THE introduction of a new performance management system at the Federal Board of Revenue (FBR) certainly merits serious evaluation. Having been involved in designing such a system in several organisations, I have a few suggestions for those instrumental in this system`s design and implementation.

First, it is most desirable to create objective data before implementing the normal distribution curve model. Without this, assigning a percentage of officers to be placed into different categories can be highly subjective and discriminatory. How to avoid this flaw? It is important to review the last five years` evaluation data of the organisation to identify the number of officers who were placed in different performance grades.

This could provide an objective basis for deciding the percentage of officers to be placed in different categories forapplying the normal distribution model.

Second, given the humongous size of FBR as an organisation, the introduction of any new management system would require some pilot testing before cascading the system across the board. This practice helps in removing glitches, if any, for successful implementation of any major change for ensuring the buy-in of the target groups.

Third, including the `honesty` parameter into a digitalised evaluation system will create more problems than it might be able to solve. Measuring honesty, it ought to be realised, is a very abstract and subjective construct. Its assessment is mostly based on perception.

Corruption in an organisation like the FBR is highly complex. It must be assessed carefully, lest it emds up creating largescale demoralisation in the organisation.

Lastly, bringing about changes within public-sector institutions and organisations is a good step, but extra care ought to be exercised, or, else, there will be a wave of resistance to the intended reforms.

Dr Zafar Iqbal Qureshi Lahore