Sunday 8 February 2009

Not burning our money

I wrote recently about an example of how the public sector burns our tax dollars. Now comes an example of how difficult it can be to burn money - and why talk of a "stimulus" often turns out to be just "talk".

The agency I worked for spent about half a billion a year on capital works. Our section got about $20 million of that, and my share in any given year was usually around $2-4 million.

We worked on a financial year of July to June, so in about January, the budget bean counters in Finance would ask us for our budget forecasts for the following financial year. Our management team would find a room with a big whiteboard, organise some coffee, and do a bit of brain storming to work out our capital bids.

All the managers would rock up with their wish list of pet projects, and we'd then rank and rate them using a simple matrix like this:

Obviously, our matrix was much bigger, often having 50 or more projects that someone or another would deem worthy of being fed by our tax dollars. Fuckwit, who I have blogged about before, would present a list of utterly ridiculous and inane projects, usually as a result of him being sold an idea by a salesman that week, or from him reading about a Bright Shiny New Idea in a magazine.

The only way to keep his mitts off your money was to cancel all his magazine subscriptions, and to banish him to somewhere like Mudgee, where salesmen would never find him. He was the greatest fadist I have ever met. No fad was too useless and ridiculous for him.

Here's the way the matrix works. You are comparing pairs of items, and giving them a score of 1 (we prefer this item) or 0 (we don't prefer this item). After working through the matrix, you total up the scores, and the item with the highest score is top ranked.

For example, the first pair to compare is a fridge vs air conditioner. The fridge is preferred, so it gets a score of 1, and the air con gets 0. And so on. If it wasn't for Fuckwit, the process would have only taken an hour or so. He could make it extend for the better part of a day.

This technique is taught on some of the more useful MBA courses, so if you are a manager faced with a plethora of competing priorities, this is a good way to do a brain dump and crystallise your thinking and show how you should be moving forward vis-a-vis your strategic direction in the context of your visioning environment. Blah blah blah. Out of all the management crap that I have been fed over my career, this matrix is one of the few gems that has produced any value at all.

So after a day of argument, usually 5 of us ganging up on Fuckwit to shoot down his stupid ideas, we'd have a clear consensus on what we wanted to do the following year. We had a rough idea of how much money we could ask for, so we'd start at the top and price up each project, and work our way down the list until we ran out of our notional budget. If we reckoned we'd get $10 million, and each project was going to cost $500,000, we'd ask for 20 projects. If the big boss came back and said we could get more money, we'd look at our list and add more projects in ranked order until the budget bucket was filled. If the budget was being cut, we'd simply lop off the lowly ranked projects.

Knowing the ranking of each project removed all argument when the amount of money in the budget bucket went up and down - we had all participated in the ranking process, so we didn't have to revisit it to add or remove projects. It also made it easy to sell our list of preferred projects to our staff, because they could understand the logic that had been applied to create the list.

Every group has a few moaners, who always think that their managers are a bunch of clueless monkeys, but once they were shown the matrix, they generally agreed that we had managed to descend from the trees for a short period and produce something halfway sensible, and earned our ration of bananas for the day.

If only our political class used a similar system for determining how they blow our money on their pet projects.

By the end of January, we'd have submitted our bid to the bean counters, and they'd put it in with all the other bids from the other divisions, and they'd spend the next few months sorting and deliberating and arguing and so forth, and then they'd go to Treasury to present the bid for the entire agency, and then once Treasury gave them a number, they'd come back and tell us what we were getting.

Around June 30, we'd be madly spending the remains of the budget for that year, and wondering how much cash we'd be getting the following financial year. In August, we'd still be waiting. In September, we'd still be waiting. By then, all the projects from last financial year would have finally wound down and been wrapped up, and we'd be cooling our heels in the coffee shop, twiddling our thumbs and looking for things to do. Some would spend the time cleaning up their files, or rearranging the artwork on their office walls, or counting whiteboard markers. I'd usually wrap my projects up long before anyone else, so I'd spend a week or two in the snow, carving first tracks each morning before the lifts opened.

Finally, in October or November, the bean counters would return and tell us what we got. By that point, we had 7 or 8 months left to spend the entire budget allocation for the year.

The first step was to write a Project Brief - a document around 100 pages long - explaining why that project should be approved. You'd think that once money had been put aside for the project, that would be enough. But that was not sufficient to satisfy the paper hungry burecaucrats.

Each year, the brief would get longer and longer, as new requirements were added by this department or that. When I started writing them, they were around 10 pages long. After a few years of Treasury, ICAC and the Premier's Department adding crap, my last one was about 120 pages - and in the scheme of things, that was fairly short! Other groups within our agency had all sorts of additional requirements to meet that we didn't. As you can see by my blog posts, I am a prolific writer, so it took me no time at all to write my briefs (comparatively) - only around a month of total slog, working 6 or 7 days a week, 12 hours a day.

Those public servants who refused to work more than a 35 hour week, and who idled their time away in frivolous meetings, could take 4 or 5 months to produce their briefs.

Fuckwit, who was essentially functionally illiterate, never produced a legible brief in all the time I knew him. Our boss had to write them for him - a case of delegation in reverse. And then our boss got me to polish them, which shat me no end, knowing that Fuckwit was downstairs in the coffee shop, having a smoke and telling everyone on his work provided Blackberry how hard he was working and how stressful his job was.

Once written, these briefs had to be approved. Depending on the amount that we proposed to spend, they might have to go all the way to the CEO for signature, collecting a myriad of other signatures along the way. Six signatures was pretty typical, but I saw some that required as many as 14. Since they could only be approved by someone writing their name in ink, the briefs had to be printed and circulated on paper. For my briefs, here is a typical travel route:

  1. me to my boss - same floor, both in building A
  2. my boss to Finance - travel from building A to building B
  3. Finance to Capital Works - travel from building B to building C
  4. from Capital Works to the boss of my boss - building C back to building B
  5. boss boss to boss boss boss - from one floor in building B to another floor in building B
  6. from there, to the Program Office - off to building D
  7. from the Program Office back to me - building D to building A
Our internal mail system was rubbish, so there was a 1 in 10 chance that the paperwork would get lost travelling between any pair of locations. Of the 8 or so briefs that I would write each year, I'd generally have to reprint half of them to replace lost copies. It was not unknown to have to reprint some 3 or 4 times. It was a really big problem for people in distant offices, who had a 50% chance of losing their paperwork at any given step.

I pitied our CEO, who had to read and approve around 500 of these things each year, some running into 300 or 400 pages. The volume of material meant of course that he never gave them more than a cursory read, so all sorts of crap slipped through. If they'd applied a bit of common sense and kept the briefs down to 1 page, chances are he would have read them properly, and that would have stopped a lot of rubbish from being approved.

No one had the guts to say that the system itself was producing and facilitating bad results - not even our CEO. Rolling back paperwork means reducing bum coverage in case of a disaster, and no one is willing to take that risk.

Bureaucrats above us thought that if they added more requirements and conditions to the briefs, that would filter out a lot of the crap. Thanks to the law of unintended consequences, the more stuff that was added, the more likely it was that crap would slip through. Sometimes, the simplest things in life are the best.

So here is our timeline so far - we find out what was approved in October, and then spend a month writing briefs to get a further layer of approvals. They get sent up for approval in November, and come back approved or rejected in December - just in time for everyone to go on holidays for 3 weeks or so.

A sensible person might think that you could save time by starting to write the briefs back in July, before anyone knew what was approved. If you wrote 90% of the briefs before the budget approvals were known, it would only take a week or so to finish them off and put them into the approval process.


Every year, the format of the briefs changed. Sections were added, deleted or moved. Or simply renamed. The font they had to be produced in changed one year. The spreadsheets of financial information that had to accompany each brief also changed. It was easier and less soul destroying to simply twiddle your thumbs for a month or two, rather than write something that would end up being a complete waste of effort. Better to save your energy for the real thing.

Everyone would come back to work in mid January, and we'd start the process of actually running a project. Except that nothing could be done until our approved project was loaded into the project tracking system, and assigned a capital works number. Each required a stack of paperwork, a long wait, and then a number of increasingly abusive phone calls to each department in order to shake them out of their post-Christmas torpor.

At last! We have a project number, a budget allocation, and account numbers that we could use to actually spend money. This would be the end of January or early February. But could we start writing purchase orders?


I now had to write a detailed Project Plan (another 100 page document), provide a detailed GANTT chart to go with it, along with monthly expenditure forecasts, a resource plan stating how many staff were required, plus a risk analysis and get all that loaded into the project tracking system. Phew.

Let's say my project was to build a bike parking station. For that, I'd need to buy some bike racks, engage a builder to install them, buy some trees and plants for landscaping, and engage a gardener to put them in.

Buying the bike racks would appear to be fairly straightforward, except that I would have to go through the following:

  1. write a submission to get approval to get quotes for some bike racks
  2. depending on the estimated value, this submission might require CEO approval (see great chain of approvals above)
  3. once approved, I'd see our Contracts Management people and write a tender for bike racks. It didn't matter if there was only one supplier in the country for bike racks - it still had to go to tender
  4. that would be reviewed by Legal (who were in a different building to us, and Contracts were in yet another building)
  5. the Contracts group would issue the tender, often to bidders that I had not asked for, and would leave out the companies that I wanted to use
  6. make many abusive phone calls to Legal and/or Contracts for their balls-up
  7. re-issue the tender
  8. give respondents 3 weeks to reply
  9. in the meantime, write a Tender Review package, and get that approved by Contracts and Legal
  10. rewrite the review package to meet their nit-picking requirements
  11. form a tender review panel, consisting of me and two other people
  12. revert back to the original review package, because they changed their minds partway through
  13. get the tenders, convene the tender review panel, review the tenders and submit the results to Legal and Contracts for review
  14. rewrite the review because they changed their minds about the format yet again
  15. Contracts would then tell us the value of each tender, and thus who had won the tender. This was usually a completely unsuitable company, so we'd spend a few weeks arguing about which tenderer to use
  16. write a submission to have the tender approved, and then get a brace of signatures on it - sometimes up to CEO
  17. get Capital Works to give that particular bit of expenditure its own account number
  18. have a long and depressing conversation with the winning bidder, who had just put their prices up because the 30 days on their bid had long expired
  19. write another submission to vary the original submission, asking for more money
  20. get Contracts to issue a contract, which of course had to be reviewed by Legal along the way (and me, because they usually made a balls up of that too, resulting in the wrong stuff being ordered).
  21. along the way, every week or so, I'd have to report on the progress of my project. I could spend as much time reporting on my lack of progress as I spent on actually fighting the project through the swamp of processes.
If I was lucky, the supplier would get an order before the end of June, and might even deliver before the end of June, meaning I could get their invoice processed before the end of the financial year. Often as not though, the stuff would be delivered in July or August, and all sorts of eyebrows would be raised as to why I was such a lazy git, and failed to expend my budget allocation in the required year.

I once got into hot water for daring to explain this process to a bunch of very senior managers, who were so far above the fray, they had no idea of what was required to get a project off the ground. Back in their day, all that had been required was a 1 or 2 page submission at the start, and that was it - you got your approval in say late August, and off you went. They had no idea about all the crap that had built up since they had been promoted into the exalted realms of upper management.

I got into trouble because I produced an honest timeline for them and showed that it was actually impossible to spend any money in the required year, even if all processes went like clockwork.

If paperwork was lost, as it invariably was, or critical staff went on leave for 2 months (as they did), or one or more departments decided to change their templates or processes partway through the project (as they always did), then it could take 16 months to undertake a process that always started 2 to 3 months into the financial year.

The only way to shoehorn 16 months of paperwork into a 7 or 8 month window was to lie, cheat and fudge - which is what we all did. They didn't want to hear that, even though it was an unavoidable fact. Once people got used to lying, cheating and fudging as the normal way of doing things, that opened the door to all sorts of unethical and corrupt behaviour.

Many managers just gave up when faced with all this crap. It was just too much for them, and they expired in despair. Projects simply fell by the wayside, and sat their gathering dust until someone noticed that nothing was happening. At that point, a manager would get really upset that no one had noticed this was occurring, and an even more onerous project management reporting system would be implemented, making it even tougher for anyone to actually get anything done. More forms would have to be filled out, and more reviews instituted.

When I left that particular agency, its projects were going so badly, a whole new team of Project Managers were recruited to help things along. I was quietly informed that if I applied, one of those $140,000 - $180,000 a year jobs would be mine.

At the same time, I was told that an additional layer of approvals was being built into the process - something I estimated would add two months to the timeline - if they were lucky.

I declined to apply, and left for greener pastures.

Our area was outsourced. I handed over my projects to the outsourcer.

4 months later, I heard that the outsourcer was recruiting 6 project managers to do the work I had been doing. The pay rate was about 50% higher than what I had been earning. Feelers were put out to see if I was interested in one of those jobs.

No thanks.

I was walking past our old offices one day and ran into my old boss. We sat down and had coffee and a chat. At that point, about a year into the contract, the outsourcer was already 9 months late on all my old projects, even with six people driving the work.

Anyone who deals with government knows that there is a collosal rush in June to spend as much money as possible. Now you know why. When so much money is being shovelled out the door in a short space of time, there is an enormous risk of completely wasting it. It was quite common for our area to buy several million dollars worth of stuff in June, and then leave it stacked up in every corner of our offices as we spent the next 9 months trying to get it installed. The time it spent in boxes was of course money wasted - but it was more important to spend the money in a timely fashion, rather than to spend it well.

I had to laugh when I read that Rudd wants to spend money on building stuff for schools. The Education Department has processes as stupid and bizarre as those I have just described, and you will have the added complication of a Federal-State relationship to deal with. If the stimulus gets approved this month, the first bricks will be laid in schools in around 2012 - by which time the economy should be growing briskly again, and this will simply add to an inflationary boom.


No comments: