Home » 2014 Categories » 2014 Durham Paper » The Corner House and Relative Property Values: An Empirical Study in Durham’s Hope Valley

The Corner House and Relative Property Values: An Empirical Study in Durham’s Hope Valley

By Nathaniel Keating   The Corner House and Relative Property Values

ABSTRACT

   This paper analyzes the effect that the positioning of a single-family home on its block has on its open market price.  First, background information is presented and the arguments for and against owning a “corner house” are offered from a homebuyer’s perspectives.  The second part includes the results and discussion of an empirical study of homes in Durham’s Hope Valley neighborhood.  Drawing from data found at Zillow.com, a hedonic model for estimating property values was employed to determine the premium to which corner houses are valued with respect to houses located in the middle of the block.  This study found no evidence that corner homes in Hope Valley are priced any differently than homes located in the middle of the block.  Nevertheless paths for future research are recommended and outlined.

I. INTRODUCTION

The determinants of the value of housing have long been studied and constitute a large portion of the literature published in journals of urban economics and real estate.  First and foremost, a house, as an asset, is a collection of various components, each of which add value to the whole.  These physical elements comprise the characteristics most often advertised on real estate websites such as lot size, living space, number of bedrooms, garage, etc.  Second, the value of a home is inextricably linked to overall market conditions, both within the region of sale by the law of supply and demand and nationally (or internationally) by interest rates and credit markets.  Finally, housing’s most unique feature as an asset, its positional permanence, profoundly affects its value and therefore, the price it might fetch on the open market.

Thus, the real estate mantra “location, location, location”, while almost laughably clichéd, does not lack for support from modern economic theory.  Studies have demonstrated that single-family homes with proximity to small neighborhood parks were valued as much as 13% higher, whereas those in close proximity to mobile home parks experienced negative effects on selling price, lending credence to the hackneyed expression.  However, the focus of much of this research has long been the value added to (or deducted from) a property due to nearby features and amenities.  This paper mitigates these effects as much as possible and adds controls for the physical differences among homes on the same street in order to examine the potential influence a home’s position on its block might have on its value.

A simple exploration into the value of corner houses—here defined as those homes built on lots at the intersection of two streets, and therefore with at least two sides exposed to the public view—yields a surprisingly deep fissure in consumer preferences.  Some Americans love the idea of owning a corner lot home, chiefly because of their generally larger lot size and accessibility and visibility from multiple sides.  On the other hand many claim they would never own a corner house, citing lack of privacy and additional maintenance and upkeep.  In addition, some states allow municipalities to assess double the taxes for two street “frontages”, though no inquiry into Durham’s policy regarding double assessment for corner lots was performed for this paper.  Realtors are also divided, some claiming corner lots are more likely to be burglarized, while others counter that a more easily visible house, like a corner home, is safer.  These contrasting views of alarming fervor and conviction provide an enticing opportunity to test Americans preferences empirically.

From an economic perspective, corner houses offer a notably distinct home ownership experience to a potential buyer.  Considering their uniqueness, homes on corner lots may fall within Haurin’s description of the atypicality effect.  Under this theory, exceptional houses will attract a higher variance in offers from potential buyers.  A homeowner’s rational reaction to a higher variance of offers is to leave the home on the market longer than the average house in order to attract a higher number of offers.  This effect was extended by Turnbull to include analysis of housing prices which the conclusion that an atypical home sells at a discount.  This follows intuitively considering atypical homes take longer to sell: controlling for time on the market, the atypical home will sell for less than a more average home.

II. DATA & METHODOLOGY

            In this empirical study, data were gathered from Zillow.com, a leading online real estate database.  The data are accessible by the public and include the year each home was built, its lot size, square footage of living space, number of bedrooms and bathrooms, and each home’s Zestimate.  The Zestimate is an estimated market valuation for a home generated by Zillow’s formula with public and private information as inputs.  For the purposes of this study, the Zestimate will serve as a proxy for the value of each home.  The data set covers 32 homes selected from the Hope Valley neighborhood, and is subcategorized into 12 corner homes and 20 middle homes.  The homes were chosen at random using a random number generator from lists compiled of corner homes and middle homes in the appropriate area.  The means and standard deviations of the relevant variables can be found in Table 1 in the Appendix along with a map of Hope Valley.

The methodology undertaken involves a traditional hedonic pricing model implemented to isolate the corner house indicator variable’s effect.  This model assumes that the price, P, of a home—or Zestimate in this case—can be described as some function P = f(Xi), where Xi represents some combination of features of the home. Following convention, the semi-log form of housing price is used.  The equation used in evaluating the hedonic regression model is:

LogP­­­­i = bo + b1Bedi + b2Bathi + b3Sqfti + b4Corneri + ei

where Pi is the Zestimate and Bedi, Bathi, Sqfti, and Corneri  are respectively the number of bedrooms, number of bathrooms, square footage of living area, and an indicator variable equal to one if the house is on a corner and zero otherwise.  Also, the bi terms are parameters and ei  represents the error terms.  Each parameter bi represents the elasticity of housing price with respect to its associated variable.  Therefore,

 This model, while simple, includes some of the most powerful indicators of overall housing price to which the indicator variable Corneri was added.  In this way, the explanatory power of Corneri can be determined while controlling as best as possible for the most important, traditional variables.  In addition, it is worth noting that both the age of the home and the lot size were intentionally discarded as explanatory variables.  Further research into the Hope Valley region revealed that the neighborhood is split between very old and brand new homes with large variation in lot size.  Statistical testing revealed that both of these variables were surprisingly terrible at predicting home values for the randomly selected sample.

III. EMPIRICAL RESULTS & ANALYSIS

Table 2

4

The results of the hedonic pricing regression analysis are displayed above in Table 2.  The overall R2 of the regression was 0.7386 demonstrating that quite a large portion of the variation in price can be explained by variation in just a few variables.   However, a single variable regression conducting using price and square footage yields an R2 of 0.6782.  This would suggest that square footage is by far the best predictor of housing prices in Hope Valley.  Further study of table 2 reveals that square footage is the only variable with a significant t-statistic at the p < 0.01 level.  In fact, the t-statistic for the indicator variable Corneri is not significant at any reasonable p-level.  The model therefore provides no evidence that a corner home is treated any different on the market than a middle home.

            One such explanation for this is the nature of the Hope Valley neighborhood.  As one of the most secluded and affluent neighborhoods in Durham, fear of an increased chance of burglary and excessive noise and light from an additional side of the house are no longer issues to deter buyers.  Additionally, assuming a potential homebuyer in this affluent neighborhood would have a high annual income, such negatives as higher maintenance (e.g. more grass to cut, more gardening, more leaves to collect) might not exist for a homeowner that hires others for these tasks.  In this sense, perhaps the nature of the quiet neighborhood is masking what would normally be a more pronounced atypicality and therefore lower demand for corner homes.

Additionally, in Hope Valley there is much more space between homes than the average street in Durham, and the lot sizes are significantly larger as well.  In this way, perhaps the major benefits of owning a corner house in a more urban setting—larger lot size and less restricted positioning—might also be mitigated.  If so, the atypicality of corner homes would be further reduced to a negligible level.  It is therefore imperative that future research on this topic focus on a greater diversity of neighborhoods to determine the aggregate effect in Durham.

           

IV. CONCLUSIONS & LIMITATIONS

A hedonic pricing model was implemented to analyze the effect that the position of a house had on its price, specifically to determine if a corner house sells at a discount to its peers.  The results of the analysis produced the unexciting result that in Hope Valley no such considerations are taken, as there is no evidence that a house on a corner is priced any differently than a house in the middle of the block.

However, more research can certainly be done on this subject.  This study was severely limited in scope, collecting a very small sample size from only a single neighborhood in Durham.  Future attempts ought to include many hundreds more observations from a greater variety of cities and neighborhoods.  In addition, the hedonic model used included only a few explanatory variables.  Most notably, there was no control for the time each home was on the market before sale as estimates for housing prices were used.  In subsequent studies, greater care could be taken to determine a more robust model to predict housing prices, in particular a time-oriented analysis of sale prices.   Lastly, Zestimates were used as a proxy for housing prices.

The degree to which these estimates are accurate has been the subject of heated debate.  The consensus seems to be, however, that while most of the estimates are reasonably accurate (particularly given the lack of knowledge and inputs into Zillow’s proprietary formula), when the estimates are wrong, they can be extraordinarily incorrect.  If this is true, alternatives should be sought as use for housing prices.  Including these factors would certainly allow for more insightful discoveries and implications in the future.

Works Cited:

Espey, M., & Owusu-Edusei, K. (2001). Neighborhood parks and residential property values in

Greenville, South Carolina. Journal of Agricultural and Applied Economics33(3), 487-492.

 Haurin, D. (1988). The duration of marketing time of residential housing. Real Estate

Economics16(4), 396-410.

 Munneke, H. J., & Slawson, V. C. (1999). A housing price model with endogenous externality

location: A study of mobile home parks. The Journal of Real Estate Finance and Economics19(2), 113-131.

 Turnbull, G. K., Dombrow, J., & Sirmans, C. F. (2006). Big house, little house: relative size and

value. Real Estate Economics34(3), 439-456.

Appendix

 Table 1

5

Map of Hope Valley Neighborhood

6

 

 

            


6 Comments

  1. Nate– You explored a very interesting and unique idea, and I value that this paper does not try to claim significance where it is absent. My feedback is to question your chosen method for statistical analysis. Given that you were only interested in the effect of a house being on the corner of a street, rather than also the effect on price of the presence of different features of a house, I think your data analysis could have been improved with a statistical method other than simple regression.
    You mentioned that abnormal houses, like corner houses, experience more variation in their duration on the housing market. I think that an interesting question is therefore to test the amount of variation in price in houses that are located on a corner versus on the center of a block. If the within-group variation in covariates of the middle block and corner houses are similar, a simple comparison of two standard deviations of home prices would indicate relative variation between the two groups. If the covariates vary more for one group than the other, you would need to find a way to account for this. To do so, you could run two regressions, one with 20 random corner houses and 20 random center-block houses, and compare the R^2 values of the regressions. Alternatively, you could take a two-stage approach to the problem in which you would use logistic regression to calculate the propensity score that each house is on a corner given its physical traits, then use matching, binning, or weighting to compare houses of similar propensity scores and determine the effect of actually being on a corner on price or variation in price. If corner houses sell for both more and less than middle houses, you could be missing their effect by finding the average effect using simple regression. I really enjoyed reading this paper!

  2. Nate,

    I really enjoyed reading your paper and watching you present it in class. After reading about how a house’s position on a street is one of its most important features, I thought back on all the houses I have lived in and realized they were all at the bottom of cul-de-sacs. This was a definitely not a coincidence, but rather important to my parents when buying a new home.

    I thought your choice of analyzing the corner house was interesting because this position on a street is not one where the pros or cons clearly outweigh one another. As you described, depending on the neighborhood and preferences of the buyer, different buyers would be willing to pay a premium or want a discount for this feature. Furthermore, you made another great point that these pros and cons become more or less pronounced, depending on what street you are looking at.

    Something interesting going forward, would be to look at one of the pros or cons (such as proximity to a busy street) and see how this effects housing values. You could isolate this study to corner houses to see how the specific street you are looking at and whether or not it is near a busy road affects the value of the corner house. I thought your discussion of the limitations of your study was very thought provoking and i enjoyed reading it!

  3. I think that the topic you wrote about is a very interesting one, as I was under the impression that corner houses were typically sold for more money. My mother and father almost decided against purchasing our current house because they were specifically looking for a corner house.

    Throughout the paper you mentioned a few of the different negatives associated with owning a house on a corner such as higher gardening expenses, especially if they are using a private company. I think that a simple way to increase the credibility and influence of your paper would be to contact a landscaping company and see if they do in fact charge additional rates for corner houses.

    I also think that by utilizing a neighborhood other than Hope Valley your paper could have been much stronger, obviously this is easy to say after you have presented your research. The reason I feel this way is because from your map and data presentations there are very few actual true corner lots in the Hope Valley Neighborhood. In addition the fact that Hope Valley is an outlier in the Durham community does not allow for the research and results you have come up with to be easily passed along to other neighborhoods in Durham.

    Overall I think it is a very interesting topic. I also think that from the results that you have shown my own previously held belief about corner lots seems to be incorrect. Well done!

  4. Really interesting paper. I enjoyed listening to your presentation in class as well. It is always fun to hear about effects on property values in the area in which you live. In my personal opinion, I would prefer not to purchase a house on a corner lot because of the extra exposure to the street. I also had no idea that there was a double tax from the City of Durham for corner houses.

    I am wondering if you might be ably to find more research on prices of corner houses. I would imagine that there exists at least one article out there about it. It would also be interesting to see if there is any statistical significance for corner town homes and condos on both home value and rent value. I would imagine that price effects here would be more significant as there is the extra variable of wall sharing which could impact negatively the price of this sort of property. Generally end-units sell at a premium for their privacy.

    It is a shame that Zillow is so hard to work with in terms of collecting data. As you mentioned, perhaps there was no statistically significant findings due to the fact that the data set was small. It would be nearly impossible to click on each corner house individually and record its Zestimate. Also, it would be interesting to look at recent sale prices of corner homes compared to similar houses sold that are not on a corner. This way you might be able to see some price premiums or discounts that are not built into Zillow’s Zestimate formula, but are reflected in the sale price of the property. Data from MLS listings may provide such a data set. You would need a much larger area, however; and it might be hard to find similar homes sold if you are expanding your area.

    Great job and I really enjoyed reading your paper!

  5. Thank for your effort. I enjoyed reading your paper. But my major concern is your sample size for the test of statistical significance of corner’s homes. In my view the simple size of 12 was minimum. I would like to see a result with larger sample size if it is possible.

  6. Given that location is so highly valued in real estate, I found your conclusions to be very interesting. I would argue, however, that your statistical analysis is lacking. An obvious criticism is that you only use 12 homes, and all of them were in the same neighborhood. I think this limitation is demonstrated by the inability of your model to find number of bedrooms and bathrooms to be insignificant.
    As a recommendation, I would first suggest you widen your scope. Furthermore, I think your analysis would significantly benefit if you used other statistical techniques. Your hypothesis is that corner location adds value to a house relative to a non-corner house. This assertion indicates that a difference test would be more appropriate than a simple OLS regression. Another interesting point made in a previous comment is that value a corner location gives is dependent on the type of intersection. Therefore, I would like to know if your results hold if these criticisms are adressed, since location always seems to play a role somehow.
    -Garrett Kingman

Leave a comment

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