$DMCY · THE DEMOCRACY DESK NEXT HARVEST: T-000 DAYS MARKET OPEN — CLOSES: NEVER 00:00:00 E.GILLISSIE / RESEARCH

$DMCY · SECTOR: ATTENTION EXTRACTION · SELL-SIDE COVERAGE OF AMERICAN DEMOCRACY

Every election
is earnings season.

American democracy manufactures one commodity with perfect reliability: a captive audience of voters. Campaigns are legally compelled to buy it, a shrinking club of media companies owns the inventory, and the Supreme Court has ruled the price can never be capped. The harvest now under way — the 2026 midterms, 108 days out — is projected at $11.6 billion in political advertising: the most expensive cycle of any kind, ever, presidential years included. This terminal tracks the trade in real time. The asset being liquidated is the republic.

LIVE WHAT POLITICAL ADS EARN, EVERY SECOND — MODELED FROM THE $11.6B CYCLE T+00:00

Booked from democracy since you opened this terminal

$0.00

That is the projected $11.6 billion spread evenly across every second of the two-year cycle: $183.92 per second, around the clock. It has not stopped for you.

 

$184/sec
All political advertising
$11.6B proj. ÷ one 2-yr cycle
$89/sec
Local broadcast TV alone
$5.6B proj. · 48% of the take
$28/sec
Just four station owners
at their 2024 pace of $1.77B

F1 · Price historyThe scissors. Two lines, one story. What each federal cycle costs keeps setting all-time highs. What Americans think of the Congress those cycles produce keeps probing all-time lows. Run your cursor across a quarter-century of the divergence — every dollar of the green line is somebody's revenue.

F1THE HARVEST — COST OF CYCLE vs. APPROVAL OF PRODUCT 2000 – 2024 · OPENSECRETS × GALLUP
CYCLE2024 COST OF FEDERAL ELECTIONS$15.9B CONGRESS APPROVAL AT ELECTION16% ⌖ trace the tape
Of each total, the slice spent on ads — the part media collects: '18 $4.0B '20 $9.0B '22 $8.9B '24 $11.2B '26 $11.6B PROJ ← ATH
Two different money numbers, on purpose: the green bars are everything campaigns spend — staff, travel, mail, and ads (OpenSecrets). The smaller chip figures are the advertising slice alone: the part that becomes media revenue (AdImpact). The red line is Gallup's approval of Congress nearest each election. Between 2016 and 2020 the price of a cycle more than doubled. Approval's answer: 16%, the lowest election-year reading in the series. And AdImpact now projects the 2026 midterm to out-spend every presidential cycle in history. Rising price, collapsing product — in any other market, that's a scandal. Here it's the business model.

F2 · Market depthThe order book. Campaign cash is the boring number — the revealing one is who fills the order. Here is the sell side of American attention from the last complete cycle: every channel quoting inventory, the four market makers of local television who cleared $1.77 billion between them, and the fill log ticking below.

F2ORDER BOOK — WHO SOLD DEMOCRACY ITS AUDIENCE 2024 — THE LAST FINISHED CYCLE

WHO SOLD THE AIRTIME · bar width = share of the $11.2B in ads

Hover a row for the detail.
◉ ORDER FLOW — what the buying looks like, one ad at a time ILLUSTRATIVE — MODELED AT MARKET RATES, NOT A LIVE FEED

    THE BIG FOUR — the companies that own "local" TV

    GRAY MEDIA#1 earner · +15.6% vs 2020$497M
    NEXSTAR"85% of contested election markets"$491M
    SINCLAIR+16% vs 2020$405M
    TEGNAAcquired by Nexstar, 2026$373M
    Four owners · $1.77B in one cycle — more political money than Google and Meta's entire reported take ($1.35B) combined.
    Insider statement · Q4 2024 earnings Nexstar's CEO credited record results to the effectiveness of local television broadcasting and our presence in nearly 85% of contested election markets. Contested elections are not the company's risk factor. They are its footprint strategy.
    Cross-reading — the local paradox

    "Local TV" is the biggest line on this book, and almost none of it is local. The stations are inventory in national portfolios; the political windfall books to corporate headquarters, services acquisition debt, and returns to shareholders. Meanwhile the layer of local journalism that actually covers the school board keeps collapsing: America has lost roughly 3,500 local newspapers since 2005, newsroom staffing has fallen by nearly half, and 213 counties now have no local news outlet at all. So hold both facts at once: local screens carry billions in political money every cycle, and the newsrooms behind those screens keep shrinking. The money for local accountability visibly exists — it is on this page. It is simply contracted to persuasion, not coverage: stations sell airtime to campaigns. They do not spend the proceeds on reporters.

    Fill report: channel figures are AdImpact's 2024 count; the digital line is the Brennan Center / OpenSecrets / Wesleyan compilation of Google + Meta disclosures — an undercount, since no federal law requires platforms to publish political-ad data at all. Station-group figures are from company earnings filings. Consolidation note: with Nexstar's purchase of TEGNA closed in 2026, two of these four rows are now one company.

    F3 · Trade historyThis trade is 2,000 years old. Selling access to voters is the oldest continuously operating market in Western politics. Only the venue changes — games, taverns, printing presses, airwaves, feeds. Open any execution for the fill details, and note the column that never empties: who collected.

    F3EXECUTED TRADES — SELECTED FILLS, 53 BC → 2024 AD TAP A ROW TO OPEN THE FILL
    Blotter note: prices as recorded by period sources; ancient and pre-modern figures are not inflation-comparable and are shown as quoted. Full citations in disclosures.

    F4 · FundamentalsThe underlying. Any honest desk quotes the asset it's stripping. Here is democracy's own board — and one risk disclosure the industry would rather not file.

    F4CIVIC INDICATORS — THE UNDERLYING ASSET LATEST PRINTS

    Read the colors the way the industry does: green ▲ = good for the seller · red ▼ = the state of the asset itself.

    CONGRESS APPROVAL · ELECTION DAY 2024
    16%▼
    Lowest election-year reading on the Gallup series — while 96% of House incumbents who ran were re-elected.
    TRUST IN MASS MEDIA · GALLUP 2024
    31%▼
    Record-low territory for the very industry booking the election windfall.
    CAMPAIGN DURATION · 2024
    ~720▲
    Days from Trump's announcement to Election Day. Mexico's formal season: 3 months. India's: 44 days. Longer campaign, longer meter.
    PRICE CAP · U.S. FEDERAL RACES
    NONE
    Buckley v. Valeo (1976): spending is speech; caps are unconstitutional. Britain caps parties at £30,000 per seat.
    COST PER SENATE VOTE CAST · 2014
    $24▲
    A record at the time — set in a cycle with the lowest Senate-election turnout since 1990. More money, fewer voters.
    SEASONALITY · GRAY MEDIA Q3 2024
    +565%▲
    Year-over-year jump in political ad revenue. The harvest is biennial, contractual, and guaranteed by the Constitution's calendar.
    Risk disclosure — filed on behalf of the seller

    Broadcasters have little incentive to cover candidates free of charge, because it is in their interest to make campaigns buy the reach instead (McChesney, Rich Media, Poor Democracy). During the McCain–Feingold debate, a U.S. senator marveled that election season had gone from broadcast's worst revenue period to one of its most profitable; the bill's free-airtime provisions were stripped to get it passed. Identified risks to the seller: none.

    F5 · The counterfactualThe money is real. The waste is a choice. Other democracies cap what campaigns may spend, and their republics have not collapsed for it. Run the American election under British rules, scaled to American population, and almost the entire bill evaporates. What's left over is not abstract — it's teachers, school lunches, ambulances, and libraries that were never funded. Below, the delta. Then spend it yourself.

    F5THE COUNTERFACTUAL — REALLOCATION DESK MODELED · SOURCES IN DISCLOSURES
    What the 2024 U.S. federal cycle cost
    $15.9B
    −
    Same election under UK-style caps, scaled to U.S. population
    ≈ $0.3B
    =
    The delta — money we agree to burn as "part of the game"
    $15.6B

    Modeled: Britain's entire 2017 general election cost ≈ £42M under its £30,000-per-seat caps; scaled by population (×5.1) and converted, ≈ $0.3B. One cycle. Every two years.

    = the ENTIRE federal Head Start program for a year — with $3.3B left over
    Head Start's full FY2024 appropriation was $12.3B, serving ~716,000 children and pregnant women (HHS).
    = 3.4 YEARS of America's entire Community Health Center program
    The federal appropriation that anchors ~1,400 safety-net centers serving 32.4M patients: $4.6B/yr (FY2026).
    ≈ every public library in America, operating for a FULL YEAR
    Total U.S. public library operating spend runs ≈ $14B/yr across ~9,000 systems (IMLS).

    Spend it better_

    0% ALLOCATED
    YOUR BUDGET: ONE CYCLE'S DELTA — $15,600,000,000 ALLOCATED: $0 · REMAINING: $15.6B
    Start allocating. Fair warning: it is genuinely hard to spend fifteen billion dollars on communities — and the political-media machine does it on persuasion every single cycle, then resets the meter.
    Unit costs, all sourced: teacher $74,177/yr (NEA 2024–25 avg) · local reporter $60,280/yr (BLS median, May 2024) · Head Start slot $14,532/child-yr (HHS/ASPE FY2024 median) · school lunches $774/kid-yr (USDA free-lunch rate ≈ $4.30 × 180 days) · ambulance ≈ $300K (2024 municipal purchases) · health-center care $1,530/patient-yr (HRSA UDS 2024: $49.6B ÷ 32.4M patients) · library system $1.6M/yr (IMLS averages) · afterschool slot ≈ $950/child-yr (21st CCLC: $1.33B ÷ ~1.4M kids). The delta itself is modeled; the unit prices are not.
    F6ANALYST NOTE — INITIATING COVERAGE THE DEMOCRACY DESK
    THE AMERICAN ELECTIONSTRONG BUY AMERICAN DEMOCRACYUNDERWEIGHT

    We initiate coverage of the American election with a strong buy. Revenue is contractually biennial. Demand is inelastic — a candidate who declines to purchase attention forfeits the race. The Supreme Court prohibits price caps, the Constitution guarantees the schedule, and the customer base regenerates every cycle regardless of product quality. Guidance is already raised: the 2026 midterm is projected to out-spend every election in American history.

    We rate the underlying asset underweight. Approval of the product sits at 16%. Trust in the seller sits at 31%. The campaign is now permanent, the meter no longer stops, and the parties best positioned to demand reform are the ones cashing the checks.

    The counterfactual desk quantifies the opportunity cost: the excess of a single cycle over a capped election would fund the entire national Head Start program with billions to spare. This is not a market failure. Rome's vote was already, in one historian's phrase, a marketable commodity — the market has simply never closed since. The conflict of interest is not a flaw in the model. It is the model.

    The contest belongs to the candidates. The revenue belongs to the media. The casualty is democracy.

    CMDCOMMAND LINE — TALK TO THE DESK TYPE HELP · OR PRESS /
    DEMOCRACY DESK v2.6 — SESSION OPEN. THE MARKET NEVER CLOSES, BUT THE PROMPT IS YOURS. TYPE HELP FOR COMMANDS.
    DMCY>

    Disclosures & citations

    Every print on this terminal traces to a primary source. Modeled figures are labeled where shown.

    1. Cycle in Review 2023–24 & 2025–26 Political Projections — AdImpact: 2024 total ≈ $11.2B; broadcast $5.35B (48%); CTV $2.3B; cable ≈$1.9B; radio $326M; prior cycles ('18 ≈$4B, '20 $9.0B, '22 $8.9B); 2026 projection $11.6B — the most expensive cycle on record, broadcast $5.6B (48%).
    2. Cost of Election — OpenSecrets: federal cycle totals 2000–2024 ($3.1B → $15.9B); 2024 state + federal > $20B; outside spending ≈$4.5B; dark money $1B+.
    3. Company earnings & SEC filings via S&P Global, TVNewsCheck, The Desk: Gray $497M (+15.6% vs '20; Q3 political +565% YoY), Nexstar $491M, Sinclair $405M (+16% vs '20), TEGNA $373M; Nexstar–TEGNA close, March 2026; Nexstar CEO on "85% of contested election markets."
    4. Online Political Spending in 2024 — Brennan Center × OpenSecrets × Wesleyan Media Project: $1.35B on Google + Meta; disclosure voluntary and incomplete.
    5. Gallup: Congress job approval nearest each election, 2000–2024 (49% → 16%); trust in mass media, 31% (2024).
    6. Elections in the Roman Republic; Electoral Bribery in the Roman Republic (J. of Roman Studies); Slate: ambitus, Milo's 1,000 asses per voter (53 BC), the 10M-sesterces pledge (54 BC), the vote as "marketable commodity."
    7. ECPPEC (Newcastle Univ.); Britannica: "treating," rotten boroughs, £8,000 contests, and the £60,000 sale of Old Sarum (1802) — a seat with no residents.
    8. U.S. Senate Historical Office; A. Winkler (HNN): Mark Hanna's $3.5M war chest (≈$100M today), 100M+ pamphlets, corporate assessment "according to stake," zero disclosure until 1910.
    9. Museum of the Moving Image, The Living Room Candidate; Wisconsin Historical Society: "Eisenhower Answers America," Rosser Reeves, the $1.5M 1952 spot blitz, Stevenson's "antithetical to democracy."
    10. PBS NewsHour; W. Horncastle, The Conversation: ~720-day campaign vs. Mexico (3 mo.) and India (44 days); UK £30,000-per-seat cap and ≈£42M 2017 general election; Buckley v. Valeo (1976). Counterfactual = UK total, population-scaled ×5.1 — modeled.
    11. Wesleyan Media Project (post–Citizens United airings: House +20%, Senate +79%); "more money than there is airtime to buy" — CMAG; R. McChesney, Rich Media, Poor Democracy (1999); McCain–Feingold floor debate; Pew Research ($24 per Senate vote, 2014).
    12. Civic unit costs: NEA Rankings & Estimates 2024–25 (teacher $74,177); BLS OOH May 2024 (reporters $60,280); HHS/ASPE Head Start brief, FY2024 ($14,532 median/slot; program $12.3B, ~716K children); USDA/SNA free-lunch reimbursement ≈$4.30; municipal ambulance purchases 2024 (≈$250–330K); KFF/HRSA UDS 2024 ($49.6B ÷ 32.4M patients ≈ $1,530; federal program $4.6B FY2026); IMLS Public Libraries Survey (≈$14B national ops, ~9,000 systems); Afterschool Alliance (21st CCLC appropriation $1.329B serving ~1.4M students ≈ $950/slot-yr).
    13. Medill Local News Initiative, State of Local News (≈3,500 newspapers lost since 2005; 213 counties with no local outlet); BLS/Pew (newsroom employment down ~47%, 2008–2018) — the collapse side of the local paradox.
    Compiled for evangillissie.com · Prints current to July 2026 Session rate is cycle-averaged arithmetic, not a live feed · The market, regrettably, is real