University of Kent

School of Economics Discussion Papers

Kaldor’s 1970 Regional Growth Model Revisited

A. P. Thirlwall

July 2013

KDPE 1311
 Kaldor’s 1970 Regional Growth Model Revisited

A. P. Thirlwall1

University of Kent

Abstract: Kaldor’s 1970 paper ‘The Case for Regional Policies’ was republished in the
sixtieth anniversary volume of the Scottish Journal of Political Economy. This paper reflects
on the model after more than forty years, and argues that even though it has been criticised
for its deterministic nature, it has lost none of its relevance. It predates the ideas of so-called
‘new’ growth theory, and the new economic geography of Krugman, and provides at least a
partial explanation of why growth rates and levels of per capita income between regions and
between countries can continue to persist and even widen in contrast to the predictions of
orthodox equilibrium theory.

Key Words: Regional Growth; Kaldor; Uneven Development; Cumulative Causation.

JEL Codes: O18; R11

1
The author is Professor of Economics, School of Economics, University of Kent, Canterbury CT2 7NP, UK.
E-mail : at4@kent.ac.uk.

1
 Kaldor’s 1970 Regional Growth Model Revisited

A.P.Thirlwall

University of Kent

Introduction

In the sixtieth anniversary issue of the Scottish Journal of Political Economy, Kaldor’s 1970
paper ‘The Case for Regional Policies’ was reprinted as the most cited paper in the Journal
during the decade 1963 to 1973 (Kaldor, 2013). I provided a commentary on the paper
(Thirlwall, 2013) but was restricted for words. In the present paper, I reflect in more depth on
the Kaldor 1970 regional growth model, and elaborate on its appeal and significance. After
more than forty years it has lost none of its relevance, and provides at least a partial
explanation of why growth rates and levels of per capita income between regions and
between countries continue to persist and even widen.

The first point to make is that the origin of Kaldor’s interest in regional economic growth is
not entirely clear, but it did coincide with his switch of interest in the mid-1960s from the
pure theory of growth (Kaldor, 1957, 1961) to the applied economics of growth when he
became involved in policy-making in the United Kingdom at the highest level as Special
Adviser to the Chancellor of the Exchequer in the 1964 Labour government. In his
Cambridge Inaugural Lecture in 1966 (Kaldor, 1966) he addressed the causes of the slow rate
of growth of the UK economy compared to the economies of continental Europe, and related
them to the slow rate of growth of the manufacturing sector which was unable to draw on
cheap surplus labour in agriculture. To compensate, Kaldor argued for a Selective
Employment Tax (SET) on the service sector to release labour from ‘over-manned’ services,
and a subsidy to employment in manufacturing, with an extra Regional Employment
Premium (REP) to manufacturing industry in depressed regions (a form of regional
devaluation).SET was introduced in 1966, and the REP in 1967, but then Kaldor soon
changed his mind on the major cause of the UK’s slow growth. It was not a shortage of
labour in manufacturing; it was a balance of payments constraint on the growth of demand in
the economy as a whole.

2
This is the background to Kaldor’s address to the Scottish Economic Society in 1970 in
which he presented a model of regional economic growth in purely verbal form, but which
can be formalised as Dixon and I did in a paper in 1975 (Dixon and Thirlwall, 1975).
Kaldor’s ideas can be reduced to a four-equation, structural, model with cumulative features,
reminiscent of a more general model of ‘circular and cumulative causation’ developed by
Gunnar Myrdal in his book Economic Theory and Underdeveloped Regions (1957). Kaldor
was more than familiar with Myrdal’s ideas having worked with him in the Economic
Commission for Europe (ECE) in Geneva from 1947 to 1949, and remained close friends.

The 1970 Model

The first proposition of the model is that regional growth is driven by export growth. Kaldor
regarded exports as the only true autonomous component of aggregate demand, not just at the
regional level but also at the national level because consumption and investment demand are
largely induced by the growth of output itself. The more specialised regions are, the greater
the importance of exports. The second proposition is that export growth depends largely on a
region’s changing price competitiveness and the growth of income outside the region. The
third proposition is that the rate of change of a region’s prices is determined by the difference
between wage growth and labour productivity growth. Lastly, labour productivity growth is
partly determined by the growth of output itself through static and dynamic increasing returns
(captured by Verdoorn’s Law)2.

In equation form, the propositions may be specified as (t is a time subscript):

gt = ϒ (xt) (1)

where gt is the growth of regional output, and xt is the growth of exports.

xt = η (pdt – pft) + ε (zt) (2)

where pdt is the growth of domestic prices; pft is the growth of foreign prices measured in a
common currency; zt is the growth of income outside the region ; η (< 0) is the price elasticity
of demand for exports, and ε (>0) is the income elasticity of demand for exports.

pdt =wt - rt (3)

2
In his 1966 Inaugural Lecture, Kaldor revived Verdoorn’s Law which had lain virtually dormant for seventeen
years since the publication of Verdoorn’s paper in 1949. Verdoorn was one of Kaldor’s research staff at the ECE
from 1947 to 1949.

3
where wt is the growth of wages, and rt is the growth of labour productivity.

rt = rat + λ (gt) (4)

where rat is autonomous productivity growth and λ is the Verdoorn coefficient.

Substitution of equation (4) into (3) and the result into (2) and (1) gives the equilibrium
growth of regional output as:

γηw t  rat  pft   εz t 

gt  (5)
1  γηλ

Remembering that η < 0, growth is shown to be negatively related to domestic wage

increases, but positively related to foreign price increase and autonomous productivity
growth. Growth is also positively related to the growth of external demand and the size of the
Verdoorn coefficient. It is the Verdoorn coefficient (λ) that makes the model ‘circular’; but
whether growth is ‘cumulative’ (i.e. departs further and further away from equilibrium)
depends on the behaviour of the model out of equilibrium. To make the model dynamic, and
to assess whether it is stable or not, it is sufficient to put a one-period time lag into any of the
equations. Dixon and I chose to put a one-period lag in the export growth equation giving xt =
η(pdt-1 – pft-1) + ε(zt-1). Successive substitution as before gives a first order difference
equation, of which the general solution to the homogenous part is:

gt = A ( -ϒηλ)t (6)

where A is the initial condition. Whether the model is stable or not out of equilibrium

depends on the value of (-ϒηλ). If exports grow twice as fast as output, ϒ = 0.5. A typical
value for the Verdoorn coefficient (λ) is 0.5. In this case the price elasticity of demand for

exports (η) would have to exceed minus 4 for (-ϒηλ) >1, and for there to be ‘explosive’
growth. It is rare to find aggregate price elasticities of demand for exports as high as that, but
in any case we don’t observe in practice regional growth rates diverging – only levels of per
capita income. This suggests that regional growth rate differences that are observed are
associated with differences in regions’ equilibrium growth rates largely determined by
differences in the income elasticities of demand for exports (ε) and imports (π) associated
with regional differences in the structure of production and trade: whether regions specialise
in primary production or manufactured goods and sophisticated services.

4
Centre-Periphery Models

If the Verdoorn coefficient is ignored, and it is assumed that regional competitiveness stays
constant, equation (5) (ignoring lags) can be written as:

gt = ε (zt)/π (7)

where π is the income elasticity of demand for imports and is equal to 1/ϒ if regional balance
of payments equilibrium is a requirement with the growth of exports equal to the growth of
imports. Equation (7) is the classic centre-periphery model of Prebisch (1959) where the
growth of one region relative to others (gt/zt ) is equi-proportional to the ratio of the income
elasticity of demand for exports and imports (ε/π).Equation (7) can also be shown to be the
dynamic analogue of the static Harrod trade multiplier, Y = X/m, where Y is the level of
output; X is the level of exports, and m is the marginal propensity to import (Harrod, 1933;
Thirlwall, 1982). Kaldor first revived the Harrod multiplier in a letter to The Times
newspaper 15th March 1977, and argued that it is more important than Keynes’s investment
multiplier for understanding the pace and rhythm of economic growth in an open economy
(Kaldor, 1981). Or, to put it another way, it is more difficult for a country to rectify an
import-export gap than it is to rectify a savings-investment gap.

Regions within a country are particularly open, of course, and they also share a common
currency. That is why there is no exchange rate in the original Kaldor model. It is possible to
include one, however, by modifying the export growth equation (2) and then, if regional data
are not available, to test the model at the national level as an alternative. That is what Dixon
and I did; we applied the model to the UK economy and found it seriously over-predicted the
UK’s historical growth rate. We surmised that this was because the UK suffered perpetually
balance of payments crises which constrained overall aggregate demand. It is a weakness of
the original Kaldor model that import growth is not modelled and there is no balance of
payments constraint. A balance of payments constraint is easily incorporated, however (see
Thirlwall and Dixon, 1979). The export growth equation (2) can be modified to include the
rate of change of the exchange rate (e) :

xt = η(pdt - pft - et) + ε (zt) (8)

We can then add an equation for the rate of growth of imports (m):

5
 mt = ψ(pft- pdt + et) + π (gt) (9)

where ψ (<0) is the price elasticity of demand for imports and π (>0) is the income elasticity
of demand for imports.

Setting equation (8) equal to (9), and substituting equations (3) and (4) into (8) gives the
balance of payments equilibrium growth rate of:

gt 
1  η   w t  rat  pft  et   εz t (10)
π  λ1  η   

If there is no Verdoorn effect (λ = 0), and relative prices measured in a common currency
remain unchanged, equation (10) collapses to equation (3).

Of course, regions within countries don’t experience classic balance of payments problems in
the sense that an exchange rate comes under pressure, but if import growth exceeds export
growth and capital transfers (domestic and international) do not finance the difference, the
balance of payments constraint will show up in slow growth and rising unemployment.
Regional problems are balance of payments problems (Thirlwall, 1980) as we witness in the
peripheral countries of the Eurozone today. A large part of the sovereign debt and private
banking crisis in the Eurozone stems from the heavy borrowing by the deficit countries of
Greece, Spain, Portugal and Italy from the surplus countries of Germany, the Netherlands and
Austria (see Priewe, 2012). The free movement of capital facilitates the financing of deficits,
but exposes countries to adverse internal and external macroeconomic shocks if the flows are
debt-creating.

Significance and Adaptation of the Model

Kaldor’s 1970 paper has been an inspiration to growth and development economists, critical
of orthodox equilibrium growth theory which predicts that if economic and social divisions
emerge between regions or countries, forces will come into play to narrow those differences.
But the Kaldor model has had its critics, even those sympathetic to Kaldor’s non-equilibrium
view of the world. In particular, the deterministic nature of the model has come under attack
which doesn’t allow for the possibility of growth ‘reversals’ or ‘catch-up’ and its
inconsistency with Kaldor’s historical view of the growth and development process.

6
Setterfield (1997) was one of the first to raise concerns. He makes the model more flexible by
allowing both the Verdoorn coefficient (λ) and the income elasticity of demand for exports
(ε) to be endogenous on the grounds that initially successful regions or countries may get
‘locked-in’ to certain techniques of production and products inherited from the past which
prevents them from adapting to a new environment leading to falls in λ and ε. Long run
growth performance then becomes path dependent. Relatively high growth can endogenously
break down depending on the specific technological and institutional environment – although
not inevitably. Argyrous (2002) argues in fact that the Verdoorn coefficient and income
elasticity of exports may rise before ‘lock-in’ sets in, particularly in the right historical
conditions and with appropriate institutional structures, in which case the cumulative nature
of growth will continue unabated. Roberts (2006) makes a similar point that the point of lock-
in at which reversals may occur should not be thought of as a constant at which ‘lock-in’
inevitably occurs, but at which it may tend to occur, so that the threshold lacks intrinsic
closure. This is the essence of ‘open systems – ceteris paribus’ modelling in which exogenous
factors are explicitly allowed to influence system outcomes in a non-defined way. In this
framework, all equilibria are conditional equilibria. There are not determinate outcomes but
intermediate positions subject to potential revision by forces that are endogenous to the
system. This keeps the possibility open that ‘lock-in’ might be avoided through human action
or institutional change, including government policy. But eventually there is a conditional
equilibrium. Even with ‘lock-in’, the economy ultimately settles down to an equilibrium.

Leon-Ledesma (2002) allows for growth ‘reversals’ and ‘catch-up’ by a different root. He
adds three extra variables to the Verdoorn equation (4) to explain the growth of labour
productivity: (i) the investment/GDP ratio; (ii) a measure of innovative activity, and (iii) a
productivity gap variable. Innovative activity is then assumed to be a function of four factors
:the growth of output; the cumulative sum of real output (to capture learning by doing); the
level of education, and the productivity gap. This specification adds a fifth equation to the
Kaldor four-equation model specified above. The parameters of the model determine whether
growth is stable or not, and whether productivity levels diverge or converge. Applying the
model to seventeen OECD countries using pooled data for the period 1965 to 1994, the
author concludes “cumulative growth arises from the effect of the Verdoorn-Kaldor
relationship and also from the induced effect that growth itself has on learning and non-price
competitiveness. The diffusion of technology arising from the productivity gap, however, is a
significant force that counteracts those forces favouring a ‘catch-up’ process.”

7
There are certain similarities between Kaldor’s model and the so-called new economic
geography (NEG) pioneered by Krugman (1991), but there is nothing in Kaldor on the
interaction between scale economies and transport costs which lies at the heart of the
centripetal and centrifugal forces that lead in the Krugman model to regional agglomeration
and dispersal, respectively. Recent work that combines NEG with endogenous growth theory,
however, has moved in Kaldor’s direction stressing technological spill-overs, learning by
doing, and Allyn Young’s (1928) concept of macro-increasing returns resulting from the
interaction between activities subject to increasing returns and a price elastic demand for their
product (Bhattacharjea, 2010)3.

The NEG doesn’t acknowledge Kaldor’s 1970 paper, probably because it lacks formal
modelling, but as Dixon and I (1975) showed, it is possible to formalise the model, and to
appreciate its rich insights into the regional growth process. It also turns out to be a very
flexible model and is easily augmented to allow for ‘growth reversals’ and ‘catch-up’ – to
allow history to matter – as shown by Setterfield and Leon-Ledesma. The presence of
increasing returns, which is central to Kaldor’s vision of the growth and development
process, makes a huge difference to the way regional and national economies function and is
the major challenge to equilibrium theory that with its narrow neoclassical assumptions finds
it difficult to explain persistent or growing regional divergence within and between countries.

References

Argyrous, G. (2001), Setterfield on Cumulative Causation and Interrelatedness: A Comment,

Cambridge Journal of Economics, Vol. 25 pp. 103-106.

Bhattacharjea, A. (2010), Did Kaldor Anticipate the New Economic Geography? Yes, but…,
Cambridge Journal of Economics, Vol. 34 pp. 1057-1074.

Dixon, R. J. and A. P. Thirlwall (1975), A Model of Regional Growth Rate Differences on

Kaldorian Lines, Oxford Economic Papers, Vol. 27 pp. 201-214.

Harrod, R. (1933), International Economics (London: Macmillan).

Kaldor, N. (1957), A Model of Economic Growth, Economic Journal, Vol. 67 pp. 591-624.

3
Kaldor was a pupil of Allyn Young at the LSE in 1928 and took a full set of his lecture notes (Thirlwall, 1987;
Sandilands, 1990). For a survey of endogenous regional growth, see Roberts and Setterfield (2013).

8
Kaldor, N. (1961), ‘Capital Accumulation and Economic Growth’ in F. Lutz (ed.) The Theory
of Capital (London: Macmillan).

Kaldor, N. (1966), The Causes of the Slow Rate of Growth of the United Kingdom Economy
(Cambridge: Cambridge University Press).

Kaldor, N. (1970), The Case for Regional Policies, Scottish Journal of Political Economy,
Vol. 18 pp. 337-348. Reprinted in Scottish journal of Political Economy

Kaldor, N. (1981), The Role of Increasing Returns, Technical Progress and Cumulative
Causation in the Theory of International Trade and Economic Growth, Economie Appliquee
No.4.

Krugman, P. (1991), Increasing Returns and Economic Geography, Journal of Political

Economy, Vol. 27 pp. 483-499.

Leon-Ledesma, M. (2002), Accumulation, Innovation and Catching Up: An Extended

Cumulative Growth Model, Cambridge Journal of Economics, Vol. 26 pp. 201-216.

Myrdal, G. (1957), Economic Theory and Underdeveloped Regions (London: Duckworth).

Prebisch, R. (1959), Commercial Policy in the Underdeveloped Countries, American

Economic Review, Papers and Proceedings, Vol. 49 pp. 251-273.

Priewe, J. (2012), ‘European Imbalances and the Crisis of the European Monetary Union’ in
H.Herr, T. Niechoj, C. Thomasberger, A. Truger, and T. Van Treek (eds), From Crisis to
Growth? The Challenge of Debt and Indedtedness (Marburg : Metropolis).

Roberts, M. (2006), ‘Modelling Historical Growth: A Contribution to the Debate’ in P.

Arestis, J. McCombie and R. Vickerman (eds.), Growth and Economic Development: Essays
in Honour of A. P. Thirlwall (Cheltenham: Edward Elgar) pp. 96-115.

Roberts, M. and M. Setterfield (2013) ‘Endogenous Regional Growth : A Critical Survey’ in

M. Setterfield (ed) Handbook of Alternative Theories of Economic Growth (Cheltenham :
Edward Elgar) pp. 431-450.

Sandilands, R. (1990), Nicholas Kaldor’s Notes on Allyn Young’s LSE Lectures 1927-29,
Journal of Economic Studies, Vol. 17.

Setterfield, M. (1997), History versus Equilibrium and the Theory of Economic Growth,
Cambridge Journal of Economics, Vol. 21 pp. 365-378.

9
Thirlwall, A. P. (1980), Regional Problems are Balance of Payments Problems, Regional
Studies, Vol. 14 pp. 419-426.

Thirlwall, A. P. (1982), The Harrod Trade Multiplier and the Importance of Export Led
Growth, Pakistan Development Review, Vol. 1 pp. 1-21.

Thirlwall, A. P. (1987), Nicholas Kaldor (Brighton: Wheatsheaf Press).

Thirlwall, A. P. and R.J. Dixon (1979), ‘A Model of Export Led Growth with a Balance of
Payments Constraint’ in J. Bowers (ed.) Inflation, Development and Integration : Essays in
Honour of A.J. Brown (Leeds: Leeds University Press).

Thirlwall, A. P. (2013), Commentary on Kaldor’s 1970 Regional Growth Model, Scottish

Journal of Political Economy,

Verdoorn, P.J. (1949), Fattori che Regolano lo Sviluppo Della Produttivita del Lavoro,
L’Industria, 1, pp. 43-53.

Young, A.A. (1928), Increasing Returns and Economic Progress, Economic Journal, Vol. 38
pp. 527-542.

10