It’s hard not to feel a little head-rush when thinking about gold as an asset class. The yellow metal’s mythical status and cultural cachet is enough to trigger a reflexive “I WANTSSS IT” from your inner Gollum. But then – after briefly checking your eyes have stopped bulging – your rational side wins back control and asks: is gold a good investment? 

Well, is it? The barbarous relic? Beloved of conquistadores, James Bond villains, and pirates with dodgy accents.

Moreover given virtually all the information out there is US-focussed, is gold a good investment for UK investors? 

The Investor previously wrote a passable (his words) introduction to gold as an investment. 

TI cited the common arguments in favour of ownership as:

• A portfolio diversifier due to gold’s low correlation with equities and bonds
• Inflation-hedging
• Insurance against financial collapse, social disorder, and hyperinflation

In this post, we’ll use GBP gold return data to test how well those claims stand up – and which are about as credible as alchemy. 

But for our first stop, let’s examine gold’s historical track record, and whether that tells us anything about the precious metal’s future potential. 

(Note: All returns quoted in this article are real returns – that is, they are adjusted for inflation.)

Gold investment returns

The chart below shows the growth of gold’s investment returns from 1900-2022:

Gold GBP data from The London Bullion Market Association and Measuring Worth.
September 2023.

Over 123 years, £1 of gold transformed into £2.74.

Which translates into a real annualised return of 0.82%.

Hardly worth travelling to the end of a rainbow for. 

For comparison, over the same period the other main asset classes delivered:

• UK equities: 4.85%
• Gilts: 0.91%
• Cash: 0.45% 

Also notice how the graph’s trend line is stoved in by some monster bear markets – especially the 1980-1999 beast. We’ll come back to that. 

But there’s a snag we need to flag up immediately. Which is that, in truth, the returns history of gold is compromised by the heavy hand of State control.

Gold annual returns 1900-2022

The fingerprints of government are easier to see in the next graph:

Prices are effectively fixed up until 1968 by various forms of the gold standard, along with other restrictions including the closure of the markets due to the World Wars. 

The free market is gradually reestablished between 1968 and 1975. After ’68, the market wakes up and gold returns oscillate wildly thereafter. 

The real return fluctuations you see before then are mostly a consequence of inflation (especially during WW1 and WW2) and deflation (during the 1920s to early 1930s). 

The gold price only otherwise unmoors when the pound devalues against the dollar (1949 and 1967), and when Britain exits the gold standard (1919-1925, and from 1931 until the war begins in 1939). 

For our purposes as budding gold investors, the track record before the dawn of the free market era is too distorted by conditions that no longer apply. They should probably be disregarded.  

But naturally enough guillotining the data at any given particular point creates its own problems. 

Goldie unlocks

The free market in gold thawed in stages from:

• 1968 – Gold bars are traded again on the New York market
• 1971 – The US ends the convertibility of the dollar to gold
• 1975 – US citizens are legally allowed to own gold again (a pleasure denied them since 1933)

The restoration of the market unleashed a shock wave of compressed change, as traders attempted to discover the real value of gold. 

Thus three out of four of gold’s highest-ever annual returns medal from 1972 to 1974. 

Notice how those years’ returns rise like skyscrapers above everything else on the chart (except the Burj Khalifa of results, 1979’s 69%.)

Gold returns in the free market era

Any of the three stages mentioned above (1968, 1971, or 1975) mark plausible starting points for gold’s modern track record. 

Start digging from 1968 and we unearth a much more impressive growth story:

Data from The London Bullion Market Association, A Century of UK Economic Trends and FTSE Russell.
September 2023.

Gold turned £1 into £6.80 and notched an impressive real annualised return of 3.5% from 1968-2022.

The precious metal also ran golden rings around its rival diversifiers:

Data from Summerhaven1, S&P GSCI TR, BCOM TR, JST Macrohistory2, JP Morgan Asset Management, The London Bullion Market Association, Measuring Worth and FTSE Russell. September 2023.

1968-2022 real annualised returns (%)

• Equities: 5.7
• Gold: 3.5
• Gilts: 2.7 
• Commodities: 2.7
• Cash: 1.1

However the yellow stuff’s performance is less than dazzling if we start from 1975 – and screen out the unrepeatable gold rush of 1972-1974. 

1975 – 2022 real annualised returns (%)

• Equities: 8.4
• Gold: 1.5
• Gilts: 4.4
• Commodities: 0.9
• Cash: 1.3

The key takeaway here is that starting points matter. Not least because gold aside, the equities, gilts, and commodities results are all way off the historical averages you get from longer-term data. 

Even 50 years’ worth of asset class returns can be misleading if that period is dominated by circumstances that aren’t likely to repeat in your investing lifetime.

For example, would you expect average future returns to look like 1914-1964, with its two World Wars and a Great Depression putting the boot in?

Hopefully not.

Golden ratios

When historical data is ambiguous, we can normally lean upon expected return models to help us form estimates of future performance that take current valuations into account. 

So does a widely-respected expected return model exist for gold?

In a word: no. A few academics have taken a stab but there isn’t an accepted equation we can pull off the shelf to guide our thinking.  

So what can we expect? What factors influence the gold price?  

The greater fool theory

The greater fool theory suggests that when an asset has no intrinsic value, your hopes of making a profit rely on a ‘greater fool’ to buy it from you.  

The question then is does gold have any intrinsic value?  

Infamously, gold doesn’t offer compounding cashflow. Your golden nuggets do not pay out dividends or interest. You’re fully reliant on selling at a higher market price to make money. 

The yellow element isn’t productive like a farm, a company – or even a gold mine.

Gold is just a lump of lifeless metal that some people think looks good dangling from their ears. 

Industrial use accounts for less than 7% of worldwide demand, and is relatively price insenstive. 

Jewellery makes up almost half of demand, though. It is somewhat influenced by the going rate for gold. 

Load of old bullion

You’ve perhaps come across theories that an emerging middle class in India and China will drive demand for jewellery in coming decades? 

But that story has been hanging around for years. Meanwhile, those two giant countries have enjoyed explosive economic growth – yet the gold tonnage required by the jewellery industry is no higher now than in the early Noughties. 

Doubtless, if below-ground gold reserves ran out, that’d do wonders for your holding’s value. But a good third of gold is still entrusted to Mother Earth, relative to the total amount ever extracted. 

And what if Jeff Bezos’ asteroid mining ship ever comes in? That’d dynamite your gold ETC for the foreseeable. 

Mining for gold in space may sound like sci-fi. But think of it as an analogy for any technological breakthrough that increases the gold supply in the future.

Not to mention if the millennials of the ‘prepper’ persuasion ever do swap gold for Bitcoin en masse…

My point is that gold’s fate is obscured by a gauze of contemporary fables. If you want to remain dispassionate, then don’t get wedded to anyone’s alternative facts. 

Fools like us?

The demand inelasticity of gold’s industrial and luxury goods customers means that much depends on your fellow investors. If the world supply of ‘fools’ for gold runs dry one day then your golden goose will be cooked. 

Tellingly, the launch of gold ETFs3 during the Noughties does appear to have boosted the gold price.

According to the academics Erb, Harvey, and Viskanta:

The historical relationship between the real price of gold and the gold holdings of the two largest gold owning ETFs is shown for the period November 2004–July 2020.

As the gold holdings of these ETFs have risen, the real price of gold has risen. These two ETFs’ gold holdings represent the majority of demand for gold by ETF investors. 

The authors go on to speculate that the emergence of these ‘massive passives’ could lead to:

…higher peaks and lower troughs for the real price of gold relative to the experience of the past.

Essentially, they’re saying that the financialisation of gold via ETFs and ETCs has led to it becoming a momentum play. Gold’s star rises when prices take-off and investors pile-on. But they’re as likely to head for the hills if prices sag. 

He who smelt it…

Another hope lies in Cold War 2.0. If you’re a heavy subscriber to What the Government Won’t Tell You style newsletters, then you’ll know all about China’s attempts to diversify its central bank reserves away from the US dollar. 

And actually, there is a nugget of truth to this one. If the figures are to be trusted then China’s gold reserves have doubled in 12 years. But then China is a mite larger than it used to be.

Feverish speculation about a gold-backed BRIC currency adds to the intrigue but – as a reason to be bullish – while I think this story is a crock, it ain’t made of gold. 

Why gold can succeed when equities and bonds fall

In my view then, we have no fundamentals-based reason to expect a positive long-term return from gold. If you agree, then that could be reason enough to strike it from your investment shopping list. 

But given the high level of uncertainty, we should also look at the other reasons why gold may be a good investment. 

For instance, there’s strong evidence that gold works as a useful portfolio diversifier and can succeed when equities and bonds fall.

Here’s how gold responds when global equities take a hit:

Data from The London Bullion Market Association and MSCI.
September 2023.

The chart shows the performance of gold whenever World equities have suffered a 10%+ fall after 1970. 

During 16 equity market slumps:

• Gold beat equities: 13 times
• Did was worse than equities: twice (plus one draw)
• Produced positive returns: eight times

Six out of 16 of those sell-offs were in bear markets: 

• Gold beat equities: six times
• Gold produced positive returns: four times

So gold was really worth its weight when equities were in headlong retreat. Even the least of those bear markets inflicted a -30% knee-drop!

Using a different methodology,4 gold also bested UK government bonds in 11 out of 15 years when equities turned negative from 1970-2022. 

And it helps too – when playing defence as a UK investor – that gold gets a bump when the pound falls against the dollar – as often happens during market strife. 

All in all, the record shows that gold helps diversify risk in a traditional equity-bond mix. 

Gold’s correlation to the other asset classes 

Looking at a correlations asset class matrix can help us assess the diversification benefit of gold. An effective diversifier would register low positive or negative numbers against the other asset classes. 

Indeed, one of the main arguments in favour of gold is that it enjoys low correlations to equities and bonds.

So let’s check if that really is the case: 

Asset class returns correlations: annual returns 1968-2022 (inflation-adjusted)

	Gold	UK equities	Gilts	Cash	Commodities
Gold	1	–0.30	-0.17	-0.1	0.44
UK equities	-0.30	1	0.38	-0.09	-0.26
Gilts	-0.17	0.38	1	0.22	-0.26
Cash	-0.1	-0.09	0.22	1	0.12
Commodities	0.44	-0.26	-0.26	0.12	1

Gold’s correlation to equities and bonds is staunchly negative. This means it has a reasonable chance of pitching up when they’re tumbling down. 

Conversely, that also means gold regularly falls when those assets rise – which will most probably be the majority of the time. 

But while gold may prove to be a drag on overall returns, its typically negative correlation to your investment mainstays can help reduce portfolio volatility. 

That’s especially useful for retirees wishing to mitigate sequence of returns risk. And it’s why small gold allocations are often recommended in investment portfolio examples. 

As an individual holding though, gold is as volatile as equities. It’ll be a wild and often difficult ride. Don’t put money into it if you only like good news. 

What’s that you ask? Just how rocky can gold get?

Gold’s biggest market crash

While gold had a glittering 1970s it didn’t take long for the shine to come off. The gold market began to meltdown from February 1980. It kept sliding for another 19 years. 

Losses peaked at -78%. Recovery took until July 2011. The whole saga lasted a brutal 31 years. 

Of course, gold isn’t the only asset class that can torch wealth on a scale. Witness the UK’s biggest bond crash and worst stock market slump. 

When people say investing is risky – they mean it. 

Hopefully such nightmare scenarios won’t come to pass in our investing lifetime, but it’s still as well to be briefed on what can happen. 

Golden fleeced

For another take on how gold can cut up rough, here’s its real drawdown chart:

The graph shows how far gold dropped from its previous peaks, adjusted for inflation. The white space on the 0% line represents the precious moments after recovery is achieved and before the next plunge. 

As we can see, gold investors got little respite. 

A couple of nasty bears even dumped on the golden age of the 1970s. 

Then came the 1980 to 1999 rout. 

Even if you held on through that – until July 2011’s breakeven point – you only had to wait a few months for the arrival of the next bear that October. Cue a -40% mauling that lasted until December 2015. 

Not to worry though, the pandemic arrived to take every gold investor’s mind off it. The yellow metal duly clawed its way back into the recovery position in April 2020. A mere nine years to breakeven this time!

The point is don’t invest in gold for kicks. Its performance could leave you as regretful as King Midas. 

But enough misery. What about the claim that gold is a good inflation hedge?

Is gold a good inflation hedge?

Sadly this is a myth that’s as persistent as El Dorado. 

Take the -78% real return gold bust we’ve just talked about. The equivalent nominal loss was -46%. If gold was a good inflation hedge, it shouldn’t have shed any additional value in real terms. 

Perhaps that was a temporary glitch? Well here’s gold’s one-year return plotted against UK inflation.

The connection looks almost random. And, sure enough, gold’s annual correlation to inflation is 0.025 from 1968-2022. 

A score near zero means the relationship between the two metrics is almost non-existent. And that’s close to the pattern we see in the chart. 

Finally, here’s the path of the real gold price during the free market era:

If inflation was the only thing that moved the gold price, then the yellow line in the chart above would be level. Gold would be a perfect short-term inflation hedge.

However the fact that the yellow line weaves around tells us that factors other than inflation have caused the price to move. After all, the nominal price minus inflation equals the real price.

This means, at the least, that gold is not a good inflation hedge – because it is non-inflationary factors that will mostly determine the return you receive.

Gold as disaster insurance

The Global Financial Crisis (GFC) was the one true Time of Darkness I’ve lived through when people really talked like the wheels could come off our decadent Western lifestyles. To me at least, the pandemic didn’t even come close for scares from a financial perspective. 

And gold definitely was the ticket during the GFC. Gold ETCs were up more than 70% at the height of the crash. 

People feared QE-induced currency debasement was on the cards. Some thought a hyper-inflationary spiral could follow. 

Which meant gold melt-ups accompanied the post-GFC aftershocks, too. 

Thankfully the cashpoints didn’t close and we didn’t have to brain each other in the streets for Chicken McNuggets. But some investors did still fear the system was teetering and sought refuge in gold. 

Call that anecdotal data. 

But what does the actual data show?

Well, hyperinflation is another catastrophe that gold is reputed to repel. And Erb and Harvey sift the evidence in their paper The Golden Dilemma. Yet the financial golden boys report:

…there is no reason to expect that the real gold return will be positive when a specific country experiences hyperinflation. 

That said, they do think that gold may be less bad than paper assets when you’re paying for bread with a ten trillion note. 

So then, is gold a good investment?

I’ve read study after study, and poured over the UK data, yet I’m still ambivalent about whether gold is a good investment.

For every feature, there’s also a gold bug. 

While I’m attracted to gold’s negative correlation to equities and bonds, I shy away from its lack of intrinsic investing value. 

It’d be okay if there was a reliable demand story that forecast a procession of fools buyers long into the future. But the decisive factor appears to be investor sentiment, and that cuts both ways. 

I think an honest appraisal has to be a shrug of the shoulders. The fact is we could end up choking on gold dust as per the 1980s and 1990s. Or maybe we’ll enjoy golden years like the 1970s or Noughties. 

On the other hand, those two decades were awful for equities (and the Seventies were appalling for bonds). 

And that’s the point about gold – you buy it for its diversification potential. If everything else is going swimmingly then you’ll probably end up loathing your gold. But if it’s not…

Let’s delve more into the diversification side in the next post in this mini-series.

How much difference does an allocation to gold make to an equity-bond portfolio? Perhaps if the risk-adjusted returns are good enough, and gold confers a juicy rebalancing bonus, then the case will be clearer.  

Take it steady,

The Accumulator

1. Bhardwaj, Geetesh and Janardanan, Rajkumar and Rouwenhorst, K. Geert, “The First Commodity Futures Index of 1933,” Journal of Commodity Markets, 2020. [↩]
2. Òscar Jordà, Katharina Knoll, Dmitry Kuvshinov, Moritz Schularick, and Alan M. Taylor. 2019. “The Rate of Return on Everything, 1870–2015.” Quarterly Journal of Economics, 134(3), 1225-1298. [↩]
3. Gold Exchange Traded Products are ETFs in the US and ETCs in Europe. [↩]
4. Due to the lack of monthly gilt returns data [↩]

What caught my eye this week.

With Storm Babet tap-dancing across the flat roof of my extension, I wondered how I’d spend the windfall afternoon a cancelled trip to the provinces had gifted me.

(Broken British trains, as usual. Flooded, this time).

Obviously I chose to indulge myself by reading the new IFS Green Budget 2023 with a latte whipped up with my infamous espresso machine.

Yes, I know how to party.

Here are five interesting charts from the report, with a few thoughts on each.

1. First some good news: lots of jobs

Two defining images from the TV broadcasts of my childhood are the AIDS iceberg, and long dole queues as three million languished unemployed.

Who says nothing ever gets better? The UK unemployment rate fell to its lowest level since the 1960s at 3.6% in early Spring 2023, although it has since rebounded to just over 4%.

True, there is shade you could throw on this achievement. Many in the UK are poorly-paid. Brexit has constrained the supply of workers, so while unemployment may be lower as a consequence, it’s also pulled down GDP growth. And lots of older workers apparently gave up working during the pandemic, which might be a good thing for some but others will live to regret it.

Still, I’d take those issues over millions feeling like they’re on the scrapheap.

2. Growing nowhere fast

Absentee workers are one potential reason why GDP growth has been so slow to rebound, after the Covid shock.

And while you’d normally expect labour markets to loosen in a sluggish economy, as we saw above they have tightened. As a result wages have actually been rising (albeit only very recently ahead of inflation) despite the foot-dragging economy.

Clearly the GDP break must mostly be a hangover from the Covid disruption: fewer workers than otherwise, people focusing on the less-productive ‘experience’ economy once we could go out again, snarled supply chains, and so on.

Moreover, to look at it glass half-full I suspect this weak performance will take the edge off any upcoming recession – which we’ve been promised for 18 months – if and when it finally arrives.

It’s harder to slump when you’re already slouching.

3. Won’t anybody think of the landlords?

Rental yield isn’t everything, as any How To Get Rich In Property seminar will tell you. The big money is made from capital gains.

And that’s true, but to get capital gains you either need to see more attractive underlying fundamentals – to boost the net present value of your asset – or you need more suckers – to buy off you in the future.

It’s hard to anticipate either in this graph. Bar a blip during the financial crisis, the last time the five-year gilt yield outpaced rental yields was the early 1990s. As interest yields fell, the buy-to-let boom took off.

True, a recession and/or lower inflation bringing interest rates down could change this picture pretty quickly.

But then again in that scenario you’d also enjoy a nice bump to the value of your gilts – and without a broken boiler or a defaulting tenant in sight.

4. Big Government is watching over you

There are many notable things going on in this pretty dispiriting chart. (Do highlight anything you feel is interesting in the comments below.)

However I’ll just note here the far chunkier blocks of fiscal support (the government giveth…) and taxation (…and taketh away) seen in the last few years

While I blame this long Conservative administration for many things since 2016, I don’t blame them for the pandemic. Nor would I particular criticise the big picture decisions they and others made to try to support the economy in 2020 and 2021.

At the micro-level in Downing Street it was clearly a shitshow. But when it came to the grown-up levers of power, every country was working without a rulebook.

I think it’s mostly unfair to blame either Central Bankers or politicians for the subsequent high inflation, for example. Nobody knew exactly what to do, nor what would happen – however it looks with hindsight.

Either way though, the result is that government has been much more directly and visibly impacting household incomes in recent years.

I can’t helping thinking that this will have long-term political consequences.

5. It’s a fix

Many Monevator readers will already know the UK has moved to a predominantly fixed-rate mortgage market, but it’s still striking to see it in a graph like this.

Indeed, given the centrality of property (and property prices) to British economic life, the changeover amounts to something of a quiet revolution.

If you’re wondering why we haven’t had a house price crash yet despite the speedy rise in interest rates, fixed rate mortgages abounding is a big reason.

(Another is that about a third of UK homes are now owned outright, with no mortgage. The third is the tight labour market we saw in the first graph).

Prices holding up despite higher rates is good for anyone who already owns their own home – and frustrating for those who can’t afford to do so.

But home buying aside, the nation’s mortgage holders continuing to remortgage onto higher fixed rates over the next couple years must surely be another dampener on economic growth.

I do also wonder what will happen if rates are cut sharply in, say, 2025? Presumably we’ll see lots of stories about homeowners being ‘stranded’ on 6% fixes!

Something to look forward to…

Anyway, there’s lots more to see in the IFS Green Budget if these are your bag. Let me know if anything catches your eye in the comments below.

And have a great – wet and windy – weekend!

[continue reading…]

Bertie ‘The Investor’ Wooster settled into his comfy and ever-motivating armchair at the aptly-named Drones Club and began one of his infamous monologues…

Alright Biffy? How about a tale from mine youth? From back when I learned the great unwashed were all agog over Premium Bonds, the state-sponsored gambling racket run by National Savings & Investments (NS&I).

Now, you know I’m a contrary so-and-so. But even as a pallid stripling in my third decade I was aghast to discover I’d been eschewing Britain’s most popular savings product.

Had I missed a trick?

Am I missing one now?

Well that depends, as we shall see.

Either way, the billions of postponed daydreams the masses have stashed in Premium Bonds – £121 billion, as of the latest count – make them the nation’s favourite way to save, according to The Telegraph.

Yet for most of my adult life, Premium Bonds were a joke among the investing cognoscenti I pow-wowed with over Mai Tais in Mayfair.

Something bought by maiden aunts overdue some luck, or given to unlucky grandchildren who’d have been better-off with envelopes stuffed with cash.

Premium Bonds are odd(s) investments

Indeed as my top pals Boffo the banker and Over Easy the currency trader would banter across the billiards table, in those days only a fool would buy Premium Bonds for the chance to win a million.

“If you must be an insufferable member of the lottery punting hoi polloi…” would chirp in Erinaceous the hedge fund manager, from atop the club’s antique rocking horse, “…then put your money into a savings account and buy National Lottery tickets with the interest. The odds of hitting the big one are far higher.”

Wise words old sport!

Because what really baffled the old bonce about Premium Bonds back then was how little the blighters actually did for your wealth.

In those halcyon times – around the turn of the century – it was easy to get 5% interest on cash. Much more if you were savvy.

In contrast, the effective interest rate on Premium Bonds with average luck was in those days far lower.

Put a few bags of sand into Premium Bonds, and you gave up a stream of interest income that you could have put to a nobler use. Like punting on the nation’s favourite weekly sweepstake, par exemple.

National Lottery tickets only cost a quid then1. Every week you could buy dozens with the interest on your cash, chance your arm on the Lotto’s superior odds – and still make it to Rules for teatime.

But, of course, those dreamy days did not last forever.

Following some less-than-stellar moves by friend Boffo and his mob, the financial system turned queasy in 2008. Interest rates were duly savaged by the pink-hued rascals at the Bank of England – thoroughly filleting the spread between the misery return from Premium Bonds and what you could get from Captain Mainwairing at the bank.

Which for a while actually made Premium Bonds competitive on their own merits – especially if you were a higher-rate taxpayer.

Strange times those were.

Which brings me to… now.

Premium Bonds: interesting again

After near-abolishing the notion of making a few pennies on one’s hard-earned by cutting rates to near-0% for a decade, the brain trust of Threadneedle Street has been hiking rates for more than a year.

Indeed Bank Rate is now a chunky 5.25%.

High Street banks have followed suit. So it’s easy today to score 5% or more on your cash – and better-than 6% if you push the operation like a barrow-boy on the make.

But have these higher rates on cash kicked Premium Bonds back into the long grass?

Well – lean in chums – not exactly.

Because seemingly keen to hasten Albion’s descent into a nation of perfidious gamblers, the gang at NS&I have made Premium Bonds more generous too.

Indeed they are now paying out very nearly 5% a year in prizes.

The number of jackpots was doubled a few years ago too, to – well – two. Not amazing, but that’s still two chances to top-up the old trust fund with a £1m bag of swag.

Consider me roused. Piqued, even!

I mean, I may be the sort of absent-minded chump who leaves his Estonian beauty queen bride on the tarmac at the airport on our honeymoon because I mistook her for a shop mannequin.

But I’m not one to turn his nose up at a penny-compounding wheeze – not without a closer inspection!

Time to look at Premium Bonds with fresh eyes.

What are Premium Bonds?

Given there’s 121-thousand million pounds invested in Premium Bonds, a majority of Monevator readers will surely own a few – and already know as much about them as I do.

Hence my picaresque introduction, which no doubt has P.G. Wodehouse turning in his grave. (We’ll leave Bertie W. in peaceful repose here).

More prosaically:

• Premium Bonds are a special savings fund administered by NS&I on behalf of the British government. (Note: they are not ‘bonds‘ as professional investors use the term).

• Premium Bonds are backed by the government. This means your capital is not at risk from bank runs. (But remember all regulated savings accounts are protected up to £85,000 by the FSCS).

• You can sell Premium Bonds at any time. According to NS&I it can take up to eight working days until you the money reaches your bank. That makes them fairly liquid, but not quite instant access.

• Premium Bonds are a form of lottery bond, costing £1 each. The minimum tranche is £25. You do not ‘spend’ these bonds to take part in the lottery. Rather, they stay in your account until you withdraw them, like with cash savings.

• The maximum holding of Premium Bonds for any one individual is £50,000. This limit has been frozen since 2015.

• The effective interest rate on the £121 billion the nation has socked away in Premium Bonds is currently 4.65%.

• However the bonds do not pay interest. Instead every bond enters a monthly prize draw for cash prizes. These range from small and far more common prizes of £25 to £100, up to £1 million.

• NS&I says the odds of winning a prize per £1 bond in any given month is currently 21,000 to 1.

• Winnings are tax-free.

• The bonds were introduced in 1956, and today 24 million – or roughly one in three – Britons owns at least a few. A dedicated machine called ERNIE supposedly draws the prizes. There have been 649 million prizes since launch. They add up to a staggering £27.7 billion in total.

If you’re easily seduced, you can tootle off and buy some Premium Bonds online from NS&I.

How the Premium Bond prize money is divided up

Still here? I’m glad to see you’re not a sucker for any old 21,000:1 odds.

The next thing to understand is the nature of the Premium Bonds prize draw. This varies depending upon, for example, how many bonds are in issue.

Here’s the October 2023 prize draw breakdown:

Prize band	Prize value	Number of prizes
Higher value (10% of prize fund)	£1 million	2
	£100,000	90
	£50,000	181
	£25,000	360
	£10,000	902
	£5,000	1,803
Medium value (10% of prize fund)	£1,000	18,834
	£500	56,502
Lower value (80% of prize fund)	£100	2,339,946
	£50	2,339,946
	£25	1,027,651
Totals for October 2023	£470,853,175	5,786,217

Source: NS&I

The first thing to note is there are vastly more small prizes than large prizes.

A whopping 80% of the prize pool is allocated to prizes of £100 or less. Or, looking at the individual lines, there are only 181 prizes of £50,000 versus well over two million £100 prizes.

This skewed distribution clearly boosts the ‘savings pot’ credentials of Premium Bonds.

Why? Well, to give an extreme counter-example, imagine if the £471 million prize pot was awarded to just one winning bond in October. For all other bond holders, the interest rate that month would be 0%.

Conversely, if the whole prize fund was distributed ‘fairly’ across each individual bond, then Premium Bonds would be just like a bank account in delivering a set interest rate. Not a lottery at all.

As things stand, however, Premium Bonds are a bit of both.

There are enough small prizes to ensure that anyone with a reasonably big holding will probably win at least a few times over the year. But there is the tantalising prospect of a big win, too.

This asymmetrical distribution is why it’s so hard to calculate the precise odds with Premium Bonds.

More big prizes nowadays

Incidentally, the allocation of the prize fund has changed quite a bit over the past decade.

Most eye-catching – there are those two £1m jackpots a month.

More meaningfully, as I showed above today 20% of the fund is split across the two larger prize bands, with 80% allocated to small prizes. A decade ago, only 5% of the fund was allocated to the highest prize band, and 4% to the medium value one. Some 91% went towards smaller prizes.

I don’t recall why NS&I made this change. Perhaps it felt a 10/10/80 distribution was easier to understand? (Premium Bonds are plagued by conspiracy theories.) Or maybe it wanted to more sharply distinguish Premium Bonds from conventional savings?

The cynic in me says someone did some game theory to determine what sucked the most money in.

On that note, there’s also far more money in Premium Bonds today. That’s another reason why there are so many more individual prizes than years ago.

Back in February 2014, for instance, the prize fund was only £51m.

And just 61 people won a £5,000 prize in February 2014, compared to 1,803 in October 2023.

What are the chances of winning with Premium Bonds?

Good question. And one that’s far harder to answer than you might presume.

A few years ago Martin Lewis of Money Saving Expert semi-famously employed a cosmologist to do the multinomial probability maths needed to solve the Premium Bond prize riddle.

The result was his Premium Bonds Odds Probability Calculator. Lewis claimed this ran for six hours every month to calculate the latest odds.

With it you could see how much you might expect to win with average luck and owning a certain amount of Premium Bonds over one to ten years.

It was nice to play with. But rather than for determining your likely – but still uncertain – return, the tool was most useful for getting a feel for the shape of the probabilities.

For example:

• Most smaller Bond holdings (say £1,000) will win nothing in a typical year.

• To maximise your chances of a smoothed savings-like return, you need to hold a full £50,000 allocation.

• The odds of winning one of the two £1 million jackpot for any individual bond is about one in 60 billion.2.

Unfortunately the Money Saving Expert calculator is currently broken. It’s apparently giving the wrong odds, and has done so for a while.

That the calculator hasn’t been fixed in day or two is a clue as to how thorny the maths is!

Median expectations

Whether they’re maths genius cosmologists or not, people get into big arguments about the expected rate of return from Premium Bonds.

Some insist that the prize fund rate is the expected return. They claim any talk about expecting any other return is hocus-pocus.

But this is to misunderstand how probability works.

It’s true that the expected rate of return on the £121bn of Premium Bonds in issue is currently 4.65% – the prize fund rate set by NS&I.

But to guarantee achieving this exact return, you’d need to own all of the Premium Bonds.

And even if you were a multi-billionaire baller and so minded to do so, you’d be scuppered by the £50,000 maximum holding size.

So in practice everyone owns only a sliver of the Premium Bonds out there. And each bond they own is a shot at winning one of those prizes in the table above.

Mr Average is not so lucky

The overwhelming majority of individual bonds will win nothing in any given month.

But what is the expected return for a particular bond owner – who will certainly own more than one Premium Bond?3

Well, in the absence of Lewis’ calculator, your maths guess is good as mine.

However it’s certainly less than the prize fund rate.

To see that, recall that only two bonds will win £1m in a month.

For those individual bonds, the realised return is through the roof! A £1 investment has turned into £1m, over some period of time.

But for those two lucky jackpot winners to get their £2m, every other bond must earn a lower return.

• The prize fund is smaller by a meaningful £2m.
• But the remaining bond pool is down by only two bonds.

True, this won’t reduce the expected rate by much – it would still round up to 4.65% or so.

But you’ll recall that 20% of the prizes are in the medium and higher value category. And those chunkier prizes are won by very few bonds – just 78,674 in October, according to the table above, out of 121-odd billion.

Even for the smaller prizes, only a tiny fraction of the total bonds in issue will win anything.

Set against all that, we normally talk of interest or returns over a year. Your bonds go into the draw every month – so that’s 12 times a year when each of your bonds might win £25 or more.

Finally, as already stated, the more bonds you own the more likely you are to get ‘average’ luck.

Own just one bond4 and you can’t expect to win in your lifetime or the next. But with £50,000 you will probably get something close to the prize fund rate.

Crunch all the numbers and the typical expected return – the so-called Median return, enjoyed by an average Premium Bond holder with average luck – will be much less than the expected return.

How much do you have? Size matters

The exact expected median return will change every month as the prize fund rate changes.

But to give you a feel for the difference, some actuaries did the sums in August 2023 when the prize fund rate was increased from 3.7% to 4%.

Here’s what they calculated:

The numbers going forward will be different again with a 4.65% prize fund rate at the time of writing – higher, obviously – but you get the idea.

Remember, it’s the skewed nature of the prize fund that results in the ‘lost’ interest here for a median bond holder. Which is what makes Premium Bonds a lottery of course.

Incidentally, statisticians have studied exactly why this distribution – especially the relatively big weighting to lower-probability, higher-value prizes – makes Premium Bonds so popular with savers.

But you don’t need to be an academic to understand the dream of winning big.

Premium Bonds versus the National Lottery

On that note, how does the Lotto compare?

Well, the odds of matching all six numbers in the National Lottery are about one in 45 million. (They rose from one in 14 million with the addition of ten more balls in 2015.)

In contrast, the odds of winning the £1m jackpot with Premium Bonds is about one in 60 billion.

Yes, that’s ‘illion’ with a ‘B’.

For what it’s worth lottery tickets cost £2. Whereas a single Premium Bond ‘ticket’ is just £1.

But that doesn’t change the magnitude of the odds difference much when you’re pitting millions against billions. You’re clearly far more likely to win the jackpot on any single draw of the lottery, compared to with Premium Bonds.

However that’s not the end of the story.

With the National Lottery, your money is gone after each draw. With Premium Bonds, you get another chance next month. In perpetuity.

On the other hand the Lottery jackpots can be far higher than £1 million.

Got a third hand? Then consider that the Lottery win might also be shared.

Jackpots aside, the National Lottery is a terrible bet. Only about 50% of the money is returned as prizes.

That means the expected return for all Lottery ticket buyers is negative 50%, versus the positive 4.65% from Premium Bonds. And nearly all Lottery ticket buyers will win nothing.

Premium Bonds and inflation

Remember that Premium Bonds are not inflation-proofed, so a £1 bond will become less valuable in real terms over time.

You’d have to reinvest your prize money to try to keep up with inflation.

You can reinvest your winnings automatically via NS&I, but only up the maximum holding of £50,000. Once you’re full up, you must redirect any future winnings to an alternative account.

At the time of writing inflation is much higher than the prize fund rate on Premium Bonds – let alone the lower median expected rate.

Your Premium Bond holding will probably decrease in value in real terms for as long as this persists.

Incidentally I suspect inflation is a big reason behind the persistent myth that long-owned older bonds win much less often than newly-purchased bonds.

Somebody’s £100 Premium Bond holding was a sizeable stash in the 1970s. But if it’s never been added to then it’s not worth much now. Just a few pizzas.

Such a tiny holding is now unlikely to win for years on end, if ever. Inflation has reduced its real terms value – and the proportionate share of a £100 holding versus a £121 billion pot.

Premium bonds and your tax bracket

There’s one more thing to consider, which is your tax bracket.

Remember that Premium Bond winnings are tax-free. This makes them intrinsically more valuable to a higher-rate taxpayer than a basic rate payer – let alone to somebody who pays no tax.

You might get just over 5% interest on the best instant access savings accounts right now. That’s obviously decently higher than the 4.65% prize fund rate with Premium Bonds.

So you should prefer cash over Premium Bonds, right – far-fetched £1m jackpot opportunities aside?

Perhaps, but remember you may be liable to pay tax on your 5% interest from cash – once you’re earning above your personal savings allowance.

For a higher-rate taxpayer, for example, a cash savings pot of more than £10,000 earning 5% or higher interest would already take some of your interest income above your meagre allowance of £500.

Additional-rate payers have it worse. They get a zero savings allowance!

In contrast basic-rate taxpayers can earn £1,000 in interest before being taxed.

Tax on savings income reduces your effective interest rate

For a higher-rate taxpayer, the tax deducted from interest income above the £500 allowance essentially reduces your interest rate by 40%, if you’re comparing it to a tax-free income.

So a seemingly juicy 5% rate is now just 3%. That is well below the current prize fund rate from Premium Bonds.

Even basic-rate taxpayers will see the effective post-tax interest rate fall to 4%. Which again is less than the 4.65% payout touted by NS&I’s lottery bonds.

And as always with our stupidly complicated tax system, complications abound.

For example, if you’re living off savings and your other income is below a set threshold, you can earn far more from savings before you’re taxed.

At-a-glance guide to how tax impacts returns

Broadly speaking though, here’s how the Premium Bond prize fund rate compares to a variety of cash-like options:

Product	No tax	Basic-rate taxpayer	Higher-rate taxpayer	Additional rate payer
Premium Bonds*	4.65%	4.65%	4.65%	4.65%
5% instant access	5%	4.0%	3.0%	2.75%
6% one-year fixed	6%	4.8%	3.6%	3.3%
Top Cash ISA	5%	5%	5%	5%
Repay mortgage	Mortgage rate	Mortgage rate	Mortgage rate	Mortgage rate

*As discussed above, the 4.65% rate in the table is the prize fund rate across all Premium Bond holders. You’re probably going to earn less than this, with average luck.

These specific rates will all date pretty quickly. But the takeaway is clear.

As always, fill your ISA allowance first. And after that, perhaps consider pension contributions. You could put the pension contributions into a money market fund in your SIPP if you want to keep it in cash.

Still have cash to spare?

I’d favour conventional savings accounts until you exhaust your personal allowance. You may as well take advantage of the tax break. Buy a national lottery ticket every week if you must have some excitement.

But for savings income earned above your personal allowance, Premium Bonds do look attractive.

If you’re in the higher tax bands, they may look like a no-brainer. But please do remember, again, that their return is not guaranteed. It’s literally a lottery!

In complete contrast, paying off your mortgage is a certain return. If you’ve just remortgaged on to say a 6% fixed rate, then you’re getting effectively a 6% tax-free return when you pay some of it down.

That’s very attractive. Although it’s not quite like-for-like, in that by paying down your mortgage you’re effectively locking that money away.5

Are Premium Bonds worth it?

Unfortunately, with Mr. Lewis’ calculator broken and neither an actuary or a crystal ball to hand, it’s hard for me to be sure if Premium Bonds are worth it even for myself.

Your exact circumstances will be different again – in terms of your tax bracket, your other income, and exactly how much you would have in Premium Bonds, which goes a long way to determining your expected return.

As a rough guide though, you might knock 0.5% off the prize fund rate to think about your expected return with a full £50,000 holding.

Increase this discount if you hold fewer bonds, as seen in the results from the actuarial maths above.

You might deduct 1% from the prize fund rate if you only hold £5,000, for instance.

Happy go lucky saving

So do these median expected return rates – from 4.15% to 3.65% say – make sense?

Perhaps, as per the post-tax return table above. But remember, again, that luck will loom large.

For higher-rate tax payers with £50,000 to put away who’ve already filled their ISAs and used up their savings allowance with cash savings elsewhere, I’d say Premium Bonds look a decent bet.

You might expect (but not assume) you will get 4% or so from Premium Bonds today, versus just 3% after tax on cash.

However this advantage diminishes at lower tax brackets.

Ironically, the legions of pensioners who own small amounts of Premium Bonds and have low incomes would probably be better off swapping them for cash savings. (Assuming the resultant interest didn’t change their tax bracket).

Finally, don’t be a conspiracy theorist about whether Premium Bonds are a good investment. The myths and inconsistencies arise from probability. Unless you believe the Government is into explicit fraud.

Premium Bonds are a lottery. If you want a sure thing, buy a gilt!

Note: Comments below might refer to an older version of this article. We last asked whether Premium Bonds are a good investment in 2014! So some things will have changed.

1. £2 today. [↩]
2. Remember there are two jackpots and roughly £121 billion in Premium Bonds. Individual bonds costing £1 [↩]
3. Remember: The minimum holding size is £25. [↩]
4. Again, the minimum is 25, or £25 worth. [↩]
5. Unless you have an offset mortgage where you can easily withdraw it again if needed. [↩]