1The literature on the interrelations between urbanization and economic growth is generally confusing. Most of what is stated about this topic entails a level of abstraction and generalization that may contribute to establishing some dynamic links, but at the same time, it may conceal many other more elementary and important ones. This means that it does not help much to picture the varied aspects at stake here, which, I believe, are certainly very easy to understand. It is just necessary to assume the fact that urbanization is in itself a variable “embodied” in GDP measurement, that is, that an important portion of output and added value derives from the “building cities” stage, which inherently comprises a series of basic economic activities. As will be seen, understanding this link between urbanization and growth – ignored by current literature – is crucial to grasp economic dynamics from a macro perspective, that is, one that is both economic and historical. And it is also necessary to see how different stages of the urbanization process impact on growth, business cycles and the distribution of wealth (output) in diverse manners.

2The link between growth and urbanization is usually described in terms coming from traditional economic analysis, which obviously explains the phenomenon from the point of view of its own analitic approach. I consider that a more precise and integrated framework is necessary to explain these phenomena. Indeed when explanations depend on analytic approaches resorting to differences in productivity between urban and non-urban activities (and to income elasticity of the demand for rural goods with respect to urban goods), one may be trapped by the temptation of drawing hasty conclusions. The same happens when statements about the link between growth and urbanization are based on correlations between levels of urbanization and output per capita. This is so particularly because the indicator used in both cases inevitably includes within wealth creation the set of goods implied in the transition from a rural to an urban society.

3This is not incorrect – in any case, what is incorrect, as is well-known, is any categorical statement based on averages with high dispersion with regard to the average value. The point is that this interpretation does not say much about the possibility of future wealth growth considering that, over the last sixty or seventy years, part of this has mainly been linked to the construction of the urban lifestyle, as has been stated. Therefore, though simplified analytic models have been able to admit part of this problem, they have done so in too abstract terms. Again, what is wrong is not abstraction in itself but the shallow analysis of the phenomenon.

4I will try to clarify my previous statements by means of a brief analysis of the literature on the topic, whether quoting some fragments or providing a summary of the texts in question.

5Polèse (2001), for instance, states the following:

There is overwhelming evidence of a positive link between cities (urban areas) and economic development. Ample literature has shown the positive contribution of urban areas to national economic growth. Several studies confirm the real link between income per capita and levels of urbanization (Jones & Koné, 1996; Lemelin & Polèse, 1995; Tolley & Thomas, 1987). Other studies have repeatedly shown the disproportionate contribution of urban areas to national income and to income coming from taxpaying (Peterson, 1991; World Bank, 1991). Others have shown the positive link between productivity, population agglomeration and economic activities in cities. (Ciccone & Hall, 1996; Glaeser, 1998; Henderson, 1998; Krugman, 1991; Rauch, 1993; Quigley, 1998)1.

6Polèse’s statement is along the lines of others that start off with econometric analyses, or that emphasize a series of explanatory aspects of these correlations. This should be considered in some detail. For it is particularly relevant when taking into account that, unlike the development that took place over the last sixty or seventy years, the bulk of the urbanization still necessary on a global scale implies almost completely exhausting in the next fifteen to thirty five years the basic mechanism of rural-urban migration that has sustained this urbanization and growth process.

7Indeed, over the period 1950-2010, urban population increased by an equivalent of 64% of total population increase (which was 51% between 1950 and 1975). But the most recent projections carried out by the United Nations estimate that urban population will rise over 100% of total population growth either between 2015 and 2030 or between 2015 and 2050.

8Thus whereas world rural population grew from 1,786 million in 1950 to an estimated 3,337 million in 2010, projections indicate that it will become stable at this order of magnitude in the next thirty five years, or that it could even be reduced to some 3,000 million. In turn, urban population, which grew from 750 million in 1950 to 3,558 million in 2010, would rise to almost 5,000 million in 2030 and to 6,250 million in 2050 (United Nations, 2012). Though this is not in itself completely determining, it implies a serious change in the understanding of the dynamics of the driving forces of future growth.

9In the field of the links analyzed in this chapter, there are, then, approaches that start off with a difference in productivity between agricultural, and industrial and service activities. In some cases, urbanization is more the effect than the cause of development. In others, there is an apparently thorough analysis of different factors deriving from the observation that cities are centers of exchange of knowledge, suppliers of different types of positive externalities for firms, economies of scale and of agglomeration, among other factors. There are many arguments, but none can completely avoid the question of rural-urban migration.

10Kumar & Brianne Kober’s description (2010) of the phenomenon is in terms of structural transformation:

Firstly, low productivity economies become high productivity economies. Secondly, primarily rural economies become urbanized. Empirical evidence suggests that no country has ever reached middle-income status without a significant population shift into cities.

11Another example of what has been stated comes from a very comprehensive study of urbanization and growth carried out by the Commission on Growth and Development and published by the International Bank for Reconstruction and Development (IBRD), a member institution of the World Bank (WB), titled Urbanization and Growth (Spence, Annez & Buckley, 2009). One of the authors contributing to that compilation states:

Industry and services are concentrated in cities. These sectors grow more rapidly than other sectors, so cities must be important to growth. […] A large body of literature explains why industry and services locate in cities. (Duranton, 2009).

12Following this kind of reasoning, the key issue is the theory of the economies of agglomeration, discussed by W. R. Helsey & W. C. Strange (1990); G. Duranton & D. Puga (2003), and S. S. Rosenthal & W. C. Strange (2004). This line of thought actually goes back to Marshall’s concept (1890) of economies of scale2 and Marshall, Arrow & Romer’s approach to localization economies (Henderson, 2002). Arguments range from the motivations individuals may have to voluntarily live close to one another in spite of the costs implied in competing for land, to the reasons why industries choose to locate in one city rather than another. Individual decisions – whether those related to the labor market or to business people’s investment and localization decisions – constitute the fabric of subsequent formalizations.

13Thus, economies of scale offer efficiency and advantages for consumers, manifested in cities in several ways. Spence, Annez & Buckley (2009) state in this respect:

Process industries, such as chemicals, steel, and automobiles, operate more effectively at higher volumes; for this reason they have traditionally been established in urban areas […] Public services such as hospitals, theaters, orchestras, and sports stadiums require a critical mass of consumers to make them economically viable. (p. 13)

14In turn, the localization theory finds its justification in demonstrations of the links between productivity and market density (Ciccone & Hall, 1993).

15Other authors, such as Tolley & Thomas (1987), consider that questions regarding urbanization and economic development go hand in hand for a number of reasons, but, they say, when trying to establish a line of causality between them, the key question should be how such causality works. Along this line of thought, they hold that cities would not grow unless there were productive things for people to do in them:

In the past hundred years the enormous growth of cities occurred in the Western world as people took the fruits of development in the form of products, other than food and fiber, which could be produced more economically in cities than in the countryside.

16There is no reference here to a distinction between capital and consumer goods, even when an important part of these goods “other than food and fiber” is precisely used in the construction of urban infrastructure (houses, buildings, streets, roads, theaters, industrial facilities, harbors, airports, etc., which are intensive in cement, steel, aluminum, copper and other products, including specific machinery). Besides, the future markets for these products depend mostly on the continuity of city construction processes and of different activities that might take place in those cities, as a consequence of the need for innovation or of consumers’ incomes.

17Another important question has to do with the fact that, as urbanization progresses, a large portion of installed capital is owned by many, but it does not generate a level of wealth as annual flow, at least not the amount it generated when such capital was being built. Railways are a simple example of this: while they are being built, an annual value added is generated that is higher than that generated by its maintenance and administration. There is a change in the production function, and also a different way of generating and using the spending. This capital is not in itself reproductive investment but a condition for the reproduction of other activities.

18Reproductive activities linked to the construction stage of railroads, as well as of electricity, gas, water grids and communication networks, in turn, may be affected. Houses may be of private ownership, but they have almost no reproductive function, unless they produce income if rented, and as a consequence of a certain amount of services they may demand. But here again the value added may be lower than the activity generated along the most dynamic stage of the real-estate market. Thus only territorial expansion on a global scale of new modernization and urbanization processes may create wealth as flow as a consequence of an important number of related economic activities. Otherwise, this wealth creation depends more on the ability to find new things to do, that is, on continuous technological innovation.

19Yet this manner of creation of new wealth may not be enough to generate full employment of the resources. In mainstream economics, the problem of not fully employing resources is usually attributed to market imbalances, which may be adjusted by freeing all the barriers preventing each factor from reaching a remuneration according to its marginal productivity. Such an approach tends to derive the cause for that situation from some type of market failure. A Keynesian analysis would solve this by means of monetary and fiscal policies promoting public and private investment, whereas the Marxists view this as a crisis of capital accumulation (under-consumption, realization crisis, and the like). Only in the Schumpeterian evolutionary approach does the technological rigidity deriving from a lifestyle (production, consumption, exchange) have a central role.

20But even here, the link to urbanization is not considered central. The rigidity of the general structure of productive supply – and its powerful influence on the world of political decisions – may have been emphasized by John Kenneth Galbraith3 in relation to the military-industrial complex. Or it may permeate the pragmatic thought of some lobbyist groups (as is the case with several initiatives of infrastructure construction coming from dominant sectors of a particular industry related to the massive use of cement, for instance). However, no systematic effort has been made to link these issues with the prevailing technological style related to massive urbanization.

21Yet, the literature on the different interrelations between the construction activity and economic growth – typical of the first stage of sustained growth discontinuation over the Golden Age (1950-1970) – already shows evidence of this, which seems to have been disregarded in recent times. Thus, for instance, a document by the Massachusetts Institute of Technology (Moavenzadeh & Koch Rossow, 1975) states the following:

A major objective of development is economic growth. During the early developmental process economic growth seems to be generally high, and construction, along with manufacturing, tends to play an increasingly important role in the economy while agriculture’s importance declines; once a relatively high level of development is achieved, however, economic growth appears to slow, and construction’s role tends to stabilize or even decline slightly while the other sectors continue as before. While construction’s direct contributions to development are significant, it also stimulates a sizeable amount of economic growth through backward and forward linkages. Construction’s requirements for goods and services from other industries are considerable; the development of the construction industry therefore stimulates these ancillary industries, thus encouraging further economic growth. In developing countries the construction of physical facilities makes up more than one half of gross domestic investment and tends to be concentrated on basic infrastructure in agriculture, mining, transportation, communication, and utilities.

22But, as has been said, the construction industry uses machinery and equipment such as concrete-mixing and other types of trucks, cranes, backhoes, and the like, and it demands for steel, cement, paint, chemical products, glass, ceramics, cables, switches, pipes, etc. Total installed capacity in these industries depends, eventually, on the size of the global market.

23The construction market may, as pointed out in the document quoted above, slow down as more mature stages are reached – and in turn, the industry supplying equipment and input may be left with an important amount of idle capacity much before that decline is even noticed –, or it may require less labor as its dynamics decreases. None of these activities will easily be reconverted, whether physical or human capital is involved.

24The assumption that the process can always be replicated is as strong as the belief that, once traditional markets have become exhausted in basic industries, such industries can sustain their level of activity by means of the continuous creation of new products. This will be dealt with below, but only by way of example here, the steel industry kept producing 736 million metric tons between 1979 and 1999, with only a 5% variability in twenty years (0.2% annual growth), whereas its dynamics between 1950 and 1978 implied multiplying production by 3.8 in 28 years (an annual accumulated rate of 4.9%). Interestingly enough, such production increased again almost twofold between 1999 and 2012 (an annual 5.3%), most likely as the consequence of Asia’s fast urbanization process. In 2002, steel production was 950 million tons a year, with China accounting for 20.1% of apparent consumption, whereas the rest of Asia accounted for 11.6%, very much like Japan. In 2012, apparent consumption rose to 1547 million tons, with China accounting for 46.3% and the rest of Asia, not including China and Japan, for 12.2%. There was also a reduction in the share of the North American Free Trade Agreement (NAFTA) bloc, the European Union and Japan, which even implies a reduction in absolute value of apparent consumption in view of a fourfold increase in China’s market (World Steel in Figures 2013, Worldsteel Association).

25But the continuity or possible increase in such production and consumption levels depends, not only on the size of any new urbanization to come, but on whether it will or will not imply the need for extending already installed capacity. Many basic products such as cement and aluminum have shown patterns of this type: highly dynamic during the urbanization periods 1950-1975 and 1995/2000-2012, and of relative stagnation between 1979 and 1999.

26Accompanying this pattern, the world agricultural production grew by an annual 2.1% over this last period, an expected impact since, despite the low income elasticity of agricultural commodities, more people living in cities have more food needs. This boosts production to levels that, for long periods, may be higher than those of other goods in turn affected by the behavior of other industries for which they are an input, as is the case with steel. No doubt this also reflects technological changes and cross impacts from other sectors. Part of the steel demand may have been replaced with other materials, or the demand for steel flats may have decreased when, for a while, the automotive industry tended to produce smaller cars as a first step towards achieving greater efficiency in the use of fuels as a consequence of the 1974 and 1979 oil crises (Aubernon, 2013)4.

27The analytic approach adopted by the literature on this topic may study the correlation between the use of steel and the level of income per capita (Warell & Olsson, 2009). But even in this case the link with urbanization is not explicit, precisely because of the use of synthetic indicators that ignore the fact that many of the variables embodied in GDP have to do with urbanization in an implicit manner, not as an activity in itself, even when, strictly speaking, urbanization is not an economic activity.

28Thus, it is true that empirical evidence at an aggregate level suggests a continuing increase in product per capita and larger urban than rural productivity (Figure 1). But the fall in urban productivity between 1980 and 1985 should trigger as much concern as the fact that the relation between both productivities reveals growth and decline cycles which are not independent of the urbanization process and its stages.

Figure 1. Trends in value added for industries and services per urban resident and in agricultural value added per rural resident on a global scale: period 1970-2012

29When considering that, for instance, the world production of precious stones (gems) grew by a factor of 23 between 1955 and 2012 and that of steel only by 5.7, it can be better understood that the value added above (e.g., “industries and services” value added) probably shows very diverging trends in things produced which, in turn, depend largely on different stages of urbanization. So, whereas an important part of the steel production is included in a large amount of urban public and private infrastructure, gems reflect preferences – and surely enough, income levels deriving or not from work – that appeared within urban markets alongside their development.

30As will be seen below, long terms of near stagnation of industrial output have been parallel to stages during which urban population along several five-year periods could not exceed a certain number of people, so no new capacities needed to be created in industries related to the construction of the urban lifestyle. In turn, the recent urbanization of Asia implied a growth of cement and steel industries which, in ten years, exceeded the increases in those industries along the two periods prior to the last, though none of them showed the dynamics they had between 1950 and 1970.5

31This can be seen, for instance, when comparing growth rates for food and mining products over periods such as 1961-2012 (Figure 2) with respect to the 1975-1990 rates (Figure 3). In the first case (including the period 1975-1999) there are two major waves of urbanization and in the second, there is an interlude of smaller relative growth of incremental urban population. In particular, over the 1975-1985 decade, this growth slows down with respect to the 1950-1970 trend.

Figure 2. Growth rate of world production of agricultural commodities and of metal and non-metallic mineral products between 1961 and 2012

Figure 3. Growth rate of world production of agricultural commodities and of metal and non-metallic mineral products between 1975 and 1990

32Comparison of these data reveals that mineral commodities such as cement, copper and steel were more dynamic between 1961 and 2012 than between 1975 and 1990, when several agricultural commodities grew well over these basic products of the heavy industry. This, I insist, is closely connected to the stage of construction of the large urban centers and their interconnections, such as systems of cities and regions and their roads (Cf. Ausebel & Herman, 1998).

33On the other hand, there is the deep-rooted tradition of considering that it is GDP growth that drives both urbanization and demand for those products, as if they were not a part of this index. For instance, a paper on the United States Geological Survey (USGS) webpage states that “the expansion or contraction of gross domestic product (GDP) may be considered a predictor of the health of the steelmaking and steel manufacturing industries, worldwide and domestically” (2014).

34In fact, these two variables (GDP and the steelmaking industry) show a level of correlation close to one (perfect correlation), whereas the autocorrelation is so high that it invalidates the model. So this means that the model is not explanatory. GDP would fall if the demand for steel and its commodities fell without other commodities generating an equivalent value added. The statistical test of correlation between steel production and GDP actually displays values that reveal the autocorrelation of these variables, as is to be expected with any component determining GDP level. The same is true, for instance, of series of steel, copper, aluminum production, and of energy consumption correlated with GDP6. On the contrary, explanatory variables of the production of these commodities are different and more complex and, all in all, determine the level of wealth created over a year.

35The question is usually confusing precisely because all the commodities produced are variables embodied in the measurement of wealth creation (GDP). Certainly, the conditions for their reproduction, whether historical, social, technological, economic, financial, political or any other, determine, in turn, their future demand (size of the market), cost-effectiveness expectations, decisions regarding private and public investment and the distribution of that wealth. Yet the previous level of urban population growth seems to be a future growth factor for some time, because in the meantime, precisely the construction of the city and its equipment is an important activity. As this large market decreases, though, the demand for consumer goods will depend on the income distributed as from the creation of that wealth in the cities already built. Now, “already built” should not be taken literally, of course, for construction, remodeling, etc. are always in progress; it should rather be understood as a mature and consolidated stage of the city and of the system of cities on a global scale.

36The case of China, first world producer of cement (Soule, Logan & Stewart, 2014), is certainly forceful. In 1980 its cement production was almost 100 ton/capita, the world average being 198 ton/capita; in 1990, it rose to 184 ton/capita, only 4% lower than the world average. This was mostly the consequence of a policy aiming to provide rural residents with houses, for which cement was produced at relatively low qualities (grade 325) in a large number of plants. In 2000, production markedly rose to 457, 68% above the world average, the industry was restructured and higher quality cement was produced (grades 425 and 525). But between 1980 and 1990, the population living in large cities totaled only 33.6 million, whereas between 1990 and 2000 it reached 100.8 million and then, between 2000 and 2012, it grew by 93 million. Will this be related to the fact that between 2000 and 2012 cement production grew by an annual rate higher than 7%, whereas between 1980 and 1992 – another 12-year period – it only grew by an annual 2%? The answer is obviously affirmative, particularly when considering that such growth in population implied also the construction of buildings, streets, roads, ports, airports, factories and hydro-electrical plants. All this means an increase both in annual wealth and in stock of private and social wealth, for such stock is shared by the people living in the cities. However, part of that annual wealth will not go on reproducing new wealth with the same dynamics, unless the processes involved in the growth of large cities never stopped or slowed down, which is obviously not the case. An alternative to this could be that the value added of other things to do exceed that of what is no longer done or produced, but empirical evidence shows that this is not so easy or replicable on a world scale. In fact, the end of the golden age and of the North-South convergence was followed by marked structural changes on the world and political scale. It now seems that such cycle was replicated by the Asian urbanization peaks that took place before the 2008 crisis, not to mention the changes in the period 2014-2015, which again seemed to jeopardize the South.

37As was previously stated, theoretical and analytic approaches such as Tolley’s, start by considering supply and demand factors, where, on the demand side, the low income elasticity of agricultural commodities reveals that people spend a growing proportion of their income, as this grows, in urban goods. The cause (of the urbanization process) should be the low income elasticity of agricultural commodities. But we have provided conclusive evidence that this statement is neither so definite nor a universal law. Precisely between 1975 and 1990, income elasticity of some agricultural commodities exceeded that of some very important non-agricultural ones. Then the statement is consonant with the data in Figure 1, but not with that in Figure 3, at least not in a conclusive manner on a global scale. It is then clear that the low income elasticity of agricultural commodities is an elusive concept, so much so because the consumption of these commodities in urban environments implies the use of many other products and services.

38Thus, bananas produced in Ecuador will require cardboard boxes and paper, tags, scanners, storage centers and containers for exportation – all of which are manufactured from different inputs and involve the use of machinery. Means of transport, fuel, roads, ports, airports, food inspection, Customs and other services are also needed. In turn, all of this is part of GDP as value added, but also of the employment required for manufacturing and its remuneration, which is related to the demand for goods and services, that is, the most important part of aggregate demand in the Keynesian sense.

39If it is necessary to explain such elementary facts – typical descriptions of seventeenth century scientific style –, it is mainly because the economic jargon has become somehow obtuse, circular, tautological and too axiomatic, giving rise, in turn, to dangerous misunderstandings and worse public policies. Lack of clear knowledge of the composition of GDP (and of annual wealth) and of what products may have life cycles that might trigger real depressions – in the absence of others that could replace them –, amounts to believing that the creativity of individual companies and their desire to grow and survive in the market are enough to avoid long depressions with unimaginable impacts on twenty-first century history. The analysis of the 2009 recession is good evidence of this (Irons, 2009).

40Going back to Tolley’s analysis, on the supply side, traditional explanations have remarked that the rapid growth in agricultural production – in turn the consequence of technological advance – is the main cause for the increase in labor supply in urban areas. However, a factor to be considered, says the author, is land scarcity, which, he warns, could trigger a decrease in agricultural production per worker. If such decrease is permanent, it could delay urbanization processes.

41The author says that evidence of such trend in the case of China, by far the most important, is not clear enough. He points out that there have been processes of expansion of the agricultural frontier in other continents and suggests that, despite the growth of agricultural productivity per worker, low elasticity of the demand for agricultural commodities could keep putting pressure on the increase in urban labor supply. But besides, he rightly considers, most of the input and machinery implied in technological progress in the field of agriculture is in turn produced in the cities.

42Tolley then refers to a typical question in the literature on urbanization and growth, which is the formation of human capital, and also to the reasons why an increase in productivity in urban areas could stop urbanization. After developing his analytical model, he concludes that the level of a nation’s ability to boost urban productivity would mostly determine its urbanization. Population growth would still be the source of urban and non-urban growth, but countries falling behind in their increase in agricultural productivity would experience growing urbanization pressure. Hence, he concludes, a better understanding of the economic factors influencing the efficiency of policies related to urbanization is necessary. Such factors include, in particular, the industrial structure, the conditions affecting international trade, the availability of greater detail on types of goods, knowledge of how certain price fixing affects the economy, the composition of the age pyramid, the level of education, the level of dependency and other aspects related to the job supply market. Finally, he says: “The assumption that the number of workers in urban areas goes hand in hand with the population of these areas would be modified”.

43Now, he exactly identifies urban unemployment as the most important concern, mainly because it is the first cause of urban poverty, radically different from rural poverty. But enumerating aspects such as those mentioned above (the industrial structure, international trade, details on the types of goods produced) actually conceals the main issue. That is, the possibility of finding, after the consolidation of urban systems, enough activities equivalent in value to substitute for those that were necessary along the stage of construction of those systems, in terms that will not affect the lifestyles and conditions of historical and inter-generational transition of a large number of people around the planet.

44Believing that every city in the world may reach levels of enough productive diversification to generate employment amounts to disregarding the geography of specialization and the economies of scale. For such differentiation is what makes it possible for some cities to be prosperous and for others to have saturated markets.

45The arguments put forward in Chapter V of the compilation on urbanization and growth (Spence, Annez & Buckley, 2009) illustrate such space inequality in urban systems. Its relation to economic development, as well as the absence of both theoretical and descriptive models, and of clear policies to deal with these phenomena are also explained there.

46This literature dangerously confuses case analyses – and general recipes – with the problem considered on a global scale. Tolley (1987) says that “cities would not grow unless there were productive things for people to do in them”. This seems to overlook the very important fact that precisely some of the “productive things” to do in the cities – in most developing countries – are related to the construction of urban infrastructure, such as communication networks through ports, airports, railroads, energy transmission grids, houses, factories and facilities whose lifespan is longer than that of consumer goods and services. Such omission derives from the fact that his approach is non-systemic and, above all, non-evolutionary.

47In developing countries the portion of labor employed in the construction sector and in the industries and services related to it may be very important (ILO, 2001)7, whereas the conditions of competitiveness for traditional or innovative industries are low. And this is so because certain manufactures cannot compete with those from the United States, Europe or China, or because the level of accumulated technological development and human and technological capital in developed countries cannot be replicated or reached on a global scale.

48In this context, China, as the factory of the world, managed to produce a growing amount of consumer durable, semi-durable and nondurable goods, whereas its urbanization process was completely dependent on the growth of developed nations until 2007-2008 (before the international “financial” crisis). For countries producing raw materials, this joint growth meant a rise in the demand for all commodities and their price, which in turn boosted industries, trade and services. This would not have happened in these countries but for a surplus in foreign trade, and not precisely of raw materials.

49However, the “north-south displacement of production” – as the phenomenon is usually referred to by economists and analysts – started to generate employment shortages even in central economies once the expansion cycle caused by China’s urban population growth before its accession to the WTO8 became exhausted. In turn, the future growth of an important number of developing countries is now in check. In developed ones the outlook may be different because of the income from foreign capital, interests, transfer of profits, patents and royalties coming from transnational companies with headquarters in these countries, but with production markets in other countries and consumption markets around the world. All of this clearly allows for the comfortable handling of the public budget. This is dealt with in Chapter IV about the specific nature of technological innovation and its links with growth and with the creation and distribution of wealth.

50The crucial question should then be: what happens with the capacity of the system to produce equivalent numbers of productive things to do in the cities to substitute for those that get inevitably exhausted or lose dynamics as the urbanization process gets saturated or tends to that?

51In previous work (Kozulj, 2001; 2005) this question was considered central. Peaks in population growth along fifteen-year periods are regarded as determining the maximum capacity needed to cater for the demand for capital goods for an important number of activities. Since such a situation cannot likely be foreseen, overcapacity crises usually manifest later, that is, when successive growth of new urban population turns out to be lower than the maximum reached previously.

52This question may not be important for China or for nations leading in technological innovation as yet, but it is crucial for many European countries and for Latin American, African and many Asian ones. In the long – and not so long – run, it is crucial for everyone to notice that, according to the 2014 WTO report, the employment gap after the 2009 crisis affects a total of between 60 and 80 million people, adding up to the 200 million who are nowadays unemployed plus those who, though not unemployed, have not been able to cater for the fulfillment of a minimum level of basic needs in urban areas. This totals some 700 million people, according to the United Nations Population Fund (1996)9 and the ILO (2014).

53What this is actually about is a change in the rules of the rural-urban social contract implied in the logics of urbanization. The fact that masses of people were attracted to the cities at an early stage of urbanization supposes that they migrated without feeling the need to be innovative business people but workers, and not even skilled ones. At a later stage they found themselves depending on their own inventiveness or on the capacity of the system to fulfill the promises that there would always be productive things to do in the cities. A similar assumption (that is, that there will always be productive things to do) is probably held by 2006 Nobel Prize Muhammad Yunus regarding his “bank for the poor” initiative, as he believes that the youth have the responsibility of being creative. Which poses the following question: is this not too heavy a burden for those who must become members of the adult world without the benefits of a material and cultural background from the societies built over the last seventy years?

54The situation of ample sectors of the urban population working in the so called informal economy adds to the difficulty of providing most of these people with services and dwelling. This, in turn, is another factor impacting on the use of already installed production capacities, among them those related to the construction (Arnott, in Spence et al., 2009).

55In 2003, Kofi Annan, at the time Secretary-General of the United Nations, revealed alarming information. Around a thousand million urban residents, most of them in developing countries, lived in slums. This meant around 32% of the world urban population. The projection warned that in the following thirty years that figure might reach two thousand million people (UN-Habitat, 2003) if measures were not taken jointly by national and municipal authorities, the civil society and international bodies. And there is serious evidence that the conditions of urban poverty are in many cases worse than those of rural poverty (UN-Habitat, 2007).

56Then we can go back to the question posed at the beginning regarding whether the poor are nowadays poorer than those living several years ago. I find this more important than Piketty’s view (2014) that today’s average income greatly exceeds that of the eighteenth century and that 60% of the richest people nowadays live well above the richest 5 or 10% two centuries ago.

57It is not an easy estimate, for certainly 2,400 million people are much wealthier than their fellow people at that time, and some 1,000 million probably are not, or at least a fraction of them certainly are not. Anyway, there is no denying that this is closely linked to the logics of wealth creation deriving from the technological and productive style of the post-Second World War period regarding urbanization and the fact that it is a variable embodied in wealth as stock and as flow.

58This will be dealt with again below, but it is necessary to point it out here. Otherwise, some basic elements of the problem might be overlooked10, since a number of assumptions could be made: that there are endless investment options at all times; that it is possible to transform financial surplus in real investment; that it is possible to exchange labor across different activities – that is, that human capital is never destroyed; that new urban goods and services can always be created; that the existence of instant – and sometimes even spontaneous – technological transitions is desirable because they renew the flow of goods; that individuals have unlimited capacity to use new goods and services; that natural resources are limitless or always renewable; that physical capital is flexible enough to be used in different activities; that prior historical experience is irrelevant.

59For GDP to be a little higher each year, it must be assumed that the total goods and services produced will increase so that the decrease in amount (value) of some of them will be compensated by an increase in the amount (value) of other goods and services. If production in some market in particular decreases and that value is not compensated by an equal value in some other market, the total output will decrease.

60In fact, the reduction in the proportion of gross fixed capital formation with respect to the total annual product is normally compensated by an increase in the proportion of the national income devoted to consumption. Now for this to occur, it is necessary that consumers have enough income, which in turn will come from the share of the agents in the creation of wealth and from the rules by which such wealth is created and distributed.

61The economic literature has always assumed that factor incomes depend on their contribution to marginal productivity and that, in any case, the State determines their distribution by means of fiscal and tax policies. But this simplification again overlooks the fact that an important portion of the profitable gross fixed capital formation (GFCF) is in a strong correlation with the construction of urbanization, or better still, with the productive capacity associated to this process. If this productive capacity becomes exhausted, the new GFCF may refer to substitutions of obsolete physical capital or to investment devoted to the creation of new goods. Public investment depends on the limit to which public spending may be financed, that is, on Keynesian policies and on the limits of their efficiency, which, in turn, depend on the creation of wealth by the private sector. Regarding the new goods, in order to recreate the flow of wealth production, a process of planned obsolescence is usually forced, and if the function of the new goods is not very different from that of the goods they are replacing, formation of the supply price may be affected as to the share of work and capital in them. This will imply a bias towards a greater incidence of capital only because it must be recovered in a shorter term11.

62If this reduces the share of the labor factor in national income, investment in consumer goods will probably be lower, and the general return on capital will not improve either. This may remain unnoticed for in such a system, the gap between the creation of value and the act of estimating that value (goods and services) in monetary terms will be wider as opportunities to convert savings in physical investment become fewer.

63It is significant, I insist, that there is no explicit reference to the fact that an important portion of the product generated by urbanization is clearly linked to the construction of cities over the time it takes to do so and for whatever reasons it may entail, while in turn such process comprises internal migrations from rural to urban areas. This last topic, however, is included in a different branch of the literature on urbanization and growth, as was postulated by the so called Harris-Todaro model around 1970.

64In his contribution to the question of urbanization and growth, Gilles Duranton (cf. Spence et al., 2009, Chapter 3) wonders whether the cities are engines of growth and prosperity for developing countries12. He uses an integrated theoretical model featuring different types of cities, some of which are engines of growth and some of which are not. His basic analytic tool uses, in the first place, the wage curve, which grows as the size of the cities and of the product generated in them grows. He explains the reasons for this by means of a series of typical arguments such as the indivisible nature of the cost of large infrastructure, specialization earnings, risk reduction, factors related to the concept of economies of scale, among others. Secondly, the cost of living curve, which grows faster as from a certain size of the city, thus correlating with the cost of houses, prices of consumer goods, cost of exchanges – in itself growing because it is the consequence of the larger size of the city in question. Thirdly, the net wage curve, defined as different from the labor supply curve, as it results from the difference between the wage curve and the cost of living curve according to the size of the city. Next, the labor supply curve, which may adopt different shapes according to the nature of labor mobility. And finally, equilibrium conditions between labor supply curves and net wage curves to be able to formulate different policies according to the behavior of the type of city in question. This tool is then used in order to analyze and recommend suitable policies to improve the purportedly spontaneous links between urbanization and growth.

65The question of whether cities drive economic efficiency is then answered affirmatively by the author (if the analysis is static), and he adds that there is no evidence that cities (urbanization) systematically hurt particular groups”13. Yet his answer is elusive when he wonders about the dynamic benefits of cities, that is, when the evolution of urbanization is studied along the years and the inevitable situation of urban poverty and unemployment has to be faced as coexisting with opulence in many developing countries (that is, those other than developed or first world countries). Here, however, the arguments focus on the link between the largest city and other smaller ones in a nation, since a disproportionately large city with respect to the rest may bring about “diseconomies of scale and negative externalities”. He admits that, though empirical evidence is not conclusive, it is still enough to lay the ground for radical political initiatives, though not to the point of taking a stand that will limit or discourage labor mobility or reject fast urbanization processes. Duranton’s recommendation is, then, to prevent the imbalances that urbanization processes might bring about, specifically focusing on the reallocation of productive factors and activities across cities, rather than trying to stop or slow down such processes.

66The Marxist thought will no doubt see such recommendations as simply aiming to sustain the growth of a “reserve force” that will increase the labor force exploitation rate, particularly if such suggestions come from World Bank analysts or from other bodies or institutions supporting at all costs a certain identity between free market and democracy. That might certainly be part of the objective or the rationale of the system, but because of the arguments that will be suggested below, this reading of reality or discourse intention would be an oversimplification. The same would be the case if the links between growth and urbanization were treated on the basis of too gross arguments to formulate policies that should be global but at the same time with clear targets.

67Indeed, part of the recent controversies about and renewed interest in the links between urbanization and growth are not outside the scope of the large-scale transformations that Asian countries – particularly China and India – are now undergoing, with a strong impact on world economy. And another reason for this interest is the fact that in certain countries, rural-urban migration may be under control, limited or not allowed, whereas fast urbanization processes may bring about huge challenges to fully incorporate those rural migrants not finding a job into modernity.

68The so called Harris-Todaro model (1970) has certainly brought about much controversy for it considers that it is unreasonable for rural-urban migration processes to continue coexisting with high levels of urban unemployment and marginality. This model poses that decisions to migrate from rural to urban areas are made on the basis of the migrants’ expectations to get a higher income in the latter, and not only on the basis of the differences between rural and urban wages. Thus these expectations may exist even under urban unemployment conditions. This paradox is solved by finding the equilibrium condition in which the minimum urban wage multiplied by the quotient between the number of urban labor positions available and the total urban workers seeking jobs equals the rural salary, which in turn represents the marginal productivity of rural labor. Then migration expectations might be deterred by a context in which the unemployment rate was so high that, multiplied by the minimum urban wage it would be lower than a minimum rural wage. This analysis is obviously static (though replicable in different periods) for it does not consider intergenerational consequences, or loses or gains accruing from knowledge and habits of the individuals and their families along periods of time (a concept that would be equivalent to that of destruction of human capital in the agricultural sector).

69Tolley and other authors have seen this more clearly. On the other hand, the Harris-Todaro model has also been criticized from another more theoretical point of view, for it considers that rural migrants are willing to take risks, when this could well not be the case. But even this condition could be included in the model without changing the type of reasoning that accounts for the market forces driving urbanization processes in current literature. In this case, the incentive perceived individually is an improvement on the income situation, whereas labor demand comes from the combination of factors and activities typical of city life.

70In the last analysis, the Harris-Todaro model and its substitutes again stress the different productivity levels in urban and rural areas. They implicitly assume that industry and services produce greater wealth per individual than a rural society can create, which is obvious since, otherwise, there would not have been any driving forces towards modernization. And it is even more obvious when irreversibility conditions are considered for each short-term technological stage. For it is well known that the urbanization process is accompanied by constant improvements on agricultural productivity by means of the introduction of inputs, machinery, stretches of land, and by the organization of the productive process. All of this reduces subsistence possibilities for those who might hypothetically want to go back to their native rural areas to take up again their former lifestyle. This will be dealt with again below, for it has to do with an issue that has arisen in answer to the rejection of the urban lifestyle and the always prevailing wish in western population to go back to an immaculate nature14.

71There is a hidden aspect here, however, related to the fact that most of the work carried out by rural migrants in the first stage of urban growth has to do with the construction of that lifestyle. Once this is established, the markets that absorbed all the initial labor supply cease to be dynamic and can even decrease in absolute magnitude. Their substitution with another set of activities is far from viable, even more if the question is approached from an overall view of the countries making up world economy (Kozulj, 2001).

72In this respect, the link between growth and urbanization developed by Gilles Duranton (cf. Spence et al., 2009, chapter 3) admits that a key limitation of this analytic model is that each city is viewed as “an island of growth” which could generate economic growth by itself. Yet it is essential to understand how information, factors and goods flow across cities in a nation and, though not dealt with by the author here, also between developed and developing countries, a topic for which he quotes Keller (2004).

73This means that the likely impact of flow and mobility across cities is now added to the so called endogenous growth factors, which were the focus of all explanations about cities as centers of human capital and knowledge, always increasing on account of intergenerational processes leading to innovation (Jacobs, 1969; Lucas, 1998; Eaton & Eckstein, 1997; Rossi-Hansbert & Wright, 2007). In developing this issue, the author quotes Duranton & Puga (2001) and points out that modern cities may be classified into two groups: those incubating innovations (nursery cities), which in turn present a great diversity of productive structures, and cities specializing in a set of particular products. He also notes that developing countries have seen, over the last fifty years, processes of separation between business centers – including firms’ headquarters and business services – and manufacturing and productive cities where industries locate (Duranton & Puga, 2005).

74The author draws very interesting conclusions from this analysis regarding the fact that policies not favoring a smaller growth of the mega-cities in favor of a more important growth of secondary cities slow down overall growth, for they deprive nursery cities from the efficiency achieved by the concentration of scientists, engineers, technicians, etc. This, in turn, stops innovation, which implies a lack of new ideas, new products and new production processes to be transmitted to secondary cities where the industries producing such new goods are located.

75He adds that it is difficult to find conclusive empirical evidence, though he mentions the examples of Seoul, in South Korea, and Brazil, which seem to be following a similar path, that of industrial deconcentration towards secondary cities. He infers that such an interpretation of the question could lead to the temptation to forcibly create cities as centers of excellence (or incubating cities), as if they could be successfully replicated on the basis of a certain kind of alchemy.

76It is certainly impossible to disagree with such a point of view considering the number of influential people who possibly imagine a world riddled with replicas of Boston or of enclaves of the “Silicon Valley” type in small cities in developing countries. It is also necessary to consider that large cities such as Tokyo, Paris, London, New York, Munich and Amsterdam are within the top ten, out of the hundred most innovative in the world. In 2010, in fact, there were 634 cities with over 750 thousand inhabitants in the world.

77Figure 4 shows the trend in urban population classified into four large groups: a) 1 Nexus innovation category (35 cities); b) 2 Hub innovation category (65 cities); c) rest of large cities (534 cities with over 750 thousand inhabitants); d) rest of the world urban population.

Figure 4. Trend in total urban population by types of cities and their size. Historical data for 1950-2010 and projected data to the year 2025 in thousands of people

78It is curious to see that, in 1950, the hundred cities nowadays considered best positioned in technological innovation accounted for 44% of the population of cities with over 750 thousand inhabitants, and in 2010 that figure fell to 25%, and to only 11% of the urban population. While these cities grew almost linearly, the rest did it exponentially. It is almost obvious that the thirty five innovation cities classified as 1 Nexus (which are central to several segments of technological and social innovation) and those classified as 2 Hub (central to innovation in very specific segments and industries) produce prototypes that are consumed and sometimes generated in the rest of the cities and urban areas.

79The fact is that before and also after the war, these cities developed all types of industries. This is why their human capital was able to learn and produce key technological transitions to survive, and thus make it possible for the growth of these cities not to depend only on their own markets, which would not be able to sustain such growth. Proof of it is, for instance, that three fourths of the industrial plants existing in the United States in 1972 had closed down by 1992, and more than half of the industrial workers in 1992 worked in plants that did not exist back in 1972 [Dumais et al. (1997), in Duranton & Puga (2001)].

80Eight of the innovation cities classified as 1 Nexus are in the United States, three in Canada, two in the United Kingdom, five in Germany, others in Europe’s most advanced nations, such as France, Sweden, Denmark, Belgium, The Netherlands and Austria, and two in Japan. Almost all these countries took part in the Second World War and had developed industrial systems at the beginning of the twentieth century. Very few of these leading innovation cities are outside Europe, the United States and Japan – only Seoul, Hong Kong, Sydney, Melbourne, Tel Aviv, Dubai and Singapore.

81Interestingly, the fast decline in the participation of these hundred innovation cities divided into one third and two thirds – as 1 Nexus and 2 Hub respectively – started to take place along the end of France’s “thirty glorious years” or the “golden age of capitalism” (and of European communism). It is also clear that for them to keep having technological returns and for innovation to pay off commercially, there must be a growing market of labor force and consumers. But this does not take away from the fact that an important portion of industry value added on a global scale still depends on the expansion of the urban market as embodied within the world scale product. This will be dealt with below, when examining the role of innovation.

82It would then be a dangerous illusion to try to replicate innovation as the driving force of growth in each city, which in turn made use of economies of agglomeration and of scale. It would be impossible for all cities, or even for the most important ones around the world, to have the necessary human capital, laboratories and sophisticated equipment. This would imply that those countries that have based their growth on the exploitation of natural resources and have boosted urban growth and consumption in an attempt to attract capital to develop their different imitative industrialization modes would be at a disadvantage in this economic world situation. For ours is a context in which wealth created year after year is the key variable to explain employment, income and life standards, and where “the productive things to do” must be reinvented for existing ones do not have solvent markets.

83It is true that this redistribution of income is a key element, but it requires a source of income and also a social contract rendering this situation understandable (or else imposing it). Another problem could then be for the dominant classes to be purely extractive. This has to do with the purpose of social spending, which contributors have to agree to finance – whether they like it or not.

84The intention of extending the health insurance system in the United States and its rejection is a clear example of this problem which, around the world, takes different forms and applies to different fields – education, infrastructure maintenance, health, justice, etc. That this still remains unnoticed is just natural for the main driver of development – urbanization and many of its associated industries –, is still growing. Of course, without profitable markets in developing countries, life in first world and innovation cities will no longer be so simple.

85A second line of thought analyzed by Duranton (Spence et al., 2009, chapter 3) deals with the relevant aspect of the diffusion of knowledge among workers. This factor had already been analyzed by Alfred Marshall (1890) as the importance of human capital as a positive externality offered by cities. Duranton quotes vast bibliography about the fact that technological progress is associated with mobility of the most skilled labor force across firms15. It is curious, in passing, that no comment be made on labor mobility within firms, which, sooner or later, usually also implies a subtle trick for wage reduction ending in dismissals. This issue has recently been studied by Huttunen, Møen & Salvanes (2011).

86It can be inferred from the arguments put forward so far that the lack of – especially highly skilled – labor force mobility between a main city (the primate city) and secondary cities may be an important factor determining the continuity of the primacy of a city, as well as the falling behind of secondary cities. But this does not add much to the argument, since this is a self-defined dichotomy. Most probably, this highly-skilled labor force does not move from the main city to a secondary one because there would not be labor opportunities there. And in turn, firms do not locate in secondary cities because there are no positive externalities there, which they do find in main cities.

87Aware of this, the author sees that diffusion of technologies and knowledge is a more powerful factor when it is related to international trade and the countries’ level of opening. From this he infers that it is possible to obtain larger profits from productivity by eliminating trade barriers in developing countries. But he obviously does not discuss what it is that these cities located around the world could exchange, for that would require research into the real possibilities and limits to diversify production on a global scale. That is, if many of China’s exports are more competitive than those from the United States and the European Union, what is to be expected from intermediate and poorer countries?

88Finally, Duranton concludes that there is strong evidence that the increase in productivity is linked to processes of creation and destruction at the firm level. Summing up, he insists on certain simplifications: “resources need to flow from less productive to more productive firms and allow new entrants to rise and challenge incumbents”16. Though he does not go deeper into this topic, he says that in developed countries there is a strong spatial dimension to processes of industrial reallocation as technologies evolve. Therefore, whatever hinders the movement of factors across firms in a city or even across different cities may have high dynamic costs in the long term.

89He suggests several policy recommendations, such as eliminating favoritism towards primate cities; improving urban efficiency in order to lower the cost of living as a way of dealing with urban crowding and the supply of public services; eliminating biases leading to squatter settlements by means of a reasonable titling policy and urban deregulation. He also suggests improving access to the markets across cities by means of the development of transport infrastructure and lowering trade barriers or impediments, and not discouraging internal migration, for it fosters an efficient allocation of the population and has equalizing effects across places. Finally, he holds that, though these recommendations are consistent with already existing urban policies, the difference is that there is emphasis on the labor mobility factor (cf. Spence et al., 2009, chapter 3, p. 106).

90What this analysis lacks, I believe, is rigor. Labor mobility across cities is produced naturally, and in the case of generalized unemployment, what this process does is simply transfer the problem from one place to another. Squatter settlements may be present in the megalopolises and also spread to other smaller cities around a country. This happens in Argentina, for instance, or is what the Spaniards now fear as a consequence of the Greek crisis, just to consider two very topical issues as of the 2009 crisis to date, not to mention the question of migration into the European Union or that of the Latin – particularly Mexican – into the United States. Obviously, such mobility creates certain additional demand in a new urban space, but this impact cannot in itself justify such a statement, for it is not clear which of the wealth creation mechanisms is affected by favoring labor mobility.

91The central problem is still avoided because it cannot be admitted that sometimes there are not enough “productive things to do” with a proper market – and even less so on a global scale –, or that, in any case, those productive things have to be done by the State if there are no economically viable private initiatives. But the fact that they are done by the State is in no way neutral. The question naturally arising would be what the maximum public budget is so that there is not a negative impact on growth and who is willing to pay for it. Though there is some literature on the topic, it is evident that the willingness to pay for public spending is not the same across societies, and that collective learning is a slow process, even more so when each wrong approach is also part of the problem, which I shall call “the ecology of ideas”. A clear example of this is the debates on the Welfare State in Europe, or on adjustment policies in countries such as Spain, Greece, Italy, Portugal, and others in the EU 27.

92Now if the emphasis is not on the link between the capacity of the system to generate a growing flow of wealth and employment, reality itself takes care of showing the consequences. Indeed, the links between economic crises and increase in crime are being monitored by the United Nations (UNOCD, 2014). Some specific institutions are stressing the fact that the lack of “things to do” particularly affects young people. The ILO 2012 Working Paper on unemployment held in the Preface:

First, the youth unemployment rate is typically twice the adult unemployment across low, middle and high-income countries. Second, youth employment is much more sensitive to business cycles and policy-induced economic downturns than adult employment. Third, short-run demand shocks mutate into long-run ‘scarring’ effects manifested in reduced employment and earnings opportunities that can last decades. Young people with limited skills and from disadvantaged backgrounds are particularly vulnerable to ‘scarring’ effects. There are also well-known negative externalities associated with high and persistent youth unemployment: higher incidence of unhappiness, higher crime rates, higher inequality, higher fiscal costs in terms of foregone output and lower tax revenues, and higher political and social tensions.

93These observations surprisingly contradict Nobel Prize Yunus’s statements when he puts on the shoulders of the youth the burden of being more creative. It is also surprising that issues as complex as those put forward here are still being discussed along the lines we have tried to summarize.

94The next chapter will show, on the one hand, how spatial reconfiguration of production, trade and consumption has modified the behaviors of key aggregates, and on the other, certain empirical evidence regarding the way in which urbanization is embodied in GDP according to the approach we have described here.