In 2024, Netflix released the four part docuseries American Conspiracy: The Octopus Murders about the mysterious 1991 death of journalist Danny Casolaro and his investigation into the Inslaw/PROMIS affair—a spiraling conspiracy theory that connected every major scandal of the 1980s to a federal software contract dispute.

Even before the Netflix series, the Inslaw/PROMIS affair had been a subject of much theorizing, including books, articles, an off-broadway stage play, and an episode of the TV show Unsolved Mysteries.

The short version is that PROMIS (Prosecutors Management Information System) was the name of software written in COBOL created by the software company Inslaw in the 1980s for federal prosecutors to keep track of cases winding through the legal system. Eventually, Inslaw fell into bankruptcy, leading to a protracted legal battle between Inslaw and the Department of Justice (DOJ).

During the bankruptcy dispute, Inslaw would accuse the DOJ of an extensive conspiracy that connected the software to everything under the sun—the Iran-Contra scandal, BCCI scandal, the October surprise, drug cartels, Israeli intelligence, etc—which is why it was dubbed “the Octopus.”

One version described the U.S. government of providing a copy of the software to Israel, who then augmented it with spyware and backdoor access and then sold it to intelligence firms across the world.

While there was plenty of speculation about what Casolaro could have unearthed during his investigation via a source connected to Lyndon Larouche, the Inslaw/PROMIS affair might be one of the most researched subjects in U.S. government history with at least five separate federal investigations—including ones from the DOJ, CIA, House Judiciary Committee, and the Senate Select Committee on Intelligence (SSC)—all of it leading to dead ends. Much of the investigations asked seemingly insignificant questions about whether government agencies used the software or shared it with others.

Compare that to the investigation into the Bank of Credit and Commerce International (BCCI), which originally got little attention or scrutiny but would eventually be described by a Senate investigation as an “international financial crime on a massive and global scale.”

Democrats in Congress at the time successfully demanded an independent counsel be appointed to investigate the affair, even though similar attempts at investigating illegal military funding to Iraq around the same time went nowhere.

The DOJ investigation ordered by then-attorney general William Barr would entail interviewing dozens of people related to Inslaw and the DOJ team handling the contract. None of them corroborated Inslaw’s accusations that an employee of the DOJ intentionally targeted the company and aimed to bring them down, steal their software, or add in spyware.

According to the Bua report—named after the DOJ counsel investigating it—Inslaw’s finances fell apart as the grants that originally funded PROMIS via the Law Enforcement Assistance Administration (LEAA)—a subagency of the DOJ—ended.

Inslaw was originally created as a nonprofit only to later switch to for-profit status, similar to the transformation done by the OpenAI Foundation. As it was funded by government grants, PROMIS was actually in the public domain, and Inslaw had no exclusive rights to it.

Prior to losing its funding, Inslaw received what were called “advanced payments,” wherein they were paid even before their invoices were processed—a unique situation for most contractors. Without government grants, advanced payments, or a unique claim to PROMIS’ intellectual property, Inslaw was unable to privately fund additional work.

Another story of the time detailed how Inslaw was way over their heads for other software projects: unable to complete contracts and far over budget.

In Inslaw’s eyes, the DOJ stole PROMIS and wanted to give it to one of their employees as a gift for their work as part of the Iran-Contra affair. But the DOJ employee, who was also an ex-Inslaw employee harboring a longstanding grudge, had little to do with the project and no known connections to the Iran-Contra affair. More often, government workers are accused of having a positive connection with contracting firms that enable them to receive government contracts.

A large source of the accusations that the government added spyware to PROMIS came from Michael Riconosciuto, a computer programmer that admitted to hacking PROMIS. Riconosciuto would be arrested by the Drug Enforcement Agency (DEA) for methamphetamine production shortly after his accusations.

Few of Riconosciuto’s accusations could be verified. The location where he professed to have edited the software was a largely empty native American reservation with no sign of a computer lab or computers in general.

The investigation did confirm one of Riconosciuto’s accusations: that the reservation was the site of a company working with the defense contractor Wackenhut to bid on the development of night vision goggles for the military. While the project to bid on night vision goggles existed, there was no evidence of a lab to develop night vision goggles on the reservation either.

Potentially the reservation was simply used for its status as a reservation as Alaskan or native companies benefit from 8(a) minority and small business preference in government contracting, like that of Wackenhut’s work with the Alaskan company Alutiiq.

An Israeli-Canadian businessman who previously worked for Israeli intelligence, Ari Ben-Menashe, gave testimony detailing how PROMIS was being sold to governments across the globe, but when interviewed under the Bua investigation he detailed how it was a different version of PROMIS developed by the NSA for signals intelligence, which Israel added backdoors to allowing it access to the systems of banks around the world. In total, the investigation found no basis for his testimony:

After a thorough investigation, the Task Force described Ben-Menashe’s testimony variously as totally lacking in credibility,” “fabricated,” “demonstrably false from beginning to end, “riddled with inconsistencies and factual misstatements,” and “a total fabrication.”

Eventually it was revealed that the DOJ gave a copy of Inslaw’s PROMIS to Dr. Ben Orr, a senior assistant state attorney with the Israeli Justice Ministry who was on an exchange program at the DOJ. When interviewed, Orr detailed that the justice ministry never installed the software as it was foreign-produced and instead contracted with an Israeli company to produce their own version called 2000.

Multiple PROMISs

Much of the confusion surrounding PROMIS was that there were multiple software packages called PROMIS made by different companies. Besides the one developed by Inslaw, another one, Project Management Information System, used by the CIA was developed by Strategic Software Planning Corporation of Cambridge, Massachusetts.

Accusations that the Canadian government was using PROMIS were based on confusion with the Strategic Software Planning Corporation’s software of the same name.

Then there was another software package used by the NSA called Product Related On-line Management Information System for signals intelligence.

Another package unrelated to government work, the Problem-Oriented Medical Information System, also existed at the time.

Origin of Wild-Eyed Claims

While little evidence pointed to a grand conspiracy surrounding the software, there was some legitimate reason that it might be investigated. The original court filing on behalf of Inslaw accusing the DOJ of a grandiose scheme was made by Elliot Richardson—a prominent D.C. attorney who would go on to win the Presidential Medal of Freedom in 1998.

Despite little evidence that the DOJ was trying to sabotage their business, one bankruptcy judge did side with Inslaw, finding that the DOJ intentionally harmed their business. That finding was eventually thrown out on appeal.

Many of the accusations seemed like they were plausible and might have parallels in other cases: compromised software and sensitive data, corporate sabotage and abuse of the DOJ.

But many of the original accusations made no sense. They involved coordinated illegal activity by the federal government—”trickery, fraud, and deceit” in the original affidavit—solely to obtain basic case management software and avoid payment on a relatively minor contract.

The accusations that the government would augment software with spyware and sell it to drug cartels implied that drug cartels would want or need legal case management software.

While Inslaw accused the DOJ of trying to steal their software, Inslaw actively avoided providing the government with a copy of the software they paid for, and instead encouraged them to use a newer, shared version of the software from a remote computer—something unique at the time in the mid 1980s.

All software has the ability to be compromised with spyware and backdoors, but the original version of PROMIS would have little ability to be hacked in a time when most computers had no connection to the internet or even other computers, whereas the newer, remote version that Inslaw wanted DOJ to use would not reside on government property and likely have more security concerns.

That doesn’t jibe with numerous anecdotal descriptions of tent cities appearing in cities following the financial crisis in 2008 that ascribe the growth in homelessness to housing affordability and other economic issues. Academic papers were already being written in 2017 about the rise in “tent wards”—large-scale semi-regulated homeless encampments—to deal with the flood of the unhoused showing up in West coast cities.

More so, much of the decline in homelessness that supposedly occurred during that period happened in Florida. The sunshine state went from almost 58,000 homeless— the majority of which were unsheltered, not in a temporary living situation—to 21,000 in less than ten years with no substantive explanation.

That decline accounted for 40 percent of the national decline over that period. Essentially, Florida, which had a substantial homeless issue for years, suddenly resolved most of the issue with little explanation of how.

Based on a report from the Florida Housing Coalition, it’s because the state’s economy turned around and the government started funding housing initiatives:

First, the economy has improved since 2010. Since homelessness is primarily due to a mismatch between income and housing costs, increasing income and employment opportunities result in decreased homelessness.

Second, the Florida Legislature has increased appropriations to preserve and create affordable housing, including housing for those with special needs and very low incomes. Again, because homelessness results from housing costs that are out-of-reach to those with lower incomes, increasing the stock of affordable housing reduces homelessness.

Third, in recent years Florida communities have embraced evidence-based best practices such as Housing First, collaborative case management, and rapid rehousing. Nationwide, these models have proven successful in decreasing homelessness.

None of those make much sense. To the first point, homelessness had been high prior to the economic collapse when unemployment was relatively low. Total non-farm employment in Florida fell by about a million following the financial crisis.

To the second, major funding for housing support disappeared as homelessness declined. They were still collecting hundreds of millions, but they weren’t appropriating any of it to housing.

The state collects hundreds of millions a year for its housing trust fund based on the 1992 Sadowski Housing Act, which funds programs like the State Housing Initiatives Partnership (SHIP) and the State Apartment Incentive Loan (SAIL) through the documentary stamp tax—a tax on real estate document filings.

But starting in 2009 and through 2014, much of that money got swept into the state’s general fund to plug other budgetary holes. In 2012 and 2014, 100 percent of the housing trust fund was used to plug the budget and none of it got spent on housing—right smack dab in the middle of Florida’s supposed rehousing initiative.

Other metrics, like the count of sheltered homeless barely budged during this period as well, showing no indication that the homeless were being moved into temporary housing, like homeless shelters.

Previously, when the housing programs were receiving and appropriating millions—over $600 million in 2006 alone—it had little affect on homelessness as the point-in-time estimates kept climbing into 2010.

The increase in incarceration may have caused a decrease in crime, but the magnitude is highly uncertain and the results of most studies suggest it was unlikely to have been large.

Crime rates exploded in the 1960s, a lot of it related to the proliferation of heroin and guns against a background of civil strife, some related to deinstitutionalization—the shift away from hospitals and asylums for treatment of mental health and towards independent treatment and prescription drugs.

Within ten years, murder and assault rates doubled. Robbery rates tripled. Drugs were considered the major culprit. In 1970 Congress passed the Comprehensive Drug Abuse Prevention and Control Act, which created the illegal drug scheduling system, to limit access to pharmaceutical drugs that can be abused.

Once it was revealed that 10 to 15 percent of servicemen in Vietnam were now heroin addicts, president Nixon declared a “war on drugs” in 1971. In 1973, the Drug Enforcement Agency was established. Other measures like the Rockefeller drug laws in New York would have a similar effect at the state level. The percentage of state and federal inmates received annually for drug law violations would go from 2.5 percent in 1940 to 9.3 percent by 1980. That would add to the overall incarceration rate, which would grow four-fold between 1970 and 2000.

Yet crime rates stayed elevated into the 1990s as various American cities battled for the title of global murder capitol. It was only in the early 1990s that crime rates would drop significantly, first in New York City, but then suddenly across the country in 1994, particularly for homicide.

Various inconclusive theories have attempted to explain the sudden drop, from lead exposure to abortion laws, cocaine markets, three strikes laws, and the elimination of discretionary parole.

But while the drop in crime in the 1990s was sudden and seemingly inexplicable, the overall crime rate had been slowing down going back to the mid-1970s following the war on drugs and start of mass incarceration based on data from the Federal Bureau of Investigation’s Uniform Crime Reporting program and prisoner data from the Bureau of Justice Statistics.

Crime rates might not have declined significantly right away, but the upward trajectory stopped and the overall the trend was downward beginning in the mid-1970s, particularly for homicide and robbery.

Crime Rates Stopped Growing

Crime didn’t immediately come to a halt, but the sharp growth in crime plateaud—whether by way of incapacitation, intimidation, or simply by coincidence—a few years after more people were being thrown in prison en masse.

Prior to 1975, the total crime rate increased annually by about 1 per 10,000, doubling over the course of ten years. Ten years later, by 1985, the prison population had doubled and the crime rate stopped growing altogether.

Homicide Rates

After growing year over year through the 1960s, the homicide rate dropped in 1976—the first time in 14 years and only a few years after prison rates started growing—trending downwards after that.

If the murder rate kept going as it had been in the 1960s, by 1994 it would have been twice as high as it was—not a weak effect.

Decennial Trend

Because crime rates are somewhat volatile—they can vary from year-to-year—it’s better to look at crime trends over the course of a decade. And there it’s easily visible how the average annual crime rate over the course of a decade initially started declining in the 1970s.

Prison Rate Eventually Followed Crime Rate

While the 1970s is considered the beginning of mass incarceration, when considering the ratio of prisoners to crimes, it was at a relative nadir. That is, what might be considered large-scale incarceration existed prior to the 1960s. When crime grew in the 1960s, the incarceration rate just didn’t keep up.

With increased imprisonment, the ratio of prisoners to total crimes in the 1990s was effectively the same as what it was in the early 1960s before crime rates exploded.

If that doesn’t seem right, you’re not alone. The Richmond Federal Reserve noted how that doesn’t seem to comport with reality in a November 2025 post showing a considerably high vacancies-to-unemployed ratio (V/U rate) of .98. And only a few years prior it was twice that.

Is something fishy going on with job openings? With 7.2 million job openings as of August 2025, vacancies seem high, especially relative to the five-year pre-pandemic average of 6.4 million. Vacancies also seem high relative to unemployed workers.

Another post from the Saint Louis Federal Reserve asked a similar question. If the labor market is so tight, why aren’t wages higher? If workers are in such high demand, then they could demand more money.

All of this is relatively new. Prior to the financial collapse, the V/U rate barely ever got above .7. But beginning in 2010, post-financial collapse, the economy has been continually creating millions of jobs that are not getting filled.

Not only are they not getting filled, they might be sitting around unfilled for years. Even as the economy adds jobs that get filled, the number of job openings rarely goes down even as more and more employees are hired.

Ghost Jobs Aplenty

But there’s good reason for skepticism as many of the openings are likely ghost jobs—fake postings intended for some other purpose besides actually filling a role. Sometimes it’s meant to collect resumes for a job that doesn’t exist just yet. Sometimes it’s done to create a façade of growth for a flailing company.

Various estimates have put the number of ghost jobs between 20 to 40 percent of all jobs or at least 2 million, which is the minimum number of job openings in Bureau of Labor Statistics (BLS) data that never seem to go away. Besides misleading applicants, ghost job openings skew the labor market, making it appear that the economy is more robust and flush with jobs than it is, that wages should be higher, and that there isn’t enough skilled labor to fill the gap.

But there’s good reason to believe that the number of ghost job openings could be much higher, or at least that the Job Openings and Labor Turnover Survey (JOLTS) that the number is based on, is significantly flawed.

That’s because the number of added job openings is unrealistically high. Since 2010, the economy has almost continually added more employment—an average of 2 million jobs added each year—yet the number of job openings keeps going up. What was 2.8 million job openings in 2009 is now regularly over 8 million.

Assuming that every added job—e.g. net hires, or total hires minus total separations—had a job opening, that means that for every job filled there is a new job opening to replace it plus a few more to account for the increasing number of openings.

That works out to a seemingly impossible 193 million net job openings added since 2000—an average of 7.7 million net job openings added per year. Considering that there are currently 156 million jobs in the economy, it would mean more than one job opening added for every job in the economy.

More Reasonable Pre-Crisis

Job opening numbers were not as specious prior to 2010. Before then, the potential number of added or removed job openings varied between plus or minus 5 million a year. That’s because the number of job openings actually declined at various points when people got hired.

If the current number of job openings was similar to what it was back then—around 5 million rather than 10 million—then the V/U rate would also be similar to what it was —about .5. Not a considerably tight labor market.

A grand undertaking, it involved digging a giant underground rainbarrel and a 2.4 mile long tunnel along the river that would ingest thousands of gallons of sewage and stormwater runoff and direct it to the city’s nearby wastewater treatment plant. It would halt the pollution from combined sewage overflows (CSOs)—when excess rainfall overwhelms the sewer system and the overflow goes straight into the river. CSOs are estimated to be the largest contributor to fecal coliform bacteria in the water like Escherichia coli, otherwise known as e.coli.

DC has already borrowed $3 billion for the complete Clean Rivers project through bond offerings, with $253 million spent on the Anacostia River section alone and the tunnel boring machine, nicknamed “Nannie”, costing $25 million.

The vast majority of the Anacostia tunnel was completed in 2018 and is considered a success for preventing 90 percent of CSOs from entering the river, including 19.5 billion gallons of wastewater and 12,265 tons of trash.

Yet despite its success on CSOs, the Anacostia river now shows e.coli levels in the river surpassing Environmental Protection Agency (EPA) standards sometimes two to five times more often than before the tunnel was available based on data from D.C.’s Department of Energy and Environment (DOEE)’s water quality reports to the EPA. The upper Anacostia is now out of compliance a full 92 percent of the time with respect to e.coli when it used to be within range under 40 percent of the time.

Water quality measurements are done across a range of times and locations to avoid bias from isolated, short-term affects like rainstorms and then averaged using a geometric mean. Not all Anacostia measurement locations provided data for the 2018 to 2023 report.

How that is possible despite the cessation of sewage overflows is an open question. The tunnel project appears to have generally improved other metrics of river health like turbidity, dissolved oxygen, and pH.

Potentially Failing Pipes

A 2022 report from Anacostia Riverkeeper gives a different perspective based on their own independent sampling:

For the single sample primary contact standard, Kingman Island (AR3), Buzzard Point (AR-6), and The Washington Channel (AR-7) passed 90% of the time or more. Given their high passing rates, these sites could be good locations to promote recreation.

But at the end of the report, it also admits what might be the source of the problem outside of CSOs:

The high bacterial loads in some of these streams on wet days could indicate more than overland contamination - perhaps a sewer line has been improperly connected to the storm system.

Similarly, the DOEE reports don’t particularly highlight the problem of growing e.coli in the Anacostia, but it is buried in the text:

Chronic E. coli percent exceedances continue to be a problem for the majority of the District’s waterbodies. Fluctuations in these constituents are due to various factors, such as weather and subwatershed activities and conditions, including failing sewer pipes and illicit discharges.

A failing pipe is also what led to the recent breakdown of the Potomac Interceptor, a large junction for wastewater near the Potomac River near Washington, DC. The ensuing spill released 240 to 300 million gallons of untreated sewage into the river and is potentially one of the worst sewage spills in U.S. history.

An NPR story from 2023 detailed how ten years later many in the city lamented the closings, and residents are now forced to deal with fewer opportunities while many of the buildings lay empty.

But data from Chicago Public Schools shows the total number of locations only declining by 30 since 2013, and it happened slowly, not all in one year. And the number of closings was relatively minimal considering that there was a growth of 41 schools in the six years prior.

Even more schools may have been added before that.Based on Department of Education’s National Center for Education Statistics (NCES) data, Chicago’s District 299 added 40 schools since the late nineties. Some schools may have closed since 2013, but the net result was actually more schools.

Ostensibly the school shutdown was intended to save money and plug that $1 billion budget hole, but the school budget has only grown since 2013. What was then a $5.3 billion budget ($7.5 billion in current dollars) is now $10.25 billion based on city budget documents.

According to data from NCES, the number of full-time equivalent teachers in District 299 has only gone up. What was 22,951 in 2003 is currently 23,532. Student-teacher ratios went from 18.93 to 13.77.

Some of this is already well known. Under then-CEO Arne Duncan, who went on to become Secretary of Education under Obama, CPS went on a public charter school creation spree in the early 2000s, while also closing a few traditional schools.

When they started closing schools after 2013, the majority were charter schools and many were only recently created. Of 47 schools that no longer appear in CPS data, 35 didn’t exist in 2007, 21 were charters, and 5 were contract schools.

It’s an important point to make since Chicago’s population has been in stark decline for the last few decades. As families leave the city, the school system has been hemorrhaging students. Over 100,000 school-age residents left the city since 2010 based on U.S. Census data. Average daily attendance is down over 100,000 according to NCES data, possibly indicating some are leaving for private schools.

With little else changing—more schools, similar number of teachers, yet a lot fewer students—it means the teacher-student ratio and spending-per-student ratio has doubled.

While being flush with cash and resources could have been an opportunity for the city’s educational system to turn itself around, Chicago’s education system still struggles. According to the independent research site Wirepoints only 29 percent of Illinois students can read at grade level. There are school districts with 7 percent proficiency in math, while employees earn hundreds of thousands.

For Douglass High School in Chicago’s West side, despite having almost no students—35 out of a capacity of 912—and 1.6 teachers per student and spending $68,203 per student, zero students were proficient in reading in 2023.

SOMAH provides incentives for solar panels on multifamily housing paid for by cap and trade funds. California Governor Newsom’s office has denied that any partisan activity was funded through the program.

Some of those accusations are false. SOMAH couldn’t have given that much to nonprofits since the project hasn’t spent most of the money it’s been allocated yet.

But part of the accusations are true. SOMAH does fund community-based organizations like the California Environmental Justice Alliance (CEJA) and the Environmental Health Coalition (EHC) as part of its marketing budget to promote incentives for rooftop solar. Those groups are often activist organizations also involved in separate get-out-the-vote operations that can walk a thin line between non-partisan philanthropy and partisan political action. For example, besides promoting alternatives to fossil fuels, CEJA lists “[conducting] civic engagement activities to strengthen the [environmental justice] voter bloc across California” as one of their main activities in IRS filings.

Yet totals for community-based marketing in SOMAH’s budget are less than $5.5 million since 2020 on a $1 billion project. It’s a significant amount for funding a nonprofit’s get-out-the-vote campaign, but not a financial drain on a major energy project.

The funding of community organizer groups part of the budget isn’t what’s burdening solar installation. Rather the major financial burden comes from paying out $10 to $15 million a year in rebates to subsidize property owners for solar installations. And rooftop solar is expensive to install.

The Inefficiency of Rooftop Solar

Based on reports, the cost of completed solar panel installations under SOMAH so far is $84 million for 41.942 megawatts of capacity. That works out to $2 million per megawatt of solar capacity.

That’s not too far removed from other common estimates of cost per capacity of solar. The Energy Information Administration (EIA) estimates it at $1,588 per megawatt for all solar types (e.g. utility-scale or rooftop). The Department of Energy estimates an average of over $2.5 million to $3 million.

SOMAH data on individual installations shows a median cost of $560,000 per megawatt just for the panels alone.

But the issue isn’t just the capacity, but the actual amount of energy generated, especially for solar. Solar panels don’t produce energy when the sun isn’t shining, and their actual generation can stray far below their potential capacity, especially for rooftop solar.

Generation numbers for rooftop solar are usually more difficult to come by because of the distributed behind-the-meter nature of rooftop solar. But SOMAH reports for Q1 to Q2 of 2025 give a glimpse behind the meter.

The program tracked 22.9 gigawatt-hours of energy produced in the first half of 2025—45.8 gigawatt-hours annually—based on 39.4 megawatts of capacity. That works out to a solar efficiency rate of 13.2 percent—less than average solar efficiency rates, which hover around 15 to 20 percent.

If the SOMAH program spent $84 million for 41.942 megawatts of capacity, with an efficiency of 13.2 percent, that works out to a whopping $1,730 per megawatt-hour of actual generation annually, easily one of the most expensive sources of energy, even before including other costs like ongoing maintenance, cost of capital, and replacement costs.

Most levelized costs of energy (LCOE) calculations list utility-scale sources with rates below $100 per megawatt-hour. EIA shows nothing above $200 per megawatt-hour. Department of Energy has a larger scale of estimates that includes lifetime costs, but even that tops out at $600 per megawatt-hour.

While solar is commonly promoted as one of the cheapest energy sources because of declining photovoltaic panel costs subsidized by China, that only applies to utility-scale solar where thousands of panels are installed at scale and panels are attached to motors that track with the sun. Rooftop solar, where solar tracking is rare and installation costs are higher, is substantially more expensive.

It’s likely the most accurate survey of how many are considered homeless at one time—either living on the street (unsheltered) or in some kind of temporary housing (sheltered).

When the pandemic hit in 2020, COVID restrictions hampered the survey out of concern for public health. As a result, data on total unsheltered homeless population in 2021 would show an enormous drop, almost 200,000, from a lack of reporting.

Fifty percent of that drop was because of a lack of reporting in California alone. Not just because California as a state is large, but it also has an outsized homeless population—27 percent of national total in 2020—and it limited its homeless survey the most in 2021. Plenty of other states—Alaska, Delaware, Virginia, and Rhode Island among others—surveyed their homeless population in 2021 with no issue related to COVID and saw little to no change in their numbers.

According to a report from California, homelessness likely declined during the pandemic in California as the state actively arranged to house individuals through Project Roomkey—a temporary measure to give housing to anybody who tested positive for COVID or similar health concerns, or over the age of 65.

Supposedly 62,000 individuals were given temporary housing as part of the program, with a peak of 16,000 committed hotel/motel rooms in use. Over 12,422 were housed in Los Angeles County alone, which is actually relatively low considering that Los Angeles has the highest homeless population in the U.S. and represented around 40 percent of California’s homeless (69,144) at the time—although the Project Roomkey estimate puts it at a higher 91,145 total homeless (65 percent)—yet L.A. represented only 20 percent of Project Roomkey participants.

The state saw no real change in the sheltered population—the kind that would fall under the housing provided by Project Roomkey and wasn’t affected by COVID limitations. There were less than 33,000 sheltered homeless in the state at the time anyhow so growth by 62,000 would certainly show in the numbers somehow.

The project ended in summer of 2022 and all temporarily housed individuals since left where they were staying. The HUD PIT survey is all done in January, so even if a few thousand housed individuals were still there, that would likely show up in the 2022 estimates for unsheltered, but no such decline is apparent.

But it’s unlikely that they are handling that much, or if they are, they are not getting paid for it.

For example, the annual report for 2024 shows an operating revenue of $978 million based on water and sewer services for D.C. and just wastewater services for nearby Maryland and Virginia.

Based on recent charging rates posted on their site for just handling wastewater of $16.14 per 1,000 gallons for residential, that would put the daily wastewater usage at 166 million gallons as an upper estimate—about half of what DC Water is estimating. And that’s without factoring in all other sources of revenue besides wastewater usage (e.g. treated water, fees).

DC Water’s water delivery numbers don’t really add up either. From 2015 to 2020, the utility reported delivering millions of gallons less per day on average when D.C.’s population was consistently growing by tens of thousands.

When the pandemic hit and D.C. saw a sharp population drop, water usage somehow grew as if fewer people started suddenly using more water.

Sewage Spill, Sewage Overflows, and Clean Rivers Project

Currently DC Water is the focus of what might be the largest sewage spill in U.S. history as a major pipe called the Potomac Interceptor broke in January spilling millions of gallons of sewage into the Potomac river.

While the details of the break are still being investigated, DC Water is being scrutinized for how it was spending money.

DC Water was in the midst of repairing the pipe close to where the break occurred as part of various system improvements, including the large scale Clean Rivers project. The latter is a $3 billion initiative to create multi-mile tunnels that can absorb combined sewage overflow (CSO)—sewage that gets sent into the Potomac or Anacostia rivers during periods of high rain because there is too much volume at one time for the system to handle.

CSO was already responsible for millions of gallons of sewage flowing into those rivers on an annual basis. The first part of the Clean Rivers project to tunnel underneath the Anacostia was completed in 2016 and the next portion dealing with the Potomac began in 2019.

Bond Pre-Funding and User Fees

DC Water is funding the project largely through bonds, but then paying back the bonds through the Clean Rivers Impervious Area Charge (CRIAC) on customer bills, which now is on the order of $24 per equivalent residential unit. While the Potomac portion has just started, DC Water already made over $3 billion in bond offerings—likely a pre-funding of the project in advance—and is paying $146 million a year in interest payments—26 percent of its expenses for the year.

Altogether, with the CRIAC fee, DC Water bills now average around $147 per month for consumption of 4,024 gallons. While that is likely extremely high for the U.S., DC Water shows it as being average considering the high average income for the D.C. area. Yet customers are substantially further behind on payments—customer receivables, net of allowance for doubtful accounts—which was $49 million ten years ago now stands at $114 million.

Even with the high user fees, DC Water has made little headway in paying off the debt. The most they paid down was in 2024 when liabilities decreased by $123 million—slightly less than their annual interest payments for the year. Since 2014, they have taken on about $200 million in debt a year.

Similarities to Jackson, MS

The story of a public water utility potentially falsifying numbers in the midst of sewage leaks and system repairs in very similar to that of Jackson, Mississippi.

In that city, records for water usage suddenly plummeted to a fraction of what they previously were following changes to the billing system. Similar to D.C., water usage didn’t follow distinct population trends.

In Jackson, the city was hemorrhaging residents, potentially related to sharp increases in water bills needed to help fund the ailing system, but the statistics kept showing high water use as if nobody had left.

With fewer residents to pay the bills, the system wasn’t bringing in enough revenue to pay for necessary repairs, leading to the complete collapse of its treatment plant and a citywide boil advisory for days in 2022.

Previously, Medicaid spending data was aggregated to totals that don’t give details about which providers are receiving the spending. This new data provides a much more granular view of how Medicaid funds are really being spent.

And without much effort it’s easy to see how egregious some of the charges are. Not just a couple standard deviations outside the median but ten and twenty times the average.

For example, the average spending to a provider for syringes (HCPCS code:A4657) in a month is less than $10.

Yet in 2018 Medicaid gave U.S. Renal Treatment Centers, Southeast over $4.7 million for syringes for 22 recipients—an average spend of $215,314 per recipient.

For nursing care (HCPCS: S9124), the average spend per beneficiary per month is $9,029. Yet Quality One Care Home Health in Silver Spring Maryland was spending $288,058 per beneficiary per month.

Miscellaneous dialysis charges (HCPCS: 9099) rarely top $3,000 per person per month, but somehow U.S. Renal Care in Kapolei, Hawaii was charging Medicaid over $200,000 per person.

While the data release to the public is new, none of this is likely new to HHS. All of these providers are listed in an HHS Flagged Service Providers data sheet for exactly what you would expect: charges far beyond average costs.

U.S. Renal Treatment Centers was prosecuted under the False Claims Act by the Justice department for Medicaid overcharging related to dialysis medicine, but that was back in 2013. All of the recently released is after 2018.

Excess spending in the data does not necessarily imply fraud or misconduct but could simply be issues with the data and how it is reported.

Just this last week, the Washington Post repeated a similar conclusion:

Many well-intentioned people answered these appeals, resulting in a boom in backyard apiaries. Celebrities such as Morgan Freeman, Jennifer Garner and Beyoncé reportedly took up beekeeping.

But it was all based on a fallacy.

Honeybees were never in existential trouble. And well-meaning efforts to boost their numbers have accelerated the decline of native bees that actually are.

CCD is a confirmed predicament for beekeepers, otherwise known as Isle of Wight disease, where bee colonies suddenly die off with no distinct explanation. But the idea that bee AIDS caused by pesticides was ravaging honey producing populations never had any basis in fact.

Honey production and honey producing bee colonies were in steep decline for years prior to the first recognition of CCD and actually stabilized after it was recognized.

Recently a whistleblower within the current governor Wesley Moore’s administration leaked to Fox 45 Baltimore that the state intentionally kept up a high error rate to delay federal penalties. Following recent changes under the Big Beautiful Bill, states can avoid new penalties for SNAP overpayments for a few years if they show they are actively remedying the problem.

States with an overpayment rate above 10 percent could be forced to pay back 15 percent of benefits. Overpayments between 6 to 8 percent would have to pay back 5 percent of benefits. States are also on the hook for subsidizing administrative costs of the program. If Maryland didn’t get their rate below 10 percent in time, they would be on the hook for close to $300 million in SNAP spending according to Maryland Matters. Since 2022, error rates have steadily declined so they might be in the clear.

Moore has denied the allegations of intentional error rates and refused to investigate the issue, which he ascribed to staff turnover who were unfamiliar with the system.

But there’s no sign that the excessively high error rate in 2022 was related to large-scale fraud or errors. Error rates are based on SNAP cases that are flagged for investigation, not necessarily those eventually found to be linked to fraud, error, or theft. They may be based on a subsample of cases that year, not all SNAP recipients.

Based on U.S. Department of Agriculture (USDA) state activity reports for SNAP, total fraud in 2023 for Maryland—the highest in reported records—was only $938,964—less than .1 percent of total issuance for the year. And that was for all types of inaccurate payments—fraud, household errors, and agency errors—the latter of which was only $518,572.

Essentially, Maryland suddenly started flagging a lot of cases as errors that did not wind up being fraud or errors. If there truly was a 35.56 improper payment rate, the state would have accidentally paid out $600 million a year—based on $1.68 billion in total SNAP issuance for 2023. That money would be owed back to the USDA and potentially sink Maryland’s budget.

SNAP Theft Since 2023

Yet, rather than agency fraud, the state is now dealing with an avalanche of SNAP theft. A separate report from the USDA’s Office of the Inspector General (OIG) researching Maryland’s SNAP theft repayment program shows that the state “replaced more than $22.6 million of SNAP benefits stolen between October 2022 and December 2024,” more than 11 times the annual SNAP errors and fraud listed in prior years.

That report appears to be the only data for 2024 published so far. One anecdotal reference pegged prior annual theft totals at around $1 million a year.

In 2023, Maryland passed House Bill 502 and signed into law by Governor Moore that required the state to replace 100 percent of benefits stolen on or after January, 2021.

That detail is still true—that current year budget documents showed increasing expenditures under Moore—but it turns out that Maryland repeatedly revised its historical expenditures after the fact. More recent budget documents now show a sharp increase in spending happening in previous years under Moore’s predecessor, Larry Hogan, who left office at the end of 2022.

For 2021 through 2023, total state expenditures would be revised upwards by between $5.8 billion to $11.5 billion a year. What was originally $49 billion in spending for 2022 eventually landed at $61 billion based on budget documents published in 2025.

By 2024, the revisions were done and expenditure estimates published in 2026 were effectively the same as what they were when published in 2024—within $1.6 billion.

In general, budgets regularly get adjusted because of unknown expenses at the time, line items that get ascribed to the different years, or inconsistent accounting. Accounting standards usually define materiality thresholds for error based on what kind of metric is being measured—e.g. assets, income etc..—that all range within 10 percent.

But a 23 percent adjustment as the case for Maryland in 2022, would go substantially beyond that. Budget documents detail that Maryland held a budget surplus post-COVID-19 because of federal relief spending, but lackluster growth had eaten into that surplus.

Not only were expenditures restated, but they were substantially higher than previous years, by $15 to $20 billion. In particular, additional billions were moved into Elementary & Secondary Education, Human Resources, and Health.

Spending on Schools

Of the additional $4.3 billion put into elementary and secondary education—a 62 percent increase in state subsidies—most of it went to Baltimore, with an additional $126 million split between Baltimore City and Baltimore County.

Frederick County was next with an additional $44 million, although that was largely because the county has seen a large upswing in population. On a per pupil basis, it was somewhat further towards the back.

But ironically enough, Homan was hired by the Obama Administration in 2013 and that administration awarded him the Presidential Rank for Distinguished Service in 2015 for leading Enforcement and Removal Operations (ERO) and spearheading the management and reform of the largest civil detention system in the nation. Under his helm as an executive at ICE and for which he was given the award, Homan helped remove 316,000 persons in the U.S. who were likely here illegally with little outcry from pro-migration activists at the time.

The irony may go deeper as the period around Homan’s award was at the lowest point of illegal migrant enforcement of the last decade. The 316,00 removals were significantly less than the 413,000 a few years earlier in 2013 when Homan started his position.

According to Department of Homeland Security (DHS) data, DHS returned 1.7 million to their home countries in 2000. Fifteen years later in 2015 it was less than one-tenth that amount—130,000. While that trend pre-dated Homan and Obama’s tenure, the number of removals—returns based on an order of removal, which had been rising throughout the decade, started distinctly declining around 2012 right before Homan was nominated for his post at ICE in 2013 and would be 25 percent lower by 2015.

Those numbers would decline further through 2019, most of it under Homan’s tenure as director of ICE, which ended in 2018.

The House Oversight Committee is reportedly investigating representative Ilhan Omar’s finances. Here is the original Investigative Economics’ story on the subject.

Similar to the Investigative Economics story on Medicare spending following enactment of the ACA, the Government Accountability Office published a report last month on fraud and abuse of subsidies under the ACA.

In a similar vein, any readers familiar with our story on the perverse incentives of the ACA’s medical loss ratio rule and how it was encouraging high premiums and consolidation in the medical world should read this recent paper in Health Affairs Forefront which echoes that same sentiment.

Readers of the recent story on the decline of the oil refining industry in California might want to also read this December 2025 preprint of a paper on a similar topic from University of California researchers. The gist is similar: California’s policies have driven refining out of the state and towards foreign imports, but there are some additional details here that aren’t in the Investigative Economics story.

A recent story on the Arabella Network of philanthropies and their funding of political campaigns has some similarities to a recent rule passed in Ohio.

In July, an appeals court upheld a law banning political donations from foreign entities. Specifically, it changed the definition of a foreign entity to include lawful naturalized migrants like those holding green cards, and also applies to indirect contributions targeting ballot initiatives.

The law appears to almost directly target the Arabella Network’s Sixteen Thirty Fund and other philanthropies connected to the Swiss billionaire Hansjorg Wyss. In the Ohio Secretary of State’s press release on the rule, it even goes so far to include an infographic about Wyss’ funding of political advocacy groups in the state.

Ballot initiatives in Ohio have been big money for a number of years, with spending in 2017 almost hitting $87 million—getting close to the large dollar numbers seen in California ballot initiatives. Most of those California initiatives tend to fail but can be considered successful in how they funnel large amounts of cash into the campaign finance system at large.

The most recent ballot initiative in Ohio in 2025—Issue 2 on raising a state bond for construction—only saw a few hundred thousand in spending—so potentially the rule has already cut off the cash, but it’s hard to tell. Not all Ohio ballot initiatives are multi-million dollar affairs.

Similarly, Nebraska’s attorney general pointed out the potential for foreign influence via ballot initiatives in his state.

A recent Chronicle of Higher Education story highlights the same issues of grade inflation found in an earlier Investigative Economics story: declining ACT scores as grades go up. The Chronicle story misses some nuance to the topic, that the sharp decline in ACT scores is likely related to changes to the exam that occurred at that very same time rather than a sudden drop in student intelligence.

Elon Musk on a recent episode of the Joe Rogan podcast talks about the same aspect of gerrymandering that was first written about in Investigative Economics: that non-citizens that can’t vote can still skew elections by changing how districts are apportioned because they are included in the Census counts that determine voting districts.

The improving health of the Great Barrier reef got some recognition in an article from the International Journal of Global Warming back in July. This is the same trend highlighted by Investigative Economics in 2024, but interesting to see it detailed in a journal that might not otherwise cover a subject that might run counter to other global warming theories.

The recent tragedy at Brown University has drawn some attention to what might have been the perpetrator’s motives—potentially a grudge against the school related to recent budget cuts.

University spending is an ongoing contentious topic with the blame commonly pointed at bloated administrative staff as the culprit.

But a prior Investigative Economics story noted how similar complaints aimed at Yale were unfounded. That non-faculty staff ratios were in line with what they were in the past with no sign of large growth in salaries. Instead, faculty costs were the ones that were growing.

And a quick glimpse at the numbers for Brown show a similar albeit different story.

Ratios of administrators to total staff are currently in line with what they were in 2015. Salary costs have actually gone down as a percentage of total expenses. Instead there are other costs like graduate student support and purchased services adding to the totals.

Originally established by Charles Keeling—the climate scientist given credit for confirming human-based global warming—while working at the Scripps Institution for Oceanography at the University of California, which was heavily funded by the major oil companies of the day.

It’s also regularly the data source that atmospheric carbon dioxide is growing unsustainably, leading to excessive global warming. Numbers from Mauna Loa have shown carbon dioxide levels consistently growing year over year without end, implying that natural sources of carbon dioxide absorption—carbon sinks like oceans and forests—are not sufficient to buffer human-based carbon emissions and therefore catastrophe is inevitable.

But Mauna Loa’s consistent growth is almost too consistent. If Mauna Loa is ostensibly a measurement of CO₂ in the atmosphere largely from human emissions, there should be significant varying fluctuations from industrial output, car emissions, and population growth.

Estimates of anthropogenic CO₂ emissions—i.e. human-sourced emissions—tend to increase year over year as the global human population grows linearly and there is more industrial consumption, but there are still periods of little growth, particularly around the COVID-19 pandemic or during severe recessions that limit industrial output that would ostensibly show up in CO₂ measurements.

In particular, little change in measurement happened in 2022 when the observatory was shut down due to the eruption of the Mauna Loa volcano. Volcanoes are one of the largest sources of CO₂ emissions, so a volcano erupting right next to the CO₂ observatory would likely show up in the data.

That consistency also applies to other Mauna Loa measurements related to climate change like methane and nitrous oxide. NOAA has a note on its site explaining why COVID-19 might not have an affect on measured CO₂ levels as there is no photochemical destruction of atmospheric CO₂ as there is for other gases.

But no matter what, Mauna Loa shows such a predictable measurement that a two-dimensional polynomial can accurately represent values within two-tenths of a percent. It increases year over year whether or not volcanoes erupt in front of it, oil consumption collapses, or industrial output comes to a halt.

For measurements in January, the following equation will predict the Mauna Loa CO₂ value given the number of years since 1959:

The measurements vary from month to month, but a similar two dimensional polynomial equation works for any month. This is the one for June:

That all came to a sudden halt in the late seventies. No nuclear plant would seek a new permit until the mid-2000s. While some plants already approved would still be constructed in the following years, many already approved plants would be canceled.

The common explanation is that the partial meltdown incident at Three Mile Island in 1979 scared off public support for nuclear power. The threat of a catastrophe near Philadelphia, combined with anti-nuclear environmental protests and the release of the movie The China Syndrome a week later—about an almost nuclear catastrophe eerily similar to Three Mile—stopped any demand for nuclear power.

The NRC immediately enacted a one year moratorium on nuclear permits. Nuclear safety regulations were increased, leading to further permitting delays, and power companies entered into new insurance pools to cover potential accidents. Besides the protests and class-action lawsuits surrounding Three Mile Island itself, anti-nuclear sentiment from environmental activists became more heated nationwide, which led to protests, concerts, federal and state litigation, and testimony against upcoming nuclear projects. The combined cost and burden of building a nuclear plant became seemingly untenable.

Proposed plants in Bailly, Indiana, Marble Hill, Indiana, Erie, Ohio, and Midland, Michigan would be abandoned over citizen opposition, lawsuits, and delays.

While Ohio Edison stood by the safety of nuclear energy, delays meant the project was dead in the water:

“The political and regulatory uncertainties affecting the future construction of nuclear plants has intensified following the accident at Three Mile Island. Nuclear construction scheduled in the future carries greater uncertainty of eventual cost.”

Three Mile Island certainly did delay nuclear construction, but construction times for nuclear plants were already growing year over year for the last decade. Changing technology and changing standards for first-of-a-kind projects meant more difficulties in building a nuclear plant—something inherent in many nuclear energy projects that continues to this day.

In 1970, average time between construction permit and operating license was 3.2 years based on NRC operating reactor data. By 1978, it was 6.7 years. By 1988, it was 12.3 years.

More Than Just Three Mile Island

Anti-nuclear sentiment and regulations alone wouldn’t completely explain why projects that had already been permitted were canceled. While 36 projects would be canceled within a few years after it happened, nine proposed plants were canceled in the two years before the incident.

Oil Crisis, Interest Rates, Recession, and Declining Energy Demand

Not just anti-nuclear sentiment and regulations, lack of demand and financial concerns were also significant issues.

While oil and gas prices spiked after the 1973 oil crisis, potentially making nuclear power a cheaper option for electricity generation, high prices also meant a drop in electricity consumption, particularly for the industrial sector, which is unlikely to produce as much when prices are high.

That drop was even more severe in 1979 during the second oil crisis. High oil and gas prices limited industrial output and industrial electricity consumption along with it. When combined with historically high interest rates driven by Paul Volcker and the Federal Reserve’s attempts to wring out inflation from the economy and the ensuing recession, 1982 would be the largest single year decline in industrial electricity sales—tied with the financial crisis in 2009. High interest rates would also hamper any nuclear project leaning heavily on borrowed money for their funding.

Altogether, the original reason to build a nuclear plant to satisfy increasing energy demands was not there, more difficult to justify financially, and much of the planned nuclear capacity was no longer necessary.

For example, the Montague nuclear plant was planned to be a 1.1 gigawatt nuclear reactor in Montague, Massachusetts permitted for construction in 1973. It saw continuous protests throughout the 1970s led by Sam Lovejoy, a figurehead in the anti-nuclear movement who would go on to help found Musicians for Safe Energy (MUSE). MUSE would organize high profile anti-nuclear concerts with the likes of Bonnie Raitt, Graham Nash, and the Doobie Brothers.

The plant was eventually canceled in 1981 after $23 million had already been spent and nothing to show for it, not because of anti-nuclear sentiment but because the region no longer needed the extra capacity.

The Illinois Power Company canceled its plans for the Clinton 2 nuclear plant in 1985 as any unserved demand would likely be met with the Clinton 1 plant already in construction and in operation in 1987.

The Virginia Electric and Power Company (VEPCO), now known as Dominion Energy, stood by the potential for nuclear energy in 1980 despite the downturn in demand, the licensing moratorium, extensive emergency drills, and increased safety regulations to open their North Anna 2 power station in 1980—the first licensed plant after Three Mile Island—but they would cancel plans for North Anna 4 and severely delay plans for North Anna 3.

A proposed nuclear plant in Forked River, New Jersey had been on the edge of survival for some time because of financial struggles following public opposition and declining energy consumption from the first oil crisis in 1973. Once Three Mile Island happened and interest rates shot up, its cancellation was all but guaranteed.