Turkey's economy grew by 8.9 percent in 2010. This is really impressive. If the growth rate is sustained at this level until 2023, the national income will reach US$ 2.2 trillion and Turkey will become the 10^th largest economy of the world. If this performance can be maintained, in fifty years, a child born in 2011 will be 46 times richer than his/her grandfather was.

Unfortunately, this performance cannot be maintained. The growth rate of 8.9 percent is considered by growth economists to be the result of an overheating of the economy and meanwhile the economic administration is scolding the banking sector and trying to cool the economy through the Central Bank. The target is to decrease the growth rate to a more modest level, to 4 percent, in 2011. It appears that otherwise the engine might overheat farther.

Macroeconomists are discussing these issues. I, on the other hand, wonder why the 8.9 percent growth rate is not sustainable and what we can do differently to sustain that rate. In a commentary in November 2010, I explained why economic growth implies the growth of firms.[1] Today let me discuss why economic growth implies the growth of cities. Lately, I have been obsessed with the question about the geography of the growth of Turkey; not in sector terms, but in spatial terms. In other words, I have been wondering which city, district or region will grow and become the engine of Turkey's growth.

Anyone who is curious about this question and wants to find an answer to it must be grateful (!) to the Turkish Statistical Institute (TURKSTAT). With precision, TURKSTAT has cut the connection between economic performance and geography. For instance, it has not reported on national income figures by city since 2001. However, it is still possible to state some findings in the absence of relevant data.

In spatial terms, economic growth can take place in three ways. The first is horizontal. If the population of a city grows and the need for land for production increases, the city expands over a larger area. Industrial and agricultural growth generally require horizontal spatial growth. The second and rarer type of spatial economic growth is vertical. If horizontal growth is for some reason impeded, the buildings become taller and the city sustains the increased volume of economic activities via vertical spatial growth. The development of the modern services sector is generally the main factor triggering vertical growth. The Levent-Sarıyer line in Istanbul, the Eskişehir highway axis in Ankara, and the recent development of Dubai are examples of vertical economic growth. Finally, economic growth within a certain spatial area can take place with productivity gains. A city, though it expands neither horizontally nor vertically, might grow since its residents increase production thanks to new skills and technologies. In real life, economic growth results from a combination of these three elements.

In the case of Turkey, we unfortunately cannot observe the development of city economies over time, thanks to TURKSTAT. But based on the assumption that the economic size of a city is an indicator of the population of that city, we can extract some findings on the basis of population figures.

First, the share of the largest 20 cities in the total population increased from 48 percent in 1965 to 65 percent in 2010. In other words, in the last five decades economic activities have become concentrated in the largest 20 cities. The populations of these cities exceed one million, excluding that of Aydın, which ranks 20^th. Given that there are 500 cities with populations greater t han 1 million in the world, we see that Turkey has 20 cities in the largest 500 cities of the world.

Everything is normal up to this point. The second finding, however, is much more striking. If we exclude Istanbul and then review the picture (fig. 1), we see that the share of population of the 20 large cities has not changed considerably over the last forty-five years. On the other hand, the share of Istanbul increased from 7 percent in 1965 to 18 percent in 2010. Thus, the locomotive of the economic growth in recent years mainly has been Istanbul.

The third point is related to whether and to what extent other countries converged with the pace of Istanbul in increasing their share in population. Table 1 summarizes the population movements in the  20 largest cities of Turkey. The last column on the right shows how many times the share of the respective province increased in terms of population between 1980 and 2010. Istanbul is in first place, with 1.7 times. Close followers are Antalya (1.60), Kocaeli (1.59) and Şanlıurfa (1.68). Six other cities (Bursa, Van, Gaziantep, İzmir, Diyarbakır, Mersin) also increased their shares in total population. The shares of 3 provinces remained the same (Hatay, Ankara, Kayseri), whereas those of 7 (Aydın, Manisa, Kahramanmaraş, Adana, Balıkesir, Konya, Samsun) decreased.

These are the immediate ideas this picture evoked in my mind. The fundamental factor driving the growth of Antalya is related closely to the introduction of the region as a "special economic zone" and the city's having become a tourism center in Özal's era. Şanlıurfa's growth might be tied to the increased functioning of the Southeastern Anatolia Project, another huge project of former administrators. It would not be wrong to suggest that Kocaeli's growth was an extension of Istanbul's growth. The growth dynamics in many other cities might be associated with internal migration based on the terror problem as much as economic performance.

If we leave the other cities aside and focus on Istanbul, we see that the city demonstrates a typical example of horizontal growth. In 1965 and 1980, the city had 17 districts. From 1980 to 2010, the population increased by only 100,000 people. Between 1980 and 2010, however, 12 new districts each with an average population of 600,000 were created (Table 2). In other words, in each year in the last three decades, a new piece of land as big as Sinop province (representing a population of 200 thousand people) has been annexed to Istanbul.

The issue of Istanbul is the subject of a separate discussion. At this point, some conclusions can be made. If, in the future, Turkey's growth implies the expansion of cities, I think that Istanbul will no longer serve as the locomotive of growth as it did in the 1980-2010 period. Such an expectation would be not only unfair to this beautiful city, but also unrealistic. The city already has reached its horizontal limits and further vertical growth might seriously harm its structure. With respect to productivity, each new vehicle on Istanbul's streets will not increase but reduce productivity due to the worsening of the traffic problem.

Therefore, Turkey has to find economic growth centers other than Istanbul. This can be enabled by launching large projects as has been done in Antalya and Şanlıurfa rather than letting the markets and cities function on their own. I hope that the agenda of economic policy starts to witness more frequently debates on the spatial dimension of economic growth. Otherwise, the only option will be waiting for Istanbul to reach the sky.

Figure 1: Share of cities in Turkey's total population: Largest 10 cities (excluding Istanbul), Second largest 10 cities, Istanbul; 1965, 1980 and 2010

Largest 10 cities: Istanbul, Ankara, Izmir, Bursa, Adana, Konya, Antalya, Gaziantep, Şanlıurfa, Mersin

Second largest 10 cities: Kocaeli, Diyarbakır, Hatay, Manisa, Samsun, Kayseri, Balıkesir, Kahramanmaraş, Van, Aydın

Source: TURKSTAT

Table 1: Population of the largest 20 cities in Turkey, share in total population and rate of increase in population (1965, 1980, 2010)

Source: TURKSTAT and TEPAV's calculations

Table 2: Old and New Istanbul: Change in the Population of Districts, Thousand People (1965, 1980, 2010)

Source: TURKSTAT

[1] http://www.tepav.org.tr/en/kose-yazisi-tepav/s/2209