Average Minnesota farm real estate sales prices just keep on climbing (figure 1). This despite low output prices, rising input costs, and continued uncertainty about the future of federal subsidies. Sales price increases were seen in all parts of the state except in the northwest.
In this annual sales price summary, I can provide only an overview, some cursory analysis, and--as always--a few opinions. I'll not bore you with text that simply repeats what's already shown in the charts. Instead I'll spend some time discussing how land transaction data are recorded, adjusted, and employed.
I think it's useful to go through some administrative and procedural details to further our understanding of what these data are and what they are not. If this prospect makes you say, "Just show me the data, Steve," then you can stop right after the sales summary section. Or, if your impatience knows no bounds, go straight to the Minnesota Land Economics (MLE) Web site at http://apec.umn.edu/faculty/sjtaff/landdata/index.html and start working the numbers yourself.
Assessors are required to report initial assessments in late fall, based on sales data to date. That's why the data are reported on a "record year" basis: these are the sales that were, presumably, available for assessor scrutiny at the time the initial estimated market-values are calculated. Final values are set by summer, to be used in the succeeding tax year. So, for example, sales made in late 1999 are used by assessors to set initial values for January 2001. These estimates are adjusted in spring 2001, finalized in summer 2001, and then used for tax purposes in 2002. The adjusted record year 2000 sales data were therefore not available until April of this year.
Figure 2 shows the distribution of all farmland sales in 2000. The bulk of the sales lie between $500 and $2,000 per acre. I excluded a small number of sales that exceeded $5,000 per acre as well as those involving parcels of land less than 20 acres in size. Both were excluded as not being plausibly "agricultural"--despite their designation on the Certificate of Real Estate Value (CRV) as "agricultural" land. (Although excluded from figure 2, these data are included in the MLE Web site data.)
Even though MAE readers and MLE Web site users can view the full distribution of sales prices, most still ask for a single number that somehow captures the story behind the figures. Obviously, for a set of sales that span such a wide range in prices, any single number fails completely to accomplish this end. Movements in averages, while arithmetically correct, usually fail to tell the entire story.
The particular average I use in this article is a location- and size-weighted mean (table 1). In last year's farm real estate report (agecon.lib.umn.edu/mn/mae699.pdf), I discussed the usefulness of such weighting as well as the desirability of examining sales data at the smallest geographic scale possible.
| Record year | Number of sales | Acres sold | Average price* |
|---|---|---|---|
| 1996 | 2,504 | 263,728 | 936 |
| 1997 | 2,641 | 296,803 | 1,039 |
| 1998 | 2,724 | 303,968 | 1,113 |
| 1999 | 2,212 | 235,359 | 1,196 |
| 2000 | 2,258 | 250,979 | 1,222 |
* Location- and size-weighted per-acre mean
Figure 1 compares the movements of actual sales price averages with those two other estimates of land value--the United States Department of Agriculture (USDA) annual state estimates (based on a farmer opinion survey) and the average assessor estimates (the location- and size-weighted mean estimated market value). The University of Minnesota sales prices averages are location- and size-weighted means.
The fact that all three (somewhat) independent estimates of farmland real estate values shown in figure 1 move in lockstep adds credence, I believe, to the conclusion that, on average, farmland values really are increasing in Minnesota.
Geographic variations in real estate values for the past 11 years are shown in the box-and-whisker plots of figure 3. (District boundaries are shown in figure 4.)
The range of sales prices for each district for each year is shown by the endpoints of the vertical lines. The ends of each box show the prices at which 25 percent of the sales were higher (or lower). The median is indicated by the horizontal bar within each box. So, for example, the median Central district farmland sale was about $1,200 per acre, with 25 percent of the sales lower than $750 and 75 percent lower than $1,850 per acre.
In previous years' reports, I've noted the wide variation in average price movements among districts. Such differences were accentuated in 2000 by the continued climb of values in the South East district combined with the continued stagnation in the North West district (figure 5).
When a Minnesota property is sold, the transaction details must be recorded at the county courthouse on a form called a CRV. On it, the seller attests that such-and-such a property was sold to so-and-so on a certain date for a specific price. Other information about the property (its size, soil characteristics, prior year's estimated market value) is often entered on the CRV as well.
Frequently, the per-acre prices that underlie this article and are also shown in the MLE Web site are not the prices entered on the CRV. Long before a land sales figure enters the official data base, it has been passed through an array of filters and adjustments designed to make comparison among transactions more meaningful and more reliable.
There are many possible slips between an ownership change and data analysis. Of course, there is always the chance that simple recording errors are made. For example, numbers may be miscopied from bills of sale onto the CRV, or into a computer file, or into a spreadsheet.
There is also a chance of misrepresentation. The person who fills in the CRV might have a reason to understate or overstate the actual sales price--perhaps to avoid a tax. This, of course, is illegal, but, as any courthouse veteran can tell you, it occasionally happens. Not every sale receives further processing. Local or state officials remove from subsequent analysis any sale not deemed "arms-length," because it was sold, for example, to a member of the seller's immediate family. Or, a sale might be pulled because the new buyer intends to convert the land to a non-agricultural use.
After this filtering, sales prices are frequently adjusted to make comparison among sales more appropriate.
First, to expunge the effects of inflation, sales prices are deflated by an officially reported rate to January 2 of the year in which they were recorded. This "adjustment for time" is fairly minor in years (like the past decade) where inflation has been low.
The second adjustment is "for terms." Not all farm real estate sales are for the full property. Some are made through a contract for deed, an arrangement that allows the buyer to pay a certain amount now and other amounts at stated intervals. Until the final payment is made, the property remains in the possession of the seller--even though it has been "sold." Because the full payment schedule is entered on the CRV, the Department of Revenue can calculate a present value of the initial and subsequent payments at an official discount rate. This becomes the official sales price of the property, regardless of what the buyer and seller had in mind when they sealed the deal.
Adjustments don't end with a time- and terms-adjusted sales price, honestly reported and accurately recorded. In most cases, users of the data are interested in per-acre prices, not per-parcel prices. That means some chosen total price must be divided by some total acreage. But which price? Which acres? Should we use the total price or should we first subtract out the value of buildings, personal property, ancillary property, or machinery to get closer to the "true" land price?
In this article (and on the MLE Web site), I choose to follow conventions established years ago in Minnesota. I report the time- and terms-adjusted total sales price, minus the value of personal property, divided by the entire acreage of the parcel. That's why, when I'm being careful, I speak of the average price of farm real estate, not of farm land.
The sales reported here are only those recorded between October 1, 1999, and September 30, 2000. These "record year 2000" sales are so bundled because of the way real estate transactions are used to help local assessors value land for property tax purposes.
Strange as it may seem, the Department of Revenue does not collect sales data merely to satisfy the data cravings of University economists like me. No, statewide sales data are collected principally to create statistics that are used to "equalize" property tax valuations across county boundaries.
Each year, county assessors are required to assign an estimated market value (EMV) to each of the thousands of real estate parcels in the county. The estimate is supposed to be based on an examination of similar properties that were actually sold recently. (The combined valuations for each township, city, or county are the source of the Land Values--in contrast to the Farmland Sales--data on the MLE Web site.)
Because every county has its own assessor who uses largely independent valuation procedures, there are inevitably discontinuities across county lines--even for adjacent properties. Farmer Brown wonders why Farmer Olson's land, just across the fence line in the next county, carries an assessed value that is lower by $200 per acre.
The state has created an equalization procedure that is supposed to smooth over such discontinuities. Assuming that nearby properties really would sell for similar prices, any observed difference in assessed values for otherwise similar properties is presumed to be evidence that one or both of the assessors is either undervaluing (that is, assigning an EMV that is too low) or overvaluing properties.
To test this, the state calculates a sales ratio (the EMV divided by sales price) for every property sold in a particular area. If an assessor systematically undervalues properties (shown by sales ratios that are consistently lower than some threshold), the state might demand the EMVs in that jurisdiction be uniformly raised, to better accord with what is thought to be "true" market conditions.
We can see for ourselves how close the final assessor estimates are by comparing actual sales prices against the previous year's estimated market values for the same property (figure 2). Each point in the figure represents one sale. For example, the rightmost point is for a property that was estimated to have a value of $4,900 per acre, but actually sold for only $2,900 per acre. While some of the estimates are obviously way off (like this example), the bulk are pretty close. In most cases, the EMV was lower than the sale price, but in a neatly predictable manner. A simple one-variable regression model, shown as the straight line in the figure, accounts for nearly 75 percent of the observed variation in farm real estate sales prices.
What accounts for the ever-onward-and-upward movement of average farm real estate prices in Minnesota? We need only to round up the usual suspects, most of which I have discussed at length in previous issues of MAE. These include 1) perennial farmer optimism about future crop and livestock prices, 2) expected extensions of federal farm subsidy programs, 3) continued favorable local property tax treatment for farmland, 4) the desire of some farmers to increase the size of their current operation by buying adjacent farmland, 5) the desire of some non-farm buyers to use land as a hedge against inflation, and 6) inflation itself.
An additional suspect that we need to add is the increasing prominence of location even in rural land markets. We simply can't explain current price levels on the basis of income potential (including subsidies) and speculation potential alone. Clearly, where the land sits with respect to job centers and what it looks like is influencing the price buyers are willing to pay for a particular parcel of land.
As always, I caution potential land buyers and sellers about reading too much into the average land prices reported here and elsewhere. If you've got land to sell or if you have a hankering to buy land--look before you leap. The financial stakes are too high for casual empiricism. Hire an appraiser. Talk with your spouse. Check your finances. Think about the children. Be careful out there!
Steven J. Taff is an associate professor and extension economist with the Department of Applied Economics at the University of Minnesota.
Return
to Minnesota Ag Economist Newsletter Index Page
University of Minnesota Extension Home
Page