JCPSG | TRAC Guidance

Part III - Annual TRAC

Section C: Staff costs

1. All staff costs should be attributed to activities – for academic and research staff; other staff in academic departments; staff in central departments.
2. Staff costs should be directly attributed to activities where possible. These would include any staff costs incurred by the institution for Research Assistants, Research Fellows, laboratory technicians and other support staff who are dedicated to one research project, and visiting lecturers.
3. Costs of academic staff should be attributed on the basis of time allocation records.
4. This is part of the pyramid of costs

   Cost attribution to activity
   Time allocation to activity

5. An informed view from the head of department or appropriate manager should inform the attribution of the costs of other staff in academic departments.
6. The costs of staff in central departments should be attributed to Support. They should then be attributed with other costs in their respective departments to academic departments and to activities, according to cost drivers.
7. The allocation of academic staff time, and costs is of particular importance in TRAC:
   • there is little history or culture of time recording;
   • academic staff costs make up a significant proportion of the costs of institutions;
   • data on staff FTEs (in particular on staff working in research) are a significant part of the calculation of the indirect and (often) the estates charges on Research;
   • the percentages of time spent on each activity provide cost drivers used to attribute some of the other costs in institutions (although this should be minimised);
   • academic staff costs recorded by research sponsor type (for example, the Research Councils) is used to check that charges made to projects for each sponsor are fair and reasonable, in aggregate – see reconciliation.

Chapter C.1 Academic & research staff time

Categories of time

1. Staff time should be attributed to activities and to sub-categories of these activities on the basis of the time actually spent on each activity. This information is then used to attribute staff costs.
2. This attribution is carried out irrespective of who is funding the salary.
3. TRAC definitions should be followed in all cases. These are given in Annex 6.
4. The activities are Teaching, Research, Other and Support. The sub-categories are as follows:
   • the time on Teaching should be attributed between PF and NPF;
   • the time spent on Research should be attributed to each research sponsor type. PFR (institution-/own-funded) and PFR (PGRs) are two of the seven research sponsor types.
5. Where staff wholly or mainly work on one of the main activities (Teaching, Research, Other) then their costs should be attributed directly to those activities:
   • the costs of Research Assistants and Research Fellows whose time (and therefore cost) is wholly chargeable to one or more research projects should be attributed directly to Research. They may carry out Teaching and Support duties, but where the sponsor of their research project agrees to include this time in the project, then it should be attributed to Research in the annual TRAC time allocation process;
   • the costs of visiting lecturers engaged in Teaching should be attributed to Teaching;
   • the costs of contract staff engaged in services rendered should be attributed to Other (or as appropriate for the activity being carried out).
6. Where staff are working on more than one activity, time allocation methods are used to attribute their time, and therefore cost, to activities. Timesheets are not required. Time allocation methods collect time spent on each activity expressed in terms of percentages of a working year or hours.
7. The following guidance – on allocating time between activities – refers to ‘academic time’. It also applies to any other researcher (see Annex 6 for definition) who is not working solely on Research (or Other) activity.
8. When academics are completing time allocation schedules it is possible to design these either to include numerous activities, or to minimise the number of activities listed.
9. As a minimum, academics should separately record time for:

   Teaching
   institution-/own-funded Research
   externally-funded Research/PGRs
   Other (Clinical Services)
   all Other activities
   Support

10. Institution-/own-funded Research and externally funded Research/PGRs should be recorded separately as there is no proxy that could be used to allocate any ‘Research time’ total between them.
11. Clinical services should be recorded separately as O(CS) to allow its subsequent attribution to T, R, O, S on the basis of services received from the NHS (applicable to medical/dental schools working under ‘knock-for-knock’ arrangements):
    • institution staff, paid by the institution, who provide clinical services to the NHS, should initially allocate that element of their time to Other (Clinical Services) activities. This should subsequently be reattributed to T, R, O and S, depending on the services being provided to the HEI by the NHS, where possible;
    • the initial allocation of clinical activities should therefore be recorded in the time allocation system to O(CS), and collected separately from the other types of work being allocated to Other, (such as consultancy and non-Research clinical trials).
12. Annex 10 gives more detail about the attribution of the costs of medical and dental schools. This Annex includes mandatory guidance.
13. Proxies can be used to attribute time further:
    • Teaching time can be attributed between PFT and NFPT on the number of students falling into each category.

      Costs of teaching overseas and self-funded students taught on UK credit-based courses are generally recorded to PFT by academics in the first instance.

      The costs of teaching these students should then be allocated from PFT to NPFT. A proxy of student numbers can be used (the number of FTE overseas and self-funded students, divided by total students being taught). However, this calculation should be made at a departmental level, not at an institutional level.

    • externally funded Research time can be attributed to research sponsor types on the basis of numbers of RAs and PGRs – perhaps weighted; or on the basis of directly incurred costs (excluding equipment purchases).

      However, this can only be done in laboratory departments. Different patterns of research activity in non-laboratory departments, with fewer RAs, make these proxies less usable, and the heads of those departments need to be involved in advising on the allocation of their departmental time between sponsor type.

    • Support time can be attributed to T, R, and O on the basis of the time that has already been allocated to those activities.
    • information about the volume and type of services provided to the HEI by the NHS under knock-for-knock is used to attribute Other (Clinical Services) to Teaching, Research, Other or Support.
14. Proxies should be applied at a department level, not institution level.
15. The following proxies should not be used:
    • income (as the funded elements vary by sponsor, with some paying indirect costs and some not);
    • direct non-staff costs (as the levels of consumables and equipment purchased are not indicators of how staff time is spent).
16. Using proxies, rather than asking academics to record their time against multiple categories of time, does reduce the complexities of the time allocation schedule. However, many institutions believe that these proxies are not sufficiently robust. It is good practice also to collect separately:
    • Support for Teaching, Support for Research, Support for Other, general Support. This provides a more robust method of allocating Support costs to activities (and Support costs are a large item of cost). There is otherwise a danger that Support will be a large figure that cannot easily be understood or attributed to T, R and O.

      However, if these are collected separately then at least four categories should be added (Support for Other should not be subsumed into general Support, for example); and it should not be assumed that scholarship can easily be assigned directly to Teaching or Research – some will relate to all activities and therefore should be assigned to a fifth category of scholarship, or included in the general Support category;

    • PFT courses (attended by both PF and NPF students), and short or overseas courses. This categorisation can often be made by academics.

      What is difficult is for them to categorise their time on credit-based courses between PF and NPF students as the cost of teaching these two types of student is significantly the same. A proxy of student numbers can be a more robust method for allocating Teaching time on credit-bearing courses between PF and NPF;

    • time by research sponsor type. This categorisation can often easily be made by academics (who may only be involved on a few research projects, and with a few PGR students, at a time). This avoids the need to ask the head of each non-laboratory department to allocate their department time to research sponsor type.
17. Some institutions also consider collecting additional information to help satisfy some of their other costing objectives, for example:
    • lower-level activities within T, R or O – e.g. outreach activities and knowledge transfer within O, collected separately from time on consultancy
    • types of Support – preparing bids, or course committees
    • projects, courses or programmes.
18. This information can be useful in course costing, for example. However, it is not a TRAC requirement. It can add significantly to complexity.
19. When recording time, care should be taken with the following definitions:
    • scholarship, as distinct from Research – see Annex 6 - definitions;
    • the inseparability of Teaching and Research. Some academics might find that a small amount of their activity cannot readily be split between T and R. They should apportion this activity on the basis of the amount of T and R that they have otherwise directly allocated;
    • academic time spent in training and supervising PGRs. This should all be allocated to PFR (PGRs). This is despite the fact that different individuals and organizations can be paying maintenance grants, bench fees, stipends, or tuition fees – and not all of these will be PFR sponsors. This apparent anomaly is corrected when income is reported, see Annex 15 (income allocation).
    • overseas and self-funded students who attend the same courses as HEFCE-funded students. It may be easier (and more robust) if the academic reports the teaching of both of these type of student under one category, and proxies are used to allocate the total to PF and NPF, as discussed above.
20. The time on which academics are asked to report should be defined as the total time they have made available for HEI work. This is different from the number of hours they are contracted to work; or any standard working year (e.g. 1650 hours).
21. The time should not include holidays taken and ‘standard’ levels of sickness (i.e. they should not report for TRAC purposes on holiday taken within the entitlement, and sickness up to say two weeks in a year). This has important implications when calculating the indirect cost rate and estates charges and then applying these to the days available to work (i.e. excluding weekends, holidays and ‘normal’ sick leave) – see standard working hours.
22. The time should include longer periods of leave. This might arise from maternity leave, paternity leave, sick leave, secondments or compassionate leave. The allocation of time (and costs) associated with longer periods of leave will depend on whether any cover is to be charged to the department and activity – see Annex 9.
23. Any extra hours over a ‘standard working week’ (howsoever defined) will not be reported under TRAC, even if information on hours is collected by the institution. The extra hours will not be reflected in the costs, as hours are converted to percentages of time when reported and used in cost allocation. TRAC leads to the reporting of costs: it does not consider productivity.
24. Care should be taken with regard to private consultancy. There are three main types of consultancy, although not all of these will necessarily apply within any particular institution. Materiality is an important consideration here – see Annex 9.

Chapter C.1 Academic & research staff time

Time allocation methods

in-year retrospective | diaries | other

25. A robust method of time allocation should be in place and used for reporting activity by academic staff. A robust method has the following elements:
    • time allocation schedules are used;
    • that cover periods that are representative of 12 months within a three-year cycle;
    • individual academics complete the schedules. These should not be completed on their behalf by their head of department or another individual;
    • response rates, and methods of correcting mistakes, and of completing data-sets for non-respondents and leavers, are appropriate to ensure ‘fair and reasonable’ results;
    • where in-year schedules are completed, there are at least three of these completed over the year (i.e. each is less than six months in length);
    • data is collected over a three-year cycle.
26. When sampling time periods or academic staff (whether deliberately, or by achieving a response rate of less than 100%):
    • they should be representative of types of staff; of each clinical, laboratory and non-laboratory group of departments; of each research sponsor type; and of the weeks or periods in the year;
    • acceptable levels of statistical accuracy should be achieved.
27. The sample size should be robust at a lower level (e.g. by academic department, or by type of staff) if institutions are calculating indirect cost or estates rates at these lower levels.
28. Acceptable time allocation methods include:
    1. in-year retrospective time allocation
    2. diaries
    3. other – proxies, workload planning models, structured interviews and workshops.
29. Methods that are not acceptable include:
    • estimating; or using performance targets for each academic;
    • annual retrospective time allocation, by either individuals or heads of department on their behalf.
30. It is not a TRAC requirement to link the collection of time with performance appraisal, and the implications of this should be considered very carefully before it is done.

i. In-year retrospective

31. This method of time allocation involves each academic completing three or more time allocation schedules, covering a year, every three years.
32. Time is generally recorded in percentages (of available time).
33. All academics can be asked to participate in a single year in the three-year cycle (a ‘big bang’ approach) or different departments can be involved in different years (a phased approach).
34. The 12-month period can be an academic year, financial year or any other set of 12 months, as long as every month in a year is covered, within the three-year cycle.
35. At least three time allocation schedules should be collected over the 12-month period. It is good practice to consider the pattern of academic activity when designing these. Four or six in-year schedules, covering each term and each vacation might be easier for academics to complete. They represent logical time frames around which academics can remember their activity, as they are based around periods which each display a constant work profile. Alternatively, each of three schedules could contain two columns (covering activities during the term, and activities during the vacation).
36. It is not a requirement to use monthly time allocation schedules – these might be regarded as timesheets. If an institution or department does wish to introduce these, their implementation should be considered very carefully.
37. Sampling can be used. If so, the samples selected should be representative of the staff type (e.g. grades), departments in each of the three discipline areas, of research sponsor type and of each different period in the year (a period being a teaching term, summer vacation, and so on) – see Annex 9 for more detail on the use of representative departments. The same population should be tested two years running, or the data confirmed in some other way (less robust processes can be used for this) to check the bias (variability) in the data to confirm that it provides acceptable levels of statistical precision.
38. Most institutions using in-year retrospective time allocation require all academics to participate. The advantages and disadvantages of sampling are set out in Annex 8.
39. However, response rates are rarely 100%. This means that sampling has taken place. An institution should collect figures that give a robust representation of the institution's activities. This should ensure that the institution’s results are robust for each of the following:
    • at a discipline level. Figures should be robust for each of the three main groups of department: clinical, laboratory, and non-laboratory. If the institution intends to calculate and apply estates charges or indirect cost rates for additional groups of departments (e.g. part-laboratory), or for individual departments, then the results should be robust for each group or department;
    • for each activity. Institutions should confirm that the responses fairly reflect the work profile of the different types of staff and types of departments/units, in terms of their T, R or O focus. 

      They need to be representative of the different types of staff in terms of their work profile. Salary bandings/grades are often used as a proxy for this, with a further check that the different types of research unit/department have been appropriately covered;

    • by research sponsor type. The sample sizes of respondents should ensure that the data is robust for Research activity for each sponsor type (e.g. Research Councils, OGDs, charities) at an institution level;
    • if an institution intends to calculate and apply separate estates charges, or indirect cost rates, for different types of staff, or for different departments, then their time allocation should be robust at these lower levels. The minimum TRAC requirements are, however, at a much higher level than this, as set out under the first two bullet points.
40. As a rule-of-thumb, if an institution has a response rate lower than 75%-80% of the whole population, then they should confirm that their response rates are acceptable. This use of a 75%-80% response rate to indicate the need for a review does not mean that each institution has to achieve this level of response.
41. The appropriateness of their response rates should be reviewed by an individual with statistical expertise. This (combined with reasonableness reviews and proactive involvement in external benchmarking) will allow an institution to confirm that the time collection exercise has been carried out robustly.
42. If an institution does find that it is short on returns for any one time period in a particular area (for example, from professors, or from laboratory research units) then they do not have to re-collect time data from all academics or areas. All they have to do is to 'top-up' time allocation returns from a few more staff of the required type/area, before finalising their data collection for that year.

ii. Diaries

43. This method of time allocation involves academics completing a number of weekly diaries. These generally involve 24 hours, 7 days, with an activity recorded against each half hour or hour.
44. The hours are totaled for each activity and expressed in terms of percentages of total available time.
45. If the diary is based on 24/7, then it should include a specific category to cover holidays, sick, and non-institutional activity including private consultancy. This should be excluded from total available time and therefore not used when calculating percentages of time.
46. A sample of weeks in a year, and/or academics, can be used. Statistical sampling techniques should be used to determine appropriate sample sizes. Expert advice should be obtained for this. Annex 8 (sampling) gives further information and a case study.

iii. Other

47. Proxies such as workload planning models, structured interviews and workshops could be used as a time allocation method. However, these can be difficult to design robustly. They should:
    • be comprehensive – cover all academics and all activities. They should therefore cover scholarship, preparation time, institution-own-funded Research – and all available time, not contracted time;
    • be robust – a minimum of three time allocation schedules or equivalent should be collected, covering a 12-month period, completed by each academic or a representative sample;
    • reflect reality – record what actually has taken place (not what was planned or ‘standard’). If planned workloads were to be used, there should be a direct proven correlation between this planned workload and actual workload. Standards would need to be tested and shown to reflect actuals.
48. It is unlikely that many institutions currently have workload-planning models with this level of rigour.
49. Structured interviewing can be used as a method of collecting time by activity. This might involve:
    • workshop discussions (of individuals or managers – using data such as contact time, but also technical estimates such as the preparation time for each contact hour). The information thus gained would then need to be tested to confirm it applied in practice;
    • staff surveys or interviews;
    • creating ‘standard’ workload profiles for staff carrying out different types of work – and confirming through appropriate data collection that these applied in practice.
50. These methods need to be used in a sufficiently rigorous way to be acceptable under TRAC. They need to cover all periods of time in the year, all staff, and include robust testing to ensure that statements reflect actual practice. These methods are likely to be costly to implement.
51. These types of models, timetables, job descriptions and/or outputs are often useful supporting techniques, for testing data collected through other methods.

Chapter C.1 Academic & research staff time

Designing the process

52. The following table sets out the main tasks in designing a time allocation process:

agree the time allocation method to be used; determine whether to use sampling or not; agree the categories or types of staff whose time is to be directly allocated; agree the categories of staff who are to participate in the time allocation process; consult with unions (on-going); communicate with staff (on-going). Newsletters, websites, workshops and other methods of communication are important in improving their understanding of TRAC and its purpose, and its benefits. These techniques can improve response rates; agree activity categories to be recorded and definitions of these (and match with institutional terms as necessary); agree proxies for further attribution as necessary (PF and NPF Teaching; to each research sponsor type; of Support); consider: the collection of FTE information (part-time working) the collection of information on grades whether hours are recorded the use of representative departments confidentiality/anonymity the distribution of the time allocation schedule; More information is provided on each of these areas in Annex 9 design time allocation schedule: This would include definitions, format, instructions, working paper or aide memoire for use by staff. Annex 7 provides examples of time allocation schedules.

Chapter C.1 Academic & research staff time

Implementing the process

53. The following table sets out the main tasks in implementing a time allocation process:

    obtain list of staff identify those who are to be directly attributed distribute time allocation schedules answer any queries, provide support receive completed time allocation schedules and follow up non-respondents extrapolate for non-respondents and leavers

i. obtain list of staff

54. This list can describe those in post at a specific date in the year.

ii. identify those who are to be directly attributed

55. Records from the research grant administration system, and payments to visiting lecturers etc., coupled with information from heads of department, would provide this information.

iii. distribute time allocation schedules

56. The time allocation schedule can be produced on paper or electronically. A schedule on e-mail or a web-site can:
    • improve the calculation (highlighting where 100% has not been achieved)
    • improve response rates, if e-mail or web usage is part of the academics’ communications culture
    • save time on keying in data.
57. However, a system to protect confidentiality may be required.
58. Numbers or names should be allocated to each schedule (depending on confidentiality). A covering letter or equivalent should accompany this – along with easy access to definitions, instructions and perhaps the aide memoire/working paper.

iv. answer any queries, provide support

59. A help desk or support line should be provided for queries.

v. receive completed time allocation schedules and follow up non-respondents

60. Institutions should ensure that those who respond to requests to complete time allocation schedules are representative of those who do not respond. If this is not the case, then specific types of the non-respondents should be identified for follow up.
61. The decision to contact non-respondents would depend on the use of the data (whether just for TRAC or to meet other costing objectives). If sampling techniques are used, the requirement to follow up non-respondents would also depend on the (useable) response to date compared with the required sample size in each staff group, time period, and group of departments (or sub-group).
62. If sampling techniques are not used, institutions should consider whether the responses provide a ‘fair and reasonable’ result. This was covered under Section C.1 earlier in this section.
63. When contacting non-respondents, the main actions are:
    • reminders
    • estimation or other action to complete gaps.
64. In both cases some information about who has or has not responded needs to be maintained centrally.

vii. extrapolate for non-respondents and leavers

65. Time allocation data could be used as a proxy to attribute the time (and cost) of leavers and non-respondents. It would have to be robust to do this.
66. This process could be informed by a study of the beginning and end-of-year staff lists – perhaps in conjunction with the head of each academic department. It would also allow the appropriate treatment of units which have been subject to significant restructuring during the year.

Chapter C.1 Academic & research staff time

In subsequent years

67. Once collected robustly in the three-year cycle, time allocation data does not have to be collected again within those three years.
68. Data from one year can be used as a proxy for the next year, provided there are no significant changes to activities and structures. Institutions do not have to repeat any one method of time allocation year after year, once they are satisfied that the method meets TRAC requirements, the data are fair and reasonable, and there have been no material changes from year to year in their activities.
69. Institutions should survey heads of department, and use high-level criteria, each year to ensure that they identify areas that have changed materially. These criteria could include mergers, numbers of staff, numbers of students, RAE rating and type of work, volume of income from different sources, changes in staff:student ratios, changes in the ratio of support staff to academic staff.
70. If there have been material changes in a department (size, type of activity etc.) then institutions should assess the impact on the allocations of time. New data (if required) can be collected through discussions with heads of department.

Chapter C.2 Academic & research staff costs

1. Staff costs should be attributed to activities on the basis of the time spent on each activity by each individual.
2. The main steps required in this are set out in the table below:

   design spreadsheet or database system collect costs for attribution record time collect and record other cost drivers attribute costs calculate totals and test for reasonableness report; review and amend methods as required reconcile Research Council costs

i. design spreadsheet or database system

3. A simple set of spreadsheets, or equivalent, can be used to record staff time and attribute staff costs. Key fields and functions would be to:

   record time and other data:

   • department;
   • staff number (indicating grade/salary banding – see Annex 9 for more);
   • each of the activity categories against which time is recorded (PF Teaching, NPF Teaching, the seven Research sponsor types, between Other (Clinical Services), and the rest of Other, Support (and types of Support if institutions wish));
   • FTE percentage of staff – see Annex 9;

   record costs from the financial accounts:

   • total staff costs;

   assist in the attribution of costs:

   • staff costs that have been charged to research projects as a directly incurred cost;
   • the names and costs of academics who are paid to only work on one activity (visiting lecturers etc.);
   • time collected from the annual time allocation process, weighted by staff FTEs – see Annex 9;
   • staff salaries to be applied to each individual’s time record;
   • other cost drivers.

ii. collect costs for attribution

4. These should be total staff costs reconciled to the financial statements – see Annex 9. They include some costs that are not derived from the payroll (e.g. visiting lecturers and visiting researchers).
5. They should be recorded at the level of individual members of staff.
6. The percentages of time allocated to activities should be weighted to recognise the different pay levels of individual staff. Either calculated average staff salaries for each academic/research staff grade or pay banding, or the actual salaries of each individual academic or research staff, can be used. They should include on-costs and allowances and aggregate to total staff costs, reconciled to the financial statements.

iii. record time

7. The time allocation schedules will provide reference number or name, percentages by activity, or hours, and, possibly, a FTE percentage.
8. If hours are recorded, percentages should be calculated. If staff time is being grouped before applied to costs, then the percentages of time should be weighted by FTEs – see Annex 9.

iv. collect and record other cost drivers

9. Cost drivers or proxies can be used to attribute time to lower levels of activity where necessary, for instance:
   • student numbers – for attribution between PF and NPF Teaching;
   • numbers of RAs and PGRs – for attribution to research sponsor type for laboratory departments;
   • head of department percentages – for attribution to research sponsor type for non-laboratory departments;
   • information about the volume and type of services provided to the HEI by the NHS under knock-for-knock – for attributing Other (Clinical Services).

v. attribute costs

10. Attributions would:
    • firstly, directly attribute staff costs where possible (e.g. Research Assistants to Research);
    • use the time allocation data, and salary, for each individual or group of individuals (weighted by FTEs) to attribute their costs to the categories of time collected;
    • attribute the costs of cover for maternity and sick leave so that the activity is assigned the costs of the person carrying out the work, or the costs of the individual who is on leave, but not both – see Annex 9;
    • use proxies to attribute time between PF and NPF Teaching, and by research sponsor type, and from O(CS) to T, R, O and S, as necessary;
    • use all other time – including for both academic and research staff – (or another cost driver) to attribute Support time to Teaching, Research and Other, as necessary.
11. This attribution should lead to academic staff costs being assigned to:
    • Teaching, Research, Other, and Support
    • a split of Teaching between PF and NPF
    • Research shown by research sponsor type
    • Other split between (Other Clinical Services), and the rest of Other
    • department (and therefore discipline group).
12. The financial accounts will already hold cost data against individual research project accounts. This should be used when attributing costs that have been directly incurred (Research Assistants, Research Fellows, other staff who are wholly dedicated to a research project, or who have completed project-level timesheets).
13. Many directly allocated Research costs will, over time, be recorded against individual project records. However, the DA academic staff costs that have been directly allocated to a project under TRAC fEC should not be used to attribute staff costs in the annual TRAC process. That charge was made on an estimate. Only actual time, recorded through the time allocation method, should be used in attributions carried out for the annual TRAC process (and therefore the calculation of the indirect cost rate and estates charges).
14. Academics sometimes receive a payment (made through their institution’s accounts) in addition to their salary, for their work on a particular course or project. The attribution of costs to that activity (e.g. NPF T or R) should be based on the proportion of the total payments to that individual that represents the amount of their time spent on that activity. However, the attribution could alternatively be based on the additional payments, if this is unlikely to lead to a materially different allocation.
15. Similarly, any attribution to Other, for academic staff time spent on trading company activity should reflect the actual time (and costs) spent. Only actual time recorded through the time allocation method, should be used in allocations for the annual TRAC process.

vi. calculate totals and test for reasonableness

16. Totals (percentages and costs) should be derived for each:
    • type of staff (salary banding/grades)
    • department
    • activity
    • lower-level activity including research sponsor type
    • Support costs as distinct from other costs
    • sample size/response rates.
17. Figures should be tested for reasonableness. Techniques for doing this are described in Part II (internal review).
18. Academic staff costs attributed to Support form one of the elements of the total Support costs in an institution. Support costs are used to calculate the indirect cost and estates rates.

vii. report; review and amend methods as required

19. Time data is used to calculate the costs reported externally in the annual TRAC return. Time data is not disclosed separately in this return. It can, however, be useful in sector benchmarking exercises.
20. The institution’s own experience of implementing, the experience of the rest of the sector, together with the QA processes, are likely to continue to identify scope for improvement. This might require a change in method, or some modifications to the current methods.
21. It may include, for example:
    • considering requesting time by sub-category of Support (Support to T, Support to R, Support to O, General Support);
    • improving the reallocation of O(CS) to T, R, O, S.

viii. reconcile Research Council costs

22. At the beginning of year five (2008/09) of TRAC fEC implementation, most projects which started prior to early 2006 (when the Research Council funding methods changed) will have been completed. Therefore, academic staff costs will be charged during 2008/09 on almost all externally sponsored projects undertaken in that year.
23. From December 2009, and annually thereafter, institutions are able to, and should, compare the total academic staff costs attributed to the Research Council sponsor group via the annual TRAC time allocation exercise, with the total academic staff costs charged to Research Council projects under TRAC fEC in the accounting records.
24. It is good practice for this comparison to be made on academic staff time as well as costs, but only if institutions are able to calculate the total of academic staff time charged to research projects (by sponsor) in a year. This may have costly systems implications for some institutions, and is not a minimum TRAC requirement.
25. This should only be regarded as a broad comparator, as there are various reasons why the two totals will be different, for instance:
    • the use of two different costing methodologies – based on actual time in annual TRAC (but ignoring the total amount of time actually worked), compared to the standard working year assumed in TRAC fEC (based on an average likely working year, which will, in practice, vary by individual);
    • fluctuations in the average time estimated on live projects in any one year compared with the actual time recorded under the annual TRAC time allocation for that year of the projects’ lives;
    • the impact of the less significant changes in project staffing, e.g. where the RA leaves three months before project-end, and the PI takes on a greater part of the work than originally estimated;
    • the normal slippage in the final completion of a research project, taking into account publications;
    • the difficulty in estimating time where the research outputs, the emerging requirements from the research as it develops, and the quality of the staff are not fully known;
    • there may still be no time or (full) cost records being kept for every project that started before the beginning of 2006 and is still ‘live’;
    • the difficulty in separating out research time from research student supervision, especially when the student is an inseparable part of the research team and is not a project studentship;
    • time spent supervising project studentships is recorded as part of the ‘PGR’ total in the annual TRAC time allocation, but this may or may not be in the time charged to Research Council project cost records. (Even if this time is charged to a Research Council project, that will not mean it will be funded on that basis);
    • the time of staff from other institutions who are collaborating on a grant will appear on the lead institution’s project costing record but the staff themselves will not have participated in the lead institution’s TRAC time allocation survey;
    • knock-for-knock, where O(CS) time reallocated to Research for the annual TRAC time allocation may not be matched against time from clinicians recorded on the project cost records (other types of resources may be provided to the HEI as part of the arrangements);
    • use of salary bandings to calculate charge-out rates on projects;
    • actual increments and pay awards differing from those assumed in the charge-out rates and project indexing.
26. RAs work partly on Teaching and Support, but all of their time is being recorded as Research. However, as this treatment is the same for both TRAC project costing and for the annual TRAC time allocation exercise, there is no reconciling item here.
27. These differences will be inherent in the comparison of total costs of time charged against Research Council projects with the costs of time actually spent (available from the annual TRAC time allocation exercise). Despite this, the comparison should provide confirmation (or otherwise), at the level of the institution as a whole, taking one year with another, that the Research Councils are not being charged more than the salary costs that the HEI is incurring on those projects.

Chapter C.3 Support staff

Support staff in academic departments

1. The cost of staff who have been charged to research projects as a directly incurred cost (e.g. some laboratory technicians) should be attributed to Research in the annual TRAC process. Where they have completed timesheets, the balance of their time not charged to Research should be attributed to Teaching, Other or Support as identified through those timesheets.
2. The Research time of all other laboratory technicians should be directly charged to research projects as a directly allocated cost (at least by 1 August 2007). As part of this, an analysis of the time of this group of staff would have been carried out, and should inform the attribution of their annual cost to activities – see Part IV Charge-out rates.
3. The costs of other support staff in academic departments is a Support cost (and forms part of the indirect cost rate).
4. To this can be added the costs of RAs who are ‘in-between’ projects and who are not engaged significantly in institution-/own-funded Research, or Teaching. Their costs should be attributed to Support.
5. These Support costs should then be attributed to activities using a robust approach. These costs should not automatically be attributed to activities on the basis of academic staff time, nor should it be assumed that they only relate to Teaching. An example of a robust approach is to provide heads of department or equivalent with a list of these costs and ask them to indicate the appropriate percentage split between Teaching, Research and Other.
6. Institutions could alternatively ask these staff to complete time allocation schedules (perhaps on an annual basis) but the administrative burden of this should be carefully considered – it is not a TRAC requirement.

Support staff in central departments

7. The costs of staff associated with residences, catering, conferences, should be attributed directly to Other. Some other staff (for example, working in university consultancy companies or other trading companies) should be attributed to Other where appropriate.
8. All other staff costs in central departments are attributable initially to Support. The costs of staff in each functional area should then be aggregated with non-staff costs in that area, for attribution to Teaching, Research and Other – see Chapter D.2. These totals (Support costs for Research, for example) are then included in the costs that form part of each indirect cost rate calculation (or, in the case of the estates staff, the estates charges).
9. In particular, staff associated with estates, or premises, should be attributed to a cost pool containing all estates costs. Staff associated with libraries or learning resources should be attributed to a cost pool for libraries.