1. This paper applies shift-share analysis, a tool often used in economic geography and regional science, to regional fertility change in Italy, 1952-1991. During this post-World War II period, Italian fertility declined by over 33 percent, but the decline varied widely from region to region. Moreover, the demographic originations of the decline in births are not fully understood. Using birth data for nineteen Italian regions, this analysis is able to break regional change in births into three main components: a national effect, a cohort effect, and a regional differential effect, which in turn provide insight into the roots of fertility change at the regional level. These three components of change are then further disaggregated to account for the differences between changes due to population change and those related to actual changes in birth rates (the number of children produced by each woman). Strong regional differences between the north and south of Italy are demonstrated.
1. INTRODUCTION
Changing demography, among other factors, has contributed to a substantially different regional picture of Italian fertility in the 1990s than in the 1950s. The downward trend in fertility rates has not been uniform for all regions or for the entire time period; declines were more precipitous for some regions, while others have experienced a slow and steady decrease in fertility rates.
Fertility behavior can be affected at a range of geographic scales. Some of the change might be due to national-level causes, other portions to cohort-specific changes in fertility behavior, and still others to region-specific preferences for a particular fertility schedule (number and timing of children). We employ shift-share analysis, which is able to provide a quantification of these different influences on fertility change, to evaluate the influence of all these different elements, for all regions, across forty years.
Shift-share analysis is frequently used by regional scientists, geographers, and planners to understand the drivers of regional economic growth. The traditional shift-share method is easily applied and understood, and its use for descriptive purposes is well established (Perloff et al. 1960; Fothergill and Gudgin 1979; Stevens and Moore 1980). Although shift-share analysis is a tool usually employed by regional economists to study regional employment change (e.g., Barff and Knight 1988), evolution of regional industrial structures (e.g., Haynes and Dinc 1997; Hanham and Banasick 2000), or policy effects on regional growth (e.g., Thirwall 1967; Tervo and Okko 1983), population geographers have used shift-share for the analysis of regional migration trends. Plane (1987) applied a variation of the shift-share methodology to U.S. internal migration trends, and later (1992) used it to assess the relative importance of age in explaining temporal variations in migration rates. Ishikawa (1999), too, made use of shift-share to evaluate the influence of cohort size on the concentration and deconcentration of population in Japan, Sweden, and Canada as a result of migration. In a newer twist, Wright and Ellis (1997) used shift-share to assess employment changes by ethnic group for various sectors of the Los Angeles economy.
The application of shift-share to demographic change has not extended to regional fertility analysis. Often, demographers know that fertility has been increasing or decreasing in a set of regions but are unable to link this change directly to national, local, or age-specific trends. Shift-share offers that ability. This paper applies the shift-share methodology to regional fertility change in Italy from 1952 to 1991. Rather than examining regional growth in employment, output, or number of firms across a set of economic sectors, we investigate the growth or decline in numbers of births in six age cohorts for nineteen Italian regions. Shift-share analysis enables us to parse regional growth or decline in births into what are commonly called in economic approaches the national, structural, and competitive effects. For fertility applications these effects control for national-level age-specific fertility trends and for the impacts of differential age structure within regions in order to expose region-specific differences in fertility by age of mother. The results indicate that, although all regions experienced a decline in births over the past fifty years, the decrease in births may be attributable to different causes for different regions and time periods.
The format of this paper is as follows: Section 2 provides a brief description of Italian post-war fertility at both the national and regional levels. It also examines some of the demographic context to the fertility decline, showing how fertility evolved between 1952 and 1991. Section 3 explains the shift-share methodology used in the analysis, and section 4 describes the data used. Results are presented in section 5. Section 6 offers some concluding remarks.
2. TEMPORAL AND SPATIAL REGIONAL FERTILITY TRENDS IN ITALY, 1952-1991
Table 1 provides background on regional fertility change over time by showing the percent change in births for the entire period, 1952-1991, as well as four shorter subperiods. The percent changes recorded for Italy act as benchmarks for the regions, showing which regions experienced more or less change than the country as a whole during each time period. For all time periods shown, some regions experienced declines in births that were more or less rapid relative to the level of national change.
For example, between 1952 and 1991, many of the traditionally higher fertility regions, including the Veneto and Basilicata, experienced greater percentage declines in births than the nation as a whole. Other regions, such as Piemonte and Lombardia, experienced lesser declines than Italy as a whole. Across the four subperiods, the most dramatic decline in births occurred between 1971 and 1981.
In common with most developed countries, Italy experienced a baby boom in the postwar years. This peak occurred in 1964 but was not sustained for an extended period of time. Unlike the boom experienced in the United States, Italy's increase in births was slightly delayed after the end of World War II and did not last for more than a few years. Figure 1 displays the national trends in fertility rates for 1952 to 1991. In 1952, the period total fertility rate (TFR) for the entire country was 2.3. (1) By 1991 it had fallen to 1.3. The highest TFR recorded for Italy during these years, in 1964, was 2.7. For the entire study period, Italian fertility never greatly exceeded replacement level.
[FIGURE 1 OMITTED]
At a subnational level, throughout the study period fertility rates in the south were considerably higher than in the north. Although the south's drop in fertility since 1952 has been impressive, its TFR remains much higher than those for the north and central area. Perhaps because fertility was already so high in the region relative to the rest of the country, the baby boom did not have a large effect on TFRs in the south. During this period, the average age at first birth increased across all regions, as did the average age of women in their childbearing years. The peak in age-specific fertility rates shifted towards later ages and lowered. When women marry later and wait longer, on average, to start having children, the end effect will be lower fertility rates. The percentage of first or second births increased during the study period, but remained lower for the south than for the rest of the nation. Although these changes took place in all regions, some regions experienced more dramatic changes than others, and these changes were then reflected in regional fertility trends over time.
Additional variation is apparent at the regional level (see Figure 2 for labeled map of Italy's regions). (2) Individually, regions in the north, with few exceptions, have had lower fertility rates than regions in the rest of the country. As Figure 3 shows, Liguria, Piemonte, Friuli-Venezia Giulia, and Emilia-Romagna already had below replacement level fertility rates in 1952, forming the core of low northern fertility. The southern regions tend to fall into the highest fertility category for all years shown in Figure 3, although by the end of the study period the "high" range is still only a TFR of 1.2-1.5--well below replacement level fertility. Trentino-Alto Adige and Veneto, in the northeastern part of the country, have tended to have fertility levels closer to those seen in many southern regions. Another outlier, Sardegna had the highest TFR in the country in 1952 at 3.8, but one of the lowest by 1991 (1.3). The average age at first birth in Sardegna, which began at 25.9 in 1952, increased to 27.3 in 1991, a figure well above the average for the south as a whole.
[FIGURES 2-3 OMITTED]
The above discussion points to two general conclusions. First, all regions experienced a decline in fertility during this period. Second, there has been a great deal of regional variation in this fertility decline as well as in its demographic causes. The south has tended to lag behind the remainder of the country, but even in this part of the country fertility change has not been uniform. With this mass of information, it remains unclear how the process of fertility decline evolved over time. For a clearer picture of how regional fertility change relates to change at the national level as well as to changes taking place in other regions, we turn to the application of shift-share analysis.
3. METHODOLOGY
Shift-share methodologies abound. The present analysis uses as a starting point the basic shift-share method--what might be termed "classic" or "traditional" shift-share--because it generates the most easily interpreted and straightforward...
http://www.accessmylibrary.com/coms2/summary_0286-3555285_ITM不全， 需要注册有一个东西，不过是免费注册
2。 我有PDF，要的话找我。