New wind farm property value study offers grist for both sides
Human impacts, News, Wind turbines Add commentsA new study of property values in the vicinity of a large wind farm in Illinois provides reinforcement for both sides in the debate. I first saw mention of the study in an American Wind Energy Association press release that touted its consistency with previous studies that found no significant price impact in homes near wind farms. After downloading and reading through the report, I find that the results do indeed match previous studies, though in my reading the results are subtler than the overall averages suggest, just as they were with the big DOE-funded study that came out about a year ago.
The new study, entitled “Wind Farm Proximity and Property Values: A Pooled Hedonic Regression Analysis of Property Values in Central Illinois, 2010” used complex multi-factor statistical analysis to compare many factors that affect the sales price of a home (that’s lay-speak for the “pooled hedonic regression analysis with difference-in-differences estimators” that were used). The bottom line is interesting and potentially reassuring: since local real estate agents were sure there was no effect beyond 3 miles, the study looked only at homes closer than that. Among all those homes, there was no decrease in sales value after the wind farm became operational. There was a noticeable decrease in prices in the “post-approval, pre-operations” phase, when local fears of negative impacts at homesites were most pronounced. But after that decline, prices of homes sold after the wind farm became operational were HIGHER than those prior to project approval. When looking at the final data charts, it’s apparent that there was, in fact, a tiny negative effect of proximity to the turbines, but it was dwarfed by other effects, including the positive effect of home size and lot size, and the negative effects of proximity to the railroad tracks. The negative effect of proximity was about the same as the increase in value if a house has a garage, or an extra tenth of an acre of land, or more trees than other lots.
However, as with the DOE study, when you drill down and look just at the closest home sites, there does appear to be more of a negative impact on sales prices once the turbines are operational. Within a mile and a half, it’s clear that there were some longer-term price effects. The mean price of homes within a half mile, and between a half mile and mile, continued to decrease when the wind farm became operational, in contrast to homes further away, where generally fears of impacts dissipated and prices rose once more. Within a half mile, mean post-operational prices were roughly 12% lower than those prior to wind farm approval, and between a half mile and mile, 8% lower. Even these figures are not without subtlety, however: it turns out that the median price in both of these distance ranges did rebound. For those of you who’ve become foggy on medians and means, the mean is the average price, and the median is the “middle” of all the prices considered. In this case, there were 5 sales in the half-mile range and 6 in the next distance group, so we can say that there were 2 sales lower than the median price and 2 sales higher, with one or two at the median. Since the median did not go down (in fact, it rose a bit), what this probably means on the ground is that one or more of the homes that sold for below the median at each distance range probably went down in price quite significantly (in the order of 15-25% or more), in order to drag the mean down so much.
Interestingly, at the mile to mile and a half range, the mean/median results reverse: this is the only range at which the median did not rebound significantly after the post-approval fear stage, and remained a bit more than 2% lower once operations began, but the mean did in fact rebound (suggesting a few of the 11 homes sold in this range must have increased in value quite a bit to pull the mean up).
The author concludes with a list of “particular circumstances (that) contributed to” the general lack of property value decreases. These are crucially important to bear in mind for anyone who may want to extend these results to other locations. Among the key factors in this location were very little local opposition to the wind farm and “Good Neighbor Payments” to nonparticipating landowners for the life of the project (which certainly could factor into home prices). Also, the wind farm developer was “very upfront with area residents and explained the wind farm was going to have a significant impact on the area;” so, no false expectations.
Most important of all, in my reading of the local situation, was that “There had not been much population growth in the immediate area surrounding the wind farm over the past century,” with very few “pocket farms” in the immediate area. Those of you familiar with my analysis of community annoyance rates (see this NEWEEP presentation) will recognize what this means: in this community, the vast majority of residents have a “place identity” that stresses working the land, with very few newcomers who have chosen to relocate to this community (in retirement or as cyber-commuters) in search of a peaceful rural refuge.
The bottom line appears to be that in a community with such high-quality relationship with the developer and few residents primarily seeking peace and quiet, only a few properties within a mile were significantly impacted by the arrival of this wind farm. It is also worth reaffirming that the biggest price hit occurred prior to operations, when fears of impacts are widespread, and on-the-ground experience is still not a factor. While this is reassuring news, we need to be mindful of the subtleties and particulars of this situation. Clearly, some properties were negatively impacted, even in this predominantly supportive atmosphere. In locations where these balancing conditions are not in place, the negative impacts within a mile or perhaps a bit more could be significantly more noticeable.
Download the report here or here. See page 49 for means and medians at various distances; pages 78, 82 for relative effect of proximity; page 84 for local circumstances.
December 8th, 2010 at 6:37 am
I live near Tehachapi, CA. Your study is on flat land, not along ridges, in canyons or in the hills. The study is deeply flawed and misrepresents the reality of the post implemented noise pollution, severe light pollution of the night sky and permanent destruction of the habitats and endangered raptors like the California condors. At least in eastern Kern County, the turbines will not replace other energy sources. It is so developers can build on every inch of the Mojave Desert, by the way, is the ONLY desert in the entire United States. Here is a recent article in on Bakersfield Californian that substantiates this.
http://www.bakersfield.com/news/local/x867865420/ROADBLOCK-Transmission-shortage-threatens-to-halt-Kern-s-alternative-energy-boom
This whole renewable energy movement is nothing but a land grab by the robber barons that that are pushing right over the properties of the people for their own use and fortunes…dirty hands that stink like old fish.
December 16th, 2010 at 7:58 pm
[…] owner who objects to living near a turbine. While large studies of property values have found no clear correlation between distance and price, they may not be as definitive as the industry would like to believe — […]
January 12th, 2011 at 2:26 pm
People are looking at the executive conclusion without drilling into the study, because, let’s face it, some knowledge of statistics is needed to determine whether or not the conclusion is trustworthy.
There are other statisticians who have commented ont the defects of the hedonic analysis. But again, it needs translation for normal readers.
Primarily, it is this: improper sample selection and pooling will inflate the variance upon which the regression is based. Select the conditions broadly enough and everything will become insignificant. To put it bluntly; selection bias can create a self-fulfilling analysis. Critiques of the study have pointed this out, and the very nature of the summary reads as more of a disclaimer then statement of fact.
The tip-off of poor modeling is when property values are determined by a boots-on-the-ground appraisal. This is from a review by an appraiser for Adams County Illinois.
“Real estate sale data typically reveals a range of 25% to approximately 40% of value loss, with some instances of total loss as measured by abandonment and demolition of homes, some bought out by wind energy developers and others exhibiting nearly complete loss of marketability.”
[ http://www.scribd.com/doc/32984818/McCann-Appraisal-LLC-Written-Testimony-Re-Setbacks-Property-Values-June-8-2010 ].
When a model does not predict reality, science rejects the model.
March 23rd, 2011 at 7:01 pm
Does this study include homes that did not sell?? If so, would that affect the conclusions. I have tried to muddle through this study, but have found it much too difficult for someone not “into” statistics.
May 5th, 2011 at 1:51 pm
[…] noted in AEI’s previous coverage of property values research, while there is little evidence of decreased property values due to […]