In the Case Studies section:
Design-Reality Gap Case Study No.3
Computerised Integration of Two Pension Funds in Southern Africa
Case Study Authors
Mmapula Kekana and Richard Heeks
Organisation
The national Welfare Agency administers a variety of social security funds, including a National Pension Fund (NPF) paid to those who retired normally from work, and a Workers' Compensation Fund (WCF) paid to those forced to retire because of workplace injury, or to their dependents in case of death. It receives contribution payments from employers and employees.
Application Description
The application was the planned introduction of IT into the Welfare Agency in order to computerise and integrate and decentralise the previously-separate, previously-manual, previously-centralised operations of the NPF and the WCF. This involved a national network of PC workstations and servers using a single national database of contributors and claimants.
Application Drivers
The Welfare Agency's two Funds were paper-based, resulting in many problems - the Agency had no idea how many members it had or overall/individual sums owed and owing. Payments both into and out of the Funds were delayed, and often made or recorded incorrectly. Because records had to be held on paper, all significant activities were centralised at the Agency office in the nation's capital, with outlying offices merely acting as claim submission points. This caused delays and errors. Finally, the two pension schemes had no way to talk to each other. This led to lack of an overall picture, to payment errors, and to duplication of a lot of activities - for example with employers all being sent two separate bills and payment records; one from each Fund. All of this created a head of steam for change. Externally, this came from pensioners who suffered from late or under-payments, from employees who feared such problems in the future, and from employers who were wrongly charged and suffered the costs of having to interact with two Funds. Internally, the change was supported by the Agency's Board for two main reasons. Politically, they wanted a better image and fewer complaints for the Agency. Financially, the Agency had a few years before been granted parastatal status which cut off government funding. It therefore had to cut costs to ensure Fund sustainability.
Stakeholders
The employees and employers who were members of the two Funds represent the largest group of overall stakeholders, with fund claimants (retired pensioners and their dependents) being the specific external stakeholders most likely to be affected. The Minister of the line Ministry of Work and Welfare was the main political stakeholder. Other external stakeholders for implementation were two multinational consultancy organisations, and an IT supplier based in Western Europe. Internal stakeholders were the Agency's Board, including its CEO, and the remaining staff of the Agency.
Design-Reality Gap Analysis
Design-reality gap analysis compares the assumptions/requirements within the application design with the reality pertaining just before that design was implemented along seven 'ITPOSMO' dimensions:
- Information : the design was based on the pre-existing basic data available in reality on paper contributor and claimant forms, but it envisaged a change in the type of information being handled within the organisation, particularly a change from the initial reality of separate data in the two pension Funds to integrated data. It also envisaged a range of new management information being made available. Overall, this represented a medium/large design-reality gap on this dimension. Gap score: 7.
- Technology : the design assumed the introduction of a network of high-powered PCs with database software. The initial reality was paper-based operations, a few old PCs used as word processors, and poor quality telecommunications infrastructure. This created a large design-reality gap on this dimension. Gap score: 8.5.
- Processes : the design assumed automation of a series of pre-existing processes, with the overall framework for service processes (payments in and out, claims, etc) remaining broadly as per the initial reality. The main difference between design plans and initial reality were integration of Fund processes, decentralisation of processes so that local offices could process and pay as well as register claims, and new management decision-making processes. This created a medium/large design-reality gap on this dimension. Gap score: 7.
- Objectives and values : the design assumed that the objectives of the project (more accurate and timely payment of funds, lower costs, coordination and decentralisation of the twin payment funds) were shared by all stakeholders. In reality, prior to computerisation, all the external stakeholders supported these objectives, and there was also support from the senior managers and politicians. Staff at lower levels were worried about possible transfers or even job losses, and so did not share the design objectives. Overall, there was a medium design-reality gap on this dimension. Gap score: 7.
- Staffing and skills : the design incorporated two main skill requirements. First, in relation to implementation, that there would be a set of skilled staff able to implement a complex, networked information system. Second, in relation to operations, that all Agency staff would have a reasonable level of computer skills. In initial reality, neither of these requirements was met. In terms of staff numbers and location, the design assumed a redesignation of roles and locations as duplications were removed, and as responsibilities were decentralised to local offices. It also assumed that there would be a number of job cuts through automation. This reflects a large design-reality gap on this dimension. Gap score: 8.5.
- Management systems and structures : the design proposed the opening of new sub-regional offices, and the integration of previously-separate management systems and structures between the two Funds. This produced a large design-reality gap on this dimension. Gap score: 8.5.
- Other resources : there were what seemed to be fairly generous allowances for the project timescale - approximately three years - within the overall design, and the initially-designed budget was within Agency resources, particularly given planned savings from job cuts. This meant an overall small design-reality gap on this dimension. Gap score: 1.5.
- Overall : there was an overall medium/large design-reality gap in this case. Total gap score: 48.
Design-Reality Gap Reductions During Implementation
A medium-large/48-score design-reality gap is likely to bring significant problems for an e-government application.
However, the Welfare Agency took actions during the implementation process that helped to reduce the size of the gaps:
- Because of the reality of a lack of in-house and even in-country skills, the Agency changed that aspect by hiring consultants from a nearby country. Likewise, because of the reality of a lack of in-house skills for data conversion (from paper to computer), the Agency changed that aspect by hiring an external firm to undertake the data conversion work. While this successfully reduced certain elements of the staffing and skills design-reality gap, it opened up the gap for other resources by raising the designed expenditure.
- An organisation-wide user training exercise was undertaken. This successfully closed the operational skills gap.
- The design was altered to remove the losses of clerical and other staff jobs that had initially been built in. This helped partly close the objectives and values, and the staffing and skills gaps. However, it did open up the other resources gap by removing the potential savings initially designed in.
- Most crucially, the whole project was undertaken in an incremental manner rather than - as initially suggested - being rolled out in one go to all offices. Thus the system was first piloted in the national office. Once that was found to work (partly - see below), the system was then incrementally rolled out to other offices. This broke down the overall design-reality gap.
As a result, what was judged to be a medium/large design-reality gap at the start of the implementation process had been reduced somewhat to a medium design-reality gap later in the implementation process.
Evaluation: Failure or Success?
The Agency's computerisation project was partly successful, and partly not, as one might well predict from a medium design-reality gap.
In terms of successes, the project has improved service delivery:
- For fairly simple benefit claims, lead times have come down dramatically. For example, processing a funeral grant used to take around three months. Now, with valid documentation, the process takes just 30 minutes.
- Employers now receive a single unified monthly bill for the two Funds that is generally accurate and timely. This has made debt-chasing a faster and more confident process.
- Numbers of complaints has been reduced, and business can be transacted direct at a local office.
- Managers can receive reports giving overall statistics such as new members added, contributions collected, and so forth.
This sounds very good, but it only relates to those parts of the system that have actually been computerised. The initial plan was for 100% computerisation within three years. In practice, after six years, only 40% of the system has been computerised, with the remaining 60% - including the main claims processing system - still undertaken manually. Thus, the benefits to claimants, reduction in complaints, availability of management information, all remain limited in scope. In part, this has related to hiring the consultants who had certain skills, but lacked others - particularly experience of developing complex claims systems.
Secondly, the envisaged cost cutting has not happened. Because of the need to hire consultants and the data conversion firm, costs spiralled from US$30m to US$60m (and far more in local terms given simultaneous and continuous currency devaluation). Because of the agreement to avoid job cuts, there has been nothing concrete to set down on the credit side against this debit.
Recommendations: Reducing Design-Reality Gaps
One of the good practices of this e-government project was that it was participative, ensuring that the design and implementation process involved a broad range of stakeholders. Quite radically, this extended to cover external as well as internal stakeholders. A project board was created consisting of ten members - a chair plus three members each from employers paying into the Funds, employees/trade unions paying into the Funds, and government. This ensured that the design represented the objectives and values of these key groups.
The other good practice was the incremental approach taken, which has broken the overall e-government project down into a set of smaller, discrete sub-projects. Within each of these sub-projects, the design-reality gap is significantly smaller than it would be for the complete project.
Finally, this project shows some of the challenges that can emerge in trying to reduce design-reality gaps in order to increase the likelihood of e-government success: making changes can result in one gap closing, but another opening up. Hiring consultants reduced the skills gap but increased the finance gap. Project implementers must therefore think carefully about the gap impacts of their actions.
Case Details
Author Data Sources/Role : Documents and Interviews; No Direct Role
Outcome : Partial Success/Partial Failure. Reform : eServices (improving public services).
Sector : Social Services (Social Security).
Region : Southern Africa. Start Date : 1996. Submission Date : January 2003