Government cloud contracts need a credible exit budget

A government may have the contractual right to leave its cloud provider and still be unable to afford the move.  

When renewal approaches, the public service must keep running, but the replacement needs testing and the team must find time alongside its existing work.  

A right that cannot be exercised within the agency’s financial and operational limits offers little negotiating room. 

The budget discussion should therefore begin before the contract is signed.  

Alongside the cost of running the service, decision makers need a credible estimate of what changing it would require.  

The purpose is to preserve a usable choice when price, performance or public needs change. A sensible renewal may still be the best decision. 

Singapore’s Government on Commercial Cloud illustrates the value of a common approach to adoption, monitoring and cost control.  

GovTech Singapore identifies MyCareersFuture (an online portal that helps job seekers find jobs) and GoBusiness (a one-stop online portal providing various government e-services from business registration to licensing and grant applications) among the services developed on the platform.  

Its 2021 discussion of cloud operations also recognises the difficulty of coordinating applications and data across providers. Flexibility has to be weighed against the complexity of maintaining it. 

Incorporating exit budget into cloud procurement 

An exit budget brings that trade-off into a procurement decision.  

I would start with one service and one plausible transition, then ask the service owner, technology team, procurement officer and finance team to produce a shared estimate.  

The destination could be another provider, a different operating arrangement or retirement of the service. Each implies different work. 

The estimate should follow that work: preparing data, changing applications, testing security and access, training staff, obtaining supplier assistanceassistance, and paying any applicable transfer or termination charges.  

It should also include the period when old and new arrangements must operate together.  

A low transfer price can distract from the larger cost of keeping a public service available during the change. 

The Welsh Government’s migration to the cloud, documented by the UK Government Digital Service in 2020, makes the preparation visible.  

As an outsourcing contract approached its end, existing arrangements continued to support ordinary operations while another team carried out the transformation. 

The case concerns the work needed to change an operating arrangement while maintaining continuity. 

The account reports that a data migration took two weeks after two years of preparation on data and records management. It also describes maintaining legacy technology during the transition.  

Those details explain why a transfer window alone cannot describe the resources a move requires. 

Planning, staff time and the continuing service all belong in the financial assessment. 

An estimate becomes useful when its assumptions are visible.  

How long might the overlap last? Which application needs rebuilding? What supplier support is included in the contract? The team should show how costs change if a difficult component takes longer than expected.  

This gives the approving officer a range to assess, with clear reasons for uncertainty, rather than a reassuring total nobody can defend. 

The next question is who could authorise the spending.  

A technical exit plan can be sound while the budget process makes its timing impossible.  

Finance should identify the relevant approval route, likely decision dates and treatment of any overlap across financial years.  

Addressing constraints early 

The agency can then recognise a funding constraint while there is still time to address it. 

This does not require placing the entire estimated amount in a separate fund. 

Some preparation can be included in normal operating budgets; larger transition expenditure may need a later investment decision.  

What matters is a recorded route from a justified decision to usable resources. An assumption that someone will find the money during a difficult renewal deserves scrutiny. 

Australia’s guidance on lock-in, portability and exit planning supports a proportionate approach. 

It calls for estimating exit costs, identifying accountable owners and testing relevant arrangements, while recognising that provider-specific services can offer worthwhile benefits.  

It also warns that diversification adds cost, complexity and capability requirements. Maintaining several clouds is therefore one possible response to a particular risk, with its own business case. 

The estimate should be tested at a useful scale.  

For example, an agency could export an approved sample of data, check that another authorised environment can read it and record the staff effort and supplier assistance required.  

A tabletop exercise could then examine the approval sequence and continuity arrangements.

Neither proves that the whole service can move, but each can expose assumptions worth correcting. 

How governments use external cloud without giving up control 

Keep the resulting record with the contract and revisit it when the service changes materially or before a renewal decision.  

A new managed database, a larger dataset or a change in the internal team may alter the work involved.  

Procurement can use the updated assessment to negotiate specific obligations for documentation, assistance, data return and deletion, subject to security and retention requirements. 

An agency can begin tomorrow by asking the owner of one critical cloud service for three things: an estimated transition cost, the official who could approve it and the evidence behind the estimate.  

Where those answers are missing, there is useful work to do before the next renewal.  

A credible exit budget helps government choose to stay because the arrangement delivers value, with a practical alternative available if that changes. 

----------------------------------------------------- 

The author is an economist and professor at Getulio Vargas Foundation, a higher education institution and think tank based in São Paulo, Brazil. His work concerns finance, regulation and public policy.