Join our community of smart investors
OPINION

The bonds that bind

The bonds that bind
February 21, 2013
The bonds that bind

Back in 1995, had I reckoned in those terms when I used a chunk of capital to buy the maximum amount of Premium Bonds that the government allows, I might have thought otherwise. As it was, there was some logic behind buying the maximum number of bonds (then £20,000 and raised to £30,000 in 2003). Buy a few hundred Premium Bonds and that's much the same as having a flutter on the National Lottery. True, you get your principal back (unlike the lottery), but the chances of winning a prize are small and the odds against winning a big payout are lottery-esque.

However, buy the maximum number of Premium Bonds and the game changes. From offering the sort of cheap thrills that come with lottery scratch cards, bonds become more like a high-return savings account or a gilt-edged stock, with a kicker thrown in.

First, they offer regular income. They must because, for anyone holding the maximum number of bonds, the chances of winning a prize each month are better than even. Basically, he or she has 30,000 chances of winning a prize on a 24,000-to-one shot. Of course almost all of these 'shots' generate just £25. Each month, 1.81m prizes are paid out by 'Ernie' - the anthropomorphic name for the computer that randomly generates winning bond numbers - and 96 out of every 100 are for £25. Still, on average that should mean 14 prizes of £25 a year, or £350. That's a 1.2 per cent return before tax, but you don't pay tax on Premium Bond winnings. So, for a 40 per cent taxpayer - and the arguments in favour of owning the maximum number of bonds only really stack up for higher-rate tax payers - the gross equivalent becomes 1.9 per cent. In these recession-ravaged times, that's almost competitive with 90-day notice savings accounts and it beats short-dated gilt-edged stocks. To get a better return from gilts, a saver buying now would have to go out to a nine-year maturity, where Treasury 4 per cent 2022 offers a 2.1 per cent return to redemption.

 

 

Second, there is that kicker - the chances of landing a prize that can transform the economics of owning Premium Bonds. This is not lost on folk in the City, where there are many informal syndicates clubbing together to share prizes. Forming a syndicate achieves two aims. First, the bigger the syndicate, the closer it will get to generating the same rate of return as that offered on the prize fund (1.5 per cent since October 2009). For City folk paying top-rate tax, this is especially valuable - the equivalent of a 3 per cent annual return. Second, the syndicate improves the chances of winning a big prize. If five people club together to pool £150,000, their chances rise five-fold. Sure, in practical terms the odds of winning a 'higher value' prize (£5,000 or over) remain prohibitively long. Each month there's still a 3,000-to-one bet against winning a £5,000 prize. Should they get lucky, then fairness dictates that the prize would only be £1,000 to each syndicate member. Even so, it's a tempting scenario, especially for people with more money than places to use it intelligently.

It was logic along these lines that persuaded Bearbull's alter ego to put £20,000 - and then another £10,000 - into the maximum number of bonds the government allows. That was 18 years ago. I bought £20,000-worth in February 1995 and, when the maximum holding was raised to £30,000 in 2003, another £10,000 in June of that year.

So, how have I done? How close has owning the maximum number of bonds been to holding a high-return savings account with a kicker, or should I have put my money into an alternative low-risk vehicle?

Being the anorak that I am, I have kept a full record of returns from my Premium Bonds and, after 18 years, should have enough data to make intelligent observations.

To jump straight to the conclusion, the judgement would be that Bearbull's Premium Bonds have done okay, but I should have put the capital into gilts back in 1995, say, Treasury 8 per cent 2013, which offered an income return of about 8 per cent a year back then.

Still, back in 1995 it was impossible to imagine the bull market in gilts that would develop. True, in the mid-1990s central banks were starting to target the control of inflation as their main mission. That implied that the rate of change in consumer prices would fall to a level that was acceptable to central banks and treasuries. But it would be years before anyone realised what instability could be caused putting the stability of consumer prices ahead of all else, and its effect on interest rates and gilts' prices.

 

What Ernie pays out each month

Prize valueNumber of prizesCost (£)% of payoutOdds on winning *Once every
£1 million11,000,00021,451,733 to 1121,000 years
£100,0005500,0001290,347 to 124,000 years
£50,0008400,0001181,467 to 115,000 years
£25,00018450,000180,652 to 16,700 years
£10,00046460,000131,559 to 12,600 years
£5,00091455,000115,953 to 11,300 years
£1,0001,0891,089,00021,333 to 1111 years
£5003,2671,633,5003444 to 137 years
£10031,9903,199,000645 to 14 years
£5031,9901,599,500345 to 14 years
£251,746,16943,654,225801.2 to 1month
Total1,814,67454,440,225100
Based on January 2013’s draw * assuming £30,000 maximum of bonds

 

So let's dig a little deeper. The first thing to examine is whether my Premium Bond holding has provided the regular income commensurate with a savings account. Without doubt, yes. In the 213 months for which I have been eligible to receive prizes, I have picked up 215. Sure, there has been variation. In 2010, I got 19 prizes, but only two in 2002 (both of which were in September). Indeed, from August 2001 to February 2003, I received only three prizes. Statistically, that was sufficiently unlikely to prompt a letter to National Savings in Blackpool, asking if warrants might have gone astray. Happily, nowadays National Savings pay prizes directly into a nominated account.

That regular flow of prizes means that my winnings now total £11,550. A good way to assess whether that's acceptable is to calculate the 'internal rate of return'. That's financial analysts' jargon for answering the question: what's the interest rate that makes all the cash flows from an investment - both money out and money in - have a present value of zero? That rate either forms, or is compared with, a 'benchmark' rate - the rate of return that should be exceeded for an investment proposition to be worthwhile.

 

 

For my Premium Bond investment, the internal rate of return is 2.7 per cent. However, it gets better. Remember that bond prize money is tax-free and I pay tax at the higher rate. Factor that into the calculation and the return rises to 4.7 per cent. Back in 1995, if someone had offered me a 4.5 per cent annual return on an investment that carried no risk except the inability to protect the principal against inflation, I'm pretty sure I would have accepted it. So the 4.7 per cent out-turn from the Premium Bonds is satisfactory.

Let's not dodge the fact that it has not been enough to protect against inflation. That much is shown in the main table, 'Premium Bonds versus inflation and gilts'. What should be quite a simple table becomes complex in practice. First, comparing my Premium Bond returns with inflation, as measured by the Retail Prices Index, is not comparing like with like. That's because the inflation rate is a 'compound' measure (one year's inflation is based on the price level at the end of the previous year), whereas returns from Premium Bonds are 'simple' (each year's return is based on the value of the original investment). That said, Premium Bonds offer something that inflation does not - the return of capital (albeit weakened by inflation's effects). Factor that into the calculation - in other words, we are back to internal rates of return - and the return on Premium Bonds (2.7 per cent) is almost as high as inflation (2.8 per cent).

 

 

The comparison with gilts is messier still. True, had I invested in gilts rather than Premium Bonds back in 1995, I could have participated in a remarkable bull market. But that's more theoretical than real. If I had opted for gilts, I would have chosen a specific stock - quite likely Treasury 8 per cent 2013 - and ran it to maturity (or when - as was the case - it was subsumed into another stock). That would have given me a fixed income return of about 4.3 per cent a year net of tax (income tax is payable on gilts' dividends). But it would have returned little in the way of capital gains because the market price and the redemption value (£100 paid back for every £100 'nominal') would converge as the stock approached maturity. Yet the table, which is drawn from a standard database of gilt returns, assumes that somehow I would have participated in that wonderful bull market. I wouldn't have - at least, not unless I had been doing some buying and selling and taking on those concomitant risks.

Still, there is no escaping the fact that a 4.3 per cent return from gilts, which is roughly comparable with the 2.7 per cent return from my Premium Bonds, skins the bonds.

Which leaves one question: should I stick with my Premium Bonds? Probability theory says 'no'. After all, just eight months after being eligible to win prizes, I picked up a £500 prize (only every 444 months on the balance of probabilities). And after almost four years, I won a £1,000 prize (on probability, once every 111 years). That's when I should have quit. Since then, it has been the occasional £100 prize (when they were paid out more often, anyway) and, for the past four years, a stream of £25 winners.

Even if I pick up another £1,000 prize next month, it will do little to my overall returns. So rationality says 'sell'. Yet then there is the dilemma of what to do with the £30,000 freed up. Sure, I can replace the income; I can even add to it. But that comes at a price - the loss of hope; the hope that, in return for risking nothing more than the effect of inflation on a bit of capital, on 1 March 'Agent Million' (that's what they call him at National Savings) will knock on my front door and say: "Here's a million quid." You can't call that priceless, but it's nice to have.

 

Premium bonds versus inflation and gilts

Premium BondsPremium BondsInflationInflationGiltsGiltsGiltsGilts
Annual returnIndexed valueInflationIndexed valueCapital return (%)Capital valueIncomeCumulative
1995*51052.59102.67.47107.54.27111.7
19964109.22.45105.1-0.11107.44.27115.9
19974.3113.93.13108.48.971174.27129.8
19984.3118.83.42112.113.62132.94.27150
19996.8126.91.56113.8-9.691204.27141.4
20003130.72.93117.24.121254.27150.6
20012.3133.71.82119.3-2.421224.27151.8
20020.5134.41.63121.35.07128.14.27162.3
200321372.91124.8-2.64124.84.27163.2
20041.8139.52.96128.51.52126.74.27169.3
20052.7143.32.83132.13.67131.34.27178.3
20062.5146.93.19136.3-4.771254.27176.3
20071.5149.14.27142.21.84127.34.27182.8
20082.5152.83.99147.810.761414.27200.8
20091.75155.5-0.53147-5.41133.44.27197.5
20101.58157.94.62153.86.21141.74.27210
20111159.55.2161.814.5162.24.27234.8
20121.42161.83.211672.12165.74.27242.5
Starting date: Feb 1995; index = 100 * 10 months