Will Data Centers Get Their Own Line on the Appraisal?

PRIMARY SOURCE — the spine of the data center appraisal argument. Fannie Mae states that UAD 2.6 requires appraisers to rate Location and View on a three-point scale (adverse / neutral / beneficial) and describe the driving factors, but that form space constraints limit appraisers to two factors each, with only eight defined View factors and ten defined Location factors, entered as cryptic abbreviations such as LtdSight or Ind. Fannie itself asks how an appraiser should rate a location affected by both adverse and beneficial influences. That is the exact condition a data center creates.

PRIMARY SOURCE — the second half of the "the box already exists" argument. For UAD-required appraisal report forms, the report must rate the location of the subject property and each comparable as Neutral, Beneficial, or Adverse. Freddie defines the rating as describing the overall effect of the property's location within the Market Area on value and marketability — explicitly not a rating of the overall market area. Establishes that appraisers already have authority to flag a data center as an adverse location influence; no new rule is needed.

THE PRECEDENT — the argument that the appraisal profession already built a framework for exactly this class of problem. Compiles studies on how high-voltage transmission lines (100kV+, towers over 45 feet) affect property value. Reported impacts range from -5% to -36% with an average around -20%; encumbered land sold roughly 23% below comparable unencumbered land; effects fell to about -15% when the line ran along a back fence rather than through the parcel. The critical takeaway for the article is methodological, not numerical: the profession concluded that PROXIMITY, VISIBILITY, and CONFIGURATION are measurable and must be distinguished — it did not conclude "power lines are bad." A data center brings the transmission lines with it, so the methodology transfers.

The missing middle term in the argument. Data centers require substations and transmission; this CRA research brief isolates the substation proximity effect on home sale prices specifically. Bridges the transmission-line valuation literature (decades old, well developed) and the data center literature (new, contradictory) by measuring the piece of infrastructure that actually sits closest to the houses. NOT YET READ — pull the effect size before citing. If this brief shows a measurable substation penalty in the same market where the Schar School note showed a data center premium, that tension is the strongest analytical material available on this topic.

COUNTER-EVIDENCE — this is the study that stops you from claiming data centers hurt values. Research note examining data center proximity and for-sale housing values across Northern Virginia's "data center alley." Found homes NEARER data centers sold at HIGHER prices, across single-family, townhome, and condo product; the farther from a data center, the lower the sale price tended to be. Authors and coverage attribute this to confounding: data centers site where power, fiber, transportation, and job markets already exist, which are the same amenities that support home values. Correlation, not a causal endorsement — but it is the single most-cited rebuttal to the property-value objection and must be addressed head-on in any piece on this topic.

THE CONTRADICTION — and the reason "George Mason found X" is a sloppy sentence. This is a DIFFERENT George Mason school (Costello College of Business) than the Waters/Clower Schar School note, and it points the opposite direction: data center arrival associated with SLOWED local home price growth. 47 pages, posted 24 June 2026. Working paper, not peer reviewed — reliability set to Medium accordingly. Citing this alongside the Schar School note is what makes the "genuinely unsettled" framing honest rather than evasive. NEEDS FULL READ: confirm the exact identification strategy, sample, and effect size before quoting any number from it.

READ AND LOGGED 2026-09-14. Fourth paper in the literature, same geography as the GMU Schar School note (Northern Virginia), which makes it the most direct comparison available. FINDING: using Loudoun County (200+ data centers) as the setting, houses within 0.5 miles of an ALREADY ANNOUNCED data center sell for roughly 2.8% less than otherwise similar houses located slightly farther away. This is the key result in the whole stack — it does not contradict Waters/Clower or the Indiana study, it operates at a much tighter radius (0.5 mi vs 1.5 mi vs county). Announcement effect, not construction effect. 52 pages, written 26 Jan 2026, posted Feb 2026, last revised May 2026. Working paper, not peer reviewed.

READ AND LOGGED 2026-09-14. Third paper in the emerging data-center-and-housing literature. FINDING: a null result. Confidence intervals rule out substantial price declines, and results are similar for recent permits associated with AI-related facilities. Author concludes data center development appears to have little measurable impact on local housing prices. NOTE THE TITLE TRAP — "Not In My Back Yard!" reads like an anti-data-center finding and is the opposite; do not cite this paper as support for a negative price effect. Useful as the strongest single piece of evidence AGAINST claiming these facilities tank values. 36 pages, written 27 Feb 2026, last revised March 2026. Working paper, not peer reviewed.

Trade-press roundup useful for one specific figure: an Indiana study using Zillow data on single-family homes within 1.5 miles of four large data centers across four counties, 2021-2026, spanning before, during, and after construction. Found homes near data centers appreciated 42% versus 41% for surrounding areas, and NO significant difference in days on market. That last finding is the direct challenge to the thin-comps / longer-DOM argument, so it must be confronted rather than ignored. Caveat worth noting in any use: the report evaluated a proposed project that was subsequently approved by local officials, so check sponsorship before treating it as neutral.

HALF ONE of the disclosure contradiction. Introduced 15 Dec 2025. Would prohibit non-disclosure agreements concerning data center development, supplementing New Jersey's Municipal Land Use Law. Defines a data center as a facility whose primary services are storage, management, and processing of digital data, housing servers, network equipment, telecommunications, and storage systems. Pair with FL SB 1118, which moves in the opposite direction. STATUS NOT CONFIRMED — introduced is not enacted; verify current posture before publishing.

HALF TWO of the disclosure contradiction, and the sharper of the pair. Committee substitute text declares a public necessity to keep confidential, for a period of time, information about persons locating a data center in Florida held in county or municipal records. Stated rationale: disclosure of siting plans, proprietary confidential business information, or related business activity could injure the person in the marketplace by giving competitors insight into strategic plans, and without the exemption companies might refrain from locating a data center in the state. Read against NJ S5003, the two states are moving in opposite directions in the same year. STATUS NOT CONFIRMED — this is a committee substitute; verify enacted vs. died before publishing, since describing a dead bill as law would undercut the whole piece.

Representative example of the material fact disclosure standard used in most states, chosen because the language is unusually clean. Requires written disclosure before the parties sign a final agreement, covering all material facts the seller is aware of that could adversely and significantly affect an ordinary buyer's use and enjoyment of the property or any intended use. Applies to residential dwellings of four units or fewer. The operative limit for the article's argument: the duty runs only to what the seller is AWARE of — which is precisely what a siting NDA or public records exemption defeats. Most state forms also carry a neighborhood section covering noise, odors, and planned developments.

Inman coverage of NAR's 2026 Data Center Impact Report, combining a member survey with a county-level analysis of 3,200+ counties. Realtor sentiment on nearby residential values is split: 25% report a positive effect, 22% negative; 19% see increased residential demand vs 26% decreased. 38% of respondents have a data center built or in development in their market. Commercial sentiment is clearly positive: 50% report higher nearby commercial values (22% up more than 10%) and 42% higher commercial demand, with industrial (58%) and land (38%) drawing the most interest. Top client concerns are energy costs (61%), water use (56%), environmental contamination (43%), and landscape impact (32%), plus noise from cooling systems and generators. Counties with 10+ data centers show a $431,750 median home value vs $174,500 in counties with none, and 95% decade growth vs 64% — but NAR explicitly frames this as correlation, not causation: those counties already had higher incomes ($89K vs $64K), more college graduates (41% vs 22%), younger populations, and cheaper electricity before the buildout. Lawrence Yun: "There is no single data center effect." NAR's practical guidance is to assess impact property-by-property, not market-by-market.