This is article 5 in a series about designing internal governance.
I once worked in an organisation which needed to procure many large construction projects. Nearly all of these were subject to European Union procurement rules, which make specific steps in the procurement processes mandatory, and which create some short deadlines despite lasting months overall. It’s a bit like a game of Snakes and Ladders (but without the ladders): if things go wrong you may have to start all over again. Consequently, inability to control the processes well is a serious risk. This article is about how to get delegation and escalation right.
Unfortunately, the general level of authority delegated to the local Board was a fraction of the contract value for most of the procurements, meaning that awarding these contracts needed the approval of the main Board. They normally met only about every 2 months, and had little knowledge of the issues. Control of that process was out of our hands. Very careful planning (and significant extra work) was required to minimise the risk of failing to get properly-authorised approvals at the necessary times, and so avoid the ‘snakes’.
Fortunately, once we were able to demonstrate that we had well-managed internal governance, more appropriate delegations were granted. Nevertheless, the example illustrates the problems that can arise when the level of delegations is not appropriate to the required decisions.
Delegation and escalation levels: getting the balance right
Deciding what authority to delegate requires balancing the need for control with the need for practical delivery, which includes the need for timely decisions. The place to start is the levels of authority to be delegated. What will meet the business need, taking account of the volume of decisions are required, and which it is practical to make in the time available at different levels? The membership of the bodies to which delegation is made then follows: it should be chosen to ensure that the members have the competence they will need. If you start with the membership, and then decide what decisions they can be trusted with, the result may not deliver what the business needs.
A delegated authority generally consist of two parts: a subject and a value or scale limit. The subjects will probably be the same at different levels in the hierarchy, but the scale limits will vary. A decision which exceeds the limit at one level must be escalated up the structure, perhaps with a transition from Individual to Collective Authority part way up. It is convenient to articulate both the subjects and the authority limits in an authority grid or scheme of delegations. Obviously there will normally be one or more columns for financial and budget matters. A programme organisation might also require columns such as changes to delivery time, changes to scope and requirements, and quality assurance (see example grid below). Obviously, other kinds of organisations will have other requirements.
How will you decide what the authority levels should be? This is another question of balance, this time between the desire to encourage ownership and accountability at lower levels, and the need to retain senior accountability for the bigger decisions. While this can never be a precise science, limits should generally be set so that the proportion of all the decisions which come to a body (or indeed an individual) and which must be escalated to a higher level is expected to be fairly small – perhaps 10-20%. This pattern, repeated at each level and interpreted intelligently, ensures that the Board (and each level below it) sees what it needs to see, without being overwhelmed by volume or delaying decisions unnecessarily. Equally importantly, each lower level feels it has the trust of its superiors and is empowered to do its job.
Authority grid
While setting such limits for financial decisions is straightforward because the escalation levels are easily quantified, it is not quite so simple (although just as important) in other areas. However, in these cases practical qualitative tests can work perfectly well. The illustrative authority grid shown includes examples of tests which are still sufficiently objective to be useful.
| Category | Requirements | Financial/Budget | Time | Quality | Control |
| Level A | Set overall strategic direction | Budget including contingency
Larger budget changes |
Approve third line assurance | Appoint Level B
Cancel a programme |
|
| Level B | Approve programme scope & requirements
Approve changes to scope & requirements |
Change allocations within portfolio contingency | Changes which delay benefits delivery | Approve second line assurance, commission third line assurance | Appoint Level C
Initiate programme and approve business case Close programme on completion |
| Level C | Approve design and outputs as meeting requirements
Approve design or output changes which meet requirements |
Change allocations within programme contingency | Changes with no impact on benefit delivery | Approve first line assurance, provide/ commission second line assurance | Appoint Level D
Options selection Approve stage gates, output acceptance & readiness to proceed Accept key milestone completion |
| Level D | Change allocations within project contingency | Changes with no impact on key milestones | Commission first line assurance | Accept project milestone completion |
Note that it is normal for Boards to have ‘Reserved Matters’ which are not delegated at all; the same principle will apply at lower levels as well. Just because there is a box on the authority grid does not mean it has to be filled!
Principles to establish:
- That the authority grid should include qualitative as well as quantitative delegation limits, and should not be restricted to areas where a financial limit can be set
- That escalation levels will be set to provide an appropriate level of filtering, decided on the basis of business need and practical delivery
- That decisions about committee membership should be made in the light of the delegation levels set