Abstract
We study the bargaining power of investors and the contagion effects of investor-owned single family homes on nearby property values. By controlling for the characteristics of both buyers and sellers, we find that investors tend to have more bargaining power than owner-occupiers — they purchase at lower prices and sell at higher prices, all else equal. We identify two types of investors: Professional Investors (e.g., corporations and partnerships) and Individual Investors. We find differences in the behavior of these two types of investors. For example, Individual Investors tend to invest in homes similar in terms of unobserved quality to those purchased by owner-occupiers. The tendency to buy lower quality homes is primarily attributable to Professional Investors. We also find that Professional Investors have more bargaining power than Individual Investors. For the contagion analysis, we use a repeat sales methodology and find that increasing ownership by investors in a neighborhood is associated with a small positive effect on nearby property values.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Change history
03 July 2020
A Correction to this paper has been published: https://doi.org/10.1007/s11146-020-09780-7
Notes
Molloy and Zarutskie (2013) discusses the potential benefits and risks of investor involvement. On the positive side, the authors observe that investors deploy capital for the purchase and renovation of homes that might otherwise remain under-maintained and/or vacant. However, they also discuss a risk that investors could overestimate rental demand or underestimate the cost to renovate creating a local risk of increased vacancy.
Previous studies have variously defined investor-owned properties as those where the tax bill is sent to an address different from the property address or as properties owned by entities that have purchased multiple properties, or used a list corporate names that have declared themselves in the business of operating single family rental properties. We use a combination of all three methods.
The bargaining framework can only be used when it is possible to identify the nature of both buyers and sellers.
A less favorable explanation is that the positive pricing advantage results from investors taking advantage of distressed homeowners facing foreclosure and/or pressure to move and therefore buy for prices below market value. However, this does not explain the price advantage observed for investor sales.
The literature on the benefits of homeownership (See Coulson and Li 2013 for a summary) has contributed to a negative perception of renters and investors. The conventional explanation for this perception has been that renters under-invest in maintenance activities because they do not participate in the investment benefit and further may not stay long enough to enjoy the full consumption benefit. Further some have argued that investors are more likely to default because of lower default costs. See Henderson and Ioannides (1983), Wolfson (1985), Williams (1993), Ioannides and Rosenthal (1994), O’Sullivan (1996), and Haughwout et al. (2010).
See Frame (2010) for a review of related foreclosure contagion literature. Other recent foreclosure contagion papers include Li (2017). A relevant contribution of Li (2017) is the finding of a “contagion” type of effect for local capital investments. For example, Li (2017) finds than nearby homeowners are more likely to invest in major maintenance projects when they observe nearby owners investing in their property.
The term “rental discount” in Turnbull and van der Vlist (2017) and others refers to the common observation that rental properties generally sell at a discount to observationally equivalent owner-occupied properties.
To the extent that investors tend to deal with lower quality properties (as measured by characteristics not included in the regression model), if the “demand” effect is not controlled for in the model, the estimated bargaining discount is likely to be biased. Further, it is important to control for the characteristics of both the buyer and seller. For example, if investors are intrinsically better bargainers and an investor buyer negotiates with an owner-occupier seller, we expect to observe a larger negotiated “discount” than if an investor buyer negotiates with a similarly skilled investor seller.
As described by Cohen et al. (2012), Atlanta experienced a much larger run up in prices, as well as a much steeper “bust”, than Denver.
Another issue with using the tax authority mailing address is the common practice of mortgage lenders escrowing property tax payments. In most cases where taxes are escrowed, the tax bill is sent to the lending institution where the address will differ from the property address.
The discussion of potential errors associated with identifying investors by differences in address fields is intended to be generally descriptive and not any specific article. By restricting our attention to single family properties, we avoid the second error.
In addition, it is likely that at least one of the multiple purchases by an individual investor is for the purpose of occupancy. Thus, some owner-occupier purchases will be classified as investor purchases.
Mills et al. (2019) obtain their “master list” based on entities that “appear frequently in media reports on buy-to-rent activity or that follow a business model that is known to be the same as the rest of the buy-to-rent investors…”
We used the Mills et al. (2019) list of names but found that those firms accounted for a very small fraction of the Denver, CO transactions during our study period.
The most common terms that identify professional investors are “LLC”, “LLP” and “INC”. A complete list of the terms used to identify professional investors is available from the authors upon request.
This error type is mitigated in our data by the fact that we exclude new home sales so that almost all our transactions are a second or higher sale transaction for the property.
As noted before, Turnbull and van der Vlist (2017) use a set of indicator variables to capture the possible permutations of buyer and seller.
Ideally, one would control for other buyer and seller characteristics. For example, one could envision measuring the experience, education and liquidity of the buyer and seller. Unfortunately, our data do not enable us to use anything more than the identification of the buyer/seller as an investor or not.
The description of the bargaining model presented here is based on the more detailed development in Harding et al. (2003).
With deep, liquid markets, both buyer and seller can costlessly find an alternative property.
The situation is even more complex if one simply includes an indicator denoting the status of only the buyer or seller, as several previous authors have done. For example, if one tries to estimate an “investor discount” by including an indicator for the buyer being an investor without controlling for the seller characteristic, unless the buyer and seller variables are uncorrelated, the estimated coefficient will not be an efficient estimator of the sum of the effects.
We consider different definitions of investors: Professional Investors, Individual Investors and both definitions combined. As discussed in the Results Section, when estimating our bargaining models, we control for location and time using tract by year fixed effects.
Based on our own analysis and other published work, we have a strong prior that dbuy < 0. If investors make larger investments in improvements and upgrades (to unobserved characteristics) than do owner-occupiers, then dsell > dbuy -- and possibly positive if investors sell properties with more valuable unobserved characteristics than do owner-occupiers.
One exception is that the repeat sales model cannot control for the age of the property. As discussed in the Data section, the average age of homes in our sample is 54 years. It is possible that older homes require more maintenance than newer homes. Since we are unable to observe the maintenance expenditures associated with the homes in our sample, a correlation between age of the property and unobserved maintenance expenditures could reduce the precision of our contagion estimates. Nevertheless, we believe the repeat sales approach provides the best possible control for unobserved characteristics.
The modification incorporates the assumption that prices of all homes rise and fall due to overall market forces. The current level of market forces is represented by γt .
The land at the northeast end of the city was annexed by the City of Denver in 1988. This land was subsequently the location for Denver International Airport, which opened in 1995. Our data covers all transaction in the City of Denver from 1986 through 2016. Therefore, some transactions related to properties in the annexed area will not be included in our dataset if they occurred before the annexation. We do not believe such exclusions are significant because of the short time period (two years) affected (relative the thirty years for the full period) and the fact that there were fewer total transactions in 1986 and 1987 and our belief that the annexed areas were less densely developed at the time of annexation. The airport was not built until 1995, so any associated stimulated transactions would be included in our data.
To the extent that bargaining or contagion effects vary systematically with such housing market characteristics, our results may not be generalizable to markets with different characteristics. Similarly, our results may also be affected by the nature of the housing cycle in Denver—which was less severe than in certain other markets such as Miami or Phoenix.
The data reportedly includes all transactions subsequent to 1993, but also includes many transactions from earlier years. The full raw data (for all property types and deed types) include just less than 600,000 transaction records dating back to 1986. After restricting attention to single family residential properties with reasonable prices and characteristics, and transactions based on warranty deeds, we are left with 251,376 records. Within this restricted sample, 1986 is the first year with more than 3000 transactions (e.g., 1985 has 157 transactions). Because the single family transaction records from years prior to 1986 appear to be sparse, we restrict our attention to 1986 and later. For some portions of the analysis, we use the full date range from 1986 through 2016, for other analyses we restrict attention to more recent data (e.g., 2003–2016). See separate discussions in the Results section.
We drop houses that are over 100 years old because many are historic properties that are subject to restrictions on redevelopment and use. We are concerned that there are unobserved characteristics associated with such properties (e.g., unique quality and design characteristics, historical designations, unique architecture, etc.) that would affect the observed transaction price. We are also concerned that such transactions may involve a special type of buyer and seller.
REO sales frequently have damage or deferred maintenance that is not captured by the reported structural characteristics.
As a robustness check, we rerun the estimation including new home sales and REO sales. The bargaining results are little changed. See the discussion below.
Smaller lots are likely associated with locations that are more densely developed.
Any underestimate of investor buyers based on the address field would bias this difference downward, so this behavioral difference is robust to that error.
Estimating equation (11) without the log transform results in the coefficients being interpreted as dollar amounts. This facilitates interpretation of some of the coefficients such as the price per square foot of living area, lot size and bathrooms. We also estimate equation (11) with the log transformation resulting in coefficients interpretable as percent changes. The overall model results are quite similar and for paper length considerations, we only present the untransformed results in Table 2. The bargaining results with the transformed dependent variable are presented in Table 3. Full results are available from the authors upon request.
The negative coefficient on the number of bedrooms is a frequently seen result. Recall, that the coefficient reflects the price per bedroom after controlling for total square footage of the home. A preference for larger bedrooms can be reflected as a negative coefficient for the number of bedrooms or total rooms after controlling for overall size.
The estimated attribute prices are generally higher in the later time period. This could reflect the failure of the fixed effects to fully control for house price inflation. We experimented with using real 2016 dollars rather than nominal dollars and the results were qualitatively similar.
Consistent with the argument that liquidity can provide extra bargaining power, Asabere et al. (1992) and Jauregui et al. (2019) report a pricing advantage for all cash offers over offers contingent on mortgage financing While the earlier work estimated an effect as large as 13%, the more recent work shows that house and neighborhood characteristics explain much of the pricing difference and estimate a cash discount on the order of 4%. Although the fact that investors have alternative sources of funds than traditional mortgage lenders may be one reason why they are able to buy at lower prices, that should not affect their ability to sell at higher prices
Without the assumption of symmetry, it is possible that dsell > 0, but the weighted combination of dbuy and dsell is negative. This would require dbuy to be much more negative than reported by the coefficient in Table 2.
Note that these restricted model specifications imply that in the results using the definition of investor restricted to just Individual Investors, if an individual investor sells to a professional investor, the bargaining variable will be +1 not zero and the demand variable will take on the value of 1 because the professional investor is classified as a non-investor in this particular regression. Similarly, if a professional investor sells to an individual investor, the bargaining variable will take of a value of +1 and the demand variable will also be +1 in the regressions that use the definition of investor restricted to just professional investors.
The results of the log transform estimations are available in a separate table upon request from the authors.
All of the additional transactions in Model 5 relative to Models 1–4 entail a warranty deed and thus do not reflect the transfer of the property from a defaulted borrower to a foreclosing lender. Such transactions typically involve a deed type other than a warranty deed.
The bargaining estimates for Professional Investors presented in Table 3 are highly statistically significant for all 7 models, while only 3 of the 7 Individual Investor models have statistically significant bargaining effects. When both individual and Professional Investors are pooled, all 7 bargaining effects estimates are highly significant.
Recall that, to the best of our ability, we exclude non-arms’ length transactions and transactions that appear to have erroneous price data. Such exclusions do not necessarily mean the deletion of all transactions for a property. If a home sells at t0, t1 and t2, but we exclude the transaction at t1, we treat the sale at t0 as the original purchase and the sale at t2 as the subsequent resale. On the other hand, homes that transact twice are excluded if one of the two sales is excluded for violating the filters described earlier.
The most significant filter restricted the sample to just those repeat sales pairs that exhibited less than a ± 50% per year annual rate of price appreciation (depreciation). Other filters included eliminating all short-term resales.
Age of the house is measured at the time of the first sale in the repeat sales sample and consequently differs from the average age reported in Table 1, which averages age at all transactions for each house.
The short cut approximation that ln (1 + r) is roughly equal to r only holds for r close to zero.
A portion of property price appreciation can be the result of unobserved maintenance instead of general price appreciation. Hayunga et al. (2019) and others show that higher probability of default deters maintenance and affects the trajectory of appreciation.
Note that the rate of price appreciation is only one component of the total return from investing in the home. In addition, investors receive a stream of rental income and owner-occupiers receive a stream of housing benefits associated with living in the home. Furthermore, we do not observe maintenance and improvement expenditures which would offset a portion of the price appreciation. See Harding et al. (2007).
The distribution of the annual return from price appreciation is skewed to the right. The median annual rate of return is 6.85%. The annual rate of price appreciation associated with a .3337 log ratio of price ratio is 6.26%.
It is important to note that we are not able to control for maintenance and improvement expenditures made by buyers after the purchase. If investors invest additional funds for repairs after purchase, the calculated rate of return will overstate the investor’s true rate of return.
Rosenthal (2018) reports that the national average for investor-owned single family detached properties is approximately 15%.
Based on an analysis of correlation coefficients (available upon request), we find that there is little correlation (ρ = .09) between the change in nearby REOs and the change in the number of nearby investor-owned properties. The correlations between the actual number of REOs and the number of investor-owned properties is higher (0.25 at t0 and 0.37 at t1). This positive correlation further supports the claim suggesting that investors are more active in neighborhoods with higher numbers of distressed properties. Much of this positive relationship is eliminated when we restrict focus to the change in the numbers of nearby properties.
We use the FHFA Annual HPI for CBSAs (All-Transactions Index) for CBSA #19740 (Denver-Aurora-Lakewood CO). The FHFA reports that these annual CBSA indexes should be considered developmental. As with the standard FHFA HPIs, revisions to these indexes may reflect the impact of new data or technical adjustments. Indices are calibrated using appraisal values and sales prices for mortgages bought or guaranteed by Fannie Mae and Freddie Mac. For more information on the calculation of FHFA indexes, see: Bogin et al. (2018).
CoreLogic collects sales data from various sources and verifies reported transactions using its own proprietary algorithms. See https://us.spindices.com/index-family/real-estate/sp-corelogic-case-shiller.
We regress the annual changes in each of the three indices against the other two. In all cases, the changes in the indices are highly correlated, with correlation coefficients ranging between 0.84 and 0.90. These correlations confirm the visual evidence supporting the notion that our index closely tracks the other two published indices.
For example, HRY report an estimated REO effect of roughly −1%.
Unobserved maintenance may vary with market cycles, and this can be reflected in bargaining coefficients. We are the first to look at bargaining effects in this way. The relationship between bargaining power and the market cycle is an important topic for future work.
References
Allen, M. T., Rutherford, J., Rutherford, R., & Yavas, A. (2018). Impact of Investors in Distressed Housing Markets. Journal of Real Estate Finance and Economics., 56, 622–652. https://doi.org/10.1007/s11146-017-9609-0.
Asabere, P. K., Huffman, F. E., & Mehdiany, S. (1992). The Price effects of cash versus mortgage transactions. Real Estate Economics, 20(1), 141–154.
Bailey, M. J., Muth, R. F., & Nourse, H. O. (1963). A regression model for real estate Price index construction. Journal of the American Statistical Association, 58(304), 933–942.
Bogin, A., Doerner, W., & Larson, W. 2018. “Local house price dynamics: New indices and stylized facts.” Real Estate Economics.
Bracke, P. (2015). “How much do investors pay for houses?” Bank of England. Staff working paper 549.
Case, K. E., & Shiller, R. J. (1989). The efficiency of the market for single-family homes. The American Economic Review, 79(1), 125–137.
Cohen, J. P., Coughlin, C. C., & Lopez, D. A. (2012). The boom and bust of US housing prices from various geographic perspectives. Federal Reserve Bank of St. Louis Review, 94(5), 341–367.
Coulson, N. E., & Li, H. (2013). Measuring the external benefits of homeownership. Journal of Urban Economics, 77, 57–67.
Epple, D. (1987). Hedonic prices and implicit markets: Estimating demand and supply functions for differentiated products. Journal of Political Economy, 95(1), 59–80.
Fisher, L., & Lambie-Hanson, L. (2012). Are investors the bad guys? Tenure and neighborhood stability in Chelsea, Massachusetts. Real Estate Economics, 40(2), 351–386.
Frame, W. S. (2010). Estimating the effect of mortgage foreclosures on nearby property values: A critical review of the literature. Federal Reserve Bank of Atlanta: Economic Review ISSN 0732-1813.
Ganduri, R., Xiao, S.C., & Xiao, S.W. 2019. “The rise of institutional Investment in the Residential Real Estate Market.” Available at SSRN 3324008.
Griliches, Z. (Ed.). (1971). Price indexes and quality change. Cambridge: Harvard University Press.
Hansz, J. A., & Hayunga, K. (2016). Revisiting cash financing in residential transaction prices. Real Estate Finance, 33(1), 33–46.
Harding, J. P., Rosenblatt, E., & Yao, V. (2009). The contagion effect of foreclosed properties. The Journal of Urban Economics, 66(3), 164–178.
Harding, J. P., Rosenthal, S. S., & Sirmans, C. F. (2003). Estimating bargaining power in Markets for Heterogeneous Products. The Review of Economics and Statistics, 85(1), 178–188.
Harding, J. P., Rosenthal, S. S., & Sirmans, C. F. (2007). Depreciation of housing capital, maintenance and house Price inflation: Evidence from a repeat sales model. Journal of Urban Economics, 61(2), 193–217.
Haughwout, A.D., Tracy, J., & van der Klaauw, W. 2011 “Real estate investors, the leverage cycle, and the housing market crisis.” Federal Reserve Bank of New York Staff Report 514.
Haughwout, A., Peach, R., & Tracy, J. (2010). The homeownership gap. Current Issues in Economics and Finance, 16(5), 1–10.
Hayunga, D. K., & Munneke, H. J. (2019). Examining both sides of the transaction: Bargaining in the housing market. Real Estate Economics. https://doi.org/10.1111/1540-6229.12272.
Hayunga, D. K., Pace, R. K., & Zhu, S. (2019). Borrower risk and housing Price appreciation. The Journal of Real Estate Finance and Economics, 58(4), 544–566.
Henderson, J. V., & Ioannides, Y. M. (1983). A model of housing tenure choice. American Economic Review, 73(1), 98–113.
Ioannides, Y. M., & Rosenthal, S. S. (1994). Estimating the consumption and investment demands for housing and their effect on housing tenure status. Review of Economics and Statistics, 76(1), 127–141.
Jauregui, A., Tidwell, A., & Sah, V. (2019). Sample selection approaches to estimating and allocating house transaction funding Price differentials. The Journal of Real Estate Finance and Economics, 58(3), 366–407.
Li, L. (2017). Why are foreclosures contagious? Real Estate Economics, 45(4), 979–997.
Mills, J., Molloy, R., & Zarutskie, R. (2019). Large scale buy-to-rent investors in the single-family-housing market: The emergence of a new asset class? Real Estate Economics, 47(2), 399–430.
Molloy, R., & Zarutskie, R. 2013. “Business investor activity in the single-family-housing market.” Federal Reserve Board of Governors FEDS Notes, December 5, 2013.
O’Sullivan, A. 1996. Urban economics. Third Edition. Irwin. McGraw-Hill: Boston.
Rosen, S. (1974). Hedonic prices and implicit markets: Product differentiation in pure competition. Journal of Political Economy, 82(1), 34–55.
Rosenthal, S.S. 2018. “Owned now rented later? Housing stock transitions and market dynamics”. Working Paper.
Schnure, C. 2014. “Single-family rentals: Demographic, structural and financial forces driving the new business model”. NAREIT working paper, march 31.
Smith, P.S., & Liu, C.H. 2018. “Institutional investment, asset illiquidity and post-crash housing market dynamics.” Real Estate Economics, Forthcoming, Institutional Investment, Asset Illiquidity and Post-Crash Housing Market Dynamics.
Turnbull, G, and van der Vlist, A. 2017. “Investor bargaining power, rental externalities and housing prices.” Manuscript.
Wolfson, M. (1985). Tax, incentive and risk sharing issues in the allocation of property rights: The generalized lease-or-buy problem. Journal of Business, 58(2), 159–171.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The original version of this article was revised: The original version of this article unfortunately contained mistakes. Errors were found within Tables 2, 3, 5, 6 and 7. These tables were corrected accordingly.
Rights and permissions
About this article
Cite this article
Cohen, J.P., Harding, J.P. The Bargaining and Contagion Effects of Investors in Single Family Residential Properties: The Case of Denver Colorado. J Real Estate Finan Econ 67, 29–64 (2023). https://doi.org/10.1007/s11146-020-09766-5
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11146-020-09766-5