Three Easy Pieces
The US right-left political convergence mimics Europe; A reality check on AI productivity gains; And a balanced take on renewables and power outages.
Right, left, left, right
In The Economic Irrelevance Of Europe’s Politics, I noted that when it comes to economic policy, Europe’s left and right are nearly indistinguishable. “Both sides support big government, a generous welfare state, high taxes and pervasive regulation. Because that’s what voters want.” — I wrote. I also acknowledged that a similar convergence is taking shape in the US. Right on cue, Trump’s pick of J.D. Vance as Vice-Presidential candidate signals a further shift to the left by the Republican Party. Vance appears to favor industrial policy as much as the Democrats do, though he has a different set of favored industries — he would subsidize gasoline-powered cars rather than electric vehicles, for example. He favors trade unions. He has praised Lina Khan’s Federal Trade Commission’s heavy-handed regulatory approach. All this when Trump’s Republicans already prefer high government spending, large deficits and protectionism.
The Democratic Party has been running further to the left, and maintains a marked preference for higher taxes. But since deficits can’t increase forever, at some point the Republicans will have to choose whether to cut spending or raise taxes, and the way things are going … who can tell.
The Republicans’ leftward lurch is partly driven by our manifest failure to manage the adverse impacts of trade globalization: the hollowing out of American manufacturing has played a key role in shaping Vance’s views. But there are two ways to support the struggling sections of America’s population: opportunities or handouts. The Vance of Hillbilly Elegy seemed to favor opportunities — the Vice Presidential candidate leans towards handouts.
When it comes to economic policies, the difference between Republicans and Democrats is becoming one of degrees rather than fundamental views — and this does not bode well for America’s economic future.
AI and productivity: dreams vs reality
Innovation will eventually give us a growth boost, but for now the AI hype runs well ahead of what the technology is delivering. A new study by the freelancing platform Upwork highlights the major gap between hope and reality: almost all C-suite corporate leaders (96%) believe AI will boost their companies’ productivity; but over three-quarters of workers (77%) say that so far it has actually lowered productivity and added to their workload. Half (47%) say they have no idea how the expected productivity gains might be achieved.
We face two problems here. The first is that the human-like chattering aptitude of Generative AI tools like ChatGPT has tricked us into vastly overestimating their actual intelligence (and reliability). We have a long way to go before these systems even approach human intelligence in a broad sense. The second is that it takes a lot of investment and adaptation to incorporate AI into business processes — we’ve seen this with digital-industrial technologies for the past decade and a half.
A flexible, competitive economy has the best chance to rapidly get economic value from transformational innovations. When the first Information and Communication Technology revolution came around in the mid-1990s, the US enjoyed much greater productivity gains than Europe. Political parties on both sides of the political spectrum — and on both sides of the Atlantic — might want to reflect on that.
A balanced take on an unbalanced grid
Last week I wrote that the electricity grid will need substantial investments in the coming years, putting upward pressure on electricity prices. Some of these investments are needed regardless of whether we want to use renewables or fossil fuels; others are needed to switch to greener generation. I also noted that the US grid has become less stable, with a greater incidence of outages. The US Energy Information Administration provides data like the System Average Interruption Duration Index (SAIDI) and the System Average Interruption Frequency Index (SAIFI), which measure respectively the minutes of electricity interruption per year and the number of interruptions per year experienced by the average customer. The following chart shows the clear increase in electricity interruptions: for the US as a whole the average SAIDI for 2020-2022 (latest available) is 80% higher than the 2013-16 average.
Source: US Energy Information Administration
My friend and favorite sparring partner on these issues, energy entrepreneur Fabio Ficano, quickly cautioned that fossil fuels have been a major contributor to power interruptions. Fabio pointed me to this Deloitte report, which notes that in 2022 gas power outages have been a greater cause of power unreliability than renewables, particularly during Winter Storm Elliott in late December 2022. Similarly, the February 2021 Texas power outage was due to natural gas plant failures, as Bill Nussey shows here. The Deloitte report also documents examples where the combination of renewables and storage played a major role in meeting peak demand and therefore strengthening grid resilience. Deloitte argues that the widespread perception that natural gas and nuclear power are vastly more resilient than renewables is outdated and runs counter to reality.
The strengthening of the renewables and storage combination is extremely welcome news. Further progress in energy storage at scale will be crucial in allowing us to rely more and more on renewables. I am happy to acknowledge that Fabio’s arguments and the Deloitte report suggest we are further along than I thought. I would still urge some caution, though. California — once again — stands out as an exception to the argument that renewables improve grid resilience: as the next chart shows, California’s interruption index (SAIDI) for 2020-2022 (latest available) is 140% higher than the 2013-16 average, almost twice worse than the national average. (Note that the 2020-22 average leaves out the disastrous 2019…) Perhaps this tells us more about California than about renewables, but it’s worth pondering.
Source: US Energy Information Administration
Also, a recent Energy Information Administration analysis shows that during last month daily natural gas electricity generation spiked to a new record, as a summer heatwave coincided with a steep drop in wind generation — a tangible reminder than stored renewable energy is not yet enough.
Source: US Energy Information Administration
The chart also shows that over the past six years the summer spikes in natural gas electricity generation have gotten progressively higher: even as renewable generation capacity has increased and storage has improved, the need for natural gas to meet peak demand kept increasing.
Looking at the balance of these data, I would say that:
I feel more optimistic that renewables and storage will be able to play an increasingly important role in creating a cleaner, stronger and more stable grid;
I am still convinced that we’ll need an important contribution from fossil fuels for quite a while longer, and that a balanced combination of fossil fuels and renewables is the best bet to meet fast-growing energy demand while reducing the incidence of outages. (Let’s maybe add to the expenditures list some investment to reduce the technical vulnerability of gas plants to extreme cold…)
I'm only going to comment on the AI productivity subject. This reminds me of the late 1990's and early 2000's when BPR & ERP became huge buzzwords for magically improving productivity in organizations! Without going into specifics, suffice it to say that these gains never really happened in the desired timelines or saved major costs. 😊