The Harrod-Domar model is described as a 'knife-edge' because:

Any deviation of the actual growth rate from the warranted rate is self-reinforcing (cumulative boom or slump) rather than self-correcting

It guarantees full employment automatically

It has diminishing returns to capital, like the Solow model

In the Solow model, the steady state k* is stable and unique specifically because:

The production function exhibits diminishing returns to capital alone

The capital-output ratio is fixed

There is no depreciation at all

Population growth is always zero

In the Solow model, steady-state per-capita income growth without technological progress is:

Zero — capital accumulation alone cannot sustain it because of diminishing returns

Always positive and accelerating forever

Exactly equal to the population growth rate

Embodied technical progress, as distinct from disembodied progress, requires:

New investment in new vintages of capital to actually realise the improvement

Only a change in the interest rate

Endogenous growth theory (e.g. the AK model) was developed specifically to address:

The Solow model's treatment of technological progress as an unexplained exogenous variable

Which of the following are among Kaldor's stylised facts of economic growth?

Zero economic growth in every country at all times

Models of Economic Growth: Harrod-Domar, Solow, Robinson and Kaldor, and Technical Progress

This chapter works through the syllabus's formal growth models with their actual equations — Harrod-Domar's knife-edge, the Solow model's steady state, and the post-Keynesian Robinson/Kaldor models — before covering the distinction between disembodied and embodied technical progress and the rise of endogenous growth theory.

1. The Harrod-Domar Model

The Harrod-Domar model derives the WARRANTED rate of growth (Gw) — the rate at which, given the economy's savings ratio (s = S/Y) and its fixed capital-output ratio (v = ΔK/ΔY, the incremental capital-output ratio or ICOR), planned saving exactly equals the investment needed to keep capital growing in step with output: Gw = s/v. This is a KNIFE-EDGE result: nothing in the model automatically drives the economy's ACTUAL growth rate (Ga) to equal Gw — if Ga exceeds Gw, capital is scarce relative to what firms would like to hold, encouraging MORE investment and pulling Ga further above Gw (cumulative boom); if Ga falls short of Gw, capital is excessive, discouraging investment and pulling Ga further below Gw (cumulative slump) — the model has NO self-correcting mechanism of the kind the Solow model provides. Harrod additionally defined the NATURAL rate of growth (Gn) — the growth rate consistent with full employment of a growing labour force, Gn = (rate of labour-force growth) + (rate of labour-augmenting technical progress) — and full-employment, steady growth requires all three rates (Ga, Gw, Gn) to coincide, which the model gives no reason to expect.

🧠 Reading the Harrod-Domar formula numerically

If an economy saves 15% of national income (s = 0.15) and its incremental capital-output ratio is 3 (v = 3, meaning 3 units of extra capital are needed to produce 1 extra unit of output), the warranted growth rate is Gw = s/v = 0.15/3 = 0.05, or 5% per year — a standard NET numerical substitution, and raising the savings rate OR lowering the ICOR (more capital-efficient investment) each independently raises the warranted growth rate.

2. The Solow Growth Model

The Solow model replaces Harrod-Domar's FIXED capital-output ratio with a NEOCLASSICAL production function Y = F(K,L) exhibiting constant returns to scale and, crucially, DIMINISHING returns to capital alone. Writing everything in per-worker terms (k = K/L, y = Y/L = f(k)), the fundamental equation of capital accumulation is: k̇ = sf(k) − (n + δ)k — capital per worker rises with saving/investment per worker, sf(k), and falls with 'capital widening' needs: population growth (n, which dilutes existing capital across more workers) and depreciation (δ). The STEADY STATE k* is where k̇ = 0, i.e. sf(k) = (n+δ)k — saving exactly covers the capital widening a growing, depreciating economy needs, and BECAUSE of diminishing returns to capital (f(k) is concave), this steady state is UNIQUE and STABLE: unlike Harrod-Domar's knife-edge, an economy below k* has sf(k) > (n+δ)k and k grows toward k, while one above k has the reverse and k shrinks back — self-correcting from either direction.

The Solow model's key implications
Result	Implication
Steady-state per-capita growth	Zero absent technological progress — capital accumulation alone cannot sustain per-capita income growth forever, because of diminishing returns.
With exogenous technology growth at rate g	Steady-state per-capita income grows at exactly g — only technological progress drives sustained long-run per-capita growth in this model.
Conditional convergence	A country further below its OWN steady state grows faster (diminishing returns give a higher marginal product of capital when k is low), so poorer countries with similar steady-state determinants converge toward richer ones.

🎯 Why an unexplained technology term motivated endogenous growth theory

The Solow model's own central conclusion — that sustained per-capita growth requires technological progress — is delivered by a term (g) the model itself treats as purely EXOGENOUS, i.e. it explains nothing about where technology growth comes from. This 'unexplained residual' (famously large in growth-accounting exercises, sometimes called the Solow residual) is exactly the gap that endogenous growth theory (covered later in this chapter) set out to close by modelling technological change, human capital or knowledge spillovers as outcomes of economic decisions rather than a free gift.

3. Robinson's and Kaldor's Post-Keynesian Growth Models

The post-Keynesian (Cambridge) growth models, associated chiefly with Joan Robinson and Nicholas Kaldor, assign DIFFERENT savings propensities to different income classes — workers (who save little, sw) and capitalists/profit-recipients (who save more, sp, sp > sw) — a distributional dimension entirely absent from the aggregate-saving Harrod-Domar and Solow frameworks. The resulting 'Cambridge equation' (in its simplest, workers-save-nothing version) gives the growth rate as g = sp × π, where π is the profit share of national income — so a given required growth rate DETERMINES the profit share needed to generate the saving to finance it, reversing the usual causal story where distribution is exogenous and growth is derived; income DISTRIBUTION (between wages and profits) is itself the equilibrating variable. Kaldor additionally proposed a set of empirical 'STYLISED FACTS' of growth — a roughly constant capital-output ratio, a roughly constant share of profits in national income, a roughly constant rate of return on capital, and roughly steady growth rates of output and capital per worker — which any acceptable growth model, Kaldor argued, should be able to reproduce simultaneously, and which Kaldor's own 'technical progress function' (relating the growth rate of output per worker to the growth rate of capital per worker, deliberately blurring the usual movement-along-versus-shift-of distinction in the production function) was designed to match.

4. Technical Progress: Disembodied, Embodied and Endogenous Growth

DISEMBODIED technical progress raises the productivity of ALL existing capital and labour immediately, regardless of when they were installed or hired — a pure shift of the whole production function upward (e.g. a better organisational method usable with any existing machinery). EMBODIED technical progress, by contrast, is carried only by NEW vintages of capital — realising the improvement requires actually investing in and installing the new machinery, so an economy with a low investment rate captures embodied technical progress only slowly, however advanced the available technology, because old capital simply does not benefit until replaced. ENDOGENOUS GROWTH THEORY (Romer, Lucas, and others) was developed specifically to overcome the Solow model's treatment of technology as an unexplained exogenous gift: by modelling R&D, knowledge spillovers or human-capital accumulation as themselves generating (constant or increasing, rather than diminishing) returns to broadly-defined capital, these models can generate permanently sustained per-capita growth from investment and policy choices alone — the simplest version, the 'AK model' (Y = AK, with NO diminishing returns to the broad capital stock K), gives a growth rate of saving-driven output directly proportional to the savings rate, with no automatic convergence to any external steady state.

⚠️ Disembodied versus embodied — the investment-rate test

The reliable way NET distinguishes these: if a question describes a technology benefit that spreads to old capital/existing workers instantly without any new investment, it is disembodied; if it describes a technology that only new machines/equipment carry, so that the SPEED of diffusion depends on how fast old capital is replaced (the investment/scrappage rate), it is embodied — a country with a low investment rate is precisely the one that would be slow to benefit from embodied progress even with excellent technology 'on paper'.

Key takeaways

Harrod-Domar's warranted growth rate Gw = s/v has no self-correcting mechanism (the 'knife-edge'); full-employment steady growth additionally requires the natural rate Gn = labour-force growth + technical progress to coincide with it.
The Solow model's diminishing returns to capital give a unique, stable steady state k* where sf(k) = (n+δ)k; steady-state per-capita growth is zero without exogenous technology, which is exactly why endogenous growth theory arose.
Robinson/Kaldor's Cambridge growth model makes distribution (the profit share) the equilibrating variable via g = sp × π; Kaldor's stylised facts constrain what any acceptable growth model must reproduce.
Disembodied technical progress benefits all capital instantly; embodied progress requires new investment to realise, so diffusion speed depends on the investment rate.
Endogenous growth theory (e.g. the AK model) models technology/knowledge as generated within the economy, with constant (not diminishing) returns to broad capital, sustaining permanent per-capita growth from saving alone.

Practice questions (9)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

If an economy's savings ratio is 0.20 and its incremental capital-output ratio is 4, the Harrod-Domar warranted growth rate is:
1. 5% (Gw = s/v = 0.20/4 = 0.05)
2. 20%
3. 80%
4. 4%
Show answer
Answer: A — 5% (Gw = s/v = 0.20/4 = 0.05)
Direct substitution into Gw = s/v gives 0.20/4 = 0.05, or 5% — a standard numerical item testing the formula directly.
The Harrod-Domar model is described as a 'knife-edge' because:
1. Any deviation of the actual growth rate from the warranted rate is self-reinforcing (cumulative boom or slump) rather than self-correcting
2. It guarantees full employment automatically
3. It has diminishing returns to capital, like the Solow model
4. It requires no savings at all
Show answer
Answer: A — Any deviation of the actual growth rate from the warranted rate is self-reinforcing (cumulative boom or slump) rather than self-correcting
Because the capital-output ratio is FIXED (not diminishing-returns-based as in Solow), a gap between actual and warranted growth is amplified rather than corrected — this instability is exactly the 'knife-edge' the model is known for.
In the Solow model, the steady state k* is stable and unique specifically because:
1. The production function exhibits diminishing returns to capital alone
2. The capital-output ratio is fixed
3. There is no depreciation at all
4. Population growth is always zero
Show answer
Answer: A — The production function exhibits diminishing returns to capital alone
Diminishing returns to capital mean sf(k) rises less than proportionally with k while (n+δ)k rises proportionally, so the two curves cross exactly once, giving a unique steady state that pulls the economy back toward it from either side.
In the Solow model, steady-state per-capita income growth without technological progress is:
1. Zero — capital accumulation alone cannot sustain it because of diminishing returns
2. Always positive and accelerating forever
3. Exactly equal to the population growth rate
4. Negative in every case
Show answer
Answer: A — Zero — capital accumulation alone cannot sustain it because of diminishing returns
At the steady state k*, capital per worker stops growing, and with a fixed technology level output per worker stops growing too — only an ongoing rate of technological progress (g) delivers sustained steady-state per-capita growth in the model.
In the simplest Cambridge (Robinson/Kaldor) growth equation g = sp × π, the variable π represents:
1. The profit share of national income
2. The population growth rate
3. The nominal interest rate
4. The exchange rate
Show answer
Answer: A — The profit share of national income
π is the profit share; because only capitalists/profit-recipients save in this simplest version, the growth rate depends on the savings rate out of profits (sp) times the share of income going to profits.
Embodied technical progress, as distinct from disembodied progress, requires:
1. New investment in new vintages of capital to actually realise the improvement
2. No investment of any kind
3. Only a change in the interest rate
4. A fall in the population
Show answer
Answer: A — New investment in new vintages of capital to actually realise the improvement
Embodied progress is carried only by new capital vintages, so its diffusion speed through the economy depends directly on the rate of new investment/replacement of old capital — unlike disembodied progress, which benefits existing capital immediately.
Endogenous growth theory (e.g. the AK model) was developed specifically to address:
1. The Solow model's treatment of technological progress as an unexplained exogenous variable
2. A shortage of labour alone
3. The need to abolish all saving
4. A fixed exchange rate regime
Show answer
Answer: A — The Solow model's treatment of technological progress as an unexplained exogenous variable
Endogenous growth theory models technology, knowledge or human capital as generated by economic decisions (R&D, learning, education) with non-diminishing returns to broad capital, closing the explanatory gap the Solow model's exogenous g left open.
Which of the following are among Kaldor's stylised facts of economic growth?
1. A roughly constant capital-output ratio
2. A roughly constant share of profits in national income
3. A roughly constant rate of return on capital
4. Zero economic growth in every country at all times
Show answer
Answer: A — A roughly constant capital-output ratio; B — A roughly constant share of profits in national income; C — A roughly constant rate of return on capital
The first three are among Kaldor's actual stylised facts (alongside steady growth rates of output/capital per worker); zero growth everywhere is the opposite of what growth theory or Kaldor's facts describe.
Name the growth rate (Harrod) equal to the rate of labour-force growth plus the rate of labour-augmenting technical progress, needed for full-employment steady growth.
Numerical answer — type the value.
Show answer
Answer: natural rate of growth
The natural rate of growth (Gn) is Harrod's term; full-employment steady growth in his framework requires the actual, warranted and natural rates all to coincide, which the model gives no automatic reason to expect.