The new National Pensions Reserves Fund represents the most important initiative in economic policy for the past decade. At least 1 per cent of GNP will be set aside every year for the next 55 years and invested to provide for future pensions.

Very few other countries have managed to put such a fund in place, even though something like it is even more urgently needed elsewhere. That we in Ireland can afford to embark on it is a reflection of the strong fiscal position coming from recent economic success.

The fund is intended to take account of the inevitability of an ageing population: paying old-age and public service pension out of current tax revenue will not be so easy (or popular) as the number of pensioners per worker becomes greater and greater. By making early provision, and by investing well, the current working generation can preserve the old-age pension from the political pressures of rising tax rates that could otherwise lead to an erosion of pension levels.

The fund will be in accumulation phase for at least the next 25 years: that's a long time for investments to grow - or to underperform. All going well, a few simple calculations show that by 2025 we could easily see the fund growing to the equivalent of more than £100 billion. That amounts to 20 times the current value of the official external reserves, and four times the value of the national debt. The fund will be by far the largest investment institution in the country.

It is clear that we need to get the institutional and statutory arrangements right from the start. Draft legislation now before the Dail has that goal. But how well is the legislation designed? Two main elements have to be got right: the investment mandate and the governance structure. The legislation addresses both of these points, but it needs to be strengthened.

The manager of the fund needs to be given a clear mandate. Nobody wants the fund to hold only speculative dot.com shares; nor would the pensioners have been well served by a manager who plumped for recent initial public offerings by green chips such as Eircom or First Active.

There are some good elements in the legislation on this: no Irish government securities are to be held and investments are to be made on strictly commercial criteria. We would go further and have the law preclude the fund from holding Irish assets.

In addition, it is not hard to foresee that, without some restrictions on investing at home, there would be continuous political and other pressure from promoters for the fund to finance worthy-sounding but unviable projects. And think of the uproar if the fund was to sell local investments, thereby depressing their price.

Better to insulate the fund from these kinds of pressure right away. Besides, most of the investments should be in foreign assets if risk is to be sufficiently diversified; during an Irish economic downturn, we would be doubly hit, not only by a decline in tax revenues but also by losses on the fund's financial portfolio. There may be other constraints on the fund's activities that should be locked in now: for instance, if there are to be ethical constraints on investment policy, these should be specified from the beginning.

But then the legislation takes a wrong turn. Instead of recognising that stock-picking is a wasteful and risky activity for such a fund, it envisages the fund manager being given target investment returns to be achieved within specific risk parameters.

Much better, in our view, to simply define the broad classes of allowable assets and require the manager to invest sufficiently widely in these classes to ensure that the fund's return equals that on the average of these allowed assets. In other words, we believe that the fund should be established from the outset as a passive "index fund", tracking the whole investment market rather than trying to pick winners.

Index funds are cheap to operate and as safe as the market allows. Based on international experience, management costs of as little as 0.1 per cent of the fund per annum should be enough, whereas an actively managed fund would cost a multiple of that - with no assurance of better performance.

It may be that the draft legislation does not prevent the fund's commissioners from adopting an index-fund approach, but there is so much money at stake here that it seems only prudent to specify such an approach in legislation from the start. After all, even 0.1 per cent of £100 billion is £100 million.

We do not rule out that there should still be room for leaving the commissioners with some discretion on investment strategy: for example, it might be sensible to deviate from simply the average of the world portfolio of assets in order to take account of a few known and large risk factors for the Irish economy.

For example, it might be wise to hedge by specifying under investment in UK stocks, given that Irish budgetary and macroeconomic performance is correlated with that on the neighbouring island. When it comes to bonds, there might also be grounds for announcing a deviation from the world currency mix.

The investment mandates based on this index-tracking strategy should, of course, be published in draft by the commissioners and widely discussed in advance of implementation. Curiously, however, the draft legislation asks very little from the commissioners by way of informing, and listening to, the public. An annual report to the minister and compulsory appearance before the Public Accounts Committee is about it. Given recent events, one might have hoped for more.

Financial institutions require transparent and unambiguous governance structures. Unfortunately, the draft legislation seems to fall down here also. It starts promisingly by distinguishing between commission and manager.

The seven-person commission is to determine the investment strategy and to control the fund. At first sight this seems to offer a desirable independence between strategy and implementation, the latter being carried out by the manager. Unfortunately, it turns out that the manager - or in fact its chief executive - will be a member of the commission, and the commission will only act through the manager. So the separation of commission and manager, of strategy and implementation, is effectively lost. It is not too late to disentangle this structure, establishing a board that is more effectively independent of the manager in setting strategy and performing its oversight function.

Many were surprised to find the draft legislation providing that, at least for the first 10 years, the manager has already been chosen. It is to be the existing National Treasury Management Agency (NTMA). One might have expected the Government at least to go through the motions of trawling widely for suitable entities for such sophisticated technical work.

Still, the NTMA is an impressive and professional organisation, and has performed its existing functions admirably. There is no reason to doubt that it can do a good job within the mandate given to it. As is appropriate for the government debt manager, it has not been the most transparent of organisations: that will have to change in its new role.

All in all, there is much that needs to be amended in the current draft legislation. Even though no party political issues are at stake, time spent in debating it will be time well spent.

Patrick Honohan is a research fellow at the Centre for Economic Policy Research; Philip Lane is an economist at Trinity College, Dublin