"The most important quality for an investor is temperament, not intellect." --Warren Buffett

THE GRAND ILLUSION (September 14, 2025): Once every few decades, far too many investors become convinced that if they throw a ball into the air with enough strength, it will go into orbit. By many measures, large-cap U.S. stocks have never been more overpriced relative to corporate earnings not only in the entire history of the United States, but for any country in the world at any time. Many investors have convinced themselves that the U.S. stock market only goes up, and have therefore put a large percentage of their retirement funds into the same stocks that just about everyone else has been buying. Regardless of how high these prices climb, or when they drop, the eventual inevitable regression to fair value and beyond will be an unpleasant event for most people who have forgotten the lessons of past bubbles and are convinced that because of one reason or another, "it's different this time."

Many stocks have proven intrinsic value; the problem is that in many cases today the fair value is far below the current price.

Costco (COST) was recently trading at several times its fair value, most analysts upgraded the stock for one reason or another, usually saying the equivalent of "you should buy it because it keeps going up." The price-earnings ratio eventually reached 62.5. I wouldn't touch Costco at or above one thousand U.S. dollars per share, but if its price went below 100 then it would become a worthwhile bargain and I would gladly accumulate it. This would be especially true if top corporate insiders were to become substantial buyers rather than selling at their heaviest pace ever recorded as they had been doing during the past year.

Any stock is a bargain whenever its valuation is substantially below what Benjamin Graham or Peter Lynch have described in their formulas as representing fair value. Earlier this year, especially in April 2025, we had numerous stocks in various sectors including energy and emerging markets which were trading near or below half fair value, and we accumulated as many of them as possible including RIG, WTI, EWY, and FLBR. During the past year we also had worthwhile undervaluations for U.S. Treasuries and TIPS including the highest yield for 30-year TIPS last month since 2001. Palladium was a compelling bargain at all-time lows in real terms and all-time bullish traders' commitments, which we repeatedly purchased via PALL. As is the case with all undervalued assets, they will generally rebound more than most other assets in percentage terms.

If you own undervalued shares then it is less necessary to sell them following a rally, since they will eventually rally again. If you own overpriced shares then one day a rally will be followed by a dramatic percentage decline in order to regress toward fair value.

More importantly, since underpriced assets have to climb rather than falling to reach fair value, it is less necessary to sell them whenever they have recently rallied. If an undervalued asset retreats, it will eventually recover because it is less than its intrinsic fundamental worth. Ono the other hand, if you own an especially overpriced asset which insiders have been aggressively selling, then you will eventually lose a large percentage of your investment. You just don't know when or how.

Cryptocurrencies have no proven worth and will end up collapsing even more in percentage terms.

It doesn't matter how many people recommend cryptocurrencies or who they are, since a cryptocurrency has no intrinsic value. Unlike a stock, it doesn't represent part ownership of a company, and unlike a bond, it doesn't represent a promise to repay principal with added interest. It is only worth whatever a bunch of other people, most of whom are among the least experienced investors, believe it is worth in their minds. It is like owning part of a fantasy, except that you have to pay real money for it. Perhaps there is some intrinsic value for someone who has to conceal their transactions, such as the black market, but whatever this value may be is far below the current valuations of nearly all cryptocurrencies.

Eventually cryptocurrencies will become nearly worthless. As with all other bubbles, it is unknowable in advance how or when this will occur, but it must happen one way or another. You don't want to get stuck holding the bag.

The long-term ownership of U.S. stocks as a percentage of total household assets has been at a median of 26% for decades, and recently climbed to more than double that long-term average.

The percentage of total U.S. household assets invested in the U.S. stock market briefly surpassed 51% in early 2000 before plummeting again. Recently over 53% of total U.S. assets were invested in U.S. stocks, thereby surpassing their previous 2000 peak. It is just a matter of time before the 26% level is reached again. There was a popular myth in 1999-2000, recently revived, in which investors "had to" put their money into the U.S. stock market because there was no alternative. As soon as the U.S. stock market, especially the most popular funds, retreated 30% or 40% in 2001 and 2002, and again several years later, investors had no trouble finding numerous alternatives including "boring" U.S. Treasuries which saw their yields plummet after 2000 and after 2007.

When any asset is rising in price, investors will do whatever they can to own it. When it is falling in price, they will similarly go out of their way to get rid of it. We had all-time record outflows at each of the past several bear-market bottoms for the U.S. stock market. Even the brief plunge in February-March 2000 experienced two consecutive weeks of the biggest-ever weekly net outflows from U.S. equity funds.

No asset has its fair value changed because a particular group of analysts like it or hate it, or due to any other popularity contest.

The price of any asset will often fluctuate, sometimes hugely, whenever analysts either love it or hate it. However, popularity has no effect on the fair value of any asset. That is based entirely upon current and future earnings for a stock, and current and future income payments for a bond. The best time to purchase any stock is when it is the most undervalued relative to its future earnings, and almost everyone is telling you why you shouldn't buy it. The best time to sell any stock is whenever it is most overpriced relative to its earnings, and almost everyone is eagerly purchasing it.

The U.S. stock market usually tracks the real U.S. economy, but lately this ratio has become absurdly out of line even when it is compared with other bubble periods like 1999-2000.

Large-cap U.S. stocks in particular have far outpaced the U.S. economy, thereby creating a temporary massive gap which will have to be resolved either by U.S. stocks slumping in price or the U.S. economy suddenly tripling or more in value. Guess which will occur:

As measured by another useful fundamental indicator, price-to-book for the S&P 500 Index surpassed its previous all-time record from March 2000:

For the first time in history, price-to-sales exceeded 10 for one-third of all U.S. companies, versus only two-thirds as many during the 1999-2000 internet bubble which was the previous record:

Real earnings yield in the U.S. recently fell to its lowest point ever measured:

Like most of the world's most experienced investors, Warren Buffett made all-time record stock sales and all-time record U.S. Treasury purchases roughly since the middle of 2024. Naturally most investors have foolishly concluded that Buffett's amply proven track record is meaningless and that he must be getting senile, since Berkshire Hathaway has been trading near its all-time record discount to the U.S. stock market:

Gold is not currently a viable alternative, and neither are gold mining shares. Some of the biggest drops in history for this sector occurred simultaneously with bear markets for large-cap U.S. stocks, because bubbles for one asset class usually inspire bubbles in other asset classes.

GDX, the most popular fund of large-cap gold mining shares, plummeted 72.1% from its March 2008 top to its October 2008 bottom. Investors who expected precious metals to be a safe haven from a retreating stock market discovered that they lost more money with gold mining shares than with many other sectors. A similar percentage drop occurred prior to the mid-November 2000 bottom for this sector. The traders' commitments and insider data show that the most experienced investors have been selling gold and silver while central banks and individual speculators have been piling in.

Central banks consistently buy gold near tops and sell gold near bottoms, and will always do so.

At the all-time bottom for gold in real terms in the late 1990s, the Bank of England sold all of its gold. The Bank of Canada sold most of its gold below 1100 U.S. dollars per troy ounce near the end of 2015, the last time that gold was so cheap. Now that we have all-time highs, many central banks have been accumulating the yellow metal. The result will be the same as when they had last piled in during the late 1970s and early 1980s; afterward, gold slid from 850 in January 1980 to 250 in August 1982.

Fortunately there is a simple signal for when you should be purchasing GDX, GDXJ, and other funds of gold mining and silver mining shares.

Here's the "secret" to buying precious metals: wait for silver's traders' commitments to show that commercials, who are those who own actual silver including miners, jewelers, and fabricators who make things from real silver, have a combined long position which is near or higher than their combined short position. As of the most recent weekly reading, silver commercials had combined longs of 40,163 contracts and combined shorts of 113,565 contracts. That is definitely not anywhere close to a buy signal.

On September 6, 2022, silver commercials were long 55,823 and short 50,768. That was a strong buy signal. Compare what GDX and GDXJ have done from then until now.

Investors are obsessed with unknowable data such as how extreme any especially overpriced asset will become, or when, rather than the percentage it will eventually have to lose.

The definition of an intelligent Dutch investor in 1640 was someone who avoided the lure of Tulipmania. In 1723, a smart U.K. investor was someone who didn't participate in the South Sea Bubble. More recently, brilliant investors in 1975 were those who avoided the siren song of the Nifty Fifty in the early 1970s. In the 21st century, an insightful investor in 2002-2003 was defined as anyone who didn't chase after the most popular stocks in 1999-2000. In 2028, an intelligent investor will be described as someone who resisted the siren call of the AI bubble in 2025.

There is no need to change my long-term outlook, since it is based upon proven fair-value principles.

I am frequently asked if I have revised my long-term price targets due to the bubble situation. People expect me to say that I no longer expect QQQ to drop 83.6% as it had done from its peak on March 10, 2000 to its bottom on October 10, 2002, or that I don't think that GDXJ will retreat all the way down to 28. I have retained these and all of my other targets, and am tempted to expect even further percentage losses, since millions of new investors have piled into these and other assets only because they're chasing after what everyone else is doing. The more inexperienced people who crowd into anything, the more that the same people will panic out of the same assets after their losses have become emotionally intolerable. This is why the biggest percentage moves in one direction are followed by proportionate shifts in the opposite direction.

My subscription service includes two 70-minute participatory Zoom meetings each week. In addition to answering questions in real time, I log onto my brokerage account and show subscribers my purchases and sales, as well as my current open orders.

Disclosure of current holdings:

Below is my current asset allocation as of 4:00 p.m. on Friday, September 12, 2025. Each position is listed as its percentage of my total liquid net worth.

I extracted the totals for each position and grouped these according to sector.

The order is as follows: 1) U.S. government bonds; 2) shorts; 3) bear funds; 4) precious metals; 5) emerging-market funds; 6) individual Brazilian ADRs; 7) energy; 8) other individual shares.

VMFXX/TIAA Traditional, TIAA money market/bank CDs/FZDXX/FZFXX/SPRXX/SPAXX/BPRXX/Savings/Checking long: 35.42%;

17-Week/52-Week/26-Week/13-Week/2-Year/8-Week/3-Year/5,10,30-Year TIPS/4-Week/6-Week/20-Year: 25.73%;

TLT long: 11.21%;

I Bonds long: 3.88%;

PMM long: 0.01%;

XLK short: 34.11%;

QQQ short: 25.12%;

GDXJ short: 1.52%;

SMH short: 1.43%;

GDX short: 0.24%;

AAPL short: 0.15%;

PSQ long: 1.24%;

SARK long: 0.38%;

Gold/silver/platinum coins: 10.55%;

PALL long: 2.62%;

EWZ long: 0.52%;

FLBR long: 0.51%;

EWY long: 0.11%;

FLKR long: 0.06%;

TUR long: 0.02%;

EWZS long: 0.01%;

UGP long: 0.42%;

VALE long: 0.30%;

GGB long: 0.17%;

BBD long: 0.15%;

RIG long: 0.44%;

WTI long: 0.07%;

PTEN long: 0.04%;

OGN long: 0.25%;

CLF long: 0.01%.

"Our advantage, rather, was attitude: we had learned from Ben Graham that the key to successful investing was the purchase of shares in good businesses when market prices were at a large discount from underlying business values." --Warren Buffett

BUY PALLADIUM, SELL GOLD (April 3, 2025): The media have been full of positive stories about why you should buy gold and related assets. Many analysts who had been bearish or indifferent toward gold a year ago have recently jumped aboard the bullish bandwagon. Almost all analysts' recent price targets talk about gold reaching 3,500, 4,000, or some higher number, with the only debate being when. Hardly anyone is talking about gold dropping to 2800 which is a modest pullback and the same distance away from 3150 as 3500 is. A tiny number of people have mentioned gold dropping to 1,820, but they say it will happen in five years which is an eternity with investing. Meanwhile, gold mining shares and their funds including GDX and GDXJ have been struggling relative to gold bullion, while the traders' commitments show commercials in the highest percentiles of short:long ratios for gold, silver, and platinum.

Almost no one has been mentioning palladium which is by far the most depressed precious metal in recent years. The price of PALL, a fund of physical palladium, slumped from a top of 298.21 on March 8, 2022 to a bottom of 76.49 at 8:30:48 a.m. on August 5, 2024 which is a loss of 74.35%. The traders' commitments for palladium approached all-time record bullish extremes, with the most recent reading showing commercials long 9,082 contracts and short 1,662 for an amazing ratio of 5.4645 to 1 long:short.

Click here for a useful Seeking Alpha article by Andrew Hecht about palladium.

Gold mining shares and their funds including GDXJ have been dramatically underperforming gold bullion as they consistently do prior to large percentage declines.

If you only saw a ten- or fifteen-year chart of GDXJ, a fund of mid-cap gold mining and silver mining shares, then you would probably conclude that the price of gold had been very high in past years and had been making lower highs in more recent years, because that is how GDXJ has behaved. Gold has gained more than one thousand dollars from its 2020 peak to its 2025 top so far, while the highest that GDXJ could climb recently was 58.595 on March 28, 2025 which was notably less than its 65.95 multi-year peak from August 5, 2020. Most of the precious metals media have been insisting that the shares of gold producers will catch up with the price of gold, but historically the shares tend to consistently lead bullion in both directions. In 2011, GDXJ topped out in April and made lower highs in September, while gold set higher highs several times from April through September. This was followed by substantial losses for the entire sector. In 2022, GDXJ bottomed in September 2022 and began forming higher lows into November, while gold bullion kept dropping from September to November.

Hedge funds have made all-time record total long:short ratios in gold, silver, and platinum, along with all-time record short:long ratios for palladium.

Hedge funds have increasingly been acting nearly identically to each other in recent decades. When I had worked at Thomson Reuters for 16-1/2 years, I sat next to a fellow whose job it was to track how hedge funds were investing and how they had been evolving through the decades. He showed me how, as recently as the early 1990s, hedge funds tended to be mostly independent of each other. Recently they have been mimicking each other's selections and algorithms with minor variations, so that at any critical reversal they are nearly unanimously piled onto the long side at a zenith, especially near a multi-decade top, and are nearly unanimously piled onto the short side at any nadir, especially when approaching a multi-decade bottom.

Hedge funds will massively close out any long or short position whenever a given asset has moved about 20% or 25% from its most recent extreme. This can lead to dramatic changes, such as in the late summer of 2024 when hedge funds had registered a huge net short position in non-internet Chinese shares. Suddenly they began rallying (see a one-year chart of ASHR) by about 60%, so that hedge funds not only closed out their massive shorts but went very heavily net long Chinese non-internet stocks. Then the Chinese market slid rapidly lower again, causing the hedge funds to once again close out their Chinese positions. Hedge funds thus accomplished the rather amazing feat of losing money on both sides of the same trade within less than one month.

With hedge funds' record longs in gold and their record shorts in palladium, it seems pretty clear what will happen next.

Even non-financial media have been jumping aboard gold's bandwagon.

I expect to read about gold on Seeking Alpha, the Wall Street Journal, or the New York Times. As gold has been frequently featured on National Public Radio and very recently on cooking and travel cable channels for the first time in many years, it is almost certain that it has become too popular. Last year there was a widespread myth that cryptocurrencies were a valuable portfolio hedge and these got touted in the most unlikely places just in time for these to experience historic losses especially for non-Bitcoin crypto. Now we have a nearly identical myth about gold being an ideal hedge against uncertainty. It is definitely true that gold has been in a very-long-term bull market which began on August 25, 1999 and will likely continue for perhaps another decade, but whenever any asset becomes very trendy then it is almost always a good idea to sell and to wait for most investors to be gloomy again before getting back in.

Gold consistently performs poorly in the first year of any large-cap U.S. bubble collapse.

Following the 1929 bubble for the most popular U.S. stocks, sometimes called the blue-chip bubble, gold mining shares experienced huge losses for less than one year, followed by an impressive multi-year bull market. If we jump forward to the 1970s then we see similar behavior for this sector during the plunge following the Nifty Fifty bubble. Going forward some more to 1999-2000, we see another example of large percentage losses for gold mining and silver mining shares which bottomed in mid-November 2000, two years before most U.S. stocks completed their lowest points in October 2002. If you had bought the equivalent of HUI, an index of gold mining and silver mining shares, at its exact bottom on November 15 or November 16, 2000 then by December 2, 2003 you would have had more than seven times as much money in just over three years. However, if you had bought HUI on the day that QQQ had topped out on March 10, 2000, then you would have initially suffered large percentage losses.

Similar behavior is likely to happen in 2025. As large-cap U.S. stocks may have completed all-time overvaluations as a group on February 18-19, 2025, the first several months to a year will likely be accompanied by greater percentage losses for gold mining shares and their funds including GDXJ than for the S&P 500 and similar large-cap indices. Following this decline, when gold, GDX, and GDXJ will once again go powerfully out of favor as these had done most recently during the summer of 2022, we will likely experience a doubling, tripling, or more for funds like GDXJ over the subsequent few years and eventually larger gains.

Gold will rise again, and probably by more than most gold bulls are currently anticipating. But don't buy it until it is once again widely detested, rather than now when it is adored.

The current U.S. equity bear market will likely last for roughly three years altogether.

Timing and price estimates, whether from me or anyone else, should be taken with a grain of salt. We have so many Bogleheads today who are convinced of their divine right to come out ahead in the long run that it will take a huge total market drop to convince them otherwise. Therefore, we might have a lengthy bear market, especially since we haven't had a serious bear market since March 9, 2009. There will be many pullbacks and just as many subsequent convincing-looking rebounds along the way. Don't believe frequent reports after each rally, including some probably during the next few weeks, about how "the market has bottomed." When we finally do reach the ultimate lows for most assets, the media and the vast majority of analysts will not be saying anything about recently bottoming. Instead they'll be telling you why you shouldn't buy since the market will supposedly be going much lower and why it will be many years before we can enjoy a true rally. That will be your buy signal.

Bear markets consistently experience the most powerful and frequent upward bounces.

Whenever there has been extended weakness for any asset, resist the temptation to sell, just as you must be equally firm about not chasing after anything which has experienced protracted strength. The market will repeatedly reward those who gradually accumulate any position into adversity rather than using lump sums or attempting momentum plays. Now that we have accelerated the downturn for the U.S. stock market, we will likely have energetic rebounds just as we did in previous severe bear market years including not only long-ago periods like 1931, but more recently 2001, 2002, and 2008.

A good rule of thumb is to track VIX and VVIX. Whenever VVIX has recently dropped to its lowest point in many trading days, while VIX has notably retreated from a recent peak in order to complete yet another higher low, this is often a useful time to add to short positions in very overpriced assets. With VIX recently surpassing 30, if it retests 20 then it could provide such an opportunity instead of jumping aboard when fear has recently become elevated.

Mark Hulbert did some useful research to confirm the thesis of the biggest bear markets featuring the sharpest short-term rebounds:

Sharp Short-Term Upward Spikes Confirm Bear Markets

Several emerging markets and commodity producers have either already become compelling buys or will likely do so at various points over the next few years.

Near the beginning of 2025, Brazilian shares and their funds including EWZ sported average price-earnings ratios below 8. Other funds of Brazilian stocks, including FLBR, BRF, and EWZS, had even more compelling ratios of profit growth to price-earnings ratio and other classic valuation measures championed by Benjamin Graham and Peter Lynch. I therefore began to purchase these and have continued to buy these into higher lows, so far in small percentages with the intention of gradually increasing these into all pullbacks especially when Brazilian insiders are doing likewise.

Recently we have experienced worthwhile bargains and depressed behavior for funds including THD (Thailand), EIDO (Indonesia), EPHE (Philippines), as well as some funds of commodity producers including REMX (rare-earth extractors). These are mentioned even less often than Brazil in the mainstream media. Chinese shares, especially those of non-internet companies like ASHR, remain worthwhile bargains after last year's early autumn spike and collapse mentioned earlier in this update; I plan to gradually accumulate them whenever they approach their 2024 bottoms.

Since June 2024 we have experienced by far the most intense insider selling by top executives in U.S. history.

Top corporate insiders sold about 2-1/2 times as much in U.S. dollar terms during the past summer, autumn, and winter than during any previous nine-month period. There was especially aggressive selling in the roughly 73 out of approximately 7300 listed U.S. companies which had become the most blatantly overpriced. If you look at the other 99% of U.S. companies, such as the Russell 2000 which can be tracked via the symbol IWM, then you will see that even before the most recent slump this index had been trading not only below its 2021 highs but below its 2021 lows. All previous five U.S. large-cap stock-market bubbles in 1837, 1873, 1929, 1972, and 1999 had extended underperformance by all but the top 1% of all shares prior to suffering severe bear markets. In all of these other five bubbles, those large-cap U.S. stocks which had been the big winners ended up losing more than 80% during their subsequent bear markets.

The following charts highlight some of the all-time record extremes that we had experienced during recent months:

The more that U.S. investors have piled into U.S. stocks as a percentage of their total net worth, the worse is the performance of the U.S. stock market during the subsequent decade:

There has rarely been any stock market in world history which had been more overpriced than the U.S. stock market was at its February 18-19, 2025 bubble top:

Whenever high-yield "junk" bonds barely yield more than U.S. Treasuries of identical maturities, it signals that investors are willing to accept far too much additional risk for a tiny additional yield:

One especially dangerous sign of overvaluation was seen at the February 2025 peak when the total amount of money in dangerous leveraged long funds was by far at an all-time record while the ratio of leveraged long assets to leveraged short assets also reached an all-time extreme:

The AI, internet, and Nifty Fifty bubbles have become increasingly extreme in how only about 1% of the most popular stocks have accounted for all of the stock market's gains:

The bottom line: whenever it is most worthwhile and profitable to buy or to sell anything, almost everyone wants to do the exact opposite. In February 2025 almost everyone wanted to be long the most popular large-cap U.S. stocks before they began what will likely become roughly three-year bear markets. Now everyone loves gold which will likely suffer a similar fate. Instead, invest in the most unpopular assets including palladium which can be purchased via the symbol PALL. Gradually accumulate emerging-market shares throughout the next few years whenever they are most disliked and have strong annualized profit growth. Eventually it will become timely to purchase funds of mid-cap gold mining and silver mining shares including GDXJ, but only when they are once again hated which may occur near the end of 2025.

Disclosure of current holdings:

Below is my current asset allocation as of 4:00 p.m. on Thursday, April 3, 2025. Each position is listed as its percentage of my total liquid net worth.

I computed the exact totals for each position and grouped these according to sector.

The order is as follows: 1) U.S. government bonds; 2) shorts; 3) bear funds; 4) precious metals; 5) emerging markets.

VMFXX/TIAA Traditional, TIAA money market/bank CDs/FZDXX/FZFXX/SPRXX/SPAXX/BPRXX/Savings/Checking long: 35.88%;

17-Week/52-Week/26-Week/13-Week/2-Year/8-Week/3-Year/5,10,30-Year TIPS/4-Week/6-Week/20-Year: 23.17%;

TLT long: 11.07%;

I Bonds long: 3.70%;

PMM long: 0.01%;

XLK short: 29.02%;

QQQ short: 22.12%;

SMH short: 1.17%;

GDXJ short: 0.47%;

AAPL short: 0.13%;

GDX short: 0.01%;

SARK long: 0.59%;

PSQ long: 0.29%;

Gold/silver/platinum coins: 8.66%;

PALL long: 2.12%;

FLBR long: 0.31%;

EWZ long: 0.22%.