Toespraak conferentie CIGAR op 13 juni 2019
Toespraak van Ewout Irrgang, collegelid Algemene Rekenkamer, tijdens de conferentie CIGAR op 13 juni 2019, Vrije Universiteit in Amsterdam.
At the NCA we have an iconic picture of a former NCA-employee who is recording changes in the National debt, while smoking (not a cigar, but a pipe). This is Mr. Dietz in the 1950s.
When I was invited to speak at the CIGAR conference about an issue relating to public accounting, the first thing that came to mind was an image of Mr W.H. Dietz, one of our former employees. I later realised that this thought was triggered by the word ‘cigar’.
The world has changed significantly since the 1950s, both inside and outside the NCA. Our chartered accountants, auditors and data scientists are working together closely to explore how we can use data techniques to make our work more relevant to our auditees and members of parliament.
The NCA audits how public funds are collected and spent. We check whether government spends its money correctly and wisely and whether the public gets value for money.
Independence is essential if we are to function properly. This means that we must be free to select audit topics and to report our findings, and that we need to have sufficient resources at our disposal. I’m sure that you, as academic researchers, cherish and adhere to the same values.
Our work includes producing and explaining audit reports, publishing articles, organising seminars and sharing knowledge. Internationally, we’re connected with almost 200 sister organisations through INTOSAI – the umbrella organisation for Supreme Audit Institutions. One of its main tasks is the development of professional standards for financial auditing, compliance auditing and performance auditing – known as ISSAIs.
CIGAR’s mission is: ‘… to promote worldwide discussion and research on comparative governmental accounting’.
Accounting and independent auditing are two sides of the public accountability coin. Accounting – the key focus of your academic discussions and research – is the pre-stage to and – I would say – a precondition for our audit work.
The issues we audit as a SAI – like compliance and performance – are primarily the responsibilities of politicians and administrators. Responsibilities which should be reflected in relevant and reliable accounting by ministers and their civil servants.
Promoting sound public sector accounting practices is a key element of our mission and day-to-day work. Our core business is adding independent audits to registrations and reports, so that users like members of parliament and other ussers can trust the information they contain and use it safely. So public sector accounting is at the heart of both your work and ours. Accurately recording, analysing and reporting on government finances, fairness and performance is indispensable.
Your academic insights contribute to worldwide progress in public sector accounting. This is also in our interests and – ultimately – in the interests of the general public. I realise that many of you are, to some extent, lone warriors in your own countries, since in many countries public sector accounting is not really ‘mainstream’ – to put it mildly. Business accounting is still often considered to be ‘sexier’, although in my opinion maximising profit can hardly be seen as more important than public wellbeing.
I think it’s only a matter of time before universities realise that public sector accounting should receive more academic attention, given the big, topical and challenging issues it deals with – which could be summed up in a motto like ‘people, planet and not-for-profit’. Until then, your international network provides a unique opportunity to raise an academic voice that could make a difference to public sector accounting.
So as the NCA, like SAIs worldwide, recognizes the importance of your profession, and we hope that in the future we can even learn more of the results of your research.
The world where Cigar’s research community is operating is changing rapidly, just like ours and I will share our observations. In order to fulfil its role: the Netherlands Court of Audit has to adopt to changing circumstances: ‘we have to change to stay the same’. This is because, as Jan Rotmans, a professor of transition science in Rotterdam, recently put it: ‘We are not living in an era of change, but in a change of era.’
Our world is being transformed. The environment in which auditors and chartered accountants operate is changing fast. There are no more Mr Dietzes in this world. The only thing which we expect to stay the same is change itself.
First of all, digitisation and digitalisation are making more and more structured and unstructured data available. In a moment I’ll give you an example involving our Central Government Audit Service in the Netherlands.
Modern computer processors have computing power that was unimaginable a few decades ago. A modern computer processor can be held between a thumb and index finger, once computers where huge.
These two developments – more digital data and more processing power – mean that more powerful analytics, such as algorithms, can now be used, because computation costs have fallen dramatically.
Other Supreme Audit Institutions around the world are benefiting from the development and doing exciting work. SAI South Africa has used data mining in fraud investigation; Oman’s State Audit Institution has incorporated digital forensics in audits aimed at fighting corruption.
In this changing world accountants and auditors have to do things differently, ‘We have to change to stay the same’.
The Dutch Government has digitized a large part of its administration and, as a result works with large datasets. Here I want to stress that that we’re no longer testing the waters. A number of ministries and related institutions have really jumped into the data pool and are learning how to swim.
These technologies have the following potential benefits:
Policy implementation can become more efficient, for instance in: Approval of acces to long-term care benefits ; identifying students with a high risk of dropping out of school; detecting social benefits fraud.
At the NCA we see this as a chance for government to become more efficient and spend public money more wisely. Personally, I feel that auditors and accountants should stimulate this development as it can help us to better scrutinize the use of public money. I think we could also do more, for instance when it comes to ensuring that the underlying data-sets are reliable and experience with various government initiatives is shared more widely.
As a SAI, we also want to make sure governments remain careful and transparent and can publicly explain their decisions. There is a real need for this, only last week one of our daily newspapers reported that only 9 percent of the population believe that algorithms will lead to a fair and just world.
Increasing use of data analytics
The Central Government Audit Service is an internal audit department with certifying competence like a public accountant. It audits the public annual accounts of the departments and gives an opinion on behalf of the ministers. It also gives assurance on the annual accounts. At the NCA we use the work of the Audit Service for our external financial audit.
The Central Government Audit Service has embedded data-analytics into its work for providing assurance on the annual accounts. It has a process for extracting financial data from the different enterprise resource management systems used by the ministries, and transforms this into data that can be used in its analyses.
The auditors work with data analysts to analyse payment processes, tenders, personell costs, double payments, and so on.
At the NCA, we are following this development closely and we considering how we can use these data to improve our work. For instance, instead of taking samples for our reviews, we can analyse the whole data-set. This makes our reviews more reliable, efficient and relevant.
In addition it may take less time to do risk-based reviews so we can supplement those reviews with additional analysis or do more reviews else where we see other risks.
We’ve only just started to explore these opportunities for our audit work. We are really ‘testing the water’. Our financial auditors are working with our data scientists and IT auditors to identify the opportunities for NCA.
An important question that our auditors are trying to answer is: what measures must be in place to guarantee that the data analytics provide reliable audit evidence?
It’s interesting to note that the primary interest of our data scientists is not in the measures that should guarantee the reliability but it is the analysis of the data that will show the reliability. In this discussion, the IT auditors have an important role to play.
Algorithms used within the whole of government
When we discuss algorithms at NCA we use a broad definition: all data techniques, ranging from simple decision-making rules to artificial intelligence, where large quantities of data are used for classification or prediction purposes. Decision-making rules include automatic decisions on issuing a traffic fine or eligibility for social security benefits. An example of a sophisticated tool to support policy implementation is an algorithm that predicts in which city an asylum seeker has the best chance of successful integration.
As with data and digitalisation, it’s not a question of whether government is going to work with algorithms; algorithms have existed for decades and government is already making increasing use of them.
Last Tuesday, 11 June 2019, we published an audit on how our tax authority uses data techniques to select tax returns that have a high risk of containing errors.
Some of the main findings. On the positive side: the risk model with the algorithm performs better than traditional selection rules. However, the algorithm does not achieve both aims of tax supervision policy. Specifically it does identify tax-returns where the taxes are too low but it does not identify tax-returns where the amount is too high. In other words, it benefits our treasury, not the individual who pays too much tax due to an error.
Overall, the tax authority currently has a limited number of prediction models. The tax authority is investing heavily to extend the scope of these models to taxes on profits and VAT.
My overall impression is that the algorithms we audited illustrate the potential benefits: they are helping our tax authority to become more efficient; citizens’ privacy issues are adequately dealt with as we did not find any profiling in this model; however, we did notice the absence of a government framework which stipulates what can and what cannot be done in using algorithms taking into account privacy and carefully dealing with sensitive information. This absence needs to be dealt with.
Digitization not only changes “what” we audit, it also changes “how” we audit this. Digitalised processes, data and the analytical models used by governments will increasingly feature in our audit work. With all the available data, we will be better able to answer different questions, or answer similar question in a better way.
When auditing government processes, we often check whether these processes are verifiable. When it comes to algorithms, this poses difficulties. This is our major challenge when we audit an algorithm.
Besides understanding the algorithm and verify how it works there’s another approach. This involves checking the systems in place to prevent accidents from happening. We can ask our client or auditee to explain how the organisation ensures that the decisions made by the algorithm are reliable and of a high quality.
As auditors we still need to develop criteria to assess the performance of the system that has been put in place. At the moment there are no clear criteria for:
when algorithms can and cannot be used; this is a very difficult matter and raises ethical issues;
what prediction errors are allowable in automatic decision-making;
what is a good process for making, training and testing or evaluating algorithms.
At the NCA we are currently thinking how we can develop these criteria.
Consequences for accountants and auditors
Looking at the fundamentals and content of financial reports, accounting and auditing, I see real changes in the near future. So far progress has been made in the visualisation of data and data analysis. But the only thing that stays the same is change itself.
In a field that’s dominated by numbers we can expect use of algorithms to have an enormous impact. It offers a chance to completely automate parts of our work, especially in financial administration.
Because of technological change, the automation of business processes and digitalisation, some traditional bookkeeping tasks are being minimised or completely automated.
So we might ask ourselves: will the accountant become extinct?
The dodo on Madagascar died out, duet o human intervention arround 1662. The quagga was a plain zebra that lived in South Africa until becoming extinct late in the 19th century. Can the accountant be added to this notorious department?
My answer to this is definitely not. At least not yet. But from the animal kingdom we also know that species which were once prevalent can become endangered quite quickly.
We have to think about how to bring more analysis to our client, ‘uncover meaningful stories’, and help government improve itself. So we can help governments to realize the potential that these data techniques offer, for instance by auditing and providing assurance about data-sets.
To realize this, as auditors we need to: have a proper understanding of what is and what is not possible with data; cooperate closely with data specialists and IT auditors; help government to improve its policies.
At the NCA we’re trying hard to make our reports better. A good recent example how we do this is our report on the National Declaration 2019.
Example of how the NCA enhances its reports
Since 2007, The Netherlands has been issuing an annual National Declaration to account for the management and regularity of expenditure declared to the European Commission for EU projects carried out here, and for the support awarded by the EU. The National Declaration 2019 renders account for €1,202.5 million in total. The money was spent via EU funds that are managed jointly by the European Commission and the Netherlands, the so-called ‘funds under shared management’.
At the NCA, we express an opinion every year on the attestations made in the National Declaration. Our report provides additional assurance on the National Declaration to the House of Representatives. The Minister of Finance subsequently informs the European Commission of our opinion.
In our 2019 report we presented, an opinion on the National Declaration 2019 and identified room for improvement. In addition, we considered the effectiveness of European farm income support.
In this infograph out of our report shows the left-hand column the income earned in 2014 by the farmers we analysed, after the deduction of income support. Without farm income support, more than half the farmers (52%) earned less than the statutory minimum wage (€19,253 per annum).
The second column shows farm incomes including income support. The gross income of just over a third of the farmers we analysed (36%) was less than the statutory minimum wage in 2014. That is 16 percentage points fewer than the number of farmers without income support. The second column also shows that the situation at the other end of the range is different. When income support is taken into account, more than a quarter of the farmers (26%) earned between the modal and twice the modal income (€33,000 to €66,000) in 2014, and nearly the same number (24%) earned at least twice the modal income (≥ €66,000).
The third column shows that over a third (37%) of the support was received by farmers who earned more than twice the median income.
By doing this analysis we have shown that not only care about compliance, but also about effectiveness and efficiency. In my view, this is becoming an increasingly important part of our work.
How NCA keeps up with technological developments
At the NCA we identified a need for a professional approach to data. At the moment we’ve organised our expertise in a ‘data hub’. The hub is tasked with increasing data analytics capabilities within the organisation, increasing data awareness, and increasing the data literacy of NCA auditors.
The number of audits involving data analysis in our office is increasing. You can find all reports on our website. Good examples of audits where we used data analyses are: a cohort study into asylum seekers 2014-2016; embassy visits to Dutch detainees abroad; the effect of the increase of highway speed-limits; access to long-term care benefits.
As the NCA, we will need to make maximum use of our unlimited access to data. This is a real asset. But when we analyse data at our office, we also struggle with correlation and causation. Let me give an example. In 2017 we published a report on the impact of a tax exemption used by parents when part-financing the purchase of their children’s house. We found that the use of this tax exemption was correlated with the reduction of total mortgage debt in the Netherlands. We’ve not found hard evidence of causation, but it was plausible.
Another example, increased funding to public schools may correlate with improvements in final exams scores, this does not automatically prove causality. Here we need to find a way to solve this puzzle.
How we try to maximize the impact of our work?
At the NCA we carry out significant number of audits that lead to audit-findings, conclusions, and recommendations. The quality of our audit work should be beyond any doubt.
An important question that we ask ourselves, right at the start of each audit proposal is: ”how do we create impact”? In this part of my opening address I will share five lessons that are, in my view, absolutely crucial.
Know your client!
For us, it is essential to be in touch with our clients continuously. You have to note here that I mean, both our auditees – such as ministries and government agencies – and the Dutch Parliament, to which we report. We are well informed when it comes to the issues and concerns of ministers and government officials at the various ministries. We can also easily adapt to changing circumstances, as there are several occasions each year when we can modify our research agenda.
Get the attention of the decision maker
At the NCA we have a portfolio of products tailored to our clients’ needs. It includes brief focused audits, conventional audit reports, infographics and animations. We try to time our publications optimally, so they appear just before a particular parliamentary session. While carrying out an audit, we try to create impact by – for instance – organising a conference, round-table discussions, etc. A recent example of this is our audit on revolving funds.
In addition, we try to pay more attention to so-called after sales. We encourage our staff to present the findings and recommendations to wider audience after the launch of the report and also repeat some of the messages when relevant debates in Parliament take place.
In our current strategy the NCA has a strong focus on healthcare and the revenue side of the national budget, and less on defence and the European Union.
We select our audit topics at several points during year. We update our audit programme at regular strategy sessions, where the NCA board and its staff discuss the options with each other. This system enables the NCA board to continuously consider different options and focus on the audits that we find most important strategically.
As mentioned before, we invest heavily in our knowledge of data techniques. In the Netherlands we are neither frontrunners nor laggards. But we are definitely a good discussion partner for our auditees. We see this for instance in our discussion with the tax authority.
We will never surrender what is essence of our work
The NCA has a long term views on some issues that, in our view, really have to change in The Netherlands. I’m talking about accrual accounting.
As some of you know, the Dutch government currently has a cash based accounting system. For a number of years, the NCA has been advocating for accruals-based accounting, because we believe that only accruals accounting tells the whole financial story in a systematic and well-organized way.
The current government made it clear that it does not want to move towards full accrual accounting yet, but has started a few pilots.
We will keep raising this issue in the future as we believe accrual accounting would enhance the information position of Parliament and improve the quality of decision making. We are counting on your support as well!
Reflection on Cigar’s work
At this point I would like to come back to the point of SAIs – including the NCA – as stakeholders of your work. The points I will now make are also relevant for other potential users of your work, like public managers, politicians and journalists. If we are to know about and learn from the results of your debates and research, it is of course helpful for these results to be easy to find and easy to understand, including by me and other people outside the research community.
In my opinion, the connection between your network and the community of practice could be strengthened. For many practitioners the world of journal articles and conference papers is very difficult to access. Which is hardly surprising, since these types of documents are not written for the primary purpose of practical use. The many footnotes and references, for example, are necessary from an academic point of view, but constitute psychological hurdles for managers, controllers and auditors who want to get to the practical message as quickly as possible – trusting the academic quality of the research.
Here in the Netherlands, the professional institute of chartered accountants – the NBA – has developed a very valuable and internationally lauded ‘knowledge sharing’ product, known as the ‘public management letter’. This letter discloses the main cross-cutting issues and problems discovered in financial audits in both the private and the public sector. It also provides recommendations on these topics, for example on risk management and on healthcare.
Following this line of thought, would it be conceivable for CIGAR to publish something like an annual ‘public management letter’ as well, based on its most important discussions and research, in which it addresses developments, opportunities and risks in public-sector accounting? Written in practical language, supplemented by images. Announced in the CIGAR newsletter. I’m quite sure that such an initiative could contribute greatly to the impact you make!
I started with a picture of the pipe-smoking Mr. Dietz who is recording changes in the National Debt. I will end this speech with a picture of some of our data scientists, Marcus, Eva, Elze and Rudi who are working together with our register accountants, performance auditors, and IT auditors on daily basis to discover how we can use data-techniques to make our work more reliable and relevant for our auditees and members of parliament.
This picture was taken three months ago during a hackaton on how our Parliament should work in 2025 in a data-driven environment where Members of Parliament can check facts in real-time. These are issues for our Parliament, for the NCA, and I suppose also for you.
For instance, imagine that we no longer need ex-post audits and move to real-time auditing? What if we spot irregularities in real-time and only report these when the annual reports are due?
I would love to hear from you on these questions, and other issues that are on your research agenda. I hope we at the NCA will learn from your work and use it in our practice. I wish you a very fruitful congress and would like thank you for your attention.