Manufacturing Stock

Document Sample
Manufacturing Stock Powered By Docstoc
					           The Stock Return Difference between
       Industrial REITs and Manufacturing Firms:
               Space Supply and Demand Effects

Executive Summary. This paper compares and studies         by Hwahsin Cheng*
the stock performance of industrial real estate invest-       Luis C. Mejia**
ment trusts (REITs) and other industrial companies. The       Charles C. Tu***
return difference between industrial REITs and manu-
facturing firms is modeled as a function of industrial
space supply and demand. A graphical analysis illus-       The industrial real estate market has been the
trates the connection between the manufacturing goods      subject of various research papers during the last
and the industrial space markets, showing that perform-    two decades. Researchers have attempted to ex-
ance differences between the two markets can be ex-        plain the performance of industrial real estate as-
plained by changes in space supply and demand. An em-      sets using property and market factors. One aspect
pirical analysis using aggregate industry data confirms     of the industrial real estate market that is impor-
this premise.
                                                           tant to investors seeking portfolio diversification
                                                           but has not been thoroughly investigated is the
                                                           stock return difference between industrial real es-
                                                           tate investment trusts (REITs) and manufacturing

                                                           Industrial REITs and manufacturing companies
                                                           have common systematic and sector components.
                                                           Therefore, their portfolio returns should have some
                                                           degree of correlation. However, although driven by
                                                           similar underlying factors, the returns of indus-
                                                           trial REITs and manufacturing firms do not ex-
                                                           hibit much co-movement over time. This paper
                                                           compares the stock returns of industrial REITs
                                                           with manufacturing companies, and examines the
                                                           difference between these two return series. The di-
                                                           vergence can be modeled as a function of industrial
                                                           space supply and demand factors. A graphical
                                                           framework shows that changes in space supply
                                                           and demand may affect the relative performance
                                                           of the industrial space and the manufacturing
                                                           goods markets. These effects are tested empirically
* Fannie Mae, Washington,    DC   20016   or   hwahsin     using aggregate industry data while controlling for                                       the funds flowing into REITs and the traditional
** Fannie Mae, Washington, DC 20016 or luis c mejia@       stock markets.
*** University of San Diego, San Diego, CA 92110-2492 or   The remainder of this paper is organized as fol-                                          lows. First, there is a brief literature review on the

                                                                       Journal of Real Estate Portfolio Management   249
Hwahsin Cheng, Luis C. Mejia, and Charles C. Tu

industrial real estate market. Second, the paper         and bond returns. At the same time, Gyourko and
discusses the stock return difference between in-        Keim (1992) find that stock returns move to some
dustrial REITs and manufacturing companies.              extent with real estate returns, while Ambrose,
Third, a graphical framework explains perform-           Ancel, and Griffiths (1992) find that equity and
ance differences between the manufacturing goods         mortgage REITs behave in some ways like the
and the industrial space markets. Fourth, there is       general stock market. Additionally, Mueller and
a discussion of an empirical test of the effect of       Pauley (1995) find that although REIT prices have
space supply and demand factors. Lastly, the paper       high negative correlations with interest rates in
closes with concluding remarks.                          falling interest rate periods, REIT price changes
                                                         are relatively flat in rising interest rate
Literature Review
                                                         More recently, the literature has studied the issue
Early research to understand industrial property         of REIT and traditional stock market integration.
performance was completed by Hoag (1980), who            Li and Wang (1995) support the integration be-
constructs an industrial real estate value index as      tween real estate and stock markets in a two-factor
a function of property and market factors. Gris-         asset-pricing framework. Okunev and Wilson
som, Hartzell, and Liu (1987) expand on Hoag’s           (1997) show a nonlinear relationship while reject-
research by explaining regional risk across indus-       ing a linear cointegration between the two mar-
trial properties. Wheaton and Torto (1990) follow        kets. Research by He (2002) concludes that real es-
with an industrial real estate investment model,         tate factors explain industrial stock returns and
defining industrial space production as an invest-        their significance is only second to the overall stock
ment decision driven by employment and cost of           market effect. Although results are conflicting re-
capital factors. Ambrose (1990), in turn, shows          garding real estate and stock market integration,
that industrial property asking prices and rents         it is recognized that real estate offers diversifica-
are explained largely by property attributes such        tion benefits, even when transacted in the form of
as building size, office space, and other physical        REIT shares.
                                                         One aspect of the literature that has not been fully
In a second group of studies, Atteberry and Ruth-        investigated is the stock return divergence be-
erford (1993) find that monetary base and indus-          tween industrial REITs and manufacturing com-
trial construction explain changes in industrial         panies. The hypothesis presented here is that the
real estate prices. Fehribach, Rutherford, and           divergence may be attributed, at least partially, to
Eakin (1993) extend the Hoag (1980) and Ambrose          changes in industrial real estate space supply and
(1990) studies by examining the effect of financial       demand. The next sections address this question
economic variables, in addition to physical char-        by reviewing relevant data and presenting graph-
acteristics, on industrial property prices. Hughes       ical and empirical analyses.
(1994) also adds to the literature by listing factors,
including labor force, infrastructure, and popu-
lation, that affect industrial property demand.
                                                         Return and Economic Data
Lastly, Lockwood and Rutherford (1996) find that
physical characteristics, market conditions, and lo-     This section illustrates the stock returns of indus-
cal factors affect industrial property value.            trial REITs and other industrial firms using value-
                                                         weighted indices based on 1-year holding period
On another front, researchers have attempted to          total returns collected from COMPUSTAT. In
explain the behavior of REIT stock returns. Liu,         addition, it presents aggregate economic data that
Hartzell, Greig, and Grissom (1990) find that the         may affect the difference between the two return
real estate stock market is segmented from the           series, such as construction put in place, industrial
stock market, while Liu and Mei (1992) conclude          production, and capacity utilization. Sources of the
that REIT returns are more predictable than stock        aggregate economic series include the U.S. Census

250   Vol. 12, No. 3, 2006
                                        The Stock Return Difference between Industrial REITs and Manufacturing Firms

                         Exhibit 1                                    Bureau, the Bureau of Labor Statistics, and the
Stocks Included in the Industrial REIT Portfolio                      Federal Reserve Board.
Ticker Symbol                        Company

AMB                                  AMB Property Corporation         Stock Returns of Industrial REITs and
CNT                                  Centerpoint Properties Trust     Manufacturing Firms
CTR.2                                Cabot Industrial Trust
                                                                      A portfolio of industrial REITs and one of manu-
DRE                                  Duke Realty Corp
                                                                      facturing firms were created to compare stock re-
EGP                                  Eastgroup Properties
                                                                      turns in the industrial space and the manufactur-
FR                                   First Industrial
                                                                      ing goods markets. The first portfolio includes 16
IND                                  American Industrial Properties
                                                                      industrial REITs, which are listed in Exhibit 1.
KTR                                  Keystone Property Trust Corp
                                                                      The manufacturing stock portfolio contains firms
LRY                                  Liberty Property Trust Corp
                                                                      with SIC codes between 2000 and 4000, and be-
MDN                                  Meridien Industrial Trust
                                                                      tween 4212 and 4231, resulting in a sample of
MPH.1                                Meridien Point VIII
                                                                      5,426 industrial stocks.1 One-year holding period
MONM                                 Monmouth Capital Corp
                                                                      return, including both dividend and capital appre-
MNRTA                                Monmouth RE Invest Corp
                                                                      ciation, is calculated every month for each portfolio
PAG                                  Pacific Gulf Properties
                                                                      based on the market value-weighted return of its
PLD                                  Prologis
                                                                      components. Although there is significant dispar-
TRI.1                                TriNet Corp Realty Trust
                                                                      ity in the size of the two samples, this is not con-
Note: The source for the ticker symbols is COMPUSTAT.                 sidered a major obstacle in the analysis.

                                                                      Exhibit 2 presents the stock return series of the
                                                                      industrial REIT and manufacturing firm indices

                                                              Exhibit 2
        One-Year Holding Period Stock Return for Industrial REITs and Other Industrial Firms:

            Source: Return series derived from COMPUSTAT return data.

                                                                                  Journal of Real Estate Portfolio Management   251
Hwahsin Cheng, Luis C. Mejia, and Charles C. Tu

                                                   Exhibit 3
                                       Industrial Construction Put in Place

         Source: U.S. Census Bureau.

between January 1993 and September 2005. The              introduced recently by the U.S. Census Bureau is
returns capture stock income and price changes.           used. This series is presented in Exhibit 3 for the
This graph shows the variability of the stock re-         1993–2005 period. Between 1993 and 1998, con-
turn spread between industrial REITs and other            struction recovers following the downturn of the
industrial companies during the sample period.            early 1990s. This recovery coincides with a period
Manufacturing firms outperformed industrial                of improving economic strength. The space supply
REITs around the 1995–1996, 1998–2000, and                series shows a slight decline between 1998 and
2003–2004 periods. Industrial REITs, on the other         1999, following periods of limited capital, recover-
hand, performed better than manufacturing firms            ing in 2000, before dropping significantly in 2001
around the 1993–1994 and 2001–2002 periods,               and 2002 as the economy weakens. The space sup-
which generally coincide with, or follow, times of        ply remains flat during 2003, recovering again in
economic weakness. The correlation between the            2004 and 2005 as economic conditions improve.
returns of industrial REITs and manufacturing
firms is only 0.10, indicating there is little corre-      As measures of industrial space demand, three
spondence between the two series.                         manufacturing activity variables are used: manu-
                                                          facturing employment, industrial production, and
                                                          capacity utilization. These variables are illustrated
Industrial Space Supply and Demand
                                                          in Exhibit 4. Between 1993 and 2000, industrial
The U.S. aggregate industrial construction put in         production maintains an increasing trend and ca-
place for manufacturing purposes is used to mea-          pacity utilization remains relatively stable. During
sure industrial space supply. Specifically, the new        most of this period, changes in manufacturing em-
national, non-residential, private, seasonally ad-        ployment are positive, although with a decline that
justed, annualized construction put in place series       led to negative employment changes in the 1998–

252   Vol. 12, No. 3, 2006
                                      The Stock Return Difference between Industrial REITs and Manufacturing Firms

                                                      Exhibit 4
     Change in Manufacturing Employment, Industrial Production, and Capacity Utilization


         Source: Federal Reserve Bank of St. Louis.

2000 period. Both industrial production and ca-              A preliminary examination of the data is useful as
pacity utilization decline between 2000 and 2001             background for the analysis of the stock return
as manufacturing employment slows down. They                 spread. The return spread of REITs over manufac-
remain flat during the 2002–2003 period, before               turing firms has a contemporaneous correlation of
increasing again after 2003 as the manufacturing               0.41 with industrial construction, suggesting the
employment recovers.                                         possibility of a space supply effect. At the same
                                                             time, the stock return spread has a contempora-
The interaction of industrial space supply and de-           neous correlation of 0.29 with changes in manu-
mand determines the level of vacancy in the in-              facturing employment, 0.22 with industrial pro-
dustrial space market. The quarterly industrial va-          duction, and 0.38 with capacity utilization. These
cancy rate reported by Grubb and Ellis is shown              correlations show that manufacturing firms tend
in Exhibit 5. The vacancy rate declines between              to outperform industrial REITs during periods of
1993 and 2000 as favorable industrial production,            economic expansion. The correlation between the
capacity utilization, and employment conditions              return spread and industrial vacancy is not clear
lead to positive space absorption. The vacancy rate          at the contemporaneous level. The contemporane-
takes an upward turn in 2000 as the manufactur-              ous correlation is 0.51, which may be counterin-
ing sector slows down but industrial construction            tuitive since it suggests that REITs tend to per-
activity remains high. The vacancy series partially          form better than manufacturing firm stocks in
levels between 2002 and 2003, and then declines              periods of higher vacancy. A reason for the lack of
again in 2004 as the manufacturing economy                   clarity in this measure is that a contemporaneous
starts to improve and construction activity re-              bivariate correlation does not capture the dynamic
mains low.                                                   interaction that REIT performance could have

                                                                         Journal of Real Estate Portfolio Management   253
Hwahsin Cheng, Luis C. Mejia, and Charles C. Tu

                                                  Exhibit 5
                                           Industrial Vacancy

            Source: Grubb & Ellis.

with past, current, and expected space supply and      The above data overview illustrates space supply
demand trends.                                         and demand factors that could affect the relative
                                                       performance of REITs and manufacturing stocks.
Stock and REIT Flow of Funds                           The next section shows a graphical framework
                                                       where changes in space supply and demand affect
In addition to space supply and demand factors,        the relative performance of industrial space sup-
the flow of funds into and out of the market also       pliers, which include REITs, and industrial space
plays a role in the performance of REITs and other     users, such as manufacturing firms.
stocks. Exhibit 6 shows quarterly data of the Flow
of Funds Accounts of the U.S. from the Federal Re-
serve Board. Specifically, these series represent net
                                                       Space Supply and Demand Effects
purchases of stocks and net increases in REIT eq-
uity. The stock market flow of funds is quite vola-     Consider a two-period framework as presented in
tile especially in the second half of the sample       Exhibit 7.2 Panel A shows the demand and supply
period. The REIT market flow of equity funds            of manufacturing goods, D0 and S0 at time 0. The
increases steadily and reaches the highest point in    intersection of D0 and S0 yields the initial equilib-
the fourth quarter of 1997. After the credit crunch    rium output q0 and unit price p0. Panel B shows
of 1998, the REIT market flow of funds declines         the demand and supply for industrial space, L0 and
sharply, but rebounds after 2000 when the tech-        C0 at time 0. The intersection of L0 and C0 yields
nology bubble bursts and investors start consider-     the initial equilibrium price of space r0 (i.e., the
ing REITs as a viable alternative to traditional       rent paid by manufacturers to use each unit of
stocks.                                                space s0). Also in Panel B is the existing stock of

254   Vol. 12, No. 3, 2006
                                     The Stock Return Difference between Industrial REITs and Manufacturing Firms

                                                           Exhibit 6
                                      Stock and REIT Market Flow of Funds

         Source: the Federal Reserve Board of Governors.

space, a0, which is the sum of leased space s0 and              demand depends on multiple factors, including
vacant space v0.                                                whether the developers are increasingly risk-
                                                                averse, are pessimistic about future manufactur-
In the manufacturing goods market (Exhibit 7,                   ing activity, or react cautiously to short-term in-
Panel A), manufacturers adjust their production                 creases in space demand. Assume that, because of
and capacity based on expectations about the de-                any or a combination of these factors, the devel-
mand for goods, E(D). Assume that manufacturers                 opers decide not to build additional space, such
expect the demand for their goods to increase at                that their supply function will not shift.
time 1, such that E(D1) is to the right of D0. The
expectation of stronger demand leads existing                   Exhibit 7 also illustrates the short-run result at
manufacturers to collectively increase production               time 1 when D1 materializes. Suppose that man-
and capacity, and attracts new producers into the               ufacturing goods demand at time 1 does not in-
market. The result is a new short-run supply func-              crease as expected by the manufacturers, instead
tion S0 and a new short-run production level q0.                remaining at the initial level, such that D1      D0
The expected short-run price at time 1 is thus                  as shown in Panel A. Since the increased produc-
E(p1).                                                          tion capacity at time 0 is partially irreversible
                                                                without a waiting time, the supply remains un-
In Exhibit 7, Panel B, the increased production and             changed at time 1, such that S1 S0. As a result,
capacity leads to a short-run space demand shift                the actual short-run price attained at time 1 is
from L0 to L0 and a higher space requirement, s0,               p1, instead of E(p1). Given D1 and S1, output is ad-
with short-run market clearing rent r0 and vacancy              justed from q0 to q1. Consequently, the space de-
v0. The developers’ decision to build or not to build           mand function at time 1 is L1 and the short-run
more space in response to the additional short-run              demand for space becomes s1. Since the developers

                                                                           Journal of Real Estate Portfolio Management   255
Hwahsin Cheng, Luis C. Mejia, and Charles C. Tu

                          Exhibit 7                                            The opposite, however, could also happen. If the
Short Term Equilibrium at Time t                              0 and t      1   developers are too optimistic about increases in
                                                                               manufacturing activity and space demand, they
Panel A: Manufacturing Goods Market
                                                                               might build excessive space, eventually causing
                                                                               higher vacancies and lower rents. Under that sce-
                                                                               nario, the developers’ decision to build additional
                                                                               space would negatively affect their performance
                                                                               elative to manufacturers.

        E(p1)                                                                  The above scenario indicates the possibility that
            p0                                                                 space supply and demand factors affect the rela-
                                                                               tive performance of industrial REITs and manu-
                                                                               facturing firms. It shows that changes in space
                                                                   E(D1)       supply and demand affect the performance of in-
                                                                               dustrial space suppliers, which include REITs and
                                                         D1= D0                industrial space users, such as manufacturing
                                                                               companies. In the context of the stock market, an
                                   q0    q1        q0
                                                                               implication would be that changes in space supply
Panel B: Industrial Space Market
                                                                               and demand affect the return performance of
                                                                               REITs relative to manufacturing firms. This view
                                                                               is tested empirically in the next section.

           r0                                                                  Empirical Analysis
          r1                                                                   The empirical analysis tests the effect of industrial
           r0                                                                  space supply and demand factors on the stock re-
                                                                   L   0       turn spread between industrial REITs and other
                                                              L1               industrial companies using a logistic regression
                                                         L0                    model. The probability that the stock returns of
                                                                               industrial REITs higher than the returns of man-
                                                                               ufacturing firms is estimated. Measuring the re-
                                   s0   s1 s0           a0                     turn spread in probabilistic terms reduces the
                                                                               noise associated with short term changes in the
                                                                               absolute spread.
                                                                               The probability, p, that the return of industrial
                                                                               REITs exceeds the return of manufacturing firms
                                                                               as a function of space supply and demand factors
                                                                               is estimated while controlling for the funds flowing
decided not to build additional space, which im-                               into the REIT and traditional stock markets. The
plies that C1 C0, then L1 yields a short-run mar-                              equation is specified contemporaneously, consis-
ket rent r1, shown in Panel B.                                                 tent with the notion of an efficient market where
                                                                               returns reflect current information. In addition,
The developers’ conservative short-run decision                                the equation includes space supply and demand
and the manufacturers’ failed expectations result                              factors but it does not include vacancy. As noted
in r1    r0 and p1   p0. In this example, the devel-                           earlier, vacancy is the result of interactions be-
opers’ decision not to build additional space, in                              tween space supply and demand, and as such is
light of the unrealized space demand, positively af-                           assumed to be implicitly captured through the var-
fects their performance relative to manufacturers.                             iables in the model.

256     Vol. 12, No. 3, 2006
                                           The Stock Return Difference between Industrial REITs and Manufacturing Firms

The probability that the return of industrial REITs                  A space supply effect on the stock return spread
exceeds the return of manufacturing firms is mod-                     between industrial REITs and manufacturing
eled through the logistic regression:                                firms is suggested by the negative and significant
                                                                     coefficient of the construction put in place variable,
            Rt     a0     a1d(CPP)         a2d(EMP)                  d(CPP). The model shows that as space supply ac-
                                                                     celerates, the probability of industrial REITs out-
                        a3d(IP)      a4d(CU)                         performing manufacturing companies diminishes,
                        a5d(FFS)       a5d(FFR)       .        (1)   controlling for space demand and flow of fund fac-
                                                                     tors. This result is consistent with the notion that
The dependent variable, R, is the logarithm of the                   the market reacts to changes in space supply by
odds that the stock return difference between the                    adjusting their expectations about the relative per-
industrial REIT portfolio and the manufacturing                      formance of REITs with respect to manufacturing
firm portfolio is positive. In a logistic regression,                 firms.
the technical label for R is the log odds ratio. A
probability can be extracted from such odds ratio                    The existence of a space demand effect is also in-
through an arithmetic calculation.3 Space supply,                    dicated through the significant coefficients of the
CPP, is measured through the construction put in                     changes in employment, d(EMP), industrial pro-
place for manufacturing firms in constant 1996                        duction, d(IP), and capacity utilization, d(CU). The
dollars. Space demand is measured through the                        coefficients of the changes in employment and ca-
changes in manufacturing employment, EMP; in-                        pacity utilization are positive, suggesting that as
dustrial production index for manufacturing firms,                    manufacturing employment and capacity utiliza-
IP; and the capacity utilization index for manufac-                  tion accelerate, the stock performance of industrial
turing firms, CU. To control for investor prefer-                     REITs improves relative to the stock performance
ences, the model also includes the flow of funds                      of manufacturing firms. Note that the signs of
into the REIT and traditional stock markets, FFR                     these coefficients are opposite to those indicated
and FFS, respectively.4 The explanatory variables                    earlier by the bivariate contemporaneous correla-
are specified as one-year differences, as indicated                   tions. Since the regression signs reflect the effect
by the notation d( ).5 The variable a represents the                 of a specific variable in the presence of other ex-
regression parameters and        is the error term.                  planatory factors, the estimation signs are rea-
Estimation results of the logistic regression are                    sonable. They indicate that improvement in em-
shown in Exhibit 8. Note that the estimation pa-                     ployment and capacity utilization enhances the
rameters represent the effect of each independent                    performance of developers relative to manufactur-
variable on the probability of industrial REITs out-                 ers, given a certain level of space supply, industrial
performing manufacturing firms. The results do                        production, and other factors. The coefficient of the
not indicate whether REITs and manufacturing                         change industrial production, in turn, has a neg-
firms have positive or negative returns.                              ative sign. This result is consistent with the notion
                                                                     that, holding capacity and other factors constant,
                                                                     increases in production come from more efficient
                           Exhibit 8                                 manufacturing processes. Such increase in effi-
                 Logistic Regression Results                         ciency, in turn, increases the chance that manufac-
Variable                     Parameter                    p-value
                                                                     turers perform better relative to developers.

Intercept                         3.7158                   0.0021
                                                                     The control variables representing the flow of
d(CPP)                            0.3446                   0.0005
                                                                     funds into the stock and REIT markets, d(FFS)
d(EMP)                            0.0046                   0.0117
                                                                     and d(FFR), also have reasonable signs. The coef-
d(IP)                             1.2229                   0.0001
                                                                     ficients are negative for the stock flow of funds and
d(CU)                             0.2839                   0.0585
                                                                     positive for the REIT flow of funds. The negative
d(FFS)                            0.0001                   0.0001
                                                                     stock flow of funds coefficient is consistent with the
d(FFR)                            0.0006                   0.0165
                                                                     notion that higher demand for stocks relative to

                                                                                 Journal of Real Estate Portfolio Management   257
Hwahsin Cheng, Luis C. Mejia, and Charles C. Tu

REITs reduces the probability of REITs outper-                       explanatory variable. The probability p is estimated as
                                                                     exp(R / 1 R).
forming manufacturing firms. Conversely, the pos-
                                                                   4. The Federal Reserve flow of funds series is generated quar-
itive REIT flow of fund coefficient indicates that                      terly. However, the model is estimated monthly. To be able
stronger demand for REITs positively affects the                      to use FFS and FFR as control variables, each quarterly
chance that industrial REITs have better perform-                     observation was split into three equal monthly values.
ance than manufacturing firms.6                                     5. For a variable x, the notation d( ) indicates xt xt 12. A 12-
                                                                      month lag is consistent with the 1-year measurement period
                                                                      of the industrial REIT and stock returns.
                                                                   6. In addition to the logistic regression presented in this paper,
Conclusion                                                            other specifications were tested, including: (a) OLS equation
                                                                      with a continuous return spread dependent variable. The
This paper shows that the disequilibrium between                      space supply and demand effects were generally in line with
the manufacturing goods and industrial space                          those shown in this paper. However, a continuous return
                                                                      spread variable seems to add noise to the model that affects
markets is partially explained by changes in space
                                                                      the significance of some coefficients; (b) logistic equation
supply and demand after controlling for the stock                     with lagged explanatory variables, under the argument that
and REIT flow of funds. The results are consistent                     space effects take time. However, results were not signifi-
with the notion that changes in current space sup-                    cant, arguably because past information already processed
                                                                      by the market does not necessarily explain current perform-
ply and demand could affect the relative perform-                     ance; and (c) equation system where one equation predicts
ance of REITs with respect to manufacturing                           vacancy as a function of lagged space demand and supply
firms. The results help investors recognize the im-                    and another equation explains the return spread as a func-
                                                                      tion of expected vacancy. Some lagged variables in the va-
portance of incorporating information about indus-                    cancy equation are not significant or have conflicting signs.
trial space supply and demand when making de-                         However, the results suggest the possibility that the space
cisions about the allocation of industrial REITs                      supply and demand effects are associated with vacancy ex-
                                                                      pectations resulting from current market conditions. Addi-
and other industrial stocks in their portfolios. Ex-
                                                                      tional research is necessary to understand the structure and
tensions of this analysis could include the use of                    significance of the space supply and demand lags and the
other space supply and demand variables that help                     formation of vacancy expectations that could help to explain
explain the stock return spread between industrial                    changes in the return spread.
REITs and other manufacturing companies, a dy-
namic specification with leads and lags that permit
the use of vacancy history and expectations, or a                  References
model that incorporates the recursive effect of                    Ambrose, B. An Analysis of the Factors Affecting Light Indus-
space supply, space demand, and vacancy.                           trial Property Valuation. Journal of Real Estate Research,
                                                                   1990, 5, 355–70.
                                                                   Ambrose, B., E. Ancel, and M. Griffiths. The Fractal Structure
                                                                   of Real Estate Investment Trust Returns: A Search for Evi-
Endnotes                                                           dence of Market Segmentation and Nonlinear Dependency.
1. The sample includes domestic as well as foreign manufac-        Journal of the American Real Estate and Urban Economics As-
   turing corporations. Among the largest companies in the in-     sociation, 1992, 20, 25–54.
   dustrial stock sample are Novartis, Toyota Motor Corpora-       Atteberry, W. L. and R. C. Rutherford. Industrial Real Estate
   tion, Exxon Mobil, Royal Dutch Petroleum, Coca-Cola,            Prices and Market Efficiency. Journal of Real Estate Research,
   Procter & Gamble, Merck & Co, Johnson & Johnson, Lilly          1993, 8, 377–85.
   (Eli) & Co, British Petroleum, Pepsico Inc., Abbott Labora-     Fehribach, F. A., R. A. Rutherford, and M. A. Eakin. An Anal-
   tories, DuPont De Nemours, Wyeth, and Bristol Myers             ysis of the Determinants of Industrial Property Valuation.
   Squibb. Some companies on the list are headquartered in         Journal of Real Estate Research, 1993, 8, 365–76.
   foreign countries but have manufacturing plants or oil refin-
   eries in the U.S.                                               Grissom, T., D. Hartzell, and C. Liu. An Approach to Industrial
                                                                   Real Estate Market Segmentation and Valuation Using the Ar-
2. In economic terms, this is a partial equilibrium framework      bitrage Pricing Paradigm. Journal of the American Real Estate
   with competitive profit-maximizing manufacturers and             and Urban Economics Association, 1987, 15, 199–219.
                                                                   Gyourko, J. and D. Keim. What Does the Stock Market Tell Us
3. In the logistic regression, the predicted variable, R, is the   About Real Returns? Journal of the American Real Estate and
   logarithm of the odds that industrial REIT returns are          Urban Economics Association, 1992, 20, 457–86.
   higher than manufacturing firm returns. The odds are de-
   fined as p / (1 p), where p is the probability that the stock    He, L. T. Excess Returns of Industrial Stocks and the Real
   returns of industrial REITs surpass those of manufacturing      Estate Factor. Southern Economic Journal, 2002, 68, 632–45.
   firms. The estimated coefficient, a, represents the change        Hoag, J. W. Towards Indices of Real Estate Value and Return.
   in the odds as a result of a change in the corresponding        Journal of Finance, 1980, 35, 569–80.

258    Vol. 12, No. 3, 2006
                                      The Stock Return Difference between Industrial REITs and Manufacturing Firms

Hughes, W. T. Determinants of Demand for Industrial Prop-        Okunev, J. and P. Wilson. Using Nonlinear Tests to Examine
erty. The Appraisal Journal, 1994, 62, 303–09.                   Integration between Real Estate and Stock Markets. Real Es-
Li, Y. and K. Wang. The Predictability of REIT Returns and       tate Economics, 1997, 25, 487–503.
Market Segmentation. Journal of Real Estate Research, 1995,      Wheaton, W. C. and R. G. Torto. An Investment Model of the
10, 471–82.                                                      Demand and Supply for Industrial Real Estate. Journal of the
Liu, C. H., D. Hartzell, D. W. Greig, and T. V. Grissom. The     American Real Estate and Urban Economics Association, 1990,
Integration of the Real Estate Market and the Stock Market:      18, 530–47.
Some Preliminary Evidence. Journal of Real Estate Finance
and Economics, 1990, 3, 261–82.
                                                                 The authors are grateful to the editors and two anony-
Liu, C. H. and J. Mei. The Predictability of Returns on Equity
                                                                 mous referees. Opinions and results presented in this pa-
REITs and Their Co-Movements with Other Assets. Journal of
Real Estate Finance and Economics, 1992, 5, 401–18.              per are those of the authors and are not intended to rep-
Lockwood, L. and C. A. Rutherford. Determinants of Industrial    resent the views, policies, or interests of Fannie Mae. This
Property Value. Real Estate Economics, 1996, 24, 257–72.         paper is not the result of Fannie Mae related research
Mueller, G. R. and K. Pauley. Interest Rate Movement’s Effects   and does not use or cite Fannie Mae data sources.
on Real Estate Investment Trusts. Journal of Real Estate Re-
search, 1995, 10, 319–26.

                                                                              Journal of Real Estate Portfolio Management   259

Shared By: