On the basis of the Defence Budget Audit led by George Pappas the Government has charged the Department of Defence with finding $20 billion over the next ten years to help fund the ambitious capital program set out in the Defence White Paper. Defence has cheerfully acquiesced in this, and earlier in the year the former Secretary gave everyone a “shape up or ship out” message.
Somewhere in the system there is someone who struggles with numbers, and concepts like lead times and lags.
To start with, what the Defence Budget Audit found was an estimated savings range of $1.3 billion to $1.8 billion per annum. This suggests savings over a full decade of $13 billion to $18 billion. So where did the $20 billion number come from?
The best answer seems to be by a simple summation of two numbers to be found in George Pappas’s letter to the Minister covering the submission of his report:
...Defence has committed to find the operational cost savings identified in the Defence Budget Audit, which will total $15 billion across the decade. Prior to the Audit, Defence had identified operational cost savings from individual groups and services, worth $5 billion over the decade from 2009/10 to 2018/19. These savings have been effectively integrated with or replaced by the Audit savings.
There seems to be a substantial element of double counting here: the $5 billion which had been found by Defence has been “effectively integrated with or replaced by” the savings found by Pappas’s review. We don’t know how much of the savings found by Defence were replaced by the Audit savings, but I would hazard that it is a large proportion.
Also, that $15 billion requires closer scrutiny. It sounds like a nice round number in between the upper and lower estimates, but when you take lead times into account it is right at the upper limit of what is achievable. The reason for this is:
- The Audit identifies annual savings of $1.3 billion to $1.8 billion, with a one-off saving of $218 to $398 million.
- It states (page 12) that the amount of reform required to capture these savings will take 3 to 5 years.
- If we postulate the most successful end of the range of possibilities, Defence could in three years make the reforms required to achieve the maximum annual savings of $1.8 billion per annum. If we assume that the efficiency gains from reform can be gained uniformly day by day over the three year period, the total recurrent savings in the ten year period are $15.3 billion. Add the maximum one-off saving of $400 million and we have an estimated maximum achievable amount of $15.7 billion.
- At the bottom end of the envelope of possibilities, Defence would take five years to achieve recurrent savings of $1.3 billion per annum. On that basis, again assuming uniform (straight line) progress towards the target, recurrent savings would be $9.75 billion. Add to that the minimum one-off saving of $218 million and the savings could be as little as $9.87 billion.
- It is worth bearing in mind that the achievable savings is the sum of savings identified in a large number of areas, which are stated in the form “savings of up to $x million”. For each of those components there is only a slight (indeed infinitesimal) probability of reaching the maximum number, because the assessment is that there is zero possibility of achieving more.
- To give an example, if it is estimated that a proposed series of measures would yield savings of $40-50 million, there will be a probability distribution of achieving any given sum across that range, probably highest at or near the centre of the range. Provided the measures are implemented, the assessment is that the savings would be at least $40 million – zero probability that it would be only $40 million, rising to maximum probability at around $45 million, declining to zero again at $50 million.
- What this means is that when we add a range of programs that will yield savings of “up to $x million” it is heroic indeed to postulate that the achievable savings are the sum of the maxima. It is much more likely to be the sum of the middle of each range: some programs will go close to realising their potential, others will fall substantially short.
I remain of the view expressed earlier, and from a different analysis, that finding anything like $20 billion from the Defence budget in the next ten years is an impossible dream.
For an earlier post on this subject see Defence savings: the impossible dream.