NO.PZ2016071602000008
问题如下:
A database of hedge fund returns is constructed as follows. The first year of the database is 1994. All funds existing as of the end of 1994 that a willing to report their verified returns for that year are included in that year. The database is extended by asking the funds for verified returns before 1994. Subsequently, funds are added as they are willing to report verified returns to the database. If a fund stops reporting returns, its returns are deleted from the database, but the database has an agreement with funds that they will keep reporting verified returns even if they stop being open to new investors. Which of the following four statements are correct?
I. The database suffers from backfilling bias.
II. The database suffers from survivorship bias.
III. The database suffers from an errors-in-variables bias.
IV. The equally weighted annual return average of fund returns will under-estimate the performance one would expect from a hedge fund.
选项:
A.All the above statements are correct.
B.Statements I and II are correct.
C.Statements I, II, and III are correct.
D.Statements II and IV are correct.
解释:
B is correct. The database includes histories before 1994 and therefore suffers from backfill bias. Next, funds that stop reporting are deleted from the database, so it has survivorship bias. Errors-in-variables bias arises in other contexts, such as regression. Finally, the average of fund returns will be too high (not too low) because of these two biases. Hence, I. and II. are correct.
关于errors-in-variables bias:
1.我看了之后,觉得因为题目中回填等操作使得获取的report return变量其实本身就有偏差(会被高估)了,为什么不属于errors-in-variables bias?
2.所以这个bias主要指在操作上出现的问题,例如漏了变量、模型运算等操作风险事件上吗,。而不是说变量input选的数据集不好?