Houses, pensions and billionaires who fell for it

A widely shared poll, claiming that baby boomers overwhelmingly support higher taxes on younger workers to keep their Social Security benefits intact, prompted Nick Maggiulli of Of Dollars and Data to answer a question younger readers often ask him: can I count on Social Security in retirement? His answer is that the program will keep paying something, and he argues that planning on roughly 70% of your scheduled benefit is reasonable.

Two trust funds, one problem

What happens then is often misunderstood. Social Security can't borrow, so after the fund is gone it must cut benefits to whatever incoming payroll taxes can pay for. That is about 78% of scheduled benefits in the first year, a 22% cut but not a 100% one. The share then falls as the number of workers per retiree keeps shrinking, reaching about 62% by 2100. For today's average benefit of about $2,000 a month, 78% means about $1,560.

How a benefit is calculated

Many people assume benefits work like a pension, where more years of work means a bigger payout. That is only true if income stays the same. The Social Security Administration (SSA) uses three steps.

• 90% of the first $1,286 of AIME
• 32% of the amount between $1,286 and $7,749
• 15% of the amount above $7,749

Maggiulli compares this to a marginal income tax schedule running in reverse: each slice of earnings is "paid back" at a lower rate than the slice before. The result is that low earners get most of their earnings replaced and high earners get a much smaller share.

Worked example (from the article): with an AIME of $5,000, the benefit is 0.9 × $1,286 + 0.32 × ($5,000 − $1,286) = about $2,345 a month at the full retirement age of 67. That is 46.9% of AIME. Claiming at 62 cuts it by about 30%, to $1,642. Waiting until 70 adds about 24%, for $2,908. The largest possible benefit in 2026 is $5,181 a month. It goes to people who earned at or above the cap for 35 or more years and waited until 70 to claim.

You can see your own projected benefits at 62, 67 and 70 on the SSA's "my Social Security" website.

The options for closing the gap

The SSA has estimated how much of the shortfall each proposed change would close. The figures below are measured against the 2025 shortfall. The 2026 shortfall is larger, so each would close somewhat less today.

The price-indexing option needs explaining. Wages usually grow faster than prices, and tying bend points to wages keeps benefits in step with living standards. Social Security has historically replaced about 40% of a median earner's wages. Tying the bend points to prices instead would gradually cut that share in half.

Raising the payroll tax is the only option on the list that closes the whole gap by itself. In a 2025 National Academy of Social Insurance survey, 85% of respondents said benefits should be kept or increased even if that meant raising taxes. Maggiulli therefore leans toward a future of higher taxes rather than smaller checks.

A planning rule

Maggiulli expects Congress to act in some way. Whether through a legislated fix or the fund running dry, he plans on receiving 70% of scheduled benefits. For the average person retiring in 2032, that is about $1,400 a month. He expects people retiring later to get more, even after inflation, because wages, and therefore AIME, tend to grow faster than prices.

His method:

Caveats

Maggiulli admits nothing is guaranteed. Congress could change the program drastically, though he doesn't expect it to. The 70% figure is his own judgment, not a projection. The article also doesn't say how taxes on benefits, spousal benefits or future changes to the earnings cap would affect any individual's numbers. His bottom line is that you shouldn't build a retirement plan around the Social Security check, but assuming it won't come at all is the bigger mistake.

What "August" actually means

The data runs well behind the calendar. Case-Shiller's "August" figure is a three-month average of closing prices from June, July and August. Many June closings come from contracts signed in April, because a home sale usually takes one to two months to close. So the release partly describes what buyers agreed to pay in the spring. That is worth keeping in mind before reading it as a sign of where the market is heading now.

The headline numbers

All three sped up by about 0.2 percentage points. The big-city composites are rising faster than the national figure, which means prices outside the largest metros are rising more slowly.

Why the monthly numbers point in opposite directions

The gap comes from the calendar. Home prices usually peak in late spring and summer and soften once the selling season ends. A raw decline in August is normal. The adjusted figures ask a different question: was August better or worse than an ordinary August? The answer is better. Rebecca Kaufman of S&P Dow Jones Indices says the usual post-summer slowdown "is obscuring some underlying strength in home prices."

What "falling in real terms" means

An illustrative example, not taken from the article: a home worth $400,000 in August 2025 would be worth about $408,400 a year later after a 2.1% gain. To keep pace with 3.4% inflation it would need to reach about $413,600. The roughly $5,200 shortfall is the 1.3-percentage-point real loss S&P describes. Fifteen months of small losses like that add up, though S&P notes the gap has narrowed as nominal price growth picked up.

A nine-point spread between cities

Regional differences are large. The gap between the strongest and weakest metro areas is nearly 9 percentage points:

• Strongest: Chicago +6.7% (top for the sixth month running), New York +6.2%, Cleveland +4.3%
• Weakest: Seattle −2.0% (largest decline for the third month running), Las Vegas −1.3%, Denver −0.8%

Midwestern and Northeastern markets are leading. Western markets, many of which rose sharply earlier in the decade, are falling. The national average of 2.1% hides very different outcomes depending on where you live.

The second opinion: FHFA

FHFA's regional figures tell a similar story. Monthly changes ranged from −0.8% in New England to +2.1% in East South Central (Kentucky, Tennessee, Mississippi, Alabama). Over 12 months, the range ran from +0.5% in the Mountain division to +4.4% in East North Central, which includes Chicago and Cleveland. That fits Case-Shiller's metro leaders.

Affordability

Kaufman points out that even as prices lag inflation, buying a home is still hard. Inflation is at 3.4%, Brent crude oil is near $100 a barrel, and mortgage rates are in the mid-to-high 6% range. Prices that are flat in real terms do not make homes cheap when financing costs this much and household budgets are already stretched.

What's left open

The release doesn't say why the regions have split so sharply. Possible reasons include housing supply, local job markets, and how much each area rose earlier, but none of these are covered. It also doesn't say whether the faster nominal growth will last as autumn data arrives. McBride's "few things to watch" section, where he might weigh in, is behind a paywall. Because of the April-to-August lag, it will take a few more releases to see whether the summer's adjusted strength holds.

The September numbers

The revisions matter. New home sales come from a survey of a fairly small sample, and monthly estimates often get revised by a meaningful amount. A 0.6% monthly drop is within the noise. The year-over-year decline and the inventory trend tell you more.

How much supply is sitting unsold

At the end of September, 476,000 new houses were for sale (seasonally adjusted). That is 1.0% below August's 481,000 and 2.3% below the 487,000 of a year earlier.

• 676,000 a year ÷ 12 ≈ 56,300 homes sold per month
• 476,000 for sale ÷ 56,300 ≈ 8.4 months

So if builders stopped adding homes, current stock would take about 8.4 months to sell at September's pace. That is down from 8.5 months in August but up 2.4% from 8.2 months a year ago. Inventory fell over the year, but sales fell faster, so the supply measure rose.

For scale:

• Normal: about 4 to 6 months
• Record high: 12.2 months, in January 2009, during the housing bust
• Record low: 3.3 months, in August 2020, during the pandemic buying rush

At 8.4 months, McBride says, supply is "well above the top of the normal range."

Three kinds of inventory

How the Census Bureau defines "for sale" matters. A house counts as for sale once a building permit has been issued (or, where no permit is needed, once work has started on the foundation), as long as no sales contract has been signed and no deposit taken. So the 476,000 figure includes many homes that are only permits on a lot.

Since 1973 the Bureau has split inventory into three groups, shown in McBride's third chart:

The groups carry different risks for builders. A completed, unsold home is the expensive kind of inventory: the builder has spent the money on land, materials and labour and pays interest and upkeep until it sells. Builders usually work that stock down with price cuts or incentives such as paying to lower a buyer's mortgage rate. (That is general industry practice; the article doesn't discuss pricing.) Completed inventory falling from its January peak suggests builders have been clearing some of it.

A record number of not-started homes is harder to read, and the article doesn't interpret it. One possibility is that builders have permits in hand but are waiting for contracts before they break ground, which limits their exposure if demand weakens. Another is that permits were pulled during a more optimistic period. Either way, those homes add to measured supply without yet being physical houses.

What the picture adds up to

Taken together: sales are steady at roughly pre-pandemic levels and slightly better than expected. They are down from a year ago, and supply is well above normal. Builders appear to be trimming the costliest inventory (completed homes) and the pipeline under construction, while not-started inventory reaches a record. A high months-of-supply figure usually points to softer new-home prices and fewer housing starts ahead. The article itself doesn't make that forecast.

What the article leaves out

The free part of the post doesn't cover prices, regional splits, or McBride's fourth chart, which compares not-seasonally-adjusted monthly sales for 2019 to 2026 against the 2005 bubble peak and the 2008 to 2011 bust lows. That analysis is for paid subscribers. It also doesn't say what drove the consensus beat or how mortgage rates figured into September demand.

Why this shock is different

Step one: Europe as a whole

The Kiel researchers (Chowdhry et al., 2026) calculated how Chinese and EU exporters' shares of third markets changed between the 2012–14 average and the 2022–24 average. Averaging over three years smooths out one-off swings in individual years.

The results are one-sided. Across nearly every product category, EU manufacturers lost share and Chinese manufacturers gained it. The few exceptions where Europe held up were leather goods (largely luxury items such as handbags), art and weapons. In those categories, brand, provenance or political alignment count for more than cost or technical specification.

The researchers also expressed the shift in growth terms: Chinese exports to these markets grew much faster than European exports over the decade. Klement's post doesn't give the headline figures. They appear only in the study's charts.

Step two: Germany against its peers

Next, the researchers tested whether every EU country lost share at the same rate. They compared Germany's actual change in each product category with the change it would have seen if it had moved exactly in line with the European average. The gap between the two is the "excess" loss or gain.

A simple illustration, with invented numbers: suppose EU exporters together lost 10% of their share in a product's third-market sales. If Germany's share in that product fell 10%, it simply moved with Europe. If it fell 18%, the extra 8 points is the excess, a loss specific to Germany.

Germany's results were bad. Its excess losses were large and covered many categories, and they were concentrated in the industries that define the German economy:

• motor vehicles
• machinery
• electrical equipment
• optical and medical equipment

These are exactly the categories where Germany has historically competed on engineering quality rather than price. Klement's post doesn't report the actual percentages.

The argument: quality, not just cost

The authors argue, and Klement says he largely agrees, that the breadth and location of the losses point to one conclusion. Chinese goods are no longer winning only on price. They now match or beat German products on quality and innovation. Klement adds that many German manufacturers still deny this.

The reasoning goes as follows. If the problem were only cost, such as Germany's high energy prices after 2022 or its labour costs, you would expect losses to concentrate in price-sensitive, commoditised goods. Instead they cluster in premium engineered products, where buyers have long been willing to pay more for German quality. When buyers stop paying that premium, the most likely explanation is that the quality gap has closed.

Why tariffs don't fix it

Klement's policy conclusion is that tariffs won't help manufacturers that aren't innovative enough. The third-market data shows why. An EU tariff on Chinese cars or machinery can protect sales inside Europe, but it does nothing in Brazil, Turkey, Southeast Asia or the Middle East, where German firms have to compete without protection. For an export-led economy like Germany's, those markets matter a great deal.

His alternative is heavy investment in future technologies. He cites the German car industry as evidence that this is not happening, though the post doesn't detail which investments he considers missing.

What the post leaves open

The evidence is suggestive but limited, and a careful reader should keep a few points in mind:

Even with those limits, the core finding is notable. Over a decade, in the markets where competition is fairest, Germany has lost ground faster than its European neighbours, and it has lost most in the products that built its export strength.

Ben Carlson, a wealth manager who writes the blog A Wealth of Common Sense, has reposted an edited chapter from his book Don't Fall For It. The trigger, he says, is the attention around Nathan Fielder's new film about Elizabeth Holmes, You Can See Everything. His case is simple. Being smart and successful doesn't protect you from bad money decisions and can make them more likely. He thinks it's worth repeating now because confidence tends to rise along with markets in boom times.

The problem: smart people get taken too

Carlson opens with a line from the economist John Kenneth Galbraith: fools are "soon or eventually" separated from their money. Carlson's reply is that fools aren't the only ones. A high income, a large portfolio, a degree from a prestigious university or an impressive job title doesn't make you good at managing money. He argues that many of the worst investors are very successful in other fields.

Theranos was Holmes's Silicon Valley blood-testing company. It promised to run a series of diagnostic tests from a pinprick of blood. Holmes dropped out of Stanford at 19 to build the device, which she called "the most important thing humanity has ever built."

Carlson describes how staged the company was. In 2015 Vice President Joe Biden toured the facilities and afterward called it "the laboratory of the future." The lab he saw was staged, the devices didn't work, and most lab workers had been told to stay home that day. Holmes also spoke in a deliberately deeper voice so she would sound more serious.

The losses Carlson lists:

Partner Fund Management, a well-known hedge fund, bought shares at a price that valued Theranos at $9 billion. On paper, that made Holmes worth $5 billion.

How it worked: borrowed credibility

Carlson argues that investors suspected nothing because of who else was involved. The board included former Defense Secretaries Jim Mattis and William Perry, former Secretaries of State George Shultz and Henry Kissinger, and retired Navy Admiral Gary Roughead. Shultz called Holmes "the next Steve Jobs or Bill Gates."

Why success breeds overconfidence

Carlson's second mechanism is "personalizing" success. He draws on Brendan Moynihan's book What I Learned Losing a Million Dollars, about Jim Paul. Paul made a fortune trading commodities when he was young, then lost it after the windfall went to his head. Moynihan's lesson is that people who treat their successes as proof of their own ability, rather than of a good opportunity, good timing or plain luck, start to believe their involvement guarantees success.

Carlson adds a social piece. When the people around you assume intelligence is all that matters, nobody holds you accountable for your decisions.

Carlson goes further than that. He says the bias blind spot was actually bigger among people with higher cognitive ability. His explanation is that intelligent people have more mental firepower to talk themselves into things.

How much weight to put on this

The argument is mostly anecdote plus one line of research. Theranos is a striking example, but it was a private company sold to a small circle of very rich people. Carlson doesn't show that wealthy investors make mistakes more often than ordinary ones, only that they clearly make them. The study supports a narrower claim than "smart people are worse investors": intelligence doesn't protect you from a blind spot about your own biases, and simply knowing about biases doesn't fix them.

That second finding is the most useful one for an ordinary investor. If knowing about a bias doesn't cure it, then good intentions won't either. What helps are outside checks: verifying claims yourself instead of leaning on endorsements, and asking whether a past win was skill or luck. Carlson's chapter diagnoses the problem but offers no checklist. He doesn't say how to tell a real opportunity from a well-credentialed fraud, beyond implying that famous names are not evidence.