For this second Asian urban Inquiry, questionnaires were sent to about 250 cities. Information was obtained on 133 cities from 10 countries. The number of cities by country is as follows.

Table 5
Number of Cities Responding
by Country

The basic description of this sample of cities will include their size their growth rates and the sources (migration or natural increase) of their growth, and their area and densities

Siz. Renecting the onginal orientation of the Asian Urban Infbrmation Center, the cities in this sample are medium-sized cities. Although only six were less than 100,000 in population, well over half (80 cities or 60%) were between 100,000 and 500,000. Another 13 were between 500,000 and l mi1ion. Thus 99 of the 133 cities (or 74 percent) were cities of less than l million.

Only 5 cities have populations greater than 5 million. Four of these were in China and one (Delhi) in India. Another 21 cities had populations between one and five million. Only Nepal, Pakistan and Thailand did not cover citjes of greater tllan one million. From China there were 5, India 2, Indonesia 6, Korea 3, Malaysia l, The Philipplnes l, and Japan 3.

Reflecting one of the serious problems in this type ofsurvey, there were completed questionnaires from ll (8%) cities Where 1990 population levels were not given. There were 2 from China (Harbin and Zhangiang), 2 from India (Bangalore and Lucknow), one from Korea (Taqion), one from Malaysia (Sibu) and 4 from Pakistan (Multan gave no popu) ation size fbrany date, and Fasilabad, Hyderabad, and Larkana provided data for 1980 but not for 1990). Thailand provided two completed questionnaires for diffrent districts of Bangkok, and for one of these there was no infbrmationon population Size. This signals a problem in orgamizing the survey that must be addressed in future rounds.

Growth Rates. A major question to be asked at the outset is how valid are the responses of our urban administrators when they are asked how rapidly their cities are growing. The questionnaire provides three different approaches to this question. Two questions ask about size and growth rates, and one asks to what extent the growth rate is a problem. We must first ask what we can learn about reports of actual growth rates.

Two different questions in the survey provided information on growth rates. Question 2 asked for total population for 1970, 1980 and 1990. From these we can calculate decadal growth rates. For 1970 data were provided for 98 cities; for 1980 there were data for l09 cities, and fbr 1990 we have data for 122 of the 133 cities. From these data decadal growth rates can be calculated for l07 (some cities provide population data on only one of the two years) of the 133 cities.

Question 6 asked administrators to provide data on the total increase for a most recent date, and to break this down by births, deaths, in-and out -mJgration. For births and deaths, 83 and 84 cities could provide information. For in-and out-mlgration, only 70 and 71 cities could provide information. Thus for the total most recent increaSe, We can have data for only 70 of the 133 cities. On the other hand, if we use Just the two reported population figures of 1980 and 1990, We can ccalculate growth rates for l07 cities. The number of cities that provided net actual increase data by country are as shown in Table 6.

Table 6
Cities Reportlng 1980 and 1990 Population
and Cities Reporting Migration
and NaturaI Increase Data.

This reflects a common data problem. The high income countries generally have more, and usually more accurate, data on their conditions than do the low income countries. China and Thailand are exceptions to this general trend. Both have more information on migration and natural increase than other countries of the same level of development. In both cases, however, serious questions can be raised about the quality of the data. Thus for the most part we are left with fairly good data on urban growth and its sources only for the high income countries. This is important to recall, since later we report on administrators' perceptions of the level of migration and the problems caused by this migration. As we shall see, administrators do have definite ideas about such problems, even if solid data are not available.

What, then, can we say about growth rates? First, it is not wise to use data reported on migration and natural increase, except for Korea and Japan. For the other countries, or for 60 of the 86, we can use decadal growth rates as calculated from their reports of 1980 and 1990 census counts. Using the calculated decadal growth rates, we can compare the reported rates of these cities, on a country by country basis, with the United Nations' estimates of national urban growth rates for the period 1980-1990. Table 7 shows the minimum and maximum and mean rates reported for our sample of cities compared with the UN estimates of overall urban growth for those countries.

Table 7
Reported Urban Decadal Growth Rates
Compared with UN. Estimates

With two or three exceptions the reported means and the UN estimates are relatively close. The Nepal report is far below the UN estimate, but this may be due to the fact that the report comes only from Kathmandu, and the smaller cities may be registering higher growth rates because of their smaller size. Korea may reflect the opposite condition. The reports tend to be for the smaller cities (12 of the 18 returns are from cities of less than 100,000), which may be expected to show higher growth rates than the national average because of their smaller size. We have no explanation for the discrepancy in China, and are not sure that the discrepancy should be considered large or only moderate.

In effect, we can have some confidence in the reported levels of city size for the past decade, and thus in our calculation of the decadal growth rate. The cities in our sample, like those for their national populations as a whole, are growing more rapidly than their total population

Given the limited response to natural increase and migration data, however, we cannot have much confidence in the reports of the sources of growth - natural increase and migration -, except for the most developed countries, Japan and Korea.

Area and Density. Two queStions were asked about area. For 1970, 1980 and 1990, data were requested on total area and inhabited area. As might be expected, data on inhabited area were avaihble only for very few cities; 35 could report inhabited area for 1970, 44 for 1980 and 57 for 1990. Ther anges Were from 3 square kilometers to 453 squarekilometers. Since the data are so sparse, We shall not attempt to use them except to illustrate a major data problem.

For total area, there are much more data, though the record is still not complete. Of the 133 cities, 122 couId report total area for 1990, 109 for 1980 and 98 for 1970. We shall use only the most recent data here. The overall mean for the 122 cities was 997 square kilometers. But the range was quite Iarge, from a minimum of 3 to a maximum of 52, 285 square kilometers. Three of the smallest cities in area were in Thailand: Maungpol (3), Bangplad (11) and Prachaup (14). The two largest in area were in China: Jijihar (52, 285) and Heihei (14, 443). By far the greater number of cities, however, were in the moderaterange of around 100-200 square kifometers. Half the cities had areas less than 175 square kilometers, and three quarters had less than 383.

For density, we simply divided the population by the total area to obtain the measure. As with area, the range was great. Population per square kilometer runs from a low of 10 in Heihei, China, to a high of 42,077 for Manila. The mean density for all 122 cities responding was 4,076 persons per square kilometer. One quarter of the cases ftll below 1113; the median value was 2,720, and thTee quarters of the cases Were below 5,000 persons per square kilometer. Later we shall use the total density figure to determine whether there is any relationship between density and various urban problems. First, however, we must say something about the reliability of the measure itself.

City boundaries sometimes include more area, and sometimes less area, than is actualIy built up as urban land use. Thus data on urban density are not necesSarily reliable or valid. Although we received responses on the inhabited as well as the total urban area for only 57 cities, we can use the two figures to illustrate at least part of the reliability and validity problems in measures of ufban density. The two largest cides in area in China well illustrate one aspect of the problem. Heihei reported a population of 5,903, 812 in 1990. Its reported total area was then 52,285 square kilometers, giving it atotal density of 113 persons per square kilometer. But the city reported that its innhabited area was only 30 square kilometers, which implies a density of 196,767. The first measure clearIy overstates the actual urban area and understates the density. But the second measure gives a densily that is many times the density of Hong Kong or Singapore. This makes it appear likely that the inhabited area is substantially underestimated.

In Thailand, which included 5 of the 6 cities under 100,000 in population, We may have the opposite case. The three smallest in area are Muangphol with 3 square kilometers, Bangplad with 11 and Prachaup Khirikhan with 14. Mauangphol also reports an inhabited area of 3 square kilometers, which would give it a total density of 4,625 even though wjth only 13,876 people, it has the smallest population in the sample. Bangplad registers a total density of 14,863, with no data given on inhabited area. Anyone familiar with Thai cities would find these density figures difficult to accept. With a preponderance of one and two story buildings, the cities have high densities of structures, and of inhabitants per room or per structure, but without high rise buildings, it is difficult to generate high densities of households or of population. Thus it appears that these cities may substantially underestimate their actual built up urban land use areas.

Until we have more accessible data on urban area by land use, we shall have problems of reliability and validity in the measurement of urban density. Fortunately the increasing use of electronic remote sensing, and computerized Geographic Information Systems (GIS), is making it possible to think of defining cities by their actual land use rather than by administrative units. Computer manipulation also makes it very easy to measure land use changes overtime.