Amazon is lazy in the one place it should be hard-working.

The New York Times published a great story about Amazon 10 days ago. I’ve thought about it a lot. Amazon treats its employees harshly. It fires a lot and forces even more out by making their lives hard.

I wrote a story last week about how this won’t work well. Most people can’t hack 80 hours a week. Our productivity drops. But a tiny percentage thrive.

Amazon wants to hire the best and brightest. And that’s terrific.

Top 2.5% of smart people are marked in red

There’s a bell curve of bright people. Amazon wants the ones on the right. The red group who represent the top 2.5 per cent. And it is prepared to pay for them. Pay is very good. Software engineers get $76,00 to $148,000 a year in salary. And there are big stock bonuses if you can hang around.

But it also wants to hire the people who want to work first and live second. The 80-hour a week crew. That’s great. Those people need a place to work that welcomes people like them.

Amazon has a right to target such people and it should expect benefits. Here’s a distribution of people by how much they want to work. Amazon wants the red people on the right.

red = diigent.
Top 2.5% of hard workers are marked in red. Emails after midnight! Woo!

But here’s the thing. The red group in graph 1 is not the red group in graph 2.

There is probably some correlation between the two groups, but it will be imperfect.

Here’s a hypothetical distribution of best and brightest, with the people from the top of the willing to work graph marked red.

red lines are tough bastards.
Where the hard workers might fall on the ability spectrum.

Amazon has a problem. The people they want to hire are very few. It can hire smart people, but they won’t all be willing to work hard enough. It can hire hard workers, but they won’t all be smart enough.

There are two ways of solving this problem. The easy way and the hard way.

The easy way is to throw people in the deep end and see who swims. This is a scatter-gun approach that says “recruiting candidates with a high chance of success is hard, and we are lazy.”

Screen Shot 2015-08-24 at 8.40.08 pmThe hard way to solve the problem is to narrowly recruit more people with a higher chance of success.

Rather than hiring “high volume recruiters” maybe they should look into hiring “high-accuracy recruiters.” This will cost money in the short-run. But it will probably be smart in the long-run.

Staff learn on the job. I’ve never had a job where I reached my potential in my first year. By burning so many staff so fast Amazon misses out on improvements. Working people to the bone in their first six months probably gets as much performance from them as they would achieve if they’d been there six months longer.

It’s not smart and the PR cost of hiring so many people that hate working there has blown up.

If you add the “externality” effect of lives damaged, marriages ruined and kids’ birthdays missed, then the cost of Amazon’s lazy recruiting strategy is even higher.

There is a moral imperative for Amazon to stop hiring so widely. Seriously. Stop it. Now.

The one piece of good news in all of this is that the New York Times piece will probably weed out some applicants who are unsuited for the company.

I suspect the names of journalists Kantor and Streitfeld are mud at Amazon HQ right now. But their expose may have actually done the company a favour.

The puzzle of where men outnumber women, and vice-versa.

The concentration of men and women in various parts of Victoria is stronger than you’d expect according to random variation.

There’s some real trends in place that seem like a genuine puzzle. (These charts are made from fascinating data released by the ABS this week.)

Screen Shot 2015-08-20 at 11.51.11 am Fully a third of postcodes have a gender ratio that’s skewed more than 5 per cent one way or the other. There are slightly more postcodes where women outnumber men by 5 per cent (79) than those where men outnumber women by 5 per cent (62.)

That’s to be expected because there are 99 men per 100 women in Australia.

But where men outnumber women they do so by a lot more.

The top result is Port Melbourne Industrial, which is a place I’m surprised anyone calls home. And indeed there are just 9 males to 4 females (ratio 2.25). Security guards who sleep among the containers? Who knows.

The next one is Braeside. A similar story with a ratio of 12 men to 6 women.  Then Alps East with 18 and 9. I’d like to imagine those 18 men have swags and wake each day to see their horse breathing steam under an old ghost gum.

Anyway, we can discount those three because the samples are tiny.

Screen Shot 2015-08-20 at 11.50.54 am
Female dominated suburbs include Chelsea and Rosebud. More men live in Seymour. Is this another case of nominative determinism?

Rosedale is the real thing. 2645 men to 1863 women (ratio 1.42). A little hamlet out in the Latrobe Valley, it is probably full of people working in the coal-fired electricity industry. A hard place for a fella to get a date, no doubt. Although the photos on the Rosedale Tavern’s facebook page suggest that’s where the local ladies go. (and it’s not as bad as East Pilbara where men outnumber women 350 per 100.)

It’s easy to explain some locations of high concentrations of men by reference to workforce pressures. They are found around heavy industries and agriculture.

Some are more tricky. Why is Footscray so full of testosterone? Why Docklands?

And why do women crowd into the expensive eastern suburbs? We see Burwood, Camberwell and Armadale in the top 10 with less than 90 men per 100 women.

Toorak, the suburb most emblematic of wealth, has a ratio of 91 men to every 100 women. Are there many young single women there perhaps? Or families whose daughters live with them for a long time?

The CBD , meanwhile, has a ratio of 107 men to every 100 women.

Perhaps we are seeing women self-select into suburbs they deem are very safe, while men are more willing to live in supposedly rough areas?

Do you have another explanation? Please feel free to share it in a comment below!


Commenter Matt points out that women live longer, which is a very good point (that I wish I thought of). This is definitely part of the explanation as we can see in the graph for the most skewed suburb, Burwood:

Screen Shot 2015-08-20 at 1.21.51 pmBut it’s not the whole explanation. If it were, Footscray would look similar up til the mid-40s, when men start dropping off. Instead Footscray has more men at every age.

Screen Shot 2015-08-20 at 1.33.30 pmI think the puzzle has had a lot of pieces added, but there’s still some blank spots… Any further ideas?

The curious case of poll-driven political reporting.

The Guardian published a report yesterday about Bill Shorten. The author set out to repent for calling  Bill Shorten a “tired accountant”. The impetus for the story was the turn around in the polls.

“Shorten is still leading the Labor party in the wake of this latest credibility disaster for the Coalition, after last week’s credibility disaster (blocking a free vote on marriage equality) and the preceding week’s credibility disaster (chopper-friendly Bronwyn Bishop). He’s now sitting atop polls from both Ipsos and Morgan that have the Coalition facing a loss of between 36 and 44 seats.

Is it time for a rethink?”

I’ve seen this kind of thing before, and I don’t like it.

Interpreting what a political leader does through the polls is intellectually vacuous. It’s easy to write. There is no need to have a view on tough questions about policy effectiveness or priorities, the merits of intriguing questions about whether the head of the AWU should be matey with big business, or the management and composition of their front bench.

The author of yesterday’s piece is not especially guilty. She has written about policy more than polls. But overall, allowing poll numbers to drive judgment of politicians’ merits is now commonplace. [1, 2, 3, 4, 5, 6, 7, 8, 9].

The rise of this sort of reporting means a swing in the polls does double business.

Not only does a poll bump get the leader kudos in their party, but it changes the tone of reporting about them. The new, glowing stories therefore amplify swings in popularity. That may be responsible for the increasingly binary popularity positions we see among our political leaders (They’re often wildly popular like Baird or old Abbott, or wildly unpopular, like Gillard and new Abbott).

This kind of reporting validates the paradigm that political hacks of the most cynical kind push inside their parties: We can do good once we’re in power. For now let’s focus on winning. It sidelines those inside a political party who think they should focus on making the country better, not just making the polls better.

Here’s a choice example of the kind of reporting I’m talking about.

The Sydney Morning Herald's Peter Hartcher thinks parties should use poll numbers not policy ideas to choose their leader. Is he right?
The Sydney Morning Herald’s Peter Hartcher thinks parties should use poll numbers to choose their leader. Is he right?

I can only imagine the cognitive dissonance some reporters must experience when they write articles demanding more policy substance and less poll-driven rubbish.

Of course, we do need some political reporting. It’s helpful to peek behind the curtain from time to time and see the way the magician performs his tricks. You feel like an insider.

But it can’t be all we have, most of what we have, or even a substantial minority of it. It’s a sometimes food.

Our meat and veg must be stories about policy.

Worried about having too many messaging apps? Boy do I have bad news for you.

Today my twitter app offered me the option to integrate my Instagram. I said no. Shortly afterward I started reading about a service called “Slack”.

Screen Shot 2015-08-17 at 3.25.02 pmThe line about the email killer made me laugh.

If you hope one day your email, your twitter, facebook, sms, skype, postal mail and home phone will all be in one app, you’re dreaming.

Likewise, if you invent a messaging app you intend to be the one to rule them all, you’ll end up contributing only clutter to the landscape.

This is not a bad thing.

We have specific needs for different types of communication. Each will be best served by different kinds of technology.

I think this argument is best served with an analogy.

The area of our lives where technology is most mature is the kitchen. What we see is not a streamlining of technology but hyper-specialisation.

I just counted and there are thirteen separate heat sources in my kitchen.

  • Oven
  • Grill
  • Microwave
  • Toaster
  • Kettle
  • Espresso machine
  • Sandwich press
  • Hot tap
  • Rice cooker
  • Four gas burners

There has been no major technological innovation there for years. And human tastes when it comes to food are very well-established. We can’t pretend the current state of technological proliferation is due to rapid developments or uninformed consumer prefereneces.

In theory I could use a single hot plate to get (almost) all the tasks done. But the cost of proliferation is small compared to my preference for getting the job done properly.

The microwave promised to do it all too
The microwave promised to do it all too.

For the same reason I have a cutlery drawer, not a Swiss Army knife.

The idea human communication needs can be met with a single “killer” app is crazy. We talk, we write, we draw, we sing, we shout and whisper. Even in the medium of “mail” we send not just letters but also bills and postcards.

In the future we’re going to have phones full of messaging apps, each subtly different. We’ll appreciate them all and look back with amusement on this naive era – an era where entrepreneurs who aimed to be all things to all people invented apps that ended up with a very specific purpose.

One of the leading candidates for President in the US is a socialist.

We all know about Donald Trump and his mad plans to do mad things, fired by a madness rich and deep.

Screen Shot 2015-08-14 at 11.08.23 amBut there’s someone else in the US Presidential race who is even further from the American mainstream and who could be even more influential.

Screen Shot 2015-08-14 at 11.28.39 amVermont Senator Bernie Sanders is ahead of Hillary Clinton in the poll for the New Hampshire primary – the vote to select the Democrats presidential candidate. He leads 44-37.

New Hampshire is the first primary vote held and is therefore seen as a key marker in the Presidential race. It gets a vast amount of media attention. The last seven Presidents have all come first or second in their New Hampshire primary.

Sanders is currently ranked about third for odds of winning in 2016. He’s rated by the betting agencies as having about a 7 to 8 per cent chance of becoming president. That maybe doesn’t sound like a lot, but at this stage of the process when the field is littered with contenders, it’s pretty good. Only Jeb Bush and Clinton are ahead of him.

All this matters because Sanders is not another cookie-cutter centrist. He offers real policy alternatives. He’s genuinely left-wing. (A socialist in American terms although maybe not by the criteria of the rest of the world)

For example, these are from his 12-point agenda for America, on his website:

Trade Policies that Benefit American Workers: We must end our disastrous trade policies (NAFTA, CAFTA, PNTR with China, etc.) which enable corporate America to shut down plants in this country and move to China and other low-wage countries. We need to end the race to the bottom and develop trade policies which demand that American corporations create jobs here, and not abroad.

Taking on Wall Street:  The greed, recklessness and illegal behavior of major Wall Street firms plunged this country into the worst financial crisis since the 1930s. They are too powerful to be reformed. They must be broken up.

Growing the Trade Union Movement: Union workers who are able to collectively bargain for higher wages and benefits earn substantially more than non-union workers. Today, corporate opposition to union organizing makes it extremely difficult for workers to join a union. We need legislation which makes it clear that when a majority of workers sign cards in support of a union, they can form a union.

Those are not the kind of things other Democrats are proposing.

Now, I’m not saying that Mr Sanders will beat Hillary Clinton. One poll ahead of one primary is not enough. Endorsements also matter and on that front Hillary Clinton is miles ahead.

But the way politics works is that a successful candidate with more extreme views pulls the whole field in their direction.

In US politics, the primaries see candidates tack out to the political extremes as they chase the votes of the party members who vote in primary elections. They then tack back to the middle as they chase the votes of ordinary citizens in the actual presidential election.

But the process of winning the primary forces a candidate to nail some colours to the mast. Sanders’ success will pull Clinton to the left as she tries to neutralise him. That effect will fade after she beats him, but it can’t be erased completely.

And as the two eventual candidates face off for President, trying to capture the centre, Clinton’s extra leftness will pull the eventual Republican nominee slightly leftward too.

So even if he loses the Democratic nomination, the effect of Sanders’ run will be to shift the whole of the US ever so slightly to the left.

Rising house prices: not a wealth fountain. A money-go-round.

RBA deputy governor Phillip Lowe gave a great speech last night. Lowe is the guy most likely to replace Glenn Stevens when Stevens quits as Governor and it is worth paying attention to what he says.

Last night’s speech was pretty radical. In the guise of a dry discussion of Australia’s balance sheet, Lowe single-handedly deflated arguments for rising house prices.

That puts him in direct opposition to noone other than Prime Minister Tony Abbott. Abbott, of course, said in June “I do hope our house prices are increasing.”

The argument Lowe makes is so smart and so obvious it’s amazing we don’t hear it more often. He starts out by showing that the rise in “house prices” is really a rise in land prices.

Screen Shot 2015-08-13 at 10.33.46 am“[T]he figures that I have presented invite the conclusion that our national wealth has risen largely because of higher land prices. But is such a conclusion really warranted? Have we really become wealthier as a nation simply because the value of our land has increased?

“The answer would clearly be yes if this increase was because we had discovered more land. To my knowledge, though, this has not happened.[7]”

Lowe argues that the rise in house prices is not a nice neat story about the returns to city life increasing. He says prices rose because of financial deregulation and supply constraints.

This creates not a wealth fountain but a money-go-round, he explains.

“from the perspective of society as a whole, much of what is gained on the one hand is lost on the other: there are windfall gains from higher land prices but then everyone pays more for housing services.”

Lowe also reveals that the “baby boomers are ripping off the kids” narrative has some credibility even in that palest of economic ivory towers, the RBA.

“For an older person who owns their own home and has no children, the capital gain from the higher land prices more than offsets the expected higher future housing costs. Such a household is better off. The same is true for owners of investment properties, since they own multiple dwellings on which they earn a capital gain. In contrast, for young homeowners with multiple children, the calculation can look quite different. If they care about the future housing costs of their children, then, in some circumstances, it is possible that the higher future expected housing costs could exceed the capital gain on their dwelling. In a welfare sense, the increase in land prices could make them worse off, even though they own land. The same is obviously true for renters as they do not have any capital gain to offset the higher future housing costs.”

“I think many Australians have an innate understanding of the concept and share the concern. Many parents around the country look at the high housing (really land) prices and worry that their children will not be able to afford the type of property that they themselves have been able to live in, even if their children were to have the same life-time income profile as they have had.”

“So it is arguable that the main impact of higher land prices is not really to increase our national wealth, but to change the distribution of that wealth.”

He goes on to argue that if parents help their kids buy houses, high house prices are perpetuated. But their wealth effect is diminished because the people that have expensive assets are using them as collateral for buying more expensive assets. That is to say the high prices bring no benefit.

If, however, parents don’t help their kids buy houses, and instead spend up big (say on trips overseas) then house prices are more likely to moderate.

This latter scenario, as unpleasant as it may seem to some, is actually the better one for social stability. Because with Australia’s strong immigration profile, not everyone has parents who own property in Australia. The divide between new migrants and established citizens will only grow larger if property wealth is transmitted across generations.

Who is to blame for the state of the labour market?

Last week unemployment was up. This week wages growth was down.

Screen Shot 2015-08-12 at 4.41.18 pm

Worst annual growth on record (since 1978)

These two series measure the most important and relevant determinants of Australia’s economic well-being. Both are deteriorating.

Forget interest rates. Forget house prices. Forget the dollar. Forget petrol prices and forget the share market.

How much money people make is the single biggest determinant of how well off they are. And we’re not doing well at all on that score.

This is a failure of economic policy. No government should be complacent in the face of a weak labour market.

The Government is silent on this and to its credit the opposition is squawking at them.

But it is to no avail. No decent policy is evident.

There’s a Productivity Commission report on our workplace relations policies, but nobody really thinks that will make a lick of difference, even if the government had the political capital to implement it.

These days it seems like some on the left actually relish a bit of weak wages growth. They use that to bash the government for hypocrisy over a wages breakout and guard against workplace reform.

I wouldn’t mind seeing a wages break-out. Isn’t that what good economic policy would produce? Wealth shared widely?

The failure of our labour market to do very much in the last few years probably comes down to macroeconomic factors. The high dollar crimped output and hiring. So did weak federal spending.

Screen Shot 2015-08-06 at 12.08.47 pmThe high dollar was a result of US quantitative easing and there was little more we could do beyond slashing official interest rates. That policy front was maxed out. But fiscally, we pulled puches.

Esteemed labour market economist Jeff Borland argues our failure to remedy unemployment is due to a shortage of aggregate demand.

“•The rate of unemployment in Australia has increased from 4.0 to 6.4 percent since the GFC. Over that period it has shown little tendency to decline. The rate of unemployment in the US is now lower than in Australia.
• This increase in the rate of unemployment in Australia appears to be explained entirely by the cyclical downturn in aggregate demand.

How could the government have increased aggregate demand? Spending more would have been one answer.

The Swan Budgets in 2012 and 2013 and Hockey’s efforts in 2014 and 2015 were all deficit-obsessed. All were focused on “return to surplus.” None of them achieved it. Instead unemployment has risen from 5.2 per cent to 6.3 per cent.

That deficit obsession hurts us all.

Why I’m going cold on rational charity optimising

People don’t decide to give, then choose where.

They see a problem they want to fix and try to help. Maybe that’s saving the world’s cutest fauna or trying to fix homelessness in their own suburb.

I’ve written lots about optimisation of charitable giving. So to me, that’s wasting money. But not to them.

Obsessing over worthiness of charitable donations assumes the amount of giving is fixed. It’s not. The amount of giving depends enormously on emotions. And not every charity causes emotions.

In fact, the ones that measure their impact and do the most good don’t necessarily create the biggest tidal wave of fellow-feeling. The Schistosomiasis Control Inititiative doesn’t exactly bring a tear to the eye.

If I criticise giving to the guide dogs, as I did in May, I’m applying an economic framework that doesn’t occur to the donor. They don’t want to optimise their return. They want to feel a good emotion. They don’t actually give a shit about the opportunity cost and the additional good they could be doing. They just want to feel warm and fuzzy.

Who am I to point and laugh at that? Warm fuzziness is what electrifies the philanthropic sector. Without warm fuzziness there would be no philanthropy.

Altruism is not rational in and of itself. Why should we expect it to be applied rationally?

Telling people warmth is weakness and fuzziness is foolishness is the quickest way to cut off philanthropic donations at their source. And we don’t want that.

There was a really amazing story in Vox today about “effective altruism” and how even supposedly rational giving can easily get trapped in a cul-de-sac where people give to what makes them excited. It’s a really terrific article and I recommend clicking it.

Scarcity: A book review

I just finished reading a book called Scarcity, by two US academics – a Harvard economist and a Princeton psychologist. It’s an economics book, sort of. But mostly it’s completely innovative.

It contains a theory of a new kind. One I had never thought of, but which I have found very useful.

The theory is that scarcity – no matter what its source – affects your brain. People affected by scarcity have their mental capacities focused on the problem at hand. And that diminishes their ability to deal with everything else.

The authors use four main categories of scarcity in the book.

  • material poverty:  a scarcity of money;
  • being too busy: having a scarcity of time;
  • being on a diet: facing a scarcity of permitted calories;
  • being lonely: having a scarcity of social connections.

For each category they find similar effects on mental capacities.

The person facing scarcity focuses on their own immediate problem, which gains them a limited upside in that sphere – perhaps they scrounge and borrow enough to pay this weeks rent, or remain disciplined enough to not eat dessert tonight. But the focus means other things are ignored to their detriment. They may not be focused at work, for example.

Anyone who has tried to play a computer game and hold a conversation at the same time knows performance and distraction are not compatible.

The theory that scarcity is distracting is well-backed by research. For example, simply reminding people of their financial constraints can lead to a drop in IQ of over 10 points in one famous study.

It is an appealling theory in part because it knits well with a socially progressive view of the world.

The poor, data shows, are worse at sending their kids to school, at taking medicines, at getting their forms filled in at Centrelink, at quitting smoking, eating well, etc. This is a puzzle that would appear to lend credence to a conservative viewpoint that says the poor are lazy.

This theory as expounded in the book helps explains these phenomena by reference to the circumstances of poverty, rather than by blaming the individual.

The authors created a study in which Princeton students had to play Family Feud. They were allocated to either a “rich” group with plenty of time to answer the questions, or  a “poor” group with little time to do so. The poor focused hard. They made more correct guesses per second. But the rich outperformed them overall. Then the authors offered the groups the opportunity to “borrow” time from future rounds for use in the current round. The “poor” group’s performance overall tumbled as they borrowed more and more. The scarcity mindset itself led to poor decision-making.

The book not only describes the problem of scarcity. It suggests interventions that could prove helpful.

The effect of scarcity is described in the book as a “bandwidth tax”. Which is to say that focusing on a single problem of scarcity inhibits the amount of mental bandwidth we have to deal with other scenarios.

Recognising that the “bandwidth” of the poor is especially thin permits better-designed social programs. Rather than intensive financial education programs, a few behavioural nudges will be more effective, for example. And where education is necessary, courses that are cumulative and that fail to accommodate students who miss a class will be far less effective than modules learners can take at their own pace.

They suggest  boom and bust scenarios, such as those issuing from monthly welfare payments or variable pay cheques, can create bigger bandwidth problems as people struggle to manage cashflow. The implication is that stable, frequent, predictable payments are better for bandwidth.

Homelessness – an extreme version of scarcity – is another good example. With nowhere to sleep, wash, prepare food, relax, read or store possessions, it is no wonder the homeless are rarely focused on their health, education or financial futures. Fixing the lack of accommodation in one fell swoop may be more effective than complex incentive structures (and evidence shows that is the case).

The book is excellent at describing scarcity traps, where we get behind – in payments or on a schedule – and constantly borrow from the future, all the while sinking deeper and deeper into the morass.

For those of us not stuck in grinding poverty, the book has some great examples of how to avoid being stuck in a scarcity trap. It is eloquent on the need for “slack” in order for systems to function. Much like when the fridge is literally full to bursting you can access nothing in there easily and things tend to fall out onto the floor, any system that has no slack is inefficient.

A great example comes from a hospital which was able to manage its surgery schedule far better by leaving one operating theatre empty for emergencies. The authors are also advocates of agreeing to fewer commitments than you can manage, and of leaving spare space in your diary so meetings can be moved around without creating disastrous cascades.

The seed of the book, in fact, comes when Professor Mullainathan, writing a book chapter about low-income Americans, sees a great deal of his own time management habits in the budgeting of Americans with payday loans and food stamps. The commonality of the scarcity mindset – which makes a top economist likely to snap at his kids because he’s stressed and busy – is reinforced throughout the book, all the while acknowledging poverty is a more severe, more binding problem than being over-committed at work.

The suggestion scarcity can be better managed with explicit provision for slack could have pay-offs in any number of realms – not least at a national level.

What is our national budget if not a circumstance where every last dollar is committed and deficits keep cropping up most unexpectedly? Is our Treasurer stuck in a scarcity mindset? Is his ability to think clearly about the big picture damaged by his burning desire to fix the budget in this time period? Is the Treasurer more like a 28 year old single mum trying desperately to pay the rent this week, or a 28 year old bond trader with a million dollar portfolio trying to optimise returns across their lifetime? I fear it is the former.

This book got me thinking thoughts like that. And that’s reason enough to recommend it whole-heartedly.

Unemployment. Something important has changed.

Unemployment numbers came out today and they’re bad. Unemployment is up. Again.

I was fossicking round in the data when I realised something important. There’s been a massive change. We are experiencing something unprecedented.

Screen Shot 2015-08-06 at 12.08.47 pm(These are raw numbers. No seasonal adjustment.)

In the past, unemployment has surged up fast and fallen back down slowly. That’s the pattern. You’d get rising unemployment only in bad economic shocks. Now we’re getting slowly rising unemployment.

Something has changed in the belly of the beast.

And the effect you see above is not just an artefact of a higher population making the slope up look bigger on the graph. Here’s the unemployment rate graph for comparison.

Screen Shot 2015-08-06 at 12.21.34 pmWhat’s going on?