146 THE ECONOMIC JOURNAL LJANUARY thousand in 1959. Then it jumped to 25.48 per thousand in 1961. From the city and country figures, we see that much of the high national death rate could be traced to rural areas where there was an astonishing 28.58 per thousand death rate in 1960. In the famine period, city death rates were above normal standards but were much lower than rural rates. The opposite patterns are found in the birth statistics where the national rate dropped considerably in the years of crisis. Again, the rural population was more severely affected with the lowest birth rate of 16.99 per thousand occurred in 1961, which was about one half of the birth rate in 1957. Reduced births also seriously affected the well-being of the population. However, since these figures do not directly represent loss of lives, our analysis shall concentrate on death rates Table 3 provides more detailed death rates between 1954 and 1966 at the provincial level. Among the existing 30 provinces and municipalities, Hainan province and Tibet autonomous region are not included in the table because the former is a newly established province and the latter does not have official death rate records for the period under consideration. The data reveal variations of death rates during the crisis. In the worst famine year, 1960, the lowest death rates are found in the three municipalities, (1)Beijing, (2)Tianjin and (9)Shanghai, which were the most important political and industrial centres in China. There are also noticeable regional variations in death rates the north and northeast provinces generally suffered lower death tolls during the famine t this point, we note a few idiosyncratic factors that may cause variations in provincial death rates but will not be scrutinised in the later statistical analysis First, the political strength of a provincial government to resist the central administration directly determined the volume of local grain extraction and consequently affected the severity of the famine. For instance, (19)Guangdong and (7)Jilin only had mild increases in excess deaths because these two provinces had been successful in reducing their grain export burdens. To the contrary, provinces that were cooperative or obedient to the state, such as (25)Gansu,(16)Henan,(21)Sichuan and(28)Hunan, all had high death rates Personalities and political strategies of provincial leaders in weighing local welfare and central orders played a direct role in affecting the death rates 24 The exception is(5)Neimonggu, which had a death rate below(2)Tianjin Neimonggu is a special province because a high percentage of its population were herdsmen whe aising activities. This province was also relatively isolated from the rest of the count See walker (1984)for s grain su 1950s and 1960s and detailed descriptions about the political struggle between the pi entral government. The general secretary of the Guangdong province, who effectively resisted th central procurement in the years of disasters, was purged during the Cultural Revolution. In othe rovinces, administrations managed to export 361 thousand tons of grain from Gansu, 935 thousand ons from Henan, 2.24 million tons from Sichuan, and 440 thousand tons from Hunan pr the severe food shortages in 1959-60. These provinces were severely hit by famine. As 1958-59 agricultural year, a procurement slogan was propagated in Sichuan: First th ocality, first external (commitments), then internal(commitments). The province organised 5 million transport grain for export and the procurement reached the highest historical level of 2.595 million tons. Because of this. the famine hit the province earlier than elsewhere, resulting 17 per thousand death rate in 1959. The obedience of the provincial government was also responsibl for the highest provincial death rate of 29.4 per thousand in 1961 C Royal Economic Society 2000thousand in 1959. Then it jumped to 25.43 per thousand in 1961. From the city and country ®gures, we see that much of the high national death rate could be traced to rural areas where there was an astonishing 28.58 per thousand death rate in 1960. In the famine period, city death rates were above normal standards but were much lower than rural rates. The opposite patterns are found in the birth statistics where the national rate dropped considerably in the years of crisis. Again, the rural population was more severely affected with the lowest birth rate of 16.99 per thousand occurred in 1961, which was about one half of the birth rate in 1957. Reduced births also seriously affected the well-being of the population. However, since these ®gures do not directly represent loss of lives, our analysis shall concentrate on death rates. Table 3 provides more detailed death rates between 1954 and 1966 at the provincial level. Among the existing 30 provinces and municipalities, Hainan province and Tibet autonomous region are not included in the table because the former is a newly established province and the latter does not have of®cial death rate records for the period under consideration. The data reveal variations of death rates during the crisis. In the worst famine year, 1960, the lowest death rates are found in the three municipalities, (1)Beijing, (2)Tianjin and (9)Shanghai, which were the most important political and industrial centres in China.24 There are also noticeable regional variations in death rates: the north and northeast provinces generally suffered lower death tolls during the famine. At this point, we note a few idiosyncratic factors that may cause variations in provincial death rates but will not be scrutinised in the later statistical analysis. First, the political strength of a provincial government to resist the central administration directly determined the volume of local grain extraction and consequently affected the severity of the famine. For instance, (19)Guangdong and (7)Jilin only had mild increases in excess deaths because these two provinces had been successful in reducing their grain export burdens. To the contrary, provinces that were cooperative or obedient to the state, such as (25)Gansu, (16)Henan, (21)Sichuan and (28)Hunan, all had high death rates. Personalities and political strategies of provincial leaders in weighing local welfare and central orders played a direct role in affecting the death rates.25 24 The exception is (5)Neimonggu, which had a death rate below (2)Tianjin. Neimonggu is a special province because a high percentage of its population were herdsmen who primarily engaged in stock￾raising activities. This province was also relatively isolated from the rest of the country. 25 See Walker (1984) for an exhaustive analysis of China's grain supplies and procurement in the 1950s and 1960s and detailed descriptions about the political struggle between the provincial and central government. The general secretary of the Guangdong province, who effectively resisted the central procurement in the years of disasters, was purged during the Cultural Revolution. In other provinces, administrations managed to export 361 thousand tons of grain from Gansu, 935 thousand tons from Henan, 2.24 million tons from Sichuan, and 440 thousand tons from Hunan province despite the severe food shortages in 1959±60. These provinces were severely hit by famine. As early as the 1958±59 agricultural year, a procurement slogan was propagated in Sichuan: `First the centre, then the locality; ®rst external (commitments), then internal (commitments).' The province organised 5 million people to transport grain for export and the procurement reached the highest historical level of 2.595 million tons. Because of this, the massive famine hit the province earlier than elsewhere, resulting in a 47 per thousand death rate in 1959. The obedience of the provincial government was also responsible for the highest provincial death rate of 29.4 per thousand in 1961. 146 [ THE ECONOMIC JOURNAL JANUARY # Royal Economic Society 2000