Doctoral Summer School of Demography

The SGH Doctoral Summer School of Demography will be held in Warsaw, Poland, on September 4-8, 2023. It aims to introduce students to selected modern methods of demographic analysis used to measure population processes (fertility, mortality and migration). In addition to lectures, the SGH Doctoral Summer School of Demography will feature hands-on workshops on developing one’s papers and reviewing papers by other scholars. Participants will also have the opportunity to receive feedback on their research proposals from the leading academics engaged in the Doctoral Summer School.

The event will offer 32 teaching hours of seminars and workshops, including:

• The inaugural lecture by Francesco Billari on the current socio-economic causes of demographic transition and the associated future challenges
• Mortality module (Carlo G. Camarda) - classical and modern mortality modelling techniques: classical parametric models, the Lee-Carter model and its extensions and semi-parametric models
• Fertility module (Tomáš Sobotka and Kryštof Zeman) - fertility trends, data, measurement and analytical approaches to and the implications of low fertility for wider demographic trends
• Migration module (Arkadiusz Wiśniowski)  - international migration analysis and bilateral migration estimation, sources of international migration data and their limitations, gravity theory of migration and migration forecasting
• Workshop on how to prepare a scientific paper
• Workshop on how to write a review of a scientific paper

Requirements

The Summer School of Demography targets PhD students with a preliminary knowledge of statistics and demography. All concepts will be introduced from the basics. However, elementary knowledge of demographic analysis (e.g., construction of a life-table, computation of age-specific fertility rates and total fertility rate), statistics (i.e., Generalised Linear Models, basics of Bayesian statistics, hierarchical models), and matrix algebra (i.e. transposing and inverting a matrix) will be extremely helpful. Moreover, participants are expected to have a working knowledge of R and bring their laptops with installed R and RStudio.

Thanks to the Polish National Agency for Academic Exchange (NAWA) funding, participation in the Summer Doctoral School is free of charge. Participants must, however, cover their travel and accommodation expenses.

Application process

Candidates should complete an online application form:

https://forms.office.com/Pages/ResponsePage.aspx?id=DhtOFuXIqUGbu29-1A7vBAntVunPUqRDh8IKHZz5dlxUMDBNMU9HS041NlU0TUkyS05PU01HM0taTyQlQCN0PWcu

In addition, the applicants should send the following documents as pdf files to: doctoral_summer_school_demography@sgh.waw.pl

• Outline of the doctoral thesis or research proposal (up to 1000 words, Times New Roman 12)
• CV with the list of publications

The deadline for applications is 30 June, 2023.

Admission decisions will be made based on the quality of the received applications (e-form and additional documents).

Programme and schedule

04.09.2023 (Monday) – 08.09.2023 (Friday)

Detailed programme description

Module 1: Mortality Modelling and Forecasting

Instructor: Carlo Giovanni Camarda (L’Institut national d’études démographiques - Ined)

The creation of the first life table about 350 years ago was not only a milestone in statistics but marked the beginning of demography and mortality analysis. Successive approaches have then shed additional light on our understanding of mortality trends. This course intends to provide an introduction to both classic and modern techniques to model mortality. Particular emphasis will be given to all features needed when forecasting mortality is the final research question. We can broadly identify three families of statistical models which will guide us during the course. 

Class 1: Classic parametric models.  

Starting from the Gompertz law of mortality, demographers and actuaries have attempted to reduce the dimensionality of mortality trends over ages by particular distributions with a defined set of parameters with clear physical meaning. Widely used to characterise with few numbers mortality age -patterns, we will learn their specific features, how to estimate them, quantify their uncertainty and forecast mortality trends by extrapolating time-series of their parameters. 

Class 2: The Lee-Carter model and its variants

A breakthrough in modelling and forecasting mortality has been achieved by the Lee-Carter model in 1992. Elegant and simple, this model is based on linear extrapolations of the logarithms of age-specific death rates, using principal component techniques. A single index coefficient captures the time trend in mortality rates and the mortality forecast is then derived from the projection of this index using standard time series methods. We will first learn how to interpret this model and when it could be employed. Statistical techniques to estimate the Lee-Carter model will be then presented and approaches for forecasting future mortality rates will be covered.

Class 3: Semi-parametric models

Both parametric and Lee–Carter approaches are based on rigid modelling structures that are often unable to capture certain features of mortality change. An alternative and more recent compromise comes from statistics. Semi-parametric approaches attempt to smooth mortality over age and time without any specific model structure and they allow for a parsimonious description of mortality development. We will learn how to exploit this powerful methodology to both model and forecast mortality. Particular emphasis will be given to P-splines as the most suitable and clear-cut smoothing approach for demographic data. 

While we will focus on the few theoretical concepts that underpin the more detailed literature, this will be a hands-on course. Handouts and routines to reproduce all outcomes presented in the course will be provided to students and the statistical software R will be used throughout the course on publicly available demographic datasets. 

Module 2: Low fertility and its population consequences 

Instructors: Tomáš Sobotka and Kryštof Zeman (Vienna Institute of Demography (Austrian Academy of Sciences) / Wittgenstein Centre for Demography and Global Human Capital, Vienna)

An increasing number of countries around the world have reported low or very low fertility levels. Until the 1980s, the experience of low fertility was largely confined to the European countries and a few higher-income “Western” societies outside Europe. Today, low fertility is rapidly spreading in other world regions, including East and South-east Asia, Latin America, but also India, Iran, or Turkey. This trend does not conform to the earlier theories and projections that envisioned that the process of mortality and fertility decline—the demographic transition—would eventually lead to a low-fertility and low-mortality equilibrium, with fertility rates stabilising around the replacement level threshold and population size stabilising. Clearly, fertility rates often decline deep below such an equilibrium, with many countries regularly reaching new record-lows in fertility. This shift has been accompanied by a long-term trend toward later parenthood, with many couples becoming parents in their late 30s or early 40s, ages when the risk of infertility and involuntary childlessness increases rapidly.

Without a broadly agreed new theoretical framework or empirical evidence, governments, policymakers and broader public in higher- and middle-income countries are increasingly concerned about persistent very low fertility and its expected consequences for population size and age structure. Fears of accelerated population decline and ageing often lead to alarmist assessments of the demographic future in low-fertility countries and regions.

Are these worries justified or overblown? How do we measure fertility and analyse its consequences in a broader context? Does very low fertility indeed accelerate population ageing? Can migration mitigate or offset some of the expected effects of low fertility? How quickly would population shrink and eventually “disappear” at different levels of low fertility? Is higher education (almost) always linked to lower fertility? Can policies reverse the fertility decline? Will delayed childbearing lead to further fertility declines in the future? 

This course will consist of three lectures combined with practical hands-on exercises that aim to enhance students’ understanding of fertility trends, data, measurement, and analytical approaches, and of the consequences of low fertility for broader demographic trends. We will cover different dimensions and layers of fertility measurement, introduce main data sources, look at the impact of the changes in fertility timing, and analyse how the combination of different levels of fertility, mortality, and migration affect trends in population size and distribution. We will use data for selected countries and from selected international databases (such as the Human Fertility Database, Human Fertility Collection, Human Mortality Database, and Eurostat) to analyse and model fertility change and its broader population consequences.

Class 1: How low is too low?

This class will introduce basic concepts and measures of fertility and population replacement and give an overview of main data sources, including the Human Fertility Database. We will cover period and cohort measures of fertility, the total fertility rate and look at different layers of fertility analysis, including age, parity, and duration (interval) between births. We will discuss data and data quality issues. We will look at the concept of population replacement, its measurement and variation across countries. We will consider the measurement of fertility timing, its impact on conventional fertility indicators (“tempo effect”) and introduce simple measures that aim to compute period fertility rates free from this tempo effect. We will also look at fertility change from both long-term and short-term perspectives, including the recent experience with fertility swings during the COVID-19 pandemic. Finally, we will review the concept of population replacement and replacement-level fertility, its measurement and the demographic factors that influence the population replacement level threshold.

Class 2: Understanding fertility change in context

Fertility change does not occur in isolation from other social and demographic trends. This class will first look at fertility differentiation in contemporary low-fertility societies, focusing especially on education differentials in fertility, their trends and driving forces. We will also consider the interplay between a shift to later childbearing and fertility levels and how assisted reproduction may change it in the future. Furthermore, we will look at the impact of crises and shocks for fertility and analyse the role of low fertility and declining mortality in driving the shift towards older population structures (“population ageing”). Furthermore, we will look at different measures and approaches to analysing the impact of migration for fertility, population change and population replacement.

Class 3: Low fertility: long-term consequences for population trends and structures

Working with the examples for selected countries, this class will look at the long-term impact of low fertility on population structure and age in Europe and possible impact of government policies to address these trends. How quickly would populations fall and age in the future if contemporary fertility levels would continue? Is migration likely to change this picture? What would be the consequences for working age population and population support ratios? Can government policies pertaining to fertility, retirement age, labour market, and migration alter this picture?

Module 3: Estimation and prediction of bilateral migration

Instructor: Arkadiusz Wiśniowski (University of Manchester)

This short course will focus on international migration and methods of estimating bilateral migration. 

Lecture (2 teaching hours): Students will be introduced to the main sources of data on international migration and their limitations. Then, you will learn about a gravity theory of migration and how to use it to predict migration. 

Laboratory (4 teaching hours): You will use R software to analyse a Bayesian hierarchical model that corrects the measurement errors in real-world migration data and imputes missing observations. You will also learn how to interpret the estimates of the model.

Funding

The Doctoral Summer School is funded by the grant received in the project “Strengthening the SGH Doctoral School’s Position Within the Framework of the International Area of Doctoral Education (Internationalisation of the SGH Doctoral School)” financed by the Polish National Agency for Academic Exchange (NAWA) under the program “STER - Internationalization of Doctoral Schools”.

Scientific Committee

Francesco Billari, Bocconi University

Carlo G. Camarda, L’Institut national d’études démographiques - Ined

Agnieszka Chłoń-Domińczak, SGH Warsaw School of Economics

Irena E. Kotowska, SGH Warsaw School of Economics

Tomas Sobotka, Vienna Institute of Demography (Austrian Academy of Sciences) / Wittgenstein Centre for Demography and Global Human Capital, Vienna

Arkadiusz Wiśniowski, University of Manchester

Marzanna Witek-Hajduk, SGH Warsaw School of Economics

Krystof Zeman, Vienna Institute of Demography (Austrian Academy of Sciences) / Wittgenstein Centre for Demography and Global Human Capital, Vienna

Organising Committee

Anita Abramowska-Kmon

Marta Pachocka

Marta Styrc

Krzysztof Tymicki

Paweł Strzelecki,

Accommodation suggestions

Gromada Hotel Warsaw Centre

Hotel Reytan

a&o Hostels Warsaw

Hotel Campanile Varsovie

E-apartments