# Composite portfolio performance

I'm trying to calculate the combined performance of two portfolios, but keep getting a nonsensical result. I have a Fixed Income & Equity portfolio and the GIPS quarterly performance number from each manager.

I want to get the return for the combined portfolio. (For this question, I use 32% Eq & 68% FI - these vary by quarter, but not by much)

Date          EQ          FI          Total
9/30/18       7.51%       0.440%      2.70%
12/31/18    -15.70%       1.735%     -3.85%
3/31/19      12.99%       2.123%      5.60%
6/30/19       3.17%       2.466%      2.69%

1 year        5.65%       6.915%      7.08%


I calculate the 1yr performance by {Product(1+B2:B5)-1} - and this matches what the manager reports.

I get the quarterly Total performance by EQ*0.32+FI*0.68. However, the chained Total 1yr performance is higher than the two portfolios.

What am I doing wrong?

Thanks, R.

It depends what you assume as to rebalancing between the portfolios.

Unless these portfolios are being actively rebalanced each quarter to bring them back to exactly 32/68 allocation you should not assume fixed weights of 32 and 68. It is misleading.

Here is an alternative calculation assuming no rebalancing: You start on 6/30/2018 with 0.32 million in equities and 0.68 in fixed income and you allow the weights to drift with the performance of each fund, with no transfers between one fund and the other.

Date            EQ    EQ AUM      FI    FI AUM     EQ+FI AUM
----------   -----  --------    ----  --------     ---------
06/30/2018          0.32              0.68         1.0
09/30/2018   7.51%  0.344032  0.440%  0.682992     1.027024
12/31/2018 -15.70%  0.290019  1.735%  0.694842     0.984861
03/31/2019  12.99%  0.327692  2.123%  0.709593     1.037286
06/30/2019   3.17%  0.338080  2.466%  0.727092     1.065172


As shown in Column F, the return on the portfolio was 6.5172%, which is between the returns on the EQ (5.65%) and FI (6.925%) sub-portfolios. This is the "buy and hold" (non rebalanced) performance.