For researchers studying the economic development of nations, the availability of human capital data is crucial. Economists making efforts to examine the impact of human resources on economic development before the end of World War II often face massive problems. Methods of collecting data, i.e., registration systems, population censuses or surveys did neither acquire sufficiently detailed data nor was the recorded data standardized in any consistent form and little attention was paid to direct information about a populations level of education. For early time spans, more than any other data, demographic records are available. This thesis proposes to use irregularities in the reporting of age to estimate the peoples level of education. Single year age data, which enables researchers to depict a populations detailed age structure, almost always exhibits irregularities in the form of heaped data, i.e., the age distribution does not run smoothly but exhibits sharp jumps and clustering at certain ages. This phenomenon is attributed to age heaping, a term which describes peoples ignorance to their age or peoples tendency to round their ages off. To a varying degree, age heaping exists in nearly all historical age statistics when people were asked for their age (as opposed to age records that were calculated from birth certificates or alike). By measuring the degree of age heaping one is able to derive a simple proxy for human capital covering a greater number of countries and regions as well as earlier periods than for instance signature ability rates. Being a proxy of numerical comprehension and diligence when responding to age questions, age heaping also serves not necessarily as a substitute but as complementary human capital measure to signature ability rates to capture as many aspects of human abilities as possible. With respect to literacy, numeracy has been understudied in economic history and its inclusion in long-run growth analysis can provide new insights into the role and development of human capital. It is the purpose of this thesis to offer profound statistical and econometric analysis to assess the usefulness of age heaping indexes to capture basic numerical skills, and to apply the age heaping concept on current research questions in economic history. The thesis comprises several co-authored manuscripts. The first paper offers a profound analysis of the statistical features of various age heaping indicators, which formed the basis for my later work. It also takes a look on the determinants of age misreporting on individual level, followed by an analysis of national numeracy levels introducing a large and newly constructed data set on age heaping stretching from the 14th to the early 19th century. These data are complemented by signature ability rates to systematically examine the relation between age heaping and literacy rates. The second paper uses the age heaping technique to construct a measure of human capital inequality in 17th and 18th century France and 19th century United States. We employ inequality values on provincial level to track the change of the distribution of human capital over time and to analyze its importance and its interaction with other variables in growth processes. The third manuscript employs the age heaping method as a proxy of cognitive ability to understand determinants of low living standards in the past. Using the quasi-natural experiment of Great Britain during the Napoleonic blockade when imports were restricted and food prices soared, the paper studies the link between net-nutrition and cognitive ability. Given that the welfare system in 18th and 19th century Great Britain was organized at county level, relief payments for the poor varied considerably within the United Kingdom. We use the regional variation of living standards for the poor in combination with the temporary food crisis to elaborate the causal mechanism between net-nutrition and cognitive skills, as well as between cognitive abilities and wage income. The fourth article documents the construction of a data set providing decadal age heaping estimates for 165 countries for the period of 1820-1940. Since the estimates are derived from census age data only, the age heaping values can be widely considered representative for the population. We assess possible determinants of age heaping such as schooling, a strong state bureaucracy, the numbers of census enumerations conducted in the past and the Chinese calendar. Growth regressions employing age heaping as proxy for human capital suggest a statistically and economically significant and positive influence of basic numerical knowledge on subsequent economic growth.
