Tuesday, 10 March 2015

The Sad Panda

Spot the difference.

Sad PandaTlouamma

Only Agang aficionados will get it.

If you do, sign the no confidence and recall petition.

Friday, 27 February 2015

Banking Liberation Movement

Back in the 1970s, when a high-end supercomputer had about the performance of today’s entry-level cell phones and networks were expensive proprietary technology, ATM transactions were … wait for it … free.

And you earned interest on a cheque account.

Banks made almost all their money on the difference between the deposit and lending interest rates.

Core computer technology, based on Moore’s Law, is about on billionth of the cost it was 45 years ago (price per transistor roughly halves every 1.5 years).

Never in all the fields of human endeavour has such a massive improvement in efficiency been so extraordinarily wasted.

So what made everything so expensive? Not having higher paid more skilled staff in the branches – that has also gone backwards. The mind only boggles at how banks have destroyed such a massive opportunity. With careful design the cost per transaction could be almost zero, and saved costs shifted to quality customer relations.

If companies like Google and Facebook can offer free services on a massive scale, only making money on a tiny fraction of total transactions, how hard can it be?

Banks have fallen into the trap many enterprises fall into of trying to maintain outdated systems on the basis that it is too expensive to re-engineer them from scratch, with the result that their software accretes more and more layers of cruft and becomes harder and harder to maintain.

If banks cut their services back to what they had on offer in 1975, carefully coded to maximum efficiency and small total software size so it was manageable, then put a web front end around what you could do back in 1975, you would have most of what you can do today and it would cost a tiny fraction of what they spend today on software. The biggest cost would be ensuring you had the best possibly security (and some banks don’t even have that…).

So why don’t they do that?

Each major bank has accumulated an army of software developers dedicated to maintaining the complexity of the existing systems to maintain the need for an army of software developers. And if they all do it, they can pass the costs on to the customer.

Nice work if you can get it.

So what can we do?

How about this for a radical idea? Free banking software. The free software movement has delivered some of the best operating systems in use today, web browsers, sophisticated database engines and the most robust network stacks available today. Why not the back-end of a banking system?

It could be done in 1975 with one-billionth of the computing power available today. How hard can it be?

Sunday, 25 January 2015

Queensland Greens and Open Tickets

Queensland election update – less than a week before polling day.

Just Vote All?
Let’s examine the reasons for Campbell Newman’s Just Vote 1 strategy. Queensland state elections have optional preferential voting: you can number 1 or more candidates.

When it has suited the major parties, usually the one in the lead, they have adopted a Just Vote 1 strategy.

This makes it possible to win on a minority of the popular vote, and takes smaller parties out of the equation.

The ostensible reason for this strategy is it prevents a hung parliament. The UK has a first past the post system (essentially the same as Just Vote 1) and their last election delivered a hung parliament and the latest polls show that as the most likely outcome next time. So that is not a really plausible reason and in any case, up to Peter Beattie’s time, Just Vote 1 was the Labor strategy. Back then, the Libs and Nats were at each others’ throats. Times change and so, apparently, do principles when they no longer suit you.

Why does it suit the LNP now? After all, conservative minor parties and independents add up to a bit more than the Greens – depending which polls you believe. It’s a matter of marginals. The LNP will have done polling that shows they gain more from stifling preference flows from Greens to ALP than they lose from stifling preference flows to them from minor parties and independents on the right. I can’t find stats for past Queensland elections, but most other minor parties in the 2013 federal election had a roughly even split of preference flows to Labor – even if they were nominally on the right (possibly an indication of where Labor is positioned?).

So overall, Labor, who derive significant benefit from Greens preferences, loses out more than the LNP from Just Vote 1.

Labor is in a poor position to complain: optional preferential voting was introduced by the Goss Labor government and it worked for them when they could exploit the divide between the Liberals and Nationals.

Open Tickets: So What?
Meanwhile Labor is attacking Greens for issuing “open tickets”, i.e., How to Vote cards that tell the voter to make up their own minds. This is nothing new, and if we look at past elections, the effect has been pretty much that the ALP gets 80% of Greens preferences once they drop out irrespective of what is on the HTV.

Even in the 2012 election where ALP most certainly did not advocate Just Vote 1, nearly 70% of Labor voters (vs. nearly 80% of LNP voters) only marked 1 on their ballot. In doing so, they created the risk that if the Greens candidate came second, there would be insufficient ALP preferences to beat the LNP candidate.

50% of Greens voters also went for Just Vote 1 despite the fact that the Greens have never advocated this position. 80% of Greens preferences go to Labor.

What Labor should really do is to think about what it takes to be attractive to Greens voters, and focus on countering Just Vote 1.

They will win a lot more votes that way than going negative. Do the math: 50% of 80% is 40% of the total Greens vote (3-4% of the total vote) that is up for grabs as ALP #2 if they attack the Just Vote 1 strategy. For LNP, the extra votes they win if every Greens voter fills in every box is a quarter of Labor’s gain so no wonder they do not want everyone to fill in the entire ballot (stats from Anthony Green’s blog).

There is therefore absolutely no basis for the ALP complaint that Greens “open tickets” risk handing the election to LNP. The stats actually indicate that Labor voters are far more likely to behave in a way that is perverse to their interests.

The 20% of Greens voters who put LNP ahead of Labor are probably converts from LNP on enviro or other issues where Greens are ahead of Labor and aren’t going to put ALP ahead of LNP no matter what. What would Labor prefer? That these people stay with the LNP?

Labor meanwhile is issuing a call to “put LNP last” which implies filling in all the squares, and is getting huffy about anyone who suggests variants. No doubt they are worrying about confusing the message. If they had not invested so much in confusing voters before (like Peter Beattie’s Just Vote 1 campaign) they would have an easier time now. Their attempts at portraying the Greens as spoilers are not matched by the historical stats.

Sunday, 16 November 2014

Time for a National Imbizo?

Back in 1955, the Congress of the People was a national gathering for those excluded from the apartheid order, which produced the Freedom Charter, a blueprint for a non-racial South Africa. By the time of our first democratic election in 1994, there was a presumption that the Freedom Charter was dated in detail, but nothing was done to replace it. Nearly 60 years on, the country is increasingly directionless and it is not clear that there is a national consensus on the kind of nation we want to be.

In the wake of the expulsion of the National Union of Metalworkers of SA (Numsa) from Cosatu there is growing talk of launching a new political party.

Before Numsa rushes into creating yet another party, is it not time for us as a nation to sit back and take stock of where we are as a nation, and what our national programme should be?

Anyway, haven’t we been there before?
Voting pattern 1994–2014: the biggest change is in the fraction not voting
There was the split of the ANC in the wake of the Zuma ascendancy that led to the formation of Cope, which fell apart amidst acrimonious lawsuits. From 6.75% of the vote in 2009, they dropped to 0.67% on 2014.

Then in 2013 the EFF and Agang were formed, each as the great new hope for an alternative to the ANC. The EFF fared a whole lot better, with 6.35% of the vote, while Agang barely made it to 2 seats with 0.28% of the vote. Collectively, Cope, EFF and Agang only scored 7.3% – not much more than Cope’s 2009 result. Agang is now following the Cope example with acrimonious lawsuits and EFF has split, with results still to be determined.

The DA, meanwhile, has gained ground – but if you look at the picture, the opposition vote is no bigger than it was in 1994, when the ANC was led into the election by Madiba, one of the most revered political figures of the last century. The only really big change since 1994 is the declining fraction of the voting population who turn out to vote (the black zone at the top – a drop of nearly 30%).

So what we can see is that parties that try to appeal directly to the ANC base are stuck in band of 5-7% even when they come with a huge plus like taking away a big section of senior leadership (Cope) or a major part of the ANC Youth League (EFF). And parties that do not have such a direct appeal to the ANC base cannot attract much more than 20%. That the biggest effect since 1994 is the collapse in voter turnout says a lot about how little appeal opposition parties have to disaffected ANC supporters.

If the main unifying force of a new party is dislike another party, that cannot lead to a sustainable movement, which is why there is so much internal strife in opposition parties. Even the DA, the most successful so far, has had its internal conflicts.

Part of the crowd at the Congress of the People, Kliptown, 1955
Back to 1955: the Congress of the People was a national gathering of those excluded by apartheid and preceding orders. It resulted in the Freedom Charter, a document so long dormant that by 1994, the ANC did not consider it relevant in detail even if the ANC pays lip service to the main ideas.

What we need today is another national gathering, this time reflecting the full diversity of our society, to talk about the problems we face today – and to map out a new national direction.

Since the political process has not delivered this and indeed is failing in many very basic ways, civil society should take a lead. I propose we call a national imbizo under the auspices of major civil society movements. We have some in this country with a huge following and that have done brilliant work, filling the gap where government has failed. Examples include various movements of the unemployed, unions, Treatment Action Campaign, Equal Education, business organizations, environmental groups, Khulumani Support Group – and there are many more.

If we can call such a national imbizo and arrive at a common understanding of what we as a society see as our priorities, only then is it reasonable to talk about a new political movement. Even if this imbizo does not result in a new movement, it will help those in existing parties to understand where we are going wrong.