Trying to count the social patterns that have not happened yet, AI job takeover edition

Posted on February 11, 2026 by legallysociable

It is hard to know how many jobs AI might eliminate when we cannot yet count many jobs eliminated:

Measurement doesn’t abolish injustice; it rarely even settles arguments. But the act of counting—of trying to see clearly, of committing the government to a shared set of facts—signals an intention to be fair, or at least to be caught trying. Over time, that intention matters. It’s one way a republic earns the right to be believed in.

The BLS remains a small miracle of civilization. It sends out detailed surveys to about 60,000 households and 120,000 businesses and government agencies every month, supplemented by qualitative research it uses to check and occasionally correct its findings. It deserves at least some credit for the scoreboard. America: 250 years without violent class warfare. And you have to appreciate the entertainment value of its minutiae. The BLS is how we know that, in 2024, 44,119 people worked in mobile food services (a.k.a. food trucks), up 907 percent since 2000; that nonveterinary pet care (grooming, training) employed 190,984 people, up 513 percent; and that the United States had almost 100,000 massage therapists, with five times the national concentration in Napa, California.

These and thousands of other BLS statistics describe a society that has grown more prosperous, and a workforce endlessly adaptive to change. But like all statistical bodies, the BLS has its limits. It’s excellent at revealing what has happened and only moderately useful at telling us what’s about to. The data can’t foresee recessions or pandemics—or the arrival of a technology that might do to the workforce what an asteroid did to the dinosaurs…

This was the point Goolsbee wanted to emphasize: Economists are constrained by numbers. And numerically speaking, nothing indicates that AI has had an impact on people’s jobs. “It’s just too early,” he said.

A lack of certainty should not be mistaken for a lack of concern.

This sounds like a classic issue facing those concerned about particular social problems: can the numbers help you build a case that this issue is important and worthy of the attention of others? With all the possible social problems that need attention, having clear data regarding the problem can help make the case to the public and leaders. But, if this is largely speculation regarding AI, how many will act based on that?

Another important factor regarding counting: it is a key way of trying to make sense of a large and complex society. When you have a country with over 330 million residents, 50 states, and numerous important social patterns occurring, having data to look at can help make sense of what is happening on the broad scale. Anecdotes offer little on a large scale; case studies might provide some insight. Having statistics on a society-wide scale is necessary.

A third way to think about this: those who could generate numerical predictions or have small sectors that could provide early data on this could be helpful for others.

The difficulty of measuring the U.S.’s housing needs

Posted on February 5, 2026 by legallysociable

How many housing units are needed in the United States? Different sources disagree and they measure the issue in different ways:

The disparate projections reflect the challenge of quantifying the nation’s housing needs, a puzzle that rests on assumptions about how much a home should cost, how many people it should hold, and how big a footprint it should have…

The U.S. has 146 million homes, Census Bureau data show. Of those, 8.1 million are “doubled up” households, meaning people are sharing space with non-relatives. Zillow’s housing estimate assumes most of those people would prefer having their own place. There also are 3.4 million vacant homes available to rent or buy, the real estate website says. So Zillow economists subtracted the number of available homes from the number of doubled-up households and concluded that the nation needs 4.7 million more homes…

Several analyses zeroed in on two questions: How many homes should be vacant, and how many consumers have delayed striking out on their own because of the cost…

For many economists, that suggests the equation should be: the number of existing households, plus the number of homes that should be vacant, plus the number of households that would naturally come into being if there was enough inventory to lower prices.

This matters for multiple reasons. First, it is helpful to have more accurate estimates. This can help policies intended to help. These are methodological questions; how do we measure what is happening on the ground? Projections that are too high or too low could lead to not addressing the issue or actions that do not have the intended consequences.

Second, the number of units needed matters because it is part of the public discourse about housing. The article describes estimates ranging from 0 units needed to 8-20 million units are needed. When discussing social problems in public discussions, these numbers can influence a sense of urgency. If people hear there are 5 million units needed, are they more likely to act compared to hearing 1 million units are needed?

Third, the numbers are part of a national discussion. Housing needs can vary quite a bit place to place. Housing is often a very local issue. These numbers are about what could be done on a national level which then has affects on local efforts.

These different measurement strategies and results could end up make it harder to reach consensus on what should be done.

“So goes Amazon/social media discourse/better schools/job growth/affordable housing, so goes the United States”

Posted on January 17, 2026 by legallysociable

In 1953, the CEO of General Motors was in a congressional nomination hearing as he had been appointed by President Eisenhower to be Secretary of Defense. Did he see think that holding the government job would be a conflict of interest given his large holdings of GM stock?

I cannot conceive of one because for years I thought what was good for our country was good for General Motors, and vice versa. The difference did not exist. Our company is too big. It goes with the welfare of the country. Our contribution to the Nation is quite considerable.

This quote ties the fate of one company to the fate of the country. Since 1953, this might fit numerous large corporations that employed many people and generated large revenues. Today, this might be Amazon or Walmart or Nvidia or other influential corporations.

At the same time, people in the United States focus on particular social issues that they think require attention. Address conversation and participation on social media and life would get better. Improve schools and education and future generations have a brighter future. Add more jobs in exciting industries and people will be excited. Provide decent or good housing at affordable prices and this can lead to other opportunities. The issue of the moment might have been different years ago and it could change in the future but there are always conversations about what should be done.

Both sets of statements are reductionistic. No single company determines the fate of the United States. One social issue could affect many yet other issues might have a broader reach or have larger effects.

Companies and social problems do evolve and change over time. A number of the companies that led the way in the United States decades ago are no more. Certain social issues vexed the country years ago but may have receded from view today or the effects were ameliorated.

In other words, the conversations of today may not be the most helpful if they limit focus to just one company or social problem and even broader conversations will change in the future. This does not mean that the conversations of now are not valuable; rather, we should seek to have a broad field of vision and a sense of the current scene even as we discuss specific firms and social concerns.

Mortgage fraud rates very low – but on the rise?

Posted on September 20, 2025 by legallysociable

With cases of mortgage fraud in the news, one source says it is rare:

About 1 in 116 mortgage applications contained fraud in the second quarter of 2025, according to Cotality’s National Mortgage Application Fraud Risk Index.

The data shows the two riskiest investments for mortgage fraud are investment properties and multiunit properties.

This is less than 1%.

But the same source says mortgage fraud is on the rise:

“The increase in the fraud risk can partly be attributed to the volatility starting to be seen in the real estate market,” Matt Seguin, senior principal of mortgage fraud solutions, said in the Cotality report. “Interest rate cuts haven’t come at the rate expected over the last year, so purchase transactions, which, historically speaking, have higher fraud risk, continue to represent almost 70% of the applications seen by Cotality.”

Cotality analyzed data in six categories of mortgage fraud: identity, transaction, property, income, occupancy, and undisclosed real estate. The research found that every category except occupancy saw an increase in the second quarter.

The largest year-over-year increases were in undisclosed real estate debt and transaction fraud risk. Undisclosed real estate debt rose 12% this year, compared with a 5.9% decline year over year in 2024. Transaction fraud risk increased 6.2% this year, following a 4.9% increase last year.

Rare and relatively small increases in the last year.

Perhaps the problem of mortgage fraud would sound more serious if this 1 in 116 mortgages was connected to the cumulative money involved. Each mortgage is connected to a good amount of money. Add all the fraud up and how much money are we talking about? Is it enough money for financial institutions or the general public to pay attention to?

Another way to think about this would be to compare fraud rates here with fraud rates with other financial instruments. How about credit card debt? Auto loans? Home equity loans? And so on. Mortgage fraud is low but perhaps it is even lower than in other areas or higher than others.

Regardless of the numbers, absolute or otherwise, fraud is still fraud. But whether it is perceived as a social problem might take more than just reporting the numbers or putting them in context.

American drivers cause many accidents and deaths

Posted on January 11, 2024 by legallysociable

Americans like to drive. And American drivers contribute to a lot of accidents and deaths:

Above all, though, the problem seems to be us — the American public, the American driver. “It’s not an exaggeration to say behavior on the road today is the worst I’ve ever seen,” Capt. Michael Brown, a state police district commander in Michigan, told me. “It’s not just the volume. It’s the variety. There’s impaired driving, which constituted 40 percent of our fatalities last year. There are people going twice the legal limit on surface streets. There’s road rage,” Brown went on. “There’s impatience — right before we started talking, I got an email from a woman who was driving along in traffic and saw some guy fly by her off the roadway, on the shoulder, at 80, 90 miles an hour.” Brown stressed it was rare to receive such a message: “It’s got so bad, so extremely typical,” he said, “that people aren’t going to alert us unless it’s super egregious.”

In 2020 and 2021, the National Highway Traffic Safety Administration has calculated, approximately a quarter of all fatal wrecks in the United States involved vehicles traveling above the posted speed limit; a significant percentage of the dead, whether passenger or driver, were not wearing seatbelts. In line with the trends documented by Kuhls in Nevada — and observed firsthand by Brown in Michigan — national intoxicated-driving rates have surged to the extent that one in every 10 arrests is now linked to a suspected D.U.I. And aggressive driving, defined by AAA as “tailgating, erratic lane changing or illegal passing,” factors into 56 percent of crashes resulting in a fatality. (Distressingly, this statistic does not cover the tens of thousands of people injured, often critically, by aggressive drivers, or the 550 people shot annually after or during road-rage incidents — or the growing number of pedestrians and cyclists deliberately targeted by incensed motorists.)

Take the bad behavior and add the perils of distraction by smartphone — responsible, by one conservative estimate, for about 3,500 deaths annually — and you’re left with what Emily Schweninger, a senior policy adviser at the U.S. Department of Transportation, described to me as a “genuine public-health crisis” on the level of cancer, suicide and heart disease.

Much could change in the coming years to address this issue. Safety features in vehicles. Changed designs of roadways and spaces for pedestrians and bicyclists. Other efforts need more time and capabilities: self-driving vehicles, a changed culture around roads, driving, and community life.

But, part of the issue is whether these accidents and deaths are a problem or not. Americans like to complain about other drivers and tend to see their own driving as okay. Driving is required in many places. Some drivers might even enjoy driving. The delivery of many of our goods requires driving. Are deaths via vehicle just the price Americans are willing to pay for driving?

Addressing this issue is a long-term project. All of daily life contains some risks but Americans tend to not think much about the risks of driving even as it impacts many lives on a daily basis. Does this mean a national safety campaign is needed? A serious conversation about how necessary driving should be? A need to invest in new technologies and options? On one hand, plenty of people would have experience with this issue. On the other hand, it will take a lot of work to convince people to support significant changes to American driving and all that goes with it.

Need to conduct a census? Invent house numbers

Posted on November 29, 2022 by legallysociable

In Democracy’s Data, historian Dan Bouk explains how the process of counting people led to the development of house numbers:

Numerical addresses owed their very existence to censuses. House numbers are essentially names state offices assigned to lodgings. According to one historian, the idea of marking houses with consecutive numbers dates to the desire of Enlightenment monarchs – like the Austrian Maria Theresa – to make a better count of men who could be conscripted into the military. Houses were also sometimes numbered to count minority groups, like the Jews of Prague in one of the earliest cases in 1727, or to make it easier for rulers to find places where soldiers could be billeted. (This practice of quartering soldiers in private homes was common enough, and widely enough disliked, to be proscribed during peacetime in the U.S. Bill of Rights, in the Third Amendment.) Odd-numbered houses on one side of the street, even-numbered houses on the other, had to be invented too. Philadelphia, the nation’s temporary capital at the time, alternated odds and evens in order to make it easier to conduct the federal census in 1790. New York City copied its neighbor in 1793, as did Paris in 1805. Though people quickly came to rely on numerical addresses to find one another, send letters, or navigate strange and growing cities, those uses were accidental benefits of an apparatus meant to serve the data makers in imperial or national bureaucracies. (58)

It is hard to imagine a world without numbered street addresses today. Perhaps it is just the scale of communities; how would you differentiate between the thousands of addresses within a small geographic area? Street addresses are not the only way properties are identified. Communities have categorization systems for parcels. The way current addresses are set up requires a zooming in approach: you start with the community or zip code, find the street, and then identify the number. Of course, there could be a system that does not use or need numbers, but it is hard to imagine such a system in the quantified world we inhabit.

More broadly, numbers are helpful with addressing particular social issues. Conducting a census helps with social matters like taxation, resource allotment, and more. Having numbers for addresses helps enable counting but also aids mail delivery and others seeking to find particular addresses.

The difficulty in getting a number on how many housing units needed in the United States

Posted on November 22, 2022 by legallysociable

One journalist set out to find out how many housing units the United States needs. The answer was complicated:

Looking at the number of American households and the number of vacant housing units, Freddie Mac, the government-sponsored purchaser of mortgage-backed securities, estimates a current supply shortage of 3.8 million units, driven by a 40-year collapse in the construction of homes smaller than 1,400 square feet.

The group Up for Growth also arrived at an estimate of 3.8 million, using data on the total demand for housing and the overall supply of habitable, available units.

The National Association of Realtors compared the issuance of housing permits with the number of jobs created in 174 different metro areas. It found that only 38 metro regions are permitting enough new homes to keep up with job growth; in more than a dozen areas, including New York, the Bay Area, Boston, Los Angeles, Honolulu, Miami, and Chicago, just one new home is getting built for every 20-plus jobs created. The NAR estimates an “underbuilding gap” of as many as 7 million units.

These numbers draw on data such as vacancy rates, household-formation trends, and building trends. But none of the estimates capture what I’ve come to think of as the affordability gap: the difference between the housing we have and the housing we would need in order to ensure that working-class people could once again live in our big coastal cities for a reasonable cost. Freddie Mac does not purport that building 3.8 million units would make New York accessible to big middle-class families and end homelessess in San Francisco. The National Association of Realtors is not contemplating whether janitors can walk to work in Boston…

To come up with that estimate, the two economists built a complicated model that assumed Americans could move wherever their wages allowed and the housing supply would adjust as it would in a place with typical permitting standards. In such a world, they estimated in some associated work, 53 percent of Americans would not live where they are currently living. San Francisco would have an employed population 510 percent bigger than it does today—implying an overall population of something like 4 million, rather than 815,000, with 2 million housing units instead of 400,000. The Bay Area as a whole would be five times its current size, the economists estimated. The average city would lose 80 percent of its population. And New York would be a startling eight times bigger. Some back-of-the-envelope math (mine, not theirs) suggests that the United States would have—deep breath here—perhaps 75 million more housing units in its productive cities than it currently has.

Considering such big numbers can be both helpful and daunting. The sheer size of these figures – multiple millions to tens of millions – highlights the scope of the problem. Housing is not a small issue; it is a large issue that needs addressing. Big numbers can help convince people this is an important social issue to address. On the other hand, these figures are daunting. That is a lot of housing units to consider. How can small efforts contribute to such a big need? Who can address this?

Even if the various methods and experts above do not agree on the same numbers, together they suggest much needs to be done. Can we get a commitment from states or cities to approve more units proportionate to their populations? I could imagine some kind of pledge drive and counting system to see the progress toward a sizable goal. Or, how about a long-term plan on the scale of a Manhattan Project or a space race to get units built? Of course, addressing housing at the federal level is difficult.

US roadway deaths rise 10.5% in one year

Posted on May 18, 2022 by legallysociable

Fatalities on American roads increased quite a bit in 2021:

Nearly 43,000 people were killed on U.S. roads last year, the highest number in 16 years as Americans returned to the roads after the coronavirus pandemic forced many to stay at home.
The 10.5% jump over 2020 numbers was the largest percentage increase since the National Highway Traffic Safety Administration began its fatality data collection system in 1975. Exacerbating the problem was a persistence of risky driving behaviors during the pandemic, such as speeding and less frequent use of seat belts, as people began to venture out more in 2021 for out-of-state and other road trips, analysts said.
Transportation Secretary Pete Buttigieg said America faces a crisis on its roads. The safety administration urged state and local governments, drivers and safety advocates to join in an effort to reverse the rising death trend…
Buttigieg pointed to a national strategy unveiled earlier this year aimed at reversing the trend. He said earlier that over the next two years his department will provide federal guidance as well as billions in grants under President Joe Biden’s new infrastructure law to spur states and localities to lower speed limits and embrace safer road design such as dedicated bike and bus lanes, better lighting and crosswalks. The strategy also urges the use of speed cameras, which the department says could provide more equitable enforcement than police traffic stops.

Americans like driving and all that comes with driving. Because of this, Americans generally accept the risks of driving. While people may have fears of airplanes crashing or being hit by lightning or other improbable occurrences, the regularity of vehicle accidents does not seem to bother many.

Would a big jump in roadway fatalities catch people’s attention in a way that a typical year-to-year change would not? That this jump is tied to COVID-19 is also an interesting twist; driving might be more dangerous during and after a deadly pandemic. Also in the article, officials note the difficulty of quickly reducing roadway deaths. When do such deaths become an acknowledged crisis or a serious social problem?

Share this:

Share this:

Share this:

Share this:

Share this:

Share this:

Share this:

Share this:

Share this:

Share this: