What do bonds and cash do for a portfolio?
If equity is the growth engine, bonds and cash are usually described as the ballast. That word does real work, but it hides three separate jobs that are worth naming individually — because a portfolio can want one of them and accidentally buy another.
Three distinct jobs
- Income. A bond is a contract: a schedule of coupon payments and a return of principal on a stated date. Unlike a dividend, it is not at the issuer's discretion.
- Ballast. Because high-quality government bonds are what frightened money often buys, they have frequently risen when equities fell. That negative co-movement — when it holds — is what softens the portfolio's worst quarters.
- Liquidity and optionality. Cash is the sleeve you can spend without selling anything at a bad price. Every forced sale during a drawdown is a loss converted from paper into permanent, and cash is what prevents it.
The two risks inside a bond
Bonds are not "the safe one". They carry two specific, nameable risks:
Interest-rate risk (duration). Bond prices move opposite to yields, and duration measures how much. The rule of thumb: price change ≈ −duration × change in yield. A bond fund with a duration of 7 years, hit by a 1 percentage-point rise in yields, loses roughly 7% of its price. A 20-year Treasury fund with a duration near 17 loses roughly 17% on the same move. In 2022 that mechanism produced a fall of roughly 30% in long-dated Treasuries — a larger drawdown than the equity index suffered in the same year.
Credit risk. The chance the borrower does not pay. It is near-zero for a major developed-market government borrowing in its own currency, meaningful for a mid-grade corporate, and substantial for high-yield issuers. Credit risk is also correlated with equity risk, which is why a bond sleeve stuffed with high-yield paper provides much less ballast than its label suggests.
A useful summary: duration is the ballast risk, credit is the equity risk wearing a bond's clothes.
What cash actually costs
Cash's nominal safety is real. Its purchasing power is not.
Take an illustrative case: cash yields 3% while inflation runs 4%. The real return is roughly 3% − 4% = −1% per year. Compounded over 10 years, purchasing power falls to about (1.03 ÷ 1.04)¹⁰ ≈ 0.908 — a loss of roughly 9% of what the money can buy, with the account statement showing a gain the whole time.
This is the trade. Cash converts variability risk into erosion risk. Neither is free; they simply show up in different places, and only one of them is visible on a statement.
The illustration that professionals keep coming back to
Two portfolios, both €100,000, both needing to fund a €20,000 withdrawal in a year when equities fall 30%.
- Portfolio A holds 100% equities. It must sell €20,000 of shares at the bottom. Those shares never participate in the recovery.
- Portfolio B holds €20,000 in cash and €80,000 in equities. The withdrawal comes from cash; the equity sleeve is left alone to recover.
Portfolio B gave up whatever the cash sleeve failed to earn in every ordinary year, and got paid for it once, in the bad year. That is the exact shape of the bargain — and whether it is worth making depends entirely on whether a withdrawal is actually coming, which is the subject of the next unit.
Return order is irrelevant to a portfolio with no flows and decisive for one paying out. Two years, +30% and −20%, a €20,000 withdrawal at each year-end, from €100,000. Good year first: 130,000 less 20,000, then 88,000 less 20,000, ends at €68,000. Bad year first: 80,000 less 20,000, then 78,000 less 20,000, ends at €58,000. Same returns, same withdrawals, a €10,000 gap from order alone; with no withdrawals both orders end at €104,000. A pension in payment, an endowment on a spending rule, a retiree: all are exposed to when the bad year lands, and an average return hides it. Hence a cash or short-bond sleeve sized to a few years of withdrawals is common for money being drawn.
In the data
The cash leg is quoted as a rate, and a Treasury bill carries two of them. Here is the latest day for the 13-week and the 52-week bill:
The discount rate is the older convention: it measures the discount against the face value, so it always sits a little below the yield. The coupon-equivalent figure is the yield, and it is the one to compare with a return. Both are what a bill bought that day would pay to maturity, not what a holding earned over any past period.
Try it now
- First the move itself. The chart below is the US 10-year government yield, and it is a yield rather than a price: a value of 4.674 means 4.674%, and nothing trades in the series itself. Read the size of the rise across 2022 in percentage points.
- Now what the move did to two pure government holdings, five years each: T-bills, then Treasuries of twenty years and longer. Credit risk is near zero in both, so it cannot be the explanation. One barely registered that rise and the other fell about a third. Multiply your figure from step 1 by a duration near 0.1 and then by one near 17, and check the rule of thumb against what the charts actually show.
- Now hold duration roughly still and change the credit instead. Across the sharpest equity drawdown visible on either chart below, which behaved like ballast and which behaved like equity risk wearing a bond's clothes?
- Estimate a real return yourself: Measure the last twelve months on the T-bill chart as a stand-in for a cash yield, and subtract the year-on-year US inflation figure below. Describe the result neutrally — it is an observation about arithmetic, not a verdict on holding cash.