A tale of two debates
Two issues have dominated debate within the Dutch pension industry over the past two years. One is the new Financieel Toetsingskader regulatory regime of De Nederlandsche Bank, the Dutch pension fund supervisor.
The other is the International Financial Reporting Standards (IFRS) specifically IAS 19 and its Dutch equivalent RJ 271, the standard which sets out the way corporate sponsors should account for the assets and liabilities of their pension funds.
A major research study, published earlier this year, suggests that both of these are unnecessarily prescriptive and are causing damage to the Dutch pension system and, eventually, to the Dutch economy.
The research, part of a series of studies by investment manager SEI of global markets over the past five years, was sponsored by its Dutch operation, based in Wassenaar in collaboration with Con Keating, principal of the Finance Development Centre.
Keating's chief criticism of the new FTK is that it is an over-reaction to a crisis that has yet to happen; specifically the failure of some Dutch pension funds.
In 2002, largely as the collapse of equity markets when the tech companies bubble burst. Dutch pension funds began to show deficits with respect to the book values of their liabilities.
Yet none of the deficits led to failures. "My main objection to FTK is I can't determine how large the problem was that prompted it. The evidence is that there wasn't actually any cost. We can't actually find a pension fund that failed. We certainly can't find a fund which failed entirely."
One effect of this has been to accelerate Dutch pension funds' transition, already under way in the Netherlands, from final salary to career average defined benefit schemes. In 2000, 60% of Dutch schemes by membership were final salary and 32% career average. By 2004 this had become 13% in final salary and 77% in career average.
In spite of the lack of any real pensions crisis, Dutch supervisors have responded with over-engineered regulations which are binding, costly and administratively complex, Keating suggests.
The most punishing of these new regulations is the minimum test. Under the current proposals, if a pension fund's assets, at realistic values, are less than 105% of the value of nominal liabilities, the fund must rectify this with a year.
The new regime also includes a solvency test, which states that the likelihood of a pension fund becoming underfunded within one year must be less than 2.5%. This requires pension funds to build up substantial buffers. The 2005 National Strategy Report suggests that an average pension fund with a typical allocation to bonds and equities will need a funding ratio of 130%. Funds will be given 15 years to reach this level.
Keating says the 30% buffer is unnecessarily large, and ultimately damaging to the economy. "An average standard of 130% funding is equivalent to a total loss of 23% of all pension schemes in the Netherlands, which suggests that the aggregate cost is excessive. That is some 30% of Dutch GDP. There has never been such a decline in GDP.
"This isn't a sledgehammer to crack a nut. This is closer to a nuclear weapon to crack a nut."
The 30% buffer can be seen as a kind of tax on pension fund members, he says. This is because no-one actually owns the surplus capital that is built up. The ownership and distribution rights of the buffer are unclear.
It is clear that the buffer does not belong to the sponsor. The presence of a buffer in a pure DB scheme, he says, resolves the accounting issue of whether a deficit or surplus is the property of the sponsor. The deficits and surpluses are not debt or assets of the sponsor, whether this is a single sponsor, as in a company pension fund, or multiple sponsors, as in an industry-wide scheme. They are consequently outside the scope of IAS 19 or its Dutch equivalent RJ 271.
Yet the ownership and distribution rights of the buffer proposed by the DNB is undefined, he says. "If you think of the FTK in terms of a quasi corporate balance sheet you've got nominal liabilities of 100 and 30 of equity - the buffer. As it is at the moment, the buffer will continue there in perpetuity. So if I'm member contributor, I'm never going to draw that 30%. So the buffer is effectively a tax on pension fund members.
"The property rights over the buffer are critical. As a member contributing to the scheme, I have my nominal liability guarantees, my nominal benefits, which are the debts in the fund. But I'm also an equity shareholder in the fund. So when do I get my share of the 30%? As it is at the moment, the buffer continues to be passed on to future generations."
Keating argues that maintaining a buffer within the scheme for the benefit of future generations - for example, by lowering contributions - is perverse from the perspective of intergenerational transfers and risk-sharing. "Leaving things to future generations is completely the wrong way around. It should be our children supporting us," he says. He suggests that the only parties who will benefit from the use of surplus capital of a buffer will be the pension funds. "What it means is that the people who will make free with it ultimately are the management of the schemes."
The minimum test of 105%, however, is the most draconian measure of FTK because its impact is immediate. Under the current proposal, due to come into force on 1 January next year, pension funds will be given a year to restore their solvency levels to
the minimum required by the regulator. Critics of the minimum test want this period extended to four years.
One of the immediate effects of the one-year requirement will be to promote a shift in asset allocation, as pension funds are forced to switch from riskier equities to less risky bonds. There are signs that this process is already under way, says Keating . The average duration of bonds in the typical portfolio has increased from five to six years.
This will have wider macro-economic effects, says Keating; specifically, raising the future tax requirement of the Dutch government.
"We're faced with a demographic problem. There are going to be fewer people working. Can we tax those people at the level necessary to support those who retire? The answer is probably not," he says.
"The only thing you can really do is either increase total production or increase productivity. If you go to bond investment rather than equity investment you don't achieve this - quite the opposite. Because if I increase savings and I invest in government debt, what I'm actually doing is raising the future tax requirement of the government. So if I'm increasing savings to put the savings to good effect, I have to invest in equity.
"If you go into fixed income, you do not solve the savings problem. That is deeply problematic for government, and is the big unintended consequence of the current wave of risk-based regulation of which the FTK is the prime example."
The central problem of the new FTK is that, like all risk-based regulation, it draws heavily on banking regulation as a model, says Keating. Yet pension funds and banks have very different time horizons.
"It is absolutely clear that the FTK is lifted from banking regulations and the economics surrounding banking regulations. In banking it is justifiable because, with a bank, your liabilities can become immediate. But with a pension fund this isn't the case. With a pension fund your liabilities arrive marginally every year over 20, 30 , 40 years, and they leave marginally, every 20,30, 40 years."
In short, the requirement with banking is immediate, while the requirement with pension is future. So imposing a banking requirement on a pension fund can be highly damaging, he says
"Let's say I have a single liability in 25 years time of a million dollars. And I'm going to have one shock. In one year, I'm going to lose 50% of the value of my assets. Which is worse for me - losing it tomorrow or losing it in year 24? In the case of pensions it is quite clear that losing it in year 24 is much more worrying than losing it tomorrow - because you've got 24 years to make it good.
"Now banking regulation, and all risk-based regulation of this form, says the one that matters is the one tomorrow. That's the one that crucifies the present values."
Keating argues that this is forcing investment managers to disregard the accepted rules of risk and return. "If you measure everything immediately, as you do in for example the banking world when you use Value at Risk, then returns don't matter because the time horizon is one day, and the return from an asset over a day is negligible.
"But if you look at the long term, the return becomes very important. It actually becomes the most important element. Risk in any financial setting is about the proportional rate of change of an asset or liability, which we call return.
"Yet everyone wants us to believe that I should accept an asset that gives a lower return because its risk-adjusted value is the same as the others. There's just no evidence for that. You maximise the likelihood that you will be able to pay something in the future by investing in the asset that gives you the highest return."
Another drawback of the 105% minimum test, with its one year requirement to rectify the situation, is that it makes investment ‘path-dependent'. Path dependency means that if A happens a pension fund must do B. The supervisor, rather than the pension fund, decides the route the fund must take to reach a certain outcome, in this case to restore solvency levels.
In contrast, a path-independent process is one where all that matters is the final outcome, and not the route by which that outcome was attained.
Path-dependency increases costs, says Keating: "If you force that, in a path-dependent fashion, you multiply costs. If you make a pension fund buy government bonds then the costs will be high, not just for the fund but for the economy as a whole. And for everyone else this will hurt the social welfare more than anything you can imagine."
One way out is to make the minimum test non-binding, he says. "These tests are informative in some ways, but the problem is that you make them binding. The regulator and the supervisor have effectively intervention and control rights. That makes the testing process costly.
"If a pension fund has to make good its 105% level within year, it could cost the fund between 0.5 and 1.5% of its value each year. Over the life of the fund that is equivalent to 10% of the value of the fund. That's a staggeringly large number."
Keating suggests that the remedy lies changing the role of the regulator, in respect of the new FTK, from one of controller to one of monitor. "A better system is to have a supervisor as monitor, someone who has a right to information and can publish that information. "If you make the FTK non-binding and allow people discretion and negotiation, then the supervisor's role becomes one of information provider more than anything else. The control rights remain with the pension fund membership. If there are disputes then the supervisor becomes the court of arbitration."
Keating recognises the risks of making the FTK non-binding. "People will stray across the boundaries, probably much more than they might otherwise. But many won't. And the likelihood is that such is the Dutch ambition with respect to pensions that within two or three years they will have far higher ratios than are implicit here."
Ideally, he says, the Dutch pensions industry should have been invited to devise a collective scheme of their own to prevent pension fund failures.
"If the government had asked the pensions industry to come up with a system whereby it collectively bailed out pension funds that were in danger of going bust, what would the cost of that be? Contributions to the UK Pension Protection Fund are capped at half a percent of liability. I'm not suggesting this is the right number, but it is a lot less than 30 %."
Keating broadens his critique of the FTK to include some of the key legislative and accounting issues behind pension reform, and identifies what he sees as some serious conceptual flaws. Foremost of these is the regulator's use of the interest sensitivity of liabilities. Keating argues that the sensitivity to interest rates of the present value of liabilities is not an economic sensitivity:
"Pension funds' sensitivity to interest rates is spurious. Where is the interest rate in the formula for the determination of my pension? There isn't one."
He argues that the interest rate element is an accounting convention rather than an economic reality or risk source. "The central role of the interest rate sensitivity of the discounted present value of liabilities in the determination of risk-based buffers is misguided, costly to the sponsor and beneficiary, and counterproductive at the national macroeconomic level," he says.
The logic of this is, he says, is that the DNB should remove the interest rate element of the standard FTK buffer calculation which applies to liabilities.
If the present value of pension fund liabilities are not sensitive to interest rates, then it follows that efforts by pension funds to hedge or neutralise interest risk with liability driven investment (LDI) strategies are also misguided.
"LDI can be seen as the hedging of an accounting convention , not any economic reality," Keating says. "By using derivatives such as swaps in the attempt to hedge liabilities a pension fund is actually introducing a new sensitivity to interest rates, since a swap merely consists of a long term fixed rate loan paid for with short term variable interest rate costs over the term of the swap.
"It is difficult to believe that this is the intention of most of those who are considering derivatives, or that it makes any sense for the majority of pension schemes. "
Keating concludes that LDI strategies a ‘folly' which will prove extremely expensive to the pension finds that adopt them. "LDI is a bit like writing insurance. The problem with being in a business like that is it's very difficult to leave. Because you've got ongoing commitments to the counterparty and he's going to say no. LDI is a business with barriers to exit rather than barriers to entry."
Accountants have much to answer for, he suggests, in creating a situation where the wrong solutions are being devised for non-problems, One of the biggest mistakes accountants are making is to accept without question the assertion that pensions are ‘bond-like', he says.
"The statement that pensions are bond-like is one of those apparently harmless statements that you feel guilty about challenging. But is there a case where a pension fund has ever bankrupted its own parent? Until the follies of the last few months and the UK's Pension Protection Fund, it has never happened. "It is also a very strange bond or debt instrument where the creditor has never foreclosed on the issuer. And in the 1990s, pension schemes in Holland, as well as the UK, were increasing benefits to pensioners voluntarily. You never come across a bond where that happens. These two characteristics say to me that pensions aren't bond-like."
Equating pensions to bonds leads to accounting errors, he says. "The moment you accept that pensions are bond-like you immediately assume that you can discount them at a bond rate and work reliably with present values. That is plain wrong.
"The simple mistake that's being made, by the accountants more than anyone else, is this question of market prices for assets and discounting present values for liabilities.
"Whenever you have a model for something and it starts telling you to do strange things, the first question you should ask is ‘what's wrong with my model?'
"When you go into an accounting standard that suddenly says from being massively over funded we've all now got ever deepening black holes all over the world, the first question that should have been asked is ‘what have we got wrong in the accounting?"
Keating suggests that IAS 19, the international accounting standard based on fair value accounting, is not fit for the purpose of a pension application and should not be used. "Fair value accounting needs a fundamental revision as it currently introduces material biases and spurious, irrelevant volatility to the pension scheme and the sponsor balance sheets and income and expense statements."
For example, assets , under both the FTK realistic valuation and IAS 19, use market prices in fair value accounting. Yet Keating points out that the measures in use differ between asset and liabilities. Market prices and interest rate curves are not a common standard measure. This, he says, introduces biases into estimates of surplus or deficit.
He concludes that "accounting measures should not influence the asset allocation of any pension scheme". Keating suggests that rather than calculate present values, accountants should instead project the future cash flows of the benefits that are contractually payable. They should then project the cash flows arising from assets and compare these future distributions at the future dates. If there is an imbalance, it can be corrected over time.
The level of contributions can be determined by comparing cash flows before and after the introduction of new liabilities and priced according to the current market price of the portfolio of assets.
"It is perfectly feasible to compare directly the future cash flow of an asset portfolio and the future cash flows of a collection of liabilities," he says. "It is also possible to estimate the marginal cost, the annual contribution, of new liabilities by reference to the assets and liabilities," he says." This simply involves comparison of the two cash-flow distributions.
"The marginal contribution is simply the market price of the proportional addition of assets necessary for balance. This would also mean that current cost of the contribution would fall as asset prices fell and rise as asset prices rose." Corporate sponsors would find this attractive, he suggest "When asset prices are low and the costs of corporate finance are correspondingly high, the contribution amount is low. The inverse is true when asset prices are high." Yet in spite of his criticisms of the current regulatory and accounting regimes, Keating is sanguine about future prospects for the Dutch occupational pension system.
"It might be easy to conclude that the Dutch private pension scheme is in crisis. It isn't.
"It is in a state of some confusion and much worry, and that can be clarified and the sources removed. It is almost certainly in a far better condition than either the UK or US systems."
He is also generally supportive of the Dutch regulatory bodies. "Many of the criticisms are predominantly
criticisms of the regulatory world in which the DNB and FTK exist, and with which there is social pressure to co-exist, and in some instances even a requirement to comply," he says.