JCPSG | TRAC Guidance

Part V - TRAC fEC

Section C: Estimating researcher costs

1. The time and costs of researchers should be included on all research project costing schedules.

Definition of researchers

2. Researchers include both academic and research staff – see researchers.
3. Academic staff include investigators (principal investigators and co-investigators); research staff include research assistants and research fellows.
   

DI or DA

4. Researcher salary costs can be directly incurred, ordirectly allocated:
   • the costs of research staff dedicated to a project are directly incurred;
   • the costs of academics that work on several projects and activities are generally directly allocated.
5. The key distinction between these is that:
   • directly incurred costs are charged on the basis of actual salaries and actual time   either the full cost of a salary is charged to a single project, or actual salaries are attributed to several projects on the basis of project-level timesheets; whilst
   • directly allocated costs are charged on the basis of standard charge-out rates and estimated time (no timesheets are required).
6. Research Assistants and Research Fellows who can be classified as directly incurred and who do not have to complete project-level timesheets are those:
   1. who are considered by a funder of a research project to be wholly working on research. All of their time can be charged against that project (including any time spent, with the funder’s agreement, on Support or Teaching activities); and
   2. all of their actual salary costs, as incurred, are being charged against that project.

   (They may be working on more than one project, and in this case, all of their time would be defined as direct research effort, charged to those projects.)

7. Other Research Assistants or Research Fellows may work on a number of projects, often alongside a Principal Investigator, and also carry out a range of other activities (Teaching and Support). Unless they meet the two criteria above, or complete project-level timesheets, below, then they would normally be classified as a directly allocated cost rather than a directly incurred cost.
8. Academics in some departments do already complete project-level timesheets, because of EU or other requirements (not TRAC related). Timesheets are not a requirement of TRAC. However, if these staff wish their costs to be treated as directly incurred, then:
   1. the department should be a discrete management unit;
   2. all academic staff in that unit should complete monthly project-life timesheets over the whole life of the project; and
   3. the timesheets should comply with TRAC (they cover all activities, and all hours worked in a year; and use 1650 hours to calculate salary and indirect costs).
9. Costs charged as directly incurred on a Research Council project can be vired between other directly incurred cost items on that project. However, preparing timesheets does involve a considerable amount of work and will not be acceptable to most academics. It is not required by TRAC.
10. This chapter mainly focuses on the methods to be used for academic or research staff that work on several activities, not just on research. Their costs are normally classified as directly allocated.
    

Standard working hours

11. During the allocation of staff time and salaries, hours may need to be converted to FTEs, or salary costs to a rate per hour. Institutions should use the following definition of a standard working year for this purpose:

    1650 hours a year.

12. This can be calculated a number of ways – most commonly it is described as 220 days a year, 37.5 hours a week, 44 weeks a year. Alternatively it can be expressed as 44.5 hours a week, 37 weeks a year.
13. 1650 hours should be used when calculating and applying salary costs, indirect cost rates and estates rates. This is a fair and reasonable figure. It has been defined as a calendar year excluding weekends, statutory/institutional holidays, and holiday entitlement for that post. (It is time available for Teaching, Research, Other and Support.)
14. This will mean that institutions will be under-charging some 5 - 15% of their academic salary costs, indirect costs and estates costs. This is because the 1650 hours only takes into account time off on bank holidays and an annual holiday entitlement. It does not allow time on sickness, paternity/maternity, jury service or private consultancy to be charged to projects, either as a direct cost or as a Support or indirect cost.
15. Against this under-charge is set a potential over-charge or over-recovery of salaries. This could otherwise have occurred as PIs apply salary and indirect cost rates calculated on the assumption of a 37.5 hour week, when staff may be working more than 37.5 hours a week.
16. This standard working year of 1650 hours is to be used even when staff are formally contracted for different hours (e.g. 44 hours a week for clinical posts, or 35 hours a week for academics).
17. TRAC includes other measures to avoid over-charging – see Chapter C.4.

Chapter C.1 Named researchers

1. Academic and research staff on the project costing schedules should include all appropriate academic and research staff, i.e. Principal Investigators, co-investigators and Research Assistants. The staff should be named where possible, or a grade included with reference to the use of pool staff, or staff recruitment.
2. The costing schedule should also include the names of Fellows, emeritus professors, visiting staff from industry, clinicians and clinical academics, as appropriate, irrespective of whether they are paid a salary through the HEI, and irrespective how much of this salary is sponsored/funded (or by whom). (It would be important to identify all relevant expertise being provided to a project so that sponsors or their representatives, e.g. Research Councils’ Peer Review Panels, can better understand the nature of the project. However, this process is unlikely to be given added value by inclusion of anyone with less than 0.05 FTE allocated to any one project.)
3. PGR students are not members of the academic or research staff. However, if their studentship is based on their input to a particular project and this is recognised by the sponsor (e.g. some project studentships), then their names should also be included.

Chapter C.2 Their time

1. The time for all of these academic and research staff should be estimated and entered on the project-costing schedule.
2. An exception to this should be made for Fellows and Research Assistants whose time or salary has already been wholly (100%) included in a previous single separate fellowship or project grant provided by a Research Council, Other Government Department (OGD), or charity
   • the time or salaries of these Fellows or RAs should not then be included in the fEC of a subsequent research project. This is irrespective of the amount actually funded on that initial fellowship or project grant. They should, however, still be named on any research project on which they subsequently work;
   • however, if their original grant will have expired by the time the new project commences, then their time and salaries should be included in the new project costs;
   • all of the time of Fellows and Research Assistants whose time on Teaching and Support has been included on the fEC of a research project should be allocated to Research, and not to Teaching or Support:
     • as part of the annual TRAC time allocation process; and
     • when FTE numbers are being established for use as a cost driver or as a denominator in the indirect cost rates.
3. However, the time that a Research Fellow or Research Assistant will be spending on a new project (and a proportionate part of their whole salary) should be included on that project where:
   • the first fellowship/project has expired; or
   • their initial fellowship grant, or other project grant, was charged with less than 100% of the Fellow or RA’s time, and the new project is outside the aim of the original fellowship.
4. Where an institution has to make a contribution to salary costs to a researcher (a ‘top-up’ to a Research Fellow’s salary, for example) they should consider whether it is appropriate to charge all of the researcher’s salary and time to that original fellowship project. This may not be appropriate if it is expected that the researcher will work on other research projects that are out-with the scope of the original fellowship (a notional estimate of time could be linked to the value of the top-up and excluded from the charges made to the original fellowship project).
5. The Teaching and Support duties of a Research Fellow and Research Assistant are usually within the scope of their research projects. The time that they expect to spend on these should remain part of the fEC for those research projects.
6. Time for all other academic staff should be included, irrespective of how they are funded. This means, for example, that the time of staff funded through university trust funds, general endowments, etc., should be included on project applications. The time of individuals who are not paid through the institution (e.g. visiting professors from industry) should be included. All hours should be included, irrespective of when the work will be done (this would include any time formally or informally acknowledged in some way as overtime).
7. Academics’ time should include all the direct time required to manage the project, undertake the work and supervise the project staff. It should not include any activity that is categorised by TRAC as Support.
8. It should not include the time spent in training or supervising of research students – see supervising PGRs.
9. The time should be that for the staff carrying out the work, and it should be costed as such, irrespective of any previous practice by sponsors to fund other costs (e.g. replacement teaching costs which are unlikely to be a reasonable proxy for the real costs).
10. Investigators can estimate their time in a number of ways – three examples are (a) a month-by-month build up of time, or (b) estimating the number of hours on average across a year; or (c) using a proxy of hours per week per RA, plus time at the end for writing up. Guidance on this is given in Annex 20.
11. The techniques and estimates used in estimating time should be reviewed by the institution in year three of TRAC fEC implementation (2006/07) to ensure that the results are a fair and reasonable estimate of the actual time likely to be taken. By 2007 they need to improve the robustness of these estimates. The first (a) represents the most robust method.
12. The time should be based on the actual hours required to work on the project, (irrespective of the total hours worked in a week). Over-charging of salary costs is avoided through a number of measures, covered in Chapter C.4 below.

Quoting time estimates

13. It is good practice for estimates of academic time to focus on the hours required for the project as a whole. These are then converted to hours per year by dividing the project total by the number of funded years. Annual hours multiplied by an hourly rate provides the chargeable salary cost for that investigator for each year.
14. An estimate of the time required on average across the whole project could be used to provide the estimate for each year, irrespective of annual fluctuations in workload. Investigators’ time does not have to be profiled across the life of the project (the actual life is often longer than the life assumed for funding purposes by the sponsor, e.g. work for a project funded by the Research Councils over three years might be carried out over six years). To ensure that project costing is not made too complex, time (and associated salary costs) can be spread evenly across the life assumed for funding purposes (e.g. three years) for costing purposes.
15. Project sponsors may occasionally ask for an estimate of ‘hours per week’. These are useful to provide a comparison of resource inputs between projects. But they can be difficult to arrive at on a consistent basis. The use of ‘average hours a week’ does not encourage academics to think through the real time required (it encourages them to assume that they do the same amount of work in every week, which they do not). It is also difficult to cost as the basis on which this calculation would be made is not clear (e.g. should working weeks or calendar weeks be used; over how many years will the academic be working).
16. Similarly, the use of ‘number of days in a year’ is difficult to cost (how many days does the academic work in a year – and how many hours a day?)
17. Under either of these, it will be difficult for academics who are using robust methods as they estimate their time (e.g. building up their project estimates week by week) to convert these into figures that can be robustly costed.
18. That is why it is good practice to estimate total hours on the project, not hours a week or days a year.
19. To avoid these problems the Research Councils, for example, require an estimate of hours required for the whole project. They are still interested in average hours per week purely for comparative purposes, but are calculating a notional figure themselves (based on standard assumptions about the number of weeks in a year).
20. It is strongly recommended that institutions do not use or refer to percentages of time in their time estimating procedures, or on their project costing schedules. This implies that there may not have been a particularly robust method of estimating time, as it is difficult to be certain that the costs that are then derived are correct.

Calculating FTEs

21. The project hours of each investigator should be converted to a FTE figure so that it can be used as a basis for charging indirect cost rates and estates charges.
22. The academics’ estimates of hours per project could be converted into hours per annum by dividing by the years of funded life of the project. These hours per annum should then be divided by the standard working year of 1650 hours – that will give a FTE for each investigator. These are then added together, along with the FTEs of Research Assistants and other research staff, to provide the total FTEs for each year of the project. Indirect cost rates and estates charges should be applied to that total FTE figure. Academic salary costs if quoted at a £/FTE basis, should be applied to each FTE figure.

Chapter C.3 Salary costs

1. A salary cost should be applied to the time estimated against any one academic’s name, which reflects the salary costs of that academic to the institution. This can be based on appropriate pay bandings (individual salaries need not be disclosed on costing schedules to maintain confidentiality). Pay bandings can be used for all academic staff, or only some grades (e.g. professors) where confidentiality may be more of an issue. However, variation in honorariums and on-costs may mean that pay bandings are useful for all staff.
2. The annual salary cost applied to a project for any one academic should be that which relates to the actual or likely salary or pay banding of that individual, whether they are in post or yet to be employed.
3. The pay bandings should be set by each institution in a way that ensures that their application is fair and reasonable.
4. The salary or pay bandings applied:

   should include:

   • on-costs;
   • allowances;
   • all honorariums (e.g. associated with the responsibilities of a dean or head of department);
   • fees or other similar payments made in lieu of salary (e.g. for visiting lecturers/professors, whose salary or pay banding may only consist of this cost element);

   should not include:

   • payments that purely relate to clinical work (often paid on an agency basis) e.g. NHS merit awards/clinical excellence awards, intensity payments, or Additional Doctors Hours (ADHs);
   • academic overtime. This generally should not be included as a direct cost on a project either by including it in the salary costs (in the £/FTE calculation) or by including it as a directly incurred cost.

     However, some sponsors may exceptionally agree to pay this as a specific part of their price. A sponsor may agree to pay a premium for a report to be produced quickly, which might require it to be done outside normal working hours (possibly leading to an overtime payment by the institution to the academic). This is not part of the salary cost (nor the pay banding) used in the fEC; it should not be included as a directly incurred cost. It is part of the price.

   • fees paid by a university to a UK academic who may be carrying out research work during their time spent on ‘private consultancy’. This is unlikely to be acceptable practice in most institutions. However, if it does occur, their time should be costed at the appropriate rate for that grade of staff, and any additional fees paid should not be included in any part of the fEC of a research project.
5. Pay bandings should be based on an average of actual salaries (and all related costs, above) that are paid to the investigators within each band in a year, expressed in prices that will apply to year one of the project. This should be recalculated at least once every three years, with indexing applied in other years.
6. Institutions can use actual salaries for some grades of staff and pay bandings for other staff (e.g. professors). However, where both methods are being used they should be applied consistently, i.e. if a pay band is being used for professorial staff, the average pay for that band should be applied to all professors’ time that is charged to projects.
7. A daily or hourly salary cost should be calculated for each salary or pay banding by dividing the average salary for that band by the hours or days in a standard working year.
8. A salary cost should then be applied to each academic’s time:
   • where academic staff are partly or wholly paid through the university then a salary cost should be applied;
   • a salary cost should be applied against the time of clinical academics whether or not their salaries are wholly or partly reimbursed (the reimbursement is part of the knock-for-knock arrangements. This means that Trusts receive clinical services (O(CS)) for reimbursed salary funding and this is deemed to balance at an institutional level. It does not affect time spent by any one clinical academic on a research grant.);
   • similarly, with Oxbridge academics where some or all of the salary might be paid by a college;
   • where staff costs are paid neither by the HEI nor by their collaborating partner (e.g. some visiting fellows, visiting professors from industry, or retired academics) then costs should not be applied to their time as there are no costs in the HEI’s accounting records;
   • the time of NHS researchers who work on a HEI research project, should be estimated and included. However, if their salary is not paid through the HEI, then a salary cost should only be attached to this time:
     • if it is deemed by the HEI to be part of the knock-for-knock arrangements (the HEI is incurring other costs in lieu, e.g. Other (Clinical Services) O(CS)); or
     • if the Trust specifically recharges an appropriate part of the salary costs;
   • where the project work (or fellowship) is specifically deemed to cover 100% of the time of a particular researcher – e.g. RA, or Fellow – then 100% of their salary cost should be included in the fEC. (In these cases the individual’s time cannot be included on another fellowship or research project. This would also mean that their time should be allocated 100% to Research, in the annual TRAC time allocation exercise, with no allocation to Teaching or Support.)
9. Where an academic is to be involved in the project, but no time can be recorded, then a cost should not be included. An example of this is a Research Fellow all of whose time (100%) has been included on a single fellowship grant from a Research Council, OGD or charity.
10. Where researchers from collaborating HEIs are to be included in the lead institution’s project costing then the time and relevant staff cost of all staff should be shown separately under the academic staff categories.
11. Staff costs should be indexed (this is covered under indexing for future years).
12. Care should be taken to quote year one prices at the price levels likely to be prevailing in year one. Some sponsors, e.g. the Research Councils, will index the prices themselves in future years (years two and three of a three-year project). When they do so, they will generally take into account pay increments and promotions on research staff, but not on academic staff (they are not told the grade and bands by applicants). Therefore, the year one salary costs for academic staff should include an element of average likely future years’ pay increments and promotions, unless the staff are already at top of scale.The year one salary costs should also include an uplift, as appropriate, to reflect an appropriate proportion of additional payments likely to be incurred from contribution-related pay (e.g. bonuses or increments when at top of scale), or from a change in the pay structure (e.g. the introduction of a single pay spine or other salary costs arising from the implementation of local pay arrangements).
13. The impact on a single individual (or a particular pay banding) will not generally be known, so the uplift should be based on the estimated average uplift in academics' salaries across the institution. Uplifts in salaries across all years of the project should be considered and built into year one. To do this, the salary at the midpoint of the project life, assuming an annual pay increment and an average uplift in academics' salaries for contribution-related pay etc., expressed in year one prices, could be used as the year one salary.

Chapter C.4 Avoiding over-charging

1. Institutions should not charge more than the salary costs they incur to research projects funded on a fEC basis by public bodies (i.e. Research Councils, and many OGD projects). A number of methods have been included in TRAC to prevent this:
   1. Each institution should have in place a process that confirms that it is likely that the staff whose names are included on the project costing form will have the time available to carry out the work, given current workloads, outstanding bids and likely success rates. This confirmation need not be in writing. The process could be based around negative confirmation only (i.e. the PI is alerted by the academic or other member of staff or management where it is considered that the academic will not be able to make the time available);
   2. The salary costs estimated and charged by a group of academics on Research Council and OGD-funded research projects(and including for this purpose an appropriate part of the time spent on Support) should not exceed the total salary cost that it is estimated that they will be paid, within the tolerances of the pay bandings used;
   3. The estimates of time made by a group of academics working on cost-based price research projects (i.e. for the Research Councils and many OGDs) should not exceed 100% of a standard working year (i.e. 220 days, or 1650 hours) in a year. (Direct project time and an appropriate part of Support time should both be considered in this.) It is not a requirement for the time for these academics to be added up for all projects being carried out and bid for, to check whether their total committed time on these activities exceeds the standard working year. However, it is good management practice to ensure that staff workloads are reasonable;
   4. (Research Councils are also likely to introduce (joint) systems that will allow them to identify any individuals who are working more than a standard working year on Research Councils’ projects. However, Research Councils also realise that, to some extent, institutions will need to over-commit academics at the proposal submission stage as success rates for proposals are not 100%.)
   5. Salary costs are to assume a standard working year of 1650 hours a year, whilst actual hours are to be charged to research projects. As actual working hours in a week are often more than a standard working week (37.5 hours) this might be seen as a potential area where over-charging might occur. But, as explained at the start of Section C, actual working days in a year are generally less than those in the standard working year (220 days). It is therefore likely that, overall, both actual and standard working hours in a year will total 1650. Cost rates based on 1650 hours, applied to actual hours worked, are therefore fair and reasonable, and double-charging is unlikely to occur;
   6. Academic overtime is excluded from the costs that are charged to research projects, unless the sponsor agrees to bear it in addition to normal salary costs;
   7. From December 2009 institutions will be able to, and should, compare:

      the total academic staff costs attributed to the Research Council sponsor type in the annual TRAC time allocation exercise

      with

      the total academic staff costs charged to Research Council projects in the accounting records under TRAC fEC.

      see reconciliation.

Chapter C.5 Supervising PGRs

1. The annual TRAC process recognises an activity called the supervision and training of PGRs. This is different from the work undertaken by an academic in a research project. A PGR student is not considered to be a member of academic staff.
2. Where a PGR is on a project studentship linked to a project, they may be recognised on the project-costing schedule. All of the costs associated with their supervision and training should, however, be kept separate from all other costs of the research project itself.
3. As funders (including the Research Councils) do not currently fund PGRs on an fEC-basis, and as establishing the fEC can be complex, it is good practice to estimate their costs on a research project, but it is not a minimum requirement under TRAC. (The Research Councils, for example, require appropriate project studentships to be recorded on their application form under ‘exceptions’ and these are funded on the basis of stipends plus fees – not on the basis of fEC.)
4. Therefore the time the supervisor spends in PGR training and development (and this would include the time of internal and external examiners, co-supervisors, etc.) and stipends paid to students, should not be included as part of the research project fEC. (And, therefore, the indirect costs and estates costs associated with this time should not be included either.)
5. Where the PGR is closely linked with one research project, it is good practice to show the fEC of this activity on the same research project costing schedule. However, this should be shown as a clearly separate category of expenditure from the costs of the research project.