Microsimulation

The experimental tool of social sciences

Microsimulation is a modelling tool applicable across all science and industry, whose importance grows with increasing computational powers and data availability. It simulates the mechanics of a large system at the level of its constituents, incorporating their individual characteristics, complex dynamics and mutual interactions. It can be used to analyse future scenarios, test and optimise strategies or design interventions. The technique challenges standard prediction methods by providing not only more accurate and precise answers, but also resolving a much wider range of possibly highly specific questions. For example, how much do we need to subsidise healthy food to reduce the T2 diabetes cost projected in 10 years by more than the expected sugar tax income? How to geographically allocate the support and target specific demographic groups over this time?

Averisera has built a fast, efficient and flexible microsimulation engine employing novel mathematical solutions and modern numerical techniques. Below we present some of its applications.

State pension reform in the UK

Presented at AESCS 2018 and IMA Asia-Pacific Regional Conference in Tokyo and PenCon2018 in Łódź.

The social and economic implications of a rapidly ageing population, a consequence of low fertility rates and rising longevity, are becoming increasingly apparent in the UK and other developed countries. A shrinking pool of workers has to fill the labour void left by the retiring ones and concurrently provide for their support, as reflected in the growing old-age dependency ratio. Several solutions are considered as a way to reduce the strain the new demographic situation places on public services, in particular the state pension and benefits system, and health care. The prevailing approach of combined spending cuts and tax rises across the board or the prolonging austerity undermine living standards of citizens, disincentivise workers and hinder the growth of businesses. Hobbled by the workforce shortages in the critical period of adjusting to the new demographic market drivers, the UK economy can slide back into recession, a situation which additionally puts private pension investments at risk. Boosting the immigration, often considered a quick remedy for the labour crisis, poses a political challenge of its own and, as we have shown, merely mitigates the problem. At the same time, fairness between generations requires that everybody spends a similar proportion of adult life contributing to and receiving a state pension. Spurred on by the above and many other interrelated issues, the UK has decided to introduce a broad reform of the state pension system, raising the retirement age for all born after 1950. The former state pension age of 65 for men and 60 for women (established by the Old Age and Widows' Pensions Act back in 1940) is already due to equalise by 2018, and next increase for everybody to 68 by 2046.

Microsimulation parameters

Initial population

5,000,000
Agent characteristics

date of birth, age, sex, ethnicity (same as mother's), fertility, reproductive history, history and likelihood of international migration (children below the age of 15 migrate with mothers), mortality
Simulation period

1991-2061
Data sources
Office for National Statistics: 1991, 2001 & 2011 UK Census, birth and death registrations, International Passenger Survey, Labour Force Survey; UK Pensions Acts 1940, 1995, 2007, 2011 & 2014

We employ Averisera microsimulation engine to compare the forecasts of pension cost dependency ratio for the UK under the following scenarios assuming different state pension age (SPA) and its reforms:

Old SPA 60m/65f
SPA equal to 65 for men and 60 for women, as established by the Old Age and Widows' Pensions Act in 1940 and ignoring the reforms, over the whole simulation period.
Equal SPA
SPA for women starts to increase in April 2010 under the Pensions Act 1995 by one month every month (by date of birth) to reach equal SPA of 65 for man and women in March 2020. The transition is accelerated in April 2016 by the Pensions Act 2011 to complete in November 2018. SPA equals 65 for everyone over the remaining simulation time.
SPA 68
SPAs for men and women equalise in 2020 following the (overridden) Pensions Act 1995 and next rise to 66 between 2024 and 2026 (one month every month), to 67 between 2034 and 2036 and to 68 between 2044 and 2046 (Pensions Act 2007).
Accelerated SPA 68 (current reform)
Following the current legislation, SPAs for men and women equalise in November 2018. Next, they rise to 66 between March 2019 and September 2020 (three months in the first month followed by one month every one of next nine months), to 67 between 2026 and 2028 and to 68 between 2044 and 2046 (Pensions Acts 2007, 2011 and 2014).

The details of the SPA transitions implemented by our model are given in the state pension age timetable published by the Department for Work and Pensions.

Additionally, we consider different scenarios of future relations with the EU, which vary migration patterns between the UK and the rest of the world.

Click items in the legend to show/hide scenarios or display their relative differences, and the date axis to change its range.

The historical pension cost dependency ratio maintained a sustainable level until 2007. Beyond that period it began to rise owing to population ageing, caused by low birth rates and extending life expectancy. The unprecedented rate of this process is propelled by post-war and 1960s baby boomers achieving the state pension age. In addition, their retirement considerably reduces the workforce, whose large part consists of cohorts born in the period of declining fertility rates, from the mid 1960s to 1970s. The trend is somewhat softened by the influx of EU workers, especially the post-2004 enlargement wave from A8 countries.

The rise of state pension age for women starting in 2010 reverses this trend sharply, creating a substantial gap between the old and the new pension scheme scenarios. In particular, the latter reduces the pension cost dependency ratio to the level from 2007. Afterwards, the ratio picks up its previous upward trend fueled by the advancing population ageing, but consecutive rises of retirement age revert it back to around 29%. The trend stops to grow after 2038 owing to subsiding life expectancies. Its final small bounce occurs in all scenarios when the children of large 1960s cohorts enter their retirement years.

The values at which the trends stabilise strongly depend on the considered state pension scheme and are moderately affected by the varying migration patterns in different EU membership scenarios. In particular, the currently introduced state pension age reforms will lead to a considerable and permanent reduction of the pension cost dependency ratio, reverting it back to the levels from before 2007.

Technical details

The details of presented methodology and data sources are privided in my recent article “Forecasting the impact of state pension reforms in post-Brexit England and Wales using microsimulation and deep learning” and presentation given at the International Microsimulation Association Congress 2018 and PenCon2018.