Putting the Supplier in Housing Supply: An Overview of the Growth and Concentration of Large Homebuilders in the United States (1990–2007)

Abstract

As housing production was ramping up in the 1990s and 2000s, some of the industry’s largest firms experienced remarkable growth primarily through mergers and acquisitions and the issuance of debt; the market share of the 10 largest firms tripled between 1995 and 2005. This article describes the role of financial firms in encouraging that growth and some of its consequences. Drawing on financial filings, news reports, investor analyses, and other relevant data, this article offers an overview of the relationship between homebuilders and investment firms, as well as a new explanation of the oversupply of housing in the 2000s. In doing so, this article seeks to bring attention to homebuilders as a missing feature in analyses of housing supply and housing markets, and proposes directions for future research.

Keywords:

• Housing
• homebuilding
• financialization
• housing markets
• housing supply
• shareholder value
• homeownership
• real estate
• economy

Acknowledgments

With thanks to Dave Barista, Nancy Brooks, Bret Christophers, Susan Christopherson, Marc Doussard, Bill Goldsmith, Dan Kuelmann, Steve Melman, Kathe Newman, three anonymous reviewers, the graduate students of Cornell University Global Finance Initiative, and the attendees of the Cornell Seminar in Regional Science, Planning, and Public Policy as well as those at other venues where versions of the article were presented, for help in one way or another. Also, a more general thanks to the past writers and editors of Builder and Professional Builder magazines, who not only served to tie together the many builders and developers of their time, but also tirelessly left a record for future researchers of the life and times of the homebuilding industry. The opinions and errors herein are mine alone.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. Ambrose and Peek (2008) summarize this literature, building on DiPasquale (1999).

2. Thanks to Bill Goldsmith for pointing this out.

3. Ball (2013) also found that the top 10 homebuilding firms in the UK have a much greater market share than their cognate firms in the United States and Australia because of the relative sizes of the territories in question (among other factors).

4. Moreover, the largest firms also received clearer market signals from the financial markets at the end of the housing boom—in fact, larger firms were encouraged to begin pulling back a full year prior to smaller firms, who received lending and market information from their local or regional banker (Haughwout, Peach, Sporn, & Tracy, 2012).

5. Ambrose and Peek (2008) use metropolitan data to show that growth in public firm market share occurs when the money supply tightens at local banks. Ambrose (2010) revisited this finding after the capital markets stopped lending following the Lehman Brothers bankruptcy. He shows that there are circumstances when large homebuilders’ access to funds will be just as constrained (at the international scale) as smaller firms (who are trying to borrow from their local lenders) as well as pointing to the significant increases in debt and equity issues among public homebuilders during the 1990s and 2000s.

6. Both took in a much higher level of profit, with smaller firms taking in excess returns of 43% and larger ones 48% (Haughwout et al., 2012, 2013).

7. In crafting their regression, however, Ambrose and Peek (2008) do not take into account firm concentration in MSAs, that is to say, the extent to which the market is dominated by one or more large public firms.

8. See also the epigraph from Ambrose and Peek (2008) at the beginning of the article. Ball notes: “real estate market models … assume a perfectly competitive market with zero transaction costs, stable technology and fixed firm behavior. The outcome is that only variations in construction input costs, selling prices and interest rates feature” (2006, p. 3).

9. The remainder of Gyourko’s piece (2009) makes no further mention of homebuilders or the industry’s changing firm structure in its discussion of housing supply. There is also no mention of homebuilders in Ferreira and Gyourko’s well-specified model of the beginning of housing booms (2011).

10. A search for the term “homebuilder” in Housing Policy Debate, for instance, found only 16 uses of the term since 1990 (searched 5 September 2015). For the most part these are trivial references, like the participation of a homebuilder association in an activity or mention of a speech given at the annual conference of the National Association of Homebuilders. The one article on suburban development which does offer an examination of suburban housing developments (in an area of Rhode Island), does not distinguish between local and national developers or relate developer costs to the size of the firm (Mohamed, 2010). Homebuilders are also not included in Ray and Guhathakurta’s examination of the role of sprawling submarkets in the financial crisis (2015).

11. DiPasquale notes: housing supply is the outcome of complicated decision-making by builders and the owners of existing housing. However, we have little direct evidence that permits us to observe the behavior of housing suppliers. To understand the micro foundations of housing supply, we would ideally want data in which the unit of observation is the supplier, with information on the quality and quantity of housing services offered, maintenance and capital improvement decisions, rents, and asset values. In the case of new supply, there is no standard data set that permits us to observe the behavior of builders of new housing ... This remarkable lack of information on the major actors in housing supply presents a significant obstacle to increasing our understanding of housing supply. More effort needs to be made in building new datasets that permit observation of the behavior of these important actors [and adds] We need to know more about the microfoundations of supply (DiPasquale, 1999, p. 10, 16).

12. Edelstein and Tsang note: “On the supply side, housing investment is affected by property values and exogenous supply shocks to the cost function. Supply shifters may be fundamentals, such as changes in construction costs, and will include credit liquidity constraints on builders” (2007, p. 302).

13. For a fuller explication of this approach see Wissoker (2013).

14. The ideas of shareholder value are grounded in key pieces by Berle and Means (1933), Milton Friedman (1970), and finance economists Michael Jensen and William Meckling (1976). Lazonick and O’Sullivan (2000) and Stout (2012) offer thoughtful histories of the idea of shareholder value and its consequences. Note also, that I will not be discussing here how growth is also a key tenet of one of the most common forms of American Capitalism, which is why when economic statistics or corporate results are presented they are always in the form of growth rates rather than in absolute terms. The book by Froud et al. (2006) looks in depth at the how shareholder value-based management was introduced into three large, multinational firms across three industries (Glaxo, Ford, and General Electric).

15. Measures such as the price/earnings ratio and the rate of growth began to become increasingly important, and on a quarterly basis, as did meeting analysts expectations (see, for example, Fox & Rao, 1997, cited by Dobbin & Jung, 2010). These expectations were not the same for all firms, however. Between 1995 and 2000, analysts saw firms as either part of the “old economy” or the new, high-tech economy, and set their expectations accordingly. While high-tech firms were valued based on seemingly unlimited potential, but no actual revenue or profit, old-economy companies who sought to create shareholder value were expected to “keep … costs steadily below sales revenue, so as to realize the 12%–15% on capital employed after tax … and, if possible … achieve sales revenue growth by organic growth or merger and acquisition” (Feng, Froud, Johal, Haslam, & Williams, 2001). Unfortunately for the homebuilders, with their product being literally bricks and mortar, they were bound to the more challenging “old economy” definition.

16. Dobbin and Zorn (2005) see firms that specialized in takeovers, like KKR, as responsible for the increasing presence of the shareholder value orientation in corporations as well.

17. Stock buybacks were done to have shares to pay executives (with some portion of the shares that were bought in) without creating more shares and in the process diluting the value of the existing ones. It was also done to decrease the total number of shares, which accomplished two things. First, it kept the earnings/share ratio lower, so it doesn’t make it look like the stock is overvalued with a price-earning (p/e) ratio that is too high. Second, it makes each remaining share more valuable, which is good for shareholders and for executives who get paid through stock options (and taxed at a lower rate than if they were receiving greater dividends) (Palley, 2007). For additional details see, for example, Lazonick and O’Sullivan (2000) and Grullon and Michaely (2002).

18. Some of this is chronicled in the works of Bluestone and Harrison in the 1980s (Bluestone & Harrison, 1982; Harrison & Bluestone, 1988). Marina Whitman, a former General Motors (GM) executive, has noted that while the press associated job cuts with a bump in share prices, in fact you only got the bump if the cuts were announced as part of a larger restructuring (1999).

19. All figures from The Institute of Mergers, Acquisitions, and Alliances, who compiled it from Thomson Financial data.

20. A study of the 328 largest corporations in America showed that by 1983 the number of acquisitions for diversification began to quickly shrink while the number of acquisitions of companies in the same industry increased 600% between 1983 and 1998; the acquisition of firms that could be vertically integrated increased by more than four times during the same period (Zorn, Dobbin, Dierkes, & Kwok, 2005). In contrast to the mergers and acquisitions of the 1950s and 1960s, many of which were centered on the idea of building a large conglomerate, corporations have, since that time, been urged to focus on horizontal and vertical integration rather than diversifying the types of firms under their umbrella (Dobbin & Zorn, 2005; Zuckerman, 2000). This pursuit of what Prahalad and Hamel called “the core competence of the firm” (1990) was facilitated by easing government antitrust regulation and followed on the heels of the break-up of conglomerates through shedding businesses that no longer fit the model, the era of leveraged buyouts, and junk bonds (Dobbin & Zorn, 2005; Bruck, 1989).

21. Along with the aspects discussed here, financialization also includes a number of other facets. One of the most recognizable is the increased use of new (or revitalized) financial products, many of which involve securitization, and many of which facilitated investment in traditionally hard-to-understand sectors, including real estate (Leyshon & Thrift, 2007). (On the public sector side, this meant the securitization of revenues from parking meters, future tax assessments, and the like (Ashton, Doussard, & Weber, 2012; Weber, 2010).) Another set of financial products were those that provided investors the means by which to hedge their investments by purchasing derivatives and other, related, products like credit default swaps, that enabled investment in financial instruments that in themselves (without the hedge) would be considered too risky. For consumers, financialization has come to mean a significant increase in household borrowing (using home equity loans, credit card debt, or anything else that was essentially a constant or predictable cash flow; Hyman, 2012), driven in part by the availability of funds through securitization (Nocera, 1994; Palley, 2007), greater consumer attention to financial rates of return, and the replacement of the traditional pension with a system in which individuals are responsible for their own financial health at retirement (Langley, 2008).

22. For the figures cited here, “Real estate-related industries” are defined as Construction (until 1990) and Construction and Real Estate, Rental and Leasing (NAICS 53) after 1990. The figures come from the BLS Employment from the Current Employment Statistics Survey and are not seasonally adjusted.

23. These figure come from the U.S. Census Bureau’s “New Privately Owned Housing Units Completed” (http://www.census.gov/const/www/newresconstindex_excel.html). Pricing comes from http://us.spindices.com/index-family/real-estate/sp-case-shiller.

24. These figures come from various years of Professional Builder magazine's annual “Housing Giants” listing of the top 400 homebuilding firms.

25. It is worth noting, however, that while the largest companies did lower their costs, the trade literature makes clear that smaller firms were actually able to work more cheaply in many cases (see, for example Even Small Builders Can Compete, 2006). Grebler (1973) offers a useful overview of some of the pros and cons of being a large firm in the industry.

26. Costantino, Pietroforte, & Hamill (2001) offer some insight into this. The use of homebuilders in the Northeast as the object of study limits the generalizability of its findings.

27. See Iskander & Lowe (2013) on skill development of Mexican workers in the construction industry.

28. Milkman (2006) makes clear that it was only after the unions had been defeated, roughly by the time of the 1982 recession (at least in the case of LA), that immigrants began to fill the jobs that Whites did not want because of the significant drop in salary and security post-union. It is also worth noting that in the 2000s there were a number of successful organizing campaigns, mostly directed at larger contractors for the large homebuilders (Rabourn, 2008).

29. Apgar and Baker surveyed a number of national builders, and are comparing national firms that build a minimum of 2,500 homes a year to regional firms that build between 500 and 2,500 homes a year (Apgar & Baker, 2006, p. 8).

30. One of the next steps to research in this area would be to investigate the motivations of homebuilders who sought to grow during the 1990s and 2000s and to test the idea that they continued building longer than common sense might dictate to generate sufficient revenues to meet investor and/or advisor expectations.

31. A short article in Builder magazine from 1995 notes: “The public builders continue to move into new markets. Spurred by the need to demonstrate 20% growth rates to Wall Street analysts, these companies are now popping up in second-tier markets, as well as crowding into first-tier cities” (Public Companies: On the Road Again, 1995). On the other hand, a somewhat philosophical piece on the merits of becoming a large firm in Builder (Donohue, 1990b) argued that small and medium size firms should realize that “Although the ‘Big’ mindset runs rampant in home building—as it does through the entire American economic system—the grass ‘on the other side’ is just as brown” (p. 198).

32. It continues: Although Wall Street apparently believes that most industries will continue to benefit from a healthy economic cycle, based on the stock prices in our industry, the Street seems to think that the home building cycle will end shortly and be followed by a severe and lengthy recession.

33. Wall Street was unimpressed, however. ‘This council is a waste of time’, says Chris Winham, building materials analyst for Goldman Sachs. ‘The individual companies have made their case already. The stock market is incredibly efficient, and generally right. It's more than a little bit naive to think investors don't understand’ (Wall Street Wimps, 2000). It is probably worth remembering that this comment was made shortly before the bursting of the tech bubble.

34. Haughwout et al. estimate that, at the peak of excess production, which didn’t actually hit until 2009, three and half million additional units were built (2012, p. 8). Not all of these were single-family homes. At the 2012 ACSP meetings, Kirk McClure offered a similar figure, based on the difference between household formation and units built.

35. Because this material is well documented, I will not address the political economy and growth of the mortgage sector, or the growth in the purchase of mortgage-backed securities (see, for example, Immergluck, 2009). One facet worth noting is the role of the Federal government in helping create the secondary mortgage market and the increased liquidity that accompanied it (Ashton, 2009; Gotham, 2009; Newman, 2009).

36. It is worth noting that homebuilders were also looking to regain some of the lending business they had lost to mortgage companies over the previous 5 years.

37. Further research would help pinpoint the exact relationship between the building firms and their mortgage division. There were also more traditional forms of reckless misbehavior, like that of the staff in Beazer Homes’ North and South Carolina division, who told prospective buyers that couldn’t afford a down payment that they were the beneficiaries of a charity that had offered to cover the down payment, when in reality the down payment was added to the total amount borrowed (U.S. Attorney’s Office’s, 2011).

38. Marxists theorists point to this as an example of capital switching (e.g., Harvey, 1982, 1985). There is a long history of debate on the topic (cf., Beauregard, 1994; Charney, 2001; Aalbers, 2008; Christophers, 2011).

39. This wasn’t the first time in recent memory that mortgage bonds became a “hot” item, that was in the early 1980s, as Michael Lewis chronicled in Liar’s Poker (1989). But financialization brought with it increasing deregulation, and the products built from the mortgages were now increasingly risky. By the end of the bubble the mortgages were as well (Gotham, 2012). As Bardhan (2009) points out, investors always need a market that is frothy. He argues that they require serial bubbles and, in this case, housing was that market (Bardhan draws on Pomboy, 2002).

40. Along with the increased presence of the large homebuilders in many markets, a second consequence of the growth of homebuilders and the influx of capital for homebuilding was the tremendous growth in the number of people living in communities that are governed by a homeowners association. This unique form of governance takes parcels of land and the citizens who occupy them and makes them into a quasi-private community, with its own strictures and laws. The citizens who live in these communities end up being what Barbara McCabe calls “dual citizens” (McCabe, 2005) with obligations both to their homeowners association and the city or town in which they live. As such, they assume responsibility for both building and managing traditionally public infrastructure such as roads and sewers, and for self-governance under extra-societal rules makes the residents of these communities and the properties and “public spaces” of those communities somehow ex-public. The numbers are staggering. In Arizona, for instance, it is estimated that “about 85% of new homes are in developments” (Stroebel, 2012, p. 2). Conservatively, one can estimate that 80% of those were single-family homes. The number of people living in places governed by residents associations doubled between 1990 and 2006, to 57,000,000 at which point it was roughly 20% of the U.S. population according to the Community Association Institute. It is worth noting that the tremendous growth of these communities, and the housing developments they represent, has its roots in the financial sector of the economy and not just in the development of a set of suburban cultural tastes, although that is important as well (cf., Bruegmann, 2005). Building such communities, then, has long-term implications for American democracy (Kohn, 2004; McKenzie, 2011).

41. Experts estimate that 24–33% of all home sales in the mid-2000s were by speculators (Anonymous, 2005; Downs, 2009).

42. The figure rises to 418.7 billion dollars if you use the 2003 average price $246,300 instead, which is slightly higher than the median home price in 2005, the last year in which the median price increased dramatically—although they would continue to rise through 2007. The average price of a new home in 2005 was $297,000, which if slotted into the equation would produce liquidity of 504.9 billion dollars, that is to say half a trillion dollars (U.S. Census Bureau, calculations by the author).

43. Or in cases where they increase buyer incentives that keep prices nominally the same, but in fact lower the cost of a home.

44. As a methodological note, it could be argued that if national firms are reacting to national economic trends, then perhaps there is no need to include them in regression analyses of housing markets, because they would be correlated with the national trends that are reflected in existing variables. This, I would suggest, would be unwise, as firms are rarely “perfect” reflections of economic trends, and their reactions are shaped, as Coiacetto (2006) notes, by their strength in any given market—essentially how much oligopolist or monopoly power they have there—as well as by differing management styles and the different types of pressure they are under from the financial sector.

45. This would be another factor to investigate with regard to homebuilders’ expansion to new markets as a way to maintain their growth rates.

46. The data from Builder are quite messy, in that market share is calculated on a different basis from year to year—sometimes they are a percentage of housing starts, other years housing permits, etc. A future project will be to create a useable data set from this material and analyze it. This is easier said than done, I recognize, given the gaps in the data. Thus, while Buzzelli and Harris (2006) argue that it is possible to generalize from their specific example of the builders in Southern Ontario, in fact their detailed case study (Buzzelli and Harris, 2006; Buzzelli, 2001; Buzzelli, 2004, and related papers) is one of the few instances where there are enough data available from which to generalize and which could be used for comparison with other metropolitan areas.

47. Thanks to one of my reviewers for this suggestion.