Inigo Fraser Jenkins — Why Gold Is Money Again: Rethink the Traditional Portfolio

1:12 Split the question: US equity exceptionalism is not dollar exceptionalism

• "I would like to take the view — I'm not sure whether it's popular or not — but to defend US equity exceptionalism but to decline to defend dollar exceptionalism."
• Most of the debate conflates the two. Treated separately they resolve in opposite directions, which is why the portfolio conclusion is an equity overweight and a currency hedge rather than a single directional "US" view.

1:47 Why AI accrues disproportionately to US firms

• Whatever the productivity gain from AI turns out to be, "it's likely that the US benefits more from that."
• Two reasons, one flattering and one not: a long history of US firms being better at exploiting IT advantages than rivals elsewhere; and US firms being able to fire people faster, since the early stages of an AI productivity gain involve rationalizing the labor force.

2:19 The slow-moving support: demographics

• The projected US working-age population is basically flat over the next 10–15 years — much worse than the last 40 years in absolute terms, but better than everyone else's.
• Europe: roughly −0.5% per annum. China: close to −1% per annum. "A real difference in base growth rates coming from demographics."

3:13 Profit share of GDP — and the mean-reversion call he abandoned

• The "extraordinary growth in profit share of GDP by US firms through administrations of both colors"; the effective tax rate is almost a monotonic line down over decades, and there is no appetite to reverse it.
• Later, explicitly 10:37: "I've given up on making that mean reversion call on profit margins or on profit share of GDP in the US." Social limits exist but "we don't seem to have reached those levels yet."
• He is candid about what that resembles — "a little bit like regulatory capture" — but treats it as a durable earnings advantage US firms hold over the rest of the world.

4:22 The case against the dollar: debt service has crossed defense

• Public debt/GDP by itself is not a limit — "Japan showed us that we can go way through that number and still function."
• The more pressing metric is the share of government expenditure spent on debt service versus things like defense — "the fact these two have crossed over is a big deal."
• Second angle: geopolitics. The weaponization of the dollar after the Russian invasion of Ukraine gave BRICS nations an incentive to de-dollarize — though "I think they can't actually de-dollarize. China can't make its currency convertible for all kinds of reasons."

5:43 The case for the dollar: no alternative, and stablecoins are a new buyer

• "There is absolutely no alternative to the dollar," and past reserve-currency transitions took decades — "we shouldn't hold our breath."
• A genuinely pro-dollar force: "the emergence of stablecoins has created a buyer of short maturity US debt that is very material."
• The resolution: "it's less a story of expecting depreciation against other currencies… the depreciation story is really a depreciation story against gold." The dollar's problem is that it has become more risky, not that it is going down against the euro.

6:40 The two portfolio conclusions

• "It pays to be strategically overweight US equities in a global equity portfolio."
• "The consequences for the dollar are more for non-dollar investors, at least to hedge more of the dollar exposure." The dollar view is a hedging-ratio decision, not an equity-allocation decision.

8:31 Concentration is uncomfortable — but diversification is the scarce good

• Two concentrations stacked: the top 10 stocks inside the US market, and the US inside the world index (roughly 60%–two-thirds of world market cap by Faber's framing).
• The real issue "runs through the very heart of portfolio construction today": it is hard to find diversification, and "harder to come by than it has been in the last 20 or 30 years."
• The price of chasing real return is explicit: "you probably have to have some more illiquid assets… be happy with higher overall vol levels, happy with more concentration."

12:24 Valuation: forward P/E is fine, Shiller PE says fully valued — and neither is a timing tool

• On forward P/E the market "does not look very extreme," but that just "shunts the problem on to what the earnings forecasts are."
• For strategic horizons he prefers the Shiller PE (price / 10-year inflation-adjusted average earnings), on which "the market looks fully valued" — but so is nearly everything else: "most assets are relatively expensive compared to history."
• The discipline: valuation caps your return forecast, it does not generate a sell signal. "If you use valuation compared to history it would have told you to be out of stocks for a long time and that would have been an absolutely terrible call." The one thing he refuses to forecast is further multiple expansion 13:55.

14:33 Relative valuation has not worked for regional calls

• "I don't think there's really good evidence that relative valuation has been a particularly useful guide to medium-term relative regional performance for a long time. And I'm not sure why that should start now."
• So anyone forecasting European outperformance has to argue European earnings growth beats US earnings growth — "I just don't see a basis for making that kind of forecast." Cheapness alone is not an argument.

15:19 Long-duration government bonds will not diversify equities

• Asked whether 4%-yielding bonds are the diversifier again: "No, I have a strong view on this… it's likely they will not perform the diversifying role that they've performed historically."
• The empirical core 16:07: the negative stock/bond correlation of the last ~20 years made 60/40 "a no-brainer" — but "if you extend the chart 200 years prior to the last 20 years that correlation was positive almost all the time." The post-2022 positive correlation "actually looks more normal."
• The number that matters for sizing 18:16: bonds still help at a +0.2 100-year average correlation — "it's just not the no-brainer it was at a minus 0.4."

16:35 The narrative he most wants investors to absorb: the era since the mid-80s was the anomaly

• Benign inflation, bond yields starting from an extremely high level, and strong labor-force growth from demographics and globalization — producing "very strong positive real returns, plentiful diversification."
• "That is something that's actually very unusual in the longer scheme of history" — and the forces behind it have "either run their course or are even going into reverse right now."
• The forward mix 17:04: deglobalization, high starting debt levels, the temptation of debt monetization, and climate push inflation up without pushing growth up — "if anything, a force downward on growth at the same time." That inversion is precisely what flips the equity/bond relationship.

18:44 Two more structural drags on bonds — plus "there is no such thing as a risk-free asset"

• The global pension system is moving DB to DC, and the biggest shift inherent in that is less demand for long-duration nominal assets and a bigger need for real assets.
• "I've long held the view that there is absolutely no such thing as a risk-free asset. People just use that term partly because it makes the maths easier… It's contingent on political and economic states of the world, and I don't think we're in those states anymore."
• Bonds keep the roles they can still play — liquidity, drawdown mitigation, cash-flow matching — but not the diversification role. Conclusion 19:52: "60/40 is in no way a passive default asset allocation strategy in that kind of environment."

20:25 Survivorship bias: rank the markets by 1899 market cap and look at what happened next

• The AB chart Faber cites is titled "false sense of security from markets that worked": Japan, Austria, Hungary, Russia, France, Germany.
• "If you rank markets by market cap in 1899, then obviously US works spectacularly well and the UK worked reasonably well too. The next six or seven went to zero, sometimes more than once."
• The implication is a haircut on the input, not a forecast of ruin: "there is this inherent survivorship bias in the way people think about the long run return from a passive cap-weighted index. It's not perhaps quite as high as you think it is."

21:04 Gold: universal interest, barbelled ownership — and he is overweight

• Gold comes up in "at least 90% if not pretty much close to 100% of client meetings" — but actual ownership "almost exactly maps on to where they sit in the financial ecosystem": family offices and sovereign wealth funds own some; pension plans mostly do not, constrained by targets and governance.
• "We've been overweight gold for some time" and are "not too worried from a strategic standpoint about the selloff that happened this year, although it's still up this year."
• Ownership is a barbell and the tail is thin: one recent survey's modal answer was zero 29:22. The honest objections he grants: no income, no price target, unclear bucket, and returns that are "very episodic — you can go for decades with no return."

22:29 The gold thesis is a correlation thesis: zero to equities at any inflation level

• "The thing that grounds a gold view is that I would like to strongly defend the idea that the correlation of gold and equities is zero and remains zero at any level of inflation you care to mention. That is clearly not true of bonds."
• The normative argument 26:40: gold "has no cash flows attached to it, hasn't really got a use — frankly, there's no reason to think there should be a non-zero correlation."
• He is careful about the size: "I'm not suggesting people can replace the 40 in 60/40 with a gold position." It is one component of a rebuilt diversifying sleeve.
• Contrast within commodities 24:56: copper is positively correlated to equities (procyclical); oil is different again (high prices coincide with equity shocks); gold sits near zero across inflation buckets.

23:25 No price target — so price gold by a long-run real return instead

• "There used to be an ability to price gold. You could price it off TIPS. That broke down on the day the Russians invaded Ukraine and I don't think that comes back."
• The substitute: a long-run return target. The 150-year real return on gold is about 0.6% per annum, "albeit very episodic"; add BRICS/China official buying — "I've got no idea how much gold and when they buy it, but it is enough to lift that return assumption. Let's call it 1% real."
• "So we've got 1% real, zero correlation to equities. That at least is the starting position for thinking about what a portfolio allocation to gold might look like." He notes the oddity openly — "here I am as an investment strategist telling people to buy an asset for which I decline to give a price target."

25:47 Why gold sold off in H1: a flow-driven correlation that was never supposed to be there

• Through to about January this year gold and equities showed a positive correlation with "strong inflows into both assets at the same time."
• "That's part of the clue as to why gold sold off so aggressively in the first half of this year" — the extra investor flow, plus CTAs chasing a price that "went up in like a straight line," then unwound.
• The read is diagnostic rather than bearish: the positive correlation was "abnormal… I wouldn't expect it to normally continue," so the unwind restores the property the allocation was bought for.

30:17 Gold is not a standalone view — it exists because the equity overweight needs a hedge

• "I'm not positive on gold in isolation. I'm positive on gold because I think people should have a strategic overweight on equities and then… we are struggling as an industry to articulate what on earth diversifies that equity position in a world where bonds no longer do it."
• This is the load-bearing logic of the whole episode: the equity call creates the diversification problem, and gold is one of the things that goes into that bucket. Kill the equity overweight and the gold case changes shape.

31:40 The "non-fiat" allocation: gold-dominated, with small Bitcoin and silver

• "As part of a non-fiat allocation that we suggest to people strategically, I think Bitcoin should be a part of that. Now a small part, because a non-fiat exposure should be dominated by gold."
• He changed this view publicly over COVID, "having previously thought it had absolutely no role whatsoever in asset allocation" — the case is explicitly "on the coattails of gold," i.e. anything plausibly gold-like earns a look.
• Silver gets a small slot on the same logic — "not because I'm fundamentally bullish on silver necessarily, but just it is different because investors play a much smaller role in that market" proportionally than in gold. Potential Bitcoin catalyst: more regulatory and custody clarity bringing new investors in.
• Client interest is currently thin — Bitcoin "ebbs and flows with the price," and Faber relays Jeffrey Gundlach's line that "Bitcoin is not cool anymore, it's only for boomers" 31:18.

27:42 Real versus nominal return targets — the change he thinks is still ahead

• "I am amazed at the number of people I come across who do have nominal return targets… most people don't, or at least shouldn't."
• The end investor is funding real-world liabilities — "retirement costs, healthcare costs" — so the target should be real. The drift to nominal is a career artifact: "people haven't had to worry about higher equilibrium inflation for a very long time."
• The symptom is benchmark choice: people try to beat the MSCI World index or the bond aggregate "rather than trying to outperform inflation." He expects more investors to be forced onto real targets — "a change that hasn't yet happened."

33:10 Inflation: higher equilibrium, not runaway — high twos to 3%

• The five forces Faber lists back to him — AI, demographics, climate, debt, deglobalization — all route to inflation.
• "It would be a brave economist who pointed to all these forces and simply put a coefficient on them and added them up, because we've never been here before. But it's enough to say that we are plausibly in a different inflation regime."
• The number: "I wouldn't want to forecast unanchored inflation. Far from it. But simply higher equilibrium inflation. Let's call it high twos, 3%" — with automation as the offsetting disinflationary force.

34:15 The 4% kink: above it, equities stop behaving like a real asset

• "Beyond four becomes a problem. For all kinds of reasons, not least socially. In portfolios it becomes a problem for a very specific reason, which is much higher than four and equities stop behaving like a real asset."
• "Above four… your portfolio that protects against inflation doesn't want to have bonds or equities in it, and it's a lot more complicated."
• Why his ~3% forecast matters so much: it "still leaves us in shooting distance of portfolios that we recognize." The 4% line is the boundary of the whole framework, and it is the number to monitor.

35:34 Healthcare — the one sector call, and it sits at a four-way intersection

• Healthcare "sits at the nexus of a few of these issues: the need for diversification, demographics, AI, and then valuation" — the defensive ballast inside the pro-equity view.
• Demographics support the sector and, specifically, its pricing power: "care costs in particular tend to be quite sticky in terms of prices." And it is "a very plausible case that it's a beneficiary of AI."
• Valuation 36:34: the sector's P/E relative to the market is low compared with its 20–30 year trading range, even after outperforming. The reason is policy uncertainty — and that is the point: "uncertainty about health policy is not particularly related to the riskiness of the AI trade. So it's a different kind of risk." The discount is an independent risk factor, which is exactly what a concentrated AI-heavy portfolio needs.

37:09 Energy and commodities: real return, inflation protection — and gold pulled out of the bucket

• Energy is now about 3% of the US equity portfolio — "a rounding error… it's less than gold I think at this point."
• Exposure "through commodities directly and through the equities linked to those commodities," and the energy sector "would stand out" for an investor focused on income and free cash flow.
• The reframing 38:33: "because of geopolitics and because of outstanding debt positions of major economies, gold is actually no longer a commodity. Gold is money in this kind of environment" — so gold goes in its own bucket, and the rest of commodities (a mix of base metals and energy) carry the inflation-protection and real-return job.

39:26 Dr. Copper at highs and nobody is talking about it — because the signal changed jobs

• Faber notes copper near all-time highs around $7 with almost no headlines; Fraser Jenkins agrees he doesn't hear it either. Copper "historically had a very clear link to the business cycle, but that business cycle regularity is… not talked about much either these days."
• The structural demand is real — "AI physical capex buildout and energy transition buildout" — but that is not his interesting takeaway.
• The strategic point 40:19: "we talk a lot about the level of inflation… but we also need to consider the volatility of inflation," which he expects to be higher — and "base metals form at least part of the potential response to managing that in a portfolio." Copper is repurposed from growth signal to inflation-vol hedge.

41:20 Why inflation volatility rises: two narratives crashing into each other

• Deglobalization "crimps the ability to cushion price shocks in the global economy" — the shock absorber is being removed.
• The collision: "AI's reawakening of serious demand for physical capex" coinciding with "limits on the US's ability or willingness to play a policeman role." Surging physical demand meeting a less-policed supply chain implies "we should expect more supply shocks."
• And the shock rotates: "this year oil is the topic, but in future years it could be cobalt or lithium or copper or something else" — which argues for a broad basket rather than a single-commodity bet.

43:05 AI productivity: don't forecast it — reverse-engineer the bar it has to clear

• "I'm relatively skeptical of anyone's ability really to make forecasts of long run aggregate productivity growth" — the TMT bubble in 2000 is the cautionary case, where permanent-productivity assumptions had to be unwound.
• So the research inverts the problem 43:50: total up the forces that depress growth, then ask "is it plausible that AI can make up for this."
• The drags: demographics, worth about −0.8% per annum versus the post-1980 trend for the US working-age population 44:14; and climate — "highly unlikely the world hits net zero by 2050," so a more-than-two-degree increase, worth "a few tens of bps" but unforecastable because of nonlinearities. Total: roughly 1% per annum lower real growth ten years out.

45:24 The steam-engine speed limit, and the answer: AI runs us in place

• Historical benchmark: the steam engine raised UK productivity by about 0.8% per annum in a sustained way. "If AI is as good as the steam engine then it can do that" — but that is "an interesting little speed limit," and anyone forecasting materially higher is claiming AI beats the steam engine.
• The academic cross-check 46:08: a "cottage industry" of productivity forecasts now averages "about 1% per annum. A huge range around it. Massive disagreement," but the central tendency clusters there.
• The conclusion, and the episode's cold-open line: "AI presumably does raise productivity, but the central case is that it just keeps us running at the growth rates that we've seen in recent decades — not an extra uplift to growth."

46:34 Labour: the optimistic productivity case implicitly assumes job losses

• AI raises productivity two ways — "enhance a unit of labor, make it more productive, or automate a role… and get rid of that unit of labor." Anyone with a more optimistic near-term productivity number "implicitly has to make some assumptions around a negative impact on the labor market."
• The standard rebuttal, fairly stated 47:31: 200 years of automation scares, and "more jobs always got created than got destroyed," with close to full employment throughout.
• Why this time may differ — and it is a genuinely novel screen: rank the sectors where AI is expected to make the biggest difference ("with our sector being number one on the list" — financial services) and overlay unionization rates. "It looks starkly different from attempts at automation in heavy industries, auto industries etc. in the last 20 or 30 years." Prior automation hit unionized sectors with organized resistance; AI hits white-collar sectors with none. "So it implies in the near term at least… some level of job dislocation."

48:32 What he is working on next: soft commodities

• The next research topic is the "broader issue around commodities," and specifically the one leg he has not covered: soft commodities.
• Drivers named: El Niño and fertilizer exports, which "point to potentially higher food prices."
• "Tends to be an asset we don't spend a huge amount of time talking about normally, but I think that's probably going to get extra airtime over the next 6 months or so." Research is published on the AllianceBernstein site and posted to LinkedIn.