Macro Watch: Topical Research from Eco3min
Macro Watch
Topical macro-financial research from Eco3min. Each study takes a current market or policy event and places it in its longer historical context: sourced from primary data, written for analysts and journalists. One new study per week.
All studies linked here remain available, with reproducible methodology and, where the sources’ terms allow, downloadable datasets. For evergreen frameworks (yield curve, real rates, inflation, liquidity), see the main data hub. For a monthly reading of the current macro regime, see the Eco3min macro barometer.
Latest studies
Most recent first. Each study covers a single topic in depth: data, structural framework, counter-arguments, and downloadable dataset.
28 studies · Updated every Tuesday
US homeownership rate by age of householder, 1982 to 2025
Households headed by someone aged 30 to 34 are the only one of the Census Bureau’s 12 age groups whose homeownership rate has been on both sides of 50% since 1982. Their rate stayed above half for 29 straight years, from 57.1% in 1982 through 2010, and has been under half every year since 2011, at 46.6% in 2025. Every five-year group from 25 to 64 owned less often in 2025 than in 1982, while the national rate went from 64.8% to 65.2%, as households shifted toward older age groups: those headed by someone 65 or older grew from 21% to 29% of all households. Held at the 1982 age mix, the 2025 rate would be 60.9%. The interactive version follows the 30 to 34 line, all 12 age groups, and 1982 against 2025 for each group.
Sources: U.S. Census Bureau · CPS/Housing Vacancy Survey, Table 12 · CSV download
Purchasing power lost by 13 currencies in their three largest inflation episodes, 1913 to 2025
For 13 currencies with consumer prices back to 1913 or 1914, Eco3min Research marks each currency’s inflation episodes with one pre-registered rule, then measures how much purchasing power its three largest episodes took. In the median currency they took 95.8%. The smallest such loss is the Swiss franc’s, 79.6%; the US dollar lost 88.3% inside its three largest episodes and the Icelandic krona 99.99%. Those episodes hold between 62% (US dollar) and 86% (Icelandic krona) of each currency’s rise in log prices; after 10,000 shuffles of the order of years, the median share falls to 27%. Purchasing power is measured at home with each country’s consumer prices, so exchange rates play no part. An interactive explorer compares the three largest episodes with all years for any currency and any start year.
Sources: Bank for International Settlements, long series on consumer prices (WS_LONG_CPI) · US Bureau of Labor Statistics, CPI-U (CUUR0000SA0) · ISTAT, Australian Bureau of Statistics and Swiss Federal Statistical Office · CSV download
Look-ahead bias in a macro regime backtest: real-time vs revised data, 2011 to 2026
Eco3min’s macro regime classifier was replayed on each day from June 2011 to September 2026 with the ALFRED vintages available on that day, then compared with the timeline it gives on today’s revised data. The two differ on 1,336 of 5,552 classified days (24.1%) and at 37 of 184 month-ends (20.1%, 95% block-bootstrap interval 11.4% to 30.4%). In 32 of those 37 months the real-time run sat in a neighbouring regime, one notch away on one axis. Of the 19 regime changes in the revised timeline, the month-end real-time runs also made 16: 8 in the same month, 1 two months early and 7 late, by one to six months. Run on final data, the code reproduces the full 586-month served timeline with 0 differences. An interactive strip sets both timelines side by side.
Sources: Federal Reserve Bank of St. Louis, ALFRED vintages (CFNAI, PCETRIM12M159SFRBDAL, SAHMREALTIME, IURSA, NFCI) · Federal Reserve Banks of Chicago and Dallas · Eco3min regime classifier v1.1.0 · CSV download
Inverted yield curve, no recession: what a real-time model showed, 1976 to 2026
Re-estimated every month since January 1976 on the data then available, a probit on the US 10-year minus 3-month Treasury spread beat the known recession frequency until 2021: from 1991 to 2021 its Brier skill score was +0.212. In May 2023 the model would have read 0.766 at a spread of −1.57 points, the third of 7 episode peaks, behind the readings ahead of the 1981 (0.992) and 1980 (0.920) recessions. As of November 2026, FRED USREC records no recession month after April 2020. Over 598 monthly origins to October 2025 the score falls to +0.054, not statistically distinguishable from the base rate, and the base model had lost 76.4% of the cumulative edge it held at the end of 2021. The interactive version slides the recession threshold from 1976 to 2026.
Sources: Federal Reserve Board H.15 via FRED (GS10, TB3MS) · Federal Reserve Bank of St. Louis (USREC) · CSV download
How are American homes heated? Main heating fuel, 1940 to 2025
In 1940, coal heated 54.7% of the US homes that reported a heating fuel in the decennial census. In 2025, fewer than 1 in 1,000 homes used it as their main heating fuel. Utility gas took the lead between the 1950 and 1960 censuses, peaked at 55.2% of occupied homes in 1970 and heated 46.9% in 2025, according to the American Community Survey. Electricity has risen at every census since 1950, when it heated 0.7% of homes, and stood at 41.8% in 2025, while fuel oil fell from 32.4% in 1960 to 3.4%. The gap between gas and electricity went from 47.5 points in 1970 to 5.1 in 2025, with the gas answers reworded since the 2024 questionnaire. The interactive version follows six fuels since 1940, two of them from 1950, the full mix of heating fuels and the yearly survey since 2010.
Sources: U.S. Census Bureau · decennial census (1940-2000) and American Community Survey 1-year, table B25040 · CSV download
Beef prices vs chicken in the US, adjusted for inflation, 1980 to 2026
In September 2026, a pound of ground chuck cost $6.48 in US cities and a pound of whole chicken $2.06, according to the average retail prices collected by the U.S. Bureau of Labor Statistics: the beef was worth 3.14 pounds of chicken. Over 2026 so far, the average is 3.3 pounds, against 1.7 in 1998, the lowest year since the series began in 1980, and no full year before 2026 reached 3. After inflation, ground chuck costs 9% less than in 1980 and whole chicken 29% less. In current dollars, chuck reached $6.98 in June 2026, the highest monthly price since 1980. The interactive version stacks the price of chicken under the beef line, compares three cuts of beef and follows seven cuts of beef, pork and chicken since 1980.
Sources: U.S. Bureau of Labor Statistics · average price series APU0000703111 and APU0000706111, CPI-U · CSV download
Crude oil in the US Strategic Petroleum Reserve, 1977 to 2026
The US Strategic Petroleum Reserve held 282.6 million barrels of crude oil in the week ending 23 October 2026, according to EIA, less than at any month-end since September 1982. This chart follows the stock at the end of every month since October 1977. It rose to a peak of 726.6 million barrels in December 2009 and now holds 39 percent of that level. Two drawdowns account for most of the decline: 309.0 million barrels from July 2020 to June 2023, a period that includes the 180 million barrels released by the Department of Energy in 2022, and 132.9 million barrels between February and October 2026. May and June 2026 are the two largest monthly declines since 1977. The interactive version shows the stock, the monthly changes and every drawdown of 15 million barrels or more.
Sources: U.S. Energy Information Administration · SPR crude oil ending stocks, series MCSSTUS1 and WCSSTUS1 · CSV download
Real US household income by percentile, 1967 to 2025
The Census Bureau put median household income at a record $87,460 for 2025. This study follows the six points of the distribution that the Census tables publish, the 20th, 40th, 60th, 80th and 95th percentiles and the median, from 1967 to 2025 in 2025 dollars. All six were at their highest on record in 2025. In 2025 dollars the 20th percentile went from $23,620 to $35,800, or 52 percent, and the 95th from $149,600 to $354,000, or 137 percent. The six points grew at comparable rates until 1979 and apart since: the ratio of the 95th to the 20th percentile went from 6.33 in 1967 to 9.89 in 2025. Two panels show growth and level side by side, with both Census estimates for the 2013 and 2017 method changes.
Sources: U.S. Census Bureau, CPS ASEC, Historical Income Tables H-1 and H-5 · CSV download
Gallons of gasoline one hour of US work buys, 1979 to 2026
Regular gasoline averaged $4.37 a gallon in the United States in the second quarter of 2026, more than in the summer of 2008. This study divides the BLS pump price by the median weekly pay of full-time workers, quarter by quarter since 1979, and asks how many gallons one hour of that pay affords. In Q2 2026 the answer is 7.2 gallons, below the 47-year median of 8.6, so gasoline is expensive by the standard of the series, and well above the 4.7 gallons of Q3 2008, when the price was lower in dollars: between the two quarters the pump price rose 13 percent and median weekly pay 74 percent. The four long episodes of expensive gasoline in work-time terms bottomed at 5.0, 4.7, 4.9 and 5.6 gallons; Q2 2026 is the first quarter below the median since 2024. The pay series is a median, not an average.
Sources: BLS Current Population Survey, median usual weekly earnings (LEU0252881500Q) · BLS CPI average prices (APU000074714) · via FRED · CSV download
Minutes of work to afford a Big Mac: 16 OECD economies, 2026
The Big Mac Index was built to compare currencies. This study holds the burger constant and asks how long the average worker has to work to afford one, on take-home pay, across 16 non-euro OECD economies. At the net average wage the answer runs from 7.3 minutes in Switzerland to 62.6 minutes in Mexico, a spread of more than eight times. The dollar price explains almost none of it: Switzerland has the most expensive Big Mac in the world at $9.08 and earns it fastest, Mexico’s $6.17 burger takes more than eight times longer, and Japan’s $3.03 burger, the cheapest of the set, still costs 11.6 minutes against 10.1 in the United States. Net hourly pay, hours worked and tax do the explaining, not the price tag.
Sources: The Economist Big Mac Index (Jan 2026) · OECD average wages, hours worked, Taxing Wages (2025) · CSV download
US customs duties collected and refunded, March 2015 to July 2026
For the decade to March 2025, the US Treasury refunded an average of $0.385bn of customs duties a month, and never more than $1.09bn in any single month. In June 2026 it refunded $49.18bn. Net customs receipts, the line that enters the federal budget, went negative for three consecutive months: minus $0.04bn in May, minus $25.56bn in June, minus $8.55bn in July. Collections did not stop. Gross duties over those same three months averaged $23.46bn, about four times the pre-2025 norm of $5.77bn. The fiscal 2026 running total peaked at $188.6bn in April and stood at $154.5bn in July, having given back $34.1bn. One primary source, 137 months, three columns.
Sources: US Treasury, Monthly Treasury Statement, Table 4 · Penn Wharton Budget Model · CSV download
How fast US prices moved relative to inflation, 2000-2019 vs 2019-2026
Between July 2000 and July 2019, tuition, other school fees and childcare ran 59.1% ahead of overall US consumer prices. Since July 2019 the same category has fallen 7.6% behind them, even though its dollar price still rose 20.3%. Prescription drugs moved further, from 23.6% ahead to 19.4% behind, the largest swing in the set. Of seventeen CPI spending categories compared across the two eras, only those two crossed from above the inflation line to below it: eight were above it in the first period, seven in the second. Hospital and related services slowed sharply but stayed above the line, and motor vehicle insurance did not slow at all. The page sets out both readings, a change in pricing regime and the arithmetic of a faster denominator, without settling between them.
Sources: US Bureau of Labor Statistics, CPI-U, not seasonally adjusted · 17 series, July 2000 / 2019 / 2026 anchors · CSV download
How Americans pay cash, by age and purchase size (October 2025)
The age gap in American cash use is almost entirely a small-purchase gap. Below $10, people aged 65 and older settle 52.2% of their in-person payments in cash, against 19.7% for people aged 18 to 34, a difference of 32.5 percentage points. Above $100 the two groups sit at 13.5% and 12.2%, a difference of 1.3 points, and no age group pays cash more than 15% of the time above $50. Cash accounts for 21.4% of in-person payments and 13.9% of all consumer payments. Every figure is recomputed from the Federal Reserve transaction-level microdata, with confidence intervals and cell counts published alongside.
Sources: Federal Reserve Diary of Consumer Payment Choice 2025 public microdata · 13,832 in-person payments, 4,053 respondents
U.S. electricity prices adjusted for inflation, state by state, 2019-2026
In the twelve months to May 2026, the residential price of electricity rose faster than inflation in 43 of the 50 states and the District of Columbia. In the year to May 2022 it did so in 11 of them. The national increase is modest, at 6.8 percent in real terms across seven years, from 16.7 to 17.8 cents per kilowatt-hour in 2026 dollars, but it conceals a spread of 54 percentage points between the District of Columbia at 41.4 percent and Nevada at minus 12.9 percent. Against the long series, the real price remains far below its 1960s and early 1980s levels, and what has ended is the four-decade decline rather than any record being set.
Sources: U.S. Energy Information Administration, Form EIA-861M · BLS Consumer Price Index · CSV download
How U.S. jobs data changes after revisions, January 2024 to June 2026
The monthly U.S. payroll figure is published once, revised twice as late survey responses arrive, then revised again each February against employer tax records. This study reconstructs the first published estimate for every month from January 2024 to June 2026 and compares it with the figure that stands after the release of August 7, 2026. Twenty six of the thirty months were revised down and four were revised up. Taken together they were first reported as 4.36 million jobs added and now stand at 2.02 million. All twelve months of 2025 moved down, taking the year from a reported 1.21 million to 116,000. Measured at the same stage of the revision cycle, the 2015 to 2019 baseline revised slightly upward on average.
Sources: U.S. Bureau of Labor Statistics (PAYEMS) · ALFRED vintages, Federal Reserve Bank of St. Louis · CSV download
Adults 65+ set to outnumber U.S. children around 2029 (1950–2060)
For most of U.S. history, children have far outnumbered older adults. On the Census Bureau’s 2023 projections (middle series), that reverses around 2029, when adults 65 and older are set to outnumber children under 18 for the first time. The older population has risen from 12.3 million in 1950 to 61.2 million in 2024, about five-fold, while the number of children has held near 73 million. The gap narrowed from about 17 million in 2020 to about 12 million in 2024. The crossover has already happened in eleven states and nearly half of U.S. counties; the earlier 2017 projection had placed it at 2034. Interactive chart with three views, plus the full Census dataset to download.
Sources: U.S. Census Bureau · decennial census, Vintage 2024 estimates, 2023 projections · CSV download
CHART OF THE WEEK · 28 JULY 2026
Where U.S. electricity demand grew, and where household bills rose
US residential electricity is up 39.6% since 2020 in current dollars and 8.9% after inflation. Across the 51 states and D.C., the places where commercial demand grew fastest are not the places where household prices rose fastest.
The U.S. cattle herd hits a 75-year low: the fewest cattle per person on record (1920–2026)
The U.S. cattle herd fell to 86.2 million head on January 1, 2026: the smallest since 1951, when the country had roughly 190 million fewer people. Per resident that is about 0.25 head of cattle, one for every four Americans, the fewest on record and down from ~0.61 at the 1975 peak of 132 million. Over the same period retail ground beef reached a record $6.90/lb in nominal terms, about 78% above its January 2020 level. The study places the 2026 inventory in its long historical context, tracks it per capita, documents the parallel move in beef prices, and lays out the competing cyclical and structural readings without resolving them.
Sources: USDA NASS · U.S. Census/FRED · BLS · CSV download
The U.S. federal deficit at each business-cycle low in unemployment
In 1969 and 2000 the U.S. ran a budget surplus at full employment; in 2023, at similar low unemployment, a deficit near 6% of GDP.
U.S. Consumer Prices vs the Fed’s 2% Target, Since 2020
Through 2024 and 2025, U.S. headline inflation cooled back toward the Fed’s 2% goal, a statement about the rate prices rise. This study separates the rate from the level. Using the PCE price index, the measure the Fed targets, it tracks the cumulative level of consumer prices against the path a steady 2% rate would have produced since January 2020. By May 2026 the level sits about 11% above that path: prices are up 25.9% versus the 13.4% a 2% path implies. The gap is not a base artifact, holding from 2012 (+5.5%), 2015 (+8.7%) and 2018 (+10.2%), and core PCE shows the same (+24.5%, 9.8% above). Headline inflation has stayed above 2% for 63 consecutive months.
Sources: BEA PCE price index (PCEPI), via FRED · Cumulative level vs a continuous 2% path · CSV download
US Wealth by Generation, 1989–2026: Boomers Hold 51.6%, Under-40 Households 6.6%
Federal Reserve data shows U.S. household wealth heavily concentrated in older cohorts: Baby Boomers held 51.6% of total net worth in 2026 Q1, Gen X 26.1%, the Silent generation and earlier 11.3%, and Millennials, the Fed's open-ended category that includes Gen Z, 11.0%. Read across a fixed age band instead of by birth cohort, households under 40 held 6.6% of national wealth, down from 12.0% in 1989. The cohort figures are partly the wealth lifecycle; the age-controlled figure is not, since the bracket is held constant. Full quarterly series, the under-40 age-band view, dollar levels, two readings (lifecycle vs structural), and counter-arguments.
Sources: Federal Reserve Distributional Financial Accounts (net worth) · 147 quarterly observations 1989–2026 · Interactive chart & CSV download
US Age at First Marriage, 1890–2025: The Mid-1950s, Not Today, Were the Low
The US median age at first marriage reached 30.8 years for men in 2025, a record in a series that begins in 1890; women stood at 28.4, just below their 2021–2024 peak of 28.6. The striking feature of the record is its U-shape: ages were higher in the 1890s (men 26.1, women 22.0), fell to an all-time low in the mid-1950s (men 22.5, women 20.1), and have climbed since. The post-war “marry young” decade, not the present, is the historical exception: Americans even married later in the 1890s than in the 1950s. The men–women gap narrowed from 4.1 years in 1890 to 2.4 in 2025. A similar shift appears in the mean age at first birth (21.4 in 1970 to 27.5 in 2023). Full series, two readings (economic vs cultural), and counter-arguments.
Sources: U.S. Census Bureau, Table MS-2 (median age at first marriage, by sex) · 85 observations 1890–2025 (decennial to 1940, annual from 1947) · CDC/NCHS (mean age at first birth) · CSV download
US Metro Home Prices: 11 of 20 Largest Cities Now Cheaper Than a Year Ago
Data from the S&P CoreLogic Case-Shiller index through February 2026 reveals a divergent landscape in the 20 largest U.S. metropolitan areas. While the composite 20-city average sits at +0.92% year-over-year, 11 individual metros are now cheaper than a year ago, led by Denver (−2.18%) and Tampa (−2.07%). Conversely, Midwest and Northeast metros like Chicago (+5.03%) and New York (+4.75%) continue to appreciate. Despite the recent 1-year divergence, the 5-year cumulative view shows every tracked metro remains substantially above its 2021 levels, with cumulative gains ranging from 19.4% (San Francisco) to 59.8% (Miami).
Sources: S&P CoreLogic Case-Shiller Home Price Index via FRED (NSA) · 20 metros · Interactive chart
US Social Security Trust Fund, 1990–2034: Full Benefits Through 2034, Then 83%
On the 2026 Trustees Report projections, Social Security’s combined trust fund can pay full scheduled benefits through 2034. After that, continuing payroll taxes still cover 83% of benefits under current law, a 17% shortfall, not zero. The reserve, which peaked at 3.6 years of benefits in 2008, held about 1.5 years at the start of 2026 and is drawn down across the three official Trustees scenarios (depletion between 2032 and 2048, best estimate 2034; the stochastic model places 95% of outcomes in 2032–2039). The retirement fund alone (OASI) reaches depletion earlier, in the fourth quarter of 2032, at 78% payable. Full documentation of the reserve trajectory, the “reserve runs dry, benefits don’t stop” distinction, the 75-year actuarial deficit (4.42% of taxable payroll), and counter-arguments.
Sources: Social Security Administration, 2026 OASDI Trustees Report (combined OASI+DI) · Table IV.B5 · Congressional Research Service · 97 data points 1990–2100 · CSV download
Wind and Solar vs Coal, 1990–2025: The 2024 Crossing in the US Electricity Mix
In 2024, wind and solar combined supplied 17.2% of US net electricity generation against coal’s 14.8%: the first full calendar year on record in which renewables out-generated coal. The crossing held in 2025 on preliminary data (18.9% versus 16.3%), with wind+solar also edging past nuclear into second place behind natural gas (40.2%). The milestone marks a shift in the generation mix, not in which source dominates it: over 1990–2025, natural gas absorbed far more of coal’s lost share (+28 points) than wind and solar did (+19 points), and coal’s share rose again in 2025 as gas prices firmed and total demand grew, the crossing held because renewables rose faster, not because coal kept falling. Documentation of the crossing, the full 35-year trajectory, the share-versus-volume distinction, and counter-arguments.
Sources: U.S. Energy Information Administration, Monthly Energy Review (Tables 7.2a & 10.6) · net generation shares, all sectors, 1990–2025 · small-scale solar PV included · CSV download
US Purchasing Power 1985 vs 2025: What One Hour of Median Work Bought
Between 1985 and 2025, US average hourly earnings rose from $8.73 to $31.34: a 3.59× nominal multiple. Six everyday items track that wage growth in markedly different ways. A median single-family home, a year of in-state public college tuition, and a month of rent now require more hours of median work to buy. A gallon of gas, a Big Mac, and a mid-range new family TV require fewer. The Big Mac, having tracked wages almost exactly since the index was launched in 1986, sits at the structural pivot of the dichotomy. A year of public college tuition now costs the wage-time of nearly 40 mid-range TVs; in 1985, it cost the wage-time of just 3.3. Full documentation per item, with primary sources, methodology, hedonic-adjustment considerations, and counter-arguments.
Sources: BLS AHE production & nonsupervisory (FRED AHETPI) · US Census & HUD (existing-home prices, rent of primary residence) · NCES + College Board (in-state public 4-yr tuition + fees) · BLS APU000074714 (gasoline) · The Economist Big Mac Index · Statistical Abstract of the US + 2025 retail (TV) · CSV download
US CPI Re-Acceleration Episodes Since 1948: When Disinflation Didn’t Stick
Since 1948, US headline CPI has fallen below 3% year-over-year and then re-accelerated meaningfully on four occasions: post-war / Korean War (1948–51), first oil shock (1970–74), second oil shock (1976–80), and the China-driven commodities cycle (2000–05). The 2024–2026 episode now in progress meets the same entry conditions: April 2025 trough at 2.33%, with March 2026 printing 3.29%, the highest year-over-year reading since April 2024. Heatmap alignment of all five cycles on their respective troughs, full documentation of mechanisms (war, oil shocks, commodities super-cycle, current US–Iran energy pass-through), and counter-arguments including the post-2011 fracking offset documented by the Yale Budget Lab. Each prior episode ended with a meaningful overshoot relative to its trough and, in three of four cases, a subsequent recession.
Sources: BLS CPI-U via FRED (CPIAUCSL, 267 monthly observations) · WTI crude oil (DCOILWTICO) · Yale Budget Lab · FOMC statements · CSV download
Real Gold Prices, 1971–2026: The January 1980 Monthly Peak Has Been Exceeded
The inflation-adjusted price of gold has decisively exceeded the January 1980 monthly average peak that stood as the all-time real high for forty-five years. In March 2026, the average London PM Fix gold price was $4,856 per troy ounce: approximately 70% above the 1980 monthly peak in March 2026 USD (~$2,860 real). The breach first occurred in February 2025 and has widened since, distinguishing the current episode from the 2011 and 2020 near-misses. Documentation of the breach, post-peak trajectories from prior episodes (1980, 2011), and the structural factors specific to the 2024–2026 advance, including central-bank reserve accumulation above 1,000 tonnes/year in 2022–2024.
Sources: World Bank Pink Sheet (LBMA London PM Fix monthly averages) · BLS CPI-U via FRED · World Gold Council central-bank reserve data · 663 monthly observations 1971–2026 · CSV download
Big Tech AI Capex 2025–2026: $1.1 Trillion in 24 Months vs Apollo, Marshall, and Manhattan Combined
The four largest US tech companies will spend approximately $1.12 trillion on AI capex across calendar 2025 and 2026: roughly 3× the combined inflation-adjusted cost of the Apollo Program, Marshall Plan, and Manhattan Project in 2025 USD. Placement against the small cluster of historic US mega-investments and post-buildout trajectory analysis using the closest private-sector analogue (1996–2000 telecom buildout).
Sources: Company 10-Qs and Q1 2026 earnings calls (Apr 29–30, 2026), Planetary Society, Wikipedia, BLS CPI
The US Dollar’s Worst First Halves (1973–2025)
H1 2025 ranked 4th worst on record for the broad US Dollar Index since the end of Bretton Woods, at −7.6%. The three deeper episodes (1973, 1986, 2003) each coincided with a structural shift in the international monetary or trade order. Statistical placement of H1 2025 against the full distribution of 53 first-half observations and analysis of post-H1 trajectories.
Sources: Federal Reserve Bank of St. Louis (FRED series DTWEXM, DTWEXBGS) · 53 H1 observations · CSV/XLSX download
Editorial standards
Each Macro Watch study follows the same standards applied to the evergreen research on the main hub:
Observatoire macro (version française) →
Frequently asked questions
How is Macro Watch different from the main research hub and the barometer?
Eco3min publishes three complementary layers. The main hub is the permanent reference layer: 97 datasets that update automatically, plus 33 evergreen studies on the foundational macro-financial frameworks (yield curve, inflation, real rates, liquidity, credit spreads). The macro barometer is the continuous framing layer: a synthetic monthly reading of the current regime. Macro Watch is the topical layer: one new study per week, anchored to a current event, with a longer historical perspective. The three layers are designed to work together: Macro Watch studies frequently cite and build on the main hub’s datasets and the barometer’s framing.
Can I republish or cite Macro Watch studies?
Yes for Eco3min text and analysis (CC BY 4.0, with attribution); data remain subject to their producer’s terms, stated on each dataset page. A suggested citation format is provided at the bottom of each study page. Direct quotes are welcome; please link back to the original study URL.
How are study topics chosen?
Each Macro Watch study is anchored to a specific current event: a quarterly earnings cycle, a half-year currency move, a major macro release, a policy shift. Topics are selected based on (1) availability of primary-source data, (2) capacity for non-trivial placement against historical context, and (3) value to the readership of analysts, journalists, and macro researchers. Topical relevance, not viral potential, drives selection.
Are the datasets behind each study downloadable?
Yes. Every Macro Watch study includes a CSV and XLSX download with the exact figures used in the analysis, source identifiers per data point, and methodology notes. The Python or R code is also published when the dataset is non-trivial to reproduce.
Are studies available in French?
Yes. Every Macro Watch study is published bilingually. The French version is accessible via the Observatoire macro hub, or via the 🇫🇷 cross-link on each card above. Both versions share the same methodology, the same data, and the same editorial standard. The English publication typically precedes the French one by 1–2 days.
Cite Macro Watch
If you reference a Macro Watch study in your research, articles, or analyses, please cite as follows:
Eco3min text and analysis: CC BY 4.0, with attribution. Data remain subject to their producer’s terms, stated on each study page. Each individual study page provides its own complete citation.
Last updated — 3 November 2026
Disclaimer – Financial and Legal Information: The analyses, commentary, and content published on eco3min.fr are provided for informational and educational purposes only. They do not constitute investment advice or a solicitation to buy or sell financial instruments. Past performance is not indicative of future results. All investment decisions involve risk and are the sole responsibility of the reader. Data, charts and calculated series are provided as is and may contain errors or lag behind their source: check any figure against the primary source cited before relying on it. Content describing laws, regulations or contracts reflects the law as of its publication date and does not constitute legal or tax advice; applying it to a specific situation requires a qualified legal or tax professional.
Regulated and tax values: the rates, indices, caps and brackets with legal effect cited on this page are set by official decision, change over time and apply according to individual circumstances. Only the official publication in force on the relevant date is authoritative (for example the Journal officiel, INSEE, Banque de France, the IRS or the Social Security Administration). To revise a rent, sign or challenge a contract, or file a return, check the value at the source or with a professional.