<![CDATA[EconWorks]]>https://blog.econworks.comhttps://substackcdn.com/image/fetch/$s_!hLrr!,w_256,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F24b47a00-2b89-47e2-9a88-7a6fc71d1555_256x256.pngEconWorkshttps://blog.econworks.comSubstackWed, 05 Aug 2026 07:43:25 GMT<![CDATA[Age of AI—Episode 2.5: The Symmetric Loop: What Happens When Authors and Reviewers Use the Same AI]]>https://blog.econworks.com/p/the-symmetric-loop-what-happens-whenhttps://blog.econworks.com/p/the-symmetric-loop-what-happens-whenTue, 04 Aug 2026 10:39:26 GMTArtificial intelligence is doing more than changing the way researchers write papers. Writers can now use AI to preempt objections, test specifications, enhance code, and polish manuscripts before submission. Reviewers can use similar tools to identify shortcomings, generate reports, and recommend revisions.

It’s a self-reinforcing loop: the author’s AI is making the paper for the reviewer’s AI.

In the short term, cheaper execution might flood academia with polished but unoriginal work. But the AI slop might not be the long-term result. As common as writing, coding, reviewing literature, and doing robustness checks are, they lose their signaling value for research quality.

What is still scarce and harder to produce is an original idea, knowledge of how the evidence was generated, proximity to the real-world problem, human judgment, and accountability.

On this episode of *The Age of AI*, we explore what happens to academic publishing when technical execution becomes a commodity—and why the age of AI might ultimately bring research back to ground truth.

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack: https://blog.econworks.com

]]><![CDATA[Price Discrimination and Railroad Competition]]>https://blog.econworks.com/p/price-discrimination-and-railroadhttps://blog.econworks.com/p/price-discrimination-and-railroadFri, 31 Jul 2026 11:22:02 GMTThe simple argument is this: rail competes with trucking, and therefore the relevant market is all freight transportation. If so, then mergers of railroads operating in different regions might seem innocuous.

But that argument assumes all railroad customers are in the same market.

In this episode of EconWorks, we explain why railroad freight is better thought of as a market with price discrimination, where different customers face fundamentally different competitive conditions. Contractual bargaining is used by industrial shippers who ship large volumes. Smaller shippers often compare trucking and rail on a price basis.

This distinction changes how economists should think about market definition, the hypothetical monopolist test, and the competitive effects of railroad mergers.

Topics:

  • Why railroad freight is not a single market Big shippers vs. little shippers

  • Substitution or negotiation

  • Definition of market and monopoly

  • Why trucking will be in the relevant market only where the merger makes little difference

  • Implications for railroad merger analysis



  • If you like clear explanations of antitrust, competition policy, AI, and platform economics, subscribe to EconWorks.

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack: https://blog.econworks.com

]]><![CDATA[Why Did an Alleged Egg Cartel End in a Civil Settlement?]]>https://blog.econworks.com/p/why-did-an-alleged-egg-cartel-endhttps://blog.econworks.com/p/why-did-an-alleged-egg-cartel-endThu, 30 Jul 2026 11:41:15 GMT

At first glance, that may seem surprising. If competitors coordinated their conduct, why would the government settle rather than prosecute them criminally?

The settlement involving Cal-Maine Foods, Hickman’s Egg Ranch, and Versova illustrates why the answer is not always straightforward. The companies agreed to pay $3.3 million, donate well over 53 milli…

Read more

]]><![CDATA[Age of AI—Episode 2.5: The Symmetric Loop]]>https://blog.econworks.com/p/age-of-aiepisode-25-the-symmetrichttps://blog.econworks.com/p/age-of-aiepisode-25-the-symmetricThu, 30 Jul 2026 11:38:57 GMT

In 30 Seconds

Artificial intelligence may initially flood academia with polished but unoriginal work. That is the short-run AI-slop problem.

But slop is unlikely to be the final equilibrium. When authors can use AI to anticipate reviewers—and reviewers can use similar AI to evaluate authors—routine academic execution becomes easier to produce and less inf…

Read more

]]><![CDATA[Age of AI — Episode 2: When Models Meet Reality: AI and the Future of Economic Analysis]]>https://blog.econworks.com/p/age-of-ai-episode-2-when-models-meet-6cdhttps://blog.econworks.com/p/age-of-ai-episode-2-when-models-meet-6cdThu, 30 Jul 2026 11:09:25 GMTEconomic models are powerful—but only if their underlying assumptions are valid.

In this episode we explore how artificial intelligence is changing the way economists test these assumptions.

Instead of settling for one model, AI lets economists produce hundreds of alternative versions, test sensitivity, and better understand which conclusions are robust.

Topics include:

  • The three-step process that economists use to test models

  • Assumptions are important

  • Sensitivity analysis assisted by AI

  • Algorithmic pricing as an example

  • Human judgment versus computer power

  • Why economists are more valuable, not less, in the age of AI

This episode is a sequel to Episode 1, and it explains how AI is changing the practice of economics long before empirical analysis begins.

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack: https://blog.econworks.com

]]><![CDATA[The Hidden Economics of Railroad Mergers: Why Market Definition Gets It Wrong]]>https://blog.econworks.com/p/the-hidden-economics-of-railroadhttps://blog.econworks.com/p/the-hidden-economics-of-railroadWed, 29 Jul 2026 11:16:16 GMTThe simple argument is this: rail competes with trucking, and therefore the relevant market is all freight transportation. If so, then mergers of railroads operating in different regions might seem innocuous.

But that argument assumes all railroad customers are in the same market.

In this episode of EconWorks, we explain why railroad freight is better thought of as a market with price discrimination, where different customers face fundamentally different competitive conditions. Contractual bargaining is used by industrial shippers who ship large volumes. Smaller shippers often compare trucking and rail on a price basis.

This distinction changes how economists should think about market definition, the hypothetical monopolist test, and the competitive effects of railroad mergers.

Topics:

  • Why railroad freight is not a single market Big shippers vs. little shippers

  • Substitution or negotiation

  • Definition of market and monopoly

  • Why trucking will be in the relevant market only where the merger makes little difference

  • Implications for railroad merger analysis

If you like clear explanations of antitrust, competition policy, AI, and platform economics, subscribe to EconWorks.

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack: https://blog.econworks.com

]]><![CDATA[Age of AI — Episode 2: Can AI Test Economic Models Better Than Humans?]]>https://blog.econworks.com/p/age-of-ai-episode-2-can-ai-test-economichttps://blog.econworks.com/p/age-of-ai-episode-2-can-ai-test-economicTue, 28 Jul 2026 11:04:06 GMTAll economic models contain assumptions.

But what if those assumptions change?

In Episode 2 of the Age of AI series, we explore how economists are using artificial intelligence to test economic models, not to replace human expertise, but to systematically test alternative assumptions and discover which conclusions are really robust.

What you'll learn:

  • Why all economic models are simplifications of reality

  • How economists study and test theoretical models

  • Why sensitivity analysis is important

  • How AI can develop and test thousands of alternative model specifications

  • Why we still need human judgment

  • What algorithmic pricing can teach us about AI and competition

AI is not meant to replace economists, but rather to give them a powerful new tool for asking better questions, exploring competing explanations, and producing more reliable economic analysis.

If you’re interested in antitrust, economics, artificial intelligence, or competition policy then this is the episode for you.

Subscribe for our next episode on how AI is helping economists grapple with messy real-world data.

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack: https://blog.econworks.com

]]><![CDATA[The Wrong Market, The Wrong Test: Railroad Mergers and the Price Discrimination Problem]]>https://blog.econworks.com/p/the-wrong-market-the-wrong-test-railroadhttps://blog.econworks.com/p/the-wrong-market-the-wrong-test-railroadFri, 24 Jul 2026 12:12:43 GMT

Read more

]]><![CDATA[Age of AI — Episode 2: When Models Meet Reality]]>https://blog.econworks.com/p/age-of-ai-episode-2-when-models-meethttps://blog.econworks.com/p/age-of-ai-episode-2-when-models-meetFri, 24 Jul 2026 11:57:33 GMT

Read more

]]><![CDATA[They Built an 87-Camera Surveillance State to Stop Cheating. It Ruined Them.]]>https://blog.econworks.com/p/they-built-an-87-camera-surveillancehttps://blog.econworks.com/p/they-built-an-87-camera-surveillanceFri, 24 Jul 2026 11:01:26 GMTWhat happens when a criminal conspiracy treats trust as an engineering problem? According to a recently unsealed DOJ indictment, four major shipping container manufacturers allegedly banded together to artificially limit global output. To keep members from cheating, they shared real-time factory camera feeds with each other.

But their biggest legal vulnerability wasn’t the cameras—it was the basic math of expanding their membership. This video explores the fascinating logic of corporate defection.

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack: https://blog.econworks.com

]]><![CDATA[Age of AI—Episode 1: Merger Enforcement Is Changing—And AI Is Accelerating It]]>https://blog.econworks.com/p/age-of-aiepisode-1-merger-enforcement-969https://blog.econworks.com/p/age-of-aiepisode-1-merger-enforcement-969Thu, 23 Jul 2026 11:07:02 GMTMost people think merger enforcement is decided in court.

The data show a different pattern.

Across the United States and the European Union, an increasing share of challenged mergers resolve before a judge ever rules. Transactions are withdrawn, abandoned, or restructured through negotiated remedies long before litigation concludes.

This shift is changing the role of economists—and artificial intelligence is accelerating it.

In Episode 1 of the Age of AI series, we examine the following:

• Why merger decisions are moving earlier in the review process

• How economists are becoming strategic advisors rather than litigation specialists

• How AI helps firms anticipate regulatory concerns

• Why better economic analysis can lead to faster and more informed decisions

• What this means for the future of competition policy

Whether you’re an economist, lawyer, regulator, student, or simply interested in AI and competition policy, this episode provides an accessible introduction to one of the biggest changes occurring in modern merger enforcement.

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack: https://blog.econworks.com

]]><![CDATA[The Box Cartel: Inside the $35B Shipping Conspiracy]]>https://blog.econworks.com/p/the-box-cartel-inside-the-35b-shippinghttps://blog.econworks.com/p/the-box-cartel-inside-the-35b-shippingWed, 22 Jul 2026 11:02:42 GMTIn May 2026, a federal indictment pulled back the curtain on an alleged global shipping container cartel. Controlling roughly 95% of the market, executives installed 87 factory cameras and signed a written contract called the “Moon Gazing Fund” to force compliance. But the economics of cartel formation dictate that the seeds of failure are baked into the architecture from day one.

We break down the industrial organization theory behind the competitive fringe, the engineering of trust under George Stigler’s framework, and why competition leniency programs inadvertently act as an “entry subsidy” for late-joining defectors.

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack: https://blog.econworks.com

]]><![CDATA[Age of AI—Episode 1: Merger Enforcement in Transition]]>https://blog.econworks.com/p/age-of-aiepisode-1-merger-enforcementhttps://blog.econworks.com/p/age-of-aiepisode-1-merger-enforcementTue, 21 Jul 2026 11:04:21 GMTMerger enforcement is no longer defined primarily by courtroom battles.

Across major jurisdictions, parties resolve many challenged mergers long before a final judicial decision through negotiated remedies, transaction restructuring, or abandonment.

In this opening episode of the Age of AI series, we explore how this shift is transforming the work of economists and why artificial intelligence is becoming increasingly valuable during the earliest stages of merger review.

Topics include:

• Trends in merger enforcement across major jurisdictions

• The evolution of the economist’s role

• AI-assisted economic analysis

• Early negotiation and remedy design

• The future of competition policy

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack: https://blog.econworks.com

]]><![CDATA[The Box Cartel: Trust, Power, and the Logic of Defection]]>https://blog.econworks.com/p/the-box-cartel-trust-power-and-thehttps://blog.econworks.com/p/the-box-cartel-trust-power-and-theFri, 17 Jul 2026 11:38:20 GMT

In 30 Seconds: The Box Cartel

  • The Allegation: Four of the biggest shipping container manufacturers in the world allegedly established a cartel in 2019 and agreed to limit output and stop capacity expansion in order to increase profits, according to a recently released DOJ indictment.

  • The Structural Necessity: The founding four were compelled to enlist tw…

Read more

]]><![CDATA[Age of AI — Episode 1: Merger Enforcement in Transition]]>https://blog.econworks.com/p/age-of-ai-episode-1-merger-enforcementhttps://blog.econworks.com/p/age-of-ai-episode-1-merger-enforcementFri, 17 Jul 2026 11:35:58 GMT

Read more

]]><![CDATA[AI Infrastructure: Capacity, Constraints, and Competition]]>https://blog.econworks.com/p/ai-infrastructure-capacity-constraintshttps://blog.econworks.com/p/ai-infrastructure-capacity-constraintsFri, 17 Jul 2026 11:09:08 GMTArtificial intelligence is creating unprecedented demand for computing infrastructure.

In this episode, we examine how data centers work, why AI-ready capacity has become a distinct market, and why infrastructure constraints—not market power—are driving today’s shortages.

We also discuss what these developments mean for competition policy, investment, and the future of AI.

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack: https://blog.econworks.com

]]><![CDATA[How a failed Silicon Valley acquisition foreshadowed today’s battles over AI, platforms, and digital gatekeepers*]]>https://blog.econworks.com/p/how-a-failed-silicon-valley-acquisitionhttps://blog.econworks.com/p/how-a-failed-silicon-valley-acquisitionThu, 16 Jul 2026 11:42:01 GMTIn 2010, Google attempted to acquire Groupon for roughly $6 billion. Groupon rejected the offer amid fears that antitrust scrutiny could trap the company in a prolonged review process during a critical growth phase.

The company later IPO’d, faced accounting controversies, and lost most of its market value.

But the larger significance of the failed deal may lie elsewhere.

This episode explores how merger uncertainty, nascent competitor concerns, and evolving antitrust theories helped foreshadow many of today’s debates involving:

  • Big Tech,

  • AI ecosystems,

  • digital gatekeepers,

  • and platform acquisitions.

Read the full article and the graphic analysis.

[Article Link]

Explore more visual economics content:

https://econworks.com

YouTube:

https://www.youtube.com/@EconWorks-d3e

Substack:

EconWorks
Industrial organization insights on antitrust, digital platforms, and competition in ecosystem markets.
]]><![CDATA[The Market for AI-Ready Data Center Capacity]]>https://blog.econworks.com/p/the-market-for-ai-ready-data-center-437https://blog.econworks.com/p/the-market-for-ai-ready-data-center-437Wed, 15 Jul 2026 11:12:02 GMTAI has fundamentally changed the economics of data centers. This episode explains why AI-ready capacity has become a distinct market, how GPU workloads are reshaping infrastructure, and why today’s shortages reflect power, cooling, and permitting constraints rather than market power.

Topics include:

* AI-ready versus traditional data centers

* Power density and liquid cooling

* The end of the efficiency plateau

* Why vacancy rates have collapsed

* What this means for competition policy

If you’re interested in AI, infrastructure, or antitrust economics, this episode provides the technical background behind today’s compute race.

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack: https://blog.econworks.com

]]><![CDATA[The Startup That Rejected Google — Then Lost 90% of Its Value]]>https://blog.econworks.com/p/the-startup-that-rejected-googlehttps://blog.econworks.com/p/the-startup-that-rejected-googleTue, 14 Jul 2026 11:32:13 GMTGoogle once offered roughly $6 billion to acquire Groupon—one of the fastest-growing startups in Silicon Valley history.

Groupon rejected the deal, partly out of concern that antitrust scrutiny could trap the company in a lengthy merger review process during a critical hypergrowth phase.

The company later IPO’d at a massive valuation before losing most of its value amid accounting controversies, competition, and structural business weaknesses.

But the story is bigger than a failed acquisition.

This episode became an early preview of many issues that now dominate debates over:

• Big Tech acquisitions

• nascent competitor theory

• digital gatekeepers

• AI ecosystems

• startup “kill zones”

• and modern merger enforcement.

This EconWorks episode explores how antitrust uncertainty itself can shape innovation markets long before regulators formally intervene.

Read the full article and graphic analysis:

Explore more visual economics content: https://econworks.com

YouTube: https://www.youtube.com/@EconWorks-d3e

Substack:

EconWorks
Industrial organization insights on antitrust, digital platforms, and competition in ecosystem markets.
]]><![CDATA[The Market for AI-Ready Data Center Capacity]]>https://blog.econworks.com/p/the-market-for-ai-ready-data-centerhttps://blog.econworks.com/p/the-market-for-ai-ready-data-centerFri, 10 Jul 2026 11:38:23 GMT

In 30 Seconds

The data center market is not a generic real estate market—it is increasingly a market for AI-ready capacity, defined by high power density and liquid cooling capability.

AI workloads have eroded the industry’s previous efficiency gains, leading to a sharp increase in electricity demand and pushing infrastructure to its physical limits.

The r…

Read more

]]>