Screen scraping – don’t believe the hype

Since bursting on to the financial services scene in the middle of last year, account aggregators or screen scrapers have been one of the hottest discussion points in the industry. In the first of a series of articles exclusive to Money Management, Robert Keavney explains his opinion on the phenomenon.

It's the biggest thing since the discovery of America!

Master trusts will be redundant. Wrap accounts will be unnecessary. Dealers investing in data management systems, and advisers spending hours in phone/fax/email communication with fund managers, will be a thing of the past.

All of this will disappear with screen scraping - or so I have been told. Screen scraping is alleged to be the panacea for financial planners' administrative problems.

Unfortunately, it is not true.

Screen scraping is the process whereby a piece of software, armed with the password of the adviser/client with each fund manager, is able to visit the Web site of all fund managers, capturing current data for the insertion in portfolio reports.

Thus, the theory goes, planners will be able to be aware of all current details, of all clients' assets, without either manual processes to capture the data, or the need of a master trust or wrap account to act as a consolidator. This will enable savings for both advisers and their clients and remove all inaccuracies in portfolio reviews.

Unfortunately, as usual, reality won't work as well as theory. There are problems with both of the key assumptions behind screen scraping, ie that it will ensure accurate data and that it will reduce costs.

Aggregate costs for the clients of the Australian financial services industry must come down. Among other things, this requires access to low cost managed funds. One flaw in the view that screen scraping is a panacea is that it does not deliver a means to access low cost, wholesale products.

Screen scraping may be useful for an adviser who has clients in a retail fund with a 2 per cent management expense ratio (MER) to be able to efficiently capture data about that fund, but it doesn't help the adviser get the client into a wholesale product with a 1 per cent MER.

To do this requires the pooling of the assets of many investors in order to meet wholesale fund minimum investment amounts. This pooling needs to be in one name, ie a custodian, and hence we are back with a master trust or a wrap account, or similar structure. This is the only means for this pooling to take place.

Thus, to the extent it presumes most clients will be in retail product, screen scraping is self-defeating as a way of materially reducing costs.

Where assets are held in a custodian's name, the problem of capturing data is significantly simplified. The custodian effectively is the investor and knows when any investments are added or redeemed; when any income is paid; or when any other transactions on the asset are carried out. Data capture for custodians is a much lower volume and cheaper issue.

Certainly, the historical cost overlays of master trust or wrap account promoters are excessive and must come down. However, screen scraping won't make their function redundant.

Another assumption behind screen scraping is that the fund manager's data is always right and therefore the objective of the adviser is to capture that data.

At first glance this seems an eminently reasonable view. However, those with long experience in trying to put accurate data in client reports know that it is false.

The clearest evidence of the potential inaccuracy of fund manager data is found in the annual tax guides of many of Australia's leading managers. For example, BT's covering letter notes, in bold, that there are circumstances where their statements "may not be relevant". They advise that, in some situations, clients "should not use the BT CGT statement either this year or in future years", and go on to list 13 such cases.

The Colonial First State tax guide identifies six assumptions they have made and ten circumstances they have not addressed in producing their capital gains tax estimates.

It is not that these managers have done a poor job in record keeping, rather they have been particularly detailed in their advising of the limitations of fund managers' data.

When fund managers feel the need to warn their clients that they should not rely absolutely on the manager's capital gains tax (CGT) data, it would be rather foolish of advisers to ignore this.

There are very many cases where fund manager CGT data is inaccurate. These include:

CGT lots - When a client has acquired part of their holding in an asset on more than one occasion, then each of these will be a separate CGT lot (savings plans are an extreme example of this). If the client needs to make a withdrawal, an adviser will usually calculate which lot(s) it is most appropriate to redeem, given the client's CGT position. Advisers may include a specific instruction to the fund manager about which lot should be sold and ensure that clients' tax agents know this to avoid them relying on the manager's disclaimed and often misleading tax advice. The accountant ultimately determines which lots have been sold by recording this in the tax return. Almost every fund manager ignores an adviser's or client's instruction about which lot to sell and assumes the sale is on a first in, first out (FIFO) basis. (The liability consequences of this are intriguing, but lie outside the subject of this article. Fund managers optimistically rely on their disclaimers for protection. Advisers should not depend on tax statements which are so specifically and reasonably disclaimed). For our purposes, the key thing is that either the adviser must always assume a redemption on a FIFO basis, even if this is the worst thing for the client, or must accept that the fund manager's data, from the time of the redemption onwards, is unreliable;

Change of nominal owner - If a client holds an asset, and later lodges it as security for margin lending, and subsequently repays the loan and has the asset released from security, there is no change of beneficial owner. The asset will have originally been in the name of the investor, then transferred to the name of the margin lender's custodian, and finally been changed back to the investor's name. However, the beneficial owner of the asset, the client, has been constant, thus there has been no change of ownership for CGT purposes. However, most fund managers, with most margin lending facilities, will treat this asset as having been acquired by the individual and later sold. A new asset, completely unrelated as far as their records are concerned, and having a new policy number, was acquired by a margin lending custodian. This was later sold. Finally, again unrelated, the original investor acquired a new asset. Nothing in the manager's systems will track the continuity of beneficial ownership. Thus, any capital gains tax information provided on this asset will be false.

Change of ownership on death - When an asset is transferred in specie on the death of an investor or where one of two joint investors dies, most fund managers close one account and treat the asset as sold and open another one. They make no attempt to keep records that accurately trade the correct CGT history of this asset for the new owner. For CGT purposes, their data is false;

Fund mergers - When one fund merges with another, most managers' systems treat the date of merger as the start date for investors in the subsumed fund, even though this merger would usually not be a CGT event.

Many other events.

For a more complete list of the cases where fund manager data may be inaccurate, refer to the notices which accompany their annual tax advices.

If advisers base their reporting on accepting as gospel anything a fund manager tells them, they will need to abandon any hope of providing accurate advice on the CGT consequences of the sale of many assets, because their data will be wrong.

They will have to live with a constant need to apologise to clients for inaccurate reporting and will also need to practice an answer to the question, "Why do you keep using that data, when we agreed it was wrong last time?"

Planners who wish to offer service to demanding clients, ie those who require quality reporting, need to face the reality that passively sucking data from managers' systems won't solve all of their problems. They need to control data quality, ie to not allow inaccurate data obtained by screen scraping, or any other means, to overwrite accurate data.

This does not mean that scraping will not increase efficiency in many cases, but it does necessitate the continued application of intelligent filters at the planner's end. There is still work to be done.

There are other difficulties involved. Screen scraping could accurately capture the information that yesterday a client had 1,000 units, but that today ten have been liquidated, leaving 990.

These ten units could have disappeared as a result of a client redemption to fund an income payment (in an allocated pension), to pay a management fee (in those funds where fees are taken partly by sale of units), to pay tax (eg a contribution tax by a super fund), to correct a data error, or for many other purposes.

Reporting to clients that they have fewer units without being able to explain why, is not adequate. Thus screen scraping will only allow the adviser to provide meaningful reports covering all of these possible events when every fund manager and deposit taking institution:

Records them all correctly;

Uses an exhaustive categorisation system; and

Makes this information fully available on their Web site.

Further, the software doing the screen scraping needs to be able to interpret every item on every manager's site. Without doubt, this is achievable but it must be some years away.

There are many other assets, which can have a legitimate place in clients' portfolios, but for which a current value cannot be routinely available. Nil per cent Residual Capital Value (RCV) annuities are an example of this. The initial and end values of this asset are known, but the daily value is not and a life insurance company would need to use its actuary to determine a redemption value at any point in time. Thus, this asset is unlikely ever to have an always current value listed on a Web site in order to be scraped.

There is no doubt that there is much room for improvement in the accessing of fund manager data by advisers. Screen scraping will have a role to play in this.

In the long term, it is in the interest of all planners to put pressure on fund managers to improve their systems to the point where they can accurately report on the details of our clients' holdings with them. It is actually quite shameful that most currently cannot.

Until this arises, planners will need to build a fence around their data and only take in, by automatic extraction, information which they know is not misleading.

Of course, this adds to our costs. Nonetheless, those of us who are offering high quality service have no alternative if our reports are to be accurate.

Apart from this, our clients will need to access lower cost product and as price is related to volume, this means they need entree to wholesale. This requires pooling, which in turn requires a custodian.

The introduction of a custodian alters the data management process, reducing the apparent benefits of screen scraping.

Thus, while screen scraping offers clear benefits, it is not the panacea that some parties are presenting it as being. If it was, of course, it would make advisers redundant as data gatherers, as clients will access screen scraping directly.

Advisers who understand where they really add value would view this, not as a threat, but as a great relief.