Household Income Growth and Firm Valuation: Evidence from REITs

Abstract

This article investigates the relationship between a firm’s market valuation and the household income growth in its asset locations. Using income tax data from the Internal Revenue Service and individual establishment information of U.S. equity real estate investment trusts (REITs) from 2000 to 2019, the article constructs an aggregated measure of household income growth for each firm based on its asset locations in various metropolitan areas. The paper adopts an identification strategy that links household income shocks to firm valuation. The results indicate that firm-specific household income growth positively affects firm value (measured as firm Q) and shareholder value (measured as market-to-book equity ratio). These findings suggest that local residents’ income is crucial in portfolio construction and operation.

Keywords:

• Household income
• asset location
• firm Q
• market-to-book
• REIT

JEL CODES:

• D31
• G32
• R3

Acknowledgment

I am grateful for the helpful comments from Jeffrey Cohen, William Hardin, Jing Li, Wenli Li (discussant), David Ling (mentor of the AREUEA Junior Scholars Program), Alejandro Pacheco, Xiaojin Sun, Chongyu Wang, Zhonghua Wu, and Tingyu Zhou, and seminar participants in the 2020 FMA Virtual Conference, 2021 AREUEA National Conference, 2022 ARES Annual Conference, and Xi’an Jiaotong University.

Notes

1 Korajczyk and Levy (2003) and Chang et al. (2019) argue that macroeconomic conditions affect firms’ profit growth and financing. DiPasquale and Wheaton (1992) provide a seminal work on the theoretical links between economic foundations and real estate.

2 Real GDP growth is used by the National Bureau of Economic Research (NBER) as an indicator of economic recessions. Other macroeconomic variables that are commonly adopted in academic studies include NBER recession, inflation, nominal interest rate, term spread, default spread, and stock market returns (Chang et al., 2019).

3 U.S. REITs hold more than 100,000 properties across the country and have properties almost in every metropolitan area. https://www.reitsacrossamerica.com/united-states.

4 One of the main challenges in studying the impacts of local income on the market value of assets in other industries is that there is no direct or easy way to measure how information on local income of an area should proportionally transmit to an individual portfolio or firm.

5 For example, consider a REIT with 20% of its property value in MSA 1, which has a 2% income growth in a given year, and 80% of its property value in MSA 2, which has a 5% income growth during the same year. The firm-specific income growth of the REIT in that year will be 20%*2% + 80%*5% = 4.6%. That is to say, this article adopts an asset-value-weighted average MSA-level income growth for each REIT each year. If a REIT has more properties located in high income growth areas, its firm-specific income growth would be high compared to others.

6 The county level income data is available at https://www.irs.gov/statistics/soi-tax-stats-county-data. The IRS provides information on the number of returns, number of exemptions, adjusted gross income, wages & salaries, dividends before exclusion, and interest received for each county in the United States, based on individual income tax return filed each year.

7 The County Business Patterns data, which provides employment information by industry for each county, is available at https://www.census.gov/programs-surveys/cbp/data.html."

8 The GDP growth data is available at https://www.bea.gov/data/gdp/gdp-metropolitan-area.

9 Negative firm value and shareholder value are replaced with missing values.

10 I also construct an alternative firm-specific income growth measure by defining each county as a local market to conduct a robustness check. The results are qualitatively and quantitatively similar to the main analysis and are not shown in this paper due to space constraints.

11 Using the growth rate of income tax data instead of the level reduces the geographical bias of the tax data. This is because some states do not tax income or wages, and taxpayers doing business across multiple states may have incentives to allocate revenue to low-tax jurisdictions. Using the growth rate of income tax data helps to address these issues and reduce the impact of states with disproportionately high or low income tax burdens on the analysis.

12 To make the total of the weight equal 100%, I drop all the properties that are not in any of the MSAs and not in the U.S. in the calculations. There is a very small number of REIT properties that are located in the rural areas, which are not part of an MSA. For robustness, I also examine the empirical results without dropping those properties. I find quantitatively similar results, which are omitted from this version of the paper for brevity.

13 The greater value of a property, which is highly correlated with a higher amount of revenue, is expected to explore more on local economic shocks. Thus, I use the net book value of the properties, rather than the number of properties, in each MSA as the weight for the local economic growth.

14 Although investment income is an important component of household income, I could not obtain a reasonable proxy for it in this study. This is because there are substantial differences in the taxation of earned income and investment income. For instance, investment income is not subject to payroll taxes such as Social Security and Medicare. Furthermore, tax rates on earned income are generally higher than those applied to investment income. See a CNN article on March 24, 2017, “Taxes: Taxes you have to pay”, at https://money.cnn.com/pf/money-essentials-taxes/index.html. Therefore, I focus on the wages and salaries component of household income, which is more reliable and easily measured in the data.

15 I use the Fama-Macbeth two-stage model to control for the time-series variation in household income growth and to highlight the effect of asset allocation on firm valuation. I also estimate equation (2)(2) $${\mathit{\text{Value}}}_{i,t}={\beta }_0+{\beta }_1{\mathit{\text{HH Income Growth}}}_{i,t}+\gamma {\mathit{\text{Controls}}}_{i,t}+{\delta }_i+{\epsilon }_{i,t}$$(2) using the panel regression approach with fixed effects for both firm and year. The results are largely consistent, suggesting that household income growth is an important factor influencing firm valuation. These fixed effects results are not reported in this paper for the sake of brevity.

16 See Baum-Snow and Ferreira (2015) and Goldsmith-Pinkham et al. (2020) for more details about this methodology.

17 This strategy is widely adopted in economics and finance literature. For instance, Saks and Wozniak (2011) and Adelino et al. (2017).

18 Manufacturing is the essence of the secondary sector of the economy. The NAICS code description for 31-33 – Manufacturing is: The Manufacturing sector comprises establishments engaged in the mechanical, physical, or chemical transformation of materials, substances, or components into new products. Establishments in the Manufacturing sector are often described as plants, factories, or mills and characteristically use power-driven machines and material handling equipment. For more details, see https://www.naics.com/naics-code-description/?code=31-33

19 The local industry-level employment shares are pre-determined (at year $t-1$) is because the instrument would need isolate variation in local employment changes due to differential impacts of national employment trends on local markets arising from the CBSA-level distinct economic structure.

20 For robustness, an alternative Bartik instrument is also adopted based on the interaction of changes in nationwide employment in the real estate sector (NAICS 531) with the preexisting regional real estate composition. The NAICS code description for 531 - Real Estate is: "Industries in the Real Estate subsector group establishments primarily engaged in renting or leasing real estate to others; managing real estate for others; selling, buying, or renting real estate for others; and providing other real estate related services, such as appraisal services. This subsector includes equity real estate investment trusts (REITs) primarily engaged in leasing buildings, dwellings, or other real estate property to others." (source: https://www.naics.com/naics-code-description/?code=531) While some regions may have little to no manufacturing employment, all areas should have some real estate employment. The untabulated results continue to hold and are available upon request.

21 Table 3, "Most Manufacturing-Specialized Metropolitan Areas among the 100 Largest Metropolitan Areas, 2010," can be found in Helper et al. (2012).

22 The economic significance of the estimated coefficients can be computed by taking the product of the estimated coefficient and the unconditional standard deviation of the household income growth and dividing it by the unconditional standard deviation of firm Q.

23 According to the REIT Industry Fact Sheet by National Association of Real Estate Investment Trusts (Nareit), as of December 31, 2019, the equity market capitalization of publicly traded US REITs is about $1 trillion, which consist of $102 billion Office REITs and $118 billion Industrial REITs. For more details, see https://www.reit.com/sites/default/files/media/PDFs/MediaBriefs/MediaFactSheetDec2019.pdf

24 The number of observations in the regression while excluding industrial and office REITs is 1,811, while that in the full sample is 2,388.