| Summary: | For investors, there are two important terms that they should understand
before deciding to invest in stock market, which are risk and return. Those two
terms is often traded off each other hence investors cannot hide from the risk to
get high return. Related to risk, since unsystematic risk, can be diversified by
forming the portfolio, investors will concern more about systematic risk or beta.
There are several ways to estimate future beta. In general, they can be
classified into historical and fundamental method, or may be the combination of
both. The historical beta methods incorporate current beta values to predict the
future beta one year later. Fundamental beta methods use several fundamental
variables to estimate the future beta. The fundamental variables can came from
same timeline with the stock beta or the data at time before. This research
investigates variables that simultaneously have significant effect to stock beta and
the best method to be applied in forecasting the future beta, for non-financial
stocks listed in Indonesia Stock Exchange. Hence, the comparison of predictive
ability for every model will be tested here. For fundamental method, the
independent variables are asset growth, financial leverage and firm size.
This research used secondary data for stock price, total asset, and debt to
asset ratio from year 2005 until 2010. The data analysis will be divided into two
observation periods, which are 2005-2007 and 2008-2010 in order to observe the
result consistency. The total sample is 50 companies, based on criteria the nonfinancial
companies that listed in Indonesia Stock Exchange since 2005 and
actively being traded.
The regression equations, both in simple and multiple forms will be used
to determine the model for predicting the future beta. Simultaneously, the
variables that consistently has significant effect to current stock beta are past beta,
or combination of past asset growth, past financial leverage and past firm size, or
could be the combination of all. To analyze the predictive ability among all
methods, the square error between estimate and actual value will be calculated and
compared. As the result, the historical method found better in predicting future
beta compares to fundamental method. When it being compared to combinational
method, at one period historical method is recognized as superior method but in
other period the combinational method is the best one.
|
|---|