Thanks to Toby Nangle for taking the time to answer my Tweets on his blog about “Strategic Asset Allocation” for Defined Benefit pension schemes – more efficient – I suspect – than a chat in pub.
Just to repeat my Tweets:
I will try to stick to first order points of economics, and leave the various second-order legal, regulatory and tax points for another day. The underlying economics also applies to all countries with DB pensions, especially the US.
Let me answer Toby’s three points:
1 “In theory there’s no difference between theory and practice. In practice there is” Yogi Berra
Toby’s major criticism of my Tweets is “It works in theory, but not in practice”.
Although matching DB pension liabilities with long dated corporate bonds works in theory, in doesn’t work in practice – there just aren’t enough long duration and/or inflation linked bonds available to match say £1.2 trillion of UK DB pension liabilities. And demand for the limited supply of matching bonds will push up their price, increasing the value of pension liabilities to be hedged.
A bit of practice and a bit of theory. Toby’s description of what Boots did misses the last piece of the jigsaw, which – as Yogi Berra may have said – completes the circle.
In the 18-months to April 2001, Boots Pensions switched all its £2.4bn of assets into long dated – 16 year duration v 17 year duration of liabilities – AAA/Aaa supranational bonds – World Bank, EBRD, EIB etc – including 25 per cent index linked. It then increased the inflation-linked to 50 per cent through inflation-linked swaps and was – I believe- the first UK pension scheme to execute a swap.
This switch to matching assets reduced financial risk for Boots plc shareholders – they were no longer running the risk of cash calls to meet pension deficits, increased security for 70,000 pension scheme members – the value of pension assets and liabilities moved closely in line – and significantly reduced the cost of managing the £2.4bn assets – LGIM charged just 1bp for holding the £2.4bn assets. (With £2.4bn assets Boots Pensions was one of the largest 50 UK pension schemes).
But then, when the dust had settled after the November 2001 announcement, crucially, in March 2002 Boots plc used £300m of its cash to buy-back its own shares. Having reduced financial risk in the pension scheme, Boots was able to increase financial risk on balance sheet by running down cash -the same as increasing debt – in a tax efficient way.
Toby missed the share buy-back which completes the circle and explains what happens if all UK companies did the same as Boots.
As company pension funds sell their equities and buy bonds, the companies themselves would effectively reverse this by selling (issuing) their own long dated bonds – bought by other company pension funds – and using the proceeds to buy-back their own equity.
Imagine a giant aircraft hangar where corporates meet to swap (buy- back) their equity held by other company’s pension funds, in exchange for their long dated bonds (sell).
This process would happen gradually, over a long period, with flexible financial markets coping perfectly well.
The Enterprise Value of each individual company (Equity + Debt) remains unchanged (ignoring second order tax benefits from the switch) – and their capital structure is unchanged – they have more debt but hold no equity of other companies.
And, at the macro level for all companies, nothing has changed.
Let’s not forget that this switch reduces the cost of capital for UK plc, as management and transaction costs and fees to third are drastically reduced – the losers are financial services third parties, and some (many?) some people will be out of a job.
So the practical objection to matching pension assets and liabilities – there just aren’t enough bonds to go round – assumes a fixed supply of bonds, rather than supply increasing as part of the whole process.
Swapping equity for debt in the aircraft hangar also answers the question, “Who will buy all the equities pension schemes would sell”. The answer is “the companies themselves”.
The idea of selling equities/buying bonds in the pension fund and selling bonds/buying equities in the company goes all the way back to the seminal Fischer Black article of 1980, “The Tax Consequences of Long-Run Pension Policy”  which is a “must-read”.
And once we include the Boots share buy-back, it becomes clear that it was NOT “LDI” which has been pushed very hard in the last few years. “LDI” views the pension scheme as a self-contained entity, rather than, in economic terms, an unconsolidated subsidiary of the corporate sponsor, and is, at best, incomplete.
(Anyone interested in the background and response to Boots Pensions should look at http://www.johnralfe.com/main_pages.php?page_num=4 )
2 Can DB pension funds “overpay” to hedge their pension liabilities?
Toby’s second point is that pension funds may be “overpaying” to hedge their liabilities through buying matching long dated bonds.
To understand if this point stands up, let’s make sure we are looking not just at hedging pension liabilities as a single process of selling equities/buying bonds, but as a two stage process, with the company effectively reversing this through selling bonds/buying equities.
Because this is a swap, if bonds are “overvalued” (whatever Toby’s phrase means, I am not sure?) so “value is lost” in the DB pension fund through buying bonds, then value (and the same value) must be gained by the company selling bonds. It is just two sides of the same coin which must net out.
So, unless I am missing something, Toby’s second point falls away when we properly look at both sides of the process.
3 Why not hold hedge DB pension liabilities with assets other than bonds?
Finally, Toby asks why not hold other financial assets, such as equities or commercial property to hedge pension liabilities. Not holding them seems “odd” to him.
Again, unless I am missing something, the answer to this is very straightforward. Suppose a company wanted to defease a 30 year bond issued 10 years ago, which, for whatever reason, it can’t buy back in the open market.
It would defease its bond by buying a portfolio of high quality 20 year bonds, with identical or similar cash flows in terms of maturity and coupons payable on the bond to be defeased.
The company just couldn’t defease or match its bond by holding equities, commercial property or any other asset, as the (uncertain) cash flows do not match the cash flows on its bond.
I think Toby accepts that DB pensions are bond like. (If not, please say so, Toby!)
Pensions promise to pay an amount based on formula related to salary and the number of years a person is in scheme. The promise includes limited annual inflation indexation – 5 per cent or 2.5 per cent – plus spousal benefits for widow/widower.
In bond terms it is a forward start limited index linked and longevity related bond. (The complexities of life expectancy, which can be estimated, are second order).
The pension promise is secured on scheme assets to provide some assurance that the pension will be paid decades into the future.
Because the DB pension is bond like, equities or commercial property don’t hedge, for exactly the same reasons they don’t defease the corporate bond.
Let’s make a final point of my own:
4 “It’s not about accounting”
*Disclosure : I was a consultant to the Accounting Standards Board on FRS17*
Toby’s blog keeps mentioning those pesky accountants who insist on valuing pension liabilities by reference to market interest rates.
This is true, but no more true than accountants insisting on valuing any bond by reference to market interest rates.
Accountants are trying to use the market value of both bonds and pensions. Both have a market value, which accountants use. And, as market interest rates change, so does the value of both bonds and pensions.
How else should DB pension promises be measured?
FRS17, IAS19 and FRS102 capture the economic value of pension promises and are a proxy for market cost. (Since the market cost of a pension buyout is higher, arguably the cost is understated).