在R中为投资组合分配权重

Question

I am trying to show in R how I can assign weights to the portfolio and then calculate the daily return. Below is the data for daily return and I have assigned 25% weight to each of them for making an equal weight portfolio and to calculate the returns. (There are 4 company stocks out of which i have just provided data for 2 stocks, hence 25% weight to each stock has been given) However, I am unbale to configure the same in R.

Workings: 

25% = 636.99*25% = 159.25
      633.79*25% = 158.45

Returns = Closing - opening 
        = 159.25-158.45
        = -0.80

Is there any way I can show this in R?


    AMZN     25%    Returns  GOOG     25%   Returns
    636.99  159.25          741.84  185.46  
    633.79  158.45  -0.80   742.58  185.65   0.18
    632.65  158.16  -0.28   743.62  185.90   0.26
    607.94  151.99  -6.18   726.39  181.60  -4.31
    607.05  151.76  -0.22   714.47  178.62  -2.98
    617.74  154.43   2.67   716.03  179.01   0.39
    617.89  154.47   0.04   726.07  181.52   2.51
    581.81  145.45  -9.02   700.56  175.14  -6.38

Answer 1

Returns should be divided, not subtracted. Your 1-day return for Amazon should be calculated as 633.79 / 636.69 - 1 = -0.0046 . An equal-weighted portfolio could just take the mean of these four returns on day 2.

Beyond day 2, you need to think about how often your portfolio is rebalanced. This is more a topic for quant.stackexchange.com, and not a programming question. But I'll give you both the (unrealistic) case where you assume the portfolio is perfectly equal-weighted at the start of each day, and the (simple & realistic) case where we set the portfolio to be equal weighted at the start and we forget about it.

As I'd mentioned in your question about stock price returns , a good R package for collecting and manipulating stock prices is quantmod . Try:

library(quantmod)
symbols <- c("GOOG", "AMZN", "FB", "AAPL")
getSymbols(symbols, src = 'google')
closing.prices <- merge.xts(GOOG[,4], AMZN[,4], FB[,4], AAPL[,4])["2016-12-30/"]

Note that everything will be done using xts time series. The time series closing.prices will be prices since the end of 2016, and can be converted to a series of returns using ROC from the TTR package:

library(TTR)
price.returns = ROC(closing.prices)

For our portfolio which is 25% equal-weighted at the start of every day, the portfolio returns will be the means:

rowMeans(price.returns)

The more realistic case where a portfolio was 25% equal-weighted at the start of the year, you might compute the overall returns of each stock, then compute the mean of the four:

mean(as.numeric(closing.prices["2017-09-12"]) / as.numeric(closing.prices["2016-12-30"]) - 1)

where closing.prices["2017-09-12"] is the closing price today, closing.prices["2016-12-30"] was the closing price at the start of the year.