Macro / Demographics
Same Age, Different Bill
Demographics set budget pressure; institutions and politics decide the outcome. Germany and Italy aged at the same speed under the same central bank and ended with very different pension bills, debt and interest costs.
By Stratifi Research
November 12, 2026
Method note. Every statistical claim uses inflation data from 2019 or earlier; the 2021–25 episode appears once, in a labelled aside, and nothing rests on it. Bracketed numbers refer to the sources at the end of the piece.
1. Two old countries, one central bank
Take two countries that aged at almost the same speed, gave up their currencies in the same year, and had their interest rates set by the same central bank in Frankfurt. Did they end up with the same budget?
In 1990 Germany had 21.5 people aged 65 and over for every 100 of working age; Italy had 21.8. When the euro started in 1999 the figures were 24 and 26. By 2019 they were 33 and 36. Over the two euro decades both ratios rose by ten points. [1]
Prices barely told them apart. Inflation averaged 1.5% a year in Germany and 1.8% in Italy from 1999 to 2019, and across eleven early euro members the spread between the lowest and highest average was six-tenths of a point. [76] One central bank produced one price level, give or take, whatever the age structure underneath.
Same aging, same central bank, different pension bills and debt
| Germany | Italy | |
|---|---|---|
| Old-age dependency ratio, 1990 / 1999 / 2019 (65+ per 100 aged 15–64) | 21.5 / 23.7 / 33.4 | 21.8 / 26.4 / 36.2 |
| Change in old-age dependency ratio, 1999–2019 | +9.8 | +9.7 |
| Inflation (HICP), average 1999–2007 / 2008–19 | 1.6% / 1.4% | 2.3% / 1.4% |
| Public old-age cash benefits, % of GDP, 1990 / 2000 / 2017 | 9.5 / 10.9 / 10.3 | 11.3 / 13.5 / 15.5 |
| Gross public debt, % of GDP, 1999 → 2019 | 60 → 59 | 113 → 134 |
| Interest paid, % of GDP, 1999 → 2019 | 3.2 → 0.8 | 6.6 → 3.3 |
Two small panels sit beside the table: pension cash and gross debt, 1990–2019, one axis each. Two countries make an illustration, not a test. Other forces: German reunification, the 2010–12 sovereign-debt crisis, pension design, Italy's inherited debt. Sources: UN World Population Prospects 2024 [1]; Eurostat [76]; OECD social expenditure data [8]; IMF historical public finance data [3].
Units: % of GDP. Frequency: Annual, 1990–2019. Formula: Plotted values are the published annual series, unchanged: public old-age and survivors cash benefits as % of GDP (OECD SOCX) and gross public debt as % of GDP (IMF Public Finances in Modern History). Table rows: change in OADR = OADR(2019) − OADR(1999); HICP averages are arithmetic means of annual rates; pension change = SOCX(2017) − SOCX(1990).The age pyramid sets the size of the bill. The rate regime sets what it costs to carry. Politics decides who pays.
The thesis this piece pursues.
That is the whole argument, and Italy’s interest line shows the middle sentence at work. Its debt rose and its interest bill still halved, from 6.6% of GDP to 3.3%; Germany’s fell from 3.2% to 0.8%. [3] The cost of carrying the budget was set by the monetary regime and by what bond buyers believed about the institutions behind it, not by the age of the voters.
2. What our tests can and cannot see
Before any evidence, a confession about tools. We ran a long battery of pre-registered correlation tests on demographics, prices, rates and debt, using only data through 2019. Hardly anything cleared every hurdle. Does that mean aging does nothing to prices?
No, and we checked why. We put thirteen relationships that economists widely document through exactly the same tests: money growth and inflation, several versions of the Phillips curve, oil and inflation. One passed: oil-price changes followed by US headline inflation a quarter later. Then we simulated data with a real link built in. For a single country, with a true correlation of 0.5 and an effective sample of nine independent years, which is typical of demographic series that move slowly, our tests found the link essentially never. Even a true correlation of 0.8 was caught only one time in eight.
So when this piece says “no reliable link,” read it as “not distinguishable from zero in this sample.” It does not mean “no effect.” Slow-moving populations give few independent observations, and our filters were built to avoid false alarms, not to catch faint signals.
That changes what the thesis can rest on. It cannot rest on aging having zero effect on prices; our data can’t show that. It rests on something larger and easier to see: the monetary regime. Across 18 advanced economies since 1871, median decade inflation was 0.3% under the gold standard, 3.9% under Bretton Woods and 2.4% under fiat money after 1971, while median population growth barely moved from one regime to the next. [73] Whatever aging does to prices, it has been small next to that.
3. The view we are arguing with
The demographics-debt-debasement view is the best-known alternative in macro today, and it deserves a fair statement. Raoul Pal is one prominent proponent; his new book, published on November 3, has no public text beyond a chapter excerpt and his own summary page, so we rely on those. [102] [103] [104]
In his summary, the case “begins where most people refuse to start. Demographics.” Shrinking working-age populations mean output “has to be grown through debt,” and debt is then serviced “through currency debasement,” which he equates with expanding global liquidity. [103] Liquidity, in his words, is “the single most dominant macro factor in the history of markets,” and debasement “creates a structural hurdle rate of roughly eight per cent a year.” [103] In the chapter excerpt, debasement “drives up the price of scarce assets” but “does not drive up wages … corporate earnings … commodity prices.” [102]
What holds. The first link is real: aging adds to old-age spending everywhere we look (section 6). The academic consensus that aging lowered equilibrium real interest rates supports the view on rates; Auclert and co-authors put the drag near one percentage point, with “no great demographic reversal.” [35] [29] [30] [31] And Pal has pointed to a repeat of the 1950s and 60s, which “collapsed debt to gdp via financial repression.” [60] That is close to where this piece ends up.
Where we depart.
Debt is financing, not a source of growth. Output grows with hours worked and output per hour. Debt is how a shortfall has been paid for.
Aging does not set the price level; the monetary regime does (sections 2, 4 and 5).
The chain has no voters in it. Germany and Italy took the same demographic input and produced different debt (sections 1, 6 and 7).
The famous chart. The view’s signature picture sets US labor-force participation against federal debt and calls the two “perfect lockstep.” [51] We return to it in section 4.
The hurdle rate depends on the definition.¹
We keep the habit of ordering slow variables before fast ones. We drop the step that turns a demographic constraint into a forecast for prices or for the debt ratio.
¹ The public texts do not define the money measure behind “roughly eight per cent.” [103] On our calculation, US M2 grew 6% a year over 2005–25; 8% is reachable only with the Fed’s net-liquidity measure from a 2005 start. [6] This is a money claim, not an inflation claim.
4. Prices follow the regime
If aging moved the price level, where would it show? Inside each monetary regime, once the shocks every country shared are taken out.
The Jordà–Schularick–Taylor macrohistory data cover 18 advanced economies from 1871 to 2019. Within each regime, decades of faster population growth did not come with reliably higher or lower inflation. [73] Across 126 country-decades since 1951, older countries had lower inflation on the raw numbers, but most of that disappears once each decade’s common level is removed. Aging tracks public debt somewhat more closely, though that link leans on Japan and weakens a lot without it.
Aging tracks debt more than inflation; older countries ran slightly cooler
Panels A and B: 18 advanced economies, decade averages 1951–2019, each variable shown relative to that decade's cross-country average. A: old-age dependency vs inflation. B: old-age dependency vs public debt. Panel C: old-age dependency in 2000 vs average inflation 2000–19, 26 advanced economies (panel A of the pre-2020 cross-section chart only). Sources: Jordà–Schularick–Taylor Macrohistory Database [73]; UN WPP 2024 [1]; World Bank WDI [2].
Units: Inflation relative to decade average, percentage points; Public debt relative to decade average, % of GDP points; Average CPI inflation 2000–19, % a year. Frequency: Country-decade averages (A, B); one value per country (C). Formula: Panels A–B: for each variable v and decade d, v relative = v − (average of v across the 18 economies in decade d); decade averages of annual data, 1951–2019 (n = 126 country-decades). Panel C: x = old-age dependency ratio in 2000; y = arithmetic mean of annual WDI CPI inflation 2000–19.Panel C is the one result that leans toward the deflation camp. Countries that were older in 2000 had somewhat lower inflation over the next two decades, and that holds with or without Japan. It is the opposite of Goodhart and Pradhan’s prediction that an aging world brings higher inflation. [25] It is also one cross-section, and other windows do not repeat it.
What about within countries? When a country aged faster than its own norm, its inflation over the following years was not distinguishable from its usual path, once the year’s common shock was removed. Look at the 2010s: the median rich country’s aging ran four to seven standard deviations above its history, and its inflation stayed below trend. Before any pandemic, the fastest aging on record came with low inflation.
Japan, 1992–2019. Japan’s working-age share peaked at 70% in 1992 and fell to 59% by 2019, while inflation went from 2.7% a year in the 1980s to slightly below zero in 1995–2012, then 0.8% in 2013–19 after a change of monetary regime, with aging accelerating. [1] [3] The deflation years also brought a near-40% fall in house prices, a banking crisis in 1997 and a strong yen, so Japan argues for the regime more than for the age pyramid. [73]
The famous chart. In levels, participation and federal debt rose together before 2000 (correlation +0.79) and moved in opposite directions after (−0.97). Correlations of trending levels like these often mean nothing. In year-on-year changes, the two are mildly negatively related in the same quarter. Participation does not lead debt at any horizon, the relationship reversed sign in 2015–19, and there is no reliable break at 2000. Before 2020, the data do not show participation driving debt.
Does fiscal strain bring inflation? Across 26 rich countries from 1980 to 2009, a rising interest bill was followed by somewhat higher inflation the next year, but the link came almost entirely from Israel in the 1980s; without it, there was essentially nothing. In the United States, the 13 years with inflation of 5% or more between 1969 and 1990 followed periods when federal debt held by the public had fallen by 0.85 points of GDP over three years. The other 41 years followed rises of 3.08 points. [12] [77] Inflation has not been a mechanical release valve for fiscal strain.
Aside: 2021–25 (not used as evidence). In our correlation scan, the year’s common shock explained 62% of cross-country differences in 2021–25 inflation, with no aging gradient. Nothing here rests on that episode; supply-chain and demand shocks dominated it.
5. Three centuries of the same pattern
If the monetary regime sets the price level and demography mostly works on the budget, history should offer cases where population changed dramatically and prices did not, while the budget moved a lot. Does it?
Britain under gold, 1821–1913. The UK population grew 2.7×, from 15.6 million to 42.6 million. The consumer price index stood at 1.26 in 1821 and 1.24 in 1913: compounded inflation was about zero. [72] The public debt ratio fell from a peak of 194% of GDP in 1822 to 28% in 1913, a decline of 85%, with average net lending of −0.1% of GDP. This was no century of big surpluses. [72] A growing population and economy outgrew the debt, and the gold standard held the price level.
England's population, prices and public debt over nine centuries
Three stacked panels, each on its own axis: England population (thousands, log scale), UK/England consumer prices (2015 = 100, log scale), UK public debt as % of GDP from 1700. Dotted lines mark 1348, 1520, 1650, 1815, 1913 and 1971. Population and productivity growth cannot be separated. Source: Bank of England, A millennium of macroeconomic data v3.1 [72].
Units: Thousands (log scale); Index, 2015 = 100 (log scale); % of GDP. Frequency: Annual. Formula: Values as published in the Bank of England millennium workbook v3.1; no transformation. Log scales on the population and price panels are labelled. Dotted lines at 1348, 1520, 1650, 1815, 1913 and 1971.After the Black Death. England’s population fell by almost half within three years of 1348, and by 1450 stood at 40% of its pre-plague level. Real wages roughly doubled: 2.2× on the Bank of England series, 1.8× for London labourers on Allen’s. Prices rose in the 1370s, then settled; by 1440–59 the consumer price level was only 5% above its pre-plague level. [72] [74] A demographic catastrophe reset relative prices, with labour dear and land cheap, but not the price level.
Between the wars. Population growth in the 18 economies barely slowed, from a median of 0.9% a year in 1871–1913 to 0.8% in 1921–38. Median prices fell, by 0.8% a year over 1921–38 and 2% a year over 1925–33. Britain’s population grew 0.1% a year while its prices fell 2.2% a year; France’s grew 0.4% while its prices rose 4.2%; America’s grew 1.1% while its prices fell 1.8%. [73] Gold, war debts and the banking collapse decided which way prices went. The median real long rate over 1925–33 was 7%, a punishing cost for any government in debt.
East Asia and Latin America, 1970–2010. Here is the experiment in reverse: similar demographic windfalls, wildly different price outcomes.
Similar demographic dividends, inflation two orders of magnitude apart
| Change in working-age share, 1970–2010 (points) | Median annual inflation, 1971–95 | Peak gross public debt, 1980–2005 (% of GDP) | |
|---|---|---|---|
| Korea | +18.1 | 7.2% | 24 |
| Thailand | +18.9 | 5.3% | 58 |
| Malaysia | +14.4 | 3.7% | 74 |
| Brazil | +14.2 | 228% | 76 |
| Mexico | +14.6 | 24% | 53 |
| Colombia | +15.6 | 23% | 48 |
Group medians of annual inflation, 1971–95: 5% in East Asia, 24% in Latin America. Other forces: monetary and fiscal institutions, indexation, the 1980s debt crisis, commodity dependence. Sources: UN WPP 2024 [1]; World Bank WDI [2]; IMF historical public finance data [3].
Formula: Change in working-age share = share aged 15–64 in 2010 − share in 1970 (UN WPP 2024); median inflation = median of annual WDI CPI inflation 1971–95; peak debt = maximum of IMF gross debt/GDP 1980–2005.What separated them was not the number of workers per retiree. It was central banks, indexation, fiscal rules and the 1980s debt crisis. East Asia saved more, too: gross saving averaged 35% of national income in Korea and 43% in Singapore over 1980–2005, against 18% in Brazil and 22% in Mexico. [2] The dividend was an opportunity. Institutions decided whether it was banked.
The postwar liquidation, 1946–80. Across the 18 economies, median public debt fell from 61% of GDP in 1946 to 23% in 1974; US debt from 119% to 33%; Britain’s from 270% to 52%. [73] Young, growing populations helped by enlarging the denominator; US population grew 1.5% a year over 1946–70. But the mix of inflation and real rates was chosen. Real short rates were negative in nearly half of all years in the median country, Britain’s averaged −1.3% over 1946–80, and the Bretton Woods peg held almost everywhere. [73] [34] Reinhart and Sbrancia call it the liquidation of government debt. [34] It was a budget outcome, delivered through institutions.
In every one of these episodes, gold, a peg, a central bank or its absence set the price level, and the number and ages of people put pressure on the budget.
6. Budget pressure: the accounting
Where does aging show up most reliably? In what governments spend. That is arithmetic, not a statistical finding, and it is close to certain.
In the United States, Social Security, Medicare and Medicaid took 20% of federal outlays in fiscal 1970, 29% in 1981, 44% in 2007 and 50% in 2019. Inflation over the same span ran the other way, from 6.9% a year in 1966–81 to 1.8% in 2008–19. [12] [77] The age-linked programs went from a fifth to half of the budget through a great inflation and a long disinflation alike.
Old-age programs went from a fifth to half of federal outlays by 2019
Share of total federal outlays, fiscal 1962–2025: Social Security plus Medicare (gross) plus Medicaid, net interest, and defense. The text uses years through fiscal 2019; pandemic-era outlays are shown but not relied on. Medicaid is not purely old-age; Medicare is gross of offsetting receipts. Source: CBO historical budget data, February 2026 [12].
Units: % of federal outlays. Frequency: Annual, fiscal years 1962–2025. Formula: Share = program outlays ÷ total federal outlays × 100, from CBO historical budget data (packet clean file). The combined age-program series starts in fiscal 1967, the first year all three programs are in the data.Run the same accounting on the euro pair. Between 1990 and 2017, Germany’s public pension cash rose by under one point of GDP and Italy’s by four. [8] Both countries aged by similar amounts, and spending did not respond in proportion. The pressure was the same; the promises were not. Japan shows the pressure in a third setting: public old-age cash benefits doubled from 4.6% of GDP in 1990 to 9.3% in 2015, gross debt rose 3.8× from 1991 to 2019, and the interest bill halved, from 3.4% of GDP to 1.7%. [8] [3]
Why not say aging “sets the budget” outright? Because the cross-country record doesn’t support that stronger form. Across 18 countries the aging–debt link is modest, leans on Japan and is not distinguishable from zero year to year within countries. Germany and Italy show why: the same pressure produced a flat debt ratio in one and a rising one in the other. The pressure is measurable and predictable. The debt outcome is decided elsewhere.
Where is the pressure heading? On CBO’s February 2026 long-term outlook, Social Security rises from 5.2% of GDP in 2026 to 6.0% in 2056, Medicare (gross) from 3.3% to 5.5%, net interest from 3.3% to 6.9%, and debt held by the public from 101% to 175%. [13] Interest grows faster than the two big age programs combined: +3.7 points of GDP against +2.9. One year of revisions moved projected 2055 debt from 156% of GDP to 172%. [13] [14] As for the retirement trust fund, the 2025 Trustees report put its depletion in 2033. The 2026 report moved it to late 2032, and CBO projects fiscal 2032. [81] [82] [83] These are projections and accounting, not inflation evidence.
What does a bond market do with predictable pressure? It prices the institution that will respond to it. Germany, Italy and Japan all saw interest burdens fall while aging rose, because their monetary regimes pushed rates down. [3] When Italy’s institutions were in doubt, the price moved fast: in November 2011 its ten-year yield averaged 7.06%, 5.18 points over Germany’s. [90]
7. Who decides: voters, wealth and the menu
Who shows up to decide? In the 2024 US presidential election, 72% of people aged 65 and over reported voting, against 44% of 18–24 year-olds, a gap of 28 points. In 1972 the gap was 14 points; in the 2022 midterm it was 39. [17] The figures are self-reported, which overstates turnout. Baby Boomers hold 52% of household net worth; Millennials hold 11%, against the 20% Boomers held at a similar age, with no adjustment for cohort size. [19]
Turnout gap between Americans over 65 and those aged 18–24, 1964–2024
Reported turnout of people 65 and over minus that of 18–24 year-olds, percentage points, November elections 1964–2024; one line. Self-reported; overstates validated turnout; the base includes non-citizens. Source: Census Current Population Survey, Table A-1 [17].
Units: Percentage points. Frequency: Biennial, November elections 1964–2024. Formula: Turnout gap: reported turnout of people 65 and over − reported turnout of 18–24 year-olds, each as a percentage of the age group’s total population (CPS Table A-1). Before 1972 the youngest group is 21–24 in most states.Some of the generational story is overstated. Under-35 homeownership in 2025 was where it was in 1994, at 37%. [18] And young voters are not a fixed bloc: among 18–49 year-olds, the margin went from 17 points for Biden in 2020 to 7 for Harris in 2024. [24]
So what is on the menu for an old, asset-rich, high-turnout electorate facing a rising bill?
Open inflation is the obvious candidate, and the evidence argues against it. Tim Vlandas finds that “Countries with a larger share of elderly exhibit lower inflation because older people are both more inflation averse and politically powerful.” [36] Mojon and Ragot find older workforces dampen wage inflation. [28] The cross-section in Exhibit 2, where older countries ran slightly cooler, fits Vlandas better than it fits Goodhart and Pradhan.
Reform happens, and it is worth being exact about what it bought. Italy’s Fornero law of December 2011 moved every worker onto contribution-based accrual for service from January 2012, set a minimum old-age pension age of 67 for those qualifying from 2021 with a link to life expectancy, and limited indexation in 2012–13 to pensions up to three times the minimum. [89] It passed the month after Italy’s ten-year spread over Germany averaged 518 basis points. [90] Italy’s pension spending kept rising to a peak of 16.4% of GDP in 2013 and then eased to 15.7% by 2018: the line bent rather than flattened. [91] Germany legislated a rise in the retirement age to 67 in April 2007, phased in by birth cohort from 2012; 67 first applies fully in 2031. Its pension spending has been broadly stable as a share of GDP since 2000. [94] [93] [92] Sweden’s parliament adopted a notional-account system in June 1998, in force from 1999, with an automatic balancing mechanism. That mechanism was first triggered in 2010, trimming the indexation of pensions. [95] [96] [97] Each of these reforms came under pressure from markets or arithmetic, and each was slow. The bill was trimmed over decades. It was not cancelled.
Financial repression is our working hypothesis for the default modern outcome. It means holding interest rates below inflation long enough that debt shrinks relative to income. That was the postwar liquidation (section 5). In Japan in 2013–19 the ten-year yield averaged 0.2% and the ex-post real yield −0.6%. [73] Indexation explains why repression is politically easy. US Social Security rises each year with the consumer price index for wage earners, measured from third quarter to third quarter. [84] [85] German pensions follow wages, with a sustainability factor and a floor on the replacement rate until 2031. [86] Italian pensions follow consumer prices in tiers: in full up to four times the minimum pension, at 90% between four and five times, and at 75% above that. [87] [88] Indexed benefits protect the old against inflation. Conventional bonds pay a fixed nominal coupon and get no such protection. Repression therefore taxes bondholders and savers far more than benefit recipients, and it does so quietly, through yields rather than a vote. We hold that as a hypothesis, not a tested result.
Our scenario hypotheses for the US fiscal choice to 2035 (these probabilities are judgments, not estimates from data):
US fiscal choice to 2035: four branches and their markers
| Branch | What happens | Probability | Marker |
|---|---|---|---|
| Quiet repression | Real yields held near or below zero; debt ratio stabilizes slowly | 40% | Real 10-year yields negative for 2+ years while CPI is 2–4% |
| Inflation bursts | Episodic inflation above 5%, tolerated after shocks | 25% | CPI ≥ 5% in a year without an energy or supply shock |
| Fiscal adjustment | Benefit growth trimmed or taxes raised near the trust-fund date | 20% | Legislation changing Social Security benefits or payroll taxes before 2032 |
| Growth and labor supply | Productivity and immigration enlarge the base | 15% | Working-age population and labor-share growth beat CBO |
These probabilities are judgments, not estimates from a model, and they are recorded on the Record page so they can be scored.
Branch weights sum to 100%. Markers are the observations that would move weight between branches.8. Migration and the geopolitics of age
Migration is the one demographic variable policy can change within a decade. On the UN’s projections, the US has 381 million people in 2050 with migration and 339 million without, a 12% difference. The working-age share barely changes (61.0% against 59.9%), but the old-age dependency ratio does (38 against 43), so migration absorbs a third of the projected rise. [1] For Japan (73 against 78) and Korea (76 against 79), plausible migration hardly moves the ratio. [1] Does it pay its way? The National Academies’ 2016 study found that first-generation immigrants are “more costly to governments, mainly at the state and local levels,” while “immigrants’ children—the second generation—are among the strongest economic and fiscal contributors in the population.” Over 75 years, it found fiscal impacts “generally positive at the federal level and negative at the state and local levels.” [101] The lever that makes the totals work quickly is a political choice, not a demographic fact.
Workers. Between 2024 and 2050, the working-age population shrinks by 35% in Korea, 26% in Japan, 24% in China and 17% in Germany, and grows by 4% in the US, 15% in India and 78% in Nigeria. [1]
China became an aged society at 30% of US income; Japan at 80%
First year the 65-and-over share reached 14%, with GDP per head at purchasing-power parity that year as a percentage of the US. The income series starts in 1990, so earlier crossers are absent. Sources: UN WPP 2024 [1]; World Bank WDI, October 2026 update [2].
Units: % of US GDP per head, same year. Frequency: One value per country. Formula: Ratio = GDP per head at PPP (constant 2021 dollars, WDI) ÷ US GDP per head in the same year × 100, in the first year the 65+ share reached 14% (WPP 2024). The US bar is 100 by construction (its own crossing year, 2014).Michael Pettis’s 2021 headline was “Debt, not demographics, will determine the future of China’s economy,” and he concluded that “China’s deteriorating demographics worsen an existing problem, but do not fundamentally change anything.” [39] China’s working-age population was roughly flat from 2021 to 2025, so the recent slowdown is not labor-force arithmetic; the steep part of the decline comes after 2030. [1] [39] China’s demographic pressure is fixed. What it costs depends on how China handles its debt and its growth model. Adam Tooze, relaying Bert Hofman, noted that if China raised its retirement age to 65 over twenty years, “the labour supply issue disappears.” [46]
Savers. Japan, Germany and Korea are old and still save more than they invest: gross saving against investment is 32% against 28% in Japan, 26% against 22% in Germany, and 36% against 29% in Korea. The US runs the reverse, 17% against 22%. [2] National saving tracks institutions and distribution more than the age pyramid.
Head counts. Capital substitutes for cohorts: on our calculation from SIPRI and UN data, military spending in 2025 came to $33,000 per man aged 18–29 in the United States and $3,000 in China. [21] [1]
9. The strongest objection, and what would change our mind
The strongest counter-case is the sixteenth century. Between the 1490s and the 1640s, England’s population more than doubled (2.3× from 1520 to 1650) and sterling prices rose more than fivefold. Measured in silver, prices still rose 3.2× in London and Valencia and 4× in Antwerp, and real wages roughly halved. [72] [74] The sharpest point is timing. Between the late 1490s and the early 1520s, English prices rose 23% while population grew 10%. As late as 1531–35, Seville was receiving only 5,000 kilograms of American silver a year; like the other Seville silver figures here, that rests on a single secondary source. [75] On that reading, prices began rising with population, before the silver arrived, which cuts directly against our claim. Henry VIII’s Great Debasement, which cut the silver in the penny by 83% and saw prices rise 141% from 1541 to 1556, was exactly the kind of monetary-regime change our mechanism predicts; yet the sharpest single-year rises came in 1556–57, after it ended. [98] [99] Spain’s long-term borrowing in juros rose from 19 million ducats in 1560 to 68 million in 1598. [100] Debasement, money velocity, urbanization and war finance all compete as explanations. Whether population led prices before the silver came is an open question, and we leave it open.
The modern version of the objection is the baby boom. Goodhart and Pradhan argue that more workers per retiree held inflation down and that the reversal will push it up; [25] Juselius and Takáts find age-structure effects on inflation in a panel. [26] [27] In the US time series, the boom fits the inflation hump uncomfortably well: changes in the 20–34 share line up with inflation five years later.
The US baby boom lines up with 1970s inflation at home, the reverse abroad
Panel A: US change in the population share aged 20–34, shifted forward five years, and US consumer price inflation, both in standard-deviation units on one axis, 1956–2019; shaded 2010–19. Panel C: change in the 20–34 share 1960–80 vs average 1970s inflation, 18 advanced economies (the youth-bulge panel of the pre-2020 cross-section chart only). Sources: UN WPP 2024 [1]; FRED [77]; Jordà–Schularick–Taylor [73].
Units: Standard deviations from the 1956–2009 average; Average inflation 1970–79, % a year. Frequency: Annual 1956–2019 (A); one value per country (C). Formula: Panel A: z = (v − average of v over 1956–2009) ÷ standard deviation of v over 1956–2009; x = annual change in the US 20–34 population share (WPP 2024) moved forward five years; y = 100 × change in the log of annual-average CPI-U. Panel C: x = 20–34 share in 1980 − share in 1960; y = arithmetic mean of annual CPI inflation 1970–79 (JST).But one boom and one hump amount to roughly ten independent observations, and after 2010 the two lines parted. Across countries the episode points the other way. The US had the largest youth bulge of the 18 economies, with its 20–34 share up more than seven points between 1960 and 1980, yet its 1970s inflation was 6.9%, below the median of 9.0%. Italy’s share fell and its inflation ran at 12.3%. [73] [1] The 1970s also brought the end of dollar–gold convertibility in August 1971, oil rising from $3.60 a barrel in 1972 to $37 in 1981, and the Volcker measures of October 1979. [80] [77] [79] [78] The saver phase undercuts the mechanism further: the US 45–64 share rose from 19% in the 1980s to 26% in the 2010s, while the personal saving rate fell from 10% to 6%.
A modest tilt remains within regimes: in the 2000s, and in the 2000–19 cross-section, older rich countries had lower inflation. So the honest form of the thesis is narrow. Within a regime, age structure tilts prices at most modestly, and where the tilt shows, it has pointed toward lower inflation. The regime sets the level.
What would change our mind (thresholds are our own markers):
Institutions stop mattering for the budget. Among euro members over 2019–35, the change in aging explains most of the cross-country change in old-age spending and debt (a correlation of 0.7 or more), so that gaps like Germany’s and Italy’s close without reform.
Pressure doesn’t arrive. An aging democracy’s old-age spending falls as a share of GDP over ten years without benefit reform while its dependency ratio rises (test by 2035).
Demography sets prices within a regime. A within-country aging gradient in inflation, after common shocks are removed, shows up for several consecutive years over 2026–30 and survives in a sample large enough to detect it.
Japan returns to persistent deflation by 2030 while its dependency ratio rises, with no credit or policy explanation.
What to watch (markers, not trades):
Italy and Germany in the next OECD pension-spending release. Does Italy’s bent line stay bent?
The next CBO long-term outlook and the next Trustees report: how far they move debt and the trust-fund date. [13] [82]
The approach to 2032, when politics has to choose among benefit cuts, taxes and borrowing. [82] [83]
Real long yields in Japan and the US. This is the repression marker; watch the sign, not a forecast.
China after 2030, when the working-age decline accelerates. [1] [39]
The 2026 midterm turnout gap, when the Census publishes it. [17]
Falsifier, horizon, confidence. The thesis is wrong if similar aging produces similar debt and spending outcomes regardless of institutions (marker 1), or if aging stops adding to old-age spending (marker 2). The horizon is 2035 for the institutional test; the pressure claim runs to 2056 on CBO’s projection. Confidence:
High that aging raises old-age spending pressure. This is accounting.
Medium-high that the monetary regime, not age structure, sets the price level. The evidence is regime medians since 1871 and three centuries of overlays.
Medium that institutions decide the debt outcome. The cases are Germany vs Italy and East Asia vs Latin America, a small number.
Low-to-medium for repression as the default modern outcome. This is our hypothesis.
So the next time someone calls a country’s demographics inflationary or deflationary, ask two questions. Which institution sets its price level? And which one will decide who pays the bill?
Sources
1
UN World Population Prospects 2024.
2
World Bank, World Development Indicators.
3
IMF DataMapper and Public Finances in Modern History.
6
FRED series (M2SL, WALCL, WTREGEN, RRPONTSYD).
8
OECD SOCX, public old-age cash benefits.
12
CBO, Historical Budget Data, February 2026.
13
CBO, The Long-Term Budget Outlook, February 2026.
14
CBO, The Long-Term Budget Outlook, March 2025.
17
US Census Bureau, CPS voting and registration tables.
18
US Census Bureau, Housing Vacancies and Homeownership, Table 19.
19
Federal Reserve, Distributional Financial Accounts.
21
SIPRI Military Expenditure Database.
24
Pew Research Center, validated voters 2024.
25
Goodhart and Pradhan, BIS Working Paper 656.
26
Juselius and Takáts, BIS Working Paper 485.
27
Juselius and Takáts, BIS Working Paper 722.
28
Mojon and Ragot, BIS Working Paper 776.
29
Carvalho, Ferrero and Nechio.
30
Gagnon, Johannsen and López-Salido.
31
Rachel and Smith.
34
Reinhart and Sbrancia, IMF Working Paper 15/7.
35
Auclert, Malmberg, Martenet and Rognlie.
36
Vlandas, “Grey power and the economy.”
39
Pettis, Fortune, 2021-06-17.
46
Tooze, 2022-05-01.
51
Pal, X post, “perfect lockstep” (2026-07-03 UTC).
60
Pal, X post, 1950s–60s financial repression (2025-12-04 UTC).
72
Bank of England, A millennium of macroeconomic data for the UK, v3.1.
73
Jordà–Schularick–Taylor Macrohistory Database, R6.
74
Allen, wages and prices in European cities, 1260–1913.
75
Munro, EH.net review of Hamilton (1934). Single secondary source.
76
Eurostat HICP.
77
FRED long US series (CPI-U, personal saving rate, 10-year yield, WTI).
78
Federal Reserve History, “The Great Inflation.”
79
Federal Reserve History, Volcker’s anti-inflation measures.
80
Federal Reserve History, end of gold convertibility.
81
2025 Annual Report of the OASDI Trustees.
82
2026 Annual Report of the OASDI Trustees.
83
CBO testimony, Social Security’s Finances, 2026-03-25.
84
42 U.S.C. 415(i).
85
20 CFR 404.272.
86
Sozialgesetzbuch VI, sections 68, 68a, 255e.
87
Decreto MEF/MLPS 19 Nov 2025, perequazione automatica 2026.
88
INPS Circolare n.153, 19-12-2025.
89
DL 201/2011, art. 24, converted by L. 214/2011.
90
FRED/OECD Italian and German 10-year yields, monthly; ECB cross-check.
91
Eurostat ESSPROS pension expenditure, % of GDP.
92
OECD Pensions at a Glance 2025, Table 8.2.
93
Sozialgesetzbuch VI, section 235.
94
RV-Altersgrenzenanpassungsgesetz, 20 April 2007, BGBl. I 2007 Nr.16 S.554.
95
Riksdag, Betänkande 1997/98:SfU13, decision 1998-06-08.
96
SFS 1998:674.
97
Pensionsmyndigheten, Orange Report 2011.
98
Munro, “The Coinages and Monetary Policies of Henry VIII,” Collected Works of Erasmus vol. 14, Table 1.
99
Munro, recalculated Phelps Brown–Hopkins basket, England, 1264–1700.
100
Drelichman and Voth, Journal of Economic History 70(4), 2010, Table 1.
101
National Academies, The Economic and Fiscal Consequences of Immigration (2016; NAP citation 2017), Summary pp. 7, 11.
102
Pal, “The Year the World Broke,” excerpt from chapter 2 of The Everything Code, Substack, 2026-10-22.
103
Pal, The Everything Code book page, raoulpal.com.
104
Harriman House and Pan Macmillan book pages, publication 3 Nov 2026.