英语论文网

of a
dummy variable to take into account the effect of the German reunification was common, but not
always significant.  We therefore used only the three factors that were assumed to be the most
important (stock market, interest rate, and real estate market) and the equation mentioned above.
Indices
Since there is no organized market for land, real estate indices are constructed on the basis
of other data that reflect the price changes of the market.  The three basic types are: 1) appraisal-
based indices, 2) transaction-based indices, 3) indices based on share prices of institutional real
estate investors.  There are several real estate indices in Germany; however, reliable monthly data is
only available from two indices of real estate companies ("property stocks").  Theoretically, if such
an index is stripped of all other influences--which is possible by employing some statistics--it is a
pure real estate index.  We used the property stock index of Westdeutsche Landesbank (WestLB)
because at that time the other index of Ellwanger & Geiger contained too many illiquid stocks.
                                                

10

  Lausberg (1997).Many studies have found a near-perfect correlation of the various stock market indices.
Our calculations confirmed this for the most widely used indices.  For our analysis we chose the
CDAX because it is a price index (like the WestLB property stock index), it has a bank sub-index,
and a broad basis.
In previous studies, a great variety of interest rate factors were successfully tested,
indicating that the negative relation between bank stock returns and interest rates is stable and

independent from the index used.

11

 This finding allowed us to use actual interest rates although they
do not meet the APT criterium of unpredictableness--primarily because we were not mainly
interested the interest rate factor.  We used the "REX-10", a 10-year bond-price-index of synthetic
government bonds because it was the only index of long-term interest rates which was available
over a longer period.
Data and Sources
The study covers the period from January 1990 thru April 1996, the longest availabilty of
the indices at that time.  To account for a variation of the sensitivities over time, we formed 2 sub-
samples; more would not have been meaningful because the statistical methods required a certain
minimum number of observations.  Mid-1992 was used for partitioning since this seemed to be the
turning point of both the long-term economic trend and the real estate market trend, as several
indices showed.
In April 1996, there were 34 bank companies listed on German stock exchanges.  For 30 of
them, information about real estate loans was available from the Hoppenstedt-database.  The banks
were divided into 3 equally weighted portfolios, depending on the percentage of real estate loans in
their balance sheet.  The separation points (10 and 55%) were chosen in a本论文由英语论文网提供整理，提供论文代写，英语论文代写，代写论文，代写英语论文，代写留学生论文，代写英文论文，留学生论文代写相关核心关键词搜索。