Graduate Theses Break New Ground
September 29, 2006
Since the founding in 1983 of the MIT Center for Real Estate, MIT/CRE students have made substantial discoveries through their thesis work. The Class of 2006 is no exception. Three students in the Center's MSRED graduating class — Arvind Pai, Kristen Wang, and John Harris Morrison — shared surprising findings during the Center's Graduation Weekend.
Arvind Pai showed that core real estate pricing violates stock market behavior — but in a systematic way. Kristen Wang's work revealed innovative financing that public housing authorities are using to fund public-private partnerships to develop housing. John Harris Morrison showed that private local investors buy the smallest properties, but they make more money doing it than larger firms.
After MSRED graduate Arvind Pai investigated real estate asset pricing, he developed a model that explains 90 percent of long-run historical core property returns in the National Council of Real Estate Investment Fiduciaries (NCREIF) index. In his thesis Stocks Are from Mars, Real Estate Is from Venus, Pai examined the factors that affect long-run real estate investment performance.
He sought to develop an investment strategy from scratch. "NCREIF investors seem to have a very narrow focus on certain property types located in certain metropolitan statistical areas (MSAs) and on very large property sizes. Potentially that could be driving down spreads. Are there superior returns elsewhere?" Pai asked. In addition, Pai pointed out that investment managers do not have an explicit asset pricing model for how the market prices real estate, and what risk the market recognizes. "Can we do better?" Pai asked. "Can we create a model that actually understands how the market prices risk in real estate? Or is it possible that the market just doesn't have the ability to systematically differentiate and price within the real estate core?"
Model Shows That Property Type and Size Matter — Not Location
Using over 20 years of institutional investment data (1984-2003) from NCREIF, Pai created two portfolios: a Property Size portfolio and a Metropolitan Statistical Area (MSA) portfolio. The Property Size portfolio had three sizes of property — small, middle, large — broken down by six types of property — offices in the central business district, offices in the suburbs, flexible industrial research and development sites, industrial warehouses, retail, and apartments — for a total of eighteen investment categories. The MSA portfolio also had eighteen categories, divided by three MSA tiers — upper, middle, and tertiary, depending on the distance from the city's center — and had the same six property types.
Using over 20 years of institutional investment data (1984-2003) from NCREIF, Pai created two sets of portfolios: a property size portfolio set and a metropolitan statistical area (MSA) portfolio set. The Property Size portfolio set had three sizes of property — small, middle, large — broken down by six types of property — offices in the central business district, offices in the suburbs, flexible industrial research and development sites, industrial warehouses, retail, and apartments — for a total of eighteen portfolios. The MSA portfolio set also had eighteen portfolios, divided by three MSA tiers — upper, middle, and tertiary, depending on the total investment value of NCREIF in those MSAs — and for the same six property types.
Pai created a measure of risk and return associated with each investment category represented by each of these portfolios, based on the Single Factor Capital Asset Pricing Model (CAPM). Pai found several surprising results:
- Portfolios with higher market beta (calculated over the collective NCREIF property set) had lower historic returns — a result that contradicts conventional theories and stock market behavior.
- Different property types tend to aggregate; offices, apartments, and the other categories clump together on graphs of return vs. market beta. "There's a strong market pricing factor by property type," Pai said. "The market requires a premium for investing in apartments, retail properties, and warehouses."
- Size matters: properties also aggregated by size, though not as distinctly as by property types, and the market required a premium for investing in large properties. "This is the opposite of what happens in the stock market," Pai said. By contrast, location doesn't matter: there was no aggregation by MSA tier.
Multi-Factor Model Explains Asset Pricing
Pai then built a multi-factor capital asset pricing model by introducing new factors to the CAPM. He found that property type was the most significant factor. Central business district offices carried the lowest risk premium; premiums increased in order for suburban offices, warehouses, research and development facilities, retail, and apartments. "We wondered why this was the case," Pai said. "It probably represents systematic personal preferences, or the risk-averse nature of investment managers — what we call the "picture on the wall" effect. [They're] more comfortable putting these big commercial office buildings in downtown somewhere — it looks good, it makes them feel comfortable with that product type — and even investors look at that and say 'OK, these guys are doing a good job.'"
Size was also significant. Pai speculated that large buildings carry an "illiquidity premium" because they can be difficult to sell. The market factor had small, negative influence on returns — the opposite of what asset pricing theory predicts. Pai said that this could be because real estate market betas are not stable, that the market does not think they are stable, or that the market simply does not care about the risk of real estate volatility.
In conclusion, Pai's multi-factor equilibrium asset pricing model explains 90 percent of long-run core property returns in the NCREIF index. "What's interesting is that the market actually does price core real estate in a systematic manner," Pai said.
Arvind Pai is founding principal of Contemporary Environment, a New Delhi-based, design and build architectural firm.
Kristen Wang, MSRED 2006 and Urban Studies and Planning, Master of City Planning (MCP) 2006, examined the role of the federal HOPE VI grant program on public housing "...one of the more innovative government programs of the last decade," Wang said. HOPE VI has emphasized public-private partnerships and a market approach to public housing, but its funding has been drastically reduced since 2004. "I was interested in seeing what the industry was thinking in an environment where these funds are likely to be cut. Do they want to continue developing in this vein?" Wang asked. "Are they interested in trying to replicate HOPE VI results without the infusion of HOPE VI funds?"
Since 1992, HOPE VI has provided substantial grants to public housing authorities to demolish and rebuild public housing and provide services to residents, Wang said. The program's goal is to revitalize public housing by improving the living environment, decreasing the concentration of very-low-income residents, and creating sustainable communities. According to Wang, HOPE VI is the only funding program specifically designed to meet the needs of distressed public housing and its very-low income residents. "HOPE VI has turned around a lot of neighborhoods," Wang said, "and transformed them into mixed-income, more livable neighborhoods by galvanizing both public and private investment in these communities."
HOPE VI will have to galvanize much more private investment in the future because federal funding for HOPE VI has been slashed. During the first ten years of the program, it was funded at $550 million a year; since 2004, the program has received less than $125 million in funding each year. The Bush administration has proposed eliminating the program entirely, but HOPE VI has survived due to strong bipartisan support. Some critics have cited high per-unit costs and construction delays as reasons for eliminating the program; others are disturbed by the net loss of affordable housing units despite HOPE VI's funding, and by how residents are treated while their housing is being rebuilt.
Housing Authorities Begin Innovative Financing
Wang interviewed over 20 developers, consultants, lenders, and representatives from public housing authorities and cities around the country about their plans for a HOPE-less future. Most of the housing authorities and developers were in large urban areas, and the developers had strong established interest in public housing. All had worked on HOPE VI projects.
She found that public housing authorities have learned to be much more innovative than in the past. Since the HOPE VI funds were reduced, the housing authorities have gained the expertise that allows them to use new financing tools and to work with the marketing different ways. "They're making much smarter, savvier decisions about real estate than they were ever able to do in the past," Wang said. Public-private partnerships give communities a chance to learn from developers, and give developers a chance to learn from communities. The result is that public funds are being creatively leveraged to redevelop housing, Wang said.
The challenge is finding the funds to do the work. Municipalities must draw together more and more sources of funding to improve public housing. "One of the big things is how many different sources of funds go into these projects," Wang said. "HOPE VI funds are already difficult to fund, and without that infusion of grant funds, it's much harder to make up that difference." Based on her interviews, Wang sorted the 18 sources of funding discussed by her interviewees into four main categories:
- Direct subsidies from the federal Department of Housing and Urban Development (HUD) — "the traditional sources," Wang said.
- Government subsidies from non-HUD funds, such as state funds and low-income housing tax credits — "the domain of affordable housing developers, low income housing tax credit and other levels of government as well as foundations," Wang said.
- Leveraged public funds, such as tax exempt bonds or HUD Section 8 housing subsidies. "They're getting creative," Wang said, "by pledging future funds as payment or security to construct the building in the first place — or they are leveraging the land and the market. They are asking, 'What will be the future increased value of this property? How can I use that now to get it to that point?' For example, rent from market-rate units can subsidize operating costs for lower-income units; tax-increment financing can help municipalities borrow against the future taxes when the redeveloped properties will be worth more."
- Leveraged land or markets, such as cross-subsidies, increased density, and surplus land proceeds.
Housing Authorities Need Flexibility and Funds
Unfortunately, simply finding financing for building projects will not replace HOPE VI. Part of the appeal of HOPE VI funds was that they were not just building funds, but also monies for a range of services to serve public housing residents. According to Wang, HOPE VI funds have traditionally been used to fund relocation expenses during construction, infrastructure improvements, and even home ownership programs — which are not allowed under many other public housing programs. "Pursuing the multiple goals of HOPE VI without the structure of HOPE VI is really difficult.... One of the great things about HOPE VI was the holistic approach. Without that structure in place, it could be more efficient to get rid of some of those goals, and then you may not actually reap the benefits of investing in the project in the first place."
Wang suggests reforming HOPE VI to give public housing authorities increased flexibility to make projects happen — getting seed money to leverage, and operating more like the private sector. "If the government wants to cut them loose, they should have the flexibility the private sector has to meet a wider range of needs and to maintain a holistic approach."
Wang concluded that HOPE VI funds are key to maintaining successful public-private partnerships. "Without HOPE VI, private developers — which are key partners in this process — are likely to look elsewhere for simpler, more profitable opportunities... a huge loss to what the industry has learned over the last few years. If the federal government hopes for housing authorities and their partners to continue development work, there must be some resources for them to leverage."
Kristen Wang came to MIT/CRE after working as an AssistantProject Manager for BRIDGE Housing Corporation, a regional nonprofit affordable housing developer in San Francisco. She is currently finishing a master'sdegree in city planning at MIT.
The thesis work of MSRED graduate Harris Morrison involved studying the behavior and performance of different real estate investors. "In particular, I was interested in looking at market timing and return trends to see who leads whom — if private locals (investors) for instance led other types of investors in trends or capital flows," Morrison said.
After meeting Real Capital Analytics (RCA) President Robert White (White visited an MSRED finance class to discuss real estate investment), Morrison went on to base his thesis on data supplied by RCA — a set of over 41,000 real estate transactions for $5 million or more from 2000-2006. Morrison was able to put 85 percent of buyers and 77 percent of sellers into one of eight categories:
- private local investors — the largest investor group
- private national
- institutional — anything from endowments to insurance companies
- public real estate investment trusts (REITs)
- foreign investors — offshore investors who invested in U.S. real estate
- users and others — entities that carried out company operations in their real estate
- condominium converters — this category is made up of the previous six types of investors; Morrison separated them out to investigate behavior during the "condo craze" of the last six years.
- syndicators — another specialized category of tenancy-in-common (TIC) investors
To examine investor performance, Morrison used the repeat-sale method to calculate investors' capital appreciation returns. Out of 41,000 properties, 10,070 had two points of sale, representing constant quality of building. Morrison examined value change over the properties' holding period to calculate a periodic growth rate for these properties. He used several statistical techniques to eliminate properties that had been improved to come up with a property-level index.
In most analyses, the periodic return from an investment is calculated from growth and income. However, the income data was not readily available to Morrison "That's private information. People are not willing to share it," Morrison said. "Only 33 percent of the properties had a capitalization rate." This information gap compelled Morrison to look only at growth. This approach had advantages; compared to stable income from long-term leases, growth is more volatile, more subject to market forces, and shows investors' skill at timing purchases and sales.
Market Leaders Change over Time
Over the last six years, different investors have shown markedly different returns at different times. Private local investors, REITs and foreign investors have all led the markets at different times — times that corresponded with high returns. REITs had excellent returns in 2002, and then a "stair-step" climb of increases and plateaus. Foreign investors surged in 2003, and then lagged — perhaps due to fluctuations in foreign currency markets. "Foreign investors were tremendous consumers of real estate, and as they continued to buy real estate that was becoming higher and higher priced, their returns tapered off," Morrison said. Private local investors had very strong returns in 2004 and 2005. "They are by nature risk-takers," Morrison said. Institutions' returns lagged throughout the five-year study period, but they were also the least volatile as well — in keeping with institution investment low-risk, low-reward investment strategies, according to Morrison.
Private Local Investors Lead Returns — Others Lag
Private local investors earned higher returns throughout the entire five-year period. Their peaks tended to lead others' returns by anywhere from 3-6 months. "Private local investors typically have asymmetric knowledge. They know something about the market that no one else knows, and they're able to jump in quickly and take advantage of return opportunities. Other investors are more fiduciary — they're managing other peoples' money, and they typically have to move more slowly, take less risk, and are not as nimble in their decision-making ability," Morrison said. "Private local investors probably take on greater risk, and therefore earn greater returns."
Morrison found that private local buyers and sellers made far more transactions than any other investor type in the market. "Private local investors play a very large role in the real estate investment market," Morrison said, though he contended that they tend to concentrate on smaller holdings. Private local buyers represented 65 percent of deals under $5 million, but less than 5 percent of buyers for properties worth over $100 million.
Private locals also had the smallest median transaction prices — $10 million per transaction — while several other investor types had a median transactions twice as high. Despite the fact that private local investors had the smallest median transaction price, they had the highest total transaction volume. Private local investors were the largest net sellers of real estate, with users and others coming in second.
Morrison concluded by saying that investors face many challenges in making rational real estate decisions. Decentralized markets are full of unique assets that cannot be easily compared, and information asymmetries abound. "Real estate is a very local business. Local people know more than others," Morrison said. Properties are expensive assets with high transaction costs and holding costs. "Investors see value very differently, if they have an opportunity to bid on a property at all — they might not even know it's available," said Morrison.
J. Harris Morrison came to MIT/CRE after working at Grubb & Ellis/Bissell Patrick in Charlotte, NC, where he focused on sales and leasing of industrial properties. Prior to brokerage, he worked for two years in office development and property management at The Bissell Companies, a full-service real estate firm in Charlotte.