Semiconductor firms are about to get showered with cash thanks to a new bill, The CHIPS and Science Act of 2022—CHIPS standing for, cleverly, The CREATING HELPFUL INCENTIVES TO PRODUCE SEMICONDUCTORS Act. Because it involved free money for capitalists, 17 Republicans (out of 50) voted for it despite their habit of voting against almost anything supported by Democrats except money for the Pentagon. Biden is scheduled to sign it on Tuesday, August 9.
It’s a $280 billion package designed to encourage semiconductor manufacturing and research in the US. Pundits and generals have watched with increasing alarm as chipmaking moved from an industry dominated by the US and to a lesser extent Western Europe to a heavily Asian affair, meaning Taiwan, South Korea, and China (the big threat lurking behind all the worry, of course). The covid crisis highlighted how dependent US and European producers—carmakers, computer companies, appliance manufacturers—have become on Asian chipmakers.
The CHIPS Act is supposed to address all that, bringing the industry back home and restoring the Silicon Valley to its rightful place of global dominance. But that’s not all. When she wasn’t stoking war with China (which would require plenty of advanced chips, presumably sourced from countries other than China), Nancy Pelosi took time out to say that the CHIPS Act would create “nearly 100,000 good-paying, union jobs.” One is skeptical.
Among its provisions is $53 billion in subsidies to the US chip industry to encourage domestic R&D. Of that, $39 billion is a direct subsidy, and the rest is for “workforce development” and speeding up the “lab to fab” pipeline. The bill also includes large tax writeoffs for companies that expand chip facilities in the US, which would theoretically close the cost gap between building here and abroad. Aid won’t be confined to American firms; industry leader Taiwan Semiconductor Manufacturing Co. (TMSC) will be eligible as long as they do the work here.
The Biden administration is touting a plant Intel is building in Ohio—he even devoted 147 words to it in his State of the Union address—as proof of the CHIPS Act’s powers, but Intel has been playing politics with the project. In June, impatient with Congressional dithering on the bill, it canceled a formal groundbreaking ceremony, emphasizing that it’s building plants in countries that have gotten their subsidy act together. (Mercifully, the EU has its own Chips Act.) Intel also said its long-term plans for the site would depend on the continuing flow of free federal cash. (TSMC said its investment plans also depended on adoption of the bill.) On its passage, Intel CEO Pat Gelsinger said, “I’m excited to put shovels in the ground as Intel moves full speed ahead to start building in Ohio,” though I suspect Gelsinger himself won’t be operating any shovels. Intel may be happy, but just last month some smaller firms were complaining that it and the other giants would get all the money and they’d be frozen out.
About $200 billion will be also devoted to basic research in computing, robotics, and semiconductors through the National Science Foundation and other federal agencies. Universities getting grants under the program would be prohibiting from participating in educational partnerships with the Chinese government known as Confucius Institutes. These have been under relentless attack for nearly a decade, as the anti-China campaign ramped up.
Worries have been raised about the return of “industrial policy,” pronounced dead in the 1980s (though the Reagan administration did craft a giant aid package for the chip industry known as SEMATECH). Rather than the return of industrial policy, what’s actually distressing is watching public money go into private coffers with so little coming in return.
The CHIPS Act bars its beneficiaries from using the money to do stock buybacks or pay dividends, though it’s not clear how one set of a few billion dollars can be distinguished from another. It would also deny money to firms expanding in China and other countries deemed unfriendly. Economic warfare is ramping up.
billions in buybacks
Criticism of the bill from the left has largely focused on the quickness of US firms to expand abroad. Bernie Sanders said, “Any company who is prepared to go abroad, who has ignored the needs of the American people, will then say to the Congress, ‘Hey, if you want us to stay here, you better give us a handout.’”
He is right, though that is pretty standard practice in American capitalism. His colleague Sherrod Brown, who might have been relied on for a denunciation, likes the bill because of the Intel plant in his state, Ohio. And few Congresspeople could turn down cash for universities in their states and districts.
It was probably an oversight, but Sanders could also have pointed out how much the likely recipients have already spent on stock buybacks. It’s a small fortune.
Between 2010 and 2021, ten top semiconductor firms spent $168 billion buying their own stock.* Here’s a graph of the running total of buybacks for those ten firms, and Intel, the biggest chip firm. It alone spent $86 billion over the same period. Both sums vastly exceed the $39 billion the gov is showering on the industry in direct subsidies.
As this graph shows, there was a burst of buyback activity in 2018 and 2019. Probably not coincidentally, 2017 was the year when Trump’s corporate tax cuts took effect. In theory, those were supposed to contribute to a burst of investment; they didn’t. Though the buyback pace eased after covid hit, 2021’s total was still higher than all but one year between 2010 and 2017.
Buybacks do little but boost stock prices, making stockholders happy and fattening CEO paychecks, which are largely based on stock prices. They could have invested the money, but that sort of thing is frowned upon in the higher Wall Street consciousness. Better to “return” the cash to shareholders, as they say, even though they actually provided no cash in the first place. Stockholders just buy their shares from previous stockholders, not the company.
Since Intel is so big, we also looked at its earlier buybacks, starting in 1990. That prehistory adds another $69 billion to the company’s total buying spree, taking the 31-year total up to $152 billion.
As I argue in today’s semi-companion post, the US economy has long been plagued by low and declining levels of net investment, a result of the shareholders-first doctrine of the last four decades. Fortunately for the semiconductor industry, Uncle Sam is effectively picking up the tab for its buyback spree.
* The ten firms are, in declining order of total buybacks, are Intel, Texas Instruments, Applied Materials, Broadcom, Analog Devices, NVIDIA, Micron, KLA, Marvell, and AMD.
It’s been a while since I looked at one of the major reasons for the pervasive sense of rot about the US: the low level of investment—investment in real things, that is, not crypto. It’s barely keeping up with the forces of decay. If you’re wondering why nothing works and everything seems to be falling apart, here are some explanations.
First a definition: investment is spending by businesses, governments, and individuals on long-lived physical assets like buildings and machinery. Gross investment is the dollar value of such spending; net is what remains after deducting depreciation, aka wear and tear. That’s not an easy process to put a dollar value on, but it’s all we’ve got. And besides, these are numbers the capitalist state produces to understand its economy, so why not take them seriously, even if the bourgeoisie seems unalarmed about them?
Graphed below is the average value of net public and private investment as a percentage of GDP by decade. Civilian public investment means expenditures on long-lived assets like schools and roads, but excluding the military. (To anticipate a question I sometimes get: yes, prisons are in there too, but they don’t count for much; almost all the costs of maintaining the carceral state come from day-to-day operations.) Private investment consists of purchases of buildings, equipment, and intellectual property (IP) by businesses and. Not shown on this first graph: residential investment, the purchase of housing by individuals and improvements to that housing.
Averages for the 1930s reflect the extraordinary circumstances of the Great Depression: private investment collapsed and New Deal-driven public investment soared. Those high levels of public investment gave us an infrastructure that we still use today—schools, post offices, and parks. (For a catalog of those projects, check out the Living New Deal.) Public investment sagged during the 1940s, reflecting World War II, but rose in the 1950s and 1960s, matching the 1930s level as the public sector expanded. It was not to last: austerity and privatization consciousness took over, and now net public investment is at a record low.
Private investment rose in the decades after World War II, peaking in the 1970s. But the Wall Street-driven imperatives of profit maximization that got the upper hand via the Shareholder Revolution of the 1980s, which transformed corporate practices, put the squeeze on investment. Investing too large a share of corporate profits in things came to be seen as wasteful—better instead to hand the cash over to the shareholders, via stock buybacks and traditional dividends.
Here’s a yearly view of the trajectory of decline, a path traced by the dotted trendlines. The graphs begin in 1950 because the extremes of the 1930s and 1940s would have distorted the scale.
These graphs show a relative stability in net private investment from the late 1950s through the early 1980s, when Wall Street’s grip on corporate cash flow tightened. There was a surge in the late 1990s, the period of the New Economy mania and the early commercial internet—an enthusiasm which was at least backed up with investment in the technology that was supposed to bring about the future. We haven’t seen much of that in the latest iteration of tech mania, the era of Uber and Airbnb.
Here’s a look at some components of private investment. The equipment and especially the structures trendlines show a persistent downward path. Against that, IP’s rise stands out—to the point where it’s surpassed investment in buildings and is rivaling equipment. Both equipment and structures used to be several times IP. Capitalists are spending less money on things that are supposed to promote general prosperity and more on legal arrangements that protect theirs.
Low levels of net private investment aren’t driven by declines in gross investment, which has been pretty stable. Instead, the major reasons for the decline are a shift towards shorter-lived equipment and the immateriality of intellectual property (IP) and a shift away from buildings. From 1950–1999, net fixed private investment averaged 32% of gross; since 2000, it’s averaged 20%—and 16% since 2020. Every asset category has seen that shift—even buildings.
Intellectual property investment, whose share of business investment grew from 8% in 1950 to 40% today, adds another layer of fleetingness to the story. Business ideologues love to tout IP as a stimulant to innovation; who’d invent anything if they couldn’t patent it? Lots of people would, actually. That aside, most basic innovation in sectors like computing and pharmaceuticals have been funded by public entities, not private companies, who then appropriate those innovations to make profits from research they didn’t pay for. Instead of supporting innovations, a lot of IP investment is about trying to establish monopolies, be it in the latest variation on an antidepressant or a Disney cartoon character. But even here the trend towards shorter-lived assets is visible: net IP investment went from 27% of gross in the 1950s and 1960s to 16% since.
For the public sector, the decline in net investment has been more dramatic, falling from around 2% of GDP in the early decades on the graph to 0.4% since 2020. (It’s 0.3% so far in 2022.) Like the private sector, we’ve seen a shift towards shorter-lived assets, but unlike the private sector, we’ve also seen a decline in gross investment, which fell by almost half between the 1960s and 2020s. Net public investment as a percentage of gross went from 67% in the 1950s and 1960s to 27% in the 2020s. Net federal civilian investment is just 0.1% of GDP so far this decade, a third its 1950–1999 average. State and local investment has fallen harder, down by almost three quarters from that 50-year average to 0.5% in the 2020s (0.3% so far this year).
And as the graph above shows, residential net investment isn’t doing too great either: it went from an average of 2.8% of GDP from 1950 to 1999 to 1.7% in the 2020s. Unlike the mid-2000s housing bubble, which took net residential investment up to 3.8%, the highest since the early post-World War II years, the latest bubble took net housing investment up to just 1.9% of GDP last year. It’s fallen back to 1.4% in 2022. That’s not the way to meet a housing deficit estimated by Freddie Mac at 3.8 million units.
The burst of net private investment in the late 1990s gave us a major productivity acceleration, but it was not to last. And the burst in civilian public investment from the early 1950s through the late 1960s gave us interstate highways, schools, and state university systems. The long declines in net investment, both private and public, have given us stagnant productivity growth and a collapsing infrastructure.
As I put it when I wrote about net investment five years ago, “If I were a debased purveyor of clickbait, I’d call this “Everything that’s wrong with America in two charts.” But I’m not, so I won’t. But still….”
More true than ever.
Photo is by me, of the 7th Ave stop on the G line of the New York City subway.
Just added to my radio archive (click on date for link):
July 7, 2022 Jenny Brown of National Women’s Liberation (and author of Without Apology and Birth Strike) on the early struggle for abortion rights that led to Roe and what we can learn from it for today • journalist David De Jong, author of Nazi Billionaires, on how respectable German businessmen became loyal Nazis
Just added to my radio archive (click on date for link):
June 2, 2022 Forrest Hylton on the first round of the Colombian presidential election, which was bad news for the leftist Petro • Olúfẹ́mi Táíwò, author of Elite Capture, on how the ruling class has debased identity politics, and how we could reconstitute it
Just added to my radio archive (click on date for link, and apologies for the late posting):
May 26, 2022 Heather Berg, author of Porn Work, on relations of production in sex work • Kevin Young and Leonard Seabrooke, co-authors of this paper, on the contrasting collegial styles of the Chicago and Charles River schools of economics
The USA is the country where everyone feels middle-class, right? No.
Gallup is out with the latest edition of a question it’s asked ten times over the last twenty years: “If you were asked to use one of these five names for your social class, which would you say you belong in?” When they did the survey in April, the largest set of respondents said “middle,” 38%—but that’s not much more than a third. Almost as many, 35%, said “working” (a term that has often been pronounced obsolete).
Here’s some more detail:
A striking thing about the chart is its upward skew. The midpoint is just 4 points into the “middle” category, and “upper-middle” is nowhere near that midpoint—it begins about 5/6 of the way to the top. Still, it’s remarkable that in a country of alleged universal middle classness, almost half the population identifies as sub-middle.
Over the last 20 years, upper-middle and middle have declined by 8 percentage points and working and lower have risen by 9. If you start the clock in 2005, the peak of the housing bubble, the “middle” share has fallen by 9 points, with most going into “working.” The Great Recession that followed the bursting of that bubble has a lot to do with that trend, but ten years of expansion following that miserable downturn did nothing to change middle-class self-identification.
Before one gets encouraged by these stats into thinking proletarian class-consciousness is on the rise, a caveat: more Republicans (38%) are likely to identify as working class than Democrats (30%). But to conclude on a more encouraging note: 49% of those aged 18–34 call themselves working class, twice the share of the over-55s. Nice to see that clarity in the young.