Research Support School of Business Administration THE EQUITY REIT UNIVERSE 1985-1992: THE ROLE OF PROPERTY TYPE AND SIZE Working Paper #9411-17 Dennis R. Capozza University of Michigan and Sohan Lee University of Michigan June 1994

I i I i

The Equity REIT Universe 1985-1992: The Role of Property Type and Size Dennis R. Capozza and Sohan Lee University of Michigan Abstract In this paper we report initial resultsfrom our National Equity REIT Database project. We provide descriptive statistics on expense ratios, leverage, total assets, and diversification of properties. We then categorize the sample by property type and size and use two-way analysis of variance to highlight significant differences among the categories. REITs are similar to closed endfunds and trade at discounts or premiums from the net asset value of the underlying properties. Retail REITs trade at significant premiums relative to the average REIT while warehouselindustrial REITs trade at discounts. Small REITs trade at significant discounts while large REITs trade at premiums. The discounts and premiums from net asset value, however, do not translate into higher cash flow yields. The most dramatic change in the real estate industry in the last few years is the rapid increase in the securitization of equity in real estate through Real Estate Investment Trusts (REITs). Many factors have contributed to the explosive growth of REIT initial and secondary public offerings. Often cited factors include changes in the capital requirements for commercial lenders that have made mortgage loans more costly and changes in tax laws that reduced the tax favored status of other ownership forms like real estate limited partnerships. The evolution of the REIT industry is now proceeding so rapidly that research and analysis will inevitably lag behind. In this paper we report initial results from our National Equity REIT Database project. The project will attempt to chronicle and analyze the evolution of the industry and will address some fundamental issues in financial economics. REITs are an unusual laboratory for financial analysis for two reasons. First, the value of the underlying real estate assets can be estimated with much greater precision than other corporate assets since real estate assets also trade individually as properties. As a result we need not resort to imperfect measures of asset value like book value. Second, REITs are not taxed at the corporate level. Issues like optimal capital structure can be studied independent of the confounding effects of taxes. Our goal in this paper is to provide descriptive statistics on the industry in four ways. First we will outline the size, property types, income, expenses, and diversification of the equity REIT universe from 1985 -92. The universe consists of 75 REITS but not for every year. The sample grows from 33 to 72 REITs during this period. Second, we provide similar data for the 32 continuing REITs for which data is available in all eight years. This allows us to separate the effect of entry to and exit from the industry on the aggregate statistics. The next two sections segment the data by property type and size class. Each REIT is classified as an apartment, warehouse, retail, or office, REIT if more than 50% of the properties owned are of one type. The size segmentation uses quartiles derived from total assets. In each segmentation we investigate differences in expenses, cash flow yield, diversification, and capital structure. To preview the conclusions, we find that expenses as measured by the ratio of general and administrative (G&A) expenses to total assets, remained constant during the period. Diversified REITs and small REITS have above average expense ratios. Leverage as measured by total liabilities / (total liabilities + market value of the equity) rose during the period. Large REITs and apartment REITs were more highly levered than average. By property type, apartment REITs are the most concentrated by location. Small REITs are more focused by property type. The stock market valuations of REITs as measured by premiums above the values of the underlying properties declined during the period. Wall Street (i.e., the stock markets) placed higher values on retail REITs and lower values on warehouse REITs relative to Main Street (the local property markets). In the size quartiles, small REITs trade at much larger discounts than large REITs. The differences are statistically significant.

June 3, 1994 (9:04am) The Equity REIT Universe 1985-1992: The Role of Property Type and Size Dennis R. Capozza and Sohan Lee School of Business The University of Michigan Ann Arbor, Ml 48109 313 764-1269 V

Capozza and Lee The REIT Universe 3 While discounts from net asset value are significantly below average for small REITs, cash flow yields are not significantly higher than those of larger firms. That is, discounts from net asset value are not being translated into higher earnings yields. The Data Criteria for Inclusion in the Sample The 1992 NAREIT (National Association of Real Estate Investment Trusts) sourcebook lists all publicly traded REITs (209 REITs) as of December 31, 1991. To focus on property, we exclude all mortgage, hotel, restaurant, and hospital REITs. REITs that do not trade on NYSE, AMEX, and NASDAQ or for which property information is not available are also discarded from our database. With these exclusions the sample consists of 75 REITs. The included REITs are listed in Table 1.0. Sample Size The 75 included equity REITs appear in the sample for at least one year. 416 total observations are available to study. 32 of the REITs appear in all eight years. We classify each REIT by property type when more than 50% of the property held is of one type (apartment, warehouse, retail, office). If no one property type is more than 50%, the REIT is classified as "Diversified." The number of REITs of each property type appears in Table 1.2. Retail REITs are the most common with 13-20 in each year. Apartment REITs are the least common with only 3-6 in each year. The jump in warehouse REITs in 1991 occurs because 18 Public Storage Equities partnerships converted to REIT status. Sources of the Data The data is compiled from 10-K filings of the REITs as well as CRSP data tapes. Most REIT specific information is obtained from 10-K reports, annual reports to shareholders, and proxy statements. Stock price data are retrieved from the CRSP daily return file. Metro and regional level property information is obtained from the NREI (National Real Estate Index) and Russell/NCREIF indices. Variables in the Database The database includes balance sheet, income statement, and property variables from the 10-K reports. The property data are classified by region using the eight economic regions as defined in Hartzell, Shulman, and Wurtzebach (1987). Table 1.1 provides descriptive statistics on the variables. Calculated variables are explained below. Net Asset Value To value the proerty holdings we assign property specific capitalization rates to each property and then calculate a weighted average capitalization rate based on each individual property's percentage of the total portfolio value. This weighted average capitalization rate is applied to the portfolio's current net operating income to derive a value for the entire portfolio. Each REIT's net asset value (NAV) is computed as follows: NAV = (market value of properties + other assets - total liabilities) / the number of shares outstanding. Value Weighted Premium The premium of each REIT is measured by (stock price - net asset value) / net asset value. Value weighted premium(VWPREM) is defined as follows: tPRE VWPRE4 I wy.prer, H (1) where nt wi=NAVitI/ NAVit, i=1 PREMh = (SP, - NAVk) / NAVE, NAV, = net asset value of REIT i at end of period t, SPi = stock price of REIT i at end of period t, and n, = the number of REITs with available PREMj and NAVj data at the end of period t. Concentration Indices To measure diversification/concentration, we use Hirschman-Herfindahl indices that are commonly used in industrial economics to measure monopoly power. We define three variables as follows: HHBOTH -:~ E i j (4) HHPROP HHRGN 2 Sand = I S 2, 'The index first acquired the name of Orris Herfindahl from work on energy in the 1950s and that of Albert Hirschman from earlier work on foreign trade patterns. See Hirschman (1964).

Capozza and Lee The RET Universe 4 where Si = the proportion of a REITs portfolio invested in property-type i and region Si= the proportion of a REITs portfolio invested in property-type i, and Sj = the proportion of a REITs portfolio invested in regionj. The above three indices measure how concentrated the properties in a REIT are, i.e., if the REIT is highly focused, the index is close to one, and if diversified, the index is close to zero. Cash Flow Yield We define cash flow as (net income + depreciation and amortization - gain on property sales + extraordinary expenses). The cash flow yield is measured by CFit/SPi,t.l, where CFi= cash flow of REIT i at end of period t, and Spit,-= stock price of REIT i at end of period t-1. The REIT Universe The Value of Property in REITs Table 1.2 describes the total property assets held by equity REITs. Book value of property assets tripled from 1985 to 1992. The square footage of retail space held by REITs doubled while the square footage of warehouse space quadrupled. The large increase in warehouse space is partly due to the conversion to REIT status of the 18 Public Storage Limited Partnerships in 1991. Figures 1.1 to 1.5 compare the proportions of property in REITs by type in 1985 and 1992 compared with the proportions in the NCREIF data and the total stock of property in the U.S. REITs are much more concentrated in retail property than either the NCREIF sample or the U.S. stock. Over half the property in REITs is retail in 1992 while only about 25% is retail in the NCREIF sample and the RREEF national estimate. Warehouse/mdustrial is underweighted in the REIT sample. The reasons for these under/overweightings are an interesting open question on which we provide some evidence below. Average REIT Size Table 1.5 outlines the average REIT size by property type and size quartile. The average REIT had $127 million of total assets during the period. By property type retail, apartment, and diversified REITs are above average in size while office and warehouse REITs are below average. The average small REIT (first quartile) has $29 million in assets. The large REITs (fourth quartile) average $279 million. Property Characteristics Total property assets average $95 million per REIT over the period (see table 1.6). The book value of the properties climbs from 71% of the property market value to 91% of market value. The average occupancy rate of the properties, reflecting general market trends, falls from 92% in 1985 to 87% in 1991 before recovering to 89% in 1992. The average cap rate rises from 8.8% to 9.4% also reflecting market trends. Figure 1.6 shows the distribution of property by economic region in 1992. Each economic region has a 14-18% share of total property except the Northeast and Farm regions that have less than 3% each. Income and Expenses REIT rental income tripled during the period to $1.3 billion (table 1.2). Total net income, on the other hand, declined through 90-91 before recovering in 1992. Average net income per REIT fell by nearly a half during the period. G&A expenses averaged 1.1% of total assets. Leverage As measured by (total liabilities)/(total liabilities + market value of the equity), leverage increased from 25% to 35% with a peak in 1990 at 48% when REIT stocks were highly depressed. ConcentrationlDiversification The Hirschman-Hirfindahl indices show conflicting trends in concentration. While concentration by property type increased (61 to 70), concentration by region declined (60 to 46). Overall (H Both) concentration declined slightly (41 to 35). Wall Street vs. Main Street Valuations of REIT stocks fluctuate widely from the value of the underlying properties during the sample period. On average REITs have traded at a discount of 8% from the net asset values of the properties; however the value weighted premium by year varies from a high of 13% in 1986 to a low of -36% in 1990. Wall Street's willingness to pay for securitized property fell by nearly 50% in just four years before

Capozza and Lee The REIT Universe 5 recovering in 1991-92. Survivor REITs The 32 REITs for which data is available in all years allow us to separate the effects of entry and exit from general market trends. Table 2 provides comparable data to table 1.6 for the survivor sample. The survivor REITs are about 20% larger than the full sample. Occupancy rates, expense ratios and cap rates are similar to the full sample. Net income, however, does not show the steep decline of the full sample suggesting that REITs that entered during the period were weaker than the survivors. Survivor REITs are also more highly leveraged and a little less focused by property type. In the next two sections we test whether there are significant differences among the REIT property type and size categories. In particular we test whether the premiums to net asset value differ significantly among types and if any differences in premiums to net asset value spill over into cash flow yields. If the premium differences are justified by the cost differences, cash flow yields should not differ significantly. REITs by Property Type Table 3 provides cross tabs on some key variables by REIT property type. In each panel we provide F-tests for significant variation in the relevant variable by year and by property type. In addition we test whether any of the category averages are significantly different from the overall sample mean and indicate those that are by asterisks. Premium to Net Asset Value In Panel A the premiums and discounts from net asset value are listed. Warehouse REITs are discounted the most heavily while retail REITs are least discounted. The differences from the sample averages are significant for these two categories. Leverage In panel B we see that apartment REITs are significantly more highly levered than average while diversified and office REITS are below average during the sample period. The F statistics for both the yearly effect and the property-type effect are significant. Concentration Panels C and D provide the concentration ratios. As might be expected diversified REITs show low concentration levels by property type. Office, warehouse and retail REITs are significantly more concentrated by property type than the average REIT. The yearly differences are not significant. By region the pattern is quite different Warehouse and diversified REITs have low focus by region while apartment REITs are highly focused. Again the yearly differences are not significant. Expenses In panel E we find that expense levels are high for diversified REITS and low for retail and apartment REITS. The yearly differences are not significant. Cash Flow Yield While some REIT types are more heavily discounted from the value of the underlying property, we might expect that in efficient markets these discounts would reflect lower earnings potential. Cash flow yields, however, might also reflect higher risk levels. We know from the previous panels that apartment REITs are more highly leveraged and that diversified REITs have higher expenses. Apartment REITs are also less diversified and carry more local market risk in addition to the higher financial risk. Therefore we might expect to see higher cash flow yields on apartment REITs. Panel F displays the cash flow yields by type. The differences among types are not significant Therefore there is no evidence that NAV premiums or risk levels are affecting the cash flow yields. However, the yearly variation is significant. Size Quartiles It is well known that many equity market anomalies are related to size. For example Banz (1981) showed that returns from buying very small firms are 20% higher than for very large firms. Roll (1981) and Reinganum (1981) present evidence that the small firm effect is partly due to errors in estimating the risk (beta) of small firms; but the effect remains even when the estimation problems are corrected. Stoll and Whaley (1983) argue that, given the differences in transactions costs between small and large firms, a round-trip transactions cost every three months is enough to eliminate the small firm effect. Keim (1983) provides evidence that 25% of the size effect occurs during the first five trading days in January. In this section we investigate whether size is related to other REIT characteristics. Premium to Net Asset Value Panel A of table 4 illustrates the dramatic effect of

Capozza and Lee The REIT Universe 6 size on the premium to net asset value of the properties. There is a monotonic increase in the premium as we move to the larger size quartiles. Small REITs (first quartile, average $29 million) are discounted 33% more than large REITs (fourth quartile, average $279 million). Both the size effect and the yearly effect are highly significant. Leverage In panel B we see that large REITs (fourth quartile) are more highly leveraged than the small REITs (first quartile). The yearly effect is also significant. Concentration Panel C: Small REITs (quartile 1) are significantly more highly concentrated by property type. The next two quartiles are less concentrated than the sample. There are no significant differences either by size or year in the regional concentration indices (panel D). Expenses Panel E: Small REITs are almost twice as costly to administer as large REITs. The G&A ratio for small REITs is 1.7% while for large REITs it falls to.9%. This accounts for some of the large discount from net asset value for small REITs. The yearly differences are not significant. Cash Flow Yield Panel F: Since small REITs have lower financial risk (less leverage) and less local market risk (more diversified by region) we might expect lower cash flow yields. However, we find instead that cash flow yields for small REITs are higher than for other REITs but not significantly so. The yearly differences, on the other hand, are significant. Summary and Conclusions In this paper we have described some preliminary results from the National Equity REIT Database project at the University of Michigan. This project will chronicle the evolution of the industry and analyze some basic issues in financial economics using REIT data. Here we have provided descriptive statistics on REIT property holdings as well as additional categorization by REIT property type and size. We find that expenses as measured by the ratio of G&A to total assets, remained constant during the period. As might be expected, diversified REITs and small REITS have above average expense ratios. Leverage rose during the period. Large REITS and apartment RElTs were more highly levered than average. By property type apartment REITs are the most concentrated by location. Small REITs are more focused by property type. The stock market valuations as measured by premiums above the values of the underlying properties declined during the period. Discounts from net asset value are higher for warehouse REITs and small REITs. Retail REITs sell at premiums relative to the average REIT. Cash flow yields, on the other hand, are not significantly different among REITs. Therefore we are unable to find evidence that differences in premiums to net asset value or in risk levels affect cash flow yields. These results may help to explain why retail property is over represented and warehousefindustrial is underrepresented in REITs. Retail property, once securitized sells at a premium while the opposite is the case for warehouse/industrial. The discounts/premiums do not affect cash flow yields. For unknown reasons, securitization adds value to retail property but destroys value for warehousefmdustrial property. References Banz, Rolf W., "The relationship between return and market value of common stock," J. of Financial Economics, 9 (1981), 3-18. Hartzell, David J., David G. Shulman, and Charles H. Wurtzebach, "Refining the analysis of regional diversification for income-producing real estate," J. of Real Estate Research, 2 (1987), 85-95. Hirschman, A. 0, "The paternity of and index," American Economic Review, 54 (1964), 761. Keim, Donald B., "Size related anomalies and stock return seasonality: further empirical evidence," J. of Financial Economics, 12 (1983). Reinganum, Marc R., "Misspecification of capital asset pricing: empirical anomalies based on earnings yields and market values," J. of Financial Economics, 9 (1981), 27 -35. "The anomalous stock market behavior of small firms in January: empirical tests for tax-loss selling effect," J. of Financial Economics, 12, 1 (1983). Roll, Richard, "A possible explanation of the small frm effect," J. of Finance, 31 (1981) 879-888.

Capozza and Lee The RET Universe 7 Stoll, Hans R. and Robert E. Whaley, 'Transaction costs and the small firm effect," J. of Financial Economics, 12, 1 (1983). Table 1.0 The 75 Included REITs Starred REITS are in the survivor sample of 32 REITs *B R E PROPERTIES INC BERKSHIRE REALTY CO INC *BRADLEY REAL ESTATE TRUST BURNHAM PACIFIC PROPERTIES INC *CALIFORNIA REAL ESTATE INVT TR CEDAR INCOME FUND LTD CEDAR INCOME FUND 2 LTD CHICAGO DOCK AND CANAL TRUST *CLEVETRUST REALTY INVESTORS *CONTINENTAL MORTGAGE & EQTY TR COPLEY PROPERTY INC COUSINS PROPERTIES INC DIAL REIT INC DUKE REALTY INVESTMENTS INC *E Q K REALTY INVESTORS 1 *EASTGROUP PROPERTIES *FEDERAL REALTY INVESTMENT TRUST *FIRST UNION REAL EST EQ&MG INVTS GRUBB & ELLIS REALTY INC TRUST *H R E PROPERTIES I C M PROPERTY INVESTORS INC 'I R T PROPERTY CO INCOME OPPORTUNITY REALTY TRUST KOGER EQUITY INC LANDSING PACIFIC FUND LINPRO SPECIFIED PPTYS *M G I PROPERTIES INC *M S A REALTY CORP *MERIDIAN POINT REALTY TR 83 *MERIDIAN POINT REALTY TR 84 MERIDIAN POINT REALTY TRUST IV MERIDIAN POINT REALTY TRUST VI MERIDIAN POINT REALTY TRUST VII MERIDIAN POINT REALTY TRUST VIII *MERRY LAND & INVESTMENT INC MONMOUTH REAL ESTATE INVT CORP *NEW PLAN RLTY TRUST *NOONEY REALTY TRUST INC *ONE LIBERTY PROPERTIES INC P S BUSINESS PARKS INC PARTNERS PREFERRED YIELD INC PARTNERS PREFERRED YIELD II PARTNERS PREFERRED YIELD III *PENNSYLVANIA REAL EST INVT TR *PROPERTY TRUST AMER *PRUDENTIAL REALTY TRUST PUBLIC STORAGE PROPERTIES VI PUBLIC STORAGE PROPERTIES VII PUBLIC STORAGE PROPERTIES VIII PUBLIC STORAGE PROPERTIES IX INC PUBLIC STORAGE PROPERTIES X INC PUBLIC STORAGE PROPERTIES XI INC PUBLIC STORAGE PROPERTIES XII PUBLIC STORAGE PROPERTIES XIV PUBLIC STORAGE PROPERTIES XV INC PUBLIC STORAGE PROPERTIES XVI PUBLIC STORAGE PROPERTIES XVII PUBLIC STORAGE PROPERTIES XVIII PUBLIC STORAGE PROPERTIES XIX PUBLIC STORAGE PROPERTIES XX *REAL ESTATE INVESTMENT TRUST CA REALTY SOUTH INVESTORS INC. *SANTA ANITA RLTY ENTERPRISES SIZELER PROPERTY INVESTORS INC *TRAMMELL CROW REAL ESTATE INVS *TRANSCONTINENTAL RLTY INVSTRS *U S P REAL ESTATE INVESTMT TRUST *UNITED DOMINION REALTY TR INC VANGUARD REAL ESTATE FUND I VANGUARD REAL ESTATE FUND II VINLAND PROPERTY TRUST *WASHINGTON REAL EST INVT TR *WEINGARTEN REALTY INVESTORS *WESTERN INVESTMENT REAL EST TR WETTERAU PROPERTIES INC

TABLE 1.1 The REIT Universe 1985 -1992: Summary Statistics Total assets and property assets are book values. Total market assets are measured by (estimated market value of properties + other assets). The leverage ratio is defined as total liabilities / (total liabilities + market value of the equity). I _ - * _. _aK....... I- -... O.. 1 I VARIABLE_ TOTAL ASSETS ($ MIL.) PROP. ASSETS ($ MIL.) PROP. AST. / MKT. VALUE OF PROP. (%) TOT. AST. / TOT. MKT. AST. (%) AVG. OCC. RATE (%) WEIGHTED CAP. RATE (%) NET INCOME ($ THOU) G&A ($ THOU) CASH FLOW PER SHARE G&A / TOTAL ASSETS (%) CASH FLOW YIELD (%) LEVERAGE RATIO (%) HH INDEX BOTH (%) HH INDEX PROP. (%) HH INDEX RGN. (%) MEAN 126.8 94.7 85.2 87.0 89.5 8.9 3,963 1,390 1.03 1.1 8.0 35.0 39.6 65.3 55.6 MAX 603.8 485.7 201.0 166.0 100.0 10.6 49,446 15,418 4.76 7.5 78.0 96.0 100.0 100.0 100.0 MIN 2.1 2.1 14.0 20.0 57.0 7.4 (58,609) 0 0 0.0 0.0 0.0 7.5 21.0 14.6 SiD. DEV. 110.2 85.3 33.0 26.0 7.0 0.5 9,209 1,478 0.64 1.1 10.0 26.0 21.4 23.5 27.2 VALUE WT. PREMIUM (DISCOUNT) (%) (8.3) 137.2 (93.6) 38.5

TABLE 1.2 The REIT Universe 1985 - 1992: Selected Characteristics by Year These statistics are based upon 75 REITs and 416 observations. Total assets and property assets are book values. YEAR TOTAL OFF. WARE. RETAIL APT DIVERSIFIED NET INCOME RENTS TOTAL TOTAL ASSETS PROPERTY ASSETS REITS REITS REITS REITS REITS REITS ($ MIL.) ($ MIL.) ($ MIL.) ($ MIL.) 1985 TOTAL 33 7 4 13 4 5 216 394 3,945 2,430 1986 TOTAL 36 8 2 14 4 8 237 488 4,483 3,069 1987 TOTAL 45 10 6 17 3 9 236 606 5,385 3,781 1988 TOTAL 52 14 7 20 4 7 189 732 6,474 4,628 1989 TOTAL 52 14 6 19 3 10 226 850 7,112 5,206 1990 TOTAL 53 15 8 20 4 6 147 910 7,284 5,685 1991 TOTAL 73 18 23 19 6 7 149 1,167 8,744 6,839 1992 TOTAL 72 16 25 15 6 10 248 1,265 9,300 7,759 YEAR SF OFF. SFWARE. SF RETAIL SFAPT OFF. VALUE WARE. VALUE RETAIL VALUE APT VALUE (THOU.) (THOU.) (THOU.) (THOU.) ($ L.) ($ MIL.) ($ MIL.) ($ MIL.) 1985 TOTAL 9,714 12,216 35,805 16,216 1,614 375 3,450 757 1986 TOTAL 11,015 17,337 44,219 16,306 1,796 552 4,294 778 1987 TOTAL 13,780 25,004 50,967 16,429 2,412 832 5,153 810 1988 TOTAL 17,650 27,389 57,911 18,740 3,044 971 6,113 939 1989 TOTAL 19,192 27,970 62,832 19,973 3,046 976 6,630 1,011 1990 TOTAL 20,812 31,963 67,240 23,366 2,999 1,106 7,264 1,193 1991 TOTAL 22,296 53,480 69,067 32,357 3,142 1,850 7,006 1,559 1992 TOTAL 23,401 51,921 75,142 45,751 2,916 1,668 7,104 1,936

TABLE 1.3 Total Market Values by Property Types in the Russell - NCREIF Data Base The Russell - NCREIF data base reflects the real estate ownership of tax-exempt institutional investors. Warehouse and R&D office value is combined. YEAR OFF. VALUE WARE. / R&D OFF. VALUE RETAIL VALUE APT VALUE ($ MIL.) ($ MIL.) ($ MIL.) ($ MIL.) 1985 TOTAL 5,188 2,328 1,775 1986 TOTAL 6,454 3,147 2,225 1987 TOTAL 5,921 3,677 2,472 1988 TOTAL 6,460 4,402 3,217 773 1989 TOTAL 7,098 5,109 3,833 1,202 1990 TOTAL 8,580 5,946 5,071 1,798 1991 TOTAL 8,066 6,645 5,956 2,408 1992 TOTAL 7,828 6,549 6,233 2,836

FIGURE 1.1 Distribution of Property Type among the 1985 REITs WAREHOUSE (12.2%)> (55.7/o)-,'_ RETAIL Figure 1.2 Distribution of Property Type among the 1985 Russell-NCREIF Data Base RETAIL (19.1%) OFFICE - (55.8%) WAREHOUSE (25.1%)

FIGURE 1.3 Distribution of Property Type among the 1992 REITs APT (12.2%) WAREHOUSE (52.1%) RETAIL Figure 1.4 Distribution of Property Type among the 1992 Russell-NCREIF Data Base APT (33.4%) OFFICE (26.6%)RETAIL ~ (27.9%) WAREHOUSE FIGURE 1.5 National Distribution of Institutional-Grade Property in 1992 These values are based on estimates by RREEF research. APT (33.0%) OFFICE (24.0%) RETAIL (28.0%) WAREHOUSE

TABLE 1.4 Location of the Properties by Eight Economic Regions The property data are classified by region using the eight economic regions as defined in Hartzell, Shulman, and Wurtzebach (1987). The eight economic regions are Northeast (NE), Mid-atlantic (MA), Southeast (SE), Southwest (SW), Midwest (MW), Farm (FARM), Southern california (SCAL), and Northern california (NCAL) region. YEAR NE MA SE MW FARM SW SCAL NCAL (%) (%)(%) (%) (%) (%) (%) (%) 1985 AVG 2.0 19.1 18.2 17.2 1.6 16.0 15.5 10.4 1986 AVG 2.2 18.0 20.9 16.8 2.1 15.4 15.6 9.0 1987 AVG 2.1 17.3 19.4 22.1 1.7 13.3 16.4 7.7 1988 AVG 2.0 15.3 20.6 22.4 1.9 14.9 14.7 8.2 1989 AVG 2.0 14.8 20.6 19.7 1.5 14.2 17.4 9.9 1990 AVG 2.8 13.6 20.2 19.3 1.9 14.5 17.5 10.2 1991 AVG 2.6 13.6 17.8 17.9 1.6 14.3 17.5 14.6 1992 AVG 2.5 13.7 16.7 18.1 1.6 15.7 16.6 15.1 1985 -1992 WT. AVG 2.3 15.2 19.1 19.2 1.7 14.8 16.6 11.2 FIGURE 1.6 Location of Properties in 1992 N ( (16.6%)SCAL (15 SW NE CAL (13.7%) k MA (16.7%) SE (18.1%) MW (1.6%) FARM

TABLE 1.5 The REIT Universe 1985 - 1992: Selected Characteristics by Year These statistics are based upon 75 REITs and 416 observations. Total assets are book values. YEAR OFF. REITS WARE. REITS RETAIL REITS APT REITS DIV. REITS TOT. AST. ($ MIL.) TOT. AST. ($ MIL.) TOT. AST. ($ MIL.) TOT. AST. ($ MIL.) TOT. AST. ($ MIL.) 1985 AVG 121 101 117 99 121 1986 AVG 75 114 141 112 154 1987 AVG 84 69 144 163 134 1988 AVG 108 67 141 149 153 1989 AVG 109 78 168 203 130 1990 AVG 109 87 167 197 139 1991 AVG 98 56 185 193 144 1992 AVG 102 51 229 231 156 1985 - 1992 WT. AVG 102 64 162 173 142 YEAR FIRST QUARTILE SECOND QUARTILE THIRD QUARTILE FOURTH QUARTILE TOT. AST. ($ MIL.) TOT. AST. ($ MIL.) TOT. AST. ($ MIL.) TOT. AST. ($ MIL.) 1985 AVG 36 75 110 236 1986 AVG 31 76 134 258 1987 AVG 27 70 124 247 1988 AVG 29 79 133 257 1989 AVG 34 83 142 289 1990 AVG 31 81 128 298 1991 AVG 25 56 111 278 1992 AVG 24 55 116 323 1985- 1992 WT. AVG 29 70 124 279

TABLE 1.6 The REIT Universe 1985 - 1992: Selected Characteristics by Year These statistics are based upon 75 REITs and 416 observations. Total assets and property assets are book values. Total market assets is measured by (estimated market value of properties + other assets). G&A is general and administrative expenses. The leverage ratio is defined as total liabilities / (total liabilities + market value of the equity) Weighted cap. rate, cash flow yield, HH index, and value weighted premium are defined in the paper. YEAR TOTAL ASSETS PROP. ASSETS PROP. ASSETS/ TOTAL ASSETS/ AVG. OCC. WEIGHTED NET INCOME ($ MIL.) ($ MIL.) MKT. VALUE OF PROP. (%) TOT MKT. AST. (%) RATE (%) CAP. RATE (%) ($ THOU.) 1985 AVG 119.6 73.6 70.5 76.1 92.1 8.8 6,534 1986 AVG 124.5 85.2 77.5 80.6 91 8.8 6,582 1987 AVG 119.7 84.0 82.6 84.1 90.5 8.7 5,252 1988 AVG 124.5 89.0 86.5 88 91.1 8.6 3,636 1989 AVG 136.8 100.1 88.2 88.4 90.1 8.7 4,352 1990 AVG 137.4 107.3 87 88.1 88.9 8.8 2,777 1991 AVG 119.8 93.7 89.6 91.3 86.8 9.1 2,037 1992 AVG 129.2 107.8 90.9 92.5 89.1 9.4 3,448 1985 - 1992 WT. AVG 126.8 94.7 85.2 87.0 89.5 8.9 3,963 YEAR G&A / CASH FLOW LEVERAGE HH INDEX HH INDEX HH INDEX VALUE WEIGHTED TOT. AST. (%) YIELD (%) RATIO (%) BOTH (%) PROP. (%)) RGN. (%) PREMIUM (DISCOUNT) (%) 1985 AVG 1.0 8.4 25 41.1 60.8 59.7 0.8 1986 AVG 1.3 7.3 26 40.6 62.0 56.3 13.4 1987 AVG 1.2 6.2 31 40.3 61.6 50.4 2.3 1988 AVG 1.1 6.8 34 44.5 66.1 62.7 1.0 1989 AVG 1.1 6.7 39 41.3 63.9 52.2 (4.7) 1990 AVG 1.1 8.3 48 40.2 62.5 55.1 (35.9) 1991 AVG 1.0 10.8 36 37.2 69.8 48.7 (23.4) 1992 AVG 1.0 10.2 35 35.2 69.5 46.4 (9.7) 1985-1992 WT. AVG 1.1 8.0 35 39.6 65.3 55.6 (8.3)

TABLE 2 The Survivor REITs 1985 - 1992: Selected Characteristics by Year These statistics are based upon 32 REITs and 256 observations. Total assets and property assets are book values. Total market assets is measured by (estimated market value of properties + other assets). G&A is general and administrative expenses. The leverage ratio is defined as total liabilities / (total liabilities + market value of the equity) Weighted cap. rate, cash flow yield, HH index, and value weighted premium are defined in the paper. YEAR TOTAL ASSETS PROP. ASSETS PROP. ASSETS/ TOTAL ASSETS/ AVG. OCC. WEIGHTED NET INCOME ($ MIL.) ($ MIL.) MKT VALUE OF PROP. (%) TOT MKT. AST. (%) RATE(%) CAP. RATE (%) ($ THOU.) 1985 AVG 119.9 74.7 70.5 76.1 92.4 8.8 6,499 1986 AVG 134.1 93.5 74.3 78.5 90.6 8.8 7,122 1987 AVG 143.0 100.8 78.3 79.5 90.4 8.7 6,148 1988 AVG 149.5 107.6 81.1 82.0 90.9 8.6 5,530 1989 AVG 164.0 119.5 79.2 81.3 90.8 8.6 5,921 1990 AVG 161.7 125.2 79.1 81.5 88.7 8.8 4,142 1991 AVG 170.1 130.2 82.4 83.9 88.7 9.1 4,553 1992 AVG 184.2 152.3 84.7 85.8 89.7 9.4 5,087 1985 - 1992 WT. AVG 153.3 113.0 78.7 81.1 90.2 8.8 5,625 YEAR G&A / CASH FLOW LEVERAGE HH INDEX HH INDEX HH INDEX VALUE WEIGHTED TOT. AST. (%) YIELD (%) RATIO (%) BOTH (%) PROP. (%) RGN. (%) PREMIUM (DISCOUNT) (%) 1985 AVG 1.0 8.2 25 40.6 60.7 60.5 0.8 1986 AVG 1.3 7.4 27 37.5 61.3 57.5 15.5 1987 AVG 1.1 6.3 32 36.0 59.9 55.6 (0.8) 1988 AVG 1.1 7.2 33 38.4 63.1 56.6 1.3 1989 AVG 1.1 6.7 40 37.2 61.5 55.8 (4.7) 1990 AVG 1.1 8.2, 47 36.9 61.0 56.4 (33.7) 1991 AVG 1.0 10.4 47 38.0 61.3 56.6 (16.4) 1992 AVG 1.0 8.8 42 36.8 61.1 55.1 1.0 1985 - 1992 WT. AVG 1.1 7.7 36 37.7 61.2 56.8 (4.7)

TABLE 3 Reits by Property Type We classify each REIT by property type when more than 50% of the property held is of one type. If no one property type is more than 50%, the REIT is classified as "diversified". The leverage ratio is defined as total liabilities/ (total liabilities + market value of the equity). Value weighted premium and cash flow yield are defined in the paper. G&A is general and administrative expenses. HH index is Hirschman-Herfindahl index. We use two-way analysis of variance (without interaction terms) to test if there are significant yearly or property-type effects, and then test if each category's mean is different from the overall mean. Asterisks (***,**,*) denote two-tailed significance at the 1%, 5%, and 10% levels, respectively PANEL A YEAR OFF. REITS WARE. REITS RETAIL REITS APT REITS DIV. REITS VALUE WT. PREM. (DISC.) (%) VALUE WT. PREM. (DISC.) (%) VALUE WT. PREM. (DISC.) (%) VALUE WT. PREM. (DISC.) (%) VALUE WT. PREM. (DISC.) (%) 1985 AVG 8.4 8.2 (1.6) (23.6) 7.7 1986 AVG (4.2) 18.9 21.7 (7.4) 16.4 1987 AVG 12.5 (4.2) 5.7 5.4 (9.9) 1988 AVG 10.9 (24.4) 2.6 4.7 (1.5) 1989 AVG 16.2 (37.1) (1.7) (1.8) (12.5) 1990 AVG (31.6) (44.5) (32.7) (63.1) (29.7) 1991 AVG (34.1) (54.7) (8.1) (26.4) (30.0) 1992 AVG (38.2) (55.5) 5.5 26.4 (10.6) 1985- 1992 AVG (7.5) (24.2)(**) (1.1)(*) (10.7) (8.7) YEARLY EFFECT F-VALUE = 5.1 PROB. > F = 0.00 PROPERTY-TYPE EFFECT F-VALUE = 2.0 PROB. > F = 0.12 PANEL B YEAR OFF. REITS WARE. REITS RETAIL REITS APT REITS DIV. REITS LEVERAGE RATIO (%) LEVERAGE RATIO (%) LEVERAGE RATIO (%) LEVERAGE RATIO (%) LEVERAGE RATIO(%) 1985 AVG 12 35 27 33 27 1986AVG 18 35 30 45 17 1987AVG 27 40 32 49 23 1988AVG 31 44 29 57 28 1989AVG 31 56 33 69 44 1990 AVG 42 62 44 76 34 1991 AVG 40 22 42 60 41 1992 AVG 45 23 36 46 39 1985-1992 AVG 31(**) 39 34 54(***) 32(**) YEARLY EFFECT F-VALUE = 4.57 PROB. > F = 0.00 PROPERTY-TYPE EFFECT F-VALUE = 9.82 PROB. > F = 0.00

TABLE 3 Reits by Property Type (Continued) PANEL C YEAR OFF. REITS WARE. REITS RETAIL REITS APT REITS DIV. REITS HH PROP. INDEX (%) HH PROP. INDEX (%) HH PROP. INDEX (%) HH PROP. INDEX (%) HH PROP. INDEX (%) 1985 AVG 75.9 58.3 65.9 53.5 34.4 1986 AVG 78.7 70.2 67.5 54.8 37.2 1987 AVG 70.3 65.8 71.3 51.6 34.1 1988 AVG 68.5 61.3 73.4 80.7 37.0 1989 AVG 67.5 65.5 72.9 82.6 35.0 1990 AVG 64.1 56.0 71.9 68.9 31.7 1991 AVG 63.9 86.0 72.1 60.9 33.3 1992 AVG 64.4 83.4 77.7 62.6 34.7 1985-1992 AVG 69.2(***) 68.3(**) 71.6(***) 64.5 34.7(***) YEARLY EFFECT F-VALUE = 0.65 PROB. > F = 0.71 PROPERTY-TYPE EFFECT F-VALUE = 27.72 PROB. > F = 0.00 PANEL D YEAR OFF. REITS WARE. REITS RETAIL REITS APT REITS DIV. REITS HH RGN. INDEX (%) HH RGN. INDEX (%) HH RGN. INDEX (%) HH RGN. INDEX (%) HH RGN. INDEX (%) 1985 AVG 62.1 52.1 59.7 64.5 69.4 1986 AVG 57.7 34.9 60.5 86.1 58.4 1987 AVG 60.7 39.2 64.3 80.2 50.4 1988 AVG 57.1 44.0 66.4 98.2 45.6 1989 AVG 57.5 36.6 63.2 97.7 48.6 1990 AVG 55.1 34.8 62.8 79.8 59.3 1991 AVG 58.8 29.2 65.7 68.5 38.3 1992 AVG 56.0 32.3 57.5 66.7 48.0 1985 -1992 AVG 58.1 37.9(***) 62.5 80.2(***) 52.3(**) YEARLY EFFECT F-VALUE = 1.23 PROB. > F = 0.32 PROPERTY-TYPE EFFECT F-VALUE = 29.48 PROB. > F = 0.00

TABLE 3 Reits by Property Type (Continued) PANEL E YEAR OFF. REITS WARE. REITS RETAIL REITS APT REITS DIV. REITS G&A/TOT. AST. (%) G&A/TOT. AST. (%) G&A/TOT. AST. (%) G&A/TOT. AST. (%) G&A/TOT. AST.(%) 1985 AVG 0.9 0.6 1.0 0.9 1.2 1986 AVG 0.8 1.0 0.8 0.6 2.4 1987 AVG 1.4 1.3 1.0 0.5 1.6 1988 AVG 1.1 1.2 1.1 0.7 1.7 1989 AVG 1.1 1.3 0.9 0.6 2.1 1990 AVG 0.8 2.3 1.0 1.3 1.4 1991 AVG 1.2 0.9 0.9 1.1 1.2 1992 AVG 1.5 1.1 0.7 0.8 1.2 1985- 1992 AVG 1.1 1.2 0.9(**) 0.8(***) 1.6(***) YEARLY EFFECT F-VALUE = 0.53 PROB. > F = 0.80 PROPERTY-TYPE EFFECT F-VALUE = 6.04 PROB. > F = 0.00 PANEL F YEAR OFF. REITS WARE. REITS RETAIL REITS APT REITS DIV. REITS CASH FLOW YIELD (%)__ CASH FLOW YIELD (%) CASH FLOW YIELD (%) CASH FLOW YIELD (%) CASH FLOW YIELD (%) 1985 AVG 8.2 7.9 7.2 9.0 1986 AVG 6.2 4.7 7.6 7.6 8.3 1987 AVG 5.9 3.3 6.3 5.8 7.2 1988 AVG 5.9 6.5 7.2 7.0 7.1 1989 AVG 5.7 8.9 7.1 5.8 6.4 1990 AVG 7.4 9.0 8.4 10.6 8.4 1991 AVG 9.9 15.5 10.1 13.2 12.1 1992 AVG 10.3 12.8 8.4 8.2 9.7 1985- 1992 AVG 7.5 8.7 7.9 8.2 8.5 YEARLY EFFECT F-VALUE = 9.99 PROB. > F = 0.00 PROPERTY-TYPE EFFECT F-VALUE = 0.83 PROB. > F = 0.52

TABLE 4 REITs by Size Quartiles The size segmentation uses quartiles based on total assets. The leverage ratio is defined as total liabilities/ (total liabilities + market value of the equity). Value weighted premium and cash flow yield are defined in the paper. G&A is general and administrative expenses. HH index is Hirschman-Herfindahl index. We use two-way analysis of variance (without interaction terms) to test if there are significant yearly or property-type effects, and then test if each category's mean is different from the overall mean. Asterisks (***,**,*) denote two-tailed significance at the 1%, 5%, and 10% levels, respectively PANEL A YEAR FIRST QUARTILE SECOND QUARTILE THIRD QUARTILE FOURTH QUARTILE VALJEWPEMDIC___)VALUE WT. PREM. (DISC.) (%) VALUE WT. PREM. (DISC.) (%) VALUEWT.PREM.(DISC.)(%) VALUEWT.PREM.(DISC.)(%) 1985 AVG 4.1 (13.7) (4.7) 20.0 1986 AVG (1.8) (1.3) 17.2 37.4 1987 AVG (9.9) 3.3 (1.6) 18.3 1988 AVG (14.9) 7.3 0.7 5.7 1989 AVG (31.0) (7.1) 6.5 3.4 1990 AVG (42.9) (42.7) (28.1) (32.1) 1991 AVG (51.5) (34.4) (20.4) (3.6) 1992 AVG (49.9) (26.4) 0.5 18.4 1985 - 1992 AVG (24.7)(***) (14.4)(**) (3.7) 8.4(***) YEARLY EFFECT F-VALUE = 9.22 PROB. > F = 0.00 SIZE EFFECT F-VALUE = 13.60 PROB. > F = 0.00 PANEL B YEAR FIRST QUARTILE SECOND QUARTILE THIRD QUARTILE FOURTH QUARTILE LEVERAGE RATIO (%) LEVERAGE RATIO (%) LEVERAGE RATIO (%) LEVERAGE RATIO (%) 1985 AVG 16 24 29 30 1986 AVG 23 22 26 35 1987 AVG 29 27 22 44 1988 AVG 34 29 36 35 1989 AVG 45 36 40 35 1990 AVG 44 53 46 47 1991 AVG 25 30 46 44 1992 AVG 25 29 52 33 1985 - 1992 AVG 30(**) 31 37 38(**) YEARLY EFFECT F-VALUE = 4.36 PROB. > F = 0.00 SIZE EFFECT F-VALUE = 2.71 PROB. > F = 0.07

A. k TABLE 4 REITs by Size Quartiles (Continued) PANEL C YEAR FIRST QUARTILE SECOND QUARTILE THIRD QUARTILE FOURTH QUARTILE HH PROP. INDEX (%) HH PROP. INDEX (%) HH PROP. INDEX (%) HH PROP. INDEX (%) 1985 AVG 69.9 43.6 70.0 60.0 1986 AVG 70.5 52.1 61.0 64.3 1987 AVG 66.7 57.9 58.9 62.7 1988 AVG 74.0 61.2 57.2 72.2 1989 AVG 69.7 58.4 58.9 68.5 1990 AVG 67.5 61.0 53.6 67.7 1991 AVG 81.4 76.3 58.7 63.2 1992 AVG 83.1 74.5 55.3 65.2 1985- 1992 AVG 72.8(***) 60.6(**) 59.2(**) 65.5 YEARLY EFFECT F-VALUE = 1.08 PROB. > F = 0.41 SIZE EFFECT F-VALUE = 6.41 PROB. > F = 0.00 PANEL D YEAR FIRST QUARTILE SECOND QUARTILE THIRD QUARTILE FOURTH QUARTILE HH RGN. INDEX (%) HH RGN. INDEX (%) HH RGN. INDEX (%) HH RGN. INDEX (%) 1985 AVG 59.7 75.8 67.7 44.3 1986 AVG 56.3 81.7 59.5 45.9 1987 AVG 50.4 73.1 58.2 52.6 1988 AVG 62.7 59.5 56.4 63.7 1989 AVG 52.2 55.3 54.3 69.3 1990 AVG 55.1 54.9 52.0 66.4 1991 AVG 48.7 37.5 50.9 62.6 1992 AVG 46.4 30.1 47.5 67.4 1985 - 1992 AVG 53.9 58.5 55.8 59.0 YEARLY EFFECT F-VALUE = 0.82 PROB. > F = 0.58 SIZE EFFECT F-VALUE = 0.35 PROB. > F = 0.79

TABLE 4 REITs by Size Quartiles (Continued) PANEL E YEAR FIRST QUARTILE SECOND QUARTILE THIRD QUARTILE FOURTH QUARTILE G&A/TOT. AST. (%) G&A/TOT. AST. (%) G&A/TOT. AST. (%) G&A/TOT. AST. (%) 1985 AVG 1.6 1.6 0.7 0.8 1986 AVG 1.7 1.3 0.8 1.4 1987 AVG 1.5 1.7 1.3 1.0 1988 AVG 2.2 1.4 1.4 0.8 1989 AVG 2.1 1.4 1.7 0.7 1990 AVG 1.8 1.3 1.4 0.9 1991 AVG 1.3 1.0 1.2 0.9 1992 AVG 1.3 1.3 1.0 1.0 1985- 1992 AVG 1.7(***) 1.4 1.2 0.9(***) YEARLY EFFECT F-VALUE = 0.99 PROB. > F = 0.46 SIZE EFFECT F-VALUE = 10.46 PROB. > F = 0.00 PANEL F YEAR FIRST QUARTILE SECOND QUARTILE THIRD QUARTILE FOURTH QUARTILE CASH FLOW YIELD (%) CASH FLOW YIELD (%) CASH FLOW YIELD (%) CASH FLOW YIELD (%) 1985 AVG 8.6 7.1 8.4 8.9 1986 AVG 6.0 7.6 8.4 7.1 1987 AVG 6.2 6.0 6.0 6.6 1988 AVG 6.2 6.3 7.1 7.5 1989 AVG 7.9 6.1 6.0 7.0 1990 AVG 9.8 8.6 7.6 7.9 1991 AVG 11.4 9.0 12.2 10.2 1992 AVG 12.4 10.9 12.2 7.7 1985 - 1992 AVG 8.6 7.7 8.5 7.9 YEARLY EFFECT F-VALUE = 9.85 PROB. > F = 0.00 SIZE EFFECT F-VALUE = 1.20 PROB. > F = 0.33 rf