使用数据帧中堆叠格式的协方差矩阵计算 R 中的投资组合方差

Question

我的 cov 矩阵和权重位于 2 个独立的数据框中，我想计算投资组合方差（对于每个日期的每个投资组合，为了简单起见，我的示例只有 1 个投资组合和 1 个日期）：

weights <- data.frame(Portfolio = c("P1", "P1", "P1"),
    Date = c("2008-03-31", "2008-03-31", "2008-03-31"),
    ID = c("Asset1", "Asset2", "Asset3"),
    Wgt = c(0.1, 0.2, 0.3))

covar <- data.frame(
    Date = c("2008-03-31", "2008-03-31", "2008-03-31", "2008-03-31", "2008-03-31", "2008-03-31"),
    ID1 = c("Asset1", "Asset2", "Asset3", "Asset2", "Asset3", "Asset3"),
    ID2 = c("Asset1", "Asset1", "Asset1", "Asset2", "Asset2", "Asset3"),
    cov = c(0.011, 0.012, 0.013, 0.022, 0.032, 0.033)
)

Answer 1

如果我正确理解您的问题，您需要按

Portfolio

和

Date

进行分组，并创建一个新列来计算

Wgt

列的方差。

library(dplyr)

P_Variance <- weights %>% 
  group_by(Portfolio, Date) %>% 
  mutate(Var = var(Wgt))

> P_Variance
# A tibble: 3 × 5
# Groups:   Portfolio, Date [1]
  Portfolio Date       ID       Wgt   Var
  <chr>     <chr>      <chr>  <dbl> <dbl>
1 P1        2008-03-31 Asset1   0.1  0.01
2 P1        2008-03-31 Asset2   0.2  0.01
3 P1        2008-03-31 Asset3   0.3  0.01

我希望这有帮助！