The ATO Longitudinal Information Files (ALife): A New Resource for Retirement Policy Research

• Author: Abraham Chigavazira,Hang To,Andrew Carter,Marc Chan,Son Nguyen,Roger Wilkins,Cain Polidano,Justin Holland,Ha Vu
• DOI: http://doi.org/10.1111/1467-8462.12388
• Published date: 01 September 2020
• Date: 01 September 2020

The Australian Economic Review, vol. 53, no. 3, pp. 429–449 DOI: 10.1111/1467-8462.12388

Data Survey

The ATO Longitudinal Information Files (ALife): A New

Resource for Retirement Policy Research

Cain Polidano, Andrew Carter, Marc Chan, Abraham Chigavazira, Hang To,

Justin Holland, Son Nguyen, Ha Vu and Roger Wilkins*

Abstract

The Australian Taxation Office release of

annual longitudinally linked individual tax

and superannuation records, known as the

ATO Longitudinal Information Files (ALife),

opens up opportunities for new research. In

this study, we provide an overview of ALife,

focusing on its use for retirement income

research. To this end, we provide the first

longitudinal estimates of superannuation out-

comes for 1‐year birth cohorts. Results show

marked increase in disparity of super bal-

ances in the lead‐up to retirement as those in

the top quartile ramp‐up their contributions,

possibly to take advantage of the favourable

tax treatment of superannuation income in

retirement years.

1. Introduction

Governments around the world are seeking to

improve the efficiency of programs and service

delivery through the curation and release of large

administrative datasets for research. In Australia,

following the Australian government's 2015

Public Data Policy Statement

1

that committed

the government to release non‐sensitive public

data, there has been a wave of activity to release

administrative data for research. Consistent with

this push, in 2019 the Australian Taxation Office

has compiled and released a 10 per cent sample

of annual longitudinally linked individual tax

and superannuation records (panel data), known

as the ATO Longitudinal Information Files

(ALife

2

). ALife data are released annually

around January each year, but due to late

lodgements in tax returns, there is a 2‐year

delay in data release. The current release,

ALife2017, tracks individuals from 1990

to 1991 in tax records and 1996–97 in

superannuation records up to 2016–17.

The release of ALife opens up opportu-

nities for new research across many policy

domains, including income distribution and

dynamics and the labour market. Perhaps less

obvious is that the data creates important new

opportunities for research on retirement. Prior

to ALife, retirement income analysis relied on

the use of nationally representative annual

survey data, especially individual panel data

from the Household, Income and Labour

Dynamics in Australia (HILDA) Survey and

repeated cross‐sectional data from the

* Polidano, Chigavazira and Wilkins: Melbourne

Institute: Applied Economic and Social Research, The

University of Melbourne, Parkville, Australia; Polidano,

Chan, Vu and Wilkins: Tax and Transfer Policy Institute,

Crawford School of Public Policy, Australian National

University, Canberra, Australia; Carter, To, Holland and

Nguyen: Australian Taxation Office, Canberra, Australia;

Chan: Department of Economics, The University of

Melbourne, Parkville, Australia; Vu: Department of

Economics, Deakin University, Burwood, Australia.

Corresponding author: Polidano <email cainp@unimelb.

edu.au>.

© 2020 The University of Melbourne, Melbourne Institute: Applied Economic & Social Research,

Faculty of Business and Economics

Published by John Wiley & Sons Australia, Ltd

Australian Bureau of Statistics’Household

Income and Wealth Survey. There are several

advantages of ALife over these surveys. First,

the large number of observations in ALife

provide new opportunities to conduct statisti-

cally robust analysis on sub‐groups of interest

that may be targeted by policy. For example,

ALife allows for analysis of responses to the

Low Income Superannuation Tax Offset

(LISTO), introduced in 2012, that targets

low‐income earners. The longitudinal dimen-

sion of ALife means that it can be used to also

examine differences in sub‐group responses

over time, which because of differences in

their life circumstances and experiences, are

often varied. Second, there is a well‐

established literature on survey non‐response

and response error and bias in relation to

questions that have clear socially (un)desir-

able answers such as income (see Moore and

Welniak (2000) for a review), that give

analysis with administrative data an advan-

tage. Third, ALife includes criteria for

program eligibility which allows for more

precise estimation of program effects. For

example, LISTO eligibility depends on tax-

able income from specific income sources and

concessional superannuation contributions,

which are not available in other datasets.

3

Finally, the new historical data in Alife2017

extends further back than existing survey data,

which allows for new analysis of long‐term

trends and policy impacts.

4

In this paper, we provide a brief overview

of ALife for researchers and showcase its

potential for superannuation and retirement

income policy research in Australia using the

initial ALife release—Alife2016, which was

the version available at the time of analysis.

To do this, we first describe briefly the

retirement income system and policy land-

scape in Australia to understand better the

utility of ALife data for retirement research

(Section 2). Following this, to understand

better the underlying Alife population and the

superannuation data, we compare the ALife

sampling frame with population estimates

from ABS Census (Section 3), provide an

overview of ALife content (Section 4), com-

pare ALife superannuation data with that from

the HILDA Survey (Section 5) and present

longitudinal superannuation information for 1‐

year cohorts (Section 6). In concluding

(Section 7), we summarise ALife's strengths

and areas for future development. Like other

administrative datasets, ALife is constantly

evolving and these developments, where

possible, will be incorporated into future

ALife releases.

2. Superannuation and Retirement Income

Policy in Australia

Retirement income policy in Australia is

based on three pillars: the means tested and

publicly funded age pension; superannuation,

a compulsory employer‐funded private retire-

ment pension; and voluntary private savings,

including voluntary private contributions to

superannuation, that are supported through tax

concessions and targeted government pay-

ments. A feature of the Australian system is

the high interdependency of the three pillars,

due mainly to the income and assets tests of

the Age Pension, which means that Age

Pension payments depend on the accumula-

tion of private savings (whether in super-

annuation or not). This interdependency

potentially incentivises people to trade‐off

higher savings from private sources for

greater access to the publicly‐funded Aged

Pension, which has fiscal implications. As

identified by the Productivity Commission

(2015), the margins where this occurs and the

fiscal implications of this are not well under-

stood. The release of ALife is an important

development in efforts to understand super-

annuation accumulation and deaccumulation

behaviour in response to tax, superannuation

and pension reforms. Research by the authors

is currently underway to examine some

of these interrelationships and their fiscal

implications.

5

Below we provide a brief introduction to

superannuation in Australia. For more detailed

information about the superannuation system,

including recent changes, see the Australian

Taxation Office website (<https://www.ato.

gov.au/Individuals/Super/>).

430 The Australian Economic Review September 2020

© 2020 The University of Melbourne, Melbourne Institute: Applied Economic & Social Research, Faculty of Business

and Economics