First Empirical Results

  • B. Philipp Kellerhals
Part of the Lecture Notes in Economics and Mathematical Systems book series (LNE, volume 506)


For the empirical implementation of our pricing model we choose to select a broad market sample of emerging market closed-end funds.45The sample examined in this study exists of closed-end equity funds for a complete five year period from January 1993 to December 1997. All NYSE traded closed-end funds are included, if their date of issue lies before January 1993 in order to avoid post-offering pricing effects.46The data is collected on a weekly basis to yield a multiple time-series of 33 closed-end funds with 260 observations each. The market prices are obtained from the last trade on Fridays, and the net asset values are calculated by the funds on Thursdays balances. However, the net asset values are reported on Fridays after the NYSE closes which makes trading on the basis of these prices impossible, but are treated as being observed contemporaneously with the market prices.47Finally, the share prices and net asset values are adjusted for distributions including income dividends and capital gains. In table 8.1 we report the figures for the total returns and their correlation along with the net proceeds and the date of the initial public offering (IPO) of the funds.


Market Price Initial Public Offering Valuation Model State Space Form Empirical Implementation 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 45.
    I greatly thank Lipper Analytical Services, and especially Donald Cassidy for kindly providing the closed-end fund data.Google Scholar
  2. 46.
    See, for example, Peavy (1990) and Weiss (1989).Google Scholar
  3. 47.
    Thus, the empirical premia will include some undetermined amount of measurement error.Google Scholar
  4. 48.
    This excess price volatility compared to the fundamental volatility is consistent with the implications of the De Long, Shleifer, Summers, and Waldmann (1990) noise trader model.Google Scholar
  5. 49.
    See, for example, Dimson and Minio-Kozerski (1998).Google Scholar
  6. 50.
    There are very rare exceptions which show this feature and no closed-end funds included in our sample.Google Scholar
  7. 51.
    For more details on state space modeling compare chapter 3 and see, for example, Aoki (1990).Google Scholar
  8. 52.
    See, for example, Bergstrom (1984).Google Scholar
  9. 53.
    Using Itô’s isometry and the fact that the expectation of an Itô integral is zero. See, for example, Oksendal (1995, ch. III).Google Scholar
  10. 54.
    The function δ i, j stands for the Kronecker symbol which takes value 1 if i = j, and 0 else.Google Scholar
  11. 55.
    The small errors ε t are introduced to capture imperfections in observed price quotations.Google Scholar
  12. 56.
    Further, we assume independent error terms ε t and η t.Google Scholar
  13. 57.
    The half-life of the dynamic premium is given by ln(2)/κ. This can be deferred from equation (8.6) solving for τ = t - s. Google Scholar
  14. 58.
    See, for example, Aitchison and Brown (1957, ch. 2.2) for the log-normal transformation.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2001

Authors and Affiliations

  • B. Philipp Kellerhals
    • 1
  1. 1.Gesellschaft für Wertpapieranlagen mbHDeutscher Investment-TrustFrankfurt am MainGermany

Personalised recommendations