Everything below started with one image:
Portfolio Charts posted this global allocation and it stopped me in my tracks. The allocation is just a thing of beauty. That one graphic is the reason this post exists, but it’s really an excuse to talk about something bigger: Portfolio Charts just launched their Global Explorer, a tool that tests asset allocations, and specifically safe withdrawal rates, from the perspective of investors living in different countries. Not just the US.
If there’s ever a Hall of Fame for personal finance websites, Portfolio Charts gets a first-ballot induction. The data depth paired with the visuals is a rare combination. And the part I think gets overlooked…maybe just by me? …is the Insights section. It’s less a chart library and more a guide on how to actually use the charts, which is exactly the kind of thing that turns a good tool into a great one.
What really sets Portfolio Charts apart, though, is its push to pull global insights out of the dataset. For as long as I can remember, personal finance content has assumed everyone reading it lives in America. Plenty of the underlying concepts travel well. But assuming a Swiss investor and an American investor face the same financial reality is a mistake. They don’t.
Then there’s survivorship bias, which nobody wants to talk about because the story it tells is depressing. The US has had an incredible run. That doesn’t mean it continues forever. Other countries have dominated global markets before and lost that spot. Anyone with heavy US-stock exposure should at least entertain the possibility of a “1990s Japan” outcome instead of another “1980s America.”
There’s a fairly simple way to hedge against that uncertainty: think like a market-cap-weighted investor. That approach would have overweighted Japan in the late 80s (hint: not great). It also guarantees you’re exposed to whatever market actually matters at any given time, without requiring you to predict which country wins the next decade. Nobody’s that good.
This is where Portfolio Charts does something almost nobody else does. It lets you answer a question like: how would a Swiss investor have done over the last 50 years holding a global market portfolio with different asset classes? Or an Italian investor? It’s genuinely strange to me that almost no platform in mainstream finance research tackles this, given that “global exposure” is supposed to be table stakes. Most research still stops at a simple domestic-versus-international split. On Portfolio Charts, you can at least dial in a market-cap-style US weighting regardless of which country you’re testing from. The catch is that those weights aren’t static: Japan’s share of global markets 40 years ago looked nothing like it does today and the tool can’t fully backtest that shifting weight. Maybe someday.
Currency does matter for returns, especially if you’re not a US dollar investor, but this is also where people get tripped up. A global portfolio hedged into a local currency and a stock index of large companies denominated in that currency are two completely different things: apples and oranges, not apples and apples.
I showed the two lines to my four-year-old and even he could tell they weren’t the same. The point isn’t that one line beats the other. It’s that they’re answering different questions entirely, and mixing them up leads to the wrong conclusion.
I made a version of this point when I reviewed the Cederburg paper a while back, and Portfolio Charts gets at it too: it’s unrealistic to model a Canadian investor holding a massive allocation to Canadian stocks. Nobody serious actually does that… Oh sorry, anyone that doesn’t listen to Ben Felix.
Which gets at the real issue with home bias. People conflate it with currency risk, but they’re not the same thing. Loading up on your local market has essentially no upside and a lot of downside (unless, by chance, you pick the winner of that century). Look at what Portfolio Charts highlights about Spain under the Franco dictatorship. A market-cap-weighted stock allocation paired with a sensible mix of AAA global bonds would have spared investors in Spain, Italy, Austria, Russia, and China (among others) from the kind of total losses that a home-country-heavy portfolio delivered at some point in history.
The one argument for home bias that seems to survive the currency conversation is the idea that you want your portfolio “correlated” with the economy you actually live in. It’s a nice theory. The problem is the stock market isn’t the economy. Ask a Danish person who didn’t happen to work at Novo Nordisk during the years the stock was flying. A big personal allocation to Novo might have paid for a slightly nicer vacation that summer, assuming they sold at the right time (why would they?) but it wouldn’t have moved the needle on their rent. Play that game and you’re benchmarking yourself against the guy who worked there and got a bonus two orders of magnitude bigger than any stock appreciation you saw.
I live in Switzerland. I understand the pull to keep pace with a functioning, expensive economy. But loading up on local stocks isn’t the answer, especially once you actually look at the downside: the genuinely ugly tail scenarios if things go wrong (hello Argentina. Hi Turkey).
This is where I think the Global Explorer earns its keep: as a way to stress-test portfolio ideas from angles that are hard to reach with something like testfol.io. Testfol.io lets you switch currencies now, which is an improvement, but if you actually live in a given country and want to model your domestic bond sleeve, you have to manually swap the ticker every time you change countries in the simulation. Often the specific ETF doesn’t exist on the platform. Sometimes it doesn’t exist at all.
Portfolio Charts takes the opposite approach, it’s ETF-agnostic, which is occasionally annoying and occasionally exactly what you need. It also applies local inflation to each simulation, which matters more than people give it credit for. Inflation is one of the biggest risk drivers in any portfolio, and testing real returns — which also reveals how different assets actually correlate with inflation — tells you far more than nominal returns ever will.
I’d expect two objections from anyone skeptical of PC analysis: the dataset only goes back to 1970, and it’s a historical backtest rather than a Monte Carlo simulation. Both are fair points, and I have two responses. First, that 1970 starting point captures all four market “quadrants”, a full cycle of conditions, which matters more to me than tacking on data from a less globalized, less interconnected era that predates the end of the gold standard. Second, the country-by-country breadth is doing a similar job to what a Monte Carlo simulation would do, just along a different axis. It’s not a full substitute, but it’s a genuinely useful addition.
Even with those caveats, the tool produces some conclusions worth sitting with. Look at any single country and you won’t find a top-100 portfolio that skips any of the three major asset classes, meaning stocks, bonds, and commodities/gold. Bonds are actually the least represented of the three if you look at the full sample. What surprised me was how much commodities mattered. Removing them doesn’t wreck the results, but the benefit of holding them is obvious once you see it side by side. What didn’t surprise me at all: removing cash (T-bills) barely moves the needle for the best portfolios.
One of the most useful things Portfolio Charts highlights is gold’s split personality. It’s a core ingredient in the best-performing portfolios, and it’s also the dominant holding in the worst ones. Regular readers of the site have probably heard this before, but it’s worth repeating: gold as a standalone bet is one of the worst things you can put in a portfolio. Gold as a complement to stocks and bonds is excellent.
I know how easily the general public falls for single-line-item thinking, which is exactly why a tool like this matters: it’s one of the better ways to show investors how their own brains mislead them. My advice is the same as Portfolio Charts’: spend real time playing with it. The goal isn’t to find the “best” portfolio or shave three basis points off some backtest. Even if you found it, you’d never actually see it play out, because the tool leads to a pursuit of optimisation across hundreds of possible future scenarios. And you only get to live one of them. Someone, somewhere, will always get lucky picking the right asset for a given decade and go around acting like it proves how smart they are. It happens constantly with single stocks too! A diversified portfolio will never win that particular argument. The best you can do is finish the race in a great shape… and maybe glance behind you at how many people didn’t.
What I am reading now:
Follow me on Bluesky @nprotasoni.bsky.social
0 Comments