HONEY, STOP THE CAR! 2015 UPDATE ON AD LANGUAGE

by John Hill

Last summer, Barnett & Hill analyzed the role ad language played in home sales as a redux of the chapter How Is the Ku Klux Klan Like a Group of Real-Estate Agents? from Steven Levitt's best-selling book  Freakonomics.  After publication, we visited with several house flippers and home builders who asked us take it one step further and place more emphasis on the role construction choices play in final home price to inform their choices as a contractor.

To oblige, this year's recalculation includes variables from Levitt's original list, but mints new variables representing construction and finish choices available to a home builder or remodeler.   The purpose of the study is to tabulate the dollar value of a contractor's choices as a share of overall price.  

The data set looks at home sales in Dallas and Taylor counties over the last 365 days.  The Dallas data set held 23,195 sales and the Taylor county data included 1,725 sales. Dallas county sales were considered because it is presumed that the Dallas market captures trends not yet seen in Abilene that might ultimately become must-haves for Taylor county homes. 

The attributes are of the following classes:

The basics:  The most obvious predictors of price are square footage (sf) and age.  The relationship between sf and price is the most obvious and predictive relationship - price grows with sf.   For age, we included both age and age squared - the thinking being that homes initially lose value when they age, but, at some point the trees provide shade or they become historic and the decline in price is blunted as the property matures.

Finishes:  "Pine", "maple", "oak" and "white" capture the stain or paint coats associated with the sold home.  We didn't make a distinction as to what oak described - perhaps the windows were milled with oak, the kitchen cabinets had an oak stain, or, possibly, the wall paper had an oak motif.   Our sole purpose was to assign value when the word "oak" was invoked.  The same is true of terms like granite, stainless, Corian, quartz or marble.

Floors and roofs:  When the property is listed, a variety of flooring and roofing choices can be coded by the realtor.  We also used this information to code the number of flooring transitions in a property with the idea that the visual clutter posed by multiple floor finishes lowered value.

Configuration:  Home builder choices include the number of covered parking spaces, bathrooms, stories, or whether or not the home needs a fireplace (and how many).  Siting of the house on a creek bank lot or within a school district also play a role in the home's value.  In Dallas County, estimations were made for all Dallas county schools except the Dallas ISD.  In Taylor County, estimations were performed for Abilene ISD, Wylie ISD and Jim Ned ISD.

Improvements:  We know big-ticket items like pools and outdoor kitchens add value to a home, but the question is: will one ever recover the costs.  Other, small-cost projects -  lawn sprinkler systems, alarm systems -  contribute to the final price, but do they add value in a way that makes sense as another amenity.

Language:  Based on last year's analysis and the original Freakonomics publication, terms like "gorgeous", "cute", "spacious", being classified as a "starter" home or the liberal use of exclamation points are hypothesized to explain closing price.  If !!! or "gorgeous" will effectively communicate the important features of a home, we are quite content to employ them,  However, it is our anecdotal perception that exuberant punctuation and empty terms can send a subtle message that detracts from the property's appeal.  Too, our brokerage draws a distinction between calling a residence a "house" or a "home", but was curious if either term is predictive of price.

The table below lists all variables used to explain the sold price of homes in Taylor and Dallas counties.  The first column provides the name of the variable, the second column provides the dollar gain or loss posed by that variable, and the third column classifies statistical validity for each parameter.  Parameters with *** are significant at the 99% level, ** are significant at the 95% level, * at the 90% leave and no asterisks means the variable fails to contribute in any significant way. 

Other things held constant, here are highlights from the table:

The basics: Each square foot in a Dallas county house adds $107.23 and each square foot in a Taylor county house adds $86.33,  A Dallas county house loses $3752.56 each year and a Taylor county house loses $2041.64 per year, but in both counties, these losses are blunted by gentrification causing the loss due to age to decrease at a decreasing rate. 

Finishes: Quartz counters matter in both locales, but are worth almost double the value in Dallas (+$37,276.44) versus Abilene (+$20,626.53).  One should recognize this isn't just the return on quartz counters alone, but are associated with a renovated kitchen.  Acknowledging that, granite has become so commonplace and expected that it is meaningless in Abilene, but in Dallas, it seems to correlate with dated kitchens that demand renovation (-$27,945.95) just like even-more-dated Corian counters (-$52,176.11) .  Interestingly, "stainless" modeled no gain or loss in either market.

If one is looking for trends between markets, marble seems to indicate a significant return (+$71,085.25) to Dallas homes, but, so far, is hasn't begun to register in Taylor county.  With regard to paint and stain, oak is comparably valuable in both Dallas (+$18,035.36) and Abilene (+$13,927.27), pine is meaningful in Abilene ($+29,150.77), but not in Dallas, and white finishes add value in Dallas (+$31,727.58), but not in Abilene.  Given the cost of stain-grade wood, white finishes might be worth considering to improve spec house margins.  Maple (homage to the original Levitt investigation) has long since jumped the shark in both markets.

Floors and roofs: Wood flooring is a favorite in both Dallas (+$45,742.22) and Abilene (+$12,520.67).  In Abilene, leaving wood floors hidden under carpet is tantamount to losing $11,788.14 on price and suggests that a homeowner should consider uncovering and refinishing their wood floors prior to sale.  Vinyl flooring is comparably unpopular in both Dallas (-$10,649.82) and Abilene (-$10,934.68).  Laminate is meaningless in Dallas, but results in a $5,059.53 loss on an Abilene house (this divergence is a little surprising & one wonders how broadly the term "laminate" is applied).   Multiple floor transitions decrease sold price in Dallas (-$10,606.48), but have no bearing on Abilene prices (good news for all those Abilene do-it-your-selfers who couldn't resist another pallet of close-out tile).

Wood shingles similarly detract from price in both Dallas (-$51,224.08) and Abilene (-$45,761.45) and suggest a complete tear-off.  Metal roofs add value in both markets, but provide enough return in Dallas (+$121,834.39) that they warrant consideration in a spec home, but not in Abilene (+$15,572.21). Tile or slate roofs add a whopping $183,463.53 bump to Dallas home prices, but have no bearing on Abilene home values - perhaps our hail storms cause people to favor something that will only get dented or can be condemned and replaced.  Surprisingly, the markets are decidedly split on tar and gravel roofs - in Dallas it adds value (+101,803.00) and it Abilene it costs the owner value (-$23,406.93).  Perhaps Dallas tar/gravel roofs are associated with new, modern. flat-roof construction or pristine mid-century architecture and Abilene tar/gravel roofs are executed in a manner that is more about expediency than style.

Configuration: In hail-prone Texas, Dallas (+$9,763.38) and Abilene (+$3,185.18) buyers will pay more for a covered parking spot. They both appreciate the convenience of powder rooms (Dallas, +$88,810.03; Abilene +$6,269.65) and full baths (Dallas, +$83,658.41; Abilene, +$6,268.77).  They dislike two-storied homes (Dallas, -$62,049.34; Abilene, -$13,902.35).  Fireplaces are meaningful in Dallas (+$66,691.04), but play no role in value for an Abilene home.

Once, I heard someone provenance a long-time Abilenian stating: "They came up Elm Creek on the Mayflower".  To give measure to that statement, the term "creek" adds $28,080.64 in Abilene, but has no bearing on price in Dallas.  In the Taylor county data, only Jim Ned ISD offered a statistically significant bump (+$21,579.95) to housing prices - an indirect reminder of Abilene ISD's squandered primacy in Taylor county and our insouciance to the notion of sprawl.

Improvements:  NTREIS (North Texas Real Estate Information System) allows realtors to code an "outdoor fireplace/pit" and an "outdoor living center" as part of the property's description.  In both markets, these improvements meaningfully add value. Given the return to these items in both Dallas (+$152,322.40 for outdoor fireplace/pit, +$122,129.21 for the outdoor living center) and Abilene (+$44,522.00 for outdoor fireplace/pit, +$74,553.24 for the outdoor living center) this might be worth considering for a spec home and resolve some heartburn about adding these features if a relocation and sale is in the cards for a homeowner.  Similarly, pools add value in Dallas (+$36,718,17) and Abilene (+$25,746.95), but not to a degree that cover the cost of installation for a spec home.  Lawn sprinkler systems add value in Abilene (+$5,120.77), but not in Dallas.  Alarm systems offer a comparable contribution in both Dallas (+$8,446.59) and Abilene (+9,756.84).

Language:  Homes that can be described as spacious (translation: badly-configured, drafty square footage with no clear purpose) negatively impact a home's value in both Dallas (-$17,724.64) and Abilene (-$4,614.83) - a death knell to any return to the formal living rooms prominent in 1960's ranch homes. "Cute" homes are appreciated in both Dallas (+$31,246.10) and Abilene (+9,803.30), but Abilene homes that are "gorgeous" merit an extra $10,520.02.  Abilene "starter" homes gain an extra +$12,301.44, but see no measurable variation in Dallas. Homes that require "ideas" (typically phrased as "bring your ideas") experience a loss in value in both Dallas (-$41,294.22) and Abilene (-$30,614.61) - good news to home flippers with a clear, well-edited plan.

A house is made of brick and stone but a home is made of love alone

As a child, I my grandmother used this trivet, printed with its folksy message, in her kitchen. This two-line poem succinctly captures the distinction between a "house" and a "home".  This sentiment implies there is nothing personal or special about a house, but home is, well, just that.  Real estate agents cleave to this linguistic distinction; in property descriptions, Dallas agents use the word "home" 7.2 times more often than house; Abilene agents use "home" 6.2 times more often than "house".  Our analysis of both terms revealed that neither term has any measurable meaning in an Abilene sale.  On the other hand, if a Dallas agent calls a house a home, it is predictive of a $4,014.47 drop in price for each use of the term.  So much for the shibboleth of "home sweet home" in Big D.

With regard to language, Barnett & Hill's personal favorite from the original Freakonomics line-up is the contribution of exclamation points as a predictor of price. According to Levitt:  "an exclamation point in a real estate ad is bad news for sure, a bid to paper over real shortcomings with false enthusiasm."  When sliced by listings, Dallas listings, on average, use 1.15 "!" per listing, while Abilene agents use 1.33 "!" per listing.   As a predictor of price (and, it would seem, our comfort with real estate listings peppered with "!'s"), the Abilene market is statically insensitive to the "!", but in Dallas, the "!'s" predictiveness corroborates Levitt's a priori expectations - each "!" used to describe a property indicates a $5,338.06/! drop in price.

In response to this estimation, we'll reprise a measure we created last year: the effusivity index.  For an individual agent or brokerage, this is the number of ! in all listings divided by the number of total listings.  In the spirit of Levitt's analysis, the idea is to provide a rough assessment of realism brought by the agent to the listing.  Judy Colvin of Diamond Properties and Keith Turner of RE/MAX tied for recognition as the agents with the most listings who resisted the inveiglement of the exclamation mark.

Results by brokerage were also compiled and are listed in the table below, ranked by most to least effusive.

An effusivity index of 1 indicates an “!” per listing and 8 brokerages limited themselves to one ! per listing.  21 brokerages scored an effusivity index of greater than 1, indicating a love of the exclamatory with multiple “!” per listing. Thirteen brokerages sparingly employed the exclamation point when making their case, and 23 brokerages  scored 0, indicating no “!” used in any property description.  

Barnett & Hill was pleased to discover that they had the most listings among the brokerages that used no exclamation points, but at the same time must acknowledge the hollow self-praise in simultaneously devising a metric and then declaring themselves winner as a measure of a professional life expressed in sentence after sentence concluded with only the humble period.  That said, here's to an unflappably taciturn and dour 2016.