Executive summary
• Benefits are notoriously difficult to slot into a pure ROI model. So a variety of alternatives have been developed to measure the ROI on different benefits.
• One of the most common ways is to measure the perceived value staff place on their benefits. An advantage of this approach, is that perceived value can be boosted through communications without spending any more on the actual benefit.
• Another measure is to look at return in risk terms. It's a question of comparing the cost of the risk benefits against the previous exposure if something had gone wrong and the likelihood of that disaster happening.
• Some feel it is inappropriate to try to measure the ROI on benefits because they are less an investment and more a cost of doing business. But ROI experts believe it's possible to measure a return on any human capital investment. They highlight that the costs are known, and they only need four things to calculate the ROI: an objective, a robust metric to calculate whether the objectives have been met, a way to bring the metric back to cash, and a system to evaluate the role benefits have played in meeting that objective.
• The other issue is that the full ROI calculation also relies heavily on opinion and assumptions, and any end result may be questioned on this basis. One way to take out the biggest assumptions is to measure the impact, but not take things back to a cash value.
Main feature
Return on investment (ROI) is the phrase of the moment, and as austere economic conditions bite, funding is increasingly only available for those projects that will provide some sort of payback.
While HR and your consultants will pepper their presentations with these three killer words as a surefire way of getting an FD's attention, finance chiefs need to know that benefits are notoriously difficult to slot into a pure ROI model.
So a variety of alternatives have been developed. The question is which metrics are the most meaningful for the job in hand? One of the most common is to measure perceived value.
It's a method that on the face of it may seem a little soft for any hard-nosed FD, but it does come with some positive gains. Darren Laverty, a partner at consultants Secondsight, says: "We measure the value people place on their benefits. We take a representative sample and email them or call them and say 'if you had to estimate the value of your benefits package, what would you say it was worth?' Then we take out the crazy figures at either end of the spectrum and find an average."
One advantage of this approach is that perceived value can be boosted through communication, without spending any more on the actual benefit. Raj Mody, a partner with PricewaterhouseCoopers, says: "I would argue that an employer contributing 10% of the payroll to pensions would get more value if they contributed 9% and spent the balance on communication."
Secondsight, for example, measures perceived value before it starts a communications project. Laverty says: "Typically people think their benefits are worth about half of the actual value." At the end it is measured again, and Laverty says: "On average they think benefits are worth 150% of the actual spend."
A second version of this is to measure relative value, compared with salary. Chris Noon, a partner at consultancy firm Hymans Robertson, says: "People ask 'if you are spending £1 on benefits are you getting more appreciation, engagement and effort than if you had given them £1 in salary?'" This can be unearthed through structured questionnaires and focus groups. Mody says: "You may ask 'would you prefer a cash bonus of 10% of salary every year to getting a pension?'."
There is a third variation. Noon says: "Occasionally organisations measure the value delivered to the employee over and above if they had bought the benefits themselves. It doesn't mean people understand or appreciate it, but it is a relative measure in terms of pounds saved."
The end result of all three is designed - in one way or another - to measure the value employees place on their benefits package. The problem is that this doesn't provide much of a business case. It simply answers the question: 'Do employees appreciate what we're spending?' The key question is: 'are we getting payback for what we're spending?" An alternative, therefore, is to look at return in risk terms. Noon says: "Benefits aren't just about making employees love you. They are about risk mitigation too, with benefits such as long-term disability insurance."
So are they proving their worth in risk-avoidance terms? "This is essentially a risk-assessment and data-mining exercise; checking where the risks lie and quantifying them, then assessing whether benefits sufficiently cover them. It may include, for example, ensuring senior managers are covered by the private medical insurance, and that employees with families are covered by life insurance. Once the risks are covered it's a question of comparing the cost of the risk benefits against the previous exposure. The cost of this exposure would, of course, be an estimation, based on the cost to the organisation if something had gone wrong, and the likelihood of that disaster happening.
This is a much stronger measure, and clearly has value in assessing risk benefits. It's not perfect, however, as not all benefits are designed to control risk, so it cannot be used to measure, for example, a company car scheme. In addition, some benefits actually pose risks in themselves to the business, such as the defined benefit pension scheme, which exposes the employer to investment and liability risks.
The weaknesses in these models for measuring return has led some to conclude that it's not possible to find a robust metric. Ben Wells, lead consultant in the managing change and engagement division of benefits specialists Buck Consultants, says: "It's less an investment and more a cost of doing business. We can't demonstrate a return on investment because it's not an investment in the first place."
However, there are other academics and consultants, who believe it's possible to measure a return on any human capital investment. They highlight that the costs are known, and they only need four things to calculate the ROI: an objective, a robust metric to calculate whether the objectives have been met, a way to bring the metric back to cash, and a system to evaluate the role benefits have played in meeting that objective.
At its most simple, all these things are known and can be measured easily. So, let's assume the objective is to control the cost of benefits in future years, and the benefit put in place is a flexible benefits package with a fixed sum to spend on benefits each year.
The robust metric can be produced simply by calculating the cost of benefits in future years, compared with the projected cost without flex in place. This doesn't need to be brought back to cash because it's already expressed as a cash sum, and there's no need to evaluate the role benefits have played in meeting that objective, because they are the only contributing factor.
This level of simplicity, however, is rare. More complex situations occur when the metric is harder to bring back to cash. So, for example, let's assume an employer has brought in a defined benefit (DB) pension scheme in order to improve retention. Let's also assume, for argument's sake, that nothing else has affected retention — either within the business, the sector, or the broader economy. A financial benefit of 'improved retention' needs to be established.
Jeremy Harrison, director of consultancy firm ROI UK, says it's a matter of consulting data drawn from the wider industry: "There are values from the Chartered Institute of Personnel and Development that can tell you what the cost of turning over a particular member of staff is. Then you can say 'conservatively the benefit will be £x'."
However, it's not strictly that simple. Mody points out: "Employees may just be hanging on for the wrong reasons. Not because they enjoy the job or because they want to be productive, but because they want to maximise their DB pension. So they may be less productive than their replacement would be, which isn't a gain at all." The nature and result of the retention, therefore, must also be scrutinised, through a combination of employee data mining and staff surveys.
The objective may be even harder to bring back to cash than retention, for example, an improvement in engagement. For this, consultants can conduct studies, or draw on the results of research, either theirs or published by others. Noon says: "There is a massive amount of proof that an engaged employee is more productive and therefore more profitable. We have measured this correlation for clients ourselves."
Once the cash benefit is established, the next step is to evaluate the role that benefits have played in meeting that objective. Assume the objective is a reduction in sickness absence, and the benefit used to achieve this is private medical insurance. The metric is clearly going to be absence, and this is easily brought back to a cash value. The challenge therefore is to establish a link between the benefit being introduced and the fall in absence. This is tricky but not impossible. The first part of this is to prove that people who use the benefit are less likely to be absent. Noon points out: "It's a link that's often missing, but can be done using surveys with questions about benefits and through analysing the data of benefit use."
Kim Honess, head of the flexible benefits consultancy at Watson Wyatt, says surveys can be effective even when aimed at softer objectives such as engagement. She says: "Vodafone has a quarterly pulse report that measures engagement. Since bringing in flexible benefits it has seen sustained improvements in these engagement scores."
And finally, it's vital to prove that it's the benefit that made the difference. Harrison explains: "It means isolating the effect of the employee benefits from any other contributing factors. This is one of the most difficult exercises. You need to ask what else could have affected it? For example, if sales have increased is it because the product has become the most competitive on the market, or is the market now right for it? "The process depends on asking people who were there what happened and what conclusions they would draw from it. How much would they attribute to the benefit? If they say, for example 30%, you would ask how confident they were in that figure. If they said absolutely confident you would use 30%, if they were only 75% confident you would scale that 30% down, so you have the most conservative estimate. You can only report what is credible, so if you have a range of values reported you can only take the lowest of them."
It is important to decide the period over which to measure. Harrison says: "There are guidelines built from experience. You want to let it settle down, but you don't want to leave it too long or it will become difficult to disentangle from other factors."
And you will need to consider trend line analysis. Harrison explains: "You find out what productivity or retention was for a year or so before. If it was already trending up you would discount the consistent curve from any benefit to get a truer picture."
The theory is sound. The practice, however, is fraught with difficulty. Each step of the process requires measuring, quantifying, assessing and monitoring, which is in itself a costly business. The sheer scale of the costs involved mean many employers may simply not want to embark on the process. Part of the problem is that the experts say it's imperative not just to do the calculation once. Noon points out: "People don't do enough in terms of ongoing measurement. They should be looking at measuring regularly and actively managing it."
The other issue is that the full ROI calculation also relies heavily on opinion and assumptions, and any end result may be questioned on this basis. One way to take out the biggest assumptions is to measure the impact, but not take things back to a cash value. So, for example, according to the Employee Benefits/Axa Pensions Research 2008, of those employers who looked at ROI on pensions, 53% looked at the impact on staff retention rates and 35% looked at how it affected recruitment of staff.
The full ROI measurement is a fairly robust way of making absolutely certain whether it's worth investing in a particular benefit, and once you have invested ongoing measurement will let you know whether it's delivering on its potential. The question, therefore, is whether it's worth the cost and complexity of a full ROI investigation, or whether the organisation would prefer to rely on the more efficient but less meaningful softer measures, such as perceived value and relative value.
After all, in leaner times FDs may themselves need to justify harder, and more expensive, metric analysis.
†
Case Study: National Steel†
Jeremy Harrison, director of consultancy firm ROI UK, recalls a study entitled 'Proving the Value of HR, ROI Case Studies', published in 2007 by US body ROI Institute, where metal firm National Steel had a problem with accidents in its Mid-West division. There were 60 incidents a year, costing around $470,000. It set a target of reducing both the total number of accidents and more serious disabling accidents.
They used an incentive of a $75 payment to each employee for every six months that passed without an accident requiring medical treatment.
And it established metrics of the number of accidents that needed medical treatment, the number of accidents that led to a loss of time, the overall accident costs and incentive costs.
They measured the year before the inventive, and a year into the scheme. The number of accidents went from 61.2 to 18.4, disabling accidents from 17 to four, cases of medical treatment from 121 to 17, lost time accidents from 23 to 5.3, and the cost of accidents fell from $468,000 to $19,000.
The total cost of the incentive plus administration for two years was $72,000. The net benefit was $336,000, a ROI of 379%, which included 20% knocked off the net benefit on the grounds that around 20% of the improvement wasn't due to the incentive - but to the extra management focus on accidents.
Measuring ROI on costly benefits like pensions
As the biggest cost in the benefits package, it's common to consider whether the pension scheme is providing a return on investment, or whether a different scheme design would provide a proportionally better return.
Many of these models can be applied to pensions:
1. The pure ROI model is simpler for a defined contribution (DC) scheme, where the cost is known. The return may be measured through retention, engagement and recruitment of staff. It can then be brought back to cash using assumptions about the cost of recruitment or the added value of engagement. For a defined benefit (DB) scheme, a projected cost will have to be used alongside the return measures.
2. Alternatively, employers can ask whether they would have more return from £1 of cash, or £1 of an alternative design of pension. This would be done through questionnaires.
3. They can measure the risk mitigation. This has a few difficulties. In a DB scheme, the pension itself does in fact pose investment and liability risks, so may cause as many risk issues as it solves. The DC scheme, meanwhile, may be ineffective in curing risk problems if not enough is put away in the scheme.
4. They can measure how much the individual will have saved , compared with if they were providing the pension themselves. This is likely to be an impressive figure when it comes to a DC scheme. It may also produce good results where an employer is using salary sacrifice, and the NI saving can be added to the benefit. Unfortunately, as Chris Noon a partner at consultancy firm Hymans Robertson says, this tends to be a meaningless figure for employers.†
5. How much spending now will save in future, especially when an organisation is moving to a DC pension scheme, and the savings can be compared against uncapped runaway DB schemes? The measure is harder when measuring the benefits of a defined benefit scheme.
6. Perceived value is often the key metric for pension schemes. This is largely because any measurement of actual value may rely on waiting decades for any realisation, when an individual actually retires on that pension, which Raj Mody, a partner with PricewaterhouseCoopers, points out: "is useless because you want the answer in the here and now". The trouble for all ROI calculations for pensions, however, is that pensions don't necessarily come on to employees' radars. Mody says: "If you ask whether the scheme affected an individual's decision to join or stay with a company you may get the odd 'yes' if you have a DB scheme. But if you ask whether it affects your engagement and performance very few people will say it has a first order impact. However, this doesn't mean you shouldn't bother with pensions, because there are a whole host of reasons why it makes sense, not least because you want your employees to be able to retire. And in that respect there may be more to life than ROI," he says.
Pensions research: How do employers measure return on investment (ROI) in their pension schemes?
They don't 74%
Look at retention 17%
Look at recruitment 12%
Plan to start looking at ROI 6%
Other 4%
Back to Employee Benefits Report for Financial Directors – November 2008