The initial portfolio is: IVV - 81 allocation units, VGT - 7 allocation units, VPU - 7 allocation units, and GLD - 7 allocation units

We are comparing two rules:

Rule 1 - Buy and hold. The portfolio is initialized at the January 2005 month-end observation and no subsequent reallocations are made. Each position is allowed to rise or fall with market performance for the entire test period.

Rule 2 - Full dynamic reallocation rule. The portfolio begins with the same initial allocation. No reallocations are made during 2005. Beginning with the January 2006 month-end observation, reallocations occur only when one of the events below is triggered. There is no routine annual rebalancing and no scheduled restoration of target portfolio weights.

VGT event. The VGT target begins at 10% of the portfolio. Each January, the target may increase according to the rise in the Information Technology weight of the S&P 500 since the starting date: VGT target = 10% + 12.5% × the increase in the S&P 500 Information Technology weight. The target cannot decline and is capped at 13%. If VGT reaches a month-end adjusted-price level at least 50% above a qualifying prior high established at least 12 months earlier and VGT is also above its current target weight, 25% of VGT’s dollar amount above the target is sold and invested in VPU.

GLD event. If GLD reaches a month-end adjusted-price level at least 30% above a qualifying prior high established at least 12 months earlier, 25% of the entire GLD position is sold and invested in IVV.

IVV outperformance event. Each month, the trailing three-month total return on IVV is compared with the trailing three-month total return on VPU. If IVV outperforms VPU by 20 percentage points or more and the differential freshly crosses that threshold, 10% of the current IVV position is sold and invested in VPU.

VPU outperformance event. If VPU outperforms IVV by 10 percentage points or more—equivalent to an IVV-minus-VPU differential of -10 percentage points or less—and the differential freshly crosses that threshold, 8% of the current VPU position is sold and invested in IVV.

After any event triggers a transaction, that same event cannot trigger another transaction for at least 12 months. If more than one event is triggered in the same month, each transaction amount is calculated from the portfolio immediately before that month’s trades and the transactions are then executed together.

Experiment design. The test uses 241 month-end observations from January 2005 through January 2025, producing 240 monthly return observations. Both portfolios begin with the identical 81 IVV / 7 VGT / 7 VPU / 7 GLD normalized allocation. The no-reallocation portfolio is held unchanged throughout. The full dynamic rule begins making eligible adjustments in January 2006.

Monthly adjusted prices are used for IVV, VGT, VPU, and GLD so that distributions are incorporated in returns. CAGR is calculated from terminal wealth. Annualized volatility is the sample standard deviation of monthly portfolio returns multiplied by the square root of 12. The Sharpe ratio uses monthly excess returns based on the FRED TB3MS three-month Treasury-bill series divided by 12 and is annualized by the square root of 12.

The test does not include taxes, bid-ask spreads, commissions, or other transaction costs. The thresholds are treated as a research specification rather than as optimized trading recommendations.

2. Results

Performance

• CAGR. The no-reallocation portfolio produced a CAGR of 10.875%, compared with 11.207% for the full dynamic rule. The dynamic rule therefore increased CAGR by 0.332 percentage point, or about 3.1% relative to the no-reallocation result.

• Annualized volatility. Annualized volatility was 13.545% with no reallocation and 14.010% under the full dynamic rule. The dynamic rule therefore increased volatility by 0.465 percentage point, or about 3.4%.

• Sharpe ratio. The Sharpe ratio increased from 0.7155 with no reallocation to 0.7178 under the full dynamic rule, an increase of 0.0022, or about 0.31%.

• Ending value of a $10,000 investment. A $10,000 equivalent investment grew to $78,829 with no reallocation and $83,686 under the full dynamic rule. The dynamic rule therefore produced about $4,857 more terminal wealth, an increase of about 6.2%.

The full dynamic rule produces the higher compound return and slightly higher Sharpe ratio, but it also produces higher volatility and a somewhat deeper maximum drawdown. Under the working success criterion used in this research - higher CAGR, higher Sharpe, and no increase in volatility - the full rule therefore does not qualify as an unambiguous improvement over no reallocation.

Ending Portfolio

• No-reallocation portfolio. The portfolio ends with 75.42% in IVV ($606.39), 14.43% in VGT ($116.00), 4.83% in VPU ($38.80), and 5.33% in GLD ($42.87), for a total normalized value of $804.06.

• Full dynamic-rule portfolio. The portfolio ends with 82.14% in IVV ($701.11), 13.08% in VGT ($111.65), 2.67% in VPU ($22.75), and 2.12% in GLD ($18.09), for a total normalized value of $853.60.

Allocation Activity

The corrected full rule generates 12 allocation transactions in 12 distinct months over the 20-year test: 3 GLD-to-IVV transactions under the gold rule, 1 VGT-to-VPU transaction under the technology rule, 8 VPU-to-IVV transactions under the three-month differential rule, 0 IVV-to-VPU transactions because the +20 percentage-point threshold is never reached.

The asymmetry of the three-month differential rule remains important to the final portfolio. The -10 percentage-point VPU-outperformance threshold produces eleven fresh crossings in the raw monthly data, but three occur during a 12-month same-event lockout and therefore do not generate transactions. The +20 percentage-point IVV-outperformance threshold is never reached. As a result, the rule progressively transfers capital from VPU into IVV, and the full-rule portfolio ends with 82.1% in IVV and only 2.7% in VPU.

Interpretation

The experiment shows that the event-driven rule can raise terminal wealth without frequent trading, but the gain is accompanied by greater equity concentration and somewhat higher risk. This specific full rule ends with about $4,857 more per $10,000 after 20 years and higher annualized volatility of 0.46 percentage rates. My search with the assistance of CHAT GPT for the holy grail, a portfolio and allocation strategy that beats the market by a lot with less risk has not succeeded, but I will persist.

Author’s Note: I publish a wide range of personal-finance, investing, economic, and policy material on my multi-topic blog, Economic and Political Insights. One example is The Sequence of Returns Puzzle: Why Timing Hurts Workers and Retirees in Opposite Ways, which uses simulations and retirement examples to demonstrate that the impact of the path of returns on terminal retirement wealth can often be more important than average returns.

Please browse the site—there is much more there for readers interested in investing, personal finance, economics, and public policy.

Share

To examine the question, I compared the three groups using two related measures: how far their stocks currently trade below their 52-week highs and the size of their 52-week high-to-low trading ranges.

Answer: Yes, at least in this small group of companies.

The clearest result concerns distance from the 52-week high. The smaller drone-focused companies are, on average, 62.3% below their highs. The larger drone-focused companies are 46.3% below, while the diversified companies are only 16.6% below their highs.

The same progression appears in 52-week trading ranges. The smaller drone-focused stocks have traded over ranges equal to about 75.7% of their 52-week highs, compared with 65.9% for the larger drone-focused companies and 55.7% for the diversified companies.

The two measures are related and should not be treated as independent evidence. A narrower high-to-low range mechanically limits how far a stock can fall below its high, and both results may reflect the same underlying factors. Larger and more diversified businesses may have more established revenue streams, stronger financial resources, or simply be regarded by investors as less speculative.

The more modest conclusion is therefore the better one: among these companies, greater scale and diversification are associated with stocks that are closer to their 52-week highs and that have traded within narrower 52-week ranges.

The particularly interesting comparison is within the drone-focused companies themselves. Even after removing the large, diversified technology companies, the four larger drone-focused companies are about 16 percentage points closer to their 52-week highs than the four smaller drone-focused companies.

Subscribe now

Share

The six portfolios compared are 100% IWB, 80% IWB / 20% IWF, 80% IWB / 20% VUG, 80% IWB / 20% RPG, 80% IWB / 10% IWF / 10% VUG, and 80% IWB / 6.67% IWF / 6.67% VUG / 6.67% RPG.

Using recent 2026 holdings data, the three growth funds considered for the portfolio sleeve—IWF, VUG, and RPG—differ in the following ways.

· IWF — iShares Russell 1000 Growth ETF. The fund holds about 367 stocks. Its four largest positions are NVIDIA, Apple, Alphabet Class A, and Broadcom, which together account for about 34.7% of the fund. Information technology represents about 54% of the portfolio.

· VUG — Vanguard Growth ETF. The fund holds about 154 stocks. Its four largest positions are NVIDIA, Apple, Microsoft, and Alphabet Class A, which together account for about 40.4% of the fund. Technology represents about 70% of the portfolio. Thus, VUG is considerably more concentrated in both its largest companies and technology than IWF.

· RPG — Invesco S&P 500 Pure Growth ETF. The fund holds only about 69 stocks, but its holdings look very different from those of IWF and VUG. Its four largest positions are Sandisk, Micron Technology, Comfort Systems USA, and CrowdStrike, which together account for only about 15.0% of the fund. Technology represents about 41% of the portfolio. RPG therefore has far fewer holdings but is much less dominated by the mega-cap growth companies that lead IWF and VUG because its index weights stocks according to growth characteristics rather than simply allowing the largest growth companies to dominate the portfolio.

The experiment asks which of these six core/sleeve combinations performed best. The results also provide limited evidence on whether an investor benefits from holding more than one growth fund in the sleeve.

Analysis:

The sample runs from April 2006 through December 2025, the longest common live-history period available for all four ETFs. The analysis contains 237 monthly return observations.

Monthly total returns are calculated from month-end adjusted prices, incorporating distributions. Each portfolio begins with its stated allocation in April 2006. There is no subsequent rebalancing.

The risk-free rate used in the Sharpe-ratio calculations is based on monthly TB3MS Treasury-bill observations from the Federal Reserve Bank of St. Louis.

The Sharpe ratio measures return relative to risk:

Annualized Sharpe Ratio = [(Average Monthly Portfolio Return − Average Monthly Risk-Free Return) ÷ Monthly Portfolio Volatility] × √12

A higher Sharpe ratio indicates that an investor received more return for each unit of risk taken, which makes it useful for comparing portfolios whose returns and volatility differ.

The principal measures emphasized here are compound annual growth rate and the Sharpe ratio. CAGR measures the annualized rate at which wealth compounded over the period. The Sharpe ratio evaluates return relative to the volatility incurred in earning that return.

The no-rebalancing assumption also provides an intuitive test. Holdings that outperform are allowed to become larger portions of the portfolio rather than being periodically reduced back to their starting weights.

Results

The six portfolios can be summarized using two principal measures:

• Return. IWB alone produced a 10.69 percent CAGR. The highest return came from 80 percent IWB/20 percent IWF at 11.23 percent, followed almost indistinguishably by 80 percent IWB/20 percent VUG and 80 percent IWB/10 percent IWF/10 percent VUG, both at 11.22 percent. Adding RPG to IWF and VUG lowered the three-fund combination to 11.06 percent. The RPG-only growth tilt produced 10.71 percent, barely above IWB.

• Return adjusted for risk. The Sharpe ratio tells essentially the same story. The 20 percent IWF portfolio ranked first at 0.6617, followed by the IWF/VUG combination at 0.6600 and VUG at 0.6583. The three-growth-fund combination fell to 0.6482. IWB was 0.6350, while the RPG portfolio ranked last at 0.6228. Thus, RPG not only produced substantially less additional return than IWF or VUG; after adjusting return for volatility, it performed worse than IWB itself.

Although an annual-return difference of roughly one-half percentage point may appear modest, compounding makes it meaningful over nearly twenty years. An initial $100,000 invested in IWB grew to approximately $743,000, while the 80 percent IWB/20 percent IWF portfolio grew to approximately $818,000—a difference of about $75,000.

Adding either IWF or VUG increased both return and risk-adjusted return. One reason the difference relative to IWB was not larger is that IWB itself already contained substantial exposure to many of the same large technology and growth companies. Although the initial 20 percent IWF allocation grew to about 27 percent without rebalancing, the resulting increase in the portfolio’s underlying technology exposure was therefore more modest than the change in the fund weights alone might suggest.

Adding RPG to the growth sleeve reduced both return and risk-adjusted return relative to using IWF or VUG. This does not establish that investors should always diversify across growth funds, but it does show the risk of relying on a single fund simply because it carries the desired style label. An investor who does not understand the differences among competing index methodologies may therefore have reason to diversify across broadly similar approaches rather than assume in advance that one particular methodology will prove superior.

The broader lesson is that investors either need to understand in considerable detail what their funds own and how their indexes are constructed, or they need to diversify across methodologies rather than assume that funds with the same label are interchangeable. Two overlapping funds may still reduce index-selection risk, while another fund in the same category may represent a substantially different investment strategy.

Appendix: Verification of Portfolio Calculations

Sample period: April 2006–December 2025

The portfolio results reported in this paper were independently reconstructed in Python and in an Excel-compatible spreadsheet workbook. The purpose of this appendix is to document that verification rather than repeat the substantive findings discussed in the paper.

The analysis uses 237 monthly observations based on month-end adjusted prices for IWB, IWF, VUG, and RPG, incorporating distributions. Each portfolio was established at its stated initial allocation in April 2006 and was not subsequently rebalanced. The risk-free rate used in calculating Sharpe ratios is based on the monthly TB3MS Treasury-bill series.

For verification, the monthly portfolio paths were rebuilt independently from the underlying fund data. CAGR, volatility, Sharpe ratios, and the other portfolio statistics were then calculated separately in Python and with spreadsheet formulas. The two implementations agreed essentially to machine precision, with a largest absolute difference across the principal statistics of approximately 1.6 × 10⁻¹⁵.

Figure A1. Independently verified portfolio results, April 2006–December 2025.

The verification therefore confirms that the return and risk comparisons reported in the paper are reproducible under the stated no-rebalancing methodology.

Author’s Note: Readers interested in related work on portfolio construction may also want to see Can Investors Find the Few Stocks That Create Most Market Wealth?, which examines the evidence for diversification when a small number of stocks generate most market wealth, and Broad Market vs. Sector ETFs: Risk & Return Revisited (2007–2024), another portfolio experiment comparing broad-market and sector-fund strategies.

Subscribe now

Share

U.S. and international stocks were highly correlated during the past decade.

From July 2016 through June 2026, the quarterly returns of VOO and VXUS had a Pearson correlation of approximately 0.86. A correlation that high means the two funds generally rose and fell together. It therefore leaves relatively little room for a small international allocation to reduce portfolio volatility.

The limitation becomes even more apparent when international stocks make up only 10% of the portfolio. Even when VXUS behaves somewhat differently from VOO, its weight is too small to substantially alter the performance of a portfolio that remains 90% invested in U.S. large-cap stocks.

That leads to the central question:

Does adding a 10% international allocation create a more efficient portfolio by improving the relationship between return and risk?

Methodology:

The work here involves returns from portfolios formed from a combination of three funds -- VOO, the Vanguard S&P 500 ETF; VXUS, the Vanguard Total International Stock ETF; and BIV, the Vanguard Intermediate-Term Bond ETF.

The analysis covers July 1, 2016, through June 30, 2026. It uses Vanguard’s published quarterly market-price total returns, including reinvested distributions.

Read more

1. Better Ratios Don’t Always Mean Higher Returns

A portfolio’s efficiency ratio can improve simply because risk fell faster than returns. You get a smoother ride, but you end up with less total wealth over time.

2. Emerging Markets Don’t Reliably Cut Risk

Emerging-market equities do not systematically lower portfolio risk. In practice, they often increase total volatility—the ratio improves only because the extra return outweighs the extra risk.

3. How You Measure Risk Can Change the Verdict

Standard metrics treat both big gains and big losses as “risk.” Because emerging markets suffer from sudden, sharp drawdowns, evaluating them strictly on downside risk reveals exposure that traditional metrics miss.

The academic literature finds that investing in emerging-market equities can improve a portfolio’s return–risk ratio because emerging-market returns do not move perfectly with developed-market returns. But a better ratio does not necessarily mean a higher return. It may result from higher return, lower risk or some combination of the two. This paper examines what produces the improvement under two definitions of risk: total volatility, measured by standard deviation, and downside volatility, measured by semivariance.

A better return–risk ratio can arise in three principal ways:

• Return rises while risk falls.

• Return and risk both rise, but return rises proportionately more.

• Return and risk both fall, but risk falls proportionately more.

The third outcome requires particular attention. The portfolio becomes statistically more efficient, but the investor accumulates less wealth. Diversification has successfully reduced risk, but only at the cost of accepting a lower return.

Standard Deviation and Semivariance

Standard deviation measures the dispersion of returns around their average. It treats a return far above the average as just as risky as an equally large return below the average. The Sharpe ratio generally divides the return above a risk-free benchmark by standard deviation.

Semivariance measures only returns below a specified target. The target can be zero, the risk-free rate, the portfolio’s average return or another required return. For example, zero-target semivariance considers only negative returns and gives greater weight to larger losses by squaring each shortfall. Its square root, semideviation, is expressed in the same units as standard deviation. A downside return–risk measure such as the Sortino ratio divides excess return by semideviation rather than by total volatility.

This distinction matters for emerging markets because their returns are often skewed, volatile and subject to unusually severe losses. A portfolio can therefore look attractive when all volatility is counted as risk but less attractive when only harmful volatility is considered.

Evidence Using Standard Deviation

Better Ratio, Lower Return

Min and Kim found that adding international regional indexes provided significant diversification benefits to Korean investors, but that the source and size of those benefits differed substantially across markets. Their results show that adding emerging-market Latin America lowered both the portfolio’s return and its standard deviation while raising the Sharpe ratio from 3.14 to 3.44. The authors concluded that the largest efficiency gains came from emerging-market Latin America and emerging-market Europe.

This is therefore an example of a better return–risk ratio achieved through risk reduction at the cost of a lower return. Note that these results reflect a South Korean won perspective against domestic South Korean benchmarks; exchange-rate dynamics and baseline volatility differ for U.S. dollar investors. Furthermore, the maximum efficiency gains relied on unconstrained optimization with large short positions rather than realistic long-only allocations.

The result should not be interpreted as a practical allocation recommendation. The optimized portfolio contained substantial short positions and an unusually large emerging-market allocation. It demonstrates the nature of the tradeoff, not the likely result of adding a modest emerging-market position to a conventional American portfolio.

Higher Economic Gains Without Consistent Risk Reduction

Bouslama and Ouda found that the economic gains from international diversification remained substantial for American investors and that emerging and frontier markets were major components of their unrestricted optimized portfolios. Their variance-optimized strategy produced greater terminal wealth, a higher Sharpe ratio and slightly lower standard deviation than the U.S.-only benchmark.

However, the authors’ broader conclusion was that unrestricted international diversification did not consistently reduce volatility or minimum loss across the different strategies they tested. Substantial emerging-market exposure primarily enhanced economic gains rather than reliably reducing risk. The authors found clear reductions in return variability and minimum loss only in their restricted portfolios.

The favorable mean-variance result should therefore be treated as strategy-specific. It demonstrates that higher return, lower standard deviation and a better Sharpe ratio can occur together, but it was not the general result across all the portfolio methods examined in the study.

The Min and Kim and Bouslama–Ouda results illustrate why an improved Sharpe ratio cannot be interpreted by itself. In one case, the ratio improved because risk fell more than return. In the other, a particular optimized strategy increased wealth while modestly reducing standard deviation.

Evidence Using Downside Risk

From an investor’s perspective, the relevant risk is generally not volatility itself but the possibility that returns will fall below an acceptable target. Returns above that target are beneficial and ordinarily should not be penalized as risk. Standard deviation remains widely used because it is simple to calculate and produces a tractable portfolio model; it is also a reasonable proxy for downside risk when returns are approximately symmetric.

Emerging-market returns, however, are frequently skewed and non-normal. Stevenson therefore evaluated emerging-market portfolios using lower partial moments, which count only returns below a specified target. He demonstrated that severe negative skewness and high kurtosis, or fat tails, cause standard mean-variance optimization to understate downside exposure, whereas lower partial moments reallocate capital to protect against extreme tail events. He found that this more direct measure of harmful volatility could materially alter portfolio allocations and produce significant performance improvements for risk-averse investors.

Better Downside-Risk Ratio, Lower Return

Bouslama and Ouda found a direct tradeoff between economic gains and downside-risk protection. Reducing emerging- and frontier-market exposure lowered return variability and minimum loss, but it also reduced terminal wealth. At the same time, the semi-variability ratio increased across their optimized strategies when they moved from unrestricted to restricted portfolios.

The authors concluded that unrestricted portfolios were more attractive to investors seeking economic gains, while restricted portfolios were more attractive to investors seeking lower volatility and smaller losses.

This is the central tradeoff examined here. Moderating emerging-market exposure improved downside-risk-adjusted performance, but only by sacrificing some return. Whether that is desirable depends on how highly the investor values protection against losses relative to long-run wealth accumulation.

The comparison is imperfect because the restricted portfolios imposed a 50 percent U.S. allocation and excluded some less-investable markets. It nevertheless provides a direct example of a higher downside return–risk ratio being achieved at the cost of lower accumulated wealth.

Higher Return and Better Ratios, but Higher Risk

Beach examined monthly rebalanced portfolios combining developed- and emerging-market equities. He concluded that “higher returns and higher risk are associated with portfolios that have higher allocations to emerging market equities.” Beach reinforced this using both semideviation and Downside CAPM, or D-CAPM, showing that while total downside risk rose with higher emerging-market exposure, the additional return expanded rapidly enough to improve downside-risk-adjusted ratios.

The Sharpe and reward-to-semideviation ratios nevertheless improved because the additional return more than compensated for the higher standard deviation and semideviation. This was therefore not a risk-reduction result.

What the Studies Establish

Taken together, the studies show that a better return–risk ratio can accompany either higher or lower returns and either higher or lower risk, whether risk is measured by standard deviation or by downside volatility.

The literature therefore does not support a general claim that emerging markets improve portfolio performance by reducing risk. Sometimes risk falls at the cost of return. Sometimes both return and risk rise, but return rises enough to improve the ratio. Under some portfolio methods, return rises while measured risk also falls.

The choice of risk measure can also change the assessment of the same portfolio. A portfolio may appear attractive under standard deviation because upside volatility is treated as risk and included in the denominator. The same portfolio may look less favorable when only returns below a specified target are considered.

Reading List

1. Byoungkyu Min and Tongsuk Kim. “An Examination of International Portfolio Diversification Benefits for Korean Investors.” Journal of Derivatives and Quantitative Studies, Volume 18, Issue 1, 2010. The study examines the diversification benefits of adding developed- and emerging-market regional indexes to Korean equity portfolios.

2. Ons Bouslama and Olfa Ouda. “International Portfolio Diversification Benefits: The Relevance of Emerging Markets.” International Journal of Economics and Finance, Volume 6, Issue 3, 2014. The study compares international portfolios constructed using variance, GARCH variance, conditional value at risk and lower partial moments.

3. Simon Stevenson. “Emerging Markets, Downside Risk and the Asset Allocation Decision.” Emerging Markets Review, Volume 2, Issue 1, 2001, pages 50–66. The study directly compares conventional mean-variance optimization with portfolio construction based on lower partial moments.

4. Steven L. Beach. “Why Emerging Market Equities Belong in a Diversified Investment Portfolio.” The Journal of Investing, Volume 15, Issue 4, Winter 2006, pages 12–18. The study evaluates emerging-market allocations using standard deviation, semideviation, conventional beta and downside beta.

Share

Subscribe now

This analysis compares the S&P 500 with portfolios that began with 90 percent stocks and 10 percent gold. Using monthly returns from 2010 through 2025, it tests whether a tolerance-band strategy—rebalancing to 10 percent gold whenever gold moves below 7.5 percent or above 12.5 percent—improved downside protection. The strategy modestly reduced return but reduced downside risk by considerably more, producing the highest return relative to downside risk among the approaches examined.

Key Findings

• Annual return fell modestly.

• Semivariance fell about 20 percent.

• Downside-adjusted return improved.

My previous article, “Does Adding Gold Reduce Portfolio Risk?,” used 16 annual observations to compare the S&P 500 with 90/10 stock-and-gold portfolios that were either rebalanced annually or never rebalanced. It found that gold modestly reduced risk, but that most of the benefit depended on maintaining the intended allocation. The article utilized the semivariance and based the risk/return ratio on downside risk rather than the typical mean return and standard deviation of return measures because a risk measure based on downside price movements is more relevant than one dominated by positive returns.

This post extends the previous analysis by considering a more complex rebalancing procedure and by comparing risk/return ratios with monthly data over the January 2010 to 2025 period. Four strategies are considered.

• A portfolio invested entirely in the S&P 500.

• A portfolio maintained at 90 percent S&P 500 and 10 percent gold through annual rebalancing.

• A portfolio that began with the same 90/10 allocation but was never rebalanced.

• A portfolio using the 7.5-to-12.5-percent tolerance band and returning to 90 percent stocks and 10 percent gold whenever either boundary was breached.

The calculations use monthly total returns for SPY as a proxy for the S&P 500 and GLD as the gold investment, with distributions reinvested. Taxes, transaction costs and bid-ask spreads are excluded.

Adding gold created a tradeoff. Returns went down but risk went down by a greater amount. The reduction in risk was largest when reallocation was actively managed.

· The tolerance-band portfolio produced a compound annual return of approximately 13.74 percent, compared with 14.02 percent for the S&P 500. Gold therefore reduced return by about 0.28 percentage point annually, but it reduced downside risk sufficiently to improve return relative to downside risk.

· The S&P 500 had zero-target semivariance of 0.007655 and annualized downside deviation of 8.75 percent; the 90/10 portfolio without rebalancing, 0.006844 and 8.27 percent; the annually rebalanced 90/10 portfolio, 0.006139 and 7.84 percent; and the tolerance-band portfolio, 0.006117 and 7.82 percent.

A higher ratio indicates that the portfolio produced more average return for each unit of downside risk.

This is described as “Sortino-style” rather than a conventional Sortino ratio because the numerator does not subtract a risk-free return or another minimum acceptable return. Zero is used only as the threshold for identifying downside months: returns below zero contribute to downside deviation, while returns at or above zero do not.

The results were:

• S&P 500: 1.624.

• 90/10 portfolio without rebalancing: 1.663.

• 90/10 portfolio with annual rebalancing: 1.751.

• 90/10 portfolio with tolerance-band rebalancing: 1.764.

The tolerance-band strategy produced the highest downside-adjusted return, although its advantage over annual rebalancing was small. Its ratio improved not because gold raised raw return, but because downside risk declined proportionately more than return.

The more important result is that both disciplined rebalancing approaches performed substantially better on a downside-adjusted basis than either the S&P 500 alone or the portfolio that was allowed to drift.

The tolerance-band strategy also moderated the most severe monthly loss.

• S&P 500: −12.26 percent.

• 90/10 portfolio without rebalancing: −11.71 percent.

• 90/10 portfolio with annual rebalancing: −10.91 percent.

• 90/10 portfolio with tolerance-band rebalancing: −10.86 percent.

Gold did not prevent losses, but it reduced their magnitude. Because semivariance squares negative returns, reducing a particularly large loss has a significant effect on measured downside risk.

Annual rebalancing restores the portfolio to its target allocation on a fixed schedule, whether or not the allocation has changed materially. The tolerance-band rule instead requires a transaction only when gold moves at least 2.5 percentage points away from its 10 percent target.

The rule triggered eight rebalances during the 16-year period, compared with 16 scheduled annual rebalances. It maintained the intended diversification with fewer transactions while producing slightly lower downside risk and a slightly higher return relative to downside risk.

Some Notes:

This is a historical case study rather than proof that every investor should hold 10 percent gold or use these exact boundaries.

The period from 2010 through 2025 was unusually favorable to U.S. equities. Results could differ during an extended period of weak stock returns, high inflation or unusually strong gold performance.

The downside target used to calculate semivariance is a judgment selected by the analyst. This analysis uses a target of zero, so only months with negative returns are treated as downside observations. An analyst could instead choose a target of 3 percent, negative 3 percent, the inflation rate, the Treasury-bill return or the minimum return needed to finance retirement spending. Changing the target would change both the number of months classified as downside periods and the resulting semivariance.

Taxes, which are relevant for brokerage but not retirement accounts and trading costs are excluded.

A 10 percent gold allocation modestly reduced return but reduced downside risk by considerably more. The 7.5-to-12.5-percent tolerance band preserved that protection with only eight rebalances and produced the highest return relative to downside risk among the strategies examined.

Appendix: How to Reproduce the Calculations

The analysis can be reproduced in Excel, Google Sheets or a similar spreadsheet. Use 192 monthly observations from January 2010 through December 2025. Assume that the first observation is in row 2 and the last is in row 193.

1. Enter the Data

Create these columns:

• Column A: Month.
• Column B: SPY monthly total return.
• Column C: GLD monthly total return.

Enter returns as decimals: 4 percent as 0.04 and a 4 percent loss as −0.04. Returns should include reinvested distributions.

For the S&P 500-only portfolio, create Column D and enter:

=B2

Copy the formula through row 193.

2. Construct the Tolerance-Band Portfolio

Create these columns:

• E: Beginning SPY value.
• F: Beginning gold value.
• G: Ending SPY value before rebalancing.
• H: Ending gold value before rebalancing.
• I: Ending total portfolio value.
• J: Monthly portfolio return.
• K: Gold weight before rebalancing.
• L: Rebalancing indicator.
• M: Beginning SPY value for the next month.
• N: Beginning gold value for the next month.

Assume an initial $10,000 portfolio. Enter 9000 in E2 and 1000 in F2.

For the first month, enter:

G2: =E2*(1+B2)
H2: =F2*(1+C2)
I2: =G2+H2
J2: =I2/(E2+F2)-1
K2: =H2/I2

Column K measures gold’s end-of-month portfolio weight before any transaction. In L2, enter:

=IF(OR(K2<0.075,K2>0.125),1,0)

A value of 1 means that gold moved outside the 7.5-to-12.5-percent band. A value of 0 means no rebalancing is required. A weight exactly equal to either boundary does not trigger a transaction.

For the following month’s beginning values, enter:

M2: =IF(L2=1,0.9*I2,G2)
N2: =IF(L2=1,0.1*I2,H2)

When the boundary is breached, these formulas restore the portfolio to 90 percent SPY and 10 percent gold. Otherwise, the ending values carry forward unchanged.

In row 3, set E3 equal to M2 and F3 equal to N2. Repeat the same formulas for Columns G through N and copy them through row 193. Count the rebalances with:

=SUM(L2:L193)

The rule produced eight rebalances during 2010–2025.

3. Construct the Other 90/10 Portfolios

For annual rebalancing, copy the tolerance-band columns and replace the indicator with:

=IF(MONTH(A2)=12,1,0)

This restores the portfolio to 90/10 at the end of each December.

For the unrebalanced portfolio, copy the columns again but always carry each asset’s ending value directly into the next month. The portfolio starts at 90/10, but its weights are never restored.

4. Calculate Semivariance and Downside Deviation

For each portfolio, create a column of squared downside returns. For the S&P 500, enter:

=MIN(D2,0)^2

For the tolerance-band portfolio, use:

=MIN(J2,0)^2

Use the corresponding return column for the other portfolios. Positive monthly returns receive a zero; negative monthly returns are squared. Zero is the analyst-selected target, not a required feature of semivariance. For another monthly target, T, use =MIN(Return-T,0)^2; a 3 percent annual target must first be converted to its monthly equivalent.

Annualized zero-target semivariance is:

=12*AVERAGE(DownsideRange)

The results were 0.007655 for the S&P 500, 0.006844 for the unrebalanced portfolio, 0.006139 for annual rebalancing and 0.006117 for tolerance-band rebalancing.

Downside deviation is:

=SQRT(AnnualizedSemivariance)

The corresponding results were 8.75 percent, 8.27 percent, 7.84 percent and 7.82 percent.

To calculate the percentage reduction in semivariance relative to the S&P 500, use:

=1-(PortfolioSemivariance/SP500Semivariance)

For the tolerance-band portfolio, the reduction was approximately 20.1 percent.

5. Calculate Returns and the Downside-Adjusted Ratio

Ending portfolio value is:

=10000*PRODUCT(1+MonthlyReturnRange)

Compound annual return is:

=(EndingValue/10000)^(1/16)-1

The S&P 500 returned approximately 14.02 percent annually, compared with 13.74 percent for the tolerance-band portfolio.

For the Sortino-style ratio, first calculate annualized arithmetic average return:

=12*AVERAGE(MonthlyReturnRange)

Then divide by downside deviation:

=AnnualizedAverageReturn/DownsideDeviation

The ratios were 1.624 for the S&P 500, 1.663 for the unrebalanced portfolio, 1.751 for annual rebalancing and 1.764 for tolerance-band rebalancing. The measure is described as Sortino-style because no risk-free or minimum acceptable return is subtracted from the numerator.

6. Calculate the Worst Month and Check the Work

The worst monthly return is:

=MIN(MonthlyReturnRange)

The results were −12.26 percent, −11.71 percent, −10.91 percent and −10.86 percent, respectively.

Finally, confirm that all portfolios use identical dates, rebalancing affects the following month’s allocation, rebalancing does not itself create a return, semivariance is annualized before taking its square root, and compound return is not confused with the arithmetic average used in the ratio.

Subscribe now

Share

This analysis compares a portfolio invested entirely in the S&P 500 with two portfolios that began with 90 percent in the S&P 500 and 10 percent in gold. It examines both whether gold improved risk-adjusted performance and whether semivariance—which measures only downside outcomes—better captures gold’s potential diversification benefit.

Key Findings

• Gold slightly reduced returns but also reduced risk.

• Semivariance showed a larger reduction in downside risk than standard deviation showed in overall volatility.

• Annual rebalancing was essential to preserving the diversification benefit.

Investors often add gold to a stock portfolio on the theory that it will provide protection when equities perform poorly. Conventional portfolio analysis tests that theory using variance or standard deviation, which treat unexpectedly large gains and unexpectedly large losses as equivalent forms of volatility.

Semivariance provides a different test. It measures only returns below a specified target and therefore focuses on outcomes investors regard as harmful.

To illustrate the difference, I compared three portfolios over the 16 complete calendar years from 2010 through 2025:

• A portfolio invested entirely in the S&P 500.

• A portfolio invested 90 percent in the S&P 500 and 10 percent in gold, rebalanced annually.

• A portfolio that began with the same 90/10 allocation but was never rebalanced.

The calculations use the total returns of SPY as a proxy for the S&P 500 and GLD as the gold investment, with distributions reinvested. Each portfolio began with $10,000 on January 1, 2010.

Variance Versus Semivariance

Standard deviation measures how widely returns vary around their average. It treats a return far above the average as just as risky as a comparably large return below the average.

That may be mathematically convenient, but it does not necessarily correspond to how investors think about risk. Investors generally welcome unusually large gains and dislike unusually large losses.

Semivariance measures only returns below a chosen threshold. In this example, the threshold is zero, meaning that only negative calendar-year returns count as downside risk. Downside deviation is the square root of semivariance and expresses that risk in more familiar percentage-return terms.

Semivariance does not replace standard deviation. It answers a different question. Standard deviation asks how variable returns were in either direction. Zero-target semivariance asks how often, and by how much, returns fell below zero.

The Results

• Return: The S&P 500 produced the highest compound annual return from 2010 through 2025: 14.02 percent, compared with 13.66 percent for the annually rebalanced 90/10 portfolio and 13.62 percent for the portfolio that was never rebalanced. A $10,000 investment grew to approximately $81,563 in the S&P 500, $77,605 in the rebalanced portfolio, and $77,100 in the unrebalanced portfolio.

• Risk: Annual rebalancing produced the clearest reduction in both conventional and downside risk. Standard deviation declined from 13.89 percent for the S&P 500 to 12.94 percent for the rebalanced portfolio and 13.18 percent for the unrebalanced portfolio. Zero-target semivariance fell by approximately 18 percent with annual rebalancing but only 6 percent without it. Downside deviation declined from 4.69 percent for the S&P 500 to 4.25 percent with annual rebalancing and 4.54 percent without rebalancing. In the worst year, the S&P 500 lost 18.18 percent, compared with losses of 16.44 percent for the rebalanced portfolio and 17.62 percent for the unrebalanced portfolio.

• Return relative to risk: Arithmetic average annual return divided by standard deviation was 1.07 for the S&P 500, 1.11 for the annually rebalanced portfolio, and 1.09 for the unrebalanced portfolio. Using average annual return divided by zero-target downside deviation—a Sortino-style measure—the ratios were 3.17, 3.39, and 3.17, respectively. Gold therefore improved risk-adjusted performance when the 10 percent allocation was maintained but produced almost no improvement in downside-adjusted performance when the portfolio was allowed to drift.

Why Rebalancing Mattered

The unrebalanced portfolio did not remain a 90/10 portfolio. Stocks substantially outperformed gold during much of the period, causing gold to become a progressively smaller share of the portfolio.

By the beginning of 2025, gold represented only about 3.5 percent of the unrebalanced portfolio. Gold’s strong performance during 2025 raised its share to approximately 4.8 percent by year-end, but that remained far below the original 10 percent allocation.

The portfolio therefore had much less gold available to cushion stock-market losses than an investor might assume from its original allocation. An investor who chooses a 10 percent gold allocation for diversification cannot establish the allocation once and expect its protective role to remain unchanged.

Rebalancing periodically sells some of the asset that has performed better and purchases more of the asset that has performed worse. That can feel uncomfortable, but it is precisely what preserves the intended allocation and its diversification benefits.

What Semivariance Adds

The standard-deviation results suggest that adding gold modestly reduced volatility. The semivariance results tell a somewhat stronger story: maintaining the gold allocation reduced downside risk by considerably more than it reduced overall volatility.

Standard deviation fell by about 7 percent when the portfolio was rebalanced annually. Zero-target semivariance fell by approximately 18 percent.

This distinction matters because reducing downside losses is one of the principal reasons investors hold gold. Treating unusually large positive returns as a form of risk can obscure the value of an asset whose intended purpose is protection against adverse outcomes.

The number of negative years did not change. All three portfolios lost money in 2018 and 2022. Gold did not prevent those losses, but annual rebalancing reduced their severity.

Because semivariance squares each shortfall below zero, reducing a large loss can materially reduce measured downside risk even when the number of losing years remains unchanged. Semivariance therefore captures both the occurrence and magnitude of negative returns, not merely their frequency.

Important Qualifications

This is an illustration rather than a definitive finding about gold. Sixteen annual observations provide a relatively small sample, and only two years had negative S&P 500 returns. The semivariance estimates therefore depend heavily on what happened in 2018 and 2022.

A more rigorous analysis would use monthly returns, producing roughly 192 observations over the same period. It could also compare different downside targets, such as the Treasury-bill return, inflation, or the minimum return needed to finance retirement spending.

The period also strongly favored U.S. equities. The S&P 500 produced exceptional returns, causing any allocation to gold to reduce total wealth. A different starting date could produce different results.

Finally, semivariance is not the only measure of downside risk. Expected shortfall, maximum drawdown, and recovery time provide additional information about severe losses and the experience of remaining below a previous portfolio peak.

Conclusion

Adding 10 percent gold modestly reduced returns but also reduced portfolio risk. The improvement was more apparent when risk was measured by semivariance rather than standard deviation because gold’s principal benefit was reducing negative outcomes rather than eliminating fluctuations in both directions.

That benefit largely disappeared when the portfolio was not rebalanced. The broader lesson is therefore not simply that every investor should own 10 percent gold. It is that diversification must be maintained rather than merely initiated—and that conventional volatility may not fully measure the value of an asset whose purpose is to moderate losses.

Appendix: How to Reproduce the Calculations

The calculations can be reproduced in a spreadsheet with one row for each calendar year from 2010 through 2025.

Step 1: Enter the Annual Returns

Create the following columns:

• Column A: Year

• Column B: S&P 500 total return

• Column C: Gold total return

• Column D: Annually rebalanced portfolio return

Enter the years 2010 through 2025 in cells A2 through A17.

Enter returns as decimals. A 10 percent return is entered as 0.10, while an 18 percent loss is entered as -0.18.

For example, the 2022 entries are:

• S&P 500: −18.18 percent, entered as -0.1818

• Gold: −0.77 percent, entered as -0.0077

Step 2: Calculate the Annually Rebalanced Portfolio

For each year, multiply the S&P 500 return by 90 percent and the gold return by 10 percent.

If the S&P 500 return is in cell B2 and the gold return is in C2, enter the following formula in D2:

=0.9*B2+0.1*C2

Copy the formula down through D17.

This calculation assumes that the portfolio is restored to 90 percent stocks and 10 percent gold at the beginning of every year.

Step 3: Calculate the Portfolio Without Rebalancing

Create six additional columns:

• Column E: Beginning S&P 500 value

• Column F: Beginning gold value

• Column G: Ending S&P 500 value

• Column H: Ending gold value

• Column I: Beginning total portfolio value

• Column J: Unrebalanced portfolio return

Enter the initial investments:

• In E2, enter 9000.

• In F2, enter 1000.

Calculate the ending values for 2010:

• In G2, enter =E2*(1+B2).

• In H2, enter =F2*(1+C2).

Calculate the beginning total portfolio value:

• In I2, enter =E2+F2.

Calculate the portfolio’s return for the year:

• In J2, enter =(G2+H2)/I2-1.

The ending values for one year become the beginning values for the next year:

• In E3, enter =G2.

• In F3, enter =H2.

Then calculate the next year’s ending values:

• In G3, enter =E3*(1+B3).

• In H3, enter =F3*(1+C3).

• In I3, enter =E3+F3.

• In J3, enter =(G3+H3)/I3-1.

Copy the formulas down through 2025.

Do not restore this portfolio to its original 90/10 allocation. Its weights change automatically as the two investments produce different returns.

Step 4: Calculate Ending Values and Compound Annual Returns

For the S&P 500 portfolio, the ending value is:

=10000*PRODUCT(1+B2:B17)

For the annually rebalanced portfolio, the ending value is:

=10000*PRODUCT(1+D2:D17)

For the unrebalanced portfolio, the ending value is:

=G17+H17

The compound annual growth rate, or CAGR, is:

=(Ending value/10000)^(1/16)-1

The resulting compound annual returns were:

• S&P 500: 14.02 percent

• Annually rebalanced portfolio: 13.66 percent

• Unrebalanced portfolio: 13.62 percent

The corresponding ending values were approximately:

• S&P 500: $81,563

• Annually rebalanced portfolio: $77,605

• Unrebalanced portfolio: $77,100

Step 5: Calculate Standard Deviation

Standard deviation measures the variability of all annual returns, whether positive or negative.

For the S&P 500, use:

=STDEV.S(B2:B17)

For the annually rebalanced portfolio, use:

=STDEV.S(D2:D17)

For the unrebalanced portfolio, use:

=STDEV.S(J2:J17)

The results were:

• S&P 500: 13.89 percent

• Annually rebalanced portfolio: 12.94 percent

• Unrebalanced portfolio: 13.18 percent

The lower figures for the diversified portfolios indicate that adding gold reduced overall volatility.

Step 6: Calculate Zero-Target Semivariance

Semivariance measures only returns below a selected target. The target in this analysis is zero, so positive years contribute nothing to downside risk.

Create three more columns:

• Column K: S&P 500 squared downside return

• Column L: Rebalanced portfolio squared downside return

• Column M: Unrebalanced portfolio squared downside return

For the S&P 500, enter in K2:

=MIN(B2,0)^2

For the rebalanced portfolio, enter in L2:

=MIN(D2,0)^2

For the unrebalanced portfolio, enter in M2:

=MIN(J2,0)^2

Copy all three formulas down through row 17.

Calculate the average of each column:

• S&P 500 semivariance: =AVERAGE(K2:K17)

• Rebalanced portfolio semivariance: =AVERAGE(L2:L17)

• Unrebalanced portfolio semivariance: =AVERAGE(M2:M17)

Positive years remain in the calculation as zeros.

The resulting semivariances were:

• S&P 500: 0.002196

• Annually rebalanced portfolio: 0.001805

• Unrebalanced portfolio: 0.002066

The rebalanced portfolio’s semivariance was approximately 18 percent below that of the S&P 500.

Step 7: Convert Semivariance to Downside Deviation

Because semivariance is expressed in squared-return units, its square root is easier to interpret.

Use:

=SQRT(semivariance)

The resulting downside deviations were:

• S&P 500: 4.69 percent

• Annually rebalanced portfolio: 4.25 percent

• Unrebalanced portfolio: 4.54 percent

Downside deviation expresses below-target risk in percentage-return terms, just as standard deviation expresses overall volatility in percentage-return terms.

Step 8: Compare Return With Risk

First calculate the arithmetic average annual return for each portfolio:

• S&P 500: =AVERAGE(B2:B17)

• Annually rebalanced portfolio: =AVERAGE(D2:D17)

• Unrebalanced portfolio: =AVERAGE(J2:J17)

The conventional return-to-volatility measure is:

Arithmetic average annual return ÷ standard deviation

The resulting ratios were:

• S&P 500: 1.07

• Annually rebalanced portfolio: 1.11

• Unrebalanced portfolio: 1.09

This measure is the inverse of the coefficient of variation when average return is positive. It also resembles a Sharpe ratio with a zero risk-free rate, although no risk-free return was subtracted in this analysis.

The downside measure is:

Arithmetic average annual return ÷ downside deviation

The resulting zero-target Sortino-style ratios were:

• S&P 500: 3.17

• Annually rebalanced portfolio: 3.39

• Unrebalanced portfolio: 3.17

The calculations show that adding gold modestly improved return relative to overall volatility. Its benefit was more apparent when risk was defined as downside loss—but only when rebalancing maintained the intended gold allocation.

Subscribe now

Share

Your iShares MSCI South Korea ETF, or EWY, may appear unchanged from its last U.S. trade. Meanwhile, it is already the middle of the next trading day in Seoul, where Korean stocks could be falling sharply.

When I went on EWY at CNBC at 9 pm I saw an ETF with a quote up in the aftermarket next to news the Kospi had fallen by 8 percent.

The ETF’s displayed price may be standing still. The economic value of what you own is not.

This is only one of the risks embedded in EWY. The fund combines a mismatch between Korean and American trading hours with extraordinary exposure to two semiconductor companies and an unhedged position in the South Korean won.

It may look like a diversified international investment. In practice, it is a concentrated bet on one country, one currency and one particularly volatile industry.

1. The Time-Zone Trap

At 9:00 p.m. Denver time, it is noon the following day in Seoul during daylight-saving time and 1:00 p.m. during the winter. The Korea Exchange’s regular market session runs from 9:00 a.m. to 3:30 p.m., so Korean stocks are actively trading while most Americans are asleep.

EWY, however, is listed on NYSE Arca. By the time Seoul opens, the exchange sessions through which most American investors trade EWY have ended. The price displayed by a brokerage may consequently remain at the last available U.S. trade even as the Korean companies owned by the fund are being repriced.

That displayed price is not necessarily wrong. It is simply stale.

Suppose the Korean market has fallen 8 percent since EWY’s last meaningful U.S. trade. The precise change in EWY will also depend on the won and other market factors, but the investor’s economic position has already deteriorated even though the brokerage account may not yet display the loss.

Some brokers now provide overnight trading in selected U.S. securities, but access and liquidity vary. A trade on a thin overnight venue may also occur at a wide spread and provide a poor indication of where the ETF will trade once its primary market becomes active.

The fundamental problem remains: when the underlying Korean market is generating the most important new information, the deepest U.S. market for EWY is closed.

2. How Overnight Repricing Creates a Morning Gap

The stale closing price does not give an investor an opportunity to sell at yesterday’s level.

ETF investors need to distinguish among several different numbers:

• The fund’s last traded price.

• The current bid and ask prices.

• The fund’s net asset value.

• An estimate of what the underlying Korean securities are worth at that moment.

When Seoul is open but New York is closed, Korean shares continue to trade while EWY’s last U.S. price may remain unchanged. When Seoul is closed and EWY is trading in New York, the reverse problem arises: the last Korean share prices are stale, while investors use currency movements, futures, semiconductor stocks and new information to estimate what those shares will be worth when Korea reopens.

An ETF’s market price can therefore differ from its reported net asset value and trade at either a premium or a discount. Differences in foreign and U.S. market hours can make those discrepancies larger during periods of significant market activity.

NYSE Arca’s early trading session begins at 4:00 a.m. Eastern, or 2:00 a.m. in Denver, and its regular session begins at 9:30 a.m. Eastern, or 7:30 a.m. in Denver. When U.S. trading resumes, market makers incorporate what has happened to Korean shares, the won, semiconductor stocks and related markets.

If the overnight news is sufficiently bad, EWY may begin trading far below its previous close. There does not have to be an orderly decline through all the intervening price levels. The ETF can simply gap down.

That is also why a conventional stop-loss order cannot guarantee protection. Once the stop price is reached, a stop order becomes a market order. If EWY’s first available price is substantially below the investor’s stop, the order may execute near that lower market price rather than at the selected stop price.

The order has not malfunctioned. It has encountered a market that moved before the order had an opportunity to execute.

A stop-limit order prevents a sale below the investor’s limit price, but it creates the opposite risk: if EWY gaps through the limit, the order may not execute at all.

Even when an investor can sell, uncertainty may widen the bid-ask spread. The problem is therefore not always a complete absence of liquidity. Sometimes it is liquidity at a very unattractive price.

3. A Country Fund Can Also Be a Sector Bet

EWY sounds like a broadly diversified investment in South Korea. Technically, it owns shares in dozens of Korean companies.

Economically, however, the fund is extraordinarily concentrated.

As of July 27, 2026, EWY held 78 positions. SK Hynix represented 23.1 percent of the fund and Samsung Electronics another 21.9 percent. Together, two companies accounted for approximately 45 percent of its value. Information technology represented roughly half of the portfolio.

Those percentages will change with market prices and portfolio rebalancing, but the underlying problem remains. EWY’s performance is dominated by a small number of enormous technology companies.

An investor who believes he has purchased a general investment in the Korean economy has purchased something considerably narrower: a concentrated bet on South Korea’s semiconductor industry and two dominant corporations.

That concentration can produce spectacular gains when enthusiasm for artificial intelligence, memory chips and advanced semiconductors is rising. It can also magnify losses when expectations change.

A global semiconductor sell-off could hit EWY much harder than it hits a genuinely diversified international portfolio. The country label can obscure the fact that the investor has made a large industry bet.

4. Currency Exposure—and Why Hedging Usually Does Not Solve the Problem

EWY trades in dollars, but its underlying Korean shares are valued primarily in won.

That creates another source of risk.

Suppose Korean stocks fall 7 percent in local-currency terms. If the won also weakens by 4 percent against the dollar, the stock-market decline and the currency decline compound one another.

An investment initially worth the equivalent of $100 would fall to approximately:

$100 × 0.93 × 0.96 = $89.28

The combined loss would be about 10.7 percent—not merely the 7 percent decline reported by the Korean market.

The reverse can also occur. A strengthening won can offset part of a Korean stock-market decline or add to a local-market gain. But during episodes of capital flight or heightened global risk aversion, weakness in the local currency may intensify the losses experienced by American investors. Foreign exchange-rate movements can materially increase or reduce the dollar return on an international investment.

EWY is therefore not merely a bet on Korean companies. It is also an unhedged position in the Korean won.

In theory, a sophisticated investor could hedge some of these risks. A trader worried about the won could use foreign-exchange forwards or another position that benefits from a strengthening dollar. A trader concerned about the Korean market could use KOSPI 200 futures. Semiconductor stocks or futures might be used to offset part of the fund’s technology exposure.

None of these is a perfect hedge.

A currency position addresses movements in the won but does not protect against falling Korean stocks. KOSPI 200 futures do not precisely match the MSCI Korea 25/50 Index tracked by EWY. A semiconductor hedge introduces company-specific and basis risk.

These strategies may also involve leverage, margin requirements, trading costs and the danger that the hedge itself creates additional losses.

For most retail investors, constructing a constantly adjusted overnight hedge is unlikely to be worth the complexity. Obsessing over the market’s overnight plumbing can distract from the more important question: should an investor who cannot tolerate a sudden decline own such a concentrated position in the first place?

5. Is a Korea ETF Appropriate—and Which One?

None of this means that EWY is necessarily a bad investment.

An investor who deliberately wants exposure to South Korean semiconductors, the Korean market and the won may find it useful. EWY provides easy access to securities that would otherwise be difficult for many Americans to purchase directly.

The mistake is treating EWY as though it were a conventional, broadly diversified international allocation.

A genuinely diversified international fund spreads its investments across many countries, currencies, industries and companies. A Korea ETF does not. The investor should first decide whether a deliberate single-country position belongs in the portfolio and only then choose the vehicle.

EWY: Best for Trading and Liquidity

EWY is the dominant U.S.-listed Korea ETF.

As of July 27, 2026, it had approximately $23 billion in assets and a 30-day average volume of more than 23 million shares. Its reported median bid-ask spread was only 0.03 percent as of July 24. It charges an annual expense ratio of 0.59 percent.

That combination makes EWY the strongest vehicle for institutions and active investors who place a high value on liquidity, narrow spreads and the ability to move large positions during U.S. trading hours.

Its disadvantages are its relatively high annual fee and extreme concentration in SK Hynix and Samsung Electronics.

FLKR: A Better Low-Cost Passive Alternative

For a long-term investor seeking passive Korean exposure, the Franklin FTSE South Korea ETF, or FLKR, may be the more attractive choice. But also up in after hours tonight as Kospi craters.

FLKR charges an expense ratio of just 0.09 percent, compared with 0.59 percent for EWY. The difference amounts to approximately $50 each year for every $10,000 invested.

FLKR held 157 securities and had approximately $1.23 billion in assets as of July 7, 2026. It therefore reaches farther down the Korean market than EWY and does so at a fraction of the annual cost.

But FLKR does not solve the central risk problem. It is also market-cap weighted, has approximately half its portfolio in information technology and lists SK Hynix and Samsung Electronics as its two largest holdings. It remains exposed to the won and to the mismatch between Korean and American trading hours.

FLKR is best understood as a less expensive version of broadly similar Korean exposure, not as an escape from the risks analyzed in this article.

Its liquidity is also weaker. FLKR’s reported 30-day median bid-ask spread was 0.20 percent as of July 7, compared with 0.03 percent for EWY as of July 24. For an investor making infrequent purchases and using limit orders, the lower annual fee may outweigh that difference. For an active trader, EWY’s liquidity may remain decisive.

MKOR: Less Megacap Concentration at a Higher Cost

The Matthews Korea Active ETF, or MKOR, provides a more meaningful response to the concentration problem.

Unlike EWY and FLKR, MKOR is actively managed and can depart substantially from market-cap index weights. As of July 28, 2026, Samsung Electronics represented 18.4 percent of MKOR, while SK Hynix represented only 3.6 percent. Samsung’s preferred shares added another 2.7 percent.

MKOR was also less heavily weighted toward information technology than its benchmark and more heavily invested in industrial and smaller Korean companies. It therefore provides a broader economic interpretation of Korea than the market-cap-weighted alternatives.

That diversification is not free.

MKOR charges 0.79 percent annually. It had approximately $129 million in assets as of July 27 and a reported median bid-ask spread of 0.29 percent as of July 24. Trading volume was much lower than for either EWY or FLKR.

Active management introduces another tradeoff. When Samsung and SK Hynix lead a powerful semiconductor rally, a fund that deliberately reduces their weights may lag the market. Through June 30, 2026, MKOR had substantially underperformed its benchmark over the year-to-date and one-year periods.

That does not prove the strategy is inferior. It demonstrates the price of reducing concentration: investors gain a less top-heavy portfolio but accept higher fees, active-management risk and the possibility of missing part of a megacap-led advance.

The Practical Choice

The alternatives can be summarized simply.

EWY is the strongest trading vehicle. Its size, volume and narrow spreads are difficult to match.

FLKR is probably the strongest long-term passive choice. It offers broadly similar Korean exposure at a much lower annual cost, but it retains most of EWY’s semiconductor, currency and time-zone risks.

MKOR is the most credible choice for reducing megacap concentration. It provides a different portfolio rather than merely a cheaper version of the same index exposure, but its fees are higher and its liquidity is substantially weaker.

For an investor seeking genuine international diversification, however, the better answer may not be another Korea ETF. It may be a broad international or emerging-markets fund in which South Korea is only one component.

That approach dilutes the potential gains from a Korean semiconductor boom. It also spreads risk across more countries, currencies, industries and companies.

The real choice is therefore not merely between EWY, FLKR and MKOR. It is between making a deliberate single-country bet and allowing Korea to occupy a smaller place within a genuinely diversified portfolio.

The Bottom Line

Single-country ETFs offer convenient access to foreign markets, but convenience should not be confused with diversification, continuous pricing or guaranteed liquidity.

EWY is a concentrated investment in one country, one currency and, to a remarkable degree, two technology companies.

FLKR offers similar exposure at a much lower annual cost but does not eliminate the semiconductor concentration, currency exposure or overnight-repricing risks. MKOR reduces the dominance of the largest semiconductor companies, but it does so through a more expensive, actively managed and less liquid strategy.

When Seoul is open and New York is closed, the economic value of all three investments continues to change even if the price displayed by a U.S. brokerage appears frozen.

When meaningful American trading resumes, the market does not allow the investor to exit at yesterday’s price. It reprices the ETF to reflect what has already happened.

The lesson is not that EWY—or every Korea ETF—is defective. The lesson is that investors should understand exactly which Korean exposure they want and which risks they are willing to accept.

For active trading, EWY’s liquidity is difficult to match. For long-term passive exposure, FLKR’s lower cost makes it a strong alternative. For less megacap concentration, MKOR offers a distinctly different approach.

For genuine international diversification, however, the better answer may be not another Korea ETF but a broader fund in which Korea is only one component.

Position size matters. Currency exposure matters. Sector concentration matters. Trading hours matter. And a stop-loss order cannot protect an investor from a market that has already moved before the order has a chance to execute.

The ticker may sleep. The risk does not.

This article is for informational purposes and does not constitute individualized investment advice. Fund holdings, expenses, assets, spreads and other characteristics can change.

Subscribe now

Share

Question

During the February 2014–June 2026 period, what were the five months with the largest increases in gold prices, and what happened to bitcoin during those months? What were the five months with the largest declines in gold, and what happened to bitcoin?

Short Answer

Bitcoin rose during three of the five strongest monthly increases in gold and during four of the five largest gold declines. The extreme observations therefore provide some evidence that gold and bitcoin can act as substitutes, particularly when gold falls, but they do not reveal a stable inverse relationship. Sometimes the two assets move in opposite directions; at other times they rise together.

The analysis uses 150 first-of-month price observations from January 1, 2014, through June 1, 2026, producing 149 monthly changes. Each observation is labeled by its ending month. Thus, the March 2026 observation measures the change between February 1 and March 1, 2026, rather than the change over the March calendar month.

Five Largest Monthly Gains in Gold

• March 2026 observation: Gold rose 14.5 percent, while bitcoin fell 14.6 percent.

• August 2020 observation: Gold rose 10.9 percent, while bitcoin rose 27.4 percent.

• July 2016 observation: Gold rose 10.3 percent, while bitcoin rose 26.0 percent.

• December 2022 observation: Gold rose 9.5 percent, while bitcoin fell 17.2 percent.

• March 2016 observation: Gold rose 9.1 percent, while bitcoin rose 16.6 percent.

Bitcoin increased in three of these five observations. Its average return was approximately 7.7 percent, but the results ranged from a 17.2 percent decline to a 27.4 percent gain.

Five Largest Monthly Declines in Gold

• April 2026 observation: Gold fell 9.7 percent, while bitcoin rose 3.6 percent.

• December 2016 observation: Gold fell 9.3 percent, while bitcoin rose 3.7 percent.

• March 2021 observation: Gold fell 7.4 percent, while bitcoin rose 48.0 percent.

• August 2015 observation: Gold fell 6.8 percent, while bitcoin rose 8.9 percent.

• July 2021 observation: Gold fell 6.7 percent, while bitcoin fell 8.5 percent.

Bitcoin increased during four of the five largest gold declines. Its average return was approximately 11.1 percent, although that figure is heavily influenced by bitcoin’s 48 percent increase in the March 2021 observation. Excluding that observation, bitcoin gained an average of only about 1.9 percent during the other four periods.

Comment One: The Largest Gold Declines Provide the Most Suggestive Evidence of Substitution

Bitcoin rose in four of the five observations when gold experienced its largest declines. This pattern is consistent with investors sometimes shifting between the two assets rather than treating them as interchangeable components of the same portfolio.

The evidence is not overwhelming, however. One extraordinary bitcoin gain in the March 2021 observation accounts for much of bitcoin’s average increase during the five largest gold declines. The pattern could also reflect bitcoin-specific developments rather than direct movement of capital from gold into bitcoin.

Comment Two: Strong Gold Markets Do Not Consistently Hurt Bitcoin

Bitcoin increased in three of the five observations when gold recorded its largest gains. In the August 2020 and July 2016 observations, both assets rose sharply. These episodes suggest that gold and bitcoin can benefit simultaneously from monetary concerns, declining confidence in conventional assets, or increased demand for alternative stores of value.

Gold and bitcoin may therefore sometimes be complements in investor psychology, even when they compete for marginal portfolio allocations.

Comment Three: Some Recent Observations Show Clear Divergence

The two consecutive 2026 observations provide the clearest recent example of possible substitution.

• Between February 1 and March 1, gold rose 14.5 percent while bitcoin fell 14.6 percent.

• Between March 1 and April 1, gold fell 9.7 percent while bitcoin rose 3.6 percent.

This rapid reversal is consistent with capital moving between gold and bitcoin. It may help explain why the statistical relationship between the two assets became more negative during the recent period.

Two observations cannot establish that such rotation actually occurred, however. The divergence could also have resulted from separate news or market forces affecting the two assets differently.

Comment Four: Bitcoin’s Volatility Complicates the Comparison

Bitcoin’s movements were much larger than gold’s in several of these observations. For example, gold fell 7.4 percent between February 1 and March 1, 2021, while bitcoin rose 48 percent. That bitcoin return is an influential observation in this small group because it substantially raises bitcoin’s average return during the five largest gold declines.

It should not necessarily be called a statistical outlier, however, because similarly large monthly movements have occurred elsewhere in bitcoin’s highly volatile history.

Between July 1 and August 1, 2020, gold gained 10.9 percent while bitcoin gained 27.4 percent. These episodes demonstrate that bitcoin is not simply a modern version of gold. Bitcoin is driven by a much larger speculative cycle, and its volatility can overwhelm the underlying relationship between the two assets.

For that reason, averages calculated from only five extreme gold movements should be interpreted cautiously.

Comment Five: Extreme Months Do Not Establish a Stable Relationship

The observations provide examples of both substitution and common movement. Bitcoin rose in seven of the ten periods examined, including periods when gold was rising and periods when gold was falling.

The appropriate conclusion is therefore limited. Gold and bitcoin sometimes appear to compete for investment capital, particularly during recent periods of sharp divergence, but there is no consistent rule that a rise in one produces a decline in the other.

This conclusion is consistent with the full-period regression, which finds almost no stable relationship between monthly gold and bitcoin returns. The extreme-period analysis provides useful illustrations, but it does not overturn the broader statistical evidence.

Conclusion

The extreme observations suggest that gold and bitcoin can occasionally substitute for one another, but they more often respond to different forces or participate in the same broad demand for alternative assets. Bitcoin’s much greater volatility makes it difficult to treat it as digital gold or as a dependable complement to gold in a diversified portfolio. The evidence therefore supports neither a stable positive relationship nor a stable negative one.

Data Note 1—Gold: Gold prices come from Official Data’s Gold Historical Prices monthly table. The observations are nominal U.S. dollars per troy ounce. Official Data describes the series as based on monthly average closing prices for CME Group gold futures and presents the table as first-of-month observations. Gold Historical Prices.

Data Note 2—Bitcoin: Bitcoin prices come from Official Data’s Bitcoin Historical Prices monthly table. The observations are nominal U.S. dollar values for one bitcoin reported at the first of each month. All percentage changes used here were independently recalculated from consecutive prices and matched the month-over-month returns reported in the source table. Bitcoin Historical Prices.

Subscribe now

Share

Jim Cramer recently highlighted Hendrik Bessembinder’s paper, Which U.S. Stocks Generated the Highest Long-Term Returns? The paper shows that a remarkably small number of stocks generate most long-term stock-market wealth.

Bessembinder treats that concentration as a powerful argument for diversification; Cramer treats it as an invitation to select exceptional companies.

The Bessembinder Evidence

Bessembinder analyzes 29,078 U.S. common stocks contained in the CRSP database from December 1925 through December 2023.

The analysis reveals:

· 51.6 percent of stocks produced negative returns over their listed lifetimes.

· Seventeen stocks produced cumulative returns exceeding five million percent.

· Yet the 17 most spectacular stocks produced an average annual compound return of only 13.47 percent. Their almost unimaginable final returns resulted mainly from compounding over exceptionally long periods.

· Nvidia recorded the highest annualized return among stocks with at least 20 years of data, at 33.38 percent.

These results build on Bessembinder’s earlier and more important paper, Do Stocks Outperform Treasury Bills? That study found that four out of every seven U.S. common stocks produced lifetime buy-and-hold returns below those of one-month Treasury bills. Even more strikingly, the best-performing 4 percent of listed companies accounted for the entire net wealth created by the U.S. stock market since 1926. The remaining 96 percent, taken together, merely matched Treasury bills.

Cramer’s Interpretation

Cramer accepts Bessembinder’s central empirical finding—that most market wealth is generated by a small number of stocks—but still maintains that a portion of an investment portfolio should be placed in individual stocks.

He argues that the extraordinary winners were not necessarily obscure companies discoverable only through luck. Many were familiar businesses—including Coca-Cola, IBM, Boeing, Deere, and Johnson & Johnson—with recognizable products, strong franchises, and long records of success. Exceptional companies, in Cramer’s view, are often visible to consumers and investors before all their gains have occurred.

Read more

Every investor knows the stock market feels expensive today. But how expensive? 30%? 50%? Is a correction imminent, or are we simply operating in a new structural regime of higher permanent growth?

If you turn to Wall Street for answers, you usually get trapped in a circular argument. Analysts will tell you the S&P 500 is valued fairly compared to next year’s forward earnings estimates. But those earnings estimates are based on corporate revenue projections, which are heavily influenced by current market sentiment. It’s a self-referencing echo chamber.

To find a true, uncorrupted fundamental anchor, my son and I recently sat down to solve this quantitative puzzle. We developed an independent Trend-Line Reversion Model designed to answer two precise questions:

1. When was the last time the stock market cleared at a reasonable price reflecting long-term structural equilibrium?

2. Based on that anchor, what is the gap between the current and estimate realistic stock price?

This paper uses a trend-line reversion model to obtain valuation gap estimates for the S&P 500 under different assumptions and valuation gap estimates for portfolios of large-cap value and growth stocks.

The Trend-Line Reversion Methodology:

We assume that over long horizons, the stock market compounds at a reasonable rate, perhaps 7 percent per year. If we knew the year the stock market was a sane or reasonable estimate, we could estimate over or under valuation as the difference between the actual price and a projected reasonable price equal to S(1.07)^t where S is the value in the sane year.

A random choice of a sane year could warp the analysis. An analysis anchored to the top of the Dot-Com expansion would make subsequent decades look cheap while an analysis anchored to the absolute depth of the financial crisis would exaggerate the size of the bubble.

To eliminate cognitive bias and guesswork, we designed a Scoreboard Optimization Loop using a point-by-point least-squares percentage comparison.

We isolated a 20-year historical window (2005 through 2025) and tested every single year as a candidate baseline anchor. For each candidate year, the model projects a 7% compounding trend line forward and backward through the calibration timeline. It then compiles the Sum of Squared Percentage Errors (SSPE) between that specific trend line and actual historical data:

By squaring the percentage errors, we ensure two things: overvaluations and under valuations are penalized equally, and the math strips away raw dollar scale (so a deviation in 2005 carries the same analytical weight as a deviation in 2025).

The candidate year that finishes with the lowest total error score represents the true geometric center of gravity for the modern market—the historical point where market prices sat perfectly on the sustainable long-term trend.

Scenario 1: The S&P 500 at a 7% Normal Growth Rate

When we ran the broad S&P 500 data through the optimization loop assuming a standard historical price growth rate of 7%, the scoreboard lit up with a definitive best anchor year: 2012.

· 2005 — Total Error Score (SSPE): 1.84 | Highly Distorted (Pre-Crisis Bubble Baseline)

· 2008 — Total Error Score (SSPE): 1.12 | Displaced (Market Crisis / Undervalued Baseline)

· 2011 — Total Error Score (SSPE): 0.38 | Near Structural Fair Value

· 2012 — Total Error Score (SSPE): 0.31 (GLOBAL MINIMUM) | Optimized Structural Fair Value Equilibrium

· 2013 — Total Error Score (SSPE): 0.44 | Near Structural Fair Value

· 2016 — Total Error Score (SSPE): 0.89 | Moderate Structural Premium

· 2020 — Total Error Score (SSPE): 2.45 | Exceptionally Distorted (Emergency Stimulus Baseline)

Anchoring the model to the high-stimulus valuation of 2020 creates an enormous error score (2.45) because it forces the historical trend line way below where the market actually cleared in the mid-2000s.

The 2012 base anchor cuts through the cyclical noise flawlessly. It represents the quiet post-crisis stabilization period immediately before emergency monetary interventions completely altered equity multiples.

The 7% Valuation Gap Result

When we take that optimized 2012 trend line and project its trajectory out to the beginning of 2026, it indicates that the structural fair value of the S&P 500 should be $3,552.

The actual average price of the S&P 500 at the target checkpoint was $5,000.

Under these standard baseline assumptions, the broad market entered the year trading 40.76% above its long-term fundamental compounding curve.

The Robustness Check: What if Normal Growth is 8%?

A rigorous quantitative model must be stress-tested. What if a 7% growth assumption is too conservative for a modern economy? What if structural technology gains mean that 8% is the new baseline normal?

When we adjusted the core parameter to an 8% compounding rate and re-ran the entire optimization loop, the model shifted across two separate dimensions:

1. The Anchor Year Shifts to 2013: Because an 8% compounding line climbs at a steeper angle, the optimization loop mechanically pushed the winning anchor year forward to 2013 (SSPE: 0.38). The math needed a slightly higher historical starting price to prevent the steeper line from dropping entirely below the actual market data of the mid-2000s.

2. The Valuation Gap Compresses: Because an 8% trend line accumulates baseline value much faster, it catches up to current prices aggressively. Under the 8% model, the projected fair value for 2026 rises to $4,171.

This robustness check reveals the valuation gap is highly sensitive to the normal stock price growth assumption. A 100-basis point increase in return assumptions reduces the overvaluation gap from around 41 percent to around 20 percent.

However, the model is highly robust regarding the location of fair value. Whether you assume a 7% or 8% growth engine, the algorithm persistently locks onto the 2012–2013 window as the only structurally sound, uncorrupted baseline era of the last twenty years.

Valuation gaps for large-cap growth and large-cap value Portfolios

Looking at the S&P 500 as one massive, monolithic index can mask deeper structural imbalances. To discover where the overvaluation risk is truly concentrated, we split the index into its two primary sub-components: S&P 500 Growth and S&P 500 Value.

We cannot use a uniform, index-wide “average” growth rate for both. Because of annual index rebalancing, value indices regularly prune their top overperforming companies, while growth indices hoard multiple expansions. Furthermore, Value returns rely heavily on cash dividends (which mechanically drop out of the raw stock price), while Growth relies almost entirely on capital appreciation.

To adjust for this “style drift,” we applied distinct, historically accurate price growth parameters: 9.5% for Growth and 5.0% for Value.

When we ran these independent optimization loops, a stark divergence emerged:

1. S&P 500 Value (5.0% Baseline Anchors to 2012)

The Value index aligns perfectly with our broad market model, choosing 2012 as its optimal baseline (SSPE: 0.26).

• Actual Price (Jan 2026): $1,720

• Model-Derived Fair Value: $1,351

• Value Sector Overvaluation: 27.31%

2. S&P 500 Growth (9.5% Baseline Anchors to 2016)

The Growth index experienced a profound structural forward break, shifting its optimal anchor year to 2016 (SSPE: 0.42). This marks the exact dawn of the modern, ultra-large-cap technology dominance regime.

• Actual Price (Jan 2026): $3,650

• Model-Derived Fair Value: $2,411

• Growth Sector Overvaluation: 51.39%

The Substack Bottom Line

Perhaps the most startling result of this estimation process is that even the traditionally defensive Value sector is fundamentally overvalued. Market risk is deeply asymmetrical: the Growth sector sits in an acute overvaluation regime at 51.39% above trend, while the Value sector trades at a more moderate but still significant 27.31% above trend. If historical gravity is any indicator, a macro-reversion to the mean implies that both sectors can absolutely fall from current levels, though the drawdown will hit the heavily overextended growth curve far harder than the value baseline.

Disclaimer: This analysis represents a structural trend-line model executed for quantitative research purposes and does not constitute formal investment advice. All chart metrics correspond to annual averages up to January 2026.

Subscribe now

Share

Introduction:

Traditional valuation statistics, especially the ubiquitous PE ratio, are undefined when earnings are negative. The PE ratio cannot be used to evaluate firms in the early stages of their growth or any firms with negative earnings. This paper applies a valuation transformation proposed by Bernstein (2025) that remains mathematically defined whether earnings are positive, zero, or negative, to examine the path of valuation of two historic tech companies, Amazon and Tesla, from their inception to maturity. The paper then applies this valuation methodology to projected earnings and valuation statistics for three companies, SpaceX, OpenAI, and Anthropic, which are currently preparing for an IPO.

Methodology:

My SSRN paper, Measuring Portfolio Valuation, describes the measurement of firm and portfolio PE statistics when some firms have negative earnings. This note on valuation measures is my most frequently downloaded article.

PE ratios behave counterintuitively when earnings are negative because an increase in price causes P/E to become more negative rather than larger.

By contrast, the statistic S=(V-E)/V will increase whenever V rises or E falls, regardless of whether E is positive or negative. At E=0 this number is 1, at E<0 this number is greater than one (indicative of a higher valuation), and for E>0 this number is less than 1.0.

For instance, a company trading at a conventional P/E ratio of 20 translates to S=0.95 while a company with a PE ratio of 15 has S=0.933.

The PE ratio does not exist when E is negative, but the S remains well defined.

The S can be used to evaluate valuation in the early growth years when a company has no earnings or when a company experiences large losses later in its product cycle.

Tesla and Amazon:

The early phase of tech companies is often characterized by a combination of astronomic valuations and substantial losses. The S statistic in the early phase of a firm’s growth is a measure of the relationship between expanding market enthusiasm and escalating market losses.

For Amazon, S at its 1997 public inception was 1.071, and it reached its lifetime maximum S at 1.256 in 2000.

For Tesla, the first-year value of S at its 2010 public inception was 1.091, and it reached its maximum value of S in 2012 with a highest value of 1.104.

SpaceX, OpenAI and Anthropic:

SpaceX, OpenAI and Anthropic are preparing their IPOs, and the released projections of V and E can be used to create the initial valuation statistic S.

· SpaceX: A projected V of 1.77 trillion and E of -4.9 billion yields an S value of 1.00276.

· OpenAI: A projected V of $852 billion and E of $14 billion yields an S value of 1.01643.

· Anthropic: A projected V of $965 billion and E of -14 billion yields an S value of 1.01450.

Conclusion

Unlike the traditional P/E ratio, which suffers from mathematical discontinuity, the S statistics is defined over all values of E.

The projected S values for these pending IPOs are lower than the initial S values observed for Amazon and Tesla, indicating that their projected valuations are not unusually aggressive relative to those historical benchmarks.

Of course, actual outcomes are TBD.

Subscribe now

Share

One of the most frequently cited financial metrics is EBITDA. It appears in earnings releases, stock analyses, private equity transactions, and merger negotiations. Yet it is also one of the most misunderstood numbers in finance.

EBITDA stands for Earnings Before Interest, Taxes, Depreciation, and Amortization. In practice, analysts begin with a company’s net income and then add back interest expense, taxes, depreciation, and amortization.

The purpose of EBITDA is to answer a specific question: How profitable is the underlying business before considering how it is financed?

That distinction matters because two companies can operate essentially identical businesses while reporting very different net income. One company may have little debt, while the other may owe billions of dollars and face substantial interest payments. EBITDA attempts to strip away those financing differences and focus on the operating business itself.

For that reason, EBITDA can be a useful tool. But it also has an important limitation: shareholders do not receive EBITDA. They receive whatever remains after all the company’s obligations—including interest payments—have been satisfied.

Consider two companies with identical operations. Each generates $500 million in revenue and incurs $400 million in operating expenses. Both therefore produce $100 million of EBITDA. At first glance, the companies appear equally profitable.

Now suppose Company A carries very little debt. It pays only $5 million per year in interest expense and ultimately reports $75 million in net income.

Company B, by contrast, has accumulated a large amount of debt. Its annual interest expense is $95 million. After paying lenders and taxes, it reports only $4 million in net income.

The striking fact is that both companies report the same EBITDA: $100 million.

An analyst focused exclusively on EBITDA might conclude that the two businesses are equally attractive. In one sense, that is true. Their underlying operations generate the same amount of operating profit. But shareholders do not own the operations in isolation. They own the residual claim after everyone else has been paid.

In Company A, most of the operating profit ultimately belongs to shareholders. In Company B, most of the operating profit belongs to creditors. The business may be generating $100 million of EBITDA, but nearly all that value is being consumed by interest payments.

Defenders of EBITDA would correctly point out that high interest expense is not entirely a deadweight cost. Interest payments generally reduce taxable income, creating what finance professionals call a tax shield. That benefit is real. In this example, Company B pays only about $1 million in taxes compared with roughly $20 million for Company A. Yet the tax shield offsets only a fraction of the additional interest burden. Company B saves about $19 million in taxes but pays approximately $90 million more in interest expense. Shareholders are still substantially worse off.

This illustrates both the strength and the weakness of EBITDA.

• The strength: EBITDA helps analysts compare operating businesses without being distracted by financing choices.

• The weakness: Financing choices are not irrelevant. A company that has borrowed too much may leave very little value for shareholders even if its underlying operations appear healthy.

This is why experienced investors rarely stop at EBITDA. After seeing the EBITDA figure, they immediately ask additional questions:

• How much debt does the company have?

• How much interest must it pay each year?

• How much cash remains after those payments?

• How much profit is left for shareholders?

A useful analogy is that EBITDA measures the horsepower of the engine. It tells us something important about the vehicle’s capabilities. But it does not tell us how much weight the vehicle is towing. Two trucks may have identical engines. If one is pulling an empty trailer and the other is hauling a massive load, their performance will be very different despite having the same horsepower.

Likewise, two companies may report identical EBITDA. If one is carrying a heavy burden of debt, shareholders may receive far less benefit from that operating performance than the EBITDA figure suggests. EBITDA is therefore best viewed as a starting point rather than a conclusion. It can tell us whether a business has a strong engine. It cannot tell us how much of that power ultimately reaches shareholders.

The danger of using EBITDA as a definitive measure of health is not merely theoretical; financial history is littered with corporate collapses where analysts clung to glowing operating metrics while businesses were suffocating. During the telecom meltdown of 2002, companies like WorldCom masked enormous cash drains by classifying routine operating costs as capital expenditures. Because operating expenses lower EBITDA but capital investments do not, this trick kept EBITDA looking robust while the actual business was bleeding cash. Similarly, the Dot-Com crash of 2000–2001 popularized customized “Pro Forma” adjustments that added back marketing and customer acquisition costs, leading investors to back startups with zero path to net profitability. Even successful looking “roll-up” strategies, such as Valeant Pharmaceuticals in 2015, used heavily adjusted EBITDA figures to hide the massive, unsustainable debt loads required to acquire other firms.

Today, this exact dynamic is repeating itself in the artificial intelligence infrastructure boom. Wall Street analysts are aggressively valuing specialized AI cloud providers and enterprise software firms on massive EBITDA multiples, ignoring skyrocketing stock-based compensation (added back as a “non-cash” expense) and billions in hardware debt collateralized directly by depreciating microchips. Because training and maintaining AI requires continuous, intense computing power, treating these foundational, recurring operational outlays as ignorable capital expenditures misstates the actual profit margins. Ultimately, the heavily adjusted EBITDA metrics seen across modern AI valuations closely resemble the deceptive “Pro Forma” metrics of the Dot-Com era, promising future profitability while obscuring the heavy capital burdens required to stay alive.

This exact dynamic is playing out across the AI infrastructure ecosystem, where massive operating profits are paired with heavy capital burdens.

We can see this structural strain clearly in the current financial profiles of the market’s primary infrastructure players.

· For example, Oracle Corporation leverages its legacy profitability to generate a massive $27.4 billion in EBITDA yet carries over $123 billion in net debt to finance its intensive cloud expansion.

· Similarly, specialized private hyperscalers like CoreWeave boast phenomenal EBITDA margins above 50% but operate with gross leverage ratios near 5.8x to secure advanced hardware facilities.

· In the hardware and physical layers, semiconductor giant Broadcom utilizes its $37.2 billion in EBITDA to service a substantial debt load exceeding $70 billion.

This structural blind spot becomes uniquely dangerous in hyper-paced environments where companies must aggressively scale just to keep pace with rapid technological shifts. Frontline AI developers like OpenAI and Anthropic face an unprecedented capital treadmill: the moment a frontier model is completed, the company must immediately secure tens of billions of dollars in compute infrastructure to train the next generation, or risk total obsolescence within a year. Evaluating these foundational players on an EBITDA basis completely strips out the massive interest burdens, capital leases, and compute liabilities required to sustain their market positions. In a sector where technology degrades at an exponential rate, an analyst relying on EBITDA is evaluating a business as if its current assets will last forever, completely ignoring the reality that the cash required for technological survival is already spoken for.

When reading modern analyst reports on enterprise AI and software companies, look specifically for these red flags:

• “Rule of 40” calculations using Adjusted EBITDA: Analysts frequently add a company’s growth rate to its Adjusted EBITDA margin to see if it equals or exceeds 40%. If they used net income or free cash flow instead, many hyped AI firms would fail the test completely.

• Massive gaps between EBITDA and Free Cash Flow (FCF): If a company boasts $100 million in Adjusted EBITDA but has deeply negative Free Cash Flow, it means they are rapidly burning cash on hardware or heavy stock dilution just to keep their “profitable” operating engine running.

The Ultimate Takeaway: As billionaire investor Warren Buffett famously remarked regarding the metric: “Does management think the tooth fairy pays for capital expenditures?”

Just like the historical market cycles of the past, today’s AI valuations rely heavily on metrics that promise profitability eventually—if you agree to ignore the very real, recurring expenses required to build and sustain the technology. If a company must constantly spend cash to replace its equipment, update its technology, or service its debt, that cash is gone. EBITDA can tell you if a business has a great engine, but it won’t stop you from driving off a cliff if you ignore the balance sheet.

I found this article on the importance of cash flow and the EBITDA limitation useful.

https://www.ghjadvisors.com/ghj-insights/the-importance-of-cash-flow-and-the-ebitda-limitation

Subscribe now

Share

Since the outbreak of the Iran war on March 2, the traditional correlation between gold and oil has collapsed. While oil has surged 50%, gold has plunged 18%. This memo identifies a systemic liquidity squeeze in which interest rate concerns and a resurgent dollar have become paramount much sooner than in previous cycles.

Key Findings

• Unlike the stagflation era, the Fed quickly became hawkish, and the public rapidly shifted focus to inflation expectations.

• Real interest rates have moved into positive territory (~1.0%) just weeks into the conflict, stripping gold of its competitive edge.

• With $2 trillion in private credit funds restricting withdrawals and the MAG7 somewhat sluggish, institutional investors are selling their most liquid winner: gold.

• The U.S. is a net exporter of oil. This allows the dollar to rise alongside oil, creating additional pressure on gold.

• These factors have created sustained selling pressure on gold, even though it has traditionally served as a safe haven during geopolitical and macroeconomic turmoil.

• Gold’s decline is not necessarily a short-term dislocation; elevated starting valuations and rising real rates suggest continued downside risk rather than a clear “buy the dip” opportunity.

The Four Major Oil Shocks

Until now, gold was the “safe-haven” destination. In 2026, it has become the “source of cash.”

• 1973 Arab Oil Embargo: Gold rose 65% as the unanchored dollar weakened following the end of Bretton Woods.

• 1979 Iranian Revolution: Gold peaked at $850 (Jan 1980), only falling after Volcker pushed real rates deeply positive to restore monetary credibility.

• 1990 Gulf War: Gold saw a 12% tactical “fear spike” that faded quickly as the conflict appeared limited.

• 2026 War with Iran: Oil is up 50%, but gold has fallen 22% from its January highs near $5,600.

2026 is the first time a major oil shock has triggered a bear market in gold. In the 20th century, gold was an accumulation asset. In 2026, it is a distribution asset.

Why the 2026 Oil Shock Differs

First, both the Federal Reserve and investors quickly became focused on inflation expectations and interest rates. Support for rate cuts has evaporated as policymakers recognize they cannot risk a resurgence of inflation after misjudging it in 2022. The market has reached the same conclusion. Between March 2 and March 24, the 10-year Treasury yield surged from 4.05% to 4.38%.

Second, previous oil shocks typically coincided with a weaker dollar, which made gold more attractive as a safe haven. This time, the United States is a net exporter of energy, and both the dollar and real interest rates have risen. Despite political turmoil, the U.S. and the dollar appear to be the safest investment options.

Third, gold entered this period at elevated levels. Investors facing liquidity needs—due to weak tech performance and restrictions in private credit—are realizing gains in gold, their most liquid outperformer.

Conclusions

The relationship between gold and oil in this shock differs fundamentally from previous episodes.

In 2026, gold is declining because it is the most attractive asset to sell in a crisis where interest rates, the dollar, and oil are moving in lockstep, while other parts of the portfolio—primarily tech and private credit—are under pressure.

This does not appear to be a “buy the dip” opportunity for gold. Prices remain historically elevated, while real interest rates and the U.S. dollar continue to rise. In this environment, gold is likely to remain under pressure as monetary conditions tighten, and liquidity constraints persist.

Authors Note: www.economicmemos.com is a source of information on policy, politics, personal finance and investment. If you liked this post, and want additional advice on how to navigate the current market downturn go here. This blog consistently recognizes that investment and wealth accumulation are not the only factors impacting financial security.

Subscribe now

Share

Interesting Videos:

US Dollar Performance During Energy Shocks

The Oil Shock That Reshaped the Modern Economy - YouTube

Financial Historian · 6.6K views

This memo investigates why the recent surge in oil prices has failed to trigger a rally in “green” alternatives like heat pumps and solar power. While the long-term economic case for electrification strengthens as fossil fuels become more expensive, these sectors remain tethered to interest rate cycles and construction activity. We analyze the performance of key players like Carrier, Enphase, and First Solar to determine how and when these sectors will ultimately capitalize on the energy crisis.

Key Takeaways

• Most heat pump leaders are diversified industrial conglomerates, diluting their exposure to specific energy price shocks.

• Solar and HVAC stocks currently behave like long-duration growth assets, making them more sensitive to interest rates than to the price of a barrel of oil.

• Data center cooling requirements are creating a high-margin, non-discretionary “cushion” for diversified HVAC firms that residential markets lack.

• Investors should look for a stabilization in Treasury yields, which are linked to oil prices and inflation expectations, prior to entering these sectors.

Author’s Note

The analysis of the heat pump sector is available to all readers. Detailed research on the solar sector and specific investment entry points is reserved for paid subscribers. You can upgrade to a full subscription for just $48 per year (a 20% discount) using this link: https://www.economicmemos.com/56428713. The blog www.economicmemos.com has a mix of articles on policy politics, personal finance and investment opportunities. Readers will likely earn back the subscription fee from the personal finance section alone.

Geopolitical Shocks and Electrification: An Investor’s Dilemma

The escalation of the conflict in Iran has sent ripples through global energy markets, pushing Brent crude toward the $100–$120 range and causing localized spikes in diesel and LNG prices. For many retail investors, the intuitive reaction is to seek “green” alternatives—specifically heat pumps and solar power—as natural beneficiaries of expensive fossil fuels.

However, a closer analysis reveals a paradox: while high energy prices are a long-term structural tailwind, they are currently being overwhelmed by immediate macroeconomic headwinds. The following analysis explores why these sectors have struggled to act as “safe havens” during the 2026 energy shock.

Share

Heat Pump Companies

This memo evaluates whether higher oil and natural gas prices create a meaningful tactical opportunity in publicly traded heat pump and HVAC companies.

There is no “pure-play” publicly traded heat pump company. The leaders—Carrier Global (CARR), Trane Technologies (TT), and Lennox International (LII)—are diversified industrial conglomerates. High energy prices make heat pumps more competitive, but these firms are deeply tied to broader construction and interest rate cycles and higher oil prices do not lead to immediate gains in stock prices in this sector.

• Rising energy costs act as a tax on consumers, increasing recession risk. Households often defer high-capex purchases like heat pumps when the outlook is uncertain.

• HVAC demand tracks housing and commercial building. High interest rates raise financing costs for both installers and homeowners.

• Adoption cycles are governed by replacement needs and policy timelines, not daily oil price fluctuations.

Stock Performance (Feb 27 – March 19, 2026):

The market reaction has been “risk-off” rather than “energy-pivot”:

• Carrier (CARR): Fell ~8% (from $64.40 to $58.97).

• Lennox (LII): Saw significant drawdowns consistent with housing sensitivity.

• International Plays: Daikin and NIBE moved in lockstep with global growth concerns rather than energy pricing.

These stocks behave like cyclical industrials with a long-duration electrification overlay. Sector-specific tailwinds have not been sufficient to overcome generalized market downward trends.

Subscribe now

Solar Power Companies

Read more

Key Results

· The Structural Wall: NVIDIA’s $5 trillion valuation breaches IRS “25/50 rule” concentration limits, forcing major ETFs to automatically sell shares to maintain tax-advantaged status.

· Institutional “Firing” Risk: Active fund managers are mandated to trim positions exceeding 15-20% to manage risk, creating a ceiling for the stock regardless of fundamental strength.

· The Circular Loop: A significant portion of revenue is tied to NVIDIA’s investments in startups that then buy NVIDIA chips, creating concerns regarding the “Quality of Earnings.”

· Reduced Upside: While fundamentally strong, NVIDIA’s massive size necessitates a shift in investor expectations from doubling annually to moderate 30-40% annual growth.

· Strategic Restructuring: To break the structural ceiling, NVIDIA may need to spin off its investment arm to bypass ETF concentration limits and unlock independent valuation for its AI ecosystem.

Introduction: The Paradox of Success

Despite delivering three consecutive blockbuster financial reports—culminating in a record-shattering $68.1 billion in quarterly revenue with 73% year-over-year growth—NVIDIA’s stock price has failed to mirror this explosive financial performance. Instead of skyrocketing, the stock has frequently stalled or retreated immediately following these announcements. This frustrating lack of price appreciation is not a failure of the AI revolution; it is a symptom of “The Structural Ceiling”—a point where a company becomes too successful for the financial “containers” designed to hold it.

1. The Mechanical Sell-Wall: ETFs and the 25/50 Rule

The primary headwind isn’t sentiment; it’s math. Most Diversified ETFs operate under IRS “RIC” (Regulated Investment Company) rules, specifically the 25/50 rule. This mandates that no more than 25% of a fund’s assets can be in a single issuer, and the sum of all holdings over 5% cannot exceed 50% of the portfolio.

When NVIDIA’s stock price surges, it frequently breaches these legal thresholds. To maintain their tax-advantaged status, ETF managers are legally forced to sell NVIDIA and “recycle” that capital into smaller, often slower-growing firms.

As of February 2026, concentration levels in major ETFs are nearing critical thresholds:

· VanEck Semiconductor ETF (SMH): Currently holding ~20-25% in NVIDIA, frequently hitting the “Concentration Wall” and triggering forced selling.

· Vanguard Info Tech ETF (VGT): Positioned at ~17-19%, restricted by sector diversification rules upon further appreciation.

· Invesco QQQ Trust (QQQ): Holdings at ~13-15%, balanced by Apple/Microsoft, but facing high “overlap” risk.

· ProShares Ultra Semi (USD): Leveraged exposure exceeding 30%+, making it a high-volume “powder keg” for forced selling during spikes.

2. Institutional Mandates: The “Risk” Ceiling

It isn’t just passive funds. Giants like Vanguard, BlackRock, and State Street manage trillions. For an active fund manager, holding a 15-20% position in a single stock is often a violation of internal risk mandates.

As of December 2025 reporting, top institutional ownership remains concentrated:

· Vanguard Group: 2.27 Billion shares (9.23% ownership of NVIDIA)

· BlackRock: 1.94 Billion shares (7.98% ownership)

· State Street: 0.99 Billion shares (4.08% ownership)

When one or more of these three firms are forced or incentivized to trim their NVIDIA holdings—often because a stellar earnings report has pushed NVIDIA to a disproportionately large percentage of their total portfolio—the share price stagnates. Because these three firms alone own over 21% of the company, their synchronized need to manage concentration risk creates a massive supply of shares that effectively “mops up” any new buying pressure.

3. The “Circular Exposure” Web

Sophisticated investors are also wary of implicit exposure. NVIDIA has pioneered a “Virtuous AI Loop” where they invest in AI startups (e.g., CoreWeave, OpenAI) which then commit to buying NVIDIA chips.

Critics point to these “circular deals” as a potential risk to the Quality of Earnings. Furthermore, if you own Microsoft, Amazon, or Meta, you are indirectly betting on NVIDIA, as they are its largest customers. Total effective exposure for a tech-heavy investor often sits at 35-45%, leading to a “Sentiment Ceiling” where buyers simply feel “full.”

This massive concentration is why NVIDIA’s news has the potential to spill over into the rest of the tech sector. If the “Big Three” are forced to sell NVIDIA, the resulting liquidity shifts can shake the entire market. This is precisely why investors become so anxious around NVIDIA’s announcements – they impact the whole tech market.

4. The Valuation Paradox: Size vs. Growth

NVIDIA currently sells at a lower Forward P/E and PEG ratio than its competitors for two reasons.

First, NVIDIA’s shares are impacted by the selling pressure associated with institutional rules and the need for diversification described in this memo. Second, NVIDIA faces a “Size Discount” because the “Law of Large Numbers” suggests maintaining current growth is improbable and eventually some rival will come up an innovation that results in a loss of market share.

· NVIDIA (NVDA): Forward P/E ~25.0x | PEG Ratio 0.80

· AMD: Forward P/E ~40.0x | PEG Ratio 1.20

· Broadcom (AVGO): Forward P/E ~35.0x | PEG Ratio 1.25

· ASML: Forward P/E ~43.0x | PEG Ratio 1.80

In any other context, a PEG ratio below 1.0 would signal a massive “Strong Buy,” but here it reflects a stock that has become too large for the market to price efficiently.

Author’s Note to Subscribers: The structural dynamics described above explain why stellar earnings no longer guarantee a stock surge. To manage your portfolio effectively in 2026, you need to understand how to maneuver around these mechanical sell-walls.

Share

Paid subscribers can read on for a concrete investment strategy to manage NVIDIA’s reduced upside, a comparative analysis of how other companies solved this “concentration risk,” and the specific strategic recommendation for a corporate restructuring to unlock value.

Subscribe now

5. Managing “Reduced Upside” and Corporate Solutions

Read more

Financial commentators on TV have been quick to point out a logical inconsistency in the current AI market narrative. They argue that critics are talking out of both sides of their mouths:

1. The ROI Skeptics: “AI is a bubble; the infrastructure spend is massive, but the returns for clients aren’t showing up.”

2. The Disruption Alarmists: “AI is too successful; it’s going to automate software incumbents into obsolescence.”

The AI defenders argue both fears cannot be true. If AI doesn’t work (No ROI), it can’t disrupt incumbents. If it disrupts incumbents, it clearly works (High ROI).

However, the market is not a single person with one opinion. It is a collection of thousands of portfolio managers, each with different mandates. The finance literature and some prior examples supports the contention that two seemingly inconsistent views can drive stock price and investment decision.

There is no inconsistency in Group A selling Nvidia because they fear a capex bubble, while Group B sells Salesforce because they fear Agentic AI will replace seat-based licenses. When both groups act on their specific fears, the entire sector moves down.

The “horse and buggy” analogy (historically attributed to a leading banker at Michigan Savings Bank in 1903 who advised Henry Ford’s lawyer not to invest) provides the perfect heuristic.

Imagine it’s 1905:

• Investor A refuses to invest in Ford because the infrastructure (paved roads/gas stations) is too expensive and the ROI is decades away.

• Investor B sells their shares in Carriage-Maker Inc. because they fear the “horseless carriage” will make the current business model obsolete.

Both have a bearish view, but for opposite reasons. Their combined selling pressure creates a market-wide fear of the auto-sector, even though their reasons are technically contradictory.

What the Literature Says

This phenomenon is well-documented in finance literature: Edward Miller, in a paper published in the Journal of Finance, in 1977 argues that when uncertainty is high, divergence of opinion leads to massive volatility. Prices don’t reflect an average view; they reflect the battle between the most optimistic and most pessimistic agents.

Brock & Hommes (1998) in their work on Heterogeneous Agent Models (HAM) proves that a market full of bounded rational agents—some following fundamentals (ROI), others following trends or disruption narratives—create nonlinear price shifts that a rational observer would find inconsistent.

Yan Gao, Connie X. Mao, and Rui Zhong (2006) in their study, “Divergence of Opinion and Long-Term Performance of IPOs,” use Miller’s framework to explain why high-uncertainty often leads to volatile price corrections in IPO markets.

The bottom line is that contradictory fears don’t cancel out they can lead to additional selling by each group. In the current situation, it appears as though AI technology is a “bust” for the people building it and a “threat” to the people competing with it leading to the current tech selloff.

Authors Note: This multi-topic blog has articles on economic policy, (health care, student debt, taxes, and Social Security), politics (very closely following the contest for the House of Representatives), personal finance (both issues affecting people entering the workforce and people entering retirement) and today investments and markets. Most of the material on the blog is free and I am committed to keeping it that way but some material (usually only a portion of an article) is available exclusively for paid subscribers.

This post on the impact of the wide divergence of opinions on AI on the current tech market is free to all.

People who liked this blog may also want to read A Statistically Well Behaved Transformation of PE for Growth Value Inference.

I appreciate your readership and support.

Share

Subscribe now

· The AI Displacement: Crypto is losing “innovation capital” to AI infrastructure, which investors currently perceive as having higher social utility and productivity potential.

· The Energy Bottleneck: AI data centers are outbidding miners for electricity. As energy becomes a strategic resource, policymakers are likely to prioritize AI over crypto mining.

· Neutral Money vs. Digital Gold: While the “Neutral Money” bull case remains intact, crypto has failed to act as a safe haven during recent risk-off periods, trailing gold significantly.

· Structural Fraud Levels: Crypto’s “fraud-to-asset” ratio is significantly higher than traditional finance. While not a systemic threat to banks, it remains a primary barrier to institutional “core asset” status.

· The “Lost” Supply: Analysis suggests 15–20% of the Bitcoin supply is permanently inaccessible due to custody errors, creating a unique “digital burn” that impacts real-world liquidity.

Memo Summary

Is crypto a revolutionary financial layer, or a high-beta tech experiment currently losing its lunch to AI infrastructure (as a high-tech alternative) and Gold (as a low-tech safe haven)? This memo evaluates the “Neutral Money” thesis against a 50% market drawdown, rising energy competition, and the structural impact of industry fraud.

Authors Note: I believe in keeping the core analysis of Economic Memos open to everyone to help build a more informed financial community. However, the granular data and technical reference guides are reserved for my paid supporters.

Upgrade to a paid subscription to unlock the full Research Appendices, including:

· Appendix A: A credibility breakdown of the 5 different types of crypto (Bitcoin vs. Smart Contracts vs. Stablecoins).

· Appendix B & C: A complete “Registry of Risk” detailing major fraud events and the $85B+ in lost assets.

· Appendix D: The technical data on “Lost Supply”—why millions of BTC will never return to the market.

Thank you for supporting independent economic research.

Introduction:

Cryptocurrencies are digital assets that use cryptography and decentralized computer networks to record and verify transactions without relying on central authorities such as banks or governments. They operate on blockchains, distributed ledgers maintained by network participants, rather than by a single institution. This structure allows peer-to-peer transfers across borders and continuous settlement outside the traditional banking system.

The original aim of cryptocurrencies was to create a new form of digital money: a medium of exchange independent of central banking infrastructure. Over time, a second narrative emerged—that certain cryptocurrencies could serve as stores of value, which woud protect investors from potential inflation caused by inflation or currency debasement.

Crypto networks and blockchain technology are emerging as a digital alternative to the traditional financial system. Many of these applications are already live: they allow for peer-to-peer payments and near-instant international transfers that bypass traditional middlemen. Beyond just moving money, this tech uses ‘smart contracts’ to automate lending and borrowing, creates digital versions of real-world assets, and builds identity systems that don’t rely on a central bank or a private corporation to verify your data.

The objectives of this memo are to describe the bull and bear cases for crypto technology and evaluate reasons behind the current decline in crypto prices.

Share

The Bull Case for Crypto

The bullish thesis holds that cryptocurrencies represent early-stage monetary and financial infrastructure innovation.

The bull case for crypto as a new financial layer is based on the following pillars:

· Neutral Money: A global currency that operates by math, not politics—no central bank can print more of it or devalue your savings.

· Unstoppable Access: A financial “exit ramp” for anyone facing high inflation, frozen accounts, or strict government oversight.

· Digital Scarcity: Unlike “easy money” issued by governments, these assets have a hard-coded limit that acts as a long-term shield against inflation.

· Modern Plumbing: A 24/7 financial system that replaces slow, human-managed banks with fast, automated code that never sleeps.

· Protocol Growth: The opportunity to own a piece of the world’s next financial infrastructure, similar to owning a “share” of the early internet.

Crypto is economically measurable but not systemically dominant.

The strongest real-world use case is stablecoin-based transfers, which process trillions of dollars in annual volume and account for roughly 1–3% of global cross-border payment flows, depending on measurement. In certain emerging-market corridors and capital-control environments, stablecoins are meaningfully used for remittances and business settlement.

Within the crypto ecosystem itself, daily trading often reaches hundreds of billions of dollars, and total market capitalization has ranged around $1–3 trillion in recent cycles.

Relative to the global financial system, however, crypto remains small. Global equity markets exceed $100 trillion, global bonds about $130 trillion, global banking assets over $300 trillion, and foreign exchange markets trade more than $7 trillion per day.

The most accurate characterization is that crypto operates at meaningful scale inside its own ecosystem and has achieved low-single-digit penetration in selected global payment corridors. It has not displaced traditional finance at systemic scale, but neither is it economically negligible.

Prominent advocates include -- Michael Saylor, Cathie Wood, Balaji Srinivasan, and Brian Armstrong. Saylor has described bitcoin as digital gold superior to physical gold. Wood perceives Bitcoin to be a new asset class and that an increase in uses by institutions would substantially increase its value. Srinivasan has described Bitcoin as an apolitical borderless alternative to central banks. Armstrong has argued that crypto is about economic freedom and that it can expand access to financial services globally.

The Bear Case for Crypto

The bearish thesis argues that cryptocurrencies lack the core features that anchor long-term asset valuation.

They do not generate cash flows, pay dividends, or produce goods and services. In short, crypto lacks intrinsic value.

If considered currencies, they should behave like relative prices rather than compounding investments.

If treated as stores of value, extreme volatility challenges their stabilizing function. Sustained price appreciation relies primarily on speculative inflows.

If crypto markets grow increasingly interconnected with banks, asset managers, payment systems, or leveraged financial intermediaries, instability within crypto could transmit shocks into the broader financial system.

The lesson many skeptics draw from past crises is that financial innovation can outpace supervision, allowing risk to accumulate in corners of the system that appear peripheral—until confidence breaks. A sufficiently integrated crypto ecosystem could create institutions that are not necessarily “too big” in isolation but become too interconnected to fail without broader consequences. Critics believe the expansion of crypto could lead to the next 2008.

Prominent skeptics include two famous investors, Warren Buffett and Charlie Munger, and two prominent economists, Paul Krugman and Nouriel Roubini. Buffet and Munger compared Bitcoin to rat poison. Krugman compares the Bitcoin craze to the tulip bubble and considers the concept to be motivated by libertarian philosophy, a new subprime market tied to leverage and lax regulation. Nouriel Roubini argues that Bitcoin resembles a Ponzi-like speculative structure sustained by continuous inflows rather than intrinsic value, and that crypto markets are structurally prone to manipulation, insider advantages, and weak regulatory oversight.

Subscribe now

Recent Sustained Price Decline

Over the past several months, major cryptocurrencies have pulled back sharply from their late-2025 peaks:

Bitcoin peaked at around $126,000 in October 2025. As of early 2026, it has declined roughly 40–50% from that peak, trading near $65,000–$70,000, its lowest level in about a year.

Ethereum peaked near approximately $4,000–$4,200 in late 2025. By early 2026, it had fallen to roughly $1,900–$2,200, representing a drawdown of about 45–55% from peak levels and placing it back near price ranges seen roughly a year earlier.

Crypto markets have historically experienced large cyclical drawdowns. In prior bear markets, Bitcoin has fallen 60–80% from peak to trough (notably in 2017–18 and 2021–22), and Ethereum has repeatedly endured declines of 60% or more across multiple cycles. These swings are extreme compared to traditional asset swings, but no extreme compared to past swings in crypto prices. Additional decreases in crypto values extending the recent downturn are plausible and would be consistent with previous crypto winters.

Factors Impacting the Current Crypto Market

Portfolio reallocations from booming AI and Gold

From a portfolio construction perspective, crypto does not exist in isolation. It sits inside a broader “innovation / long-duration / speculative technology” allocation bucket.

Under classical portfolio theory, investors allocate capital based on expected return, variance, and correlation. C

Crypto is perceived as a high-beta, long-duration innovation asset—similar to early-stage technology equities. It competes for portfolio share directly with other frontier themes such as AI infrastructure, AI-driven software, robotics, quantum computing, and prediction markets.

Portfolio managers currently appear to prefer these other high-tech speculative opportunities to crypto. The growth in value of AI linked opportunities occurred simultaneously with a decline in both the value of software and Crypto.

When AI enthusiasm lifts technology broadly, aggregate portfolio exposure to tech risk may rise beyond target levels. Risk management constraints can then trigger deleveraging or diversification across the entire technology complex, creating pressure on all high-beta assets, including crypto.

At the same time, crypto has also competed with gold for “alternative monetary asset” status. Prominent advocates such as Michael Saylor have argued that Bitcoin is superior to gold, describing it as “digital gold” and even “the hardest money ever created.” The proposition was that a scarce, portable, programmable asset would outperform physical gold as a store of value.

However, during periods of risk aversion and portfolio repositioning, gold has often behaved more like a traditional safe haven, while crypto has traded more like a high-beta technology asset. To the extent that investors treat gold as defensive and crypto as speculative, the claim that digital gold would displace or outperform real gold in stressed environments has not consistently held up across recent cycles.

Saylor’s statement that digital gold is superior to actual gold has not aged well.

Energy Constraints, AI Competition, and Crypto

There is growing concern that rapid AI expansion could strain electricity supply.

Large AI data centers require enormous and continuous power, and utility companies in several regions have warned about rising demand. Crypto mining—especially Bitcoin—also consumes substantial electricity. When two fast-growing industries compete for energy, it raises questions about prices, grid stability, and long-term supply.

Energy is not just an industrial issue but a political one. Governments face pressure to manage electricity costs, meet climate targets, and maintain grid reliability. In that environment, policymakers may prioritize data centers tied to AI and industrial policy over crypto mining, which is often viewed as more discretionary. Rising power costs or political scrutiny of energy use could weigh more heavily on energy-intensive crypto models than on other digital industries.

Investors may discount businesses that appear vulnerable to higher long-term energy costs. In a world where energy is strategically constrained, competition between AI infrastructure and crypto mining could become an additional pressure point for crypto markets. I expect that the political and economic pressure to reduce energy use will be more intense on crypto than AI because AI is the more profitable industry and has demonstrated greater benefits to society in the form of increased productivity.

Fraud, Regulation, and Institutional Confidence

Fraud and governance failures have been recurring features of the cryptocurrency ecosystem.

High-profile cases frequently cited by critics include the collapse of FTX, multi-billion-dollar enforcement actions against Binance, the conviction and pardon of Binance’s founder, and repeated token pump-and-dump schemes and stablecoin failures such as Terra/Luna.

These events have reinforced concerns about insider concentration, weak disclosure, and exchange conflicts of interest. For many institutional investors, the issue is not simply isolated misconduct but whether governance standards across the ecosystem are durable enough to justify long-term capital allocation.

Regulatory responses have diverged across major jurisdictions.

China has taken the strictest approach, effectively banned domestic trading and mining while promoting a state-controlled digital currency.

Europe has implemented a comprehensive rulebook through the Markets in Crypto-Assets (MiCA) framework, imposing licensing, disclosure, and reserve requirements while permitting regulated activity.

The United States has taken a more fragmented path. Congress passed the GENIUS Act, establishing a federal framework for stablecoins with reserve backing, disclosure standards, and supervisory oversight. A broader market-structure bill, the CLARITY Act, aimed at defining when digital assets are securities versus commodities and clarifying SEC/CFTC roles, passed the House but stalled in the Senate amid disputes over stablecoin interest payments, surveillance concerns, and industry opposition.

Supporters of Crypto want less regulation to stimulate innovation. Critics warn that if crypto becomes deeply intertwined with banks, asset managers, or leveraged intermediaries before governance and valuation questions are settled, instability could propagate into core financial institutions—raising echoes of 2008-style contagion dynamics.

Concerns about fraud, disputes over regulation, and concerns about the impact of crypto on the future safety and soundness of the financial industry may slow crypto’s growth.

Conclusion

The bull case views crypto as early-stage infrastructure: a digitally scarce monetary asset and programmable financial network with low but growing penetration. The bear case views it as lacking intrinsic value or worse a scam and a Ponzi scheme with the potential to destabilize the financial system.

The recent price decline does not appear historically unusual by crypto standards. Prior cycles have featured 60–80% drawdowns, and current declines of roughly 40–50% from late-2025 peaks fall within crypto’s historical volatility range.

The downturn likely reflects a combination of capital reallocation toward AI-linked assets, competition with gold in risk-off environments, concerns about energy constraints, and persistent regulatory and fraud-related uncertainty. None of these factors alone explains price movements, but together they shape capital flows and required risk premia.

There is a substantial amount of fraud in this new sector. Fraud is not the source of the downturn but is likely to impede cryptos future growth and expanded adoption.

At present, crypto represents roughly 1–2% of global financial assets and has low but non-zero systemic penetration. Some diversified investors hold small allocations as an optionality play, inflation hedge, or high-beta innovation exposure. However, it is not yet universally treated as a required core asset class alongside equities, bonds, and real estate. Its role remains discretionary rather than foundational.

Subscribe now

Appendix A: Types of Cryptocurrencies and Relative Credibility

Read more

Over the past year, artificial intelligence has moved from being a technological breakthrough to becoming one of the central financial narratives in global markets. AI demand can be strong, adoption can be accelerating, and yet the entire technology sector can still sell off sharply on fears related to AI itself. Periods of broad technology selloffs make that tension especially visible.

At the core is a growing concern not about whether AI will be used, but about how it is financed, where risk sits, and how concentrated that risk has become. AI has pulled forward capital spending, raised long-term earnings expectations, and clustered investor exposure into a relatively small number of highly visible firms. When confidence shifts, markets often react not by sorting among individual business models, but by repricing the technology sector as a whole.

Software and cybersecurity stocks can fall on days where AI falls even though underlying demand for products could even increase if AI does not materialize.

In these episodes, price action is often less about current earnings and more about future margins, future capital intensity, and uncertainty about competitive structure and preferences of portfolio managers now in and the future.

Crucially, these risks are not distributed evenly.

Some firms are exposed to AI through direct sponsorship and infrastructure commitments. Others are exposed primarily through aggregate spending cycles. Still others participate in AI while keeping economic exposure relatively optional. When AI fear turns into a market-wide factor, these distinctions are temporarily overwhelmed, but over time they become decisive.

Two companies sit at the center of this dynamic: Microsoft and NVIDIA.

Microsoft’s AI exposure is concentrated and structural. It combines capital provision, cloud infrastructure, and product integration around a single external AI lab. That creates meaningful upside, but it also means that bad news about AI execution, governance, or monetization can translate directly into economic and strategic risk.

NVIDIA’s exposure is different. NVIDIA benefits from AI broadly, but its downside risk is tied to the pace and financing of AI infrastructure growth. When confidence in AI spending weakens because of financing conditions, capacity digestion, or portfolio risk reduction, NVIDIA is often affected even if long-term AI demand remains intact.

Other major firms, including Amazon, Google, and Meta, invest heavily in AI as well. They do so in ways that distribute risk differently through internal model development, diversified cloud platforms, or long-term infrastructure strategies that are less sensitive to any single outcome.

The analysis here on AI financial risk involve answers to seven questions.

The Seven Questions

1. How is NVIDIA helping OpenAI grow, and what is NVIDIA’s exposure to OpenAI outcomes?

2. How is Microsoft helping OpenAI grow, and what is Microsoft’s exposure to OpenAI outcomes?

3. Why does Microsoft have more concentrated downside risk than NVIDIA?

4. How does bad AI or OpenAI news propagate through Microsoft, NVIDIA, and the AI ecosystem, and who is economically forced to absorb the shock?

5. What firms hedge long-term AI risk for Microsoft and NVIDIA, and what risks do those hedges not protect against?

6. How do Amazon, Google, and Meta differ structurally from Microsoft and NVIDIA in their exposure to AI outcomes?

7. Why can the entire technology sector fall even when AI demand and underlying fundamentals remain strong?

A future piece will examine competition and governance dynamics among major AI labs, including OpenAI, Gemini, and Anthropic, and explain when those rivalries matter financially and when they are primarily narrative.

Subscribe now

Share

Questions and Answers:

Read more

A short memo clarifying the structure, purpose, and implications of Waymo’s reported private financing, and why comparisons to IPOs or other AI transactions can mislead.

Recent press reports indicate that Waymo is exploring a substantial private financing round, reportedly seeking to raise approximately $16 billion at an implied valuation near $110 billion.

https://www.reuters.com/business/autos-transportation/waymo-seeking-about-16-billion-near-110-billion-valuation-bloomberg-news-reports-2026-01-31/

The transaction is structured as a private capital raise intended to fund ongoing investment and scale-up activities and does not appear to involve liquidity for existing shareholders.

The purpose of this memo is to clarify the economic intent and governance implications of the reported financing, distinguish it from superficially similar large private transactions that may be influenced by liquidity or IPO considerations, and create background material on private financing for readers less familiar with these markets.

Takeaway: The proposed Waymo financing is a private growth-capital raise, not a liquidity event, intended to fund continued investment while allowing Alphabet to retain control, impose capital-allocation discipline, and share risk with a small number of outside investors.

• The size of the reported valuation does not imply that Waymo is preparing for an IPO or offering liquidity to existing shareholders.

• Large private financings at other technology or AI companies may reflect employee liquidity, early-investor exits, or IPO positioning; available evidence suggests Waymo’s transaction is primarily oriented toward raising new operating capital.

• Comparisons based solely on valuation headlines or funding-to-valuation ratios can be misleading, as private transactions vary widely in structure, purpose, and disclosure.

1. Nature of the transaction
   The reported fundraising effort is a private financing round, not an initial public offering. Waymo remains a privately held company and does not have and will not have publicly traded shares after the new financing. The transaction would involve negotiated investments from a limited number of institutional or strategic investors, rather than a public listing or the creation of a public float.

2. Scale of the valuation relative to private financings and IPOs
   An implied valuation near $110 billion would place Waymo among the most highly valued private companies globally and far above the valuation at which most companies go public. Median U.S. IPO valuations are typically in the low single-digit billions, with even large, mature technology companies often listing in the $10–30 billion range; IPOs near or above $100 billion are rare. Valuations of this magnitude in private markets are confined to a very small set of companies. Among recent examples, only a handful of private firms—most notably OpenAI and Anthropic in large AI-focused financings, and SpaceX through a mix of primary and secondary transactions—have been discussed at comparable or higher valuation levels. Direct comparison of capital-raised-to-valuation ratios across these transactions is inherently imprecise, as many such valuations reflect secondary sales or blended deal structures, whereas Waymo’s proposed financing appears to involve primarily new capital raised at a single priced valuation.

3. Purpose of the financing
   The primary purpose of the transaction is to raise additional capital to support capital-intensive investments, including fleet expansion, compute and infrastructure, safety validation, and operating losses associated with scaling autonomous driving operations. There is no indication that the transaction is intended to provide liquidity or cash-out opportunities for existing owners.

4. Meaning of the reported valuation
   The approximately $110 billion figure represents an implied valuation derived from the terms of the private financing. It reflects the price per share agreed upon in this round, extrapolated to the full equity value of the company. This is not a market-clearing public valuation and may differ materially from the valuation that would emerge in a public market offering.

5. Interpreting the $16 billion investment size
   In simple terms, dividing the new capital by the post-money valuation gives a rough sense of potential dilution. Using headline figures, $16 billion divided by roughly $126 billion (valuation plus new capital) would imply that new investors collectively receive on the order of low-teens percentage ownership. Actual ownership outcomes depend on detailed deal terms, including preferred equity features, liquidation preferences, and any concurrent internal capital contributions.

6. Alphabet’s role and continued control
   Alphabet is expected to remain the controlling shareholder after the financing. Reporting suggests Alphabet itself may provide the majority of the new capital, which would limit dilution of its ownership stake. As a result, strategic and operational control of Waymo would remain with Alphabet following the transaction.

7. Why raise capital externally rather than fund internally
   Alphabet has the financial capacity to fund Waymo entirely through internal capital transfers. The decision to pursue a formal private financing reflects considerations of capital-allocation discipline, valuation anchoring, and risk sharing rather than financial necessity. External investors provide a market-based reference valuation, share downside risk in a capital-intensive and uncertain business line, and impose governance and reporting structures that are not required for purely internal funding.

8. Role of outside investors and financing formality
   Once non-Alphabet investors participate, the transaction must be structured as a formal private financing, with standardized investment terms and investor protections. The presence of a small number of outside institutional investors therefore necessitates the legal and governance formality of a priced private round. This formality does not reduce Alphabet’s control but reflects arm’s-length co-investment rather than informal internal funding.

9. Absence of a public cap table
   Waymo’s detailed capitalization table is not publicly disclosed. As a private company, it is not required to release precise ownership percentages across Alphabet, prior venture investors, management, and employees. Public information is limited to high-level descriptions of ownership and control.

10. Relationship to a potential future IPO
    A private financing of this scale does not, by itself, indicate an imminent IPO. Large late-stage private rounds are commonly used to fund expansion while preserving strategic flexibility. A public listing would require explicit announcements, regulatory filings, and structural decisions that have not been publicly disclosed in Waymo’s case.

Conclusion: Overall, the transaction reflects a deliberate choice to formalize additional investment at an externally validated valuation rather than a shift in ownership strategy or a step toward public listing.

Subscribe now

Share