Should you invest in gold as a UK investor?

As a UK investor myself, I struggle to find useful analysis aimed at sterling investors at the best of times. Almost all of it is US-centric, and this is especially true when it comes to gold. The purpose of this article, then, is to examine gold from a UK perspective, and answer the question of whether you should invest in gold as a UK investor.

One of the things that still amazes me about gold is how divisive it is.

Nobody gets worked up over government bond funds or global equity funds. Why would they? But there seems to be something about gold which attracts Mad Max style fanatics on both sides. It’s the most extreme Rorschach test the investing world has to offer.

On the one hand, show a lump of gold to a gold bug and they’ll see an investment which:

  • Beats inflation,
  • Is uncorrelated to stocks and bonds,
  • Is the ultimate hedge against excessive central bank money-printing, and
  • Will protect you when markets crash.

On top of all that, it’s great at preventing the spread of coronavirus:

Gold and coronavirusSource: The Economic Times

Not only can gold be used to parade your status as someone with the means to afford what’s probably the heaviest, least practical, and least effective face-covering possible, but it’s also an investment which generates positive real returns, is uncorrelated with both stocks and bonds, and serves as portfolio insurance.

Put that way, gold sounds like a no-brainer both in your wardrobe and in your portfolio.

But show that same lump of gold to another investor and they’ll see something which:

  • Produces no cash,
  • Is useless in a real crisis,
  • Doesn’t hedge short-term inflation, and
  • Has no intrinsic worth, so is reliant on a greater fool.

This investor argues that by buying gold, what you’re really doing is buying an ownership stake in rocks. And rocks aren’t particularly known for their ability to generate cash.

So why would anyone ever own gold?

The purpose of this article is to explore both sides of the argument, and answer the question of whether gold belongs as a core holding in a UK-based investor’s long-term portfolio.

It is not an article about whether you should buy gold now.

As regular readers will know, this blog is anti-market timing, and I see little reason why investors should be better at timing their investments in gold than in any other assets. But it still raises an interesting question of whether gold belongs as an additional asset class alongside stocks and bonds in a buy-and-hold portfolio.

The majority of gold proponents argue gold is useful in a portfolio because it’s a good hedge against rising inflation, it can protect a portfolio from an equity market crash, and it’s generally uncorrelated to either stocks or bonds.

To address each of those points, I’ve split the article into sections.

I’m aware that there are zealots on both sides of the gold divide, and by writing this I’ll probably attract hate mail from both the rock collectors and the Buffett Brigade.

If you’ve already made up your mind on gold, and your opinion has now ossified into an immutable truth, this article isn’t for you.

But for those who haven’t yet sworn a blood-oath to one of the gold factions, and aren’t clinging, Gollum-style, to your opinion, please read on. 


Beware the backtests


Understanding gold’s history is probably more crucial than any other asset class for understanding its past performance.

The trouble with gold is that its nature as an investment is split into two distinct periods – before the gold standard, and after.

Before 1971, each country valued their currency based on a fixed weight in gold. For example, for the century in the United States from 1834 to 1934, the gold price was legally set at exactly $20.67 per ounce. Gold was literally considered money, and paper bills were directly exchangeable for gold at the established rate. This was known as the gold standard.

Then, In 1944, 44 countries created the Bretton Woods Agreement. Bretton Woods was designed to set global monetary policy based on a worldwide gold standard. All individual currencies were priced in a set amount of dollars, which in turn was backed by a fixed amount of gold.

This all lasted until 1971, when Richard Nixon eliminated the convertibility of cash to gold and subsequently ended Bretton Woods.

Although the gold price was no longer fixed with the dissolution of Bretton Woods, at this point it was still illegal for private citizens to own gold.

If people think the Fed are doing a good job manipulating markets today, at least they haven’t made it outright illegal to own an asset class.

To stem the outflow of gold from the US during the Great Depression in the early 1930s (and to prevent people from hoarding gold and hence stalling economic growth), President Hoover signed Executive Order 6102 in 1933. This executive order forced all Americans to convert their gold coins, bullion, and certificates into U.S. dollars.

It became illegal to own gold.

Executive Order 6102Source: Wikipedia

Amazingly this lasted for over 40 years, and although Bretton Woods ended in 1971, it wasn’t until 1975 that private ownership of gold was fully legalised in the US.

In 1975, gold was finally able to free-float and also act as an investable asset class.

So what does that mean for backtesting?

Gold and Bretton Woods

Source: Portfolio Charts

Here you can very clearly see the problem with mixing gold datasets before and after 1975.

The gold price was artificially fixed before 1971, then suddenly set free. Although Bretton Woods ended in 1971, the initial price surge following gold’s release sought the “true value” of gold, before it was finally available for investors to buy in 1975. As a result, the 1971-1975 melt-up in the gold price formed a massive, unrepeatable return that skews any analysis.

While pre-1975 gold and post-1975 gold may both be called “gold”, they track completely different assets following very different rules.

One of the major problems with backtests is that more data doesn’t always equal better conclusions. A big problem with old data – data which reaches further back in time – is that it often portrays an environment which isn’t representative of the world today.

For another example, the graph below shows the evolving nature of the US stock market:

US market sector evolutionSource: Visual Capitalist

If an investing strategy relies on data going back 200 years, it might look impressive on paper. But does the fact that it worked when financials were over 90% of the market mean that it must also work today? The technology sector didn’t even exist then.

And so the problem is with gold. Pre-1975 data is irrelevant for today’s investors, because 1) we no longer have the gold standard, and 2) it was literally un-investable between 1933 and 1975.

We must treat any gold backtests which rely on pre-1975 data with a bucket-load of salt.


Why is gold valuable?


I’ve often wondered what makes gold so special.

Investors go nuts for gold, but I’ve never seen anyone have strong opinions on palladium, silver, or platinum. Why gold?

Luckily, Nick Maggiulli of Ritholtz Wealth Management has already written the post on why gold is valuable. He always writes far more eloquently than I do, so I’ll paraphrase his conclusions here, but I highly recommend reading his full article.

In short, gold is valuable because:

  • It’s scarce. But not too scarce. It has a goldilocks level of scarcity which makes it rare, but common enough to encourage plenty of competition for its mining, and for it to be well-known as a status symbol.
  • It’s durable. It doesn’t rust or tarnish. Durability means gold can act as a long-term store of value. As Peter Bernstein wrote, “In Cairo, you will find a tooth bridge made of gold for an Egyptian 4,500 years ago, its condition good enough to go into your mouth today.”
  • It’s malleable. It’s so flexible that one ounce can be made into a wire over 50 miles long. Think about how useful of a quality this is for a store of value.  You can divide it, hide it, or transport it from place to place with very little effort.
  • It’s difficult to acquire. Gold is hard to mine, and hard to keep hold of. As Joe Weisenthal noted in a speech from May 2018:

To get gold you:

      • Have to be good at warfare
      • Be able to marshall an extensive human workforce to mine it
      • Mastery of global supply and logistics routes
      • Be able to command guards who will watch your gold, and not steal it
      • Have the technical know-how to get gold out of the ground, which is expensive

In other words, when you have gold you’re communicating all the different things you’re capable of (mastering supply routes, commanding an army, scientific endeavor, marshalling labor, etc.)

Gold, then, is a very specific proof of work. If you can get gold, you’ve proven that you have the ability to run a state or some state-like entity.

In his post, Nick provides an interesting parallel with the history of aluminium. (Because Americans can’t spell, however, he repeatedly calls it ‘aluminum’.)

When aluminium was first discovered in the late 1700s, it was initially more valuable than gold because of how difficult it was to obtain.  In fact, aluminium was used to cap the Washington Monument, when it was first completed in 1880.

However, as the refining process improved and aluminium extraction became easier, the once rare metal became much less rare.  The increased availability of aluminium decreased its proof of work and, thus, reduced its value relative to gold.  What started as the capstone to one of America’s most treasured landmarks eventually became the go-to container for sugar water

That’s all well and good, but what is gold actually used for?

There’s no point having something that’s scarce, durable, malleable, and difficult to acquire if it’s not used for anything.

The chart below shows the share of demand between gold’s four major uses:

Sources of demand for gold

Source: Bloomberg

Jewellery is gold’s largest use by a long way, constituting almost half of all demand for gold. Investment comes in second at almost a third of demand, followed by central banks, and finally technology.

What the chart makes clear is that gold’s value doesn’t come from its industrial uses. Only 7% of gold’s demand comes from its function as an electrical conductor in technology.

Central banks hold gold because they believe it’s a safe-haven asset, it’s liquid, it’s a hedge against inflation, and it’s an uncorrelated asset. In short, they hold it for the same reasons investors do (all of which will be covered below). So combining central bank holdings with private investment holdings means that 45% of gold demand comes from some form of investment.

In total, this means over 90% of gold’s demand comes from two sources: jewellery and investment. Essentially, gold is in demand because a) it looks pretty, and b) investors believe other investors believe it’s a good investment.

Let’s dig into the second reason gold has value, and investigate whether it lives up to its reputation as a good investment by examining its function in a diversified portfolio.



Gold’s function in a portfolio



Gold as an inflation hedge


There are a few things to think about when assessing’s gold’s ability to hedge inflation risks, so I’ve split this section up into parts:

  • Gold as a cash generator
  • Gold as a long-term inflation hedge
  • Gold as a short-term inflation hedge
  • Gold as an unexpected inflation hedge
  • Gold as a hyperinflation hedge

Gold as a cash generator


“Gold generates no cash” is one of the most common arguments from the anti-gold crowd. It’s one of the reasons Warren Buffett was famously against gold (until he went and bought a slug of a gold mining company), so is a Buffett Brigade favourite.

It’s true, owning gold earns you no income. Shares pay dividends – or at least the promise of future dividends – and property and bonds pay an income, too. Even cash in the bank earns you some interest.

Not only does gold not generate any cash, it costs you money to keep it.

The lack of cash generation means gold provides nothing to reinvest. Companies use their cash to reinvest into projects with positive NPVs, which increase the company’s value. Bonds generate cash in the form of coupons, which can be reinvested.

While cash generation isn’t necessary for compounding (anything with a positive growth rate, given enough time, will compound), cash generation is usually associated with the ability to compound because cash generation makes a positive growth rate more likely.

(Note, although dividends represent cash returned to shareholders, they shouldn’t be viewed as generating positive return. Dividends are returns of the investor’s capital, not a return on capital. In this instance, I’m referring to cash generated by the company, reinvested by the company.)

I think this is the crux of the argument used by the anti-gold crowd. Cash generation increases the likelihood of positive returns, so gold’s lack of cashflow means a positive growth rate is less likely.

But is this the case? Gold has a reputation of being an excellent way to protect against rising inflation.

If gold is able to beat inflation, and therefore has a positive real return, why does it matter whether it generates cash or not?

Gold as a long-term inflation hedge


The chart below is from Burton Malkiel’s classic, A Random Walk Down Wall Street. What makes the chart interesting is its time-frame. It shows returns of four major asset classes over the last 200 years.

Gold vs stocks and bonds

Source: A Random Walk Down Wall Street (p293)

Clearly, stocks have been the best place to be for the long-term investor. But we already know that.

While the anti-gold crowd will use the chart to say “gold’s long-term returns are far lower than any form of bonds, and even worse compared to stocks”, the exact same chart can be used by gold bugs to argue that “gold’s long-term returns are excellent at beating inflation”.

Both look like they’re true.

The great thing about gold as an asset class is that we can go back even further than 1800.

Going back to when gold used to be the method of payment for the military – an institution that’s been around for thousands of years – we can see that a Roman Legionary was paid roughly the same in ounces of gold than a Private in the US Army would be today. Similarly, a Roman Centurion would’ve been paid roughly the same in gold as a US Army Captain would be today.

Military pay in ounces of goldSource: Financial Analysts Journal

Now let’s go back even further. In 562 BC, during the reign of the Babylonian king Nebuchadnezzar, one ounce of gold could purchase 350 loaves of bread.  At the recent price of $1,900 an ounce, one ounce of gold today could buy 350 loaves of bread priced at about £4 a loaf – the same price as an upmarket bakery’s sourdough.

Gold looks to be extremely strong at maintaining its purchasing power over enormously long periods of time.

However, as astute readers will have noticed, all of that data relates to before 1975. And we know that any data before 1975 isn’t much help to us, as it relates to what’s essentially a different asset class.

Gold is no longer fixed to the value of a country’s currency, as it was in Babylonian and Roman times, and is now a free-floating independent asset, untied to any single currency or country. So although it looks great at keeping up with long-term inflation, that fact isn’t incredibly useful for today’s gold investors. Firstly, their investment time horizon isn’t likely to be in excess of a thousand years, and secondly gold today is a very different asset class than it was 60 years ago – let alone 2000.

Gold as a short-term inflation hedge


It’s interesting to see how well gold has done at maintaining purchasing power. But not only does the data represent a different asset to the one we’ve got today, it’s also unreasonable to expect anyone to wait 2,000 years for their investment to keep up with inflation.

Even a 40 year time horizon is short compared to the history of gold, but most investors operate under time horizons closer to 5-10 years. And that’s being generous. From my experience in the institutional world, the more intermediaries are involved in managing an investor’s money, the shorter the time horizon becomes. Activity is used to justify each intermediary’s existence, and activity is more easily justified based on short-term performance – sometimes a couple of years, or even a couple of quarters.

So how does post-1975 gold do as a hedge against inflation over shorter time frames?

Using data from Bloomberg, which probably wasn’t available to the Romans or Babylonians, we can explore the correlation between gold’s price and inflation. The scatterplot below shows the correlation between gold (in sterling) vs UK RPI.

Gold as a hedge for UK inflationSource: Bloomberg. All figures in GBP.

You don’t need any of those statistics on the right hand side to conclude that there’s no correlation between gold and UK inflation. Let alone a statistically significant positive correlation, which is what we’d be hoping to see if we’re hoping for gold to keep up with inflation as a UK investor.

If you were a US investor, and lived in the land where only US inflation mattered, then the relationship looks a little better:

Gold as a hedge for US inflation 1 2Source: Bloomberg

Here we’ve got a slightly positive correlation between gold (in USD) and US inflation. This is likely the basis for the popular theory that gold is a good inflation hedge, even in the shorter term.

But the correlation is clearly skewed by a few years where inflation was extremely high at the same time as gold produced very strong returns. Those three dots in the top right of the scatterplot represent the years 1979, 1974, and 1980.

We have two ways of mitigating the effect of these outliers – looking at monthly data, and running the analysis from 1980.

First, the monthly data:

Gold as a hedge for US inflation 2 2Source: Bloomberg

That’s a correlation of 0.13. Hardly a ringing endorsement of gold’s positive correlation with inflation.

And now, since 1980:

Gold as a hedge for US inflation 3 2Source: Bloomberg

Now the correlation’s negative!

Gold’s relationship with inflation is clearly not as simple as it’s made out to be.

For UK investors, gold has shown no evidence of being able to hedge against inflation.

Its famous correlation with inflation is true for US investors, but that’s mainly due to a few specific years where gold did well against exceptionally high inflation. For UK investors (and in all other instances for US investors) it looks like a pretty poor inflation hedge.

Whether or not you believe that to be useful in combating inflation is for you to decide.

You can either take the view that:

  1. Gold isn’t useful as an inflation hedge because it doesn’t work most of the time. Or you can take the view that
  2. Gold is useful as an inflation hedge, because it might protect you in the very few times where inflation is exceptionally high.

For those who believe that gold would be useful to include in a portfolio because of its ability to hedge against the small number of years with unusually high inflation, we can further explore that point by looking at:

2.a. How gold fares as a hedge against unexpected inflation, and

2.b. How gold fares as a hedge against hyperinflation.

But before we get to point 2, it’s worth having a look at some additional evidence which explores the first point – that gold isn’t a useful short-term inflation hedge. Specifically, what sort of price do gold investors have to pay to reap the benefit of hedging the few high-inflation periods?

Gold’s drawdowns are real


If I asked you which asset class had a higher maximum drawdown: stocks or gold, what would be your answer?

I imagine 99% of people would say stocks.

Everyone knows stocks are the riskiest asset class, and with high risk comes high drawdowns. You never hear about “gold crashes”, but equity market crashes happen all the time. Gold is safe – at least compared to stocks.

But you’d be wrong.

The chart below shows gold’s maximum drawdowns after inflation, compared to stocks. To ensure it’s relevant for UK investors, the chart is priced in sterling:

Real drawdowns - gold vs stocksSource: Bloomberg. Stocks are the MSCI World Index (GBP). Gold is also priced in GBP.

Someone investing in gold in 1980 would’ve suffered an 80% drawdown between 1980 and 2000. 80%!

By comparison, the largest equity market drawdown was 56%. Gold’s largest real drawdown absolutely crushed the equity market drawdown.

Not only was it larger in magnitude, but gold’s drawdown lasted for 30 years. Gold reached its high in 1980, and didn’t reach that point again until 2011. While it did well during the final 10 years of the drawdown at keeping up with inflation, the 20 years between 1980 and 2000 must have been psychologically unbearable for those investors holding gold for the sole reason of keeping up with inflation.

On January 21, 1980, the price of gold reached a then-record high of $850. On March 19, 2002, gold was trading at $293 – a loss of about 65% before inflation. After factoring in the positive inflation over the same period, gold’s loss in real purchasing power was about 80%.

I can’t imagine holding an asset for one specific reason, then having to sit through a period of 20 years where it fails to perform that function.

Although gold may have now recovered from the drawdown, a 20 year period where gold fails to keep up with inflation doesn’t support the idea that gold is useful as an inflation hedge, at least in the short(er) term. No investor can stick with something that fails to perform its function for 20 years.

Gold as an unexpected inflation hedge


It doesn’t look like gold is great at protecting against shorter term inflation. But what about short bursts of unexpected inflation?

Gold and unexpected inflationSource: Financial Analysts Journal

We can see that 1980 year of high inflation/high gold price as a clear outlier. But apart from that single year, gold has done a poor job of protecting against unexpected inflation.

Not only has gold done this job poorly, but for those looking for an asset to protect against unexpected inflation, inflation-linked bonds are designed to do exactly that.

Why would you use gold as a hedge against unexpected inflation when there’s already an asset which was created to serve that exact purpose? With inflation-linked bonds, there’s no need to worry about multi-decade real drawdowns, the global demand for jewellery, or getting lynched by rock collectors.

Gold as a hyperinflation hedge


If gold isn’t any good at protecting against short-term inflation or unexpected inflation, how about hyperinflation? It’s definitely a tail risk, but is still possible.

There aren’t many examples of countries experiencing hyperinflation, and fewer still that occurred after 1975 – so we’re running the risk of the n=1 problem, and making inferences based on a very small sample size.

Still there are a few interesting examples which come to a similar conclusion, so it’s worth looking at them here.

The hyperinflationary period for Brazil between 1980 and 2000 is a good place to start. The table below shows the state of the Brazilian currency versus gold during its 20 year period of hyperinflation.

Gold and hyperinflationSource: Financial Analysts Journal

The table shows that gold lost about 71% in real terms over those 20 years in local currency terms. However, investors who kept their cash under a mattress or invested in a portfolio of Brazilian nominal bonds likely lost almost all of the real value of their assets. Compared with an almost 100% decline in real value for cash and nominal bonds, the 70% decline in the real value of gold was a great alternative.

So, if purchasing power declined 70%, was gold a successful hedge against Brazilian hyperinflation?

It depends on what you consider to be a “good hedge”. One thing is clear though, and that’s that there’s no reason to expect the real return for gold will be positive when a country experiences hyperinflation.

In fact, a better hedge against hyperinflation, as shown in the cases of Weimar, Zimbabwe and some empirical research is to keep your money in stocks.

Stocks represent the liabilities of the entities which are able to maintain their pricing power even in the case of a very high inflation, which intuitively makes it more likely they’d be able to keep up with extreme inflation.

Overall, the rarity of hyperinflationary environments makes drawing conclusions difficult, but the few examples we do have indicate that gold is likely to fall short as a hyperinflation hedge, at least compared to equities. 


Gold as crash protection


Does gold really protect you in a crash?

To me, this is one of the most important arguments in favour of gold, as tail-risk hedging assets are incredibly valuable in portfolios. I prefer pictures to words, and I think the two charts below do a good job of showing how gold has reacted in the past during equity market selloffs.

The first chart shows every market selloffs in excess of 10% since 1975, along with how gold performed during each of those drawdowns:

Gold performance in a market crashSource: Bloomberg. Equity is the MSCI World Index (GBP). Gold is also priced in GBP.

Gold looks like it does a pretty good job during a crash. It protected portfolios particularly well during the two most severe crashes in 2000 and 2008, posting positive returns during both. Gold’s return during the 2008 crash was particularly impressive, and would’ve made an excellent form of insurance.

Even during smaller drawdowns, gold fell less than the market, so would have provided at least some form of protection. In fact, gold fell less than stocks in every market selloff over 10% since 1975. In 6 out of 10 of the drawdowns, gold posted positive returns, and in the remaining 4 it fell less than the market.

While it’s not a perfect hedge (it doesn’t always go up when stocks go down), gold looks like it does perform reasonably well as crash insurance.

The chart below shows how a 20% allocation to gold would have mitigated those drawdowns:

Gold effect on market drawdowns 2Source: Bloomberg. Equities are the MSCI World Index (GBP). Gold is also priced in GBP.

As expected, the largest improvement in drawdowns was in the 2008 crash. An allocation to gold also did well in 2000, as we know, and helped dampen the fall in all other drawdown scenarios.

So it’s done a good job mitigating drawdowns in two scenarios, and a respectable job in the remainder. But at what price? Given gold has a much lower expected return than equities, how much are we paying in returns to benefit from gold’s crash protection?

The table below shows how a portfolio of equities fares when increasing allocations to gold are introduced:

Gold and risk adjusted returns 2Source: Bloomberg. 1975-2020. RAR = Risk adjusted return. MDD = Maximum drawdown. All figures in GBP.

Surprisingly, at least to me, gold makes little difference until you start allocating c.50% of the portfolio to gold. Risk decreases slightly until the portfolio is roughly evenly split between bonds and gold, then starts to increase again as the portfolio becomes dominated by gold. Maximum drawdowns also see a U-shape, with high MDDs for both pure equity portfolio and pure gold portfolios, but lower drawdowns for a blend of the two.

Here you can see the power of introducing an uncorrelated asset class. Gold produced a much lower return than equities, along with even higher drawdowns and higher risk. It’s worse than stocks on all three metrics. Yet introducing this higher-risk, lower-return, higher-drawdown asset increases risk-adjusted returns and lowers maximum drawdown. A blended portfolio resulted in higher risk-adjusted returns and lower drawdowns than either of them individually.

We should be wary of drawing too many conclusions from backtests (for all the reasons listed here), but it looks like a moderate allocation to gold, while only slightly reducing returns, does well at reducing risk and lowering maximum drawdowns.

For a more detailed investigation into how gold allocations affect a portfolio’s risk/return characteristics, I dive into more detail in the Gold as an uncorrelated asset section.

Gold in the zombocalypse


Now we have some idea how gold fares during a stock market crash.

But what if a stock market crash was the least of your worries? How useful would gold prove to be if war breaks out, and you’re forced to rely on it to preserve your wealth? Or worse, what if rage-infected monkeys cause the next Zombie apocalypse and you need to go to the Winchester, have a pint, and wait for the whole thing to blow over?

In either case (war or the zombocalypse), the weight of gold limits its portability. Although gold is valuable because of its durability and malleability, it’s still heavy to carry, which makes it less than ideal in a take-what-you-can-carry scenario. Thinking in terms of gold’s value to weight ratio (a doomsday version of the Sharpe ratio), there are plenty of other options, including precious gems, which have a higher value to weight ratio than gold.

In addition to being difficult to transport, how useful would gold actually be if you’re struggling for survival? Most likely, you’ll be fighting (perhaps hand to hand) for resources. The only metal you’re going to be worried about is lead. For bullets. You’re not going to be lugging around gold bars so you can trade.

When a person with lead meets a person with gold, the person with lead ends up with the gold, and the person with gold ends up with the lead – probably in the head.

Any value gold might have in this breakdown scenario assumes you’ve already managed to get hold of your gold. For most people, their bars would be stored in a bank. In a doomsday scenario, how likely would it be that you’d be able to access that gold? I’d wager not very likely. Not only would the bank be locked down to prevent looting, anything accessible would have already been nicked by the employees trying to look after their family. And you have to imagine that banks would be well monitored by opportunists looking to ambush those fortunate enough to retrieve anything from inside.

But all of this assumes you’re holding gold in physical form, and the gold allocation in your portfolio is held in the form of gold bars. Most people have their exposure not via gold bars, but via funds (usually ETFs). Although technically backed by gold reserves, would anything but physical ownership be useful in case of a complete breakdown? Depending on the exact type of disaster, there may not be a functioning stock exchange where you can sell your gold backed securities.

Overall, I’ve never found the doomsday scenario particularly persuasive as an argument against gold. I’ve never encountered any gold fans honestly argue that gold would be useful in a fighting-for-your-life type of environment. More often I see it used by the anti-gold crowd as a strawman argument, unfairly extrapolating it from the more reasonable “gold will protect me in a crash” argument.

Those truly worried about a dystopian future should worry less about the nuances of portfolio construction, and worry more about stocking up on guns, ammo, and the Batman soundtrack.


Gold as an uncorrelated asset


I think we can safely rule out gold’s usefulness in the zombocalypse. If we return to the land of sensible portfolio construction considerations, and thinking about things which are actually useful for investors, we can start to examine gold’s function as an uncorrelated asset in portfolios.

To start with an obvious point, you wouldn’t want to bet on gold beating stocks over the long-term:

Gold performance versus stocksSource: Bloomberg

It’s no contest. Stocks for the long run. But despite its lower returns, one of the major arguments in favour of gold is its ability to zig when stocks and bonds zag.

This point isn’t so much about crash protection, but more about introducing an asset which is uncorrelated to the rest of the portfolio, which helps smooth volatility, reduce smaller drawdowns, and helps make a portfolio easier to stick with.

It’s true gold looks uncorrelated to equities, bonds, and real estate:

Gold's correlationsSource: Bloomberg. GOLDS = Gold, NDDUWI = MSCI World, LGCPTR = Global hedged corporate bonds, LGAGTR = Global hedged government bonds, MXWO0R = MSCI World Real Estate Index. All priced in GBP.

So how does this uncorrelated return impact portfolios risk/return characteristics?

We saw above how introducing gold into an equity portfolio helped reduce risk and drawdowns while only slightly lowering returns over the full time period (1975-2020).

But realistically most investors base their investing decisions on time periods far shorter than that. So how do gold allocations look when assessed on rolling 5 year bases?

I ran the numbers, taking each monthly rolling 5 year period since 1975, and finding the percentage of rolling 5 year periods where a 20% gold allocation was as an improvement on 100% equities:

Gold's effect on a portfolio 1 Source: Bloomberg. RAR = Risk adjusted return. MDD = Maximum drawdown. All figures in GBP

Of the 490 rolling 5 year periods between 1975 and today, a 20% gold allocation would have improved returns in only 40% of them. It would’ve reduced risk in almost 90% of periods, and improved risk-adjusted returns in about half of them. It would’ve reduced maximum drawdowns in almost 90% of 5 year periods.

But the statistics above don’t give any idea about the magnitude of the differences. In the 40% of periods where gold improved performance, it could have improved returns by hundreds of percent. In the 60% of periods where it didn’t, it could have only taken a few basis points off performance. In which case, gold’s return profile would look extremely attractive, despite detracting from portfolio performance the majority of the time.

The table below puts some numbers to it, comparing a 100% equity portfolio to a portfolio of 80% equity, 20% gold.

Gold's effect on a portfolio 2Source: Bloomberg. RAR = Risk adjusted return. MDD = Maximum drawdown. All figures in GBP

On average (using the median), a 20% allocation to gold slightly reduced returns, slightly reduced risk, slightly improved risk-adjusted return, and reduced maximum drawdowns.

Taken together, the numbers show that gold is good at consistently reducing the risk of an equity portfolio and dampening the effects of crashes. Although it detracts from returns over 50% of the time, the reduction in return is slight, resulting in a roughly 50/50 chance of a gold allocation improving risk adjusted returns.

What I also found surprising was that, even with a relatively large 20% allocation, the improvements were marginal. The improvements in risk and RAR were slight, and likely a result of the few times gold managed to mitigate the effects of large equity drawdowns. The benefits of gold only start to become apparent if the investor is willing to carve out a meaningful portion of their portfolio to allocate to it.

Gold, while not harming the portfolio, is far from guaranteed to improve return, or risk-adjusted returns in any given period. Its track record is mixed, but its risk-reduction ability does look attractive.

When viewed holistically, in conjunction with an equity portfolio, I can see how gold has the potential to add value.

However, as the chart below shows, gold has gone through long periods of underperformance:

Gold's rolling 5 year returns vs stocksSource: Bloomberg. All figures in GBP.

The chart above shows how cyclical investing in gold can be. Between 1984 and 2002, gold underperformed stocks in every rolling 5 year period. It provided crash protection for all the drawdowns during that period, but it still must have been a difficult asset to hold given its underperformance.

While gold might be useful when held alongside a traditional equity portfolio, gold is only useful if the investor is able to hold it through the periods of maximum pain.

Gold heavily underperformed stocks right up until the dot-com crash, with a 5 year underperformance of almost 200%. Then, right at the point most investors would surely be itching to sell their painfully underperforming gold allocation, not only did gold go on to help mitigate the crash, gold went on a rampage for the next 10 years, culminating in a 5 year outperformance for gold of almost 250% in 2010.

Overall, gold might improve a portfolio’s return metrics, but only if:

  • The investor is willing to carve out a meaningful chunk of their portfolio in order to see any benefits, and
  • The investor has the stomach to hold it through some long, painful periods of underperformance. I don’t know many people able to handle 5 year underperformance of over 300%. But if you’re able to, gold has the potential to be a useful source of uncorrelated returns.


Gold and sterling


A final note on currency considerations before we head into the practicalities of investing in gold, and some concluding thoughts.

Given that gold is essentially a free-floating currency, the value of any gold position in a sterling-based portfolio will rise when sterling falls, and fall when sterling rises. Gold is like having a position in an unhedged foreign currency.

The currency impact that gold has on your portfolio will therefore depend on how much of your portfolio is already exposed to FX fluctuations.

If you hold an equity portfolio of an unhedged position in a global index-tracker, for example, then gold will do little to dampen the portfolio’s vulnerability to currency swings, as the index-tracker’s underlying securities are mostly non-sterling – just like gold.

If, however, your portfolio is sterling-heavy, and contains mostly UK companies which derive the majority of their earnings domestically, then gold will likely benefit the portfolio. If sterling fell, the value of the equities would likely fall too (as a weakening currency is usually a result of the market having less faith in the future of the country). Given gold isn’t priced in sterling, the value of the gold position would help offset the losses. Of course, the opposite is true if sterling strengthens, but that’s diversification for you.

Overall, gold could be a good option for diversifying a sterling-heavy portfolio, but these investors should also consider diversifying through way of non-sterling equities. Equities priced in foreign currencies would do an equally good job of diversifying a portfolio by currency. For those investors whose portfolio already consists of non-sterling assets, gold won’t provide much in the way of currency diversification.


How to own gold


There are three main ways to add gold exposure to your portfolio: ETFs, physical gold, or shares in gold miners.

Gold ETFs are the easiest way to get exposure to the gold price. ETFs are cheap to hold, and are backed by gold bullion. A synthetic ETF does not own any gold, just derivatives linked to the gold.

Gold bars are another option. Physical gold removes any uncertainty around ownership, and provide more “peace of mind” value than an ETF. However, gold bars will likely cost more than an ETF as the bars need to be transported, stored, and insured. There are also costs involved in testing the gold before you buy it, which also costs money, as well as fees charged by the intermediary for the transaction.

Buying shares in gold miners is a third option. Shares in gold miners do give you exposure to the gold price, as the value of a gold mining stock is heavily dependent on the gold price. However, they also give you exposure to company-specific risks, tied to the outcome of the companies you’re investing in – including the risk that the company is mis-managed, makes poor decisions, and suffers problems unrelated to the price of gold. With shares in miners, you’re exposed to market risk, which means the companies are likely to fall when the rest of the equity market is falling – exactly the kind of risk most gold investors are trying to avoid.

How to own gold 2

Source: State Street Global Advisors



  • Gold looks to be extremely strong at maintaining its purchasing power over enormously long periods of time. However, unless your expected lifespan is over a thousand years, this isn’t incredibly helpful for investors.
  • Any data before 1975 isn’t much help to us either, as it relates to what’s essentially a different asset class. Gold is no longer fixed to the value of a country’s currency, as it was in Babylonian and Roman times right up until 1971, and is now a free-floating independent asset, untied to any single currency or country.
  • As an inflation hedge, gold has shown no evidence of being able to hedge inflation in the UK. The evidence is more mixed in the US, but a 22 year period where gold fails to keep up with inflation doesn’t support the idea that gold is useful at protecting against rises in inflation, at least in the shorter term. Not only did it fail to keep up with inflation, gold suffered an 80% real drawdown versus inflation. No investor can stick with something fails to perform its function for so long, or in such spectacular fashion.
  • Gold has also done a poor job of protecting against unexpected inflation. Not only has gold done this job poorly, but for those looking for an asset to protect against unexpected inflation, inflation-linked bonds are designed to perform this exact function, so are likely to be a more reliable form of hedging unexpected inflation.
  • The rarity of hyperinflationary environments makes drawing conclusions difficult, but the few examples we do have indicate that gold is likely to fall short as a hedge, at least compared to equities.
  • Gold looks like it does a pretty good job during a crash. It protected portfolios particularly well during the two most severe crashes in 2000 and 2008, posting positive returns during both. Overall, gold fell less than stocks in every market selloff over 10% since 1975. In 6 out of 10 drawdowns, gold posted positive returns, and in the remaining 4 it fell less than the market. While it’s not a perfect hedge (it doesn’t always go up when stocks go down), gold looks like it does perform reasonably well as crash insurance.
  • Gold is not likely to be much use in a zombocalypse – you’d be better of with lead for bullets. But that doesn’t have much relevance in portfolio construction decisions, and is really a strawman argument. It shouldn’t be confused with gold’s ability to protect you in a more run-of-the-mill market crash.
  • Gold has the ability to improve a portfolio’s risk adjusted return metrics, but only 1) if the investor is willing to carve out a significant portion of their portfolio, and 2) if an investor has the stomach to hold that allocation through long, painful periods of underperformance.
  • As a currency diversifier, gold could be a good option for diversifying a sterling-heavy portfolio, but these investors should also consider diversifying through way of non-sterling equities. Equities priced in foreign currencies would do an equally good job of diversifying a portfolio by currency. For those investors whose portfolio already consists of non-sterling assets, gold won’t provide much in the way of currency diversification.
  • There are three main ways to add gold exposure to your portfolio: ETFs, physical gold, or shares in gold miners. Each have their pros and cons, but ETFs are likely to be the main vehicle for investors looking for gold exposure thanks to their liquidity and low cost.




We started off this article with the goal of examining both sides of the argument for gold. Gold’s proponents argued that gold serves three primary functions: to protect against inflation, to protect against crashes, and as a source of uncorrelated returns.

What the numbers say

As a way to hedge against inflation, gold comes up short – at least over time frames relevant to most investors. It’s simply not a reliable hedge for inflation, and investors holding it for this purpose may well end up disappointed.

As a way to protect against crashes, gold has done relatively well. While it’s not a perfect hedge (it doesn’t always go up when stocks go down), gold has performed reasonably well as crash insurance in the past. It might not have always gone up when equity markets fell, but it almost always fell less than the market.

As a source of uncorrelated returns, gold does have the ability to improve a portfolio’s risk adjusted return metrics, but only if an investor has the stomach to hold it through long, painful periods of underperformance.

The numbers say that gold could have a place in portfolios, but it’s not for the faint of heart. If you’re willing and able to include a relatively significant allocation to gold in your portfolio and forget about it for the next 40 years, then it could serve a useful purpose when held alongside equities. It could be particularly valuable in times of crisis, and at reducing portfolio volatility.

If, however, you’re someone who focuses on individual positions (over portfolio returns), or are prone to reassessing your asset allocation every few years, then gold is likely not for you.

What do I think?

Statistical analysis and back testing can be helpful, but it’s also useful to step back and think about the bigger picture. The primary purpose for the non-equity portion of a portfolio is (in my opinion) to help protect the equity portion when markets crash, and to help reduce portfolio volatility. The ultimate aim is to ensure you can stick with the highest allocation to equities possible, and hold them for the long term.

Equities are the driver of returns, everything else is there to protect them.

The biggest threat to equities is the Black Swan. The unforecastable tail risk. The unknown unknown. The event with a low probability but high impact.

Will gold help mitigate the effects of future Black Swans? I don’t know. Any form of certainty would require clairvoyance. However, it seems like a reasonable candidate given its track record, and I certainly couldn’t blame someone for including it for this purpose.

A part of me is still hesitant about the fact that 45% of gold’s demand comes from investment. I’m not sure how comfortable I am holding something which derives almost half its value from the fact that everyone believes everyone else believes it’s a useful asset to hold. It seems a bit Emperor’s New Clothesy to me.

But can I see the consensus on gold changing? Not really. People have loved gold for thousands of years, and I imagine people will continue to love it for thousands more.

Ultimately, the best portfolio in the universe will do no good if you sell it at the first sign of trouble, so it’s important to invest in something you truly believe in and can stick with in both good times and bad.

Personally, I’m not sure my conviction in gold is high enough to:

  1. Allocate enough of my portfolio to it that it starts having a meaningful impact on performance, or
  2. Stick with it through its inevitable periods of underperformance. I’m not sure the argument that gold is valuable because everyone else believes it’s valuable will be enough for me to keep holding it as I’m staring into the face of an 80% drawdown.

As we said at the start of this article, gold is the best Rorschach test investing has to offer. You can play with the data and come to whatever conclusion you want to support your opinion. This article alone has provided enough data to make a reasonable argument for both sides.

The decision to allocate to gold is, as with all investing decisions, a personal one. There’s only one person who knows whether they find gold’s merits attractive enough to outweigh its drawbacks, and that’s you.

The most important thing is to make a decision you’re able to stick with. If gold works for you, then great. If it doesn’t, then don’t force it!

Share on Facebook
Share on Twitter
Share on LinkedIn

Past performance does not guarantee future performance and the value of investments can fall as well as rise. The information on this site is provided for information only and does not constitute, and should not be construed as, investment advice or a recommendation to buy, sell, or otherwise transact in any investment including any products or services or an invitation, offer or solicitation to engage in any investment activity. Please refer to the full disclaimer on the disclaimer page.

Notify of

Oldest Most Voted
Inline Feedbacks
View all comments
February 3, 2024 3:01 pm

Outstanding article. Thank you.

August 18, 2022 2:06 pm

It would be interesting to know your thoughts on gold in a time of negative interest rates, as discussed here

August 8, 2022 1:00 pm

Thoroughly enjoyed this article – really insightful and offers a balanced argument for all the points made. We particularly enjoyed the section on gold acting as a stable investment during stock market crashes and reached a similar conclusion ourselves when investigating the relationship between gold prices and the stock market which your readers may find useful, it summarises why gold prices go up when the stock markets go down.

May 24, 2022 1:05 pm

An admirable and successsful stab at objectively weighing the arguements for and against.

I am firmly in the ‘no’ camp, having worked for 2 large investment managers I saw all too often tiny (2-4%) allocations to the metal for cosmetic and / or marketing purposes.

Such a small allocation is insufficient to protect to the downside or add much to the upside; I feel you need 0% or 20%.

May 9, 2021 6:40 pm

thanks for that – I really enjoyed the article.
The main interest I would have would be the Gold As Uncorrelated Asset.
In your examples of a split gold/stocks portfolio, are we to assume they are regularly balanced to maintain a given ration (eg 10/90 or 20/80)? As a regular investor, I would be interested in the idea of putting 10% of my regular investment into gold and the rest in equities and when there is a crash in stocks, selling the gold and using it to buy more stocks at a lower price. Then the monthly investments continue to build up the gold portion again. Has that been back-tested or described?

April 20, 2021 8:41 am

I didn’t think this was a fair and balanced article.

In particular, it confuses gold-as-an-investment-asset with gold-as-a-real-option. The latter is the one you want in the Zombie Apocalypse.

Even that scenario is rather derided in this article. Zombies don’t really exist. The personal Apocalypse happened within living memory (just).

Which currency do you think worked best for Jews trying to bribe their way out of Nazi deportation in Central and Eastern Europe in 1941-44? Having gold coins was a lot better than having one’s emergency fund in Polish or Soviet paper currency.

However, since this is primarily an investment-oriented article, you’d do better to delete all the jokes about gold bars not being much use in a personal crisis, since that aspect of gold’s utility isn’t relevant.

November 18, 2020 11:22 am

Very well written, Thanks Occam.

Think I’d land on the same closing remarks as you did. Gold, in my view, is more like a secondary deference behind government bond (US, Germany especially) when things go really bad.

Then the question is that it is worth holding 20% for the tail event (as you pointed out), OR forget about it/wait for portfolio recovery instead, unless the 20% can rise 4x more to protect the 80%, OR buying gold opportunistically when market has a crash (as repeatedly evidenced in 2008 and the march falls).

Perhaps a bit cynical here, but the exact link between gold price and currency value in a post gold-standard era isn’t that clear. Guess the safe heaven status is maintained for being able to borrow against for an insolvent state in a crisis, and no other metals can do that.

October 30, 2020 3:28 pm

Beautifully written and explained. Well done.

I just wonder if the analysis is right given the unusual current environment. It does beg the question why highly successful investors invest in gold like Buffett (Barrick Gold for the first time), Dalio (around 15% allocation in his hedge fund), Sokoloff, Pal etc. Maybe consider how the debt cycle (from 1981 stocks bloomed as interest rates have gone to zero and debt in all its forms is at record highs) have given the illusion of stock market outperformance will continue for decades to come. The UK stock market was cyclical between 1700-1950, despite having the largest Empire the world has ever seen. 😉

A diversified portfolio including hard assets (think 1970s Keynesian policies on steriods with MMT), crypto, stocks and bonds seems prudent given currency debasement, ongoing fiscal stimulus for years and central banks starting to embrace crypto as they try to move away from dollar dependency.


October 24, 2020 9:20 am

Really insightful as always! Thank you always look forward to your posts.

October 21, 2020 8:12 pm

Great post mate, really enjoyed the points you’ve laid out above.

I’ve recently starting investing in gold alongside equities (through bullionvault and not ETFs) at around 10% of my monthly investment. Your analysis and conclusions have given me a lot of food for thought.

Im planning on holding gold for the long term as a hedge against crashes (encouraging that your data above draws similar conclusions), however, is 10% enough? Now i’m not so sure.

It would be interested to get your view on golds position in the coming decades. (I know, how long is a piece of string..) However, as you noted above in regard to inflation, can you see a scenario where gold could actually be an inflation hedge due to the significant amount of QE that is occurring and perhaps due to continue by govs in the coming years?

Cheers and looking forward to your next post.