Performance Budgeting Course, Guatemala, May Prepared by Marc Robinson 1

Transcription

1 Performance Budgeting Course, Guatemala, May 2014 Prepared by Marc Robinson 1

2 What are performance targets? When do targets become a form of performance budgeting? What are the potential downsides of targets? How best to set targets Case study of UK Public Service Agreements system 2

3 Expected level of performance Numerically specified objective Timeline for achievement Examples: Increase literate 11 year-olds from 63% to 80% by 2002 Reduce unit cost of passport processing by 5% by

4 Performance indicators All targets are based on indicators But targets are not indicators Indicators don t set goals or timelines Indicators are only measures Unquantified objectives E.g. increase police investigation success rate E.g. improve efficiency of visa processing Valid on a continuing basis Statements of aspiration e.g. zero complaints; abolition of poverty 4

5 Targets become a form of PB when Targets are linked to the level of funding provided Targets are set in the budget process Targets focus on results outcomes and outputs Targets are often part of program budgeting systems Targets set for some or all program performance indicators Requirement that targets set in budget submission Relationship between target and funding No mechanical link between targets set and level of funding required to meet the target 5

6 Are targets damaging because of perverse effects? Should targets be set for all performance indicators? o Or are some indicators unsuitable for targets? o Because of, e.g., too little control? What type targets are best? o Output targets? Outcome targets? Should targets always be annual? o Or sometimes only medium-term? How to ensure that ministries take targets seriously? Consequences of failing to meet targets? Funding cut? Managers fired? Something else? 6

7 7

8 Public Service Agreement system As operated before radical change in 2007 Performance targets at centre of system Set as integral part of budget process Targets intended to be related to resourcing A multi-year budget process Spending Review every 3 years, sets 3 year budgets for ministries Targets set as part of the Spending Review 8

9 Proportion 11 year-olds meeting a defined literacy standard to increase from 63% to 80% by 2002 Proportion meeting numeracy standard to increase from 62% to 75% by 2002 Reduce health inequalities by 10% as measured by infant mortality and life expectancy at birth Reduce crime by 15% and further in high crime areas. 9

10 Focus on small number of targets Outcomes target focus Objective from beginning, but took time Initially, many output & even process targets Improvement targets mainly Raises contextual factors issue Limited controllability of some targets E.g. unemployment and poverty rates Government position: Outcomes are what matters 10

11 Comprehensive Spending Review (1998) CSR targets First set of PSAs published Spending Review 2000 SR targets Significant changes to CSR model Spending Review 2002 SR targets Greater continuity, architecture refined Spending Review 2004 SR targets Introduction of standards focus on outcomes Comprehensive Spending Review CSR07 30 Objectives A very different framework: Objectives linked to set of performance indicators. Very limited use of targets Source: drawn partly from Prime Minister s Delivery Unit 11

12 Great success against many targets, e.g. Education (literacy, numeracy etc) Waiting times in health system cut Even where the targets weren t fully met And in areas where role of external factors was unclear Cancer & heart disease mortality rate reductions Crime rates. 12

13 2007 Spending Review Targets completely de-emphasized National performance objectives set 30, reflecting key policy priorities Delivery Agreement for each PSA objective Set of indicators for each PSA objective 2-6 per objective A few of these indicators have targets But most don t 13

14 Supposed to be linked: Additional funding tougher targets But outcomes can t be linked closely to funding Contextual factors In reality, link is pretty loose Can t punish just because targets not achieved Process to consider why not achieved Negotiation as part of budget process gives targets muscle 14

15 Gaming and perverse effects Of course, some truth in this But on the whole, exaggerated Selective target setting & perverse effects Wrong targets chosen or emphasized E.g. waiting times less important that care Micro-management Loss of professional and local autonomy No space of local choice On service priorities 15

16 16

17 Is it really a good idea to set targets for, e.g. Childhood obesity? Unwanted teenage pregnancy rates? Controllability issue Indicators which are highly effected by external factors Targets only for relatively controllable indicators? Outputs, and more controllable outcomes E.g. literacy levels Does it make sense to set annual targets for, e.g. Disease incidence rates? Set only medium-term targets for some indicators? 17

18 Agency incentives to set targets too low Arbitrary target setting errors A bit better Big leap: stretch target Informed target-setting Start with good baseline numbers Identify trends and contextual factors Consultation with front line Achievement of more effective units Studies of scope for improvement 18

19 Manipulation or falsification of data Some proven episodes of falsification Example of manipulation: Storming of accident and emergency wards at time waiting times are measured Data is not falsified, just misleading Proper design and measurement needed 19

20 Unintended adverse consequences A problem with PIs in general Increase with pressure to perform on PIs Issue particularly with targets Types of perverse effects: 1) Unmeasured performance dimensions suffer e.g. quality 2) Lack of incentive for excellence 3) Avoiding difficult clients/cases Careful design to minimize perverse effects But can t eliminate problem entirely 20

21 Perverse effects appear exaggerated Waiting times example Research of Kelman & Friedman (2007) Ethical commitment to service Public service motivation Professional ethics Quality assurance mechanisms 21

22 Need for plans to achieve targets Dialogue on achievement versus target Performance report Discussion of what went wrong Intervention/support for serious failure Business Process redesign support PM s Service Delivery Unit in UK Incentives for target achievement Financial, managerial freedoms etc 22

23 Target setting is a great performance tool Despite the critics Set targets carefully and selectively Gradualism in increasing target role Targets appropriate to level at which set 23