Analysis of 200M newspaper pages: Sentiment has collapsed over the past 50 years
An analysis of 200M newspaper pages finds that economic and general news sentiment has grown markedly more negative over the past 50 years, even as many aggregate indicators like GDP and global living standards have improved. Commenters debate whether this reflects real deterioration in conditions—citing wage stagnation for median workers, housing affordability, inequality and deindustrialization—or primarily rising expectations, partisan framing, and the incentives of profit-driven, 24/7 media to emphasize outrage and bad news. Others question the study’s methods and note that press tone may track only loosely, and often pessimistically, with the average person’s actual material situation.
Data, scope, and methodology
- Several ask if the underlying dataset is public; others doubt it due to copyright.
- One commenter links the NBER working paper PDF.
- Some are wary of “big data” claims, asking how newspapers were selected, whether wire stories were deduplicated, and whether mix‑shift across outlets could drive the trend.
- It’s noted the study covers both economic and non‑economic sentiment, and that both trend more negative.
Is life improving while sentiment falls?
- One camp argues many objective indicators (GDP, consumption, inflation control, employment) have improved; falling sentiment may reflect rising expectations or media bias rather than actual decline.
- Others counter that stagnating real wages, weak middle‑class prospects, and insecure living conditions mean conditions have not meaningfully improved for many, even if aggregates look good.
- Several emphasize that subjective security (e.g., not living paycheck-to-paycheck) matters as much as material levels.
Wages, purchasing power, and inequality
- Dispute over whether “purchasing power has collapsed”:
- One side: real wages (especially median) stagnated, cost of living and housing rose, gains went to top earners; wage–productivity divergence graphs are cited.
- Other side: labor share of GDP is roughly stable, implying real wages broadly track productivity; some of the famous graphs are criticized as methodologically misleading (e.g., mixed deflators).
- Distribution vs averages is a recurring theme: aggregates can rise while median workers feel stuck.
Housing, land, and vacancies
- Long subthread debates whether the core problem is not building enough housing vs. misallocation (vacancies, investment properties, oversized new builds).
- One side emphasizes zoning/NIMBY constraints and “filtering” (luxury builds free up cheaper stock).
- The other stresses vacancy rates, investor ownership, geographic imbalance, and preference for single-family homes.
Media incentives and negativity bias
- Many blame profit-driven, 24/7, rage‑optimized media for increasingly negative tone.
- A cited example: a Russian outlet lost most of its readers when it ran only positive news for a day.
- Comparisons are made to persistent public belief that crime is rising even when data show declines.
Politics, policy, and globalization
- Some link the timing (1970s–80s) to neoliberal policies, deregulation, and weaker unions, arguing these channeled gains to a minority.
- Others respond that globalization and the rise of countries like China were not “caused” by Western leaders and that governments have limited control over such macro trends.
- Thread notes partisan bias increasingly shapes perceived economic conditions.
Historical context and “collapse” framing
- A few question calling a 50‑year trend a “collapse”; others note collapses (empires, cosmological) can be very slow.
- It’s highlighted that reported economic sentiment since the 1970s is often more negative than during the Great Depression, which many see as evidence of perception/media effects rather than worse fundamentals.