December 10, 2025 – Welcome to another post on the ERN blog. This is the third installment in the “How to Lie with Personal Finance” series (please also check out Parts 1 and 2). As always, this is not an instruction manual for deception, but precisely the opposite: it points out the misunderstandings circulating in personal finance. Think of it as an homage to the classic book “How to Lie with Statistics.” On the program today are the lies and misunderstandings surrounding diversification. Don’t get me wrong, I worked in finance, math, and statistics long enough to appreciate the beauty of diversification. But diversification seems to be one of the more misunderstood and misrepresented concepts in the personal finance world. I want to highlight some of those misunderstandings in today’s post.
Let’s get started…
Lie 1: More Stocks = More Diversification
The idea of diversification is that if we add more uncorrelated, or at least low-correlation assets to a portfolio, we can diversify away a large portion of the risk inherent in those individual assets. Some rules of thumb circle around on the interwebs, such as, “You need at least thirty individual stocks to have a sufficiently diversified portfolio.” And there is some truth to that. But does this insight extend to the number of stocks in your ETF? Does that mean that the ITOT (about 2,500 holdings) or VTSAX/VTI (3,500+ holdings) are more diversified than their smaller cousins, IVV or VOO (about 500 holdings each)? Would VTSAX have materially lower risk than IVV and VOO? That’s undoubtedly the “lie” you hear from certain corners of the personal finance world, and why some folks have an obsession with VTSAX.
It turns out that more stocks in your ETF don’t necessarily lead to more diversification. For example, I looked at the returns of three iShares ETFs: OEF, IVV, and ITOT, which replicate the S&P 100, S&P 500, and a U.S. Total Market index, respectively. The standard deviation for the most recent 120-month window (11/2015-10/2025) is very similar across the three ETFs. In fact, the Total Market fund had the highest standard deviation. The 500-share ETF had the lowest standard deviation, the least dramatic worst-case calendar year return, and the lowest drawdown among the three funds.
Return stats for three equity index ETFs: 11/2015-10/2025 (120 months). Source: Portfolio Visualizer.
Similar results hold when comparing Vanguard offerings, for example, VOO vs. VTI. VTI tends to have a marginally higher standard deviation. The worst calendar year and worst top-to-trough drawdown were also slightly worse in the VTI:
Return stats for two Vanguard equity index ETFs: 11/2015-10/2025 (120 months). Source: Portfolio Visualizer.
Why doesn’t the risk go down as we add more stocks? Here are the main reasons:
- Most indexes are weighted by market capitalization, so there is significant overlap among them. For example, the 100 largest constituents in the S&P 500 account for more than 70% of the index. And the S&P 500 accounts for almost 90% of the U.S. Total Stock Market. So, we can’t just assume that we’ve expanded the universe by 5x when we move from 500 to 2,500 stocks. We’ve really only increased the breadth of our portfolio by about 10%, not multiples.
- Even the non-overlapping portions are highly correlated. In other words, most mid- and small-cap stocks, i.e., part of the total market but not part of the large-cap index, are still highly correlated with the S&P 500. For example, the iShares Mid-Cap ETF (MDY) has an S&P 500 correlation of 0.956. Intriguingly, small-cap stocks, as represented by the iShares Small-Cap ETF (IWM), correlate even more closely with the S&P 500 (0.982) than mid-cap stocks do. (All correlations are for 10/2015-09/2025. Source: my Google Sheet, see tab “ETF Factor Exposures.”)
- And a technical side note: even in the best possible case, where we had access to N uncorrelated return streams with the identical return distributions, the equal-weighted portfolio standard deviation only declines in the square root of N. So, if we increase the number of assets by 5x, the portfolio variance goes down by 80%, so the standard deviation declines only by a factor of the square root of 5. Thus, even in this ideal case, 5x assets reduce risk by only 55% not 80%.
For the technically inclined, the reason there are limits to diversification is as follows. Imagine we represent every stock i’s return through a simple factor model:
R(i,t) = alpha(i) + beta(i) x MarketReturn(t) + IdSyncRisk(i,t), i=1,…,N and t=1,…,T
Any portfolio of stocks i=1,…,N will always maintain its market beta (=market exposure) equal to the weighted average of the underlying stock betas. You might reduce idiosyncratic risk by broadening the index. Still, you might also add more volatility through the backdoor: the smaller stocks you add may have higher betas and/or higher idiosyncratic risk and/or higher exposure to other common factors (like Fama-French’s SMB, HML, etc.) than the large blue-chip stocks. This explains why U.S. Large Cap ETFs seem to have the lowest risk sweet spot, and if you keep adding more exotic small-cap stocks, risk will start rising again, albeit by tiny percentages.
For the record, I hold many different U.S. equity funds. I have two very large retirement plans from former employers. One has the large-cap index as its default, low-expense-ratio offering. The other plan has the total market index as the best (=cheapest) equity option. So, while I’m slightly biased towards the S&P 500 index funds because that’s the bellwether US equity benchmark with slightly lower volatility, I will gladly hold the Total Market index if a large-cap fund is not available or is only available at a higher expense ratio.
Lie 2: More ETFs = More Diversification
There are now more ETFs than publicly traded companies in the US. So, lots of gullible people might conclude that we should have expanded diversification so much that we no longer need to worry about risk. Specifically, I frequently hear this personal finance lie that adding more ETFs will give you much better diversification. Often, that lie comes in the form of someone commenting, “Karsten, you use U.S. Large Cap index equity returns in your withdrawal simulations, but you could achieve much better diversification and higher safe withdrawal rates by properly diversifying the equity risk.” I find this argument highly dubious, though.
The most recent prominent advocate of this idea has been Bill Bengen with his new 5.5% safe withdrawal rate. About half of the gain in his purported new safe withdrawal rate is coming from a more diversified portfolio (his claim, not mine, more on that later)—the other half from shifting the goalpost, i.e., being happy with a less secure success criterion. See my brief post on ChooseFI on this topic.
In any case, to achieve more diversification, Bengen splits his proposed 55% equity portion into five equal parts: U.S. Large-Cap, U.S. Mid-Cap, U.S. Small-Cap, U.S. Micro-Cap, and international stocks. The remaining assets are in an intermediate Treasury Bond fund, e.g., iShares IEF (40%), and 5% in T-bills, e.g., an ETF like BIL or SGOV. So, how much diversification would we have achieved during the last 10 years? None. Quite the opposite, the more diversified ETF portfolio would have had a higher standard deviation and a deeper drawdown than the simple portfolio with 55% S&P 500 index stocks.
Return stats for a simple vs. “diversified” (or di-worse-fied?) ETF portfolio: 11/2015-10/2025 (120 months). Source: Portfolio Visualizer.
We can also go back further than 10 years and observe the same pattern. Small-cap stocks and the mid-cap and micro-cap, by extension, will increase your observed volatility. So, if small-cap stocks don’t really lower your portfolio risk, why would Bengen find that small-cap stocks are the panacea to Sequence Risk? Very simply, small-cap stocks outperformed significantly between the 1920s and early 1980s, thereby alleviating much of the Sequence Risk headache in safe withdrawal simulations for some of the worst historical retirement cohorts (1920s, 1960s/70s). For example, the outperformance of small-cap stocks (and, by extension, mid-cap and micro-cap) is evident in the Fama-French dataset; please see the chart below. Between 1926 and about 1982, small-cap stocks outperformed the overall market by over 200%.
The Fama-French SMB (Small-Cap) factor significantly outperformed between 1926 and 1982. But then, it added zero alpha for the last 40+ years.
Also evident in this chart is that small-cap stocks did not offer any diversification during the Great Depression or the 1970s recessions. Quite the opposite: the Fama-French SMB factor normally declined during significant US recessions, and most of the starkly upward-sloping portions occurred during economic expansions, when the overall stock market also did well.
So, small-cap stocks offered worse risk characteristics, but significant alpha on average. Unfortunately, this alpha has since disappeared. The SMB factor has been flat over the last 45 years because the efficient market has arbitraged away this free-money spring, ever since academics pointed it out to the general public.
So, Bengen’s diversification narrative is a total myth. The diversification story is neither true in the first half of the sample, where SMB was all stock-picking alpha (at the cost of more, not less volatility), nor in the second half where SMB was basically zero extra return plus more risk, nor would any sane person extrapolate the SMB alpha from 40+ years ago into the future, now that the factor has been flat for so long.
Lie 3: Low Correlation = Better Diversification.
Imagine we are looking to add other assets to our portfolio to diversify. What assets would be good candidates? Is there a magic number or indicator we should look for? The correlation between the new asset and the existing portfolio would be a good indicator. The lower the correlation, the better the diversification, right? Not necessarily. Whether adding a new asset lowers your portfolio’s volatility depends on two variables: the new asset’s volatility and the portfolio’s correlation with it. Let’s look at the following numerical example: We start with a portfolio that has a 15% standard deviation. Which of the three assets would be best at lowering our portfolio volatility?
- ETF 1 with a correlation of 0.7 and a standard deviation of 25%
- ETF 2 with a correlation of 0.8 and a standard deviation of 20%
- ETF 3 with a correlation of 0.9 and a standard deviation of 15%
Let’s plot the standard deviations for different asset mixes between the old portfolio and new ETFs; see the chart below. Although it has the highest correlation with the initial portfolio, the third ETF still offers significant diversification. The blue line slopes down and finds a minimum when we add 50% of the new ETF. In contrast, the lower-correlation ETFs are all sloping up, so there is no positive allocation to the new ETFs that lowers the portfolio volatility.
Volatility when adding a new ETF to a portfolio: Lower correlation may not always imply more diversification.
The explanation: I provided the mathematical derivation for this phenomenon last year in my Di-WORSE-fication post, see the screenshot below. The necessary and sufficient condition for the new asset to lower the portfolio variance is that its standard deviation multiplied by its correlation must be less than the existing portfolio’s standard deviation. That’s certainly true for the third ETF (0.9×15%=13.5%<15%), but not for the higher-volatility ETFs 1 and 2, e.g., 20%x0.8 and 25%x0.7 are all larger than 15%.
Many exotic equity ETFs, such as small-cap and small-cap value, have this weakness: their correlations with the S&P 500 remain relatively high, typically around 0.8 to 0.9. They also have much higher standard deviations, usually around 20-25% annualized, compared to only 15-16% in the S&P 500. That makes those ETFs unattractive from a pure risk-reduction perspective. And that’s not even mentioning the underperformance of small-cap value stocks over the last two decades.
Similar arithmetic often applies to individual stocks. For example, many stocks, even large blue-chip companies, have relatively low correlations with the S&P 500. Tesla had a 0.45 correlation with the S&P 500 recently (10/2015-09/2025). But a standard deviation of over 60%. So, since 0.45×60% is much larger than the S&P 500 standard deviation, you would have increased your portfolio volatility. But of course, TSLA would have improved your average returns.
Lie 4: Diversification helps during Bear Markets
When I worked in asset management, we’d often use the phrase “When the sh!t hits the fan, all correlations go to one.” Meaning: I might have return sources that appear uncorrelated, or at least only mildly correlated, during regular times. But during periods of economic and financial stress, those ostensibly uncorrelated assets all feel the pain—one such example: international equities. For instance, in an average day or week, we may see the US index down by 1% and the non-US stock fund up by 1%. But how often do we see the US index down 20% while my non-US equity fund is up 20%? Never! In the chart below, I plot 12-month rolling returns, the S&P 500 on the x-axis, and the bellwether non-US stock index (MSCI World ex USA) on the y-axis. All returns are adjusted for US CPI inflation. Do you notice the funnel shape of returns in this scatterplot? If the US market is down significantly, so is the rest of the world, with very little variation around the 45-degree line. In fact, the standard error of World-ex-US minus the US return is the smallest, only 6.1%, when the US index is down the most (20% or more down). So, non-US returns look most similar to US returns during a bear market. We get the least diversification when we need it the most. And we get the most diversification when we need it the least.
S&P 500 vs. ACWIexUS returns (rolling 12m). Total returns adjusted for US CPI inflation.
That’s not to say that international diversification is useless. Even if international stocks have gone through the same (or even worse) bear markets as the US, they may recover faster, as they did most recently in the early 2000s after the Dot-Com bust. So there is a diversification benefit, but international stocks are no panacea to market volatility. Anybody who tells you that you can wave the magic wand with international stocks to diversify a US equity portfolio is lying to you. Also see my 2017 post on international diversification.
I should also stress that for investors outside the U.S., you absolutely want to diversify away your idiosyncratic country risk. But that’s much harder for U.S. investors because our business cycle affects most other economies around the world.
Lie 5: With enough Diversification, you don’t have to worry about Expected Returns.
Some people falsely claim that with the proper volatility control, you can easily raise your safe withdrawal rate. Bill Bengen at least kept a high enough equity exposure to give you a chance to make it through a 30-year retirement. Though 55% seems a bit low for the early retirement crowd with a 50+-year horizon.
But some charlatans want to go even further than the new Bill Bengen proposal and come up with more exotic portfolios – Risk Parity, Golden This, Golden That, etc. – involving even lower equity shares. But very often, shifting out of high-volatility/high-return assets like equities and into diversifying assets like bonds, cash, and commodities will reduce your expected returns. If you reduce your equity holdings to only 25% as in the Permanent Portfolio, or 30% as in the “All-Weather Portfolio,” you might have put the (volatility-)cart before the (return-)horse. Or you throw out the (return-)baby with the (risk-)bathwater. Or whatever figure of speech you prefer. In other words, if you replace too much of your risky but high-return assets with diversifying assets that have much lower expected returns, you might hurt your portfolio in the long term. In historical safe withdrawal rate simulations, both of these exotic portfolios underperformed a simple 75% stock + 25% bond portfolio, as I showed in Part 34 of my Safe Withdrawal Rate Series.
Admittedly, while I did find that commodities in general are unhelpful, gold is the one commodity that has shown intriguing diversification properties in most past bear markets, performing well in both deflationary and inflationary recessions. So, mixing in about 10-15% of gold in a retirement portfolio would have slightly increased your failsafe withdrawal rate. After the recent runup in gold prices, though, I’m a bit worried that valuations are stretched and this winning streak might soon end. As a compromise, you might want to pick a dynamic asset allocation strategy, like the Momentum strategy I wrote about last month.
Lie 6: You need Diversification while accumulating!
I always bring up my personal FIRE accumulation journey that included two deep bear markets: the Dot-Com and the Global Financial Crisis, which saw the equity market drop by over 50%. Was that unsettling at the time? You bet! But in hindsight, those deep bear markets helped me financially because of Dollar-Cost Averaging. In other words, while you’re still young and accumulating assets for retirement, you should even embrace volatility, at least to a degree. As I’ve written numerous times in my Safe Withdrawal Rate Series, while retirees should fear volatility, especially in the near-term, because they face Sequence of Returns Risk, savers can be more cavalier with it, because their Sequence Risk profile is literally the flipside of the retirees’. That’s because the cash flows of a saver plus those of a retiree are again those of a buy-and-hold investor when monthly contributions and withdrawals are equal. I made this point in Part 14 of the SWR Series. So, a temporary equity market meltdown followed by a swift recovery can be beneficial to the saver, at least in the long run. The implications: 100% equity equity weight during a good chunk of your accumulation path is perfectly fine.
For more details, please see Part 43 of my Safe Withdrawal Rate Series Pre-Retirement Glidepaths: How crazy is it to hold 100% equities until retirement?
Conclusion
There you have it. Six more personal finance lies and myths debunked. And again, I’m not against diversification. I have applied it both personally in my portfolio and professionally when working for a large global asset manager. However, diversification is often abused by financial service industry salespeople to push new products on you. They want you to believe that your simple retirement portfolio with one single equity index fund is not diversified enough. They are lying to you. They want you to replace your low-expense-ratio VTI or VOO with their new offerings, which charge 10x the fees.
So much for today and likely for the rest of the year. I wish you all a peaceful Christmas season and a healthy, prosperous, and Happy New Year!
Thanks for stopping by today! I’m looking forward to your comments and suggestions.
Picture Source: Pixabay.com.
Like this:
Loading…
