Today’s crash of Inland Revenue’s website highlights a wider IT vulnerability that must be addressed – urgently.
The IRDs current tax collection system was designed in the early nineties. It was effective for many years. But it has become increasingly obvious that patches applied over recent years are a failing stopgap. New Zealand needs real progress on a replacement system.
The Government has already written off millions of dollars after a failed attempt to rebuild the student loans component of the IRD online collection system. Since then no progress, only bad news. Earlier this year, it was revealed that the Government estimates a necessary IT rebuild of IRD’s whole tax collection structure will cost up to $1.5 billion dollars.
New Zealand is not unique in having a tax system. Adapting successful systems used elsewhere is the logical approach to replacing our ailing IT infrastructure. This problem has been on the horizon for some time. It beggars belief that the Government has not yet outlined a convincing plan or timeline for the development of overdue IT solutions.
Voluntary compliance with a transparent tax system is critical to the efficient and effective functioning of our society.
Businesses rely upon our tax collection system working properly when they put their GST returns in. They don’t have time to waste reentering data when systems go down. We cannot afford to risk the patience of those citizens who wish to comply with their civic duty.
And testing the patience of businesses is not all that is at stake here. Our hospitals and schools rely upon a trustworthy tax-collection system. The IRD must not be allowed to fail.
The Minister of Revenue has not put forward a credible timeline for the project, and it is becoming clear that the government is sitting on its hands whilst New Zealanders face the consequences of an IT time bomb.
Not good enough Minister.
Isnt the website separate from the IRDs core computer system – mainframe type system.
Yes, the website is separate from but connected to the mainframe – filing over the internet has become quite common, so the crash is a major business continuity issue. In terms of upgrades, the need to update the mainframe systems – ie the backend processing – is separate from the frontend processing of the website.
The claim that because other countries have tax systems we should be able to just ‘adapt’ their system ignores the fact that NZ is one of the few countries that has integrated welfare payments (ie working for families) tightly with income tax. Add the complexity of NZs integration of GST with provisional income tax and it doesn’t look so easy. As for student loads – again, this is tightly integrated with income tax, being based on the taxable income. The sheer size of the problem comes from the greatest asset of the existing system: its tight integration. The ideal time to have updated the systems would have been in the early 2000s before 24/7 internet filing took off – certainly before the Kiwisaver tack-on added even more integration between income tax and other policies.
http://xkcd.com/932/
Well done David.
@gwwnz @Nobody staff at IRD are stretched and being asked to make do with less – when much more needs to be done. In my experience, civil servants will go to extraordinary lengths to keep things going if there is a clear path mapped out that shows the extra work will be worth it in the long run. This requires leadership from the Minister and that seems to be lacking in respect of the necessary systems upgrade.
So the current back end system is a whole lot of complex computer programs running on large mainframe systems.
A new back end system is whole lot of complex computer programs running on large mainframe systems but costs $1 billion dollars
This seems to be a classic nice to have.
The core data of tax records of individuals and companies in some sort of database would still be essentially the same.
If IRD is unable to run a web payment interface for its peak load then what faith do we have that they can even define a future system properly and run the process through to completion.
If you look closely at many retail core systems which show on the screen when the transaction is processed , eg Warehouse, and it has the look and feel ( bright colours) of something dating from the predecessor of Windows, DOS.
Old systems work, yes they have problems but leading edge systems will have problems too and seem to only advance the careers of those involved, who do the smart thing and pull out once its under way so their career isnt affected by the giant stuff up. Seen it happen many times.
Surely open-source is an option? Why does it cost that much money?
I’m in favour of removing the ‘Take’ from ‘Te Tari Taake’
David Clark
Are you saying the staff at IRD do not have the skills to operate a simplified computer system. Do they need upskilling perhaps ?
Having some experience with IRD both facially and via technology over the years I believe that they are first class.
@Matt: I suspect the issue with the website is as much hardware – the transaction volumes at peak times are huge and govt doesn’t have the money that banks etc have to build in lots and lots of redundancy. Re opensource, I believe they already use a lot.
@ghostwhowalksnz: the old system has become unnecessarily complex because of the way tax and social policy have developed in incremental adhoc ways. The intent with any upgrade would be to re-architect to a still complex system, but only complex because of the business requirements of doing tax and social policy, not because of the history of the system’s development. The third option (after new system and continue as is) would be to incrementally redevelop portions of the existing system over time to achieve the same objectives – but that isn’t sexy enough for senior executives who like to point in interviews to how they lead such-and-such an $X bn project.
Part of the problem with student loans is the legacy ‘law’ – whereas tax law generally only looks back 7 years (with a few exceptions) and each year is fairly stand-alone (only loss carry forwards to worry about), the student loans system has to look back all the way to 1992 and implement all the various law changes over that period as well. This makes it the worst possible candidate for switching to an entirely new system.
To give some credit, the IRD’s system is vastly superior to ACC’s and when you know your way around its eccentricities, can be pretty easy to negotiate.
An issue that I noticed (apart from website crashes) are that they have never figured out how to incorporate Working for Families into the tax return system properly – it got to the stage last year where it seemed that if the system spat out a Working for Families refund for any taxpayer, it would be put in someone’s too-hard basket because there were some questions that needed to be answered – but there was no front-end system for answering those questions when the return got filed (for example, a tickbox in a tax return). And they had no back-end system triggering the asking of the appropriate question – those returns just sat in someone’s anonymous basket until the taxpayer questioned it.
They also had problems with being unable to deal with a tax return and a donations rebate form being filed on the same day.
They and ACC together also had problems with designing tax returns that answer the appropriate questions needed in order for ACC to charge a correct levy – ACC liasons were in fact telling us to fill tax returns in wrong in order for the levy to come out right.
But all in all, IRD’s software system is far more modern and better functioning than ACC’s.