,

• The corn price follows its Futures Forward Curve (FFC),

• Three scenarios for natural gas prices (Henry Hub):

  • FFC: natural gas price follows the FFC.

  • FFC -25%: natural gas price declines by 25% from its FFC.

  • FFC +25%: natural gas price increases by 25% from its FFC.

Figure 5 shows the forecasts for the three scenarios. The models forecast fertilizer prices to continue declining. This is not too surprising given the sharp drops in natural gas prices and barge rates. Thus, if current economic conditions hold, we should observe prices for the three types of fertilizer to steadily decline in the next six months.

These forecasts come with a caveat based on recent price movements in the United States. Prices for certain fertilizers were higher in the United States at the end of May. Nonetheless, US fertilizer prices are about 20% lower than they were a year ago.

Conclusion

The forecasts show that conditions are there for fertilizer prices to continue their decline. In addition to what was mentioned above, fertilizer inventories in March, as Figure 6 shows, will support lower prices.¹ Overall, the news is good for farmers as they should see their fertilizer costs continuing their decline in the next six months.