Buying the NASDAQ-100

Risk disclosure: The annualized returns of the various ETFs, the concept of the economic moat, the index structures, and so forth described in this chapter are all compilations of publicly available historical data, and do not constitute a recommendation to buy or sell any individual security.

In the capital markets, where everything changes in an instant, most investors tend to lose themselves in the tangle of technical analysis and financial statements, trying to earn excess profits by predicting where the market will go. Yet the true path to accumulating wealth is often hidden within the plainest, most unadorned discipline. “The great way is simple” is not merely a mode of operation; it is a complete investing system in which the mindset, the philosophy, and the tools are all highly unified.

To root out the anxiety of watching the ticker for good, you must undergo a brutal reshaping of identity, becoming “the boss of 100 bosses.” When you buy a NASDAQ-100 index ETF, you are not simply buying a single stock; you are, with one click, firing yourself from your own post as a trader, and in an instant becoming the substantive owner of 100 of the world’s top enterprises that change the world, such as Microsoft, Nvidia, and Apple. This means that the 100 smartest CEOs in the world are working around the clock, year-round, to earn money for this asset of yours.

Once you have confirmed this identity, the short-term rises and falls of the market are nothing more than the daily fluctuations of your employees’ operations. There is simply no need to feel anxious over price. This is the true composure of a capitalist. (Video 00542)

Many people think that buying QQQ means buying tech stocks, but they have utterly failed to grasp the true defensive mechanism of the index’s design: the deliberate exclusion of the financial sector. The profits of the financial industry rely heavily on the interest-rate environment and monetary policy, and its growth curve is bound tightly to that of traditional value industries. The NASDAQ-100, by contrast, pursues growth enterprises that possess pricing power and zero marginal cost. Excluding financial stocks is meant to prevent the interest-rate cycles of the traditional economy from interfering with the capital expansion of innovative enterprises. This is also why QQQ’s explosive growth power is more pure than that of the S&P 500; its source of growth is the productivity of innovative enterprises, not the cycle of bank interest-rate spreads. But note that the cash flows of growth assets skew toward the distant future, so they are still affected by interest rates and discount rates, and are not immune to interest-rate movements.

QQQ, the Strongest Asset-Growth Engine

The NASDAQ-100 index fund (such as QQQ in the United States or 00662 in Taiwan) is today’s most explosive growth engine. Starting from the same NT$1 million, if you buy endowment insurance, after 30 years you have about NT$2.1 million; if you buy government bonds, about NT$4 million; but if you invest in QQQ, after 30 years it will grow to NT$30 million to NT$50 million! This gap of as much as 25 times is precisely the yawning difference that compounding makes at different rates of return.

▲ Figure 8-1 The vast difference the same NT$1 million makes over 30 years, where whether you pick the right vehicle is the difference between heaven and earth in compounding at different rates of return (the rates of return are long-cycle historical estimated averages, not fixed guaranteed values)

This gap becomes more intuitive if you reverse it into “how much must you invest each month to accumulate your first US$1 million after 30 years.” Under the neutral parameter settings, QQQ at 12% annualized and VT at 7% annualized, to reach the same US$1 million goal over the same 30 years, a QQQ investor need only invest US$325 a month, while a VT investor must invest US$851 a month, fully about 2.6 times more. In other words, picking the wrong vehicle is not as simple as “earning a little less”; it means “being forced to invest more than two and a half times the principal every month” just to chase the same finish line. The price of choosing the wrong vehicle is using the most precious time and principal of your life to fill a gap that never needed to exist.

Why does this book use 7% for the VT projection? Because VT tracks the FTSE Global All Cap, the global stock market, about 60% US stocks and 40% non-US, and its long-term return is naturally lower than that of pure US stocks; moreover, Vanguard’s own model has already markedly revised its forward-looking expectations for the coming decade downward. A summary of the various definitions is compiled below for readers to weigh for themselves:

The S&P 500 (VOO / SPY / IVV) adopts the same valuation logic, and its long-term performance sits between global VT and NASDAQ QQQ:

The S&P 500’s nominal annualized return over nearly a century is about 10% (real about 6.5% to 7%); over the trailing 10 years, lifted by tech stocks, it is about 13.5% with dividends, still above the long-term mean, a product of a bull market with elevated valuations. Vanguard and J.P. Morgan even lower their forward-looking view for the coming decade to 4% to 6% (chiefly because of elevated valuations), which is why this book does not adopt the high trailing-10-year figure in its forward projection. (S&P 500 nominal historical data over nearly a century: Aswath Damodaran, NYU Stern School of Business, “Historical Returns on Stocks, Bonds and Bills: 1928 to present” dataset.)

The NASDAQ-100 (QQQ / 00662), as the strongest growth engine of the three, is likewise summarized across various definitions below:

QQQ’s return slope steps higher period by period (trailing 20 years 14.7%, trailing 5 years 20.4%), but this book takes a neutral 12% for its forward projection, below the trailing-20-year measured value and far below the trailing 5 years, leaving ample margin of safety for when the tech cycle turns. It is worth noting that the reason QQQ’s “since inception” annualized return is only about 10%, actually lower than 12%, is that when it was launched in March 1999 it was stuck right on the eve of the dot-com bubble peak, an extremely poor entry point. In other words, 12% is a conservative forward-looking assumption that sits between “about 10% since inception” and “about 19% over the trailing 10 years.” (QQQ returns for each period follow Invesco’s official standardized performance: since inception on March 10, 1999, annualized about 10.1%, trailing 10 years about 19%, total expense ratio 0.18%; data as of March 31, 2026, NAV with dividends.)

Taking all of the above together, all compound-interest projections in this book adopt a unified set of return definitions: QQQ / 00662 at 12% annualized, VOO / SPY / S&P 500 at 8% annualized, VT global stock market at 7% annualized; leveraged QLD / 00670L is not assigned a separate annualized rate (daily reset makes a single annualized figure inapplicable), and is uniformly expressed through a Beta of 2.0 together with historical backtests. After deducting 2% to 3% inflation, the real returns of the three are about 9% / 5% / 4%. Every forward-looking projection in this book uniformly adopts this neutral baseline; historical actual backtest data (such as SPY’s trailing-10-year 11.88% or QQQ’s trailing-20-year 14.7%) are uniformly labeled with their actual period, kept separate from the forward-looking assumptions, and never mixed.

Many people question whether QQQ’s trailing-20-year historical annualized return of 14.7% is too optimistic to sustain. Retail investors love to invoke the 2000 dot-com bubble to warn that today’s tech stocks are overheated, but this is complete ignorance of the structure of financial statements. In 1999 the average net profit margin of the S&P 500 was only 5% to 6%; at that time it was a pure bubble of high valuations plus negative earnings. But by 2025, the market had evolved into an extreme right-skewed fat-tail structure. This means that the market’s wealth creation has become extremely concentrated in the handful of tech empires that possess zero marginal cost. The probability of a market-wide bubble has fallen, but the concentration risk of a few companies holding the whole index hostage has grown higher.

This is precisely why you must hold QQQ heavily to enjoy the fat-tail effect. The trailing-20-year annualized return is 14.7%, and over the trailing 5 years it has soared to 20.4%. As the AI and computing-power revolution advances, the slope at which the tech giants create profits is climbing exponentially.

Therefore, this book’s forward projections uniformly use 12% as the baseline for calculation. Set against the measured 14.7% over the trailing 20 years and 20.4% over the trailing 5 years, this is not merely un-optimistic; it is an extremely conservative floor. You cannot use the tortoise-slow pace of the old economy to measure the speed at which tech hegemony harvests the wealth of the world.

That said, the dividend of the AI revolution will not be realized in a straight line, but released in three stages.

• The first stage is the hype phase, when the market discounts the imagination of the next twenty years into the present all at once, valuations briefly distort, and themes erupt in succession.
• The second stage is the bubble-correction phase, when excessive optimism is squeezed out, and even quality companies are wrongly hammered down.
• The third stage is the productivity-landing phase, when AI genuinely permeates processes such as coding, customer service, and legal work, materially raising labor productivity, and only then does true wealth appear, just as the steam engine did for the Industrial Revolution.

But the key point is this: someone in the midst of it can almost never tell in real time which stage they are standing in, and this is precisely the fundamental reason CLEC does not rely on prediction. The awakened capitalist does only one thing: buy whenever there is money and keep holding the NASDAQ-100 (QQQ / 00662), letting the cash and short-term bond positions built into the allocation (such as 00865B) hold the maintenance ratio during the correction stage so they are not forced into a fire sale, and then leaving a predetermined proportion of annual rebalancing to mechanically restore the allocation to its original position, rather than judging by feel where the bottom is and adding subjectively. The stages will rotate, but a person standing on the main channel need not guess which stage they are in; they need only refrain from getting off.

Whenever retail investors hear that “the Buffett indicator (total stock market capitalization divided by GDP) is too high,” a panic of “the market is in a bubble” wells up. But this indicator is itself built upon a mistaken mathematical assumption:

GDP is an increment, meaning however much is added this year, that is this year’s GDP; the GDP accumulated over a hundred years is not counted in this GDP. Does the wisdom accumulated by humanity over a hundred years simply cease to exist? So the stock market will always be higher than GDP, and increasingly so; in the future it may be a hundred times, a thousand times, ten thousand times GDP, because this thing accumulates. (Video 00564)

The key difference lies in “GDP is a flow, while the stock market is a stock.” GDP tallies the total new output added nationwide “within this one year”; the intellectual property, brand value, technology patents, and corporate competitiveness accumulated yesterday, last year, or ten years ago are not counted in this year’s GDP at all. But stock market capitalization reflects “the accumulated wisdom of all humanity”: Apple’s brand value has accumulated from 1976 to today, Microsoft’s Windows ecosystem from 1985 to today, Google’s search algorithm from 1998 to today. As human wisdom keeps accumulating and industrial competitiveness keeps strengthening, stock market capitalization should “always be higher” than that year’s GDP flow, and this ratio will stretch ever larger over time. Reaching 100 times or 1,000 times GDP in the future is not a bubble, but the mathematical inevitability of the thing called “accumulation.” Any academic indicator that judges a market bubble by “total market cap divided by GDP” commits the basic error of comparing a “flow” as though it were a “stock.” The composure with which the CLEC system can hold QQQ tightly and not move is built precisely upon this first principle: “human wisdom only accumulates; it is never reset to zero.”

▲ Figure 8-2 A comparison of the long-term annualized returns of QQQ against the S&P 500 and the global index (VT)

Viewed over a 40-year historical span, the S&P 500 index has an annualized return of about 8.95%, and the NASDAQ-100 index about 14%; even shortened to 25 years, the NASDAQ-100 index still has an annualized return of about 9.88%, while SPY has about 6.3% (this is on the index’s own basis; the 9.14% in Figure 7-5 of Chapter 7 is the realized annualized return of a lump-sum strategy starting in 1999, and is slightly lower because the starting window differs, so the two do not conflict). Compared further with a traditional target-date fund, the gap is even more pronounced: from 2008 to 2025, VTIVX’s total return was about 96%, while QQQ’s was about 985%. The point is not to pursue what “looks stable,” but to choose an asset that can magnify the slope of your net worth over the long term.

The Financial Definition of a Moat Company

Why is QQQ so strong? The answer lies not in its ticker, but in the fact that its constituents generally possess an extremely wide “economic moat.” The economic moat means this: this company’s ability to earn money far exceeds its cost of borrowing money. To break it down in the plainest terms, a business each year really only needs to answer two questions: “For every NT$1 I put into doing business, how much can I earn back?” and “For every NT$1 I borrow, how much interest must I pay?” If what it earns back is far more than what it must pay (for example, earning 20% and paying 5%), then this company is continuously creating value for its shareholders; if the gap is tiny or even inverted, then this company is actually running in place, or even slowly bleeding out. Apple, Microsoft, Google, and Nvidia, these core holdings of QQQ, have long belonged to the former; their money-earning efficiency exceeds their cost of financing by more than 10 percentage points, which means “no matter where they borrow money or from whom they raise capital, this group of companies can take the money and keep earning far, far more money.” This is why world-class ratings agencies (such as Morningstar) treat this spread of “money-earning ability minus cost of borrowing money” as the standard for measuring a company’s economic moat. Investing in QQQ amounts to handing your funds to a group of world-class management teams that are “extremely efficient at earning money, with extremely wide moats,” rather than betting on the individual luck of any single company.

The Organism-Like Metabolic Mechanism of the NASDAQ-100

Going a step further, this is also precisely why CLEC emphasizes that “the index itself is the highest form of active management”: when a company’s moat narrows and its money-earning efficiency can no longer keep pace with its cost of borrowing money, its market cap will naturally decline, and it will ultimately be automatically kicked off the list by the index’s market-cap-weighted rules, while newly rising high-efficiency enterprises are automatically added in. The investor need not pick stocks themselves; the market mechanism does the work of weeding out the weak and keeping the strong on your behalf, forever concentrating the funds in the strongest competitors of the moment.

Approaching from the common skepticism about diversification: QQQ is the top 100 NASDAQ-listed companies in the United States, while SPY is the top 500 companies by market cap in the United States. Although 100 companies are indeed not as diversified as 500, 100 companies are still far more diversified than picking individual stocks yourself. This is why directly holding QQQ already leaps entirely beyond the level of “individual-stock concentration risk.”

The teachings of the CLEC system have mentioned many times: if an investor worries that holding only QQQ is not diversified enough and too volatile, they may also pair it with SPY (VOO), but they must understand that QQQ’s long-term performance still beats SPY. The point is to strike a balance between risk and return and find an allocation you can hold at ease for the long term. (Compiled from Video 00405, 00505.)

There are two key contexts in this passage that need clarifying:

• Teacher James’s statement that “pairing with VOO is not out of the question” is not the standard allocation of the CLEC system, but a psychological compromise offered to those “with extreme fear of a single index.” The mathematical price remains clear: adding VOO / SPY amounts to actively diluting QQQ’s excess return, sacrificing long-term performance in exchange for the psychological comfort of “looking more diversified.”
• The disciplinary direction of the CLEC system is the three elements of “using QQQ / 00662 as the underlying core, paired with leverage and cash,” rather than splitting the underlying position into two halves of QQQ + SPY.

In other words, this supplementary option is merely a back door left for investors who “genuinely cannot sleep,” and is not the main path. Investors are advised to first honestly confront whether their fear of QQQ comes from “a true risk preference” or from “unfamiliarity with volatility”: the former may consider a QQQ + SPY dual-engine setup, while the latter should build composure toward QQQ and volatility by reading the rest of this book, rather than escaping learning by diluting returns.

Let the data speak: over the long term the NASDAQ-100 index has an annualized return of about 14%, the S&P 500 about 9%, and the globally allocated VT about 7%. First see clearly the essence of VT: it is about 60% US stocks and 40% non-US, a market-cap-weighted passive product, not a bad product in itself; the problem is that it is “only half right”: so-called global diversification, mathematically, dilutes the high-efficiency growth of US stocks with the mediocre assets of non-US markets. Traditional finance always teaches you to diversify and buy the whole world, but according to classic academic research by scholars such as Elton and Gruber, “Risk Reduction and Portfolio Size: An Analytical Solution,” and Statman, “How Many Stocks Make a Diversified Portfolio?,” once holdings exceed 50 names, the marginal benefit of risk reduction approaches zero, and it instead drags on performance because funds are averaged out across low-return targets. VT buys nearly 9,000 stocks worldwide, which amounts to forcibly allocating funds to thousands of mediocre enterprises on the verge of being eliminated.

What is more, over 40% of the revenue of the S&P 500 and NASDAQ-100 constituents comes from global markets outside the United States: Apple sells phones in China, Microsoft sells cloud in Europe, Google’s ads span the globe. Buying the core US stock index amounts to buying the global economy and demographic dividend in the most efficient way, with no need to allocate any other country’s index for “fake diversification.” Buying the top US index has never been “putting all your eggs in one basket”; it is directly buying the most productive half of the entire world.

▲ Figure 8-3 A comparison of the annualized returns of QQQ / SPY / VT

Source: Yahoo Finance

The NASDAQ-100 handpicks the 100 most pricing-power-rich innovative enterprises, and has already perfectly reached the limit of risk diversification; buying VT is not hedging, but paying the bill for the world’s mediocrity. Even the banking system certifies this kind of strong asset: in Taiwan, a strong target like 00662 can usually obtain more favorable pledging interest rates in the market (for instance, Yuanta Securities Finance once had a program starting from 2.48%), better than ordinary targets; the actual rate varies with the securities finance company, the brokerage program, the collateral, and personal credit terms, and must be based on each institution’s latest announcement. This also reflects the high unity, in both philosophy and tools, of placing funds in the strongest market. It becomes even clearer if you look at absolute scale: according to the 2023 statistics of the World Federation of Exchanges (WFE), the two largest stock exchanges in the world by market cap, the New York Stock Exchange (about US$25 trillion) and NASDAQ (about US$21 trillion), are both located in the United States, and the US stock market alone accounts for nearly half of the total market cap of the world’s stock markets, larger than the next thirteen countries combined (Shanghai in China, ranked third, is only about US$6.6 trillion, not even a third of NASDAQ alone).

▲ Figure 8-4 The market cap of the world's top nine stock exchanges (2023), where the two largest by market cap are both located in the United States

Source: World Federation of Exchanges (WFE), end of September 2023

The Index Weeds Out the Weak and Keeps the Strong

“The index weeds out the weak and keeps the strong” is not an abstract slogan, but something that can be verified with concrete cases. Take a look at the top 10 constituents of the NASDAQ-100 in recent years. The actual list and rankings float with market cap and must be based on the official latest fact sheet, but over the long term they are highly concentrated in large platform, cloud, semiconductor, and innovation giants such as Nvidia, Apple, Microsoft, Amazon, Broadcom, Meta, Tesla, and Alphabet, most of which are direct or indirect beneficiaries of the AI, cloud, semiconductor, or digitalization trends, and the top 10 together account for about 46% of the index weight. Yet two or three years ago, the top ten list still included “Costco,” a traditional retail giant. As the AI wave pushed up the market cap of computing-power and semiconductor enterprises such as Nvidia and Broadcom, Costco has already been naturally squeezed out of the top ten. The investor need not actively sell Costco, nor actively buy Nvidia; the market-cap-weighted index structure completes this “paradigm shift” automatically, forever concentrating funds on the strongest competitors of the moment.

Many investors will ask: “Since the top ten account for 46%, wouldn’t buying the Magnificent 7 (M7) or FANG+ directly be even more concentrated and higher-return?” The answer is no, for two reasons.

• First, the fewer the constituents, the greater the volatility; the damage to your account from a single company’s negative news (such as a CEO departure, a product-line failure, or an antitrust investigation) is magnified in proportion.
• Second, “you do not know which one will be kicked out”: today’s top 7 may not still be in the top 7 five years from now; Costco was still in the QQQ top ten a few years ago, and has now fallen off the list.

The index automatically weeds out the weak and keeps the strong, but a “hand-picked 7-plus combination” does not; once one company in the combination begins to decline, the investor must decide for themselves when to sell, when to switch, and when to add, which is precisely the “active stock-picking” trap that the CLEC system repeatedly stresses must be avoided. QQQ is already the “best concentrated version” protected by the weed-out-the-weak mechanism; shrinking it further to M7 not only takes on more risk, but also takes on all the decision cost of active judgment.

Wall Street is forever inventing new toys to harvest retail investors. Recently the market has been touting the so-called NASDAQ-100 enhanced index (such as 009820), claiming to eliminate the bottom-ranked companies through multiple factors. This is a classic case of outsmarting oneself. The essence of index investing lies in fully embracing the market’s unknowns. When you eliminate the bottom-ranked through artificially set financial metrics, you very likely miss the disruptive innovation launched by small and mid-cap companies. Setting down the greed of trying to beat the market and embracing pure market-cap weighting is the highest form of wisdom.

Regarding 00757 (President FANG+ ETF): the CLEC system once accepted a limited allocation “with the dimensional strike controlled within 10%” in its early days, but starting in 2025 Teacher James has formally recommended “transferring all of 00757 to 00662.” The core considerations behind the transfer recommendation are summarized below:

The core conclusion is that this design of “equal weight plus few constituents” mathematically “pulls the flowers and plants the weeds”: at each rebalancing it sells off the position in the rising good companies and subsidizes the falling poor companies, running counter to the CLEC system’s first principle of “weed out the weak, keep the strong, let the winners run.” For investors who originally hold 00757, it is advised to gradually convert to 00662 during the annual rebalancing period; for investors who do not yet hold it, simply start allocating directly from 00662, with no need to take the detour of 00757. This transitional path of “dimensional strike control on extreme assets” can be formally closed.

The Positioning of Taiwan-Market Cap-Weighted and 2× Leveraged Products

| Metric | 00662 / QQQ | 0050 (market-cap-weighted) | High dividend (0056 / 00713) | |—|—|—|—| | Long-term annualized return (CAGR) | 14% to 20% | 9.5% to 14% | 6% to 9% | | Historical maximum drawdown (MDD) | -83% (2000 dot-com bubble) | -58% (2008 financial crisis) | -25% (general crash level) | | 2025/4 crash measured | -15% (US stocks strongly resilient) | -25% (concentration pressure of Taiwan’s heavyweight stocks) | -12.8% (00713 once hit limit down) | | Core value positioning | Source of excess return | Regional economic growth | Cash-flow tool |

Taiwan investors also have a set of local index tools on hand, and you must first clarify their positioning relative to the core QQQ / 00662, so as not to mistake an auxiliary tool for the main axis.

On the market-cap-weighted Taiwan-stock side, 0050 (Yuanta Taiwan 50) and 006208 (Fubon Taiwan 50) track the same “top 50 Taiwan companies by market cap” index, differing mainly in the issuer and the internal fee (006208’s expense ratio is usually slightly lower). Although both have excellent long-term returns, 0050’s TSMC weight in recent years has approached 60%, about 58% (varying by the latest holdings date), and the single-point risk has already strayed from the original intent of diversification.

This concentration can be quantified with the “effective number of constituents (EN),” which converts the actual weight distribution into “equivalent to how many equal-weighted companies.” 0050’s EN is only about 2.8, meaning that although it nominally holds 50 names, the risk it actually bears is like betting on fewer than 3 companies; even the EN of Taiwan’s broad market index itself is only about 6.2 (TSMC’s weight is about 40%).

The key mechanism is this: when a single company’s weight is as high as about 60%, no matter how many names, even tens of thousands, the remaining 40% is spread among, the EN can barely be moved. This is also why 0050, though it bears the name of “index,” is in essence already close to actively over-betting on TSMC. (The effective number of constituents EN is estimated by the reciprocal of the Herfindahl index HHI, 1/Σw_i²; the above EN figures are cited from the article “‘Buying 0050 = buying the broad market?’ 60% bet on one company is not called diversification” in “A Plain Talk on Insurance Concepts,” calculated from weight data as of 2025/08/06, a third-party estimate, not an official diversification indicator from the index company.)

By contrast, applying the same yardstick to US stocks: the NASDAQ-100’s EN is about 26, the S&P 500 about 51, and the Dow about 22, all far higher than 0050’s 2.8. This too confirms with numbers that QQQ / the NASDAQ-100 is the true core that genuinely spreads risk, while 0050’s “diversification” does not live up to its name.

▲ Figure 8-5 Concentration seen through the effective number of constituents (EN), where 0050 and Taiwan's broad market index are far less diversified than the NASDAQ-100

An honest, era-appropriate correction is due for 0050 / 006208: they are not the equivalent of the NASDAQ-100, but a mixed index of “Taiwan local tech + finance + traditional,” and originally do not belong to the CLEC main-axis “locomotive” targets. However, during this wave of AI supply-chain prosperity from 2024 to 2026, Taiwan happens to stand on the most critical hardware links (wafer foundry, server assembly, ABF substrates), which has made the market-cap-weighted Taiwan-stock funds the “first carriage” behind the locomotive in the short term, temporarily keeping pace with QQQ. Within this window of AI capital expenditure (estimated to last another 10 to 20 years), a small allocation as an auxiliary is acceptable, but you must be clear that this is merely “tactical tolerance limited to the AI cycle,” not an option to replace core QQQ over the long term. Once the AI cycle turns, market-cap-weighted Taiwan-stock funds will return to the original pace of a local mixed index, retreating from the “first carriage” back to a “middle carriage.”

As for 2× (double-leverage) tools, common in Taiwan are 00631L (Yuanta Taiwan 50 2×, Taiwan-stock leverage) and 00670L (Fubon NASDAQ-100 2×, equivalent to the Taiwan version of QLD, corresponding to the leveraged locomotive of the core QQQ). The 2× is an “attack engine” rather than the core itself; it pursues a daily return twice that of the underlying, its volatility is magnified, and long-term sideways movement produces compounding decay, so it must always be paired with a sufficient cash position and rebalancing discipline to be harnessed, and must never be bet on alone, still less used as naked leverage. When picking a Taiwan-stock 2×, there is also a reminder related to par value: do not be seduced by the cheap par value that becomes smaller after a split, for par value has no bearing whatsoever on growth potential; what you should truly compare is scale and internal fees. The details of each fund’s expense ratio, the spread against Taiwan index futures, and other stock-selection points are collected in the appendix “A Complete Guide to Taiwan-Stock 2× ETFs.”

To sum up the CLEC iron law in one sentence: the core position is always QQQ / 00662; market-cap-weighted Taiwan-stock funds (0050 / 006208) can only serve as a secondary auxiliary limited to the AI cycle, and the 2× (00670L / 00631L) can only serve as an attack position carrying cash for defense, and neither may ever exceed the core proportion. For Taiwan investors who have already over-weighted 0050 within the AI cycle, it is advised to slowly shift the center of gravity of funds back to 00662, to avoid being trapped in a middle carriage when the cycle flips.

00662 Need Not Wait for a Cheap Par Value

The most foolish behavior on the investment road is missing out on millions of future growth (dropping the big watermelon) in order to quibble over a few dollars in today’s higher stock price (picking up a small sesame seed). Many retail investors, seeing a single lot of 00662 approaching NT$100,000, balk at the price and cannot bring themselves to buy, and turn instead to buy some variant ETF issued at only NT$10, which is a complete inversion of priorities born of the poor person’s mindset. Whether the par value is NT$10 or NT$100, as long as the broad market falls 10%, the asset shrinks by the same 10%. More importantly, in order to meet market demand, 00662 has already prepared to carry out a split. Once the split is complete and the unit price drops, you can just as well buy at a cheaper price. Do not compromise the core strategy of investing in humanity’s strongest tech giants for the sake of an illusory cheap par value. When you clearly foresee that decades from now this one lot of stock will be worth tens of millions, will you still hesitate over whether it is NT$89 or NT$90 right now? As long as the trend is forever upward, any “historical high” of today is, in the future, merely a trivial line on the horizon.

Buying and selling in the short term, earning only that little bit of spread, that is called being short-sighted; what you earn is sesame-seed money. (Video 00530)

The true power of this system lies in using asset allocation and rebalancing to hold up the position, so that even in the face of an extreme market of -80%, you can still sleep soundly and hold the winning hand. This is a system that can ignore the market’s noise, letting the investor genuinely achieve a leap of wealth across class.

Here first is a concept of the risk floor: the true key to investing lies not in “winning every year,” but in “not being knocked out in extreme years.” When you see Taiwan stocks surge while the US stocks in your hand lie flat and consolidate for half a year, you often itch to sell the US stocks and chase the Taiwan stocks. This is the highway-traffic-jam blind spot that retail investors most often fall into. You turn the wheel and cut across, only for the original lane to start racing ahead while the lane you cut into jams solid, and you lose on both sides. Consolidation is actually the market accumulating momentum; as a long-term capitalist holding the broad market index, you must keep your hands under control and resist the urge to frequently change lanes.

When the market suffers consecutive heavy blows, the defensive position and rebalancing discipline will determine whether you can survive to the next round of rises. The first priority of investing is not to pursue the prettiest curve, but to ensure you always remain in the market.

With an ETF you do not need to calculate value; there is no way to calculate it. You are just buying the whole of America, and you simply never sell it. (Video 00404)

The Ten-Year Class Divide Between QQQ and VT

Even with the same “buy the index, hold for the long term,” whether you choose QQQ or the globally diversified VT leads to a completely different class ten years later.

Over the trailing 10 years, QQQ has 18.17%, SPY only 11.8%, VT only 7.8%; QQQ is more than double VT. QQQ is the tycoon, SPY is the rich man, VT is the poor man; this is a difference between heaven and earth. It is not the case that casually picking any index fund and investing long-term will make you rich; get the direction wrong and you will go to hell. (Video 00526)

To project the future, you must first have an empirical foundation. The following data uses the AQR factor model in the Portfoliovisualizer tool to run a regression analysis, quantifying the return structure and excess return (Alpha) of QQQ and VT, and the measured data produced is as follows: QQQ annualized 18.5%, model explanatory power R² reaching 91.2%; VT annualized 9.3%, R² reaching 94.9%.

▲ Figure 8-6 AQR factor model regression analysis, where QQQ's excess return does not come simply from bearing risk

Source: AQR Global Factors (analysis period: 2010 to 2024, 180 months in total)

The “AQR regression result” first sets the direction clearly: within the regression window of 180 months from 2010 to 2024, QQQ’s Alpha is a positive 3.84%, while VT’s is a negative 2.22%, and the gap is not luck but a quantitative result. The “factor exposure comparison” then further breaks apart the factor-exposure structure of the same set of data, contrasting whether the market factor, the value factor, or the growth factor drives the “return gap.”

▲ Figure 8-7 AQR four-factor exposure comparison (QQQ vs VT), confirming that QQQ heavily weights high-growth, low-book-value tech leaders

Source: AQR Global Factors (2010 to 2024)

The conclusion of the “factor exposure comparison” is that QQQ’s outperformance is not luck, but comes from higher market sensitivity and a clear growth-style exposure. Through factor decomposition it can be found that QQQ’s loading on the “market (MKT)” factor is about 1.08, higher than VT’s 0.95, meaning it can extremely sensitively capture the broad market’s rise; while its loading on the “value (HML)” factor is a deeply negative −0.48 (VT is only +0.05), reflecting the essence of its growth assets. This means that in a cycle of productivity expansion, this system can provide explosive power far exceeding the mean.

QQQ’s standard deviation is about 18.9%, with a maximum drawdown of 32.6% over the range since 2010; its volatility is indeed higher than SPY’s, but in the universe of long-term compounding, this kind of volatility is the price paid for excess return, not risk. As long as the system does not break, time and the productivity of the AI revolution are the greatest economic moat.

The true outcome is decided by systematic allocation discipline, not by simple gambling.