The Treadmill

Essay · economics · September 2026

Every government promises growth and every recession is treated as a disaster. Why? Not because economies cannot stand still in principle, but because the arrangements built during two centuries of growth depend on it: jobs, when each hour of work produces a little more every year; debts and pensions, when they are measured against a growing income; and politics, when everyone can gain at once. Portugal’s stagnant decade shows what happens when growth stops by accident.

Reading time ≈ 12 min Jobs → debts → pensions → politics Part I of two

The question

Growth is the one economic goal every party shares. Budgets are built on forecasts of it, pension reforms on projections of it, and the worst thing that can be said of a government is that the economy shrank on its watch. Yet for most of history economies did not grow at all, and a long line of economists, from John Stuart Mill to the degrowth movement, has asked why a rich society should need to keep getting richer.

This essay asks the practical version of that question: what, in a modern economy like Portugal’s, actually depends on growth, and how much? It goes through the usual answers one at a time, with the numbers, and then looks at what happened in the years when Portuguese growth stopped. The second part, Growth of What, asks what exactly is growing, and whether GDP is the right thing to count.

0.02%average yearly growth of world income per head from year 1 to 1820; since then, about 1.3%

~1.5%growth Portugal has needed each year just to stop unemployment rising, 1996–2025

137%Portugal’s public debt in 2025 as a share of GDP if nominal GDP had stayed at its 2020 level; actual, 90%

−1.3%change in Portugal’s real GDP per head from 2000 to 2013, while unemployment reached 17%

The long run

Growth is new

On Angus Maddison’s reconstruction, world income per head was about the same in the year 1000 as in the year 1, and rose by about half between then and 1820: an average of 0.02 per cent a year over eighteen centuries. Since 1820 it has multiplied by about fifteen, at 1.3 per cent a year (Figure 1).[1] Portugal followed the same curve late and fast. Its income per head barely moved in the nineteenth century, grew by 5.5 per cent a year between 1950 and 1973, the fastest in its history, and has grown by about 1 per cent a year since 2000.[1]

Fig. 1 — GDP per head, world and Portugal, 1000–2008, in 1990 international (Geary–Khamis) dollars. Estimates before 1820 are sparse and rough. Data: Angus Maddison, Historical Statistics of the World Economy, 2010 release.

The economists who first watched growth begin did not assume it would last, or that it should. John Stuart Mill, in 1848, expected the economy to settle into a “stationary state” and was glad of it: “I cannot, therefore, regard the stationary state of capital and wealth with the unaffected aversion so generally manifested towards it by political economists of the old school.” A stationary economy, he wrote, “implies no stationary state of human improvement”; there would be as much room as ever “for improving the Art of Living, and much more likelihood of its being improved, when minds ceased to be engrossed by the art of getting on.”[2] Two centuries later rich economies are still growing and still anxious when they do not. The reasons are more concrete than a taste for getting on.

The first reason

Jobs: the treadmill

The strongest reason is arithmetic. Each year, on average, an hour of work produces a little more than the year before. If the economy’s output grows more slowly than output per hour, the economy needs fewer hours of work. Those hours can be shed in three ways: fewer people working, each person working fewer hours, or fewer people of working age. In practice, in a downturn, the first happens fastest.

Economists call this output per hour: real GDP, the value of everything the economy produces at constant prices, divided by all the hours worked by employees and the self-employed. It rises with better tools, skills and organisation, and also when work moves to more productive sectors or low-productivity jobs disappear.

Portugal ran this experiment between 2000 and 2013. Real GDP ended the period where it started. Output per hour rose by about 18 per cent. So total hours worked fell by 15 per cent, and employment by 12 per cent (Figure 2). Part of that rise was the disappearance of low-productivity jobs in construction and small commerce, which raises the average without anyone producing more. After 2013 the treadmill ran the other way: GDP grew by 2 per cent a year, and almost all of it came from more hours of work, with output per hour barely rising.[3]

Fig. 2 — Portugal: real GDP, total hours worked and GDP per hour worked, index 2000 = 100, 1995–2025. Data: European Commission, AMECO database (spring 2026); author’s calculation.

Economists call the link between growth and unemployment Okun’s law, and it has held up well: a study of twenty rich countries found it “strong and stable by the standards of macroeconomics”, though with very different strengths. Unemployment in Spain, with its temporary contracts, responds five times more to a fall in output than in Japan, with its tradition of jobs for life; Portugal lies in between, and became more sensitive after 1995.[4] A simple version for Portugal, fitted to the years since 1996, says that unemployment has fallen only in years when the economy grew by more than about 1.5 per cent (Figure 3).[3]

Fig. 3 — Portugal, each year 1996–2025: growth of real GDP and change in the unemployment rate (percentage points). Dashed line: least-squares fit excluding 2020 and 2021, when furlough schemes broke the usual link; slope −0.36. An illustration, not an estimate of the natural rate. Data: AMECO; author’s calculation.

The treadmill is not a law of nature. Japan’s total hours of work fell by 16 per cent between 1991 and 2025, yet its unemployment stayed between 2 and 5 per cent, because its working-age population shrank and people worked fewer hours each.[3] A careful survey of macroeconomic theories concludes that they allow for a zero-growth economy, but that in the Keynesian ones “zero growth leads to increasing unemployment under the common assumption of increases in labour productivity” unless working time falls.[5] That is the choice the Keynes essays on this blog discussed: take productivity as more income, which needs growth, or as more free time, which does not. Portugal has taken it almost entirely as income: average hours per worker were about 1,760 a year in both 1995 and 2025, while in Germany and Ireland they fell by more than 10 per cent.[6]

The second reason

Debts: the denominator

Debts are fixed in euros; the income that services them is not. Whether a debt gets lighter or heavier depends on the gap between the interest rate on it and the growth of nominal income, which economists write as r minus g. When growth exceeds the interest rate, the debt shrinks relative to income without anyone repaying a cent; when interest exceeds growth, it grows by itself (Figure 4).

Fig. 4 — Implicit interest rate on general government debt (interest paid over the previous year’s debt) minus the growth of nominal GDP, percentage points, Portugal and the euro area, 1996–2025. Data: AMECO; author’s calculation.

Portugal has had both. Between 2008 and 2013 the gap averaged plus 4.4 points, and on the Commission’s accounting it added 28 percentage points of GDP to the public debt between 2009 and 2014. Between 2021 and 2025, with inflation and a strong recovery, it subtracted 36 points.[3] The fall in the public debt ratio from 134 per cent of GDP in 2020 to 90 per cent in 2025, discussed in The Inheritance, came from the denominator: the stock of debt in euros rose by 2 per cent while nominal GDP rose by 53 per cent. Private debt tells the same story. Portuguese households and firms owed 210 per cent of GDP in 2012 and 119 per cent in 2025, with almost the same number of euros (Figure 5).[3][7]

Fig. 5 — Portugal, 2025: public debt (general government, Maastricht definition) and private debt (households and non-financial corporations, consolidated) as a percentage of GDP, actual and recomputed with nominal GDP held at its 2020 level (public) or 2012 level (private). Data: AMECO; Eurostat, tipspd20; author’s calculation.

This is not an argument that growth is needed to repay debt. Olivier Blanchard showed that interest rates below growth rates have been “more the rule than the exception” historically, which is why debt can often be rolled over rather than repaid.[8] But that comfort depends on g. An economy that stopped growing in nominal terms would turn every existing debt, public and private, into a burden that could only be reduced by paying it down.

The third reason

Pensions: the wage bill

A pay-as-you-go pension system, like Portugal’s, pays today’s pensions out of today’s contributions. Paul Samuelson showed in 1958 that such a system pays its members a “biological rate of interest”, the growth of the population; Henry Aaron added in 1966 that the return is the growth of the population plus the growth of real wages.[9] The promise each generation makes to the one before is only as good as the growth of the wage bill that will pay for it.

The European Commission’s projections put a number on it. In the baseline, Portuguese public pension spending falls from 12.2 per cent of GDP in 2022 to 10.4 per cent in 2070, largely because pensions will replace a smaller share of wages, as The Sustainability Story discussed. If productivity grows more slowly, converging to 0.6 per cent a year instead of 0.8, spending in 2070 is 0.8 points of GDP higher.[10] Slower growth does not break the system on its own. It shifts the choice between higher contributions and lower pensions.

The fourth reason

Politics: when everyone can gain

The softest reason may be the most important. When the economy grows, one group’s gain need not be another’s loss. Benjamin Friedman argued that “economic growth – meaning a rising standard of living for the clear majority of citizens – more often than not fosters greater opportunity, tolerance of diversity, social mobility, commitment to fairness, and dedication to democracy.”[11] The evidence since is suggestive rather than conclusive, but it points one way.

People feel losses more than gains. Across 2.3 million American respondents and European surveys, a 10 per cent contraction of the economy lowered life satisfaction about six times more than a 10 per cent expansion raised it; the authors estimate that 2 to 6 per cent of growth is needed to offset 1 per cent of contraction.[12] After financial crises, the vote for far-right parties in advanced democracies has risen by about 30 per cent, relative to its previous level, over five years; ordinary recessions have had much smaller effects.[13] In Europe after 2008, regions where unemployment rose lost trust in their national parliaments and voted more for anti-establishment parties.[14] And a large American study finds that people who grew up in years of fast growth for ordinary families are less likely to see economic life as a zero-sum game, in which one person’s gain is another’s loss; “older cohorts appear much less zero-sum than younger ones”.[15] These are correlations, but they suggest why politicians fear stagnation even when the arithmetic could be managed.

The test

Portugal’s stagnant decade

Portugal has already lived through the absence of growth, not by choice. In 2007 Olivier Blanchard wrote that “the Portuguese economy is in serious trouble: Productivity growth is anemic. Growth is very low”, and predicted “a period of sustained high unemployment until competitiveness has been reestablished”.[16] Real GDP per head in 2013 was 1.3 per cent lower than in 2000; Italy’s was 7 per cent lower, while the euro area’s was 8 per cent higher (Figure 6).[3]

Fig. 6 — Real GDP per head, index 2000 = 100: Portugal, Italy, the euro area (20 countries) and the EU (27), 1995–2025. Data: AMECO.

Every mechanism above showed up. Unemployment rose from 4.8 per cent in 2000 to 17.2 per cent in 2013. The public debt went from 54 to 131 per cent of GDP, as interest outran a shrinking economy. Real pay per employee was 3 per cent lower in 2014 than in 2000.[3] Emigration surged: permanent emigrants peaked at about 54,000 in 2013, and all emigrants at about 135,000 in 2014.[17] The years of zero growth were not a calm stationary state. They were a crisis, and when growth returned, unemployment, debt and emigration all fell with it.

Advocates of lower growth reply, fairly, that this proves little about their proposals. Portugal did not choose to stop growing; it stagnated with its institutions unchanged, under a currency it could not devalue and a debt it had to refinance in a panic. Japan, whose income per head has grown by less than 1 per cent a year since 1991, shows another version of slow growth: low unemployment, flat wages, and a public debt twice its GDP held mostly at home.[3] Neither case tells us what a planned slowdown would look like.

The dispute

A law, or an arrangement?

Is growth built into a modern economy, or only into the way we have organised it? One school argues that credit money itself requires growth: the economist Mathias Binswanger argued that, since firms borrow to invest and must earn back more than they borrowed, “a zero growth rate is not feasible in the long run”.[18] Tim Jackson and Peter Victor tested the claim in a model of a monetary economy and rejected it: “neither credit creation nor the charging of interest on debt creates a ‘growth imperative’ in and of themselves.” In their simulations it was an austerity policy, not the existence of debt, that crashed the economy. They add that other pressures remain: firms’ pursuit of higher productivity, speculation in assets, and people’s wish for higher incomes.[18]

A review of the literature by Oliver Richters and Andreas Siemoneit reaches a similar verdict: market competition and profit do not in themselves force growth, but rising productivity forces firms to keep becoming more efficient, which produces a “political growth imperative” for states that must keep people employed. Most theories of a growth imperative, they write, “can be boiled down to the argument of ‘jobs’”.[19] Robert Solow, one of the founders of growth theory, reportedly said that rich economies could live with slow or no growth if they chose to take their rising productivity as leisure.[20]

The question may be settled by events rather than theory. Growth in the rich world has been slowing for decades. Robert Gordon found that American total factor productivity, the part of growth not explained by more capital and labour, grew after 1970 at “barely a third” of its 1920–1970 rate; Lawrence Summers argued that rich economies now find it “increasingly difficult” to combine adequate growth with financial stability; and the IMF expects world growth to fall to 2.8 per cent by the end of this decade, a percentage point below its pre-pandemic average.[21][22] The OECD’s long-term scenarios assume that Portugal’s potential income per head will grow by about 1.7 per cent a year to 2060, more than twice its pace of 2003–2019; that is an assumption of catching up, not an observed trend.[22]

The balance

What depends on growth

The case, set out plainly, is this. A modern economy does not need growth in the way a body needs food; the theories that say it must grow collapse under close examination. But nearly everything it has promised was written on the assumption of growth. Employment depends on growth as long as productivity rises and hours of work do not fall. Debts, public and private, become lighter only when incomes rise faster than interest. Pensions paid from wages depend on the wage bill. And politics is easier when everyone can gain at once, harder when a gain for some must be a loss for others.

Each dependence could in principle be loosened: shorter working hours instead of unemployment, less debt, pensions designed for a stable wage bill, taxes that share out a fixed income. Whether that could be done without the pain that stagnation brought to Portugal after 2000 is the real question in the debate about growth, and the Portuguese experience says only that stagnation without those changes is costly. The other question is what the growth everyone wants is actually made of, and whether the figure that measures it counts the right things. That is the subject of Part II.

On method and tools

This piece was written collaboratively with Claude Opus 5.5 (Anthropic): human specification, editorial direction and critical review; machine data research, analysis and drafting. The figures and derived numbers are computed by scripts/growth.py, shared with the second part, mainly from the European Commission’s AMECO database, the Maddison Project, Eurostat and the OECD. The unemployment-growth line in Figure 3 is a simple fit for illustration, not an econometric estimate; the published Okun coefficients are cited in the text. Quotations were checked against the primary sources where these could be obtained; the Solow remark is known only as quoted in a magazine interview, and Friedman’s sentence was checked in his own conference paper of the same title rather than in the book. The downloaded sources, with a table or page reference for every number, are kept with the script’s data; results are in docs/growth-results.json and the figures in docs/growth-figures.html.

The cover photograph is Rush hour, the Vasco da Gama bridge, by F Mira; CC BY-SA 2.0, via Wikimedia Commons, cropped.

Authored by: Luis Matos Ferreira — Physicist, Developer, Writer

Related essays on this blog

1. Growth of What — Part II: what GDP measures, what it misses, and the alternatives.
2. The Fifteen-Hour Week — Keynes’s prediction, and where a century of productivity went.
3. The Inheritance — public debt, public capital and pensions.
4. The Sustainability Story — the Portuguese pension system and its forecasts.

Sources

1. Angus Maddison, Historical Statistics of the World Economy: 1–2008 AD, 2010 release; Maddison Project Database 2023 (Bolt and van Zanden), world and Portugal; author’s growth rates. The two releases use different price bases and are compared only in growth rates.
2. John Stuart Mill, Principles of Political Economy (1848; Ashley edition, 1909, from the 7th edition of 1870), Book IV, Chapter VI, “Of the Stationary State”, §§2 and 9.
3. European Commission, AMECO database (spring 2026): real and nominal GDP, total hours worked, employment, unemployment, real GDP per head, real compensation per employee, implicit interest rate and snowball effect on general government debt; author’s calculations.
4. L. Ball, D. Leigh and P. Loungani, “Okun’s Law: Fit at Fifty?”, Journal of Money, Credit and Banking, 2017; NBER Working Paper 18668, pp. 2–3, 22 and Tables 5–6.
5. S. Lange, Macroeconomics Without Growth, Metropolis, 2018, pp. 499–500 and 527–529.
6. OECD, average annual hours actually worked per worker, total employment (the OECD advises comparing trends rather than levels).
7. Eurostat, private sector debt, consolidated, households and non-financial corporations (tipspd20).
8. O. Blanchard, “Public Debt and Low Interest Rates”, American Economic Review 109(4), 2019; PIIE Working Paper 19-4, abstract and p. 7.
9. P. A. Samuelson, “An Exact Consumption-Loan Model of Interest with or without the Social Contrivance of Money”, Journal of Political Economy 66(6), 1958, pp. 472–474; H. Aaron, “The Social Insurance Paradox”, Canadian Journal of Economics and Political Science 32(3), 1966, p. 371 ff.
10. European Commission, 2024 Ageing Report, Institutional Paper 279, pp. 7, 147, 204 and 218; Portugal country fiche, p. 48 and Table 30.
11. B. M. Friedman, The Moral Consequences of Economic Growth, Knopf, 2005, p. 4; wording checked in the author’s paper of the same title, 12th Dubrovnik Economic Conference, 2005.
12. J.-E. De Neve, G. Ward, F. De Keulenaer, B. Van Landeghem, G. Kavetsos and M. Norton, “The Asymmetric Experience of Positive and Negative Economic Growth”, Review of Economics and Statistics 100(2), 2018; CEP Discussion Paper 1304, section 4.1.
13. M. Funke, M. Schularick and C. Trebesch, “Going to extremes: Politics after financial crises, 1870–2014”, European Economic Review 88, 2016; CESifo Working Paper 5553, abstract, p. 17 and Table 5.
14. Y. Algan, S. Guriev, E. Papaioannou and E. Passari, “The European Trust Crisis and the Rise of Populism”, Brookings Papers on Economic Activity, Fall 2017; conference draft, p. 3.
15. S. Chinoy, N. Nunn, S. Sequeira and S. Stantcheva, “Zero-Sum Thinking and the Roots of US Political Differences”, American Economic Review 116(3), 2026; NBER Working Paper 31688 (revised April 2025), pp. 4, 6 and 28–30.
16. O. Blanchard, “Adjustment within the euro. The difficult case of Portugal”, Portuguese Economic Journal 6, 2007, pp. 1–2.
17. Observatório da Emigração, Emigração Portuguesa 2025: Relatório Estatístico, Quadro 1.3, p. 32 (INE data).
18. M. Binswanger, “Is there a growth imperative in capitalist economies? A circular flow perspective”, Journal of Post Keynesian Economics 31(4), 2009, as quoted in T. Jackson and P. A. Victor, “Does credit create a ‘growth imperative’? A quasi-stationary economy with interest-bearing debt”, Ecological Economics 120, 2015, pp. 32, 33 and 44–46.
19. O. Richters and A. Siemoneit, “Growth imperatives: Substantiating a contested concept”, Structural Change and Economic Dynamics 51, 2019 (working paper version, Oldenburg V-414-18, 2018); and “Fear of stagnation? A review on growth imperatives”, VÖÖ Discussion Paper 6/2017, p. 10.
20. R. Solow, as quoted in S. Stoll, “Fear of Fallowing: The Specter of a No-Growth World”, Harper’s Magazine, March 2008 (not checked in the original; paraphrased).
21. R. J. Gordon, The Rise and Fall of American Growth, Princeton, 2016, ch. 1, p. 2; L. H. Summers, “U.S. Economic Prospects: Secular Stagnation, Hysteresis, and the Zero Lower Bound”, Business Economics 49(2), 2014, p. 66.
22. IMF, World Economic Outlook, April 2024, ch. 3, p. 65; OECD, Economic Outlook 117 long-term scenarios (business-as-usual), potential GDP per head, Portugal; author’s growth rates.