A web-based Revolution in Australian Public Administration?

Delivered at the 2nd AustLII Conference on Computerisation of Law via the Internet, Australasian Legal Information Institute (AustLII), University of Technology, Sydney, Australia, 21-23 July 1999.

Contents Abstract 1. Introduction 2. Determinative Processes in Public Administration   2.1 What are Determinative Processes? 2.2 Problems of Complexity, Change, Accuracy and Consistency 2.3 Determinative Processes and Service Delivery 3. The Platform for Revolution 3.1 Rulebase Systems 3.2 Decision Support Systems 3.3 The Internet 4. Case Study: Department of Veterans' Affairs   4.1 ELMNet 4.2 Special Rate 4.3 Just in Case 5. The Revolution: Implications for the Public   5.1 Intelligent Data Collection 5.2 Absorption of Complexity 5.3 Outsourced Services 5.4 Self-service 5.5 Ownership of the Process 5.6 One-stop Shops 6. Conclusion  Download

Abstract

This paper discusses the potential for government agencies to move complex business transactions to the Internet, facilitating true self service. It argues that merely allowing simple interactions with government over the Internet only utilises a fraction of the potential offered by web-based systems. In particular, decision support systems which remove logical and other complexity for users offer a realistic mechanism for bypassing the traditional and flawed forms of interaction with government. The paper illustrates how one government department is moving down the revolutionary path to true self service.

1. Introduction

This paper discusses the profound implications of decision support technology for expanding service delivery options. It argues that ubiquitous public sector aspirations for one-stop shops, electronic service delivery and life-event based service delivery will only lead to a shallow level of service delivery, unless technology is used to enable deep transactions, complex determinative processes, to be completed in a highly automated way. The paper shows how various technologies are converging to enable the delivery of these transactions through the Internet.

For ordinary people, this represents a revolution in the way in which they interact with government. More than half the Australian population directly relies on government agencies of one form or another for a significant portion of their weekly income. Other sectors of the population rely on government agencies for decisions that critically affect their lives or businesses. The use of public-access decision support systems offers the opportunity to reshape the way they interact with government.

2. Determinative Processes in Public Administration

2.1 What are Determinative Processes?

Most government agencies that deliver services organise much of their work around determinative processes . These are processes of assessment against a body of rules: determinations of entitlement, of obligations and of rights.

Where a government agency administers legislation, the staff of that agency are likely to make a great many determinations as to how the legislation applies to specific client cases. This is often the primary function of the operational staff.

Every level of government does this. At the federal level, we have agencies that determine pension or benefit entitlements, compensation, medical benefits, immigration visas, taxation liabilities, rights to subsidies and bounties, compliance with regulatory schemes and entitlement to grants. At the state level, there are agencies that assess and collect revenue, administer workers' compensation, determine environmental and land planning rights and obligations, adoption rights and child welfare rights and obligations, administer investment guidelines, fair trading laws, residential and commercial tenancy rights and determine their own levels of grants, subsidies and payments. At the local government level, the determination of rights to use land consumes vast amounts of staff time and client anxiety.

For most members of the community, this is largely what government does. Most people's direct contact with government comes at a time when they are obliged to seek some determination on how legislation applies to them. These processes largely define the relationship between a government and individual members of the community.

How do members of the community feel in this situation? What do they experience? We suggest that there are three experiences and perceptions that are almost ubiquitous:

The more significant the determination and the more vulnerable the client will be in the face of the determination, the more these perceptions will be accentuated.

How correct is the perception that government agencies administer legislation correctly? How inevitable is it that clients feel a sinking loss of control and a general sense of reliance when dealing with government agencies that determine their rights?

2.2 Problems of Complexity, Change, Accuracy and Consistency

Inside government agencies, the issues are obviously different. In our experience, the following sketch summary is true:

Our experience has been that agencies do not conduct audits on the accuracy and consistency of their staff's determinations. They audit performance against key indicators, but these are usually quantitative, focusing on throughput. To make some measure of the quality of primary decision-making, they look to rates of administrative review and the results of those review processes. At best, this is hopeful, evincing a sense that the incidence and subject matter of reviewed cases accurately reflect underlying decision patterns. At worst, it is naïve or dishonest, where there is some sense that the rate of inaccuracy is reflected in the volume of cases going to review.

On the basis of file audits that our clients have conducted, we can make the following generalisation: if only 25% of an agency's primary decisions are wrong, then the agency is doing well. This figure flies in the face of community perceptions. However, in discussions that we have had with senior managers in many government agencies, this does not appear to ring untrue. Rather, it appears to reflect the suspicions and concerns of those senior managers.

Complexity and change exacerbate the problems of securing accurate and consistent determinations. It is difficult for front-line staff to keep up-to-date in a volatile policy environment. It is particularly difficult where they have stressful workloads, and high levels of public contact.

Agencies deal with complexity and change through a number of devices. They stovepipe processes, so that each staff member only needs to remember the rules and procedures for a limited range of policy. They use delegation levels to segment case loads, channeling the more complex cases to more experienced staff. They use training, policy manuals and instructions to try to keep their staff up-to-date.

Yet there is something that is inherently wrong in these attempts to create a determinative assembly line. It is immediately apparent when a case moves out of the administrative domain and into the legal. This usually happens on review, although it can happen for some processes at the start, when legal advisors take charge. Lawyers (good lawyers) seldom make snap judgements from memory as to how legislation applies. They read the legislation. They identify what provisions apply and why. Where there is an issue of how a word or phrase applies to the immediate situation, they review policy or case law to see whether there is any guiding precedent. Naturally, the indulgence in terms of time and thought that is afforded within the legal process is not available in the administrative process.

But this is why the primary decisions are often wrong. The decisions can be quick, good or cheap: you can have two but you can't have all three. In some agencies, you can't have any.

2.3 Determinative processes and service delivery

There are currently two standard service delivery methods for determinative processes.

There are some areas in which true self-assessment is a reality. A prominent example is personal income tax assessment. However, this is not a dominant method.

The Commonwealth Ombudsman recently released a discussion paper, Balancing the Rights: Providing information to customers in a self-assessment income support system, that argued that the Australian income support system is in fact a self-assessing system. This argument suggested that where it is incumbent upon a person to know their rights and entitlements, to seek the appropriate entitlement and to volunteer sufficient and appropriate information in order to secure that entitlement, then this is effectively a self-assessing scheme.

On this analysis, most government determinative processes could be characterised as flawed systems of self-assessment.

In the discussion paper the Ombudsman reiterated the concern expressed on a number of other occasions, including in annual reports and in a special report on oral advice issues, that the self-assessment systems which certain agencies operate, 'unfairly transfers a high level of risks to its customers'. (para 2.23)

The impact of this transfer of risk to customers is potentially significant. As customers put to the Ombudsman:

In the Ombudsman's view government agencies on whom clients rely for income support entitlements should:

The standard service delivery techniques, and the observations of the Ombudsman, illustrate a truism that lies behind the traditional service delivery models: you have to know the rules, in order to be able to apply them.

The fundamental techniques and structures of much public administration are built on this assumption. Only trained staff can administer legislation. Only experienced, knowledgeable people can determine rights and entitlements. Only specialists can accurately administer complex policy and statutory material.

This assumption represents a major barrier to expanded service delivery techniques. While agencies speak of electronic service delivery, they limit this to shallow transactions: updates on the status of a case, down-loading forms, on-line registration or lodgment. While agencies speak of single points of contact and one-stop shops, do they mean anything more than similar shallow transactions delivered over the counter?

Determinative processes are deep transactions . They are complex processes that form the meat of much government service delivery. Radical client benefits, profound electronic service delivery and major efficiencies in service delivery will only come with the capacity to deliver these deep transactions in new ways.

And the key to this is breaking a time-honoured axiom: making it possible to make determinations, without knowing the rules.

3. The Platform for Revolution

Three significant technology components are merging to enable change in the delivery of services:

3.1 Rulebase Systems

Rulebase systems automate the administration of rules. Rulebase technology is a programming technique. It is not particularly new, nor is it particularly clever. However, its application can be clever. It is profoundly appropriate for determinative processes.

Before discussing how rulebase systems work, and how they can be used, it is salutory to remind ourselves of some other areas of specialist automation that have taken place over the last 25 years.

Calculators were the first ubiquitous device to automate complex mathematical/logical processes. Initially, they were anathema in schools and universities. If you could use a calculator, you wouldn't learn how to calculate cos, sin or tan, you wouldn't be able to determine a square root from first principles, you wouldn't be able to use logarithm tables effectively. Clearly, the calculator was a seriously deskilling device.

Spreadsheets became the next ubiquitous device to automate otherwise specialised or time-consuming mathematical/logical processes. They revolutionised accountancy. It might be said that they deskilled some accountants. What they certainly did was provide the means for large numbers of non-accountants to undertake tasks that were previously beyond them.

Both calculators and spreadsheets are now completely accepted tools for particular tasks. They have redefined how those tasks are performed. They have made it possible for people to operate more effectively and efficiently, and for agencies to organise certain types of work in different ways. Just as the word processor spelt the end of the typing pool, so spreadsheets have changed the role and expectations of managers.

With both devices, there is a common acceptance that if the parameters that are input are correct, then the result will be correct. This may be subject to some known constraints, but it is generally true. They are trusted as the legitimate means of performing common tasks accurately and efficiently.

Rulebase systems are simply another means of automating logic. They automate the structural logic of sets of rules (such as legislation). They are frequently and unfortunately known as expert systems, a hangover from early research and commercial characterisations. This nomenclature hides their character. They simply automate logic processes.

The administration of determinative processes, the application of rules, involves two types of interpretative skill. The first is dealing with logical intricacy. Legislation is intricate: provisions interact with each other, interpretative provisions must be applied within substantive provisions, complex alternative provisions interact with each other, lengthy provisions contain complex sub-sections, paragraphs, sub-paragraphs and so on.

The complexity of legislation largely derives from its logical intricacy. When government agencies, Ministers, staff or clients bemoan the complexity of legislation, they are almost always referring to the logical intricacy of the material: the apparently impenetrable web that has grown through countless amendments, targeting exercises, fixes and extensions.

The second type of interpretative skill that is required to administer determinative processes is the capacity to determine whether a brief phrase accurately characterises some aspect of a fact situation. This is the classic legal issue of interpretation: whether a product is 95% fat, whether a houseboat is a house, whether a person took reasonable steps. These are exercises in judgement.

Logical intricacy can be reliably automated. Judgements cannot.

Rulebase systems can reliably model complex or intricate sets of rules. They can automate the process of investigating those rules, asking appropriate questions of the user to find out whether the facts required exist in the immediate case. They can apply the rules to determine a conclusion: whether the legislation applies. They can explain which rules apply: how the legislation covers or excludes the immediate case.

Rulebase systems make it possible for a person who does not know a set of rules, to apply them accurately. Along the way, that person will need to make judgements: interpreting phrases to see whether they cover the immediate case. However, this is a general human skill, and can be aided by the provision of explanatory material: elaboration on the meaning of the phrase, summaries of precedents, examples and so on.

Rulebase technology has the capacity to change the fundamental techniques of determinative processing. Its use is highly analogous to that of calculators and spreadsheets. It has the capacity to change the methods and the quality of many functions in public administration. It is a specialised and an appropriate technology for administering rules.

3.2 Decision Support Systems

A decision support system augments a calculation device like a rulebase with significant features to fully support users in the exercise of complex transactions. They take the computer from being a smart calculating device into the realm of a powerful tool for decision-making.

To fully support someone in the exercise of a complex transaction a decision support system must incorporate (at minimum):

• an intuitive easy interface;
• access to relevant supporting information;
• document generation facilities; and
• links to long term storage.

Users must have an intuitive user interface to help them navigate, enter information and access supporting material. This is particularly important for self service, where the level of user can vary widely.

Wherever the user must make a decision or enter information, they need access to relevant supporting material tailored to the particular issue before them. Ideally this information should not be hidden two or three clicks away from them, but be available directly where it is required. If the person has to search through several levels of user help for the correct material then it is unlikely to be much use.

The combination of an easy user interface, access to research information on individual issues, and rulebase technology that applies complex decision logic enables inexperienced users to reach a reliable conclusion on how complex legislation applies to a particular fact situation.

Once the system has helped the user to reach a conclusion about a particular right or entitlement, document generation facilities are important to provide full documentation of the decision. This documentation will generally include tailored claim forms, a proper audit of the decision and potentially a letter or report summarising the entitlements. The word document in this context has a broad meaning – it range from something as simple as a Word processor document to a spreadsheet with summaries of amounts payable.

A decision support system can be a useful reference tool for providing information and advice about entitlements. However, if it is linked to a relational database back end, then it can truly form the platform for improved service delivery. For example, someone can log into an Internet site, use the system to determine that they are eligible for a payment, bounty, license or other entitlement and electronically lodge their claim for that entitlement while they are there. This removes the need to fill out any paper forms and go through the standard cycle of information collection and decision-making which has proven so problematic in the past.

3.3 The Internet

The technology supporting decision making has been around for some time now. During the 1980's and 1990's, numerous decision support systems were developed. However, there were several problems with distribution of such systems, both internally within organisations and also to the general public.

The first problem was one of access to processing power. Many organisations would have a central mainframe computer which did not have sufficient surplus processing power for this type of activity. On the other hand, the early personal computers struggled to make a Word processor perform well, let alone a complex system incorporating rulebases, calculations, document generation and data storage. In the modern age, client-server computing has gone a long way to resolving these problems. In comparative terms, processing capacity abounds at both the server end (for example your Internet service provider's web server) and the client end (for example a standard home computer).

The second problem was a consistent operating system. Prior to the advent of the Internet and web browsers, it was essential to make publicly available software accessible on multiple operating systems: DOS, Windows, Mac and Unix. Even if you could make things run on the operating system, it was difficult to ensure that the user interface was consistent across different systems.

The third problem was purely physical: how could you distribute a system widely? Floppy disks did not offer storage for substantial systems, networks were in their infancy and CD-Roms were just emerging as a viable technology for digital distribution.

Then came the Internet.

In many ways, decision support systems were dormant while the world technology stage reached a point at which they could be delivered with ease. With the advent of the Internet and its global internetworking capabilities, the distribution problems were solved. Now, an organisation wanting to take advantage of decision support system technology really only needs:

• a customer base with direct or indirect access to the Internet (in our experience, at this stage this is more likely to be indirect through an agent, but this is changing);
   
• access to a server machine capable of running a Web server and the decision support software; and
   
• domain experts with access to computing skills to enable the capturing of their knowledge in a decision support system.

4. Case Study: Department of Veterans' Affairs

The Department of Veterans' Affairs administers a range of entitlements for veterans, including:

• service pensions, payable to veterans who have rendered particular types of service in connection with the Australian, Commonwealth or allied defence forces;
   
• disability pensions payable to veterans to compensate them for service-related illness, injury and death - these include a difficult determination: whether a claimed illness or injury resulted from service; and
   
• different types of concession cards, payable on the fulfillment of detailed eligibility criteria.

In 1992, the Department had been criticised by the Australian National Audit Office for its administration of the compensation program in particular. It was clear that primary decision-making in the program was highly inconsistent, that liability was sometimes being accepted on inadequate grounds, and that the calculation of entitlements was unreliable.

The Department redesigned the way in which the program was to be managed. The new design had three key elements: new and far more detailed policy, a new computer system incorporating rulebase technology to support the decision-making, and a re-engineering of the processing function. The Department was one of the first to adopt rulebases as a mechanism for implementing complex policy. The Department was one of the first to adopt rulebases.

In the first full year of operation, the program processed 60% more claims in an average of 60% less time, with 30% fewer staff, operating at lower delegation levels than before. Overall, the Department achieved an 80% productivity improvement, while simultaneously improving the quality and consistency of primary decisions.

Other Australian organisations committing to the use of decision support systems include the Department of Defence (in particular the Military Compensation and Rehabilitation Scheme), Comcare, Centrelink and the Australian Tax Office.

Given the background of improvements to service delivery achieved by their new rulebase-supported administration, the Department now has the confidence to take its first steps down the revolutionary path towards self service. It is now moving a number of its deep transactions to the Web. These include:

• the initial determination of eligibility for pensions administered by the Department;
   
• the determination of eligibility for special rates of pension where a veteran's disabilities are particularly severe; and
   
• in conjunction with the Department of Defence, the determination of which compensation scheme is appropriate in the event of an illness or injury (the Veterans' Entitlements Act or the Military Compensation Scheme).

A number of user groups are expected to take advantage of this shift to the Web. These include:

• ExService Organisations (ESO's) which generally represent particular groups of veterans;
• the veteran community itself, including the dependents and spouses of veterans;
• third party advisers to veterans such as tax agents or accountants.

4.1 ELMNet

The rules relating to eligibility for service pensions and disability pension are quite complex, both in terms of their logical intricacy, and also the historical knowledge required about Australia's involvement in various global and regional conflicts.

In 1997, the Department began to redress this problem by capturing the rules in an eligibility module for use by in-house staff. The module captured:

• all of the rules relating to service history, residency criteria and a person's status as a veteran under the Veterans' Entitlements Act;
   
• maps and other material to help decision makers identify whether people had served in eligible conflicts;
   
• supporting help text written by historical and domain experts; and
   
• a suite of standard letters automatically generated by the system.

The module has subsequently been used in-house by decision makers to help them determine eligibility for service pension and disability pension.

In 1998, the Department sought to expand the use of the module and make it available externally online as ELMNet . The initial module was expanded to include:

• full determination of eligibility for service pensions, and eligibility to claim disability pension (the module now has about 5000 rules in it);
   
• expanded help text and historical material;
   
• distribution through the Internet via a central web site;
   
• a tailored advice letter generated to the circumstances of the user; and
   
• the automatic generation of a subset of the standard claim forms used by the department.

ESO's and veterans were involved in the development of the user interface for the module, so that it would be easy for them to use over the Internet. Selected groups have had access to the module for a number of months to assist with testing the module. The module will be released to the public shortly, allowing veterans and their advisers to electronically complete most of their claim documentation.

At this stage, the veteran still needs to submit the paper claim forms for verification as a formal claim. They can attach the tailored letter generated by the system to substantiate the claim. The next logical progression is to remove the need for the paper forms completely. This will happen once the electronic form is formally recognised as a claim form, which the Department can do once the appropriate security infrastructure is in place.

4.2 Special Rate

A special rate of pension is aimed at veterans with significant incapacity to work as a result of war-caused injury or disease. There are a number of different categories of recipient, including those who could have worked past the age of 65, permanent, temporary and intermediate rates.

The determination of eligibility for a special rate of pension is particularly significant to the Department and its client community:

• the legal rules for determining whether or not someone is eligible for a special rate, and if so when that payment commences, are very complex;
   
• the policy underlying the legal rules is not always very clear;
   
• the revenue implications of incorrect decisions are very severe – the granting of a special rates of pension costs the taxpayer potentially hundreds of millions of dollars; and
   
• as a result of the complexity of the legislation, regional practice has sprung up over the years, with a concomitant inconsistency in decision-making.

In 1998 the Department revised and clarified its policy on special rate and embarked on a program to deliver the new policy into the veteran community. There were two key elements to this process.

Firstly, there was a consultative process over the life of the policy development. In its early stages, the Commission (responsible for determining the application of the Veterans' Entitlements Act by the Department) worked on a policy paper. Later on this paper was distributed not only internally within the Department, but also externally to ESO's and members of the veteran community for discussion and comment.

Secondly, in parallel with the development of the policy, a decision support system was implemented to help capture the policy. This also assisted in the process of clarifying the rules to be applied in the determination of special rates. The rigour of capturing the rules in a central rulebase helped to highlight where there were differences in opinion between the leading experts, let alone the front line of decision makers and the veteran community.

Now that the special rate policy has been settled, it is being implemented internally within the Department through the decision support system. In conjunction with ELMNet and Just in Case, that system will also be deployed externally onto the Web for use by veterans and ESO's. The complexity of the determinative process has been absorbed to the extent that self service is not only possible, but the best way of educating the veteran community in the new policy.

4.3 Just in Case

The Department of Defence runs a program related to the scheme under the Veterans' Entitlements Act - the Military Compensation and Rehabilitation Scheme. This scheme is administered under the Safety, Rehabilitation and Compensation Act, applied to members of the forces.

In conjunction with the Department of Defence, DVA has a publication entitled Just in Case which advises veterans and serving members as to which Department they should approach with their claims. The main decision making aid in the publication is a table cross referencing different service types and dates with potential eligibility under either scheme.

Most people have difficulty in using the paper publication to make the decision about which organisation to approach, so the underlying rules were captured in a decision support system this year. The system will be released on the Internet shortly after ELMNet and the Special Rate Module. Development of the system has been a collaborative effort between DVA and Defence, and it links in quite naturally to the other two modules.

5. The Revolution: Implications for the Public

The adoption of rulebase systems by the public sector can have profound implications for the public itself. Below, we discuss some of the major opportunities:

• enhanced data collection processes;
• absorption of complexity allowing transparent decisions;
• outsourcing of services;
• self-service;
• greater ownership of the process; and
• one-stop shops and similar government service delivery points.

5.1 Intelligent Data Collection

Some agencies tout the provision of access to existing forms on the Internet as their move into electronic service delivery.

Static electronic interaction is not a great deal more than a gimmick. The shift to the electronic media should herald a radical shift in the way that things are done - the way that services are delivered. To give a client the capacity to print a standard paper form over the Internet may be convenient, but it misses the point. Electronic media provide the opportunity for dynamic, rather than static interaction.

Decision support systems provide an opportunity for agencies to replace data collection based on generic paper forms with far more dynamic and effective electronic processes. They can provide intelligent, interactive, tailored data collection processes, which collect more detailed and more relevant data in less time than it takes to complete a generic form. They will be far easier for the user to navigate than a conventional form. And at the end of the data collection process, the user actually gets the result. ELMNet is a good example of such a system – the user actually has some idea of their entitlements at the end of the process, and is a long way down the path to formally lodging their claim.

An appropriate decision support system will conduct an intelligent dialogue with a user. It will investigate the legislation directly, asking questions of the user. As each question is answered, the system will draw any conclusions possible from that conclusion. So for example, if a particular subsidy cannot be paid to a company with more than 15 staff, and the user identifies that their company has 30 staff, the system will immediately stop investigating that subsidy, and will only look at alternative subsidies to which this limitation does not apply.

This intelligent, directed conversation is more natural and less confusing to users than forms. The user does not have to be given directions such as 'If YES, go to question 12'. The system applies that logic. The questions can be framed in terms of the client's personal situation: 'When was Lisa born?' The user will only be asked questions that are actually relevant to his or her case, in the light of information already provided. So, for example, the system may ignore blocks of questions or even pages of questions that would appear on a paper form, because they are irrelevant to the case at hand.

This intelligent data collection process does not simply make form-filling less intimidating and less difficult for the user, it allows the agency to collect more detailed and more useful data. So, for example, in a workers' compensation context, if it is clear that the claimant has a back injury, a host of special questions that only apply to backs can be asked. If the claimant suffers from stress, a host of questions on that subject can be asked. The fact that the data collection process is dynamic rather than static makes it more natural and more powerful.

Furthermore, because the data collection process is intelligent, it complies much more closely with the Privacy Principles for each jurisdiction. The relevant information, and generally only the relevant information is collected about each person. Compare this to a paper form in which it is common for a proportion of the information to be irrelevant to the person's circumstances.

If the data collection process is simultaneously the determinative process, then the agency has profoundly changed the dynamic of interaction with the client. The client will obviously see more benefit and will be more disposed to seek out this means of interaction in the future.

5.2 Absorption of Complexity

For most staff administering complex legislation the underlying policy is too difficult to remember, particularly if it is volatile. As evidenced in the Ombudsman's reports referred to earlier, this is even more true of clients themselves. Most members of the public are not sure why they need to provide the mass of information requested on most forms, and have no idea whether the decision made by the Department is the correct one.

Decision support systems can absorb the complexity associated with the process of applying complex legislation in two significant ways for members of the public.

The first is that by providing comprehensive instructions with every question or data collection field required by the online system, the process is considerably more transparent. If someone knows that they must answer questions about income in order to assess them against an income threshold for payment, they are much less likely to be suspicious about providing the information.

Using the traditional paper format, it is too costly in terms of space to explain in detail all the questions on a claim form. Most people already complain about the shear size of the documents they are required to fill out. As with something like Taxpack, for most organisations their claim forms would become books if they explained themselves fully. In the electronic medium, these problems are removed. The user can be given quite a clean interface with easy entry points into all of the supporting material.

Secondly, the system can actually explain the reasoning behind any conclusions that are reached. This is very significant when allied to the capacity to change answers in a what-if style of interaction. In a non-threatening environment, the person can see what effect changes in their circumstances will have on their entitlements and work to arrange their affairs to get the best result for themselves.

5.3 Outsourced Services

The use of rulebase technology provides an agency with the capacity to outsource determinative processes with a measure of well-founded confidence.

Because a decision support system guides the user and encapsulates the logic of the decision process, there is less scope for error. Errors will not arise from lack of knowledge of what the rules are, or from lack of knowledge about changes to rules. They will not arise through workers applying local folklore instead of the correct rules, nor through the application of irrelevant considerations.

Errors can still arise through poor judgement on a particular issue. However, the capacity of a decision support system to automatically record a complete audit trail for any decision provides an appropriate means of managing this risk. It is possible to monitor decision-making practice in detail, and to identify errors precisely, so that a policy agency can effectively audit an outsourced decision-processing agent.

Decision support systems enable outsourcing of determinative processing, because they support a peculiarly appropriate model for outsourcing. The policy agency can develop and maintain the decision tool, largely regulating the manner in which the agent goes about the decision-making process. The policy agency has control of policy change and implementation, through the provision of new releases that encapsulate the new rules. The policy agency has the capacity to precisely monitor the performance and decision quality of the agent, and can therefore effectively hold the agent accountable.

This enhanced capacity for outsourcing is complemented by an enhanced array of outsourcing vehicles. While an obvious style of outsourcing is to commission a single external agency to provide all of the processing, this is not the only model available. Because a decision support system enables precise auditing, a keenly competitive outsourced environment is possible. Multiple agents can be licensed, and their performance very accurately assessed.

However, decision support systems open the possibility for a far more creative approach to off-loading determinative processing. Widely dispersed, local agents, who provide convenient and trusted environments for clients, can safely be licensed using this type of system. They can be commissioned to provide a service on a continuing basis, or on a fee-per-transaction basis. This opens the way for significant government services to be reliably provided through neighbourhood centres, community centres, accountants, pharmacists, doctors, schools, childcare centres, service stations, gardening shops – whatever outlet makes most sense for a particular type of transaction.

These community-based service providers may rightly see no significant value in simply providing their clients with access to government forms or pamphlets. However, the provision of a capacity to complete transactions with government agencies may be highly complementary to their own services. This capacity can be very attractive to those providers, if the transactions effectively enhance the services that they wish to provide for their clientele. When the punter gets a result, rather than a pamphlet or form, they will see more value in the transaction.

5.4 Self-service

The ideal service delivery vehicle for determinative processing is, in many cases, self-service.

Within self-service we include the range of mediated or assisted self-service that effectively provides the same result. It therefore includes self-service with help from a spouse, a son or daughter, a parent, a friend or relative, a neighbourhood person who 'knows how to do it'. It is also self-service, rather than outsourcing, where a community or commercial agency provides the facilities and assistance for its clientele to complete transactions, at no cost to the government agency. We suggest that this model of self-service will become very widespread over the next 3-5 years.

Using a decision support system, it is demonstrably as easy for a person to process their own determination as to complete the relevant form. The provision of information through a tailored, intelligent on-screen interview is usually simpler than the navigation of a form of any substance. And with a decision support system, at the conclusion of this process the determination is a by-product. This determination may need to be supplemented by provision and sighting of appropriate evidence (or it may not in some cases).

Self-assessment, or assisted self-assessment, obviously requires changes to a government agency's approach to and support for determinative processes. But the benefits to the client, in terms of convenience, confidence, a sense of control and the capacity for hypothetical scenario testing are compelling. Simultaneously, the benefits to the government agency are compelling – vast reductions in unproductive data entry, vast reductions in staff time spent on mundane aspects of processing, fewer interruptions through phone calls and enquiries, and the capacity to apply more resources more effectively to essential tasks of compliance auditing and verification.

5.5 Ownership of the Process

The real significance of outsourced services and self service, is the shift in ownership of the process. This fundamentally changes the power relationship between a member of the general public and the government body to whom they are beholden.

Many people are concerned about answering forms because they are worried that they will be held to account for any mistakes they make on the forms. This risk can be significantly reduced when all the questions are explained fully, and the consequences of the information being entered are immediately apparent.

Where a person must come into an organisation with very little knowledge of their own, and present to someone who is keying information in about them behind the bureaucratic wall, they are in a significantly weakened power situation. Where they can work with someone with whom they are comfortable, in relaxed surroundings, then they are much more likely to feel that the process is for their own benefit.

Some agencies are uncomfortable with this power shift. Inevitably it requires a greater focus on audit and checking that people have been providing the correct information. However, much legislation is targeted for the benefit of people in the community, not for the convenience of the organisation administering the legislation. So the new relationship is a more accurate reflection of the way things should be.

5.6 One-stop Shops

One-stop shops run the risk of being pamphlet counters. They are currently a ubiquitous response to the stove-piping of traditional government service delivery. They can be cross-agency service delivery points, or single points of contact within an agency.

So, how are you really going to provide clients with a staff member who can confidently and sequentially resolve a client's taxation problem, the grandparent's immigration application, the daughter's social security claim, the son's hotel licensing issue and the car registration?

This example is perhaps fanciful. But if a one-stop shop is to be anything more than a pamphlet counter, it must actually deliver services, rather than simply delivering information about services. The conventional means of delivering government services is through trained, knowledgeable staff, who apply that specialist knowledge to resolve specific issues.

Decision support systems are not magic bullets. But they do address a primary problem facing a one-stop shop. They reduce the need for specific, arcane knowledge, and accentuate a requirement for generalist, pragmatic skills. They empower a government officer to be able to reliably perform a process with which he or she is broadly familiar, without requiring knowledge of the minutiae of the content or effect of the process.

If decision support systems enable government officers to complete processes and make determinations with which they are not specifically familiar, then a one-stop shop can truly look to offer an array of services across divisional and portfolio boundaries. It is at this point that the shop becomes an effective service delivery point, meeting the aspirations of government rather than simply providing a shadow presence for continuing, stove-piped bureaucracies.

6. Conclusion

Electronic service delivery has a very limited scope, unless it can truly include deep transactions: those determinative processes that currently require detailed knowledge of policy or process. When a government agency can administer deep transactions electronically, when it can provide reliable access to immediate resolution of a complex application or claim, then it is providing real electronic service delivery. It is at this point that the government agency shifts to a primarily electronic basis for the delivery of this type of service, rather than simply having a convenient adjunct to continuing traditional determinative processing.

This paper has discussed how three technologies are converging to enable true electronic service delivery: rulebases, decision support systems and the Internet. The intersection of these technologies provides the opportunity to deliver a level of electronic service delivery that profoundly and productively changes the nature of government and its relationship with the community.

For most members of the public, this represents a revolution in the way they interact with government. This approach to service delivery does not simply provide a client with convenience. It can effectively address entrenched and intractable problems with the quality of manual primary decision-making in government. It can provide the client with a range of services not previously possible – testing of hypothetical scenarios, cross-portfolio packages of determinations, determination of optimal packages of services.

Ultimately, the most important client benefit is dignity: it provides a means whereby the client can control the process of seeking entitlements, within a comfortable and trusted environment and with well-founded assurance that the decision is unprejudiced and equitable. It is this change in the relationship between government and citizens that constitutes the real revolution that this style of service delivery offers.

---

This is a Conference Paper published on 29 February 2000.

Citation: Dayal S et al, 'A web-based Revolution in Australian Public Administration?', Conference Paper, 2000 (1) The Journal of Information, Law and Technology (JILT). <http://elj.warwick.ac.uk/jilt/00-1/dayal.html>. New citation as at 1/1/04: <http://www2.warwick.ac.uk/fac/soc/law/elj/jilt/2000_1/austlii/dayal/>

---