- About MassCEC
- About Clean Energy
- Catalyst Program
- Commercial-Scale Biomass Boilers
- Commonwealth Home Heating and Cooling
- Commonwealth Hydropower
- Commonwealth Organics-to-Energy
- Commonwealth Small Pellet Boiler Program
- Commonwealth Solar Hot Water
- Commonwealth Solar II
- Commonwealth Wind
- Community Energy Strategies
- District Energy
- Geothermal Heating and Cooling
- Green Workforce: Energy Efficiency
- Investments in the Advancement of Technology
- Investments in Job Creation
- Mass Solar Connect
- Massachusetts Clean Energy Internship Program
- Massachusetts Israel Innovation Partnership
- Marine Commerce Terminal in New Bedford
- Pathways Out Of Poverty
- Production Tracking System
- Solarize Mass
- Woodstove Change-Out
- Workforce Capacity Building
- Wind Technology Testing Center
Relationship Between Wind Turbines and Residential Property Values in Massachusetts - Responses to Questions
Researchers Responses to Questions Received During the Webinar
“Relationship Between Wind Turbines and Residential Property Values in Massachusetts” Jan. 22, 2014
Q: Were all the sites equipped with only one turbine?
No. there were also sites with two, three and 10 turbines, although most have only one turbine (~ 77 percent of the 26 facilities). It should be noted, though, some facilities are located quite near others (e.g., Falmouth Wastewater one and two are 0.26 miles apart; and, Mass Military Reservation one and two are 0.18 miles apart).
Q: Why was the study restricted to turbines above 600 kilowatts (kW)?
In Massachusetts, there is one installed turbine in the 100-600 kW range (250 kW). Therefore this was recognized as a distinct gap in Massachusetts. Additionally, 100 kW turbines typically have faster development timelines which reduce the number of transactions that could occur during each of the separate periods analyzed in the study. While some turbines in the 600 kW range were included in the study, most turbines, especially those brought online in 2009 and later were much larger (closer to or above 1.5 MW).
Q: Of the similar studies discussed by the authors, all but three found no impacts. Two of these three studies focused on non-U.S. countries (Canada, Germany). Is it possible that there was a difference in findings due to cultural differences between the various countries?
Possibly, but it is unclear without a comparative study between the countries.
Q: Please give an example of what the difference in coefficients means?
In the prior announcement period (more than two years before turbines announcement) property values for homes within half a mile of where turbines ultimately would be located, and after controlling for the suite of home, site, market and location variables, were estimated to be -5.1 percent lower than those homes outside of half a mile but inside of five miles.
After controlling for this pre-existing difference, and, again, after controlling for the suite of home, site, market and location variables, home prices, homes that sold after turbine construction and operation and were located within half a mile of the turbines, were 0.5 percent more in value than homes outside of a half mile and within five miles, and which sold in the same period.
Q: The p-values for some of those coefficients were over 0.5 and yet were marked as statistically significant. Can you clarify?
For each coefficient there is a corresponding standard error (see the full set of results in the appendix) that defines the confidence interval around the estimate (in our case the 90 percent confidence interval). For coefficients with larger numbers of transactions, the standard error is often, but not always, smaller. For example, the number of homes near highways is very large as compared to the number of homes near prisons, and the standard errors for the highway coefficients (0.01) are much smaller than that for prisons (0.02-0.09). A smaller standard error will allow a smaller coefficient (i.e., effect) to be statistically significant, or, put another way, a smaller standard error will allow one of two coefficients of the same size to be statistically significant while the other is not.
Q: Was there a difference between the larger Berkshire Wind development and the smaller developments?
This was not explored, only the average effect across all facilities was calculated.
Q: Were enough MA "disamenity" structures -- prisons, e.g., -- built during your analysis period, or were you drawing on other sources for that part of the analysis?
All disamenity and amenity spatial variables, except for wind facilities, were assumed to exist before the study period began. All data on these variables was collected from MassGIS as map layers, from which distances to homes could be calculated.
Q: Is the slide presentation available as a pdf?
Yes, the presentation may be accessed here.
Q: What if anything does this study tell us about how larger projects (15+ turbines) with larger viewsheds would impact property values?
The study examined how homes prices are affected for those that are near (within half a mile) of an operating turbine. For homes in the vicinity of larger projects and are within half a mile of single turbines in those projects, the study might be applicable. Regarding homes with views of turbines, this study did not examine those effects directly. That notwithstanding, it was assumed that most homes within half a mile would either be able to see the turbines from the home, or would be in regular visual contact with it as a homeowner came and went from the home. It was therefore also assumed that distance could be used as a reasonable proxy for view of turbine (as well as a host of other factors such as sound and flicker that are experienced within close distance of turbines). These assumptions were based on direct view impacts that were explored extensively in Hoen et al. 2009, where they found a very high percentage (90 percent) of homes within a close distance (3000 feet) could see the turbines. Moreover, this 2009 study found no evidence of an impact of view of turbines on sale prices, further lessening the need to study those effects in Massachusetts.
Q: What is the closest distance [from a home] to [a] turbines?
Approximately 550 feet, but many of the homes examined within half a mile were located close to the half mile cutoff.
Q: Can you tell whether there is a greater impact on home values located near wind turbines constructed in more “upscale areas”, or areas in which there are fewer other "disamenity"?
This was not explored, only the average effect across all facilities (and therefore all neighborhoods) was calculated.
Q: Is there a difference between property values downwind of the prevailing wind direction, compared to those on the upwind side?
This was not explored, only the average effect across all facilities (and therefore all wind regimes) was calculated.
Q: What database did you use to find the wind turbine installations?
The wind turbine data were provided by MassCEC and corroborated via visual inspection via satellite imagery in a GIS.
Q: Was the proximity effect limited to simply distance, or was visibility of the turbine also taken into account? In other words, was visual blocking (e.g., trees, hills, or other buildings) considered in the study?
These view of turbine characteristics were not taken into account, only if the home was within half a mile of a turbine, which, as discussed above, was assumed to serve as a reasonable proxy for view of turbines.
Q: How many post-construction transactions were from Hull, and how many were from communities where noise and flicker are nuisances?
5,069 post-construction transactions out of the 21,907 were located near either Hull 1 or Hull 2. Of those, 88 were within half a mile out of 235 in total. Results are robust to the removal of all transactions near these two turbines.
Q: What was the year timeframe of study and were outside economic conditions accounted for? Did you consider controlling for impacts that the recession may have had?
The study period was January 1, 1998-November 23, 2012. The recession and housing market bubble were accounted for in the year fixed effect variables (see full results in Appendix of report). Although not included in the report, results are robust to allowing different inflation/deflation adjustments occur at the county level (by interacting county with sale year).
Q: How can 2003 be considered announcement when it was two years before even a feasibility study was done and the 'expectation' was a single 660 kW, not three 1.65 MW turbines in close proximity?
The question appears to refer Falmouth Wastewater 1, which had an announcement date of April 1, 2003. Announcement date was assumed to be the first instance where the project was discussed publically. This might include a permitting hearing, the release of a request for proposals to identify a construction firm, or a newspaper article. For Falmouth Wind 1, in April 2003 the Energy Committee sent a letter to abutters of the wastewater treatment plant indicating that the University of Massachusetts Amherst would be installing a meteorological tower to measure the wind speed and that the site might be a viable location to install a wind turbine.
While the announcement date preceded the feasibility study this would not influence the results. The authors were concerned that the transactions occurring more than two years before announcement (“prior-announce”) NOT be affected by the eventual build-out of the turbines. If an earlier announcement date were used, as the question suggests, it would have the effect of decreasing the number of transactions captured in the “prior-announce” or control period, but increasing the likelihood that they were not impacted by the eventual announcement and construction of the turbines. Additionally, while there are currently three 1.65 MW turbines installed in Falmouth, each were developed as separate projects and as a result have different announcement, construction, and commissioning dates.
Q: It appeared as though a number of projects were commissioned in 2010-2012. During this period, however, there were probably very few sales of close-proximity homes of the turbines. Might this small number disguise an increasingly negative view on the part of the market towards close proximity to turbines?
A majority (approximately 60 percent) of the post-construction transactions within a half mile of operating turbines occurred in the 2010-2012 period (and 80 percent from 2008 and later) and therefore likely would have captured those effects. That notwithstanding, if an increasingly negative view of turbines prevails and dominates an otherwise positive view, it would be expected that prices would reflect that. It has been shown that home prices can reflect/serve as a proxy for public sentiment towards changes to the environment. In the period studied though, it does not appear that a dramatically negative view of turbines has been priced into the market.
Q: Does the pre-existing -5.1 percent value for homes near future wind turbine site, already control for disamenities like landfills and highways? If so, what else would cause this -5.1 percent value?
Yes, the prior announcement price difference for homes near turbines does take into account for the whole suite of amenity/disamenity variables, as well as the other controlling variables. It is assumed that the -5.1 percent difference reflects effects of commercial uses such as waste water treatment sites, which were not directly controlled for in the model.
Q: Did you also look at whether the properties that transacted could see the turbines from the property? Or, if the turbine was blocked by topography, structures, or foliage how might that affect results?
The views from the homes were not factored into the analysis. It is assumed that many homes within half a mile could see the turbines from the home, or could do so on the way to and from the home site and therefore were aware of its location.
Q: Could this study help with understanding sound issues?
Correlation between average sound impacts and sales prices were not tested.
Q: How many of the houses in the study within a quarter mile of the turbine were also near another disamenity (i.e. a landfill), vs. houses that weren't near other major disamenities?
Eleven transactions within a quarter mile of a turbine that occurred in the post-construction period (out of 38) are also near a landfill, but another 110 transactions are near a landfill but are not within a quarter mile of a turbine, and therefore serve as controls to those that are close. The a quarter mile results are robust to the removal of the landfill variable from the model. Similarly the base half mile results are robust to the removal of land fill, and, for that matter, all of the amenity/disamenity variables.
Q: What did the study cost?
MassCEC spent approximately $75,000 on this study.