Capital weight is not risk weight

A friend showed me his portfolio last week, proud of how balanced it was. Sixty percent in stocks, forty in bonds, a proper cushion for the rough years. Then I asked him how much of his actual risk was sitting in those bonds. He guessed forty percent, same as the money. The honest answer is about five.

Here is the part the pie chart hides. Capital and risk are not the same split. Stocks swing around 18 percent a year, bonds maybe 5. So every rupee in stocks carries roughly three times the weight of a rupee in bonds. Run the decomposition on a 60/40 portfolio and the risk does not come out 60/40. It comes out 95/5. Ninety five percent of the swing is the equity sleeve. The bonds are along for the ride.

It gets worse before it gets better. Move to a 50/50 split and you are still at about 91 percent equity risk. To genuinely split the risk down the middle you would have to hold roughly 22 percent in stocks and 78 in bonds, and the whole portfolio would swing about 6 percent a year instead of 11. That is the allocation that is actually balanced, and in fifteen years I have met almost nobody who holds it.

None of this means 60/40 is wrong. It means the label is. A 60/40 is a growth portfolio with a small shock absorber, not a balanced one, and the day equities fall 30 percent your balanced fund will tell you exactly where its risk lived.

Three things I keep coming back to. Capital weight is not risk weight. Risk pools in the loudest sleeve. And balanced means balanced in risk, not in rupees.

The pie chart shows the money. The drawdown shows the truth. Read the contribution to risk before you call anything balanced.

How risk weight differs from capital weight:

Finding the split that is balanced in risk

For two assets there is a short rule for an equal risk split: hold each in proportion to the other's volatility. The equity weight is the debt volatility divided by the sum of the two volatilities. With 18 and 5, that is 5 divided by 23, or 21.7 percent. That is where the opening's 22/78 comes from.

Use the volatilities from NSE data instead. On quarter end closes from 14 Oct 2024 to 30 Sep 2026, the Nifty 50 swung about 16.5 percent a year and the Nifty Composite G-sec Index about 4.4 percent. The equal risk equity weight comes out at about 21 percent. Different inputs, nearly the same answer, because equity stayed close to four times as volatile as government bonds.

The two mixes on the same two years

Now test both splits on the same NSE closes, with the weights reset at each quarter end. One is the classic 60/40. The other is the risk balanced 21/79.

Do not read the return line as a lesson. Those two years were poor for Indian equity and decent for government bonds, so the low equity mix won. Over long stretches equity has earned more, and a 21/79 mix gives most of that up. The worst quarter line is the lesson. Balancing the risk halved the deepest fall, because the bonds were finally large enough to matter.

That is the honest trade. A risk balanced mix is calmer and earns less in strong equity years. A 60/40 earns more and behaves mostly like equity. Both are fine choices. Only one of them is balanced.

Reading your own pie chart

Write down your split, find a rough volatility for each sleeve, and multiply. On a 60/40 with 16.5 and 4.4, equity carries 60 x 16.5 = 990 units against debt's 40 x 4.4 = 176. Before correlation, that is already 85 percent equity on the simple product, and the correlation term pushes it higher. The free Learn module on modern portfolio theory works through risk contribution and the variance formula, with a quiz and a certificate. The Sharpe and volatility calculator handles the arithmetic. The companion post, 60/40 is a 94/6 bet, splits a real quarter's loss by sleeve.