We measure the impact of a major event in the advancement and dissemination of Generative AI technology, namely, the public release of ChatGPT, on equity returns at the firm level. Our key finding is that the arrival of ChatGPT had a sizable positive effect on the value of firms whose labor forces are more exposed to Generative AI and related Large Language Models (LLMs). Firms with higher exposure to the release of ChatGPT, as measured by the exposure of their labor force to being made more productive by tools like ChatGPT, outperform firms with lower exposures by over 40 basis points in daily excess returns during the two weeks following its release. Notably, these return differences are not only due to differences in labor force exposures across industries. Returns of firms with high labor force exposures also outperform firms with low exposures by about 40 basis points daily in industry-neutral portfolios.
Our main finding is that corporate tax cuts generate a significant boost in investment and employment for the economy overall, but the benefits are spread unevenly across sectors and groups. In particular, goods producing companies —such as manufacturing firms— expand both capital expenditure and wage bills following a cut in corporate taxes, but do not alter dividend payments. The left column [of Figure 1] shows that a 1% cut in the marginal tax rate stimulates an increase in capital expenditure for goods producing firms peaking at 8% and 5% in year two following a cut in the marginal tax rate and an increase in the investment tax credit. In contrast, firms in the service sector—which are far less capital-intensive—do not increase investment or employment at all but use most of their windfall to pay dividends. In short, we find important differences in the effect on workers vs. shareholders across sectors of the economy.
We took a look at all PE/VC-owned public companies, or companies with public debt, that were 30%+ sponsor-owned, had IPOed since 2018, had a recognizable sponsor as the largest holder, and were headquartered in North America. Today, the entire sample trades at 22.4x pro-forma EBITDA, which is roughly in line with the Russell 2000 Growth, where the median company trades at 20.7x EBITDA. However, for companies in our sample that reported both pro-forma and GAAP EBITDA, this includes 500bps of adjustments. On a GAAP basis, these companies trade at 33.4x EBITDA with 8.8x net debt/EBITDA. The sample of companies we looked at is nearly unprofitable on an EBITDA basis, mostly cash flow negative, and extraordinarily leveraged (mostly with floating-rate debt that is now costing nearly 12%). These companies trade at a dramatic premium to public markets on a GAAP basis, only reaching comparability after massive amounts of pro-forma adjustments. And these are the companies that most likely reflect the better outcomes in private equity.
Labor cost growth does not appear to fuel inflation through higher demand. The light blue bars in Figure 3 show the cumulative impact of a 1pp increase in the employment cost index (ECI) on the contribution of goods, housing services, and nonhousing services (NHS) inflation to core PCE inflation over a four-year time horizon. The estimates confirm that labor costs have a stronger impact on NHS inflation than either goods or housing services inflation. The impact of the ECI on NHS inflation is statistically significant, but the magnitude is quite small. A 1pp increase in the ECI increases the contribution of NHS inflation to core PCE inflation by 0.15pp over four years—an effect of 0.04pp per year.
About 3.66 million babies were born in the U.S. in 2022, essentially unchanged from 2021 and 15% below the peak hit in 2007. The trend of decreasing birthrates among younger women continued in 2022. For teens ages 15 to 19, the birthrate fell 3%, and for ages 20 to 24 it was down 2%. The rate for the next oldest group, 25 to 29, edged up only slightly. Increases were mainly seen among women 35 to 44. If trends continue, the birthrate for women ages 35 to 39 might soon eclipse the rate for ages 20 to 24.
The millennials of 2008 are not the same as those of 2016. For instance: six additional years of even more heavily Democratic millennials became eligible to vote after the 2008 election, canceling out the slight Republican shift among older millennials. The shift to the right appears largest among the oldest “young” voters — the older millennials who came of age in a very different political era from today. Many of the issues that drew young voters to the Democrats in 2004 or 2008 — like the Iraq War or same-sex marriage — may no longer be issues at all. Related: Millennials are Shattering the Oldest Rule in Politics and What Happened In 2022
There is no long-term trend of USD decline. It always depends on which dollar index you look at. But regardless of the index, you will find that the USD is above its long-term average, in nominal and real terms currency, despite all the Dollar Hatred.
The largest homeowner insurance company in California, State Farm, announced that it would stop selling coverage to homeowners. That’s not just in wildfire zones, but everywhere in the state. State Farm, which insures more homeowners in California than any other company, said it would stop accepting applications for most types of new insurance policies in the state because of “rapidly growing catastrophe exposure.” California’s woes resemble a slow-motion version of what Florida experienced after Hurricane Andrew devastated Miami in 1992. The losses bankrupted some insurers and caused most national carriers to pull out of the state. In response, Florida established a complicated system: a market based on small insurance companies, backed up by Citizens Property Insurance Corporation, a state-mandated company that would provide windstorm coverage for homeowners who couldn’t find private insurance.
James Liang, a Chinese economist and demographer, notes that entrepreneurship is markedly lower in older countries: an increase of one standard deviation in the median age in a country, equivalent to about 3.5 years, leads to a decrease of 2.5 percentage points in the entrepreneurship rate (the proportion of adults who start their own business). That is a huge effect, considering the global entrepreneurship rate was around 6.1% in 2010. As recently as 2010, Japanese inventors were the leading producers of patents in 35 big global industries according to the World Intellectual Property Organisation, an agency of the un. As of 2021, Japan is the leader in just three. The country has fallen behind not only China, which now occupies most of the top spots, but America too.
More than ever, our economy runs on ecosystems. Not just broad partnerships, large commercial deals, or expansive networks, but ecosystems. It’s to our collective interest that there’s maximum competition at the foundational platform layer and every application and service layer, too. But the cost of competition is now in the tens of billions, if not more. In a weird way, Google Cloud is cheap at $35B, but most can’t spend that much, let alone wait fifteen years for losses to end and another decade or more to generate a net return. In fact, it’s likely Google couldn’t have invested as heavily were it not for near-zero interest rates over the past decade. And that the investment wouldn’t have worked at all were it not for Google’s own ecosystem, including the consumer-facing (and Android-integrated) Google Photos, Gmail, and Drive, as well as its enterprise offerings such as Workspace. The prohibitive cost of entry has led to phenomenal profits. Apple, Microsoft, Alphabet, and Amazon are now 21% of the S&P (Apple and Microsoft are so large that alone, they would be the second largest sector of the S&P, behind the rest of big tech, and ahead of energy and healthcare).
Are young people really skipping traditional four-year colleges for other opportunities? The answer is a big fat No. And we can even use the same data the WSJ used (from the CPS) to prove it, but slice it more finely. The percent of recent high school graduates enrolled in 4-year colleges and universities in 2022 was 45.1%. That’s slightly higher than 2019 (44.4%) and is, in fact, the second highest level ever in this data, with only 2016 being higher at 46%. So what gives? The decline that the WSJ is reporting is entirely driven by a decline in enrollment at 2-year colleges, though you would never get a hint of that in the article. You might even think it was the opposite: perhaps young people are forgoing 4-year colleges in favor of trade schools! Nope. Here’s the data. Notice that this is in fact the exact same data as the WSJ is using. If you add the two figures for 2022, it’s the same 62% in the WSJ article. But 4-year college enrollment is up slightly since 2019, while 2-year enrollment is down about 5 percentage points. Also, it doesn’t really seem like a pandemic-induced change: this is a long-run decline in 2-year college enrollment of about 12 percentage points since the peak in 2012.
Related: More High-School Grads Forgo College in Hot Labor Market
As the Fed has been raising rates, US households have been big buyers of US Treasuries. The appetite for Treasuries from foreigners has been more limited because of higher hedging costs. Foreigners have instead increased their holdings of equities by $3 trillion during the pandemic. With a 5% budget deficit combined with QT and the Treasury’s need to replenish cash in the Treasury General Account, markets will soon begin to focus on who will be buyers of US government debt.
Major subway crime is on track to be 13.4% lower this year than in 2019 and lower than in all but a few years during the past quarter century. Then again, subway ridership is still down about a third from 2019 levels. Adjust for that, as one should, and the picture changes. Measured this way, subway crime rose 86% in the first year of the pandemic. It has fallen since but is still running much higher this year than from 2006 through 2019. It’s not exactly what you’d call out of control, though, just settling back toward normal after a once-in-a-century shock. My less-than-scientific take is that there are more troubled people on New York’s subways than before the pandemic, and less-crowded stations and cars have made it easier for them to make trouble. For-profit subway crime hasn’t risen; senseless acting-out has.
Even if you believe that excessive government spending played a big role in the initial rise in inflation, future inflation will reflect more persistent factors — which, in Olivier Blanchard and Ben Bernanke analysis, means focusing on the labor market, which they say is still overheated and needs to cool. The six-million-job question is whether this cooling off needs to involve a large rise in unemployment. The paper is actually fairly optimistic on that question, suggesting that “immaculate disinflation,” inflation coming down without any significant rise in unemployment, may be possible, and that even if it isn’t, those grim projections we were hearing a year ago about the need for many years of high unemployment no longer seem plausible. The ratio of vacancies to the unemployed as their measure of labor market tightness. And what has been really striking since late 2022 is that vacancies have come way down without any rise in unemployment. Related: What Caused the U.S. Pandemic-Era Inflation? and Inflation: A Series of Unfortunate Events vs. Original Sin
The college enrollment rate for recent U.S. high-school graduates, ages 16 to 24, declined to 62% last year from 66.2% in 2019, just before the pandemic began, according to the latest Labor Department data. The rate topped out at 70.1% in 2009. The unemployment rate for teenage workers ages 16 to 19 fell to a 70-year low of 9.2% last month, fueling larger pay increases. Average hourly earnings for rank-and-file leisure and hospitality workers were up nearly 30%, seasonally adjusted, from April 2019 to April 2023, compared with roughly 20% during the same period for all workers. College enrollment has declined by about 15% in the past decade, according to federal data. The reasons include the high cost of university education, colleges closing and uneven returns from getting a degree, as well as the hot job market.
The awful truth remains: By 2019—that is, even before COVID—the Class of 1980’s death rates at age 39 were higher than their counterparts’ from the Classes of 1970 and 1960. At age 29, the Class of 1990’s death rates were indistinguishable from those of the Class of 1950. In Japan, the Class of 1990’s mortality rate at age 29 was over 80 percent lower than those of their “grandparents” from the Class of 1930. In the US, the corresponding differential was less than 8 percent. If we look carefully at Figure 2, we can see unwelcome “crossovers”—where death rates in adulthood exceeded those of earlier cohorts—for every US cohort from the Class of 1950 onward in the years after the Berlin Wall fell (1990 onward). This is what prolonged stagnation—at times, even reversal—in national health progress looks like. Related: Who Won the Cold War? Part I and Who Won the Cold War? Part II
From most families’ perspectives, the excess savings accumulated during the pandemic are now entirely gone. The share of households saying they had savings worth 3 months of expenses dipped to just above pre-pandemic levels, with higher-income households being noticeably less likely to have a sufficient emergency fund than they were in 2019.
Last year 1.2m people moved to Britain—almost certainly the most ever. Net migration (i.e., immigrants minus emigrants) to Australia is currently twice the rate before the covid-19 pandemic. Spain’s equivalent figure recently hit an all-time high. Nearly 1.4m people on net are expected to move to America this year, one-third more than before the pandemic. In 2022 net migration to Canada was more than double the previous record. In Germany it was even higher than during the “migration crisis” of 2015. The rich world as a whole is in the middle of an unprecedented migration boom. Its foreign-born population is rising faster than at any point in history.
Defying the adage among practitioners and scholars of politics that voters become more conservative as they age — millennials (those born between 1981 and 1996) and Gen Z (those born in 1997 and afterward) have in fact become decidedly more Democratic over time, according to data compiled by the Cooperative Election Study. The graphic below, which is derived from the study, shows a significant increase in voting for House Democratic candidates among millennials and Gen Z. Brian Schaffner, a political scientist at Tufts observed “Because the population is very big and turnout rates tend to be much higher for older adults, these trends can be slow to lead to significant gains. For example, in 2018, I applied a life expectancy model to our C.E.S. data and using that model I calculated that it would take more than 20 years for Democrats to gain just 3 percentage points on their vote share from differential mortality. Those gains could easily be offset by Republicans doing a bit better among other groups. For example, part of what has helped them in recent elections is that even while the share of the population who are non-college white people is in decline, it is still a large group that (1) has come to vote more Republican in the past decade and (2) has seen its turnout rate increase during the same period.” Related: Millennials are Shattering the Oldest Rule in Politics and What Happened In 2022
A German start-up has secured initial funding to develop a revolutionary fusion energy machine that it hopes can provide a future source of abundant, emissions-free power. Proxima Fusion, incorporated in January, aims to build a complex device known as a stellarator and is the latest company to join the emerging fusion industry’s effort to generate electricity by fusing atoms. Although the amount of funding is small at only €7mn, it is significant as Proxima is the first fusion company to spin out of Germany’s revered Max Planck Institute for Plasma Physics. Little known outside the world of plasma physics, a stellarator is an alternative to the better-known tokamak device. The twisted structure of the stellarator is more complicated to design and build than a traditional tokamak but produces a more stable plasma that could enable scientists to sustain the fusion reaction for longer.
If past research informs current work, then firms face a choice. Is it more valuable to learn from exploring new areas, or to exploit the information revealed by past work? In this paper, we show that despite novel drugs generating more learning, firms prefer to invest in incremental drugs, which are easier to evaluate. This presents a dynamic tension: while firms value the ability to discard incremental drugs that are unlikely to succeed, they are reluctant to make the types of exploratory R&D investments that improve future screening decisions. our results suggest that policymakers could provide stronger incentives for firms to develop novel drugs by making it easier for them to appropriate indirect revenues from potential successors. These revenues, which can exist even when the focal project fails, disproportionately increase the overall returns to investing in newer, more uncertain research areas.
Another high reading for core PCE inflation, up 4.7% at an annual rate in April–and a 4.3% annual rate over the last three months, continuing its sideways move. But, under the hood it looks a little better as noisy items and imputations drove a lot of the extra high reading. Overall where does this leave us? The 4.7% annual rate for core PCE in April is too pessimistic a read of the inflation situation. But there is nothing in this report that would give any comfort that inflation is on track to fall to 3.5% or below without a further easing of the labor market.
A spike in the Kansas City Fed’s Financial Stress Index historically leads to higher unemployment and somewhat lower inflation. A one standard deviation increase in financial stress typically portends an increase in the unemployment rate of 0.7 percentage points. Financial stress historically reduces inflation as well: inflation, as measured by the consumer price index (CPI), falls by about 0.4 percentage points in the first year following an increase in financial stress. By this metric, financial stress is about half as disinflationary as monetary tightening. Could the recent increase in financial stress generate the same disinflationary effects as tighter monetary policy? The left panel in Chart 2 shows that monetary policy tightening, as measured by an unexpected increase in the expected path of the federal funds rate (blue line), slows economic activity and raises the unemployment rate, similar to an increase in financial stress. However, despite a similar increase in the unemployment rate, inflation falls by 0.65 percentage points in the first year following monetary tightening (right panel), a considerably larger decline than we estimate following an increase in financial stress.
Work requirements are an imperfect method to try to replace the incentive to work that social programs eliminate. Our government does this sort of thing all over to transfer income but contain the disincentives: Subsidize gas, and then regulate against its use for example. By the way, supposedly socialist Europe, after its experience with “the dole” in the early 1990s, is much more heard-hearted about these sorts of incentives than we are. Is there a better way? I’ve long played with the idea of limiting help by time rather than by income. That’s how unemployment insurance works. We understand that replacing people’s paycheck forever if they lose their job has bad incentive effects. Unemployment is understood as a temporary misfortune, and understanding the incentives, you get unemployment checks for a limited amount of time. Could not many other programs aimed at misfortune also be limited by time — but then allow you to keep each extra dollar of earnings? Perhaps even unemployment should be a fixed amount of time, and you can keep receiving it for the full (normally) 26 weeks even if you get a job.
Compare mortality in young adulthood for America’s Class of 1990 with their counterparts from affluent Cold War allies. Since breakdown by sex does not add much information here, we display overall mortality rates at ages 20 through 31 for the US and select Western European allies in Figure 3. Death rates at age 20 were universally lower for the treaty allies than Americans in the Class of 1990—usually much lower. Further, mortality curves for these allies generally remained much “flatter” over the course of their 20s than for the US. Consequently, the divergence in mortality risks between the US and the allies tended to increase over young adulthood for the Class of 1990, even before COVID. By age 29—i.e., in 2019, before the pandemic—mortality rates for the Class of 1990 were almost twice as high in the US as in New Zealand; two and a half times higher than in France; three time higher than in Japan; four times higher than in Italy or Spain.
Related: Who Won the Cold War? Part I
Sociologists Karen Benjamin Guzzo and Sarah Hayford found that when millennials (born 1981 to 1996) and the oldest members of Generation Z (starting in 1997) were surveyed in their late teens and early 20s, they said, on average, that they wanted to have at least two children. But the gap between women’s intended number of children and their actual family size has widened considerably. The researchers found that by the time women born in the late 1980s were in their early 30s, they had given birth, on average, to about one child less than they planned. That is roughly double the size of the shortfall for women born two decades earlier, and it is likely too large to be erased by a spurt of childbearing in their late 30s. The median age at which women give birth is 30, three years older than it was in 1990. Despite advances in fertility treatments, women who delay having kids until their final childbearing years reduce their chances of doing so—not just because it narrows their biological window but because other priorities and roadblocks can more easily derail their plans.
We revalue the stock of New York City commercial office buildings taking into account pandemic-induced cash flow and discount rate effects. We find a 45% decline in office values in 2020 and 39% in the longer-run, the latter representing a $453 billion value destruction. Higher quality office buildings were somewhat buffered against these trends due to a flight to quality, while lower quality office buildings see much more dramatic swings. These valuation changes have repercussions for local public finances and financial sector stability. The decline in office values and the surrounding CBD retail properties, whose lease revenues have been hit at least as hard as office, has important implications for local public finances. For example, the share of real estate taxes in NYC’s budget was 53% in 2020, 24% of which comes from office and retail property taxes.
The 2020-21 numbers here were released in late April by the Internal Revenue Service. They sort taxpayers by whether and where they moved between filing their taxes in 2020 and filing them in 2021; the adjusted gross incomes are for the 2020 tax year. It has been two years since May 17, 2021 — that year’s belated income tax filing deadline — and a lot has changed. But New York has continued to lose population, and if the trend depicted above were to continue, even in less extreme form, it would be disastrous for the finances of a state that relies on income taxes paid by those making $200,000 or more a year for almost half its revenue. That the loss of affluent taxpayers didn’t lead to disaster during the pandemic mainly had to do with how much the prices of stocks, houses and other assets rose in 2020 and 2021.
In an i > g world [nominal interest rates higher than nominal growth,] growth in the revenues, wages or tax receipts that a debtor earns will be slower than the interest accumulating on their borrowing, meaning debt levels have the potential to explode. An i > g world is unfamiliar to America and most of the West. Since the end of 2009 nominal growth has been higher than nominal rates (aside from the first half of 2020, when the covid-19 pandemic crashed the economy). Now America is about to cross the threshold, [based on a panel of economists surveyed by Bloomberg.] It is far easier to swallow a high cost of capital when it is matched by high returns on said capital. And that will not be the case for much longer.
Was inflation the result of unpredictable shocks (i.e., unfortunate events) or was it predictable (i.e., original sin)? My view: core predominantly original sin but excess of headline was due to an unfortunate accident. An elegant paper but it does not answer whether inflation was “a series of unfortunate events” or “original sin”. The paper does suggest that food and energy shocks do not explain core inflation. Shortages are consistent with demand increases—and demand supporting higher overall consumption. It may be more fruitful to ignore the labor market in assessing large non-linear shocks. Regardless, we all agree about the present situation–and the unlikelihood of a soft landing to 2% inflation.
Last year defence spending worldwide increased by nearly 4% in real terms to over $2trn. The share prices of defence firms are performing better than the overall stockmarket. We estimate that total new defence commitments and forecast spending increases, if implemented, will generate over $200bn in extra defence spending globally each year. It could be a lot more. Imagine that countries which currently spend less than 2% of GDP per year meet that level and that the remainder increase spending by half a percentage point of GDP. Global defence outlays would rise by close to $700bn a year. There is little reason to believe that the new cold war will be sharply inflationary. Not even the fiercest hawks are calling for defence spending, as a share of GDP, to return to the levels of the 1960s or 1970s.