Category: Science

People hospitalized with seasonal influenza can suffer long-term, negative health effects, especially involving their lungs and airways.

The new study comparing the viruses that cause COVID-19 and the flu also reveals that in the 18 months after infection, patients hospitalized for either COVID-19 or seasonal influenza faced an increased risk of death, hospital readmission, and health problems in many organ systems. Further, the time of highest risk was 30 days or later after initial infection.

Since the COVID-19 pandemic began, extensive research has emerged detailing the virus’s ability to attack multiple organ systems, potentially resulting in a set of enduring and often disabling health problems known as long COVID.

“The idea that COVID-19 or flu are just acute illnesses overlooks their larger long-term effects on human health.”

“The study illustrates the high toll of death and loss of health following hospitalization with either COVID-19 or seasonal influenza,” says senior author Ziyad Al-Aly, a clinical epidemiologist at Washington University in St. Louis. “It’s critical to note that the health risks were higher after the first 30 days of infection. Many people think they’re over COVID-19 or the flu after being discharged from the hospital. That may be true for some people. But our research shows that both viruses can cause long-haul illness.”

The findings appear in The Lancet Infectious Diseases.

The statistical analysis spanned up to 18 months post-infection and included a comparative evaluation of risks of death, hospital admissions, and 94 adverse health outcomes involving the body’s major organ systems.

“A review of past studies on COVID-19 versus the flu focused on a short-term and narrow set of health outcomes,” says Al-Aly, who treats patients within the VA St. Louis Health Care System and is an assistant professor of medicine at Washington University. “Our novel approach compared the long-term health effects of a vast array of conditions. Five years ago, it wouldn’t have occurred to me to examine the possibility of a ‘long flu.’ A major lesson we learned from SARS-CoV-2 is that an infection that initially was thought to only cause brief illness also can lead to chronic disease. This revelation motivated us to look at long-term outcomes of COVID-19 versus flu.

“We wanted to know whether and to what degree people with flu also experience long-term health effects,” Al-Aly says. “The big answer is that both COVID-19 and the flu led to long-term health problems, and the big aha moment was the realization that the magnitude of long-term health loss eclipsed the problems that these patients endured in the early phase of the infection. Long COVID is much more of a health problem than COVID, and long flu is much more of a health problem than the flu.”

However, the overall risk and occurrence of death, hospital admissions, and loss of health in many organ systems are substantially higher among COVID-19 patients than among those who have had seasonal influenza, Al-Aly says. “The one notable exception is that the flu poses higher risks to the pulmonary system than COVID-19,” he says. “This tells us the flu is truly more of a respiratory virus, like we’ve all thought for the past 100 years. By comparison, COVID-19 is more aggressive and indiscriminate in that it can attack the pulmonary system, but it can also strike any organ system and is more likely to cause fatal or severe conditions involving the heart, brain, kidneys, and other organs.”

The researchers analyzed de-identified medical records in a database maintained by the United States Department of Veterans Affairs, the nation’s largest integrated health-care delivery system. They evaluated information involving 81,280 patients hospitalized for COVID-19 at some point from March 1, 2020, through June 30, 2022, as well as 10,985 patients hospitalized for seasonal influenza at some point from October 1, 2015, through February 28, 2019.

Patients represented multiple ages, races, and sexes.

Regarding both viruses, patient vaccination status did not affect results. Those in the COVID-19 cohort were hospitalized during the pre-delta, delta, and omicron eras.

During the overall 18-month study period, patients who had COVID-19 faced a 50% higher risk of death than those with seasonal influenza. This corresponded to about eight more deaths per 100 persons in the COVID-19 group than among those with the flu.

Although COVID-19 showed a greater risk of health loss than seasonal influenza, infection with either virus carried significant risk of disability and disease. The researchers found COVID-19 exhibited increased risk of 68% of health conditions examined across all organ systems (64 of the 94 adverse health outcomes studied), while the flu was associated with elevated risk of 6% of health conditions (six of the 94)—mostly in the respiratory system.

Also, over 18 months, COVID-19 patients experienced an increased risk of hospital readmission as well as admission to an intensive care unit (ICU). For every 100 persons in each group, there were 20 more hospital admissions and nine more ICU admissions in COVID-19 than flu.

“Our findings highlight the continued need to reduce the risk of hospitalization for these two viruses as a way to alleviate the overall burden of health loss in populations,” Al-Aly says. “For both COVID-19 and seasonal influenza, vaccinations can help prevent severe disease and reduce the risk of hospitalizations and death. Optimizing vaccination uptake must remain a priority for governments and health systems everywhere. This is especially important for vulnerable populations such as the elderly and people who are immunocompromised.”

In both COVID-19 and the flu, more than half of death and disability occurred in the months after infection as opposed to the first 30 days, the latter of which is known as the acute phase.

“The idea that COVID-19 or flu are just acute illnesses overlooks their larger long-term effects on human health,” Al-Aly says. “Before the pandemic, we tended to belittle most viral infections by regarding them as somewhat inconsequential: ‘You’ll get sick and get over it in a few days.’ But we’re discovering that is not everyone’s experience. Some people are ending up with serious long-term health issues. We need to wake up to this reality and stop trivializing viral infections and understand that they are major drivers of chronic diseases.”

Source: Washington University in St. Louis

Black patients are significantly less likely than white patients to receive the gold standard of stroke care, according to new research.

Almost 800,000 Americans suffer a stroke each year, according to the Centers for Disease Control and Prevention.

African Americans and other people of color have a substantially higher risk of experiencing a stroke than their white counterparts. And they’re also significantly more likely to die from those strokes.

“Racial disparities exist in all levels of stroke care,” says Delaney Metcalf, a third-year medical student in the Augusta University/University of Georgia Medical Partnership and lead author of the study in the Journal of Stroke and Cerebrovascular Diseases.

“There are many studies that show quality of medical care in general can be poorer in minority populations. But as health care professionals, we are not doing a good enough job of getting these lifesaving treatments to these patients.”

The researchers analyzed data from more than 89,000 stroke patients across the US and found that Black patients were significantly less likely to receive tissue plasminogen activator (tPA), a clot-busting medication that helps restore blood flow to the brain after a stroke.

Additionally, Black patients were less likely than white patients to undergo endovascular thrombectomy (EVT), a minimally invasive procedure in which the blood clot is surgically removed.

When minority patients do get needed treatment, the researchers found non-white patients had substantially longer hospital stays, which may signify poorer health outcomes or lower quality of care.

tPA and EVT are the go-to medical treatments for ischemic strokes, which are caused by blood clots. They dramatically reduce death after a stroke and lead to significantly better health outcomes.

But both options are extremely time sensitive.

To be effective, tPA should be given within a few hours of a stroke. And patients requiring the EVT surgical procedure need to be on the table within about six hours. For rural patients or those living in underserved areas, that’s a tall ask.

Previous research has shown that Black and minority patients are less likely to call emergency services for an ambulance, which can delay medical care significantly.

But even once they arrive, Black and minority patients experience longer wait times to be seen by health care providers to get brain imaging. Brain scans are necessary to determine a treatment course.

“Why is that? It could be that the hospital is really overrun and doesn’t have the staff to be able to do it in a timely manner,” Metcalf says. “There can also be implicit racial biases that lead to health professionals not treating their Black patients the same way they treat their white patients and not taking minority patients’ symptoms seriously.”

And not every hospital has the staff, medicine, and equipment to make sure stroke patients receive this specialized care. That lack of access can have deadly consequences.

“This is a problem, but it’s a targetable problem,” Metcalf says. “Increasing education on what a stroke looks and feels like is one tiny thing we can make an improvement on.

“There are many small things we can do to make advances in minority stroke care. Increasing community education on recognizing stroke symptoms may help patients get to treatment centers faster. Additionally, providing the training and technology needed for these treatments to underserved areas can improve access to stroke care.”

Donglan Zhang, a former assistant professor at the University of Georgia now at NYU Long Island School of Medicine, is the study’s coauthor.

Source: University of Georgia

Trees growing in wetter regions are more sensitive to drought, according to a new study.

Scientists have long debated whether arid conditions make trees more or less resilient to drought. It seems intuitive that trees living at their biological limits will be most vulnerable to climate change, since even just a little extra stress could tip them past the brink.

On the other hand, these populations have adapted to a harsher setting, so they might be more capable of withstanding a drought.

According to the study in the journal Science, greater water availability could “spoil” trees by reducing their adaptations to drought.

“And that’s really critical to understand when we’re thinking about the global vulnerability of forest carbon stocks and forest health,” says ecologist Joan Dudney, an assistant professor at the University of California, Santa Barbara Bren School of Environmental Science & Management and in the Environmental Studies Program. “You don’t want to be a ‘spoiled’ tree when facing a major drought.”

Dudney and her coauthors expected trees growing in the most arid regions to be more sensitive to drought, since they’re already living at the edge of their limits. What’s more, climate change models predict that these regions will experience more rapid drying than wetter regions. This shift in climate could expose trees to conditions beyond their adaptive capacity.

To measure drought sensitivity, the authors analyzed 6.6 million tree ring samples from 122 species worldwide. For each year, they measured whether the tree grew faster or slower than average based on its ring width. They linked these trends with historic climate data, including precipitation and temperature.

The team then compared drought responses across different regions.

“As you move to the drier edge of a species’ range, trees become less and less sensitive to drought,” says lead author Robert Heilmayr, an environmental economist also in the environmental studies program and at the Bren School. “Those trees are actually quite resilient.”

Dudney, Heilmayr, and coauthor Frances Moore were inspired, in part, by the work of UCSB professor Tamma Carleton on the effects climate change has on human populations.

“This paper highlights the value of cross-disciplinary scientific work,” says Moore, an associate professor at UC Davis. “We were able to adapt methods from economics originally developed to study how people and businesses adjust to a changing climate and apply them to the ecological context to study forest sensitivity to drought.”

“A heatwave is likely to kill more people in a cool place like Seattle than in hotter cities like Phoenix,” Heilmayr says. The Southwest is already quite hot, so heatwaves there are scorching. But the region’s cities are adapted to an extreme climate, he points out.

Now we know that forests display similar trends. Unfortunately, warmer regions are slated to get disproportionately drier in the coming decades.

“There is a pretty large portion of species’ ranges that are going to face a completely novel climate, something that those species don’t see anywhere in their range today,” Heilmayr explains.

The authors found that 11% of an average species’ range in 2100 will be drier than the driest parts of their historic range. This increases to over 50% for some species.

“Broadly, our research highlights that very few forests will be unaffected by climate change,” Dudney says. “Even wetter forests are more threatened than we thought.”

But there is a flip side of the coin. Species have a reservoir of drought-hearty stock in the drier parts of their range that could bolster forests in wetter areas.

Previous research from UC Santa Barbara revealed that many species do have the capacity to adapt to environmental change. However, those researchers also found that trees migrate slowly from one generation to the next. That means human intervention—such as assisted migration—may be necessary in order to take advantage of this genetic diversity.

Source: UC Santa Barbara

Precipitation levels during nesting season are not related to reproductive success for wild turkeys, according to a new study.

The findings runs counter to conventional wisdom regarding the role that rainfall plays in wild turkey nesting success and shed new light on how climate change may affect wild turkey populations.

“We wanted to know how weather influences nesting success right now, and then use that data to assess how climate change may influence wild turkey populations in the future,” says Wesley Boone, a postdoctoral researcher at North Carolina State University and corresponding author of the study published in the Journal of Wildlife Management.

“Wild turkeys are fairly tolerant of a wide range of conditions, but there are a host of factors that can affect their reproductive success,” says coauthor Chris Moorman, a professor of forestry and environmental resources at NC State. “This work focused on two of those conditions, precipitation and temperature, and how they may influence nest survival during the incubation period.”

For the study, researchers focused on daily nest survival, which is whether the eggs in the nest survive any given 24-hour period. Over the course of eight years, researchers monitored 715 turkey nests and collected daily precipitation and temperature data for each nest during the entire incubation period.

For temperature, the researchers looked specifically at the extent to which temperatures at each nest varied from historical averages.

The researchers analyzed all of this data to determine the extent to which precipitation and temperature were associated with daily nest survival.

“The most surprising finding was that precipitation during nesting was not a good predictor of daily nest survival,” Moorman says. “It had been widely believed that particularly rainy weather made it more likely that eggs wouldn’t survive.”

“We also found that temperatures which were higher than historical averages were associated with higher rates of daily nest survival during incubation,” says Boone. “Peak nesting season is generally in April, so we’re talking about warmer than average spring weather.”

“Taken by itself, this might suggest that climate change could benefit turkey reproductive success and, by extension, turkey populations,” Moorman says. “However, we also looked at precipitation and temperature data for the months leading up to nesting season, and at the overall likelihood that a turkey nest will successfully hatch at least one egg. And when we looked at both of those datasets, things get a lot less clear.”

“For example, the data suggest that more precipitation in January—long before nesting season—is associated with greater nest survival,” Boone says. “The data also suggest that higher temperatures in January are associated with worse nesting survival.

“But there is so much uncertainty related to those findings that it’s not clear whether there’s a real relationship there, or if it’s an anomaly. However, it does temper any enthusiasm we might have about the likelihood that climate change will benefit turkey populations.”

Additional coauthors are from NC State, the University of Georgia, Louisiana State University, and the US Geological Survey.

Source: NC State

Most parents of infants and young children in the United States plan to have their children vaccinated against this year’s “tripledemic,” according to a new study.

In addition, 40% percent of parents intend to have their children vaccinated against COVID-19.

For the study, published in the journal Vaccine, the researchers asked parents about their intention to vaccinate their children against COVID-19, influenza, and RSV—the so-called “tripledemic” that experts predict for this fall and winter.

Simon F. Haeder, an associate professor of health policy and management in the Texas A&M University School of Public Health, conducted this study and also the first national study on vaccine mandates for kindergarten through 12th grade students, finding that most Americans overwhelmingly support them.

Haeder developed an online survey that was given to 5,035 parents across the United States this September 27 and 28. For each of the three vaccines, the survey asked respondents about a number of disease-specific issues that are commonly used in research on vaccine hesitancy.

First, for example, they were asked if they were concerned about their child getting the respective disease, answering using a five-point scale. Haeder also accounted for the increasing politicization of vaccines by asking about political preferences and controlled for demographic factors such as race and ethnicity, gender, age, and income.

A statistical analysis found that about 40% of parents intended to vaccinate their children against COVID-19, 63% against influenza, and 71% against RSV.

“Interestingly, these results aligned with the results of our survey of pet owners, although those were not as politically polarizing,” Haeder says. “In this case, parents made the decision because they were concerned about these diseases, they trusted health providers and their children were previously vaccinated. Those who were opposed said they were concerned about vaccine safety and necessity, and believed they lacked information about the vaccines.”

While vaccination hesitancy or outright refusal was thrust into the national spotlight after COVID-19 and its vaccination emerged nearly simultaneously in late 2019, Haeder says the issue was a growing concern even prior to that, primarily surrounding the MMR (measles, mumps, and rubella) and HPV (human papillomavirus) vaccines.

“From a public health perspective, vaccine hesitancy has a ripple effect that extends beyond the individuals involved, even though they might tragically become ill or even die from a preventable disease,” Haeder says.

“Low vaccination rates place marginalized communities, such as people living in poverty, and vulnerable populations, such as the elderly, at greater risk for contracting a disease. To prevent a disease outbreak, we need high vaccination rates across the entire population.”

This makes vaccine requirements for school-age children especially important in containing the spread of a disease, for example, since this group overall is healthier than other groups and has fewer preventive visits with doctors, and thus would be less likely to get vaccinated without a mandate.

“Now, with the newly developed antibody immunization against RSV, in addition to vaccines against influenza and COVID-19, we have a unique opportunity to avoid illness and death this fall and winter,” Haeder says.

“But vaccine hesitancy—along with the elimination of COVID-19 funding and a belief that the pandemic is behind us—make it likely that the United States will experience an excessive amount of preventable illness from COVID-19, influenza, and RSV this fall and winter.”

Source: Ann Kellett for Texas A&M University

SAN FRANCISCO – Researchers will have rare opportunities over the next year to study solar flares and coronal mass ejections, observations that could lead to improved space weather forecasts.

During the April 8, 2024, total solar eclipse, NASA will gather data on the sun’s corona from instruments mounted on satellites, aircraft and sounding rockets.

“The moon is a perfect coronagraph,” Kelly Korreck, NASA eclipse program manager, said Dec. 11 at the American Geophysical Union conference here.

NASA researchers will compare computer models of the sun’s corona with imagery gathered during the eclipse from NASA’s Parker Solar Probe and the European Space Agency-NASA Solar Orbiter.

The various perspectives will give researchers a “3D understanding of the corona as well because you can triangulate a lot of different structures that we see from Earth and in another direction from the Solar Orbiter,” said Amir Caspi, Southwest Research Institute principal scientist.

In addition, NASA will gather multispectral coronal imagery from instruments sent aloft on a pair of WB-57F jets. And sounding rockets launched from Wallops Island, Virginia, will eject magnetometers, accelerometers and other instruments to monitor the ionosphere.

“Now instead of validating models with just one view, we can validate them with multiple views,” Nour Raouafi, Parker Solar Probe project scientist the Johns Hopkins University Applied Physics Lab.

Parker’s Close Approach

Launched in 2018, the Parker Solar Probe’s closest solar approach is set for Dec. 24, 2024. At that time, the probe will be 6.1 million kilometers from the sun’s surface. For comparison, Mercury is 58 million kilometers from the sun.

By traveling closer to the sun than any previous spacecraft, the Parker Solar Probe could obtain observations that help explain perplexing solar activity.

“The solar corona by nature is so mysterious,” Raouafi said. “It is over 300 times hotter than the solar surface. That is absolutely counterintuitive because from everyday experience, if you are getting away from the heat source, it cools down.”

Another mystery is the flow of ions and electrons traveling more than 3 million kilometers per hour from the sun.

“Where do they get the energy to be accelerated so fast?” Raouafi said.

Atmospheric Waves

Additional data related to space weather is being gathered by the NASA Atmospheric Waves Experiment. AWE traveled to the International Space Station in November aboard a SpaceX Falcon 9 rocket cargo resupply mission.

“Gravity waves that arise from tropospheric weather can grow in amplitude enormously because of decreasing density and can have enormous effects at high altitudes,” said Dave Fritts, vice president of atmospheric research for GATS, a company in Newport News, Virginia. Data gathered by AWE “will enable predictions of these waves propagating to very high altitudes,” he added.

At high altitudes, gravity waves may have “very important impacts in the thermosphere, the ionosphere and the lower space weather environment,” Fritts said. “There’s even the potential that capabilities like this can contribute to predictive capabilities down the road.”

Related

Researchers report the oldest record by far of domestic yak, dating back 2,500 years.

The high-altitude hero of the Himalayas, yak are among the few large animals that can survive the extremely cold, harsh, and oxygen-poor conditions of the Tibetan Plateau. In the mountainous regions of Asia, yak and yak-cattle hybrids serve as vital sources of meat, milk, transportation, and fuel. However, little is known about their history: when or where yak were domesticated.

As reported in Science Advances, the researchers zeroed in on this date using ancient DNA from a single male yak that lived alongside domestic cattle and yak-cattle hybrids in a settlement known as Bangga, a community in the southern Tibetan Plateau located at an elevation of approximately 3,750 meters (12,300 feet) above sea level.

“Many scholars have speculated that yak was first domesticated in the high-altitude regions of the Tibetan Plateau,” says Xinyi Liu, an associate professor of archaeology at Washington University in St Louis. “It was a well-informed speculation, but up to this point, there hasn’t been robust evidence for that,” Liu says. “This is the first evidence supported by both archaeology and ancient DNA.”

Wild and domestic yak

Once widespread in the Tibetan Plateau, wild yak are now listed as “vulnerable” by the International Union for Conservation of Nature, with only an estimated 7,500 to 10,000 mature individuals left in the wild.

Domesticated yak, however, are prevalent across much of the world. An estimated 14 million to 15 million live in the highlands of Asia alone.

Scientists have previously traced the origins of other domestic bovine species found in Asia. This includes the taurine cattle found primarily in Europe and temperate areas of Asia; indicine cattle, or zebus, found primarily in India and tropical areas of Asia; and water buffalos in East and Southeast Asia. “Yak remains an open inquiry,” Liu says.

“Identifying domestic yak and yak-cattle hybrids at Bangga is not only essential to the understanding of the origin of this charismatic creature, the yak, but also informs us in general about animal domestication pathways, in which gene flow between related stocks is increasingly appreciated,” he says.

Bangga is one of the earliest agro-pastoral settlements in the southern Tibetan Plateau and the only site in the region with abundant animal remains to have been systematically excavated in recent decades. This work at Bangga, led by Hongliang Lu from Sichuan University, has provided scientists with a glimpse of daily lives at extremely high elevations between 3,000 and 2,000 years ago.

The excavations at Bangga also offer a rare opportunity to explore the history of early yak, cattle, and their hybrids. For this study, Liu and his fellow archaeologists paired up with livestock geneticists. The team used ancient DNA sequencing as well as zooarchaeological analysis and radiocarbon measurements to help answer questions that could not be resolved with field analysis alone.

Bones from Bangga

Starting with more than 10,000 pieces of mammal bones collected at Bangga, Zhengwei Zhang, a postdoctoral researcher at Sichuan University, identified and sorted out 193 specimens belonging to the genus Bos, a group that includes all domestic cattle, zebus, and yak, as well as their wild progenitors, aurochs.

The researchers then selected five well-preserved bones from these Bos samples to sequence for whole-genomic ancestry. The sequencing work was led by Ningbo Chen and Chuzhao Lei, two geneticists at Northwest A&F University who specialized in Bos domestication.

Genetic analysis revealed that only one of the ancient bones came from a yak, a male individual, while the other four bones were from female taurine cattle. Even figuring out that the cattle were taurine cattle was a surprise, Liu says, as Bangga is located close to the Himalayas and within the range of zebus and Indian aurochs, which were not found at Bangga. Instead, the cattle belonged to the taurine lineage that was introduced to the region from Anatolia via the silk route and northern Tibetan Plateau.

Additional analysis helped clarify that the bone from the male yak was truly a domestic variant, and not just a bone from a wild yak that hunters had killed and brought back to the settlement as food. The researchers also saw evidence for hybridization between the two species.

This new discovery of domesticated yak from 2,500 years ago fits into the larger story that is beginning to emerge about how humans adapted to living in a high-altitude environment on the Tibetan Plateau. For example, Liu and his colleagues have previously documented how people in this region grew barley as they faced a challenging environment.

“Bangga provided us with a unique window into lifeways at high elevations 3,000 to 2,000 years ago,” Liu says. “They cultivated barley in an intensive way, provisioned sheep with fodder and water, and consumed milk. All these resources were introduced to the Tibetan Plateau from other world regions as part of prehistoric food globalization and had become part of the Tibetan legacy. Now we know they had domestic yak.”

This discovery of the genetically confirmed evidence of domestic yak does not yet solve all questions about yak domestication, nor does it necessarily represent the very beginning of that domestication process. However, it hints at what motivated herders to bring yak home.

Early herders in this region were likely faced with harsh conditions, where animals died out quickly because of prolonged winters and severe snowstorms. One would need to be innovative to live in such conditions. A possible solution is intensive corralling, which would have allowed herders to provision their herds with agricultural by-products and water year-round. This strategy is attested by recent zooarchaeological and isotopic work led by Liu and Zhang.

The other solution would be to combine the environmental hardiness of yak with the productivity of cattle. “Dzomo (female hybrid) and dzo (male) are still the most common stocks in the Plateau even today for that reason. Cattle produce more milk and meat, but they are not as good at adapting to the high-altitude environment as yak,” Liu says. “Hybridization allows cattle to move high, and yak to move low at the same time they produce more milk.”

Study coauthor Fiona Marshall, professor emerita of Washington University, says the study draws attention to the genetic continuity among domestic yak and taurine cattle on the Tibetan Plateau. In many regions of the world, early domesticated animals were replaced by later varieties. The genomic data suggests such faunal turnover did not happen on the Tibetan Plateau.

“This suggests a successful and long-lasting legacy of early Tibetan communities who were cosmopolitan in subsistence strategies and resilient and innovative in facing a challenging climate,” Liu says. “Bangga provides the best example of such a community.”

Source: Washington University in St. Louis

Legislation aimed at preventing large companies from avoiding United States taxes by shuttling money to foreign subsidiaries hasn’t worked as well as anticipated. A new study reveals how companies are responding to the provisions—and the potential costs associated with their tax avoidance strategy.

The Tax Cuts and Jobs Act, which went into effect in 2018, included base erosion and anti-abuse tax (BEAT) provisions designed to discourage companies from circumventing US tax law. Historically, some companies have created subsidiaries outside the US in countries that had lower tax rates. And doing business with those subsidiaries has the effect of allowing the parent companies to avoid higher US taxation. The BEAT provisions—which apply only to large companies—essentially disincentivize this loophole by adding an additional surtax to payments companies make to foreign subsidiaries.

However, actual BEAT tax collections have been less than half of what was predicted. Why?

“One way a company can potentially avoid the BEAT provisions and reduce its tax liability is by reclassifying the payments it makes to foreign subsidiaries,” says Christina Lewellen, coauthor of a paper on the work and an associate professor of accounting in North Carolina State University’s Poole College of Management. “For example, one type of payment not subject to the BEAT provisions is classified under the category of ‘cost of goods sold.’”

Cost of goods sold is a broad category that covers the overall costs attributable to the production of the goods or services a company sells. Cost of goods sold includes direct costs, such as the cost of feedstock used in manufacturing. But it also includes indirect costs, such as product design, management, and sales and marketing services.

“If a company says that the payments it makes to a foreign subsidiary are for services covered under cost of goods sold, it can avoid the BEAT provisions,” Lewellen says. “That raises two questions.

“First, are companies actually reclassifying payments to foreign subsidiaries as cost of goods sold? Second, if companies are reclassifying these payments to subsidiaries, how is that affecting the companies?”

To address these questions, the researchers looked at data from 24,982 foreign subsidiaries incorporated in 48 countries. Only 9,944 of those subsidiaries are subsidiaries of companies that are subject to the BEAT provisions, with the remainder serving as a control group. Specifically, the researchers looked at financial data for those subsidiaries from 2011 through 2018—the seven years prior to the implementation of the BEAT provisions, and the year the BEAT provisions took effect.

“Companies don’t disclose a detailed breakdown of how they distribute their cost of goods sold, so we had to identify a proxy that would give us a clear idea of whether companies were shifting some cost-of-goods-sold expenses to foreign subsidiaries,” Lewellen says. “One way to do this is to measure whether there was an increase in sales by foreign subsidiaries when the BEAT provisions took effect—and there was.

“On average, sales for foreign subsidiaries subject to the BEAT provisions were 6.8% higher than the control group in the year that the provisions took effect,” Lewellen says. “This suggests that companies were, indeed, reclassifying payments as cost of goods sold expenses in order to avoid paying the BEAT. We estimate that this reclassification allowed the companies in our sample to avoid paying approximately $6 billion in US taxes.”

So why wouldn’t every company do this? Well, that’s because the researchers found that the reclassification process has costs of its own.

“For one thing, we found that the reclassification process can require a significant overhaul of the company’s internal accounting processes,” Lewellen says. “That is a significant task in itself, but modifying these processes can also result in managers across the company having to make decisions with lower quality data. This is more likely to be the case if the true cost of goods sold varies significantly from the way cost of goods sold is being reported for tax reasons.”

The researchers also found that companies whose foreign subsidiaries reported a marked increase in sales when the BEAT provisions took effect were also more likely to disclose higher levels of uncertainty about their tax positions in financial statements.

“That means those companies were required to maintain higher tax reserves in case the IRS overturns those tax positions,” Lewellen says. “It also means those companies are likely to face greater scrutiny from the IRS in the first place.”

The paper appears in the Journal of Accounting and Economics. Additional researchers contributed from the University of Wisconsin-Madison and Singapore Management University.

Source: NC State

A study evaluating the bias in OpenAI’s CLIP, a model that pairs text and images and operates behind the scenes in the popular DALL-E image generator, finds that CLIP performs poorly on images that portray low-income and non-Western lifestyles.

“During a time when AI tools are being deployed across the world, having everyone represented in these tools is critical. Yet, we see that a large fraction of the population is not reflected by these applications—not surprisingly, those from the lowest social incomes. This can quickly lead to even larger inequality gaps,” says Rada Mihalcea, a professor of computer science and engineering at the University of Michigan, who initiated and advised the project.

AI models like CLIP act as foundation models, or models trained on a large amount of unlabeled data that can be adapted to many applications. When AI models are trained with data reflecting a one-sided view of the world, that bias can propagate into downstream applications and tools that rely on the AI.

“If a software was using CLIP to screen images, it could exclude images from a lower-income or minority group instead of truly mislabeled images. It could sweep away all the diversity that a database curator worked hard to include,” says Joan Nwatu, a doctoral student in computer science and engineering.

Nwatu led the research team with Oana Ignat, a postdoctoral researcher in the same department. They are coauthors of a paper presented at the Empirical Methods in Natural Language Processing conference on December 8 in Singapore.

The researchers evaluated the performance of CLIP using Dollar Street, a globally diverse image dataset created by the Gapminder Foundation. Dollar Street contains more than 38,000 images collected from households of various incomes across Africa, the Americas, Asia, and Europe. Monthly incomes represented in the dataset range from $26 to nearly $20,000. The images capture everyday items, and are manually annotated with one or more contextual topics, such as “kitchen” or “bed.”

CLIP pairs text and images by creating a score that is meant to represent how well the image and text match. That score can then be fed into downstream applications for further processing such as image flagging and labeling. The performance of OpenAI’s DALL-E relies heavily on CLIP, which was used to evaluate the model’s performance and create a database of image captions that trained DALL-E.

The researchers assessed CLIP’s bias by first scoring the match between the Dollar Street dataset’s images and manually annotated text in CLIP, then measuring the correlation between the CLIP score and household income.

“We found that most of the images from higher income households always had higher CLIP scores compared to images from lower income households,” Nwatu says.

The topic “light source,” for example, typically has higher CLIP scores for electric lamps from wealthier households compared to kerosene lamps from poorer households.

CLIP also demonstrated geographic bias as the majority of the countries with the lowest scores were from low-income African countries. That bias could potentially eliminate diversity in large image datasets and cause low-income, non-Western households to be underrepresented in applications that rely on CLIP.

“Many AI models aim to achieve a ‘general understanding’ by utilizing English data from Western countries. However, our research shows this approach results in a considerable performance gap across demographics,” Ignat says.

“This gap is important in that demographic factors shape our identities and directly impact the model’s effectiveness in the real world. Neglecting these factors could exacerbate discrimination and poverty. Our research aims to bridge this gap and pave the way for more inclusive and reliable models.”

The researchers offer several actionable steps for AI developers to build more equitable AI models:

• Invest in geographically diverse datasets to help AI tools learn more diverse backgrounds and perspectives.
• Define evaluation metrics that represent everyone by taking into account location and income.
• Document the demographics of the data AI models are trained on.

“The public should know what the AI was trained on so that they can make informed decisions when using a tool,” Nwatu says.

The research had funding from the John Templeton Foundation and the US Department of State.

Source: Patricia DeLacey for University of Michigan