Vyn and Property Values

In January 2014 Dr. Richard Vyn, a professor at Guelph University in Ontario, published a study concluding that wind turbines have no significant effect on surrounding home prices.  The report was published formally in the Canadian Journal of Agricultural Economics in September 2014.  It didn’t attract much attention until fairly recently, when a spate of news articles appeared.  They pretty much all had the same headline, saying that a new peer-reviewed study showed wind turbines had no effect on home prices.

Vyn’s study applied a hedonic approach to studying the issue, much in the same way that previous academic studies have done.  I’ve studied many of them at some length and posted my critiques on this site.  It is important to note that none of these authors are real estate professionals.  Many of them are sponsored by or associated with an institution that has skin in the wind industry.  In every case they use statistical techniques built on questionable assumptions and come up with the results that please their sponsor/associates.

My quick look at Vyn when it first came out told me his report was more of the same, so I didn’t waste much time on it, never bothering with a posting.  But the recent media flurry got me to look at it again, and I can confirm it really is more of the same.


Vyn centered his study on the Melancthon project in Melancthon Township, Ontario.  He got sales data from the Ontario assessor, MPAC, for a total of 11 townships surrounding the project.  He looked at two basic effects of wind turbines on the neighbors – distance from the nearest turbine and visibility of the turbine(s).  He split the data up into farm properties and residential properties and analyzed them separately.  He ran a series of regressions that produced a series of sloped lines on graphs, i.e. sales prices vs. square footage.  The slopes of what the paper was primarily interested in, i.e. sales prices vs. distance to a turbine, were all close enough to zero that they “suggest that these wind turbines have not significantly impacted nearby property values“.

It isn’t until you read through the entire 26 pages that you discover that Vyn, to his credit, has what appear to be significant reservations about his study.  Examples:

  1. Page 11/375: These relatively low numbers [in close proximity to turbines] of post-turbine period observations, which may impede the ability to detect significant effects, represent a potential limitation of this study.
  2. Page 24/388: However, while the results indicate a general lack of significantly negative effects across the properties examined in this study, this does not preclude any negative effects from occurring on individual properties.  [Followed by a segue to Lansink]
  3. Page 25/389: While surveys have indicated that residents often perceive that the existence of wind turbines within their viewshed will reduce the value of their property, such perceptions have not often been corroborated by analyses of sales data, perhaps due, in part, to data limitations with respect to sales in close proximity to turbines.
  4. Page 25/389: The existence of limitations in the analysis undertaken in this paper should not be overlooked.

However, overlooking these self-acknowledged limitations is exactly what the media and proponents will invariably do.  One has to wonder how many reporters just read the abstract, not wanting to pay for the entire report.


If you were going to try to figure out if wind turbines actually lowered house prices, how would you go about it?  I think most of us would start by setting up a baseline of house prices at different distances from a project at a point in time well before the project was on anybody’s radar.  Then we’d take the current prices and see how they compared.  MLS and MPAC both have this type of historical data, albeit I don’t know how granular it might be.  Regardless, you’d think this would be a starting point.  And this is exactly what Lansink, a real estate professional, did.  But for reasons that we can all speculate upon this is not what academics (i.e. Vyn) and other government-supported researchers (i.e. Hoen) do.

It would have been nice, for example, if Vyn had published the mean sales prices of homes in each of the 11 townships he studied, unmodified by all his statistical manipulations.  While Melancthon Township and the Melancthon project don’t overlap entirely, I’d think they overlap enough for an effect to appear, if there was one.  In Vyn’s particular case, he used GIS to look at sales within different distances of the project itself, so he could have published those averages as well.  But he didn’t, and I have to wonder why.

The most charitable answer is he thought other factors were present that would distort these averages, things like different ages and sizes of houses close to the project when compared with those further away.  Too bad he doesn’t discuss what those distortions might be.  Or perhaps since other academics have been using hedonic techniques he felt his study wouldn’t be looked at by them if he didn’t also.  Or perhaps he did look at those averages and didn’t like what he saw.

The paper is full of detailed and barely understandable (for me, at least) talk about things like spatial correlation, continuous specification, multicollinearity and so on.  It was obviously written for an academic audience and is pretty much an academic exercise.  But what is appropriate in an academic setting isn’t necessarily appropriate for the general public, and Vyn had to know that his study would be used as a bludgeon by wind energy proponents against that public.

It seems that academics and policy wonks tend to think in grand terms.  This type of study seems to imply that property value decreases aren’t worthy of policy consideration unless they are widespread.  And even though both Hoen and Vyn are careful to note the possibility of individual property losses, their main thrust is to dwell on the larger picture and that is certainly what the media and industry care about.  Individual tragedies be damned, no matter how many there are.


So right off the bat I’m leery of this type of study.  But there are other problems with how this study was put together and some of the rather basic assumptions that went into it.  I mentioned upstream that Vyn looked at distance and visibility.  Academic property studies that look at visibility are common, generally claiming that “how the turbines look” is an important part of the opposition to them.  Certainly for the larger public visibility is an issue – after all they can be seen for miles.  But visibility as an issue for house prices pales in comparison with noise.

Vyn mentions noise in passing: “While earlier literature also examined the issue of noise, the reduced emphasis on the noise disamenity appears to reflect improvements in turbine technology (Moran and Sherrington 2007).”  Is he kidding?  Using a reference from 2007, when the sizes of turbines were a fraction of what they are now?  Is he so encapsulated on his campus he hasn’t seen, for example, CBC’s (hardly an opposition organization) Wind Rush?  And how did Moran know there were improvements?  He doesn’t say, it is simply an assertion, one that Vyn has carelessly adopted.

I’ll be charitable and say I think the reason so many academics focus on visibility goes back to their thinking on grand terms, where visibility has the potential of affecting large numbers of homes, while noise has a much smaller radius.  Also, it is difficult to write impressive-sounding studies when you’ve got a handful of properties that are rendered uninhabitable (and quite often unsellable).  You can’t do multiple hedonic regressions with just a few points; you’re forced into (gasp!) doing comps and repeat sales, just like Lansink did.  It sounds much better to have thousands of data points, regardless if they convey the reality of what the neighbors are faced with.

And Vyn certainly had lots of data points – 5414 residences.  And, true to form, very few within 5 km of a wind turbine – 123 (I think).  As an aside, another benefit of using visibility is making the radius so large that any house price effects are greatly diluted, often into insignificance.  Which typically makes the sponsors/associates very happy.

But even that wasn’t good enough for Vyn.  He used Melancthon plus the surrounding 10 townships.  And Ontario townships are very large.  Some of his sales are 50 km from the project.  He mentions using them as a control group, but I see no sign that he ever did so in his analysis.  A picture:


The yellow line defines his area.  The yellow push pins are the approximate extents of the Melancthon project, while the red line is the approximate 5 km boundary around the project.  That area covers about 300 sq km, or roughly 7% of the total area he studied.  The sales within that area, 123, represent just 2.3% of the total.  Even considering that Melancthon is the least-densely populated township of the 11, that still is quite a low rate.  Vyn was right to comment about how this might skew his conclusions.

Picture one of Vyns ‘sales prices vs. distance’ regression lines stretching from the center of the project to the edge of his study area.  To show an effect, that line would have to be sloped off of horizontal.  Imagine how difficult it is to slope an otherwise flat line when only 10% of it (maybe 5 km out of 50) is subject to the effect you are looking for.


Sadly the media, government, proponents and industry will all use this study to justify the continuing assault on rural Ontario.  Almost all will do so without actually reading the study, not to mention taking the time to think about what it really says and the basis upon which it was written.  When you have an agenda truth isn’t important; truthiness is.  And this report supplies truthiness in spades.  Whatever truth it supplies is buried deeply enough to not bother wind’s supporters.


Vyn studies, sorry I couldn’t find a free copy of the entire report.

For samples of media reports:

And the industry:

My previous postings:



One thought on “Vyn and Property Values”

Leave a Reply

Your email address will not be published. Required fields are marked *