2004). Recognizing the limitations of manual coding (e.g. small sample sizes, subjectivity),
whether the MD&A is truly informative remains an open empirical question.
With the rise of behavioral finance in the last decade, this discussion was resumed and
textual analysis has garnered increased attention, with the aim of assessing the
informativeness of corporate disclosures. Thereby, researchers most frequently relied on
textual tone analysis to examine firms’prospectuses (Feldman et al., 2010;Jegadeesh and Wu,
2013;Li, 2010;Loughran and Mcdonald, 2011,2015). Furthermore, most studies reviewed a
random sample of firms, restricted only by the availability of necessary data. However,
Callahan and Smith (2004) find evidence that the impact of language varies across industries.
Thispaper adds a new dimensionto the discussion by analyzingMD&As for a sampleof US
Real Estate Investment Trusts (REITs). In contrast to a sample randomly drawn from the
capital market, the US REIT market provides a number of beneficial characteristics. First,
equityREITs are fairly homogeneousregarding characteristics thatusually vary widely across
differentindustries (Hartzell et al.,2008).Second, US REITs are requiredto pay out a minimum
of 90% of taxable earnings to shareholders as dividends. Consequently, in order to take
advantage of growth opportunities, US REITs must turn to the capital markets. As such,US
REIT managers have an unusually strong incentive to be transparent and maintain investor
trust (Danielsen et al., 2009;Doran et al.,2012;Price et al., 2017).Third, the underlying assetsof
US REITs are realestate, which is an illiquid, slow-moving asset and thusmore compatible to
analysis over a relatively large time span(e.g. from one-quarter to the subsequentquarter).
These unique characteristics suggest that the MD&As of US REITs are particularly
informative. However, the asset class’s peculiarities also indicate that results from previous
studies cannot automatically be extended to US REITs. Thus, we investigate the information
content of the MD&A for a US REIT sample by answering the following questions: Does
textual sentiment in the MD&A reveal managers’expectations regarding future firm
performance? If so, does the market process the information efficiently?
To extract sentiment from the MD&A, we rely on a dictionary-based approach.
Specifically, we employ the Loughran and Mcdonald (2011) financial dictionary and a custom
wordlist for the real estate industry created by Ruscheinsky et al. (2018) to determine the
overall sentiment inherent in each filing. Our findings suggest that higher levels of
pessimistic (optimistic) language in the MD&A are associated with lower (higher) future firm
performance. This holds even after controlling for the information released in other
concurrent disclosures that may be predictive of future performance. Hereby, the use of a
domain-specific real estate dictionary, namely the dictionary developed by Ruscheinsky et al.
(2018) leads to superior results. Moreover, we find a significant market response to
pessimistic language in the MD&A at the time of the SEC filing. However, corresponding to
the notion that individuals are affected more strongly by negative than positive news, we
cannot find a significant impact of optimistic language. Overall, to the best of our knowledge,
this is the first study providing evidence that the use of language in the MD&A reveals US
REIT managers’expectations regarding future firm performance and that the market
responds to this information. We demonstrate that the market can benefit from textual
analysis, as investigating the language in the MD&A should decrease information
asymmetries between US REIT managers and investors.
The remainder of the paper is organized as follows. Section 2 discusses related literature.
Section 3 introduces the data, that is, sample and variables. Section 4 defines the specific
sentiment measures and presents empirical methods for the analysis. Finally, Section 5
reports the empirical results, and Section 6 concludes.
2. Related literature and hypothesis development
Textual analysis has recently attracted increased attention to address many pivotal
questions in behavioral finance. Not least because in the current world, a huge amount of