By Brent Gloy
The value of farmland is dependent upon expectations of the future revenue and the costs associated with the land. Expected earnings are difficult to accurately predict. For example, the income generated by farmers started rising well before farmland values shot up. As market participants began to realize that farm incomes might remain at elevated levels for some time, land values increased rapidly. Now, farm incomes are falling and one is left to wonder what the implications of these lower incomes are for farmland values. The answer will likely depend upon how market participants expect revenues and costs to adjust going forward.
Farmland values are a bellwether of the U.S. farm economy. In previous posts we looked at recent farmland value surveys and examined how some of the economic fundamentals for farmland values are shifting. We suggested that if commodity prices moderate to levels slightly higher than currently observed, farmland prices might experience downward movement but would not likely experience a dramatic fall. In this post we make a further examination using long range commodity price projections and examine their possible implications for farmland values.
The USDA compiles a 10 year baseline projection of a variety of farm economic variables. These projections are based upon extrapolation of the current situation over a 10 year period and are described here. While we all know that things will definitely change in unpredictable ways in the future, we thought it would they might be helpful in developing a better understanding the drivers of farmland values. Last year, we also looked at what the lower baseline might mean for farmland values and concluded:
…either market participants are expecting greater returns than implied by the USDA baseline price forecasts or that they are willing to capitalize earnings at lower capitalization rate.
Both scenarios are plausible. The USDA forecasts of long-range prices have not been particularly accurate (see our post or farmdocDAILY’s). However, with prices now trading below the baseline projections it is becoming clear that if these prices hold for several years there will be pressure on farmland values to end their upward march and likely fall.
The approach used in this post follows the same methodology used in the October 20, 2014 post. Basically, the 10 year USDA price, cost of production, and trend yield forecasts are used to establish a return to a hypothetical piece of farmland. We then discount those returns to estimate the present value of farmland from these returns.
Prices, Costs, and Returns
The USDA projections for corn price, variable costs and return over variable costs are shown in Figure 1. Each data point on the line is the simple average of the next 10 year projection. For example, in 2015 the average corn price projection for the next ten years was $3.57. This was $0.15 lower than the projection made in 2014 which came in at $3.72. There are several critical observations that can be made from this chart. The first, and most obvious, is the significant decline in commodity prices projected. Just two years ago the 10 year average was over $4.61 per bushel. For farmland capable of producing 200 bushels per acre this amounts to roughly a $200 per acre decline in projected revenues. This also illustrates how quickly 10 year forecasts can change as conditions evolve.
The second key observation is that USDA is not projecting that the variable cost of production to fall with commodity prices. (In 2014 the projections did not include the variable costs of production so we forecast them ourselves, thus the blip down in 2014.) We expected that variable costs would fall with the projected decline in corn prices.
In 2015 the projections again included USDA’s view on variable costs and they apparently do not expect the same kind of reduction as we did. In fact, they project variable costs at roughly the same level as they did in 2013. Recall that those costs were associated with an average projected output price of $4.61. If USDA’s projection were to come true, variable costs would consume 59% of revenues as opposed to 46% in the 2013 projection. As we will see shortly the persistence of high variable costs would have very negative implications for farmland values.
So what would these projections mean for farmland values? We compared these hypothetical projections to the case of high quality Indiana farmland (roughly 200 bushel per acre production capacity) so that we could compare the projected values to actual farmland survey values. The results are shown in figure 2. We show the projected values capitalized at a low 2.5% capitalization rate (in yellow) and a 4% capitalization rate (in green). The actual Purdue farmland survey values are shown in black. One can see that USDA projection would imply much lower farmland values than currently observed in the 2014 Purdue survey ($9,765 per acre).
Figure 2. High Quality Indiana Farmland Values with 60% Capture of the Return over Variable Costs from the USDA 10-Year Baseline Projections, 1.5% Annual Yield Growth and Alternative Capitalization Rates, 2003-2015.
The 2.5% capitalization rate produces a value of $7,609 and the 4% rate results in values of $4,756 per acre. Both are significantly lower than the $9,765 shown in the Purdue survey. At a 2.5% capitalization rate, the decline is roughly 22%. At a 3% capitalization rate the decline would be 35%. These are big differences from the observed values, with the latter falling into dramatic decline territory. The declines from 2014 to 2015 projections are also very large. For instance the USDA price projection in 2014 combined with our earlier estimate of variable costs and a 2.5% capitalization rate would have put values close to the observed values.
So why did the projections decline so much from the previous year? First, the projected price decline from 2014 to 2015 is $0.15 per bushel which is roughly worth $30 per acre annually. But the persistence of high variable costs is even more significant. In this case they average $0.31 higher than we forecast them last year. The impact of the higher costs is roughly double the impact from the corn price decline. Combined, the hit of higher variable costs and lower output prices results in a reduction in returns over variable costs of $92 per acre per year from the analysis conducted using the 2014 projections.
Wrapping it Up
The price bands in Figure 2 are just theoretical estimates, to help provide insight into the drivers of potential adjustments in farmland values. If (and that is a big if) the USDA projections were to materialize, farmland values would come under significant pressure. A major contributor to this pressure would be the persistence of the high variable costs in the face of lower output prices. USDA’s projections appear to imply that the bulk of the cost structure adjustment would come from farmland values and other fixed costs. In fact, this projection gives us a view of what a persistent margin squeeze would look like for farmland values.
In our view, if corn prices remain low we would expect that variable costs would eventually show some downward adjustment as well. This would alleviate some of the downward pressure on farmland values. It is also important to point out that the analysis does not consider any revenue that would be received from government program payments. As you can see from this handy spreadsheet developed by Gary Schnitkey at farmdocDAILY (you can find it in the middle of this post), these payments may be substantial and would also alleviate some of the negative price pressures on farmland values.
At the end of the day, it appears that declines in variable costs and government payments are not likely to entirely buffer the output price declines currently being experienced. If this is the case, farmland values are likely to come under price pressure. How much depends on where market participants expect output and input price to adjust in the long-term. These expectations can change rapidly. However, if variable costs remain high and output prices remain low downward adjustment is likely. Time will tell.
Photo by Johnny Klemme