Weekend Reading & Selected Links

Happy weekend! Here are some links to things I've been reading that you might also enjoy:

1. My new podcast conversation, with Ken Henry. At the bottom of this email, I've included six excerpts from the conversation.
2. Map of R&D gaps in science.
3. 'Biology is a Burrito', by Niko McCarty.
4. 'Growth is Getting Harder to Find, Not Ideas', new working paper by Teresa Fort et al.
5. 'Reexamining lackluster productivity growth in construction', paper by Daniel Garcia and Raven Molloy.
6. 'Jane Jacobs Can Fix American Cities, Even Though She Helped Break Them', new post by Maxwell Tabarrok.
7. Rosalind Dixon and Richard Holden on the Australian healthcare model.
8. 'Why Starship Matters', Casey Handmer in Palladium.
9. Nihilism is not enough.

Have a great weekend,‌
‌‌
‌‌
Joe

Excerpts from my podcast with Ken Henry

1. John Howard introduced the GST precisely because it was such a difficult reform

KEN HENRY: My take on – and I've not had this conversation with John Howard – but my take on why John Howard decided to go after the GST following the 1996 election, having gone into the election campaign not just with a small target strategy, but actually saying on several occasions “we will never ever have a GST, never ever, it's an ironclad guarantee”, and then within 12 months to have flipped his position on that and said “we're going on a big tax reform adventure” … And I was appointed to lead that thing, that task force. And by the way, it was one of the best policy development experiences of my professional life. It was a fantastic process. Why? And people forget. But, you know, similarly, here was a man who had crafted an identity as a frustrated reformer in the Fraser government. And I believe this to be true. I'm not saying that there's … I'm not questioning this at all – that there are a lot of big things that he wanted to do, and Fraser wouldn't let him do any of them. And reforming the financial system, floating the currency and reforming the tax system are three that I have heard him identify for himself.

And so when [Howard] was elected prime minister eventually in 1996, all those years later, there were big expectations – I mean, huge expectations, particularly from the business community – that it's all going to happen now, right? Now, of course, the dollar had been floated and the financial system had been reformed, liberalised and so on, capital controls abolished. All of that stuff had been done by Hawke and Keating. But the one that had not been done was the one you referred to, Asprey, right. That had not been done. And in fact so much political blood had been spilled on that if you ever were going to demonstrate to any audience you like that you had what it took to be a reformist leader, that was the one you had to do. And I reckon that's why he did it. And it was a near death experience for him, right? 1998 election, he lost the popular vote, and I believe that a lot of his colleagues were not happy with him for having taken such a risk. But there's no doubt in my mind that he did the right thing and he did it for the right reason. It was to demonstrate that he was up to the leadership task.

And so why do we no longer appear to have ... I mean, I think Rudd was a bit like that, right? But maybe Kevin was trying to do too many things on too many fronts. Probably … I mean, not probably, certainly. Certainly. Obviously he was. But he, it was the same thing, like he was motivated to demonstrate that he was that type of politician. He wanted that legacy. Recently we've had politicians who have said things like “I don't want to have a legacy. What do you mean, legacy? That's a vanity project.” You remember. Of course you remember. And I don't get that. I just don't get that. I think we are better off when our leaders do want to leave an impressive legacy. And that legacy is going to be based on doing the hard stuff, not the easy stuff, right?

2. On how using AGI in government will (not much) affect the speed of policymaking

HENRY: I'm not sure [AGI] would have any impact other than it may very well make politicians even more cautious than they are. And I'm not sure that's a good thing, right. Because my finding has been over a long career in the public service, the more they know, the less courageous they are, huh? Yeah. And you know, maybe that's a good thing. But … you kind of need, you need a bit of courage, you need a bit of that, not stupidity, not madness, but something people used to … I don't know if the expression is still used, but crazy brave, you know, crazy brave. You're prepared to push through on the hard stuff, you know. You want some of that...

And then the other thing I'd say to you on this is, I think it would be a mistake to think that – and I know you're not going there, but I think some people do – to think that the reason for poor policy outcomes is a lack of cognitive skills in the public service. And I can point you to a tax review published 15 years ago, of a thousand pages, that … well, you know, I mean, maybe there's an AI engine that could do a much better job of that and certainly do it in fewer pages. I don't doubt that. That would make it even less likely that it would be implemented. Not more likely, less likely. It was actually the elegance of some of the policy proposals that was their biggest flaw. It's not a lack of cognitive ability that is limiting policy development.

3. TFP: Australia could aspire to 95% of the US level.

JOSEPH WALKER: If we take total factor productivity in the US to represent the kind of technological frontier, and other countries can measure themselves against that benchmark, I think generally, Australia sits around 80% of the US level. That might have peaked a bit above 85% in the late 90s, but it came back down. How likely is it that a mix of policies exists that could help us achieve parity with the US level? Or do you think we'll always be constrained by other factors like geographic isolation from major economies, the kind of geographic fragmentation of Australia, the small size of our national market, et cetera?

HENRY: No, no, it's a really good one. So [Treasury] did some work on this in the late 1990s, asking exactly that question.

WALKER: Oh, wow.

HENRY: Yeah, well, we did, and we came up with the view that 95% is about the best we could hope for, because the other ...

WALKER: What explains the 5%?

HENRY: The last stuff that you were talking about, geographical isolation, separation, blah, blah. We figured that simply putting a rope around the Australian continent and towing it up to sit adjacent to California, that alone would lift productivity by at least 5%, right (laughter)? Yeah, just doing nothing else.

WALKER: Mainly through building all the tug boats. (laughter)

HENRY: And there is some literature on this, right … the impact of geographic location on national productivity. And that was the consensus position of the literature back in the … And look, you know, your AI assistant would be able to answer this like that right now, whereas it took us months to figure this out. But, so, but realistically, you'd have to think 95%. I would still think 95%. And who knows? The US could be falling off dramatically at the moment. And so maybe something far in excess of that is feasible. Not that that's a good outcome necessarily for the world, right? But anyway, which means that we can do a lot better. All right?

WALKER: It's exciting to know that ceiling is there, and that's what it is, and that's how much better we can do.

HENRY: Yeah. Anyway, I think that's a reasonable aspiration for policymakers in Australia.

4. The first thing Ken would fix in the tax system? Bracket creep.

WALKER: Okay, so, Ken, imagine you win the lottery. That is, imagine overnight we increase the GST to say, 20%. We're bringing in an extra, I don't know, $50 to $100 billion in revenue a year. If that happens, what would your kind of dream shopping list of dream tax changes look like?

HENRY: Well, hang on. We do have a fiscal problem to sort out first. We really do...

But if your question is, okay, I got a $50 billion surplus, let's say we get to that point, what would I do? I'd go personal income tax first.

I mean, the very first thing I would do is index personal income tax scales. When I found myself saying that recently for the very first time publicly – having argued against it for nearly 40 years – I was quite surprised at myself. But the reason, and the reason that I'd argued against it is because, you know, surely there are better things you can do, right? I mean, there must better things that you can do. But increasingly I've come to the view that the reliance that government is placing upon fiscal drag in the personal income tax system is undermining social cohesion. It's got to undermine social cohesion. I'm talking about intergenerational harmony. That's what I'm talking about …

I used to think, as a young Treasury tax policy person, that there was a cogent economic argument for preferring capital income over labour income, right? And it's, in simple terms, that capital income gets double-taxed under an income tax system. So it's an easy enough thing to talk about and to come up with a case for applying lower taxation to capital income.

But if you think of it intergenerational terms, you know, with the population bulge of the baby boomers going through and then those left to pick up what's left … this distribution of taxes across the various tax bases – labour income being the principal one, capital income being much more favourably taxed, and of course, capital gains very very favourably taxed – you can understand why young people feel that they’ve been robbed a bit. And I think we've got to deal with that, right? And I think that journey starts by indexing the personal income tax scales.

I just think it's extraordinary, for example, that in recent years, whilst the average worker has – not in the most recent years, I know it's turned around a bit now, but for many years recently – the average worker, whilst experiencing an increase in nominal wages, nevertheless experienced a reduction in real wages. And yet, because they had an increase in nominal wages, their average tax rate went up because of fiscal drag. So a reduction in your real income, and you're paying a higher rate of tax on your income.

And that is kind of – I mean, in any course I ever taught on public finance, that would have been laughed out. Nobody would tolerate that. That's complete nonsense. And yet we have tolerated that in Australia. And we've got to do something about it.

5. Should government have a plan for dealing with huge job losses from transformative AI?

HENRY: There is a very real possibility, isn't there, that the production process is no longer, or at some point is no longer, secure, or offers even an insecure form of income for most people. Let's say for the 50% [of Australian workers, who are knowledge workers] that you're talking about. Now, maybe they happen to hold shares in one of these … I don't know how many of these companies there would be that are offering these AGI services to industry and governments all around the world. There may not be many of them, right?

And I can see them making a lot of money. I can see how they can make a lot of money. But the conundrum is this. There's not going to be so many workers who are going to be in receipt of income. Those workers are not going to have the capacity on their own to buy the services that are being provided by these production machines that are heavily into AI. It's a very different structure of an economy. I mean, Say's Law would still hold, of course, that the value of production and the value of consumption broadly defined must be equal. Right? Say's Law would hold. So that's all right. But the … pool of people, yes, people who are actively engaged in the production/consumption/saving/investment space, that is – and that pretty much describes the entire economy as we currently think of it – that pool just shrinks, and shrinks, and shrinks, and shrinks.

So what, you know, people. But I know people have been wondering about this for years and years – wondering about, well, what does that mean for those who have lost their jobs? How do we get them to continue to participate? And you know, that's where the idea of these big redistributive taxes come from. And like, holy hell, do we really think Australia is going to be able to tax this stuff? I mean, we're trying, right? And with other countries, we're trying. And what's been delivered to date is really not very impressive. So it's a really important question. That's what I would say to him. And I'd say “you don't want to be remembered as the first Dave”.

WALKER: Do you think the government should be thinking about this right now?

HENRY: I hope they are. I imagine they are ... I do know that they're thinking about what forms of taxation need to be developed in order to affect that sort of income redistribution. I don't think it's because they currently fear that humans … are going to lose their jobs on the scale that you're talking about.  I don't think it's that. But it is nevertheless this realisation that more and more of the potential income tax base in particular, but also consumption tax base, in Australia is beyond reach, right?

6. Ken learned to use Microsoft Excel when building PRISMOD — and we wouldn't have got the GST without these tools

WALKER: So, first question, when you joined Treasury in 1984, you still would have had typist pools, right?

HENRY: Yeah, yeah, it's true.

WALKER: Can you describe what a typist pool is and how it works?

HENRY: (Laughs) Yeah, I mean it's extraordinary, right? And somebody of your age could not possibly have any idea unless you're watching some television show, I guess. But I mean, it is literally the case that if you produced a piece of work, a document in a documentary form, right – obviously not a piece of computer modelling or something, but a piece of advice, let's say, to go to the treasurer or to go to somebody else in the organisation – obviously it had to be typed. And not only did it have to be typed, but duplicates had to be made. And this was for filing purposes. And there were pools, I mean pools of people sitting in government departments, all government departments who spent their entire working days just clackity clackity clack on typewriters. That's what they did. Yeah, it was … I'd come from a university where the same thing was going on. The typing pool was smaller, we were a small department, economics department. But nevertheless it was the same thing. It was weird.

WALKER: And so what are your memories of when Microsoft Excel and Microsoft Word rolled out in government?

HENRY: So I've got a few rather … well, I've got a few memories of that. But actually my first introduction to spreadsheets came when I joined the Treasury at the end of 1984. Like as an academic I was. It's hard to believe, right, but in the early 1980s, when I was writing my PhD, most people were still using punch cards and carrying bundles of punch cards down to computers that have, I mean, obviously much less computer power, computing power, than your mobile phone has. I mean, much less …

And you go through this process where you'd feed the punch cards into the computer and, I don't know, an hour or two later you get some output that had come out on a piece of paper about this wide. And then you'd realise that you'd made some coding error and you get access to the computer maybe three days later to go back and fix one of those coding errors. I remember … this was in the very early days of computable general equilibrium modelling, and I was developing computable general equilibrium models not as black boxes – I shouldn't use that expression – not to inform public debate, but actually as a teaching tool for a graduate course in international trade theory that I was teaching, I thought “this is kind of a neat way of just demonstrating how all this stuff, all the algebra fits together”.

And so I built these little, just little two-sector neoclassical general equilibrium models. And it took me a month to get one of these things coded, right, through this elaborate computer process.

And I joined the Treasury in, I think, August or September 1984. And I saw somebody sitting down … In the area that I was working in, there were two desktop computers, and there were probably 70 staff. This was the only computer facilities available to the 70 staff – unless you were important enough to access the mainframe, and I knew what that was like, and I wasn't going to go into that, right? And I saw this guy sitting there and he had … it was a Lotus 1-2-3 spreadsheet. I guess most of you have never seen that thing, right? But it's the old, you know, well, it's anyway, black screen, green lines. And he showed me what it did, and I was dumbstruck.

Anyway, I said “can I have a go at that?” And I wrote on that thing in, I think it was about two hours, one of these computable general equilibrium models, right, from scratch. Two hours! Like, holy hell! Fully debugged, blah blah. I was blown away. We worked … In Treasury in those days, we worked Lotus 1-2-3 to death. And then when Excel came in, and I remember that too …  I was told that we were the first users of Microsoft Excel in Canberra.

WALKER: That would make sense.

HENRY:And that was because we had this project to develop a … This was on the instructions of the late John Kerin, who was briefly Australian treasurer. Some of you might remember that, [in] the second half of 1991 and after Keating lost the first challenge against Hawke. Anyway, he issued an instruction to the department to build a modelling capability capable of assessing any change to the indirect tax system, any change to income taxes, to a whole range of things, right down to the distributional details – so how different households of different types are going to be affected, blah blah blah blah blah. And we had six months to do it.

And that was the first time I ever encountered Excel. The IT people at Treasury said “oh, and by the way, we've got this whole new software thing that you have to learn”. And I started again, building … what economists call a price input-output model, although very few economists have used these things. They are embedded in all computable general equilibrium models, but really very few use them on their own. And I sat down and started. I got blown away by the power of this thing. And one day I … Some of you will understand what I'm talking about here. But the input-output tables in Australia [at the time had] 107 industries using outputs produced by 107 industries in intermediate usage. And a lot of the action in indirect tax changes, and therefore the price impacts of them, occurs within that intermediate usage table, believe it or not, in the input-output matrix. That's where most of the action occurs.

And so anyway, I wrote the matrix algebra and then thought “okay, I'm going to do this on Excel because it's really bloody powerful”, right? And I tried to invert – some of, you know what I mean by this, but – tried to invert a matrix that was 107 by 107. And of course it crashed. And I thought “why the hell would this crash?” So our IT people … this was the early days, right, and our IT people had a direct phone (line) to Microsoft on the west coast of the US and so they spoke to them overnight and got back to me in the morning.  And the response was “the fellow reckons you're crazy”. Like, what the hell are you trying to do? Why are you trying to use Excel to invert a matrix of that size?

And I said “well, did he explain to you why it won't work?” “Oh yeah, they put an arbitrary limit on the size of the array.” Purely arbitrary, right; it was just purely arbitrary. And I said “well, can they change it?” You know, of course they were not going to change it, right? But anyway, look, the power of that stuff was ... It was kind of mind-blowing. Well, anyway, it blew my mind.

And by the way, without that capability, I will say that there is no way that we would have got consumption tax introduced in 2000. Absolutely no way. So it can be profound, the impact of this stuff.