Nothing seems to polarise opinion in investment circles more than the level and future direction of interest rates. This is problematic, as moves in long-dated rates have been the main driver of Defined Benefit pension fund financial health over the last decade.
DB pension schemes that hedged their interest rate risks before the summer of 2016 will have had the pain of recent falls in interest rates partly or fully mitigated. But many schemes have not – with long-dated gilt rates now below 2%p.a. have these schemes now missed the boat? Is it too late to hedge? In this piece we take a balanced look at the arguments for and against interest rate hedging at these levels and provide a framework to help all stakeholders move discussions forward.
So what now..?
We know many trustees and corporate sponsors wrestle with the tough decision of whether, how much, and when to hedge.
In this paper, we present information and frameworks that will help pension scheme stakeholders ‘unlock’ the hedging debate. The paper is structured as follows:
- Context – The Hedging Dilemma
- Context – The Rationale for Hedging
- The Case Against Hedging Now
- The Case for Hedging Now
- Hedging and Behavioural Biases
- Bringing it All Together – A Risk Management Mindset
- Practical Steps to Unlock the Debate for Your Scheme
1. CONTEXT – THE HEDGING DILEMMA
Industry-wide data produced by KPMG shows, over the last decade, more and more schemes have implemented interest rate hedges. As of July 2016, KPMG estimate around £746bn (or half the total liabilities) have been hedged. 1,200 schemes have implemented such strategies (fully or partly), while many have not. Yields on long dated gilts (used for discounting liabilities) have been both low and falling for years, reaching levels that seem extreme relative to history. This is in spite of a remarkably high supply of gilts by historical standards. A similar – but even more extreme dynamic is playing out in other countries, such as Germany and Switzerland (see Fig. 1).
The size and influence of DB pension funds on bond markets is much smaller in these countries compared with the UK. This suggests the yield decline is a global macro-economic phenomenon, rather than a UK Defined Benefit story.
The chart below illustrates the yield on a particular bond – the 2.5% “Consolidated Stock” which existed as a perpetual bond for over 200 years before its redemption in 2015. It serves as one of the more reliable measures of long-term UK yields, and neatly puts today’s picture into context on three fronts:
- Today’s yields are very low by historical standards
- There have been long periods where yields remained low in the past (e.g. late 1800’s)
- The period following the 1970’s, when yields were very high appears an exception when viewed in the longer-term historical context (see Fig. 2).
2. CONTEXT – THE RATIONALE FOR HEDGING
The rationale for pension scheme interest-rate hedging is really driven by several key financial economic principles relating to risk management and the value of future cashflows:
- The desire to fund pension schemes in such a way as to have sufficient assets to meet future liabilities (a 2015 survey by Aon Hewitt1 found that 75% of trustees held either buyout or self-sufficiency as their objective).
- The financial risk management principle of matching financial liabilities with assets (of similar or identical economic sensitivities).
- That a market-consistent discount rate is the most appropriate, impartial and most easily defended basis for cashflow valuation. It is also the most consistent with a market valuation of the assets. Accepting these broad principles leads to an approach to pension scheme liability valuation that uses bond prices and yields as a basis for determining discount rates. It also drives an approach to risk management that partially or fully removes the risk associated with fluctuations in the interest rates used to value the schemes’ liabilities. These ideas were first put forward within the UK actuarial profession in the late 1990’s and we would argue that they now form the basis of standard practice among UK actuaries, as evidenced by:
- Aon Hewitt’s 2016 analysis of actuarial valuations2, where 85% of valuations among their clients were found to be based on a yield curve methodology
- Willis Towers Watson’s 2015 Analysis of Funding Trends3, which showed a yield-curve methodology overtaking the dual-discount rate as the most common approach to valuations among their clients (the dual-discount rate is of course itself frequently driven by bond yields)
- Both the Aon analysis mentioned above and The Pension Regulator’s 2016 funding statistics4 point to an average equivalent spread over gilts being used as of 2016 of around 0.9%. Aon comments that “a convergence of the post-retirement discount rate to gilts+0.5% is evident through time”, which they consider a reasonable target for self-sufficiency
In the interests of balance, it is worth noting that some within the pension fund community take an alternative perspective that suggests that, as long-term endeavours, pension funds could simply focus on meeting future cash flows while ignoring changes in the present value of assets and liabilities. This is a complex philosophical debate on which we will shortly publish a separate thought piece.
The context for this paper is this; we believe hedging is a prudent strategy for managing risk. However, with interests rates so low and the future unclear, it brings us full to circle to the decision – should I hedge now?
3. THE CASE AGAINST HEDGING NOW
■ Rates are at multi decade lows, even century lows, both in the UK and other countries
As long-dated rates (10 year government bonds are often used as a benchmark) have fallen over recent years, many studies have set out to identify similar periods in the past to extract lessons and predictions. The two most obvious periods were:
- The 1930’s and after, following the Great Depression.
- The late 1800’s which suffered multiple financial crises.
However, in several key countries including the US and the UK, bond yields have fallen lower than the levels they reached during these periods. Some data on interest rates is also available going back several centuries. It is fair to question their relevance, given the fundamental difference in the economy and setting of monetary policy. That said, it also intimates that a stable level of interest rates in the mid-single digits appear to have been the norm in previous centuries.
At the end of 2014, Deutsche Bank released a research piece showing bonds in developed markets such as the UK, US and Netherlands were in the lowest percentile compared to histories of several hundred years (yields have since fallen further.) Whichever way you look at the data, the conclusion seems clear; we are living in extraordinary times.
■ Returns available on bonds relative to inflation seem unsustainable
It isn’t only nominal yields that are relevant to this picture. The yield on fixed coupon bonds should (in theory) compensate investors for expected inflation over the period of the bond. It is true that declining bond yields over recent years have also coincided with declining expectations for inflation. However, yields have declined faster. This means the real yield (the return on a bond after allowing for the effect of inflation) has become substantially negative in many developed countries.
The real yield on a 10 year gilt at the time of writing in August 2016 was about -1.6%. This expected capital erosion in real terms from holding a bond to maturity, coupled with a volatile market value and potential credit risk, are at odds with most conventional economic theory. It is reasonable to ask the question “why would I buy an asset that locks in a real-terms loss on a buy-and-hold basis?”
■ The Bank of England (BoE) has been buying bonds, this won’t last forever
The BoE’s Quantitative Easing (QE) programs of 2011 and onwards saw it purchase a significant amount of in gilts. It stands to reason this extra demand for gilts would generate upward pressure on prices (indeed this was the stated effect of the program). It also will not be permanent – QE was intended as a temporary stimulus measure (although the Bank has not been explicit around how it is to be unwound). As of August 2016 this program has been restarted.
■ Isn’t there some sort of lower bound on long dated rates?
- Basic economic theory suggests that if interest rates available on bonds and bank deposits become negative, two things will happen:
- 1. Sensible depositors will remove their money, preferring to keep it under the proverbial mattress.
- 2. This would mean, at some point, rates would stop being at risk of large further falls implied by most risk models. Consequently, there would be less of a need to hedge against such falls, as the risk would be less.
- Conventional theories suggest interest rates are driven, to some extent, by mean reversion. This would make a case for interest rates rising in the future. Indeed, many experts have made compelling cases for higher interest rates, and continue to do so.
4. THE CASE FOR HEDGING NOW
■ We could be in a “lower for longer” environment, even if rates ultimately rise
It is a reasonable view that rates will be higher at some point in the future, but it could be a long way off. Whilst rates are undeniably low, they have the potential to go lower and stay low for longer than many expect. For example, see the graph below showing yields on long-dated Japanese Government Bonds from 1970 (see Fig. 3).
There are a couple of salient features of this data, which are worth drawing out. Japanese government bond yields first touched levels comparable to today’s gilt yields in 2003 (when 7-10 year Japanese government bonds yields were around 0.6%p.a.). Today’s yield for the comparable Japanese bond is -0.2%.
In the 13 years since Japanese bonds first moved below a yield of 0.6%, the yield has never risen by more than 1.2% at any point. Indeed, the most Japanese yields have risen from any point in the last 20 years is just 1.5% (see Fig. 4).
Whilst some might argue Japan is a special case, the data suggests that, in many markets, long-term interest rates are not mean-reverting. It has been a reasonable view, based on some of the fundamentals described earlier, that long-dated interest rates will rise. But in practice this has not happened. The Federal Reserve in the US has raised the base rate (the “Fed Funds” rate) in December 2015, but this did not result in a dramatic move in longer-dated rates – although long-dated rates in the US remain higher than in the UK and Europe, due to the different expected path of the base rate.
■ Experts have generally been very bad at predicting the direction of interest rates
In one infamous Bloomberg survey in 2013, all 67 economists surveyed thought interest rates would rise; in fact they fell over the course of the year. Surveys like this have consistently been wrong about the direction of interest rate for years, showing the difficulty of predicting the bond market using fundamental economic logic – often calibrated to and informed by pre-global financial crisis experience.
■ Interest rate forward curves themselves have been bad predictors of future rates
The long-dated interest rate contains the market’s estimate of the short dated rate at all future points. Since short-dated interest rates fell to their current low levels in 2009, long-dated rates in the UK have consistently contained the assumption that the short-dated rate will begin to rise in the near future. As this assumption has failed to be borne out again and again, long-dated rates have fallen lower and lower, with the curve becoming flatter and flatter. However long-dated rates today still contain the assumption that short-dated rates will rise.
■ Little evidence of lower-limits on rates in practice
It’s reasonable to argue there is a point below which long-dated interest rates probably cannot fall. But experience elsewhere in the world (e.g. in Switzerland where the 50 year bonds are trading at zero yields) has challenged this view and forced people to recalibrate where this “floor” might be. It seems, in the UK, we are some way away from it at the moment. In fact, relative to some other bond markets, such as Switzerland or Germany, gilts offer relatively more attractive levels of yields.
■ Incorporating strong interest-rate rising views still implies some hedging
We have a model that lets us look at the impact on optimal hedging levels of a strong view that interest rates will rise. Even with relatively optimistic views on how much long-dated interest rates might rise by (say 2 or 3%), we find it is still logical to hedge half or more of your interest rate risk. This is because a view is not the same as an assurance that something will happen, and the risk around the view not being borne out is still significant.
Even if your interest rate view is so strong as to imply only a small amount of hedging now, it makes sense to set up the mandates and infrastructure to capture any rise in interest rates as an when the view comes to fruition.
5. HEDGING AND BEHAVIOURAL BIASES
In the section above, we set out some of the key arguments for and against hedging in the current environment. Whilst we would all ideally like to make objective, rational decisions on these issues, we are all potentially subject to biases that can distort our views.
In discussing whether to hedge at current levels, the most powerful impact is likely to be felt from status quo bias.
Status Quo Bias
Status quo bias can be summarised by the old proverb, “When in doubt, do nothing.” The bias describes findings that show when faced with choices with uncertain outcomes, there is a powerful tendency for people to stick with what they have. This effect can impact something as simple as individuals selecting a brand of toothpaste, to trustees deciding not to hedge key risks they are exposed to, such as interest rates or inflation. The status quo bias is driven by several interrelated factors – two of the most relevant factors in the current context are the impact of past experience on future expectations and fear of regret.
There is good evidence5 showing individuals who have experienced certain macroeconomic environments are more likely to see the future in certain ways. For example, those who have experienced high equity returns, tend to invest more in equities, or those who have experienced high interest rates tend to expect them in the future.
Most trustees are likely to be in the age range of 40-65 and thus will have experienced periods of much higher interest rates (see Fig. 5).
So although longer-term data, or evidence from other markets may show that rates can remain low for a long period of time, many trustees may have strong beliefs that rates will return to previous levels.
We should try not to fall into the trap of framing decisions in the context of our own experiences. We’ve all lived through periods of much higher interest rates, but markets can undergo regime shifts. The current macroeconomic outlook is significantly different in many important ways to that which dominated previous decades.
As discussed above, analysis of Japanese interest rates shows they can remain low for decades.
Regret is something we all seek to avoid. Unsurprisingly, regret avoidance can have a large impact on decision-making. For example, most parents would not consider leaving their baby asleep alone in their house whilst they undertook a short car journey to deliver a package. For most, the regretful thoughts of guilt, blame and stigmatisation if the house were to catch fire, or the baby abducted etc., would be too much to bear. Most would therefore choose to take the baby in the car with them, as although the car journey has greater potential for harm, it has less potential for regret. As John Maynard Keynes famously noted, “Worldly wisdom teaches that it is better for reputation to fail conventionally than to succeed unconventionally.”
Trustees always face the problem of regret. In the context of hedging interest rate risk it is particularly acute at the moment given rates are at an all time low. A trustee who has not previously hedged is likely to feel more regret if they took all the pain as rates fell and then felt none of the upside as rates rose.
Whilst these feelings can understandably create a state of inertia, our experience with clients suggests that, with the right framework, decision-making can be ‘unlocked’.
6. BRINGING IT ALL TOGETHER – A RISK MANAGEMENT MINDSET
As we have seen, the situation is challenging. There are compelling arguments for both sides of the hedging argument and powerful behavioural drivers that can influence our choices.
How can trustees and other stakeholder resolve this challenge and make a decision that all parties can understand and support?
Our experience shows the most effective way of dealing with this challenge is by altering the frame of the question by adopting a risk-management mindset.
Often we see trustees get stuck in a loop trying to make a call on interest rates, which is incredibly challenging.
We believe that it helps to adopt a fund manager’s or company treasury’s approach to expressing market views and risk, rather than getting trapped into thinking in terms of binary outcomes. That prompts questions such as:
■ How strong is your conviction that interest rates will rise?
■ How much risk do you want riding on this view vs. other views?
■ Is it reasonable to let some risk ride on this view, but probably not so much as to swamp all other risks in the portfolio (e.g. equities, credit)?
Set-out below is a simple but powerful structure that trustees and other stakeholders can use to gain an alternative perspective and make an informed decision.
7. PRACTICAL STEPS TO UNLOCK THE DEBATE FOR YOUR SCHEME
■ Understand behavioural gremlins – Recognise these can influence your beliefs and decision-making, e.g. how your past experiences can unconsciously condition your future expectations and your fear about regret.
■ Whose role is it? – Discuss the extent to which the trustee board believes its role should be to take investment views such as the direction of interest rates (in our experience, some do, some don’t). There isn’t always alignment within trustee boards. If you come to the conclusion that you shouldn’t be taking a view, that implies a certain element of hedging.
■ Get clear on the facts – Establish any unfounded fears regarding the behaviour of interest rate hedges in practice – use expert advice to address or dispel these.
■ Begin with the end in mind – Do you have a clear route for your scheme? Research by Hymans Robertson showed whilst 88% of lay trustees stated their schemes have measurable plans in place, 40% of these trustees haven’t determined a timeframe for reaching their end destination6. You need to be sure you are clear on your schemes goals, and importantly the timeframe.
■ Align Objectives – ensure all the key people managing and accountable for the scheme share the same vision. 31% of Financial Directors don’t believe trustees share their objectives. It’s vital to ensure everyone is aligned.
■ Experiment with different approaches to framing this decision – For example, most trustees would not be comfortable with the scheme’s equity manager placing the entire portfolio in a single stock, however much confidence the manager had in this being a great investment. However, from a risk perspective, many schemes do something very similar with their approach to hedging interest rates (by leaving them unhedged and thereby creating a large risk exposure to one single risk factor).
■ Get consensus – If you are going to take a view, try to get consensus on the drivers of the long-dated real yield and how it is determined.
■ Acknowledge the arguments for and against – really push yourself and your board to consider all angles and make a decision with all the facts.
■ What if – Ask the question “what if we were to make the assumption yields will rise, and are wrong?”. Typically this will result in a larger deficit and some combination of greater sponsor contributions and a longer period required to recover the deficit before the scheme is in a stable position.
In an environment of remarkably low interest rates, hedging is an emotional business. We believe that, in principle, the concept of hedging is sensible and the evidence is clear that those schemes who have hedged are indeed better off than those that have not. We believe that the question is not whether schemes should hedge, but when… is now the best time?
Given the wide nature of schemes’ funding positions, objectives and beliefs, this paper can’t provide a blanket answer. So to be able to answer these questions for your scheme:
- Have clear objectives – it sounds easy, but we have worked with and are aware of many clients who don’t have this. And it needs to be clear. By this we mean what are you trying to achieve and over what timeframe? Without this strategic focus, the hedging decision can lack context.
- Align stakeholders – some trustees and sponsors will want to take an explicit view on interest rate direction and it won’t always be the same view.
- Challenge your thinking – meet and talk to people who you disagree with. That might mean getting a new consultant to give you a health check, or it might mean going to different places or reading different articles or publications.
- Right-size your risk – if you want to take an explicit view on rates, use a risk-management mindset to size those views. A good consultant should able to help you express your views in a way that is consistent with your scheme’s aims and wider approach to risk management. For example, it might be perfectly reasonable to take a view that rates might rise by 1% in the next 18 months. The key issue is understanding how much risk, in whichever way you to choose to measure it, you want riding on this view relative to other strategic decisions.
- Be prepared to be wrong – run the opposite scenario you have planned for. What if that happens, are you prepared for it?
No process can guarantee the right the answer with a decision as challenging as interest rate hedging in the current environment. We hope the insights in this paper will give you the tools and mindset to tackle this decision with clarity and confidence.
1. Aon Hewitt Global Pension Risk Survey 2015 http://www.aon.com/unitedkingdom/retirement-investment/pensions-stability/global-pension-risk.jsp ↩
2. http://www.aon.com/unitedkingdom/products-and-services/human-capital-consulting/thought-leadership/all-topics/in-depth-pension-scheme-funding-analysis-2016.jsp ↩
3. https://www.towerswatson.com/en-GB/Insights/Newsletters/Europe/UK-Corporate-and-Trustee-Briefing/2016/01/Current-funding-trends ↩
4. The Pensions Regulator 2016 Scheme Funding Statistics http://www.thepensionsregulator.gov.uk/docs/scheme-funding-2016.pdf ↩
5. Malmendier, U., & Nagel, S. (2009). Depression babies: Do macroeconomic experiences affect risk-taking? (No. w14813). National Bureau of Economic Research. Malmendier, Ulrike, and Stefan Nagel. “Learning from inflation experiences.” The Quarterly Journal of Economics 131.1 (2016): 53-87. ↩
6. Hymans Robertson’s Trustee Barometer 2015: The road to a resilient pension scheme. https://www.hymans.co.uk/media/uploads/trustee-barometer-2015_final.pdf ↩