Economics•Chapter 11•8 min read•Updated September 28, 2026

Macroeconomics — The Solow Growth Model

O
OiyoContributor
11/23

Macroeconomics — The Solow Model Shows That Saving Cannot Raise the Growth Rate Permanently

The previous chapters dealt with the short run, in which output rises and falls around its potential. From this chapter on, we ask why potential output itself differs across countries and grows over time. By showing that capital accumulation alone cannot keep raising income per person, the Solow model reveals that the source of long-run growth is technology. The stability of differential equations is covered in Economics Math Chapter 4.

1. The dynamics of capital per worker

Let the production function per worker be y=f(k)y=f(k), with diminishing marginal product of capital. Each year a fraction ss of output is saved and invested, capital depreciates at rate δ\delta and the number of workers grows at rate nn. The change in capital per worker is

The fundamental equation of the Solow model
Δk=s f(k)−(δ+n) k\Delta k = s\,f(k) - (\delta + n)\,k
The first term is investment per worker; the second is the investment needed to keep capital per worker constant (replacing worn-out capital plus equipping new workers). The k* at which the two are equal is the steady state.

With f(k)=kf(k)=\sqrt{k}, s=0.2s=0.2 and δ+n=0.1\delta+n=0.1, the steady state solves 0.2k=0.1k0.2\sqrt{k}=0.1k, so k=2\sqrt{k}=2, that is k∗=4k^*=4 and y∗=2y^*=2.

The steady state of the Solow model
k Investment per worker k* = 4k* = 9 0.2√k 0.3√k (δ+n)k

When the investment curve lies above the required-investment line, k rises; when it lies below, k falls. A rise in the saving rate from 0.2 to 0.3 moves the steady state from 4 to 9.

When kk is below the steady state, investment per worker exceeds required investment and kk rises; when it is above, kk falls. Wherever it starts, the economy converges to k∗k^*.

2. The saving rate changes the level, not the growth rate

If the saving rate rises to 0.3, the steady state solves 0.3k=0.1k0.3\sqrt{k}=0.1k, giving k∗=9k^*=9 and y∗=3y^*=3. Output per worker is 50% higher. But once the new steady state is reached, output per worker stops growing again. Because the marginal product of capital diminishes, each additional unit of capital adds less output, until investment only just covers what is required.

A higher saving rate raises growth only during the transition. If the saving rate becomes 0.3 at k=4k=4, capital per worker grows by 0.3×2/4−0.1=5%0.3\times 2/4-0.1=5\% in the first year, and this rate falls towards 0 as kk approaches 9.

What changes what
ChangeSteady-state level of output per workerSteady-state growth per worker
Higher saving rate sRisesUnchanged (0)
Higher population growth nFallsUnchanged (0)
Rate of technological progress g—Equals g

A. Closed form and convergence speed: how long does the transition take?

With f(k)=kαf(k)=k^{\alpha} and labour efficiency growing at rate gg (measuring kk per effective worker), the steady state has a closed form.

Cobb-Douglas steady state
k∗=(sn+g+δ)11−α,∂ln⁡y∗∂ln⁡s=α1−αk^*=\left(\frac{s}{n+g+\delta}\right)^{\frac{1}{1-\alpha}},\qquad \frac{\partial\ln y^*}{\partial\ln s}=\frac{\alpha}{1-\alpha}
In the example above (α=0.5, g=0) the elasticity is 1, so a saving rate 1.5 times higher gives y* 1.5 times higher (2 → 3). With a capital share of α=1/3 the elasticity is 0.5: doubling the saving rate raises y* only by √2 ≈ 1.41.

Linearising the fundamental equation in ln⁡k\ln k around the steady state shows that the gap closes at a constant rate.

Convergence speed and half-life
dln⁡ydt≈−λ (ln⁡y−ln⁡y∗),λ=(1−α)(n+g+δ),t1/2=ln⁡2λ\frac{d\ln y}{dt}\approx-\lambda\,(\ln y-\ln y^*),\qquad \lambda=(1-\alpha)(n+g+\delta),\qquad t_{1/2}=\frac{\ln2}{\lambda}
In the example above λ = 0.5 × 0.1 = 5%, a half-life of about 14 years. With a capital share of 1/3 and n+g+δ = 6%, λ = 4% and the half-life is about 17 years. The gap with empirical estimates (around 2% a year) is taken up in chapter 22.

Computing the actual path year by year for the economy whose saving rate rises from 0.2 to 0.3: output per worker starts at 2, closes half the gap around year 12 (the linear approximation says 14; the economy moves faster when it is far from the steady state), and after 60 years is 2.95, still just short of 3. Consumption drops to 1.4 in the first year, regains its original 1.6 in year 7, and then rises towards 2.1. The cost of a policy that raises saving is concentrated in the early years. Empirical estimates of convergence speed and the human-capital correction continue in chapter 22.

Transition path after saving rises from 0.2 to 0.3
years per worker 712 output y consumption c y* = 3

Output (blue) approaches the steady state of 3 ever more slowly; consumption (green) falls to 1.4 in the first year and passes its original level of 1.6 (grey dashed) in year 7.

3. The Golden Rule: the capital stock that maximizes consumption

Saving more raises output but can lower consumption. Steady-state consumption per worker is c∗=f(k∗)−(δ+n)k∗c^*=f(k^*)-(\delta+n)k^*, and the condition that maximizes it is

The Golden Rule condition
f′(kgold)=δ+nf'(k_{gold}) = \delta + n
Consumption is highest where the marginal product of capital equals the cost of maintaining capital. In the example above, 1/(2√k) = 0.1, so k_gold = 25; the saving rate that achieves it is 0.5, and consumption is 2.5.

Saving rates of 0.2 (consumption 1.6) and 0.3 (consumption 2.1) both leave less capital than the Golden Rule. In this economy a higher saving rate raises long-run consumption, but early in the transition consumption falls by as much as saving rises. In the year the saving rate rises from 0.2 to 0.3, consumption is 0.7×2=1.40.7\times 2=1.4, below the original 1.6. The current generation pays the cost and future generations reap the benefit. Conversely, an economy with more capital than the Golden Rule can raise consumption both now and in the future by saving less.

4. Convergence and technological progress

The model predicts that countries with less capital grow faster. But only countries with similar saving rates, population growth and technology converge to the same steady state (conditional convergence). Absolute convergence, in which all countries converge to the same income, does not show up well in the data. The data fit conditional convergence better.

If labour-augmenting technology progresses at rate gg per year, output per worker keeps growing at rate gg in the steady state. In the Solow model, technology is the only source that explains sustained growth in income per person. Yet the rate of technological progress is given from outside the model.

Growth accounting splits output growth by factor. In gY=αgK+(1−α)gL+gAg_Y=\alpha g_K+(1-\alpha)g_L+g_A, the remainder gAg_A after subtracting the contributions of capital and labour is called the Solow residual (total factor productivity growth). The residual is a measurement, not an explanation. What drives technological progress is the subject of the endogenous growth theory in Chapter 12.

Check your understanding

Problem 1. An economy with f(k)=kf(k)=\sqrt{k} and δ+n=0.1\delta+n=0.1 saves 20% of output. Find steady-state output and consumption per worker. What happens if population growth rises so that δ+n=0.2\delta+n=0.2?

Solution. k∗=4k^*=4, y∗=2y^*=2 and c∗=0.8×2=1.6c^*=0.8\times2=1.6. With δ+n=0.2\delta+n=0.2, 0.2k=0.2k0.2\sqrt{k}=0.2k gives k∗=1k^*=1 and y∗=1y^*=1: output per worker halves.

Problem 2. An economy with capital share α=1/3\alpha=1/3 and n+g+δ=0.06n+g+\delta=0.06 is at half its steady-state income. How many years does it take to close half the gap? If the country doubles its saving rate, by what factor does steady-state income rise?

Solution. λ=(1−1/3)×0.06=0.04\lambda=(1-1/3)\times0.06=0.04, so the half-life is ln⁡2/0.04≈17\ln2/0.04\approx17 years. The saving-rate elasticity is α/(1−α)=0.5\alpha/(1-\alpha)=0.5, so steady-state income rises by 20.5≈1.412^{0.5}\approx1.41. That is why saving alone struggles to explain income gaps of many times between countries.

References

  • Robert Solow, “A Contribution to the Theory of Economic Growth,” Quarterly Journal of Economics (1956)
  • N. Gregory Mankiw, David Romer and David Weil, “A Contribution to the Empirics of Economic Growth,” Quarterly Journal of Economics (1992)
  • N. Gregory Mankiw, Macroeconomics, ch. 8–9
  • MIT OpenCourseWare, 14.02 Principles of Macroeconomics
O

Oiyo

Editorial Desk

The OIYO editorial desk researches money, law, lifestyle, and self-understanding topics against primary sources and public statistics. Every piece carries source notes and is reviewed on a regular cycle for accuracy and usefulness.