Leadership in the public sector

Leadership in the public sector in the 21^st century – some learnings from the 2016 Census experience

Institute of Public Administration Australia (Tasmania Branch), 13 September 2017

David W. Kalisch, Australian Statistician

Introduction

Thank you for the invitation to speak to you today

In December 2016, I spoke at an IPAA ACT event about the 2016 Census experience when we were still within the Census process. Those comments I made then, now nine months’ ago, are still pertinent and accurate today, with a rawness you will only get at that time.

ABS was devastated by the Census events. ABS staff pride themselves on getting everything right, first time every time. The ABS as an organization has a strong results-orientation and was committed to delivering quality 2016 Census data, whatever it took.

What I plan to talk about today is more reflective, about the context and role of the ABS, as well as some further comments about the Census learnings for ABS and possibly other organisations.

Context of the ABS

Let me start with a brief overview of the ABS – Australia’s national statistical organisation.

The ABS produces around 500 statistical releases every year (around 2 major statistical products every working day), and then we also do a Census of Population and Housing, as well as a Census of Agriculture every five years.

ABS is a complex organization, with around 3000 staff with a range of skills including statistical, mathematical, economic, IT, information management, operations management, field interviewing and corporate (such as finance, risk management, HR). We provide national statistical services from 9 locations (every capital city and Geelong), with many of our work programs located across several locations.

For example, we do our agriculture statistics largely from Hobart, our financial statistics from Sydney, have our national data acquisition centre based in Geelong, and a new Centre of Excellence for Indigenous Statistics in Darwin. This helps us have a truly national perspective and source expertise from across Australia, but there are collaboration and efficiency challenges for ABS as an organisation from having such a dispersed footprint.

The Australian Statistician, rather than our Minister, is accountable for the operations of the ABS. This independence helps deliver the integrity of Australia’s statistical system – but the ABS is not truly independent of Government or the Parliament in some very important ways. The Australian Government sets our budget funding (as well as a few Census related choices) and the Commonwealth Parliament establishes our legislative framework, while we provide statistical services to all governments and the broader community.

The ABS is expected to measure an economy, society and environment that is becoming more complex, and more complex to measure. Globalization is challenging how we measure economic activity and international trade, we are putting more attention to measuring the large service sector of our economy and productivity is a key policy and measurement conundrum. Changes in our economy, society and environment need to be measured accurately.

Key data users are demanding new statistical products as well as continuation of most of the existing ones. It is becoming more difficult to get information from households and businesses through traditional survey approaches, but new information sources such as big data are emerging. New statistical techniques are also being developed in Australia and internationally.

The operating context of the ABS is that we have 20% less funding and staffing compared to 15 years ago and our fragile, ageing statistical infrastructure is being replaced over a five year program. We are putting increased attention to statistical risk management of our key statistics, as we need to use our old patchwork systems for some time. Some technology and data enhancements are helping ABS deal with this demand/funding dilemma. ABS has a workforce that is very skilled in what it does but is being asked to operate differently from the past.

The main strategic focus of the ABS leadership is to maximise public value from the resources we receive. We assess this across the five competing dimensions of the provision of quality, timely national statistics; producing new statistical insights that have policy relevance; enabling effective, safe use of ABS data; pursuing efficient and less intrusive data capture; and continuing to build ABS capability for the future. Inevitably, we make trade-offs and judgments around how much we can pursue within our budget and our staff capability.

Transformation of the ABS

The challenge for the ABS is to keep transforming what we do and how we work.

The ABS has not stood still over its past 110+ years, and has significantly changed what it does and how it works. However, these changes could be characterised as gradual over past decades.

However, we now operate in a fast-moving information age, where opportunities, expectations and technology are all moving at considerably faster pace than previously. This necessitates a more significant modernisation of the ABS, and transformation of the ABS will need to be an ongoing focus over future years.

When I began as Australian Statistician in mid-December 2014, I was very clear with senior staff from day one that the ABS:

I have been using a framework for organisational change that focuses on six key components: understand the environment in which the organisation operates, direct your strategy to most effect, and set both the hard-wiring of the organisation (governance and infrastructure) and the soft-wiring of the organisation (people and culture) for the ABS to deliver on its purpose.

In the ABS, and within the context of our purpose to “inform Australia’s important decisions”, we have further refined objectives for these six areas:

These have been enshrined within our Corporate Plan since 2015. We have been regularly reporting key initiatives to staff (6 times in the last 2 years). Progress is also reported publicly in the ABS Annual Report.

This provides clarity to staff and our key stakeholders about the breadth of our transformation, identifies progress that has been made, highlights performance and opportunity gaps, and expectations of further change.

To provide some concrete examples, over the last two years we have improved many of our key stakeholder relationships, have changed our governance committees and organisation structures significantly to improve our decision making, have improved the diversity of our senior workforce, completed two years of our five year project to refresh our statistical infrastructure and implemented activity based working across most of our nine locations.

We are now expanding our focus to cultural change, workforce planning, and helping staff understand in tangible ways what the changes will mean for them and what our expectations of them are and will be. This is being assisted by bringing in new staff with other experiences and expertise, to complement those who have worked at ABS for a long time, through short term secondments or permanent placements. ABS staff are being assisted to work elsewhere, and then return to the ABS with new insights and strengthened stakeholder relationships.

While the ABS has received extra investment from Government for transformation of our statistical business systems that has enabled part of our infrastructure change to proceed, the ABS is also self-funding part of our transformation program from our own resources.

The challenge for the ABS is to deliver quality, timely official statistics while we also transform our business operations on so many fronts. Key data users still expect all of our usual quality statistics, while we also modernise the ABS, become more responsive to user data needs and build our future capability.

So moving onto the 2016 Census

First, some background about planning and undertaking a Census of Population and Housing, one of the few statistical processes required under the ABS legislative framework.

Each Census takes about six-seven years from start to finish. In the case of the 2016 Census, initial planning began in 2011, the new digital first strategy was agreed in 2012, funding from government was set in 2012-13, major external procurements signed in 2014, with a number of decisions (Census date, Census content, advertising approach, longer retention of names) in the second half of 2015. Early results are out now, with more to be released in coming months.

We have already started planning for the 2021 Census, drawing upon our recent 2016 Census experience and the international experience of Census taking in comparable countries (eg, Canada, New Zealand, etc).

1. Manage a big program, commensurately

The Census is a major exercise, with the 2016 Census cost just below $500 million, involved up to 38,000 field staff, and complex delivery models for different populations, such as Indigenous, remote, homeless, and elderly.

The decision in 2011-12 to change from the traditional drop and collect paper forms model to a digital first Census, was a bold change, particularly for a cautious organisation such as the ABS.

It was a justifiable and reasonable change, given Australia’s adoption of technology, the increasing difficulty of physical access to secure apartment buildings and gated communities, and it was becoming more difficult and costly to secure enough temporary staff with the traditional model. The change was consistent with the Government’s digital transformation agenda, and community expectations about efficient, simple on-line transactions with government.

The overall strategy, and its ambition, was not the problem – this new Census approach needs to be continued. The challenge is to effectively govern and manage such a major program, which also has many features of a major event.

The ABS had a detailed plan and had properly envisaged and developed the 2016 Census as a major process transformation. It was not just digitising a paper form.
In hindsight there were a number of lessons for managing such a major program in contemporary times, and learnings we have already taken across to our early planning for the 2021 Census (as well as the ABS more generally):

Managing a major event over an extended period does bring some challenges. Over the five years that the 2016 Census was planned and implemented, there were 3 Australian Statisticians (one of these temporarily in the role for around 12 months), 7 Ministers consecutively oversighting the ABS and 4 Prime Ministers. The 2016 Census cycle was also the first time the ABS was expected to deliver a Census with less funding than the previous Census.

Plans for major events or programs need to include considerable preparedness around things that might go wrong. Outside of the cauldron of the particular real-time challenge, contingency planning can increase the likelihood that comprehensive and timely responses are put in place if stuff happens. The ABS could have done better around August 2016, and we learnt from this, with much stronger contingency plans in place for the release of Census data in June 2017.

2. Innovation and risk

Risk and imperfection is a part of human life. We face risks every day, from when we climb stairs, prepare and consume meals and drinks, use any form of transport, and the examples go on. Everyday life can be fraught with dangers and, even if you are careful, you do not mitigate all the risks.

We need to put the same lens and expectations on organisations. They do need to effectively manage risks that are reasonable within the funding envelope. Resource cuts and efficiency dividends in the public sector contribute to prioritisation of activities, but it can also expose governments and the public sector to greater risk taking as not all risks can be mitigated away.

At the ABS we are taking action to improve our risk maturity, as we set our risk appetite across our respective activities and more effectively monitor and manage risks in real time. The 2016 Census experience highlighted an aspect where the ABS as a whole needs to enhance its risk management capability.

New risks are emerging all the time, as the environment in which we operate keeps changing, so risks do need to be actively managed, and risk mitigation regularly tested and shifted if required.

It is often said that innovation brings increased risk, and this might be the case in some instances. However, lack of innovation, sticking to the past approach when the world is changing rapidly, will also expose organisations to risks.

At the ABS, we have no choice but to innovate and improve – if we do not, our service offer to the community will become smaller and we will become less useful to the community.

ABS staff share this perspective. In a recent staff survey, 97% of ABS staff believe it is one of their responsibilities to continually look for new ways to improve how we work, and 90% of staff believe their immediate supervisors encourage this. ABS as an employer is perceived to be well above the APS average in inspiring, recognising and supporting innovation.

So the choice we confront is not about whether to innovate or not to innovate. It is rather one of how to innovate in smart ways, and face the emerging risks.

Mistakes are very visible in the public sector where there is considerable transparency and accountability – through disclosure requirements, Parliamentary scrutiny and political contest.

Any organisation – public or private sector - will never be able to stop all risks from eventuating, even if budgets were unconstrained.

The ABS managed 2016 Census operations with its usual approach of keeping Government informed of operations and engaged on matters where Government was required to make decisions. In the future, the ABS does need to have a more active dialogue with Ministers about the risks we are confronting, providing Government with the opportunity to mitigate some risks if they choose to do so.

3. Importance of external advice and critical friends

With any organization, there is generally a predominant paradigm of “this is how we do things around here”. The delivery of a Census every five years poses particular challenges, given the tendency to repeat what we did before, and believe this is the safe pathway.

Over the last few years, the ABS has actively sought greater diversity of perspective, through refreshing membership of our Australian Statistical Advisory Council, having more independent members and an Independent Chair of our Internal Audit Committee, and introducing a new Population and Social Statistics Advisory Group (complement to Economic Statistics Advisory Group).

At the ABS, we are seeking to seamlessly blend new leadership perspectives and other major project and event experience. ABS has traditionally had very good contact with other National Statistical Organisations undertaking Censuses at different times, and that international Census community is very useful, but we are also engaging with those undertaking major events, such as the 2018 Commonwealth Games.

In another major change program underway at the ABS, the five year $250m refresh of our statistical systems and processes that commenced in mid-2015, we deliberately drew in external expertise to combat potential internal complacency and group think. In addition to the mandatory gateway reviews, we engaged a major consultancy firm, KPMG, to provide ongoing Independent Assurance to the program and also invited prominent independent experts onto the Program Board that I chair. Both aspects have been introduced into governance structures recently established for the 2021 Census.

Organisations do need to be careful around how they rely on external expertise, as our experience with IBM and the 2016 Census has showed. It is inevitable, with prevailing APS budget trends and staffing arrangements, that public sector organisations will use external suppliers to help deliver required services. However, we also need to have sufficient internal expertise and external assurance that we are receiving the quality external services that are expected.

Building greater diversity into your senior leadership team will also help. Over a three year period, there has been considerable change to the senior executive of the ABS.

By itself, it is not unusual for an organisation to have considerable change in their leadership group. We have deliberately sought out and attracted senior expertise from outside of the ABS, to add better balance to those who have spent most of their career at the ABS. We have also improved the gender balance of our SES, from 24% when I started in late 2014 to just over 50% now.

4. Towards deeper engagement, beyond communication

With every Census, there is a major challenge to alert the community to the Census and encourage full participation. Five years is a long time between Census processes and we have new potential respondents that need to be informed and engaged.

With each Census, the ABS engages with key stakeholders who can assist the ABS to deliver quality, comprehensive Census information from their interest group (eg, culturally and linguistically diverse populations, those in remote locations, Aboriginal and Torres Strait Islanders, people with disabilities, homeless populations, those in residential aged care, etc). These key stakeholder engagements were well planned, meet a mutual need, and contributed to increased Census response from many groups across the population.

In hindsight, there were a number of problems with our mass communication strategy in 2016. We underestimated the communication challenge, did not allocate enough dollars or time to the communication program, and communication campaigns need to be more agile to respond to emerging or likely future challenges.

The ABS should have communicated more clearly at an early stage why retaining names and addresses was important. We clearly didn’t persuade people about the security of their information. ABS needs to not only produce quality statistics, but engage more effectively with the community about the why and what we are doing. We could have communicated more to the public on the evening of 9 August when we appropriately took the Census system off-line given the prevailing circumstances.

The 2016 Census also got caught up in the more combative political environment that has emerged in Australia over recent years, epitomised by slogans such as “Censusfail” and “worst Census ever”. The lengthy Federal election from May to early July precluded the usual factual briefing of Federal MPs and Senators in the lead up to the Census. The usual considered perspective of parliamentarians supporting Australia’s essential statistical resources (that they rely upon) was not a feature of the 2016 Census experience. Early and extensive engagement with our political leaders needs to be a feature of the lead up to the 2021 Census.

We need to work with the changing nature of traditional media, being impacted by both technology shifts and compressed time and resources. There is little you can do to stop the media and social media reporting claims, even blatantly false ones. ABS staff, including myself, did 380 media interviews in the weeks leading up to the Census – this may have contributed to the strong level of public participation in the Census, however could not extinguish the fear and concern. We had positive elements of our communications cut from pre-recorded media interviews, and third party voices supportive of the ABS and the Census found it hard to be published in newspapers.

We have a different media and political environment than in the past, where impact and speed to market rather than accuracy appears to be the predominant paradigm. We are in the world of many “experts” who can gain prominence on TV, radio or social media. Contrary community assertions get as much attention as evidence from official sources. However, in this contemporary environment, we also now have access to new communication approaches, and can utilize novel communication approaches to get our perspectives out and about.

5. Privacy and public benefit

The ABS exists to provide information that informs important decisions.

We need to collect sensitive information from households and businesses to produce essential statistics. For example, we collect sensitive personal information around family violence, health, disability, income and wealth, family relationships, and demographic details to produce social statistics. We receive information from the Census and other official sources to deliver essential population estimates and other local community characteristics. We use commercially sensitive information from many sources to deliver reliable economic and industry statistics.

The ABS enjoys high levels of trust from the community to keep their personal information secure. Our legislation requires this, but it is our practices, training and culture that deliver this outcome. We do not take this community trust and community cooperation for granted, as we largely rely on voluntary participation of households and businesses in our data collections to produce quality statistics, drawing upon our compulsion powers to a very limited extent and only where necessary.

Australian Censuses have a long history of prompting privacy debates, particularly in the 1970s, around the late 1990s and then in the lead up to the 2006 Census. The privacy debates connected to the 1976 Census led the Australian Law Reform Commission to undertake a review on Privacy and the Census. Their report released in the late 1979 is a useful reference, and explains why names have been collected in every Census for essential statistical purposes.

Australia appears to have quite a different public privacy discourse from other generally comparable countries who also undertake 5-yearly Censuses, such as Canada, New Zealand, the Netherlands and Ireland. All of these countries retain their Census data permanently, and use it to generate quality statistics.

Community sentiment testing showed that 97-98% of the community expected to fully and accurately complete the Census right through the media and social media debate about privacy, and the 2016 Census had stronger support for a compulsory response than was recorded in 2011. This public intention carried through to strong Census completion rates, contrary to the expectations (and hopes?) of the privacy lobbyists.

The Privacy Impact Assessment process we had for the 2016 Census was sub-optimal, and the ABS will have more comprehensive arrangements in the future. However, this was no reason for false claims to be made that this was the first time the ABS had collected names in the Census as we had been doing this for over a hundred years for sound statistical reasons. Safe data integration techniques are now part of the contemporary statistician’s toolkit, and provide new statistics to inform important policy issues and better understanding of the effectiveness of government programs and policies.

Data agencies such as ABS need to navigate a more complex contemporary environment where there is reduced public confidence that personal information will be kept secure, against the backdrop of data breaches from predominantly private organisations such as Google, Red Cross, and Sony, and the recent Equifax case in the US. The prevalence and impact of cyber attacks are heightening this concern, at the same time as some people are cavalier about sharing their personal details through social media. In the recent Census, the ABS took a pro-privacy stance to protect people’s data and none was compromised.

One feature of this recent Census experience was that many of those who would be able to safely use the Census data for research or statistical purposes were comparatively silent. If reputable users of data want improved access to Census and other data, they in turn need to be more prominent advocates for effective and safe use of data. They are best placed to demonstrate how important this information is and how it could be used to benefit Australia.

6. Fix the problem!

We now know that the 2016 Census has produced information for all required uses, and more. We achieved a community response comparable to previous Censuses, with the main outcomes from the on-line form outage delaying some people from completing the Census and likely more households using paper forms than would otherwise have been the case. The quality of 2016 Census data was good, supported by more on-line responses in 2016 compared to previous years. Refusals to complete the 2016 Census were lower than in 2011.

However, this was not a fait accompli on the evening of 9 August 2016. We were facing the potential of not delivering a quality Census, essential for a range of policy purposes.

Instead of becoming self-absorbed by the Census process and its problems, the ABS as an organization came together marvelously, and in a number of very significant ways that will build future capability:

Once the Census on-line form was back operating from the afternoon of 11 August, we then received over a million more household responses on-line over coming weeks. By the Sunday after Census night, overall Census responses were back to our original expectations of the number of Census returns received by that time, with confidence even at this early stage that we would have a quality 2016 Census.

Graph: Showing the final Approach, Reminder and Visit phase, forecase and actual

Census household response trajectory: forecast and actual

The ABS also put additional effort to our communications of Census data in April and June 2017. We prioritised better engagement with the media and key stakeholders around the time of the Census release so they were able to effectively use and communicate the plethora of Census data when it became available. We were also better prepared for a possible denial of service attack, or any business continuity disruption – including testing our response. Compensation from IBM enabled this to occur without any impost on taxpayers.

7. Not just resilience but anti-fragile

Many aspects of the 2016 Census went well, but some certainly did not.

ABS has delivered quality data from the 2016 Census process, comparable to past Censuses and international Censuses, and the organization has demonstrated considerable resilience through this experience.

But from the perspective of the ABS being a learning, improving organization, being resilient is not good enough!

The ABS should not just cope and have dealt with this situation – we need to also become stronger in order to be better placed to deal with future requirements and pressures on government agencies. We have experience responding to a markedly changing environment, and in some ways a more hostile and questioning environment.

This draws on the concept of “Anti-fragile” – things that gain from disorder, and organisations that benefit from volatility, as written about by Nassim Nicholas Taleb in 2012.

ABS staff now have real life experience dealing with a major event that did not go to plan. They have skills and expertise that others may have only read about or heard others talk about. They are match-hardened professionals!

The Census experience also further encouraged the ABS leadership and organisation to consider some performance gaps. We have changed some of our key governance structures, our senior leadership group has been further refreshed, we have put more resources and senior leadership time towards our culture change program and we are enhancing our risk management maturity across the organisation.

8. Personal leadership learnings

I have taken many leadership learnings from this experience, and too many for the time we have today, so I will highlight some of the main ones:

Leadership can be dangerous. The 2016 Census experience was a very raw and wrenching experience, with many comments not well grounded and without perspective. Many of our colleagues in the public sector have their own comparable experiences.

The anthem from Paul Simon – “I am a rock, I am an Island” – should not apply to any leader. In times like the Census, you soon learn who your true friends are, and their support is invaluable. I, and many of my senior staff, received much support from past and current work colleagues, friends and family, and this was invaluable.

It is also from these very difficult experiences that you learn the most in the shortest amount of time, learning how to be more effective as a leader in the future.

Are you match fit? Have you done the prior training and invested in your development and skills, so you are better able to deal with the “black swan” events that sometimes do happen. It is so easy to put off your own professional development, responding to the urgent rather than the important, but one day you might need to draw upon it big time!

Is your organization in good shape if, but more likely, when something happens? Actions that I had already taken in the previous year to bring in some serious talent/expertise to the ABS from other agencies, have in place a better structure and reporting arrangements, and start the culture change journey helped the ABS deal with this experience. The ABS had not completed its necessary changes, but it was better placed having commenced its wide-ranging transformation than if the Census experience had happened several years’ earlier.

Leaders need to lead, be upfront and out and about, with honesty, authenticity and integrity. Leaders do need to stand up and face the media fray. Through this experience, it was invaluable to be amongst staff in the most challenging times (management by walking around has a lot going for it), for them to know you care for their wellbeing and for the organization to have measures in place to assist them if required.

Leaders need to also provide space and opportunity for others to shine. Your job as the leader is not to make every decision! I have around 3000 staff who should be expected to make sound decisions and choices every day! Decisions need to be taken at the right level on the basis of good (but rarely perfect) information. This makes use of the expertise of the organisation and builds future capability.

I am determined to leave the ABS in a better place than when I started, as I am sure all of my predecessors sought to do. And when I am no longer the Australian Statistician, I hope I will have self-awareness and leadership insight to understand that circumstances facing the ABS have more than likely changed significantly from when I was the Australian Statistician.

Census 2016 Postscript

To end up, I am sure you are wanting to know if the 2016 Census failed or was it the worst Census ever?

A resounding NO on both counts. If someone talks about Censusfail or the worst Census ever, they haven’t any regard to the evidence, particularly now we are in 2017 when Census outcomes are known rather than speculated.

As I have said many times, very clearly, the 2016 Census was not conducted as well as the ABS would have liked. It was not up to the ABS’ usual high standards.

But through the cooperation of the Australian public, and the skill and dedication of ABS staff, we have delivered quality data from the 2016 Census.

You do not need to take my word for this – following the Census night events, I established an Independent Assurance Panel of eminent domestic and international experts to provide an independent, third party view of the 2016 Census data. Their conclusion is that “the quality of the 2016 Census data is comparable with the Censuses conducted in 2011 and 2006 and can be used with confidence.” It “… will support the same variety of uses of Census data as was the case for previous Censuses.”

We are already seeing 2016 Census data being used for important public policy purposes and leading to improved understanding of contemporary Australia:

We have further Census data releases and data products scheduled for October and later, that I know will be important for key decision makers.

Closing remarks

With the Direction I have received from the Treasurer on 9 August 2017 to request the view of those on the electoral roll as to whether the law should be changed to allow same-sex couples to marry, the ABS has an early opportunity to put some of these learnings into practice in a high profile process.

Some of the key features of how the ABS is progressing the Australian Marriage Law Postal Survey, drawing on our learnings from the 2016 Census, include:

In turn, the conduct of the Australian Marriage Law Postal Survey will provide key insights that we can then take across to our planning and implementation of the 2021 Census over coming years.