Changing particular cell values of multi-level columns in a for loop

I have multi-level column data like this:

   price           flag              shares         balance      finalbalance
  A      B                          A      B       A     B 
  100   100          0             .4     .6      40     60          100
  80    150          0             .4     .6      32     90          122
  90    120          1             .48    .54     43.2   64.8        108

So my problem is that of re-balance.

• Price column shows prices of stocks "A" and "B"
• flag column is either 0 or 1. When it's 0, we don't do anything. When it's 1 we rebalance the portfolio in a 40%:60% ratio. For example, at the start of the portfolio, $100 was divided into $40 and $60, and then the prices of the security went up/down. When the flag became = 1, we wanted to rebalance our total balance into 40%:60%. Since the total balance was $108 when the flag became 1, we divide the money into these stocks in the 40%:60% ratio, which was 43.2 and 64.8 depending on the price of that day for A and B.

How do I achieve this in python? Since its multi-level columns, I am not able to use ".loc" and ".at" commands.

This is what I was able to code:

tickers=['A','B']
ticker_weights = {
    "A": 0.40,
    "B": 0.60 }

for i in range(0,len(stock_data)):
    if(stock_data.flag.iloc[i]==1):
        for ticker in tickers:
            stock_data['shares'][ticker].loc[i] = ticker_weights[ticker]*stock_data.finalbalance.iloc[i]