Why Gilt Funds Are Not Risk-Free

Lakshmi bought the safest thing available

Lakshmi moved part of her retirement corpus into a gilt fund because somebody explained that it lends only to the government, and a government does not default on money it borrows in its own currency. That is correct, and it is a genuine advantage rather than a marketing claim.

The fund had also produced excellent returns over the preceding few years, which is why it appeared on the list she was shown. Both facts were true and neither of them described what she was actually buying.

Lending money carries two main dangers: the borrower may not repay, and interest rates may move against you while you wait. A gilt fund removes the first entirely. It leaves the second untouched — and government bonds are frequently long-dated, which makes them among the most rate-sensitive instruments available.

So the honest description is that a gilt fund removes credit risk and concentrates interest-rate risk. Calling it risk-free swaps one for the other and produces exactly the wrong expectation.

Where the volatility comes from

The mechanism is the ordinary bond mechanism, amplified.

When prevailing rates rise, an existing bond paying a fixed coupon becomes less attractive, and its price falls until a buyer would earn the going rate. How far it falls depends on how long the buyer is committed to that below-market coupon, which is duration.

Gilt funds frequently hold long-dated government paper, so their duration is often high — a fund with a duration of seven or eight moves roughly seven or eight per cent for each percentage point of rate change. That is equity-like movement in a product many people hold precisely to avoid equity-like movement.

And it works in both directions, which is how the misunderstanding takes root. During a period of falling rates, gilt funds produce excellent returns, attract money from people looking at those returns, and then behave very differently when the rate cycle turns.

The pattern that catches people

It is worth stating explicitly, because it repeats and because Lakshmi is standing in the middle of it.

Rates fall over a period. Gilt funds post strong returns. Those returns appear in league tables and recommendations, and investors — often conservative ones, attracted by the word "government" — move money in, frequently money that was meant to be the safe part of their portfolio.

Then rates stop falling, or rise. The funds decline, sometimes sharply and for an extended period. The investor, who believed they had bought safety, discovers they had bought a bet on interest rates.

Nobody misled anybody. The fund did exactly what it says it does. The mismatch was between the word "government" and the assumption of stability.

What it has actually looked like

Until now this page could describe the mechanism and not show it. It can now show it.

Below are two government bond indices. Both carry no default risk of any kind — the borrower is the government in both cases. The only material difference between them is how long the money is committed for.

Over the same period	Lending for ten years	Lending for five years
Return a year	+6.45%	+8.02%
Worst fall from a high point	-11.73%	-7.16%
Time from that high to regaining it	17 months	8 months
Share of days below an earlier high	82.7%	69.1%
Worst twelve months	-5.10%	-0.56%
Share of twelve-month periods that lost money	14.0%	0.1%

Read the second row first. An instrument with no credit risk whatever fell 11.73% from a high point, between 27 May 2013 and 19 August 2013. Counting from that high, it took 17 months to be level again — and over the whole period it lost money across 14.0% of all twelve-month stretches.

Nobody defaulted. Nothing went wrong. That is simply what interest-rate risk looks like.

Now compare the columns. The shorter commitment fell 7.16% where the longer one fell 11.73%, was level again in 8 months against 17 months, and lost money in 0.1% of twelve-month stretches against 14.0%. Duration is the whole of the difference, and it is the number Lakshmi was never shown.

There is a further result in that table which deserves care, because it is easy to over-read. Over this period the shorter commitment also returned more — +8.02% a year against +6.45% — while falling less. That is a fact about 15.7 particular years, in which rates broadly rose, and not a law: over a stretch of falling rates the longer commitment would have won and would have been the more volatile all the same. What does generalise is the risk side, because that follows from the arithmetic of duration rather than from which way rates happened to move.

The risks that remain

Interest-rate risk is the main one, as above. Reinvestment risk follows from it: if rates fall, the income the fund receives gets reinvested at lower rates, so future returns decline even while current values rise.

Inflation risk is quieter and can matter more over long holdings, since a fixed coupon buys less over time and that erosion never appears as a loss on a statement.

Liquidity in stress is worth naming even though government securities are generally the most liquid rupee bonds available — that is a genuine strength, and liquidity is still a matter of degree, with even government bond markets trading at wide spreads under severe conditions.

And there is a risk specific to how these funds are run. Many gilt funds actively vary their duration according to a view on rates, which means Lakshmi is exposed both to rate moves and to whether the manager positioned for them correctly. That is a second risk dressed as risk management, and it is covered in its own page.

What a gilt fund is genuinely good for

This is not an argument against them, because used for the right job they are excellent and the reasons are specific.

They provide credit-risk-free exposure to interest rates, which is exactly what you want if you have a long-horizon liability and wish to be certain the borrower will pay. They are the cleanest ballast in a portfolio, since government bonds are the asset most likely to hold up or gain when equities fall in a growth scare — unlike corporate credit, which tends to fall alongside equities precisely when you need it not to. And they behave predictably given a rate move, which makes them easy to reason about once duration is understood.

What they are not is a substitute for cash, a home for an emergency fund, or a place for money needed within a couple of years. Money with a short horizon in a long-duration fund is a horizon mismatch, and the loss arrives at the moment you are forced to sell.

Using them sensibly

Check the duration before anything else, since it is the single number determining behaviour and it is on the factsheet. Translate it into rupees on the amount you plan to invest, at a one-point rate move, before investing rather than after.

Match it to your horizon: if you can leave the money for longer than the duration, a rate rise is a temporary mark-down you can wait out, and if you cannot, it is a realised loss. Do not judge by recent returns, because in this category more than any other, past returns describe the rate cycle that has just occurred rather than the fund.

Consider a shorter-duration government option if you want the credit safety without the volatility, since government exposure and long duration are separable choices and many people who wanted the first have accidentally bought the second. And watch costs, because in a fund whose entire return may be a few per cent the expense ratio takes a significant share of it.

What this still cannot tell you

These are indices, not funds. A fund holding government paper charges a fee, holds some cash, and in many cases varies its duration according to somebody's view on rates. Its outcome will differ from the table above, and the last of those adds a risk the table does not contain at all: whether the manager positioned correctly. That is dynamic bond fund risk.

Two benchmarks are not the whole curve. The table shows five years against ten. A fund holding longer paper would have moved more than either column, and this record has no thirty-year index in it to show you.

The longer series covers one and a half decades, containing one large rate shock rather than several full cycles. It establishes that a fall of this size happens. It cannot tell you how often.

And the figures are before inflation. A period in which a bond returns a little and prices rise by more is a loss in what the money buys, which never appears as a negative number on a statement — nominal return, real return and purchasing power.

What to take away

A gilt fund removes default risk and keeps interest-rate risk, often in a concentrated form because government paper tends to be long-dated. "Government backed" describes who repays you, not how steady the value will be along the way.

Check the duration, translate it into money, and match it to your horizon. Use gilt funds as long-term ballast, where their behaviour during an equity fall is genuinely valuable — and not as a place to keep money you might need soon.