One leading survey says the European Union's 15 member countries have nearly $1.6 trillion in pension assets. Another survey asserts there are 57% more in assets, while a third comes down between the first two. Which is right?
The answer might be that all three are correct, but they go about measuring pension assets in different ways. Still, the varying data cause pension experts to scratch their heads in puzzlement.
The European Federation for Retirement Provision, a Brussels-based trade association, comes in at the low side, at $1.562 trillion as of Dec. 31, 1995. Koen de Ryck, consultant to the group and managing director of Pragma Consulting NV/SA, Brussels, said his data largely are collected from the member associations of each EU country and includes only "second pillar" pension plans - private and public retirement systems that augment each country's state system.
London-based William M. Mercer Ltd.'s 1997 European Pension Fund Managers Guide states there were $2.454 trillion in EU pension assets, as of June 30, 1996. While the cut-off period is six months later than that of the EFRP data, it still does not justify a 50% difference in the totals.
Mercer officials include insured pension assets, while EFRP and the third provider, InterSec Research Corp., do not.
In Sweden, for example, the Mercer guide includes $130 billion in insured assets, while InterSec data basically include just the giant Allmanna Pensionsfonden system.
"From our point of view, we want to look at where there are assets available for external management," explained Stephen Oxley, vice president in InterSec's London office.
Insured assets are becoming a real quagmire, as continental European life insurers are starting to outsource management of some of their investments. Recently, for example, Aegon Nederlands NV handed two portfolios to PanAgora Asset Management Ltd., London (Pensions & Investments, May 12).
"Insurance companies are a real minefield," Mr. Oxley acknowledged.
InterSec data from year-end 1995 are closer to the EFRP numbers than the Mercer data. In contrast, InterSec's year-end 1996 data is virtually midway between the EFRP and Mercer figures. That's because InterSec officials estimate returns and cash flow factors in each market; data are revised later when harder figures become available.
Last year, the U.S. dollar's appreciation against most European currencies hurt European pension fund totals - with the notable exception of Great Britain, where a 14% average return in 1996 combined with sterling's gains led to a 25.7% jump for British pension assets.
There are a number of other notable differences among the three sets of figures.
In general, the Mercer survey tends to be more inclusive. That bias is most apparent in Germany, where the consultant shows $360 billion in pension assets, compared with $127 billion for the EFRP and $130 billion for InterSec, at year-end 1995. (For the benefit of comparability, InterSec's 1995 figures will be cited in the remainder of the story.)
Germany looms so large in Mercer's survey because 60% of the consultant's figure - some $216 billion - stems from book reserve plans, in which money is reinvested in the corporate plan sponsor. (Mercer also includes $72 billion in pensionskassen, $36 billion in direct insurance and $36 billion in support funds.)
Elsewhere, InterSec's $58 billion figure for Denmark includes the state Arbejdsmarkedets Tillaegs-pension system, with about $20 billion in assets, because some of the system's assets are managed externally. EFRP, on the other hand, excludes ATP because it is considered a "first-pillar" system.
Still, it can become confusing. InterSec excludes Italy's first-pillar system, the Instituto Nazionale della Previdenza Sociale, while Mercer includes the system's $52 billion in assets. In Sweden, InterSec includes the first-pillar AP Fonden system because some of its assets are externally managed. EFRP includes only the portion of the system that is considered second pillar. The data have different meanings to different sources. For money managers, they represent valuable market information, providing guidance on where to focus marketing efforts.
For policy-makers, however, the figures are important indicators on how well European countries have prepared for providing retirement pay. Most EU countries rely heavily on state pensions. The EFRP says 85% of retirement income in 1993 came from state systems, down from 89% in 1989. Even inclusion of insured pension assets does not greatly change the relatively low ratio of pension assets to gross domestic product, a widely used measure of how well prepared countries are for retirement.
Some experts believe the data might not be inclusive enough. E. Philip Davis, deputy head of the Frankfurt-based European Monetary Institute's monetary, economic and statistics department, questioned whether personal pensions, which generally are provided through insurance policies, should be included. However, Mr. Davis acknowledged it might be difficult to obtain accurate data, particularly in countries such as France where many individuals take out insurance policies to provide capital for their old age because of tax-free treatment of lump sums.
Valuation issues among insurers also are difficult, noted Karel Lannoo, a researcher at the Centre for European Policy Studies, Brussels. He said in four or five countries, certain assets - such as bonds or real estate - are valued at cost, making it hard to compare assets. A 1993 EU directive, however, requires insurers to show current values in an appendix, he added.