Sir David Hendry, the renowned British econometrician, talks to hosts Gene Tunny and Tim Hughes about the state of economic forecasting and the transition to net zero greenhouse gas emissions. Among other things, Sir David talks about how to avoid major economic forecasting failures (e.g. UK productivity), forecasting global temperatures after volcanic eruptions, and the role of nuclear energy in the net zero transition. Sir David is currently Deputy Director of the Climate Econometrics group at Oxford.
Please get in touch with any questions, comments and suggestions by emailing us at contact@economicsexplored.com or sending a voice message via https://www.speakpipe.com/economicsexplored.
You can listen to the episode via the embedded player below or via podcasting apps including Google Podcasts, Apple Podcasts, Spotify, and Stitcher.
About Sir David Hendry
Sir David F. Hendry is Deputy Director, Climate Econometrics (formerly Programme for Economic Modelling), Institute for New Economic Thinking at the Oxford Martin School and of Climate Econometrics and Senior Research Fellow, Nuffield College, Oxford University. He was previously Professor of Economics at Oxford 1982–2018, Professor of Econometrics at LSE and a Leverhulme Personal Research Professor of Economics, Oxford 1995-2000. He was Knighted in 2009; is an Honorary Vice-President and past President, Royal Economic Society; Fellow, British Academy, Royal Society of Edinburgh, Econometric Society, Academy of Social Sciences, Econometric Reviews and Journal of Econometrics; Foreign Honorary Member, American Economic Association and American Academy of Arts and Sciences; Honorary Fellow, International Institute of Forecasters and Founding Fellow, International Association for Applied Econometrics. He has received eight Honorary Doctorates, a Lifetime Achievement Award from the ESRC, and the Guy Medal in Bronze from the Royal Statistical Society. The ISI lists him as one of the world’s 200 most cited economists, he is a Thomson Reuters Citation Laureate, and has published more than 200 papers and 25 books on econometric methods, theory, modelling, and history; computing; empirical economics; and forecasting.
What’s covered in EP198
Conversation with Sir David:
- [00:02:27] Economic forecasting: are we any better at it?
- [00:05:56] Forecasting errors and adjustments.
- [00:08:04] Widespread use of flawed models.
- [00:12:45] Macroeconomics and the financial crisis.
- [00:16:30] Indicator saturation in forecasting.
- [00:21:02] AI’s relevance in forecasting.
- [00:24:23] Theory vs. data driven modeling.
- [00:28:09] Volcanic eruptions and temperature recovery.
- [00:32:26] Ice ages and climate modeling.
- [00:37:09] Carbon taxes.
- [00:40:10] Methane reduction in animal agriculture.
- [00:44:43] Small nuclear reactors: should Australia consider them?
- [00:49:08] Solar energy storage challenge.
- [00:54:00] Car as a battery.
- [00:57:01] Simplifying insurance sales process.
- [01:01:19] Climate econometrics and modeling.
Wrap up from Gene and Tim:
- [01:03:23] Central bank forecasting errors.
- [01:07:12] Breakthrough in battery technology.
- [01:11:18] Graphene and clean energy.
Links relevant to the conversation
Climate Econometrics group at Oxford:
https://www.climateeconometrics.org/
Conversation with John Atkins on philosophy and truth mentioned by Tim:
https://economicsexplored.com/2021/10/16/ep109-philosophy-and-truth/
Info on solid state batteries and graphene:
https://www.topspeed.com/toyota-745-mile-solid-state-battery/
https://theconversation.com/graphene-is-a-proven-supermaterial-but-manufacturing-the-versatile-form-of-carbon-at-usable-scales-remains-a-challenge-194238
https://hemanth-99.medium.com/graphene-and-its-applications-in-renewable-energy-sector-333d1cbb89eb
Transcript:
Sir David Hendry on economic forecasting & the net zero transition – EP198
N.B. This is a lightly edited version of a transcript originally created using the AI application otter.ai. It has also been looked over by a human, Tim Hughes from Adept Economics, to pick out the bits that otters might miss due to their tiny ears and loud splashing. It may not be 100 percent accurate, but should be pretty close. If you’d like to quote from it, please check the quoted segment in the recording.
Gene Tunny 00:06
Welcome to the Economics Explored podcast, a frank and fearless exploration of important economic issues. I’m your host Gene Tunny. I’m a professional economist and former Australian Treasury official. The aim of this show is to help you better understand the big economic issues affecting all our lives. We do this by considering the theory, evidence and by hearing a wide range of views. I’m delighted that you can join me for this episode. Please check out the show notes for relevant information. Now on to the show.
Hello, thanks for tuning into the show. In this episode, Tim Hughes and I chat with the legendary British econometrician, Sir David Hendry. We talk with Sir David about the state of economic forecasting, and about the transition to net zero greenhouse gas emissions. Sir David Hendry is co director of Climate Econometrics and Senior Research Fellow at Nuffield College, Oxford. Previously, he was Professor of Economics at Oxford, and before that he was Professor of Econometrics at the London School of Economics. After the interview with Sir David, Tim and I go over our main takeaways from the conversation. Okay, let’s get into the episode. I hope you enjoy our conversation with Sir David Hendry.
Gene Tunny 01:26
Sir David Hendry, welcome to the programme.
David Hendry 01:31
Thank you very much, Gene. Thanks for inviting me.
Gene Tunny 01:34
Oh, of course. It’s a pleasure to have you on to talk about forecasting. So forecasting’s something that Tim and I have been thinking a lot about. And we’ve chatted with Warren Hatch who’s a super forecaster with I’ve also spoken with John Kay, about radical uncertainty and how you deal with that. And I’ve also read your book on forecasting, the one with Jennifer Castle, and Michael Clements, and I thought that was very good. And who better to, to have on to talk about forecasting than someone who has really transformed forecasting and economics, someone who’s had a major impact on forecasting? So to begin with, David, I’d like to ask, how has economic forecasting developed over your career? To what extent has it improved? To what extent are there still areas for improvement? Could you talk to us about that, please?
David Hendry 02:34
So Gene, I don’t think it has improved. I think technology has but the actual practice hasn’t. The time that I got really interested in forecasting was acting for the select committee of parliament that was looking into economic forecasting, after the debacle of Nigel Lawson’s budget and then crashing the economy in the early 1990s. And what I discovered, and acting for them as an advisor, is that 90% of the evidence he got was people actually forecasting and only 10% was looking at how you should forecast, what should you do, what goes wrong when you forecast with no analysis at all? So we started a long programme of analysing what can go wrong in forecasting and why. And once you know that, what can you do about it? Well, obviously, there’s nothing you can do about things that are unpredictable. Right, so the pandemic, unpredictable, forecasters shouldn’t kick themselves because you’ve got it completely wrong forecasting December 2019, for 2020, to discover that it’s vastly different. I mean, the biggest ever fall in GDP in Britain, you couldn’t possibly have forecasted that, that’s not a problem. And we can’t do that, it’s you can start to improve the forecast as you go through 2020. and realise that things are going badly wrong, but you can’t forecast in advance. So we isolated two key features that go wrong in forecasting. One is unpredictable events like that, that shift the data. So data is going along, and then either shifts up sharply, like inflation, or shifts down sharply, like output. But once it has changed, you can do a great deal about it. Some methods now don’t work. And some methods do work. And the methods that don’t work are the methods that stick to what went on before. So they carry on at the same level. And that’s completely wrong relative to the new level. So you have to have very adaptable methods that jump as soon as the forecast has gone badly wrong. You use methods to try to adjust for that. We call them robust methods. Right? So they’re after the shock to GDP. They’re robust. So the Office of Budget Responsibility in Britain, forecasts productivity per decade, completely wrongly, every year, they were wrong for 10 years, if they’d used our methods of adapting because productivity had been growing at about 1.7% per annum up to 2012, and suddenly it stopped, we don’t know why it stopped. But it’s come back to the levels that we had in the 19th century. Point seven. But if you keep forecasting 1.7, we just get massively wrong forecasts all the time, very bad advice for governments. And our methods would have adjusted to that within a year, saying, Okay, it’s changed, it may change back. But meantime, you better forecast along this direction. So the actual, if you like, the forecast errors that people make today are very similar in size to the kind that were being made in the 1960s.
Gene Tunny 05:56
Right. Yeah, that’s a that’s a shame. I should I forgot to introduce Tim. Tim, do you have any questions for Sir David on that?
Tim Hughes 06:04
No, it’s, it’s interesting. I mean, this isn’t my level of expertise. I’m here as the layman in this partnership with Gene. So I tend to look at things from a macro view and more from a guy on the street sort of perspective. But I’m really interested in that when you say that, well, for instance, it hasn’t changed since since the 60s. What’s the delay in the take up of these modelling systems for government?
David Hendry 06:27
Well, one of the reasons it hasn’t changed is that the frequency of large, unpredictable events hasn’t changed. And they’re very common and much more common than people realise, except to see the pandemic has been, Oh, quite unusual. Of course, we’ve had lots of pandemics, some of them happened like SARS too, not to go anywhere. Others like the COVID have gone everywhere. Inflation in Britain in the 1970s. It’s very similar to what it is today. And for very similar reasons. Now, I think a lot of forecasting that you hear about comes from central banks. And that’s the kind of forecasts we can analyse because they’re made to publish it. We don’t see the forecasts within many major institutions like JP Morgan, or Citibank, or whatever, they tend to keep them to themselves unless they do really well, in which case, they tell you oh, we were doing really well. But when you look at Central Banks, say we take the Bank of England as a paradigm, their model collapsed with the financial crisis, it just fell apart, and they said it fell apart. So we started to build a new one, we pointed out to them why it had fallen apart. They’re using a method of mathematical analysis that works fine if things don’t change, but becomes like navigating around the globe using Euclidean geometry when things do change, that just, it just doesn’t apply. And its widespread use has been a disaster in my view, for macroeconomics, and is the reason so much of it has gone wrong, because it assumes that the method that these models are built on assumes there are no sudden, unexpected large changes. Whenever they occur, the models fall apart. And we had a letter recently in the Times saying the bank should try testing their models from the 1970s. And they would find it’s a shambles. It doesn’t work at all. Because the 1970s in Britain was filled with crises, 3 day weeks, IMF coming in, interest at 25%, inflation, etc. And their model just wouldn’t cope with that. And we’re now in is not quite such a bad situation, but we’re now in a similar sort of situation where a wage price spiral is kicking in, these models don’t have wage price spirals. They didn’t allow for the fact that people had saved a great deal during the pandemic, because they couldn’t spend it wasn’t, it was forced saving if you like, and as soon as the pandemic ended they started spending, the supply side had improved to meet this high level of expenditure. So of course, you have all these factors coming and they’re not in their model. So naturally, the model was A they said inflation wouldn’t go up and B they said when it did go up it would be transitory, whereas we were saying, it will go up and it will not be transitory, it will be very persistent and very hard to dampen down.
Gene Tunny 09:24
Right. So this is a letter in the Times I’ll have to have a look for that. That sounds interesting. And it’s a bit of a concern that the Bank of England hasn’t improved its, it doesn’t sound like it’s improved its models very much at all, because in 2010, so you gave a talk to the Institute for New Economic Thinking, and you were talking about the problems with the models that central banks were using. And this was in your conclusion, you said that “there are huge costs to underspecified models and I think the financial crisis is partly due to central banks having very badly under-specified models in their repertoire.” Would you be able to explain what what you meant by that? Is that what you’re talking about here? They’re not allowing for structural breaks. But are there also are there variables they’re not including? Could you just unpack that a bit, please?
David Hendry 10:16
Yeah, there are variables they’re not including and often including variables in the wrong way. So for example, the Bank of England includes wealth. Now some wealth is expendable, like your house, some wealth is potentially spendable like money invested in stock markets and bond markets. In some it’s very spendable, which we call cash, deposits and demand at financial institutions. And it makes a huge difference, to break these up, because wealth itself can change a lot but it doesn’t change expenditure because house prices go up, or house prices go down. But it can also change a little bit and hugely changes expenditures because people run out of money, they have to start borrowing, and they haven’t got time to sell their house or the bond markets in disarray. And financial markets have fallen hugely, and you don’t want to make big losses. So you need to think very carefully about how you include variables in models, as well as which variables to include in models. I was referring to the fact that the housing models in the US when the financial crisis started, were very weak, they didn’t cover all the aspects that that matter, because in some States, if your house price falls greatly, and leads to a large indebtedness, if it was sold, you can just hand back your keys and walk away. You can’t do that in other States. And the subprime crisis generated articles, even from central banks, saying that it’s really important to get poor people onto the housing market, because that’s where how you build that wealth, of course that led to all sorts of speculation, and then house prices crashed. And that’s poor people who end up suffering most and we got a very bad financial crisis. But you guys didn’t have it. Right Australia avoided it, because it hadn’t got engaged in quite such nebulous activities as the AAA assets that were worth nothing.
Gene Tunny 12:16
Yeah, yeah, we avoided it. I mean, partly because of mining. And then the Treasury and the government here, they would say that they had a timely fiscal policy response. I mean, there’s debate about the extent to which that was relevant. But yeah, we were we were lucky. And maybe we hadn’t had as much crazy financial activity as in the States and Britain. We’ve got our regulators too. So yeah, a variety of reasons. But yeah, that’s, it’s fascinating.
David Hendry 12:46
I was gonna say, the way macro economics is taught in almost all major universities around the world still relies on this approach of believing agents optimise across time into the future. And you can’t do that in a world in which you suddenly get big shifts, right? You’re what looked optimal one day becomes a disaster the next, for example, Royal Bank of Scotland trying to buy this Dutch Bank looked optimal to them in the state of the world before the financial crisis and did become an absolute massive disaster after it. And that isn’t something that’s taught in macroeconomics courses that I know off.
Gene Tunny 13:29
Yeah. Yeah, unfortunately, a lot of the macros become very mathematical. And you’ve got all of these forward looking models, these Ramsey type models, and yeah, but I wonder about the just how applicable, they are. So good point. Can I ask you about your methodology David? So you’re famous for having promoted this general to specific methodology, if I’m getting that right. Could you just explain roughly what that is and how its implemented and what the modern implementation of it is? I mean, you’ve got this automatic forecasting system. Could you tell us a bit about that, please.
David Hendry 14:11
The whole idea started in the 1970s, when it was quite clear that the then big models in the US and Britain didn’t really incorporate enough information. And if any, if you leave a variable out of a model that matters, say you didn’t include housing in a macro model, and suddenly you get a big change in house prices, the model will go wrong, because it should be, housing should be in the model, and it’s not there, and it shifts and that then shifts the reality relative to the model. So it became clear you needed to think very hard about all the things that might matter. And that then required you to put statistical method that could discriminate between what does matter, and what you thought might matter but does not matter. And so we had this paper in the mid 70s, on the consumption function in Britain, showing that you could explain everybody else’s consumption function failures by a more general consumption function that pointed out why they went wrong. And that led us to develop this general to specific as a very general approach. Now, it evolved greatly in terms of, as we realise, more and more the importance of shifts and outliers in forecasting, we began to develop these methods, which at first, I have to say were greeted with not scepticism but total disbelief that you could do it. So that to take the basic idea. Say you’ve got a relatively short time series that’s got 10 observations. And you think that within those 10, there might be a discrepant observation, somebody wrote down 10, when he meant one, right? You just fit the model to it, or it’ll go very badly wrong. So what we do is we create an indicator variable for every observation. So it’s one for that observation and zero elsewhere. So you get 10 of them. And you put them in in big blocks, say five, and then the other five, and they won’t do anything, if there is no shift, but they’ll pick up the shift when it happens. And we call this indicator saturation, because you put in as many of these indicators as observations. Now why would anyone think of doing that? Well, it was serendipitous. I was asked to participate in an experiment in econometrics, to model food demand in the United States, from 1929, which isn’t a great date to start, any time see, through to 1986. And I looked at what everybody had done, and they had all thrown away the data before 1946, they couldn’t model it. So I built a model of it and looked what had gone wrong in the interwar period, and discovered there were two gigantic outliers in I think, 1932 and 33 but don’t guarantee that it could have been, but round about that period. And Mary Morgan kindly went to the archives and discovered, guess what, the US had a food programme? Well, will a food programme affect the demand for food, you bet it will. So I put in indicators for those observations and immediately got a very good model for the whole period, for the period up to 1946. So then I thought, right, let’s fit the cost period, including the early one. But we’ll put in indicators for all the observations, which is the kind of forecast test and found the Korean War I think had one big outlier, but otherwise, it was fine. And then about a year later, thought that’s funny, I had put in indicators for every observation. All the ones for the pre war period and all the ones for the post war period. And it had worked, I got the best model of anybody. So I started talking to Soren Johansen, a famous econometrician statistician, he said, “You’re nuts. You can’t do that!” And about a month later, he emailed to say, “Yeah, I think you can do that and I think I know how to analyse it,” which because if you don’t analyse it in economics, they just ignore it. And so we published several papers showing detailed analytics of why it would work for impulses, we then extended that to steps and then trends. So we can pick up trend breaks, step breaks etc. So for our 10 observations, we might end up with 40 variables. Most statisticians look at you, you’re nuts. But actually, you can show it will work. Because if there’s no break, no trend, they’ll all disappear. If there’s no step shift, they’ll all disappear. There’s only one outlier, you’ll be left with one outlier. And that’s it. So that’s how we do general to specific now. And that’s why you need automatic modelling. Because a human can’t do that. The number of possible things is far, far too big. The computer programme can of course, do it in seconds, at worst, maybe minutes if it’s a huge data set, because it’s got many, many things to look across.
Tim Hughes 19:27
Actually, this probably feeds into one of my questions for you, David, which was, you mentioned about the modelling and the mathematics, and the current uptick in artificial intelligence, in AI, is that something that has made a big difference with the work that you do?
David Hendry 19:45
Now our programme is a sort of AI programme date back a long way. Because it’s experimenting with everything. It’s a programme that’s designed for data that keep changing. Most AI programmes are not. Most AI, it takes all the cases and trains the computer to identify things in those cases. But if the cases suddenly change, that’s not going to work. And so AI has itself, the way people have used, it has not made a big impact on forecasting yet. They have to adapt AI to learn from the data, and be ready for it to be adaptable into the future in a way that if you were trying to teach a programme to identify measles, you probably would just take all the cases of measles and the programme would be able eventually to look at the spots and say, Yeah, that’s measles. But if Measles can suddenly evolve, as say the pandemic did, what you’re trying to pick up by AI would no longer be relevant. It would look different, and AI would misclassify. So AI has got to be hugely changed to be relevant for forecasting, which is about a changing world. We’ve got climate change, we’ve got pandemics, we’ve got wars, we’ve got crises, we’ve got inflation, we’ve got changing population levels, etc, etc. Unless it can adapt to that it won’t be useful in forecasting.
Tim Hughes 21:15
In your view, do you think that that is quite likely that AI will get to the point where it will be more predictive and not just reactive?
David Hendry 21:22
Well, we’ve shown you can do it, ours is very simple AI, it’s nothing like the kind of complicated neural networks that are being used in some areas. But it does show that you can do it for forecasting, and it does matter. And in the M4 Forecasting competition, which was run from Melbourne, the AI ones or machine learning, as they were then called, did not do terribly well. We came seventh in our very simple one. And and it turned out that we spent about a 50th of the time that most of the other teams did.
Gene Tunny 21:57
Was this of a motorway was it was at the M4 motorway?
David Hendry 22:01
no no. M for Makridakis fourth forecasting competition. The M four we’re now at five. It’s currently ongoing. Makridakis is a Greek forecaster, who decided the only way to improve forecasting is to find what worked. So he asked people, here’s 1000 time series, we’re not going to tell you what they are, model them, and send us your forecasts for the next 10, 20, 30 observations. we’ll analyse those and see who did best. So at the M4 there was 100,000 time series to model. And you then have to forecast I think, up to 20 years ahead for some of them. And you’ve to send in all your forecasts. And they then worked out who did best and got closest to the actual numbers in the future. Actually Uber did, Uber won the competition, Uber, yeah, the car hire people got algorithms of the kind that could be applied to forecasting. But what they did, we think was accidentally wrong, that they looked across, say, nobody knew what the time series were. But it does turn out that some of them were, say, GNP from 1950 to 1980, and somewhere from 1990 to 2010. Right. Now, they looked across, do some series help us in forecasting other series. And we think they actually included the future of the series they were to forecast in the, seeing if these series helped it, which is why they forecast much better, because we’ve mimicked their method, when all the series are completely independent, and it doesn’t help. So they had to be doing something like that accidentally, I don’t think they realise that, some of the series where the future of others of the series…
Gene Tunny 24:00
Okay, yeah. Can I ask you a question that’s related to that? It just reminded me, because you were saying that they don’t tell you what the data series are. Now. There’s this debate about, well, to what extent do you use theory and you’re modelling, you know, theory driven versus data driven, is it the case that you can get a reasonably good forecast without any theory whatsoever or without any understanding of what the underlying what the data are actually measuring? Or do you need theory? How do you think about the role of theory in your modelling? David?
David Hendry 24:33
Well, when we were forecasting week ahead Covid deaths and cases in the UK, the model only used the past data. And for the first six weeks we were by far the most accurate forecasters relative to epidemiologists with their big models and taking account of whether, you met people who had it and all the rest of it, and that’s because their models needed about 10, 12 weeks of data before they even began to be useful, whereas we could forecast immediately without any theory. I mean, I understand the big models and why they work, but we thought you can’t use that. And it’s because the way COVID hit, it did big jumps, measured on very few cases. And suddenly, like Bergamo, you had 50 people dying in a day, right? And so you get these big jumps and our methods adapt rapidly. So in that area, you could do extremely good modelling without theory. But when it comes to economics, how many variables are there? 5 trillion, possibly in the economy, if you think of everything that’s going on, so you have to have some theories and say, well, most of those don’t matter. We just can’t deal with that. So we use a lot of theory in our models, but we embed it in the general. So say you have a theory, let’s take a very simple theory that only income causes consumption, consumers spend their income and that’s it. So consumption is related to income, period. Okay, we keep that and embed that within a model in which things like well, maybe interest rates matter, maybe wealth, maybe liquidity, maybe, etc, etc, etc, matter. But when you’re searching, you don’t search over the relationship between consumption and income, you always keep it there. And if it doesn’t matter, then it will turn out to have a very small coefficient, and you can decide to drop it in the end. But if it does matter, and it’s the only thing that matters actually our method will give you the same answer as your theory model. So we embed theory in such a way that if it’s correct, that’s what you will get. And if it’s wrong, you’ll get a better model. So it’s both theory driven and data driven.
Gene Tunny 26:53
Okay, we’ll take a short break here for a word from our sponsor.
Female speaker 26:58
If you need to crunch the numbers, then get in touch with Adept Economics. We offer you frank and fearless economic analysis and advice, we can help you with funding submissions, cost benefit analysis studies, and economic modelling of all sorts. Our head office is in Brisbane, Australia, but we work all over the world. You can get in touch via our website, http://www.adepteconomics.com.au. We’d love to hear from you.
Gene Tunny 27:27
Now back to the show.
And so you start off with a very general specification, lots of data in your database, lots of variables, if I’m getting this right, and then you allow for the potential potentially all of these structural breaks things where things go a bit crazy, you jump up to a new level, like during the pandemic or, or whenever, and we we dropped down from the trend growth path we were on maybe we were cycling around it and suddenly we’re somewhere we’re in this hole and so you’ve got models that can adjust for that sort of thing is that if a fair way of thinking about it,
David Hendry 28:05
Yes, it’s post, yeah. Yes. Once it’s happened, then the model will pick it up. So quite a good example that might intrigue you is finding out where all the volcanic eruptions occurred over the last 1000 years. And when the when one of my colleagues gave a paper in our methods at our General Environmental Conference, all the volcanologists were intrigued and asked us can you use these adaptive methods to show where volcanoes were and measure how they work? Well, the answer is yes, we adapt our methods instead of them being one zero zero. They’re kind of like a V. Because when a volcano erupts, the temperature drops immediately. But it then recovers roughly half a half a half a half of what’s left. So the V shape picks that up. So we found all the volcanoes from 1200 AD in the data set of tree ring growth dendrochronology. And the key thing about that Gene is as soon as you’ve got the first observation of the volcano erupting and the drop in temperature, you can very accurately forecast all the remaining observations to the recovery, by having this V shaped go half a half a half. And we showed we could forecast, okay, forecasting after a volcano in 1650 isn’t all that interesting today, but it tells you that the next time we get a world explosive forecast like Tambora or Krakatoa, we will be able to tell the world after it’s stopped erupting, how quickly the temperature will recover to the previous level. It also lets us adjust the so called baseline temperatures that IPCC use. That in fact have been several quite big volcanoes that have dropped the temperature a little bit for a few years, and that actually means that they’re cheating by using a slightly improper average over the periods they’ve picked, as they shouldn’t include the volcanic eruption, right? Because that’s when you should use the natural level that had been over that period overall, if that makes sense.
Gene Tunny 30:22
Yeah, that’s fascinating. And so your, so is that an application of your method? Or are you, the point you’re making about the volcanoes there? Or are you saying that you can apply some theory to get a better forecast? I’m just trying to understand
David Hendry 30:36
It’s our methods purely. And it’s just the knowledge that when volcanoes erupt, the temperature falls, but it goes back again. So the question is, what form do you use for that? We just invented one that says V shape, and then we put in a V for so, it’s just over, I think we had about 900 observations. So we’ve put in 900 of these Vs in big blocks, but it only picks up a significant g if there was a Volcanic Volcanic eruption, right, because otherwise, it doesn’t help fit the model. So it then just picks up all the volcanic eruptions, and the volcanologists started using this method, we’ve done one to get a new archive of volcanic eruptions since zero, like 2000 years ago,
Tim Hughes 31:24
Actually this probably leads us on, do you have anymore questions Gene?
Gene Tunny 31:30
No not at the moment, go ahead Tim.
Tim Hughes 31:33
David, I was gonna ask you about climate econometrics. So you’ve written a book on that with Dr. Jennifer Castle. So I was interested to see exactly what climate econometrics is, and how it might be able to help us tackle climate change.
David Hendry 31:44
Yes, climate change is caused by our economic behaviour. All our methods were developed for modelling economics. So it would be quite unsurprising that they would work from modelling climate change, which is due to economic behaviour, CO2 emissions, and nitrous oxide emissions, the way we travel, the way we live, the way we eat, the way we warm our houses, etc. All these things are economic decisions. And so if the methods work for the economic behaviour, they’ll work for explaining climate, they’ll actually also work for claiming, explaining things like ice ages, even though there’s no humans around then, because they, the kind of dynamics of ice ages how the amount of co2 in the atmosphere, the amount of past sunlight falling on the Earth, that’s created the temperature, the amount of ice that’s around, etc. All of these carry forward into the future and there’s really good data on ice ages, I mean, 800,000 years of pretty accurate data and how it evolved. And we can fit our models to that, again, very general. Now, why would you want to put in indicators? Well, of course, there’s often a lot of dust in the atmosphere. And dust falling on ice turns it black, which turns up the amount of heat that absorbs. So if you have a period of massive volcanism, which does occur, I mean, often you can have 50 years of vulcanism puts up so much dust, it actually changes the pattern, and you can pick that up, and the sudden jumps in temperature that were unexpected, for example. So it can be applied to all these issues. We’ve been applying it to modelling ‘How well is the UK doing in getting to net zero?’ Now we were at a particular point that we had very good data on all the ingredients that lead to CO2 emissions, the amount of coal, which was huge in the 19th century and up to about 1970 was pretty large in Britain, but then began to drop dramatically, because it became inefficient relative to other sources, but also because it was banned in household fires. When you were not allowed to have fires based in smoky coal because smoke, so you get the demand for coal falling, and that led to the discovery of natural gas in the North Sea. Prior to that the gas system was coal gas, which required you to burn coal to get the gas but it’s very inefficient so that got rid of coal and natural gas is much more efficient. And oil was throughout beginning to replace the use of coal in many industries particular. And then in 2008, the government banned it from being used to produce electricity. And that’s the death knell for coal in Britain’s there’s almost none used nowadays. Now, 2008 is something The Climate Change Act of 2008 amazes many people, both parties unanimously voted for it as did the Lib Dems is completely we need this, let’s do it. And you get a huge, very rapid drop in the amount of CO2 emissions in the UK. Now Britain’s been moving towards a service economy from a manufacturing one. But it hasn’t been doing that to get rid of CO2. It’s been doing that because World Trade Organisation rules meant you couldn’t put extra taxes on people who are cheating in the way they were pricing their products. And so they killed off a lot of British industry. So I don’t accept that the offshoring has anything to do with climate change and claim that our domestic reductions. So Britain’s come down from 12 tonnes per person per year to four and a half tonnes per person per year over that period, which is a very dramatic reduction. America is still at 15. So it’s still above the highest it ever was in Britain. And one of the explanations we came across recently is that in Britain, cars went about 20 miles to the gallon in 1920. Now on average, they’re going 55. In America, they went about 20 to the gallon. And now they’re going about 20 to the gallon. And there’s many, many more cars, and they’re driving much further. So they’re consuming vastly more oil, and therefore gasoline. And therefore pumping out much more CO2, nitrous oxide, particulate matters, etc. They’ve had no efficiency gain, whatever, because they’ve gone for these bigger SUVs, much heavier, much bigger engines and petrol, gasoline has never been taxed in the US, whereas in Britain, the tax is about two thirds of the price of a pump.
Tim Hughes 36:54
Yeah, it’s expensive. Yeah, it’s a lot.
David Hendry 36:57
Yeah, it hasn’t discouraged people from driving. Right, people are still driving, there more and more kilometres on aggregate in Britain driven every year despite these high taxes and gasoline is one of my reasons for believing that carbon taxes will not by themselves solve the climate change problem. We need technology we need to adapt until we’ve written several papers, proposing a system of what we call five sensitive intervention points. That can be used to exploit how people behave without trying to change their behaviour, but to make them do things that will then be climate optimal. So for example, cars in Britain last nine to ten years on average, and then become obsolescent. So instead of buying another internal combustion engine car, price electric cars so that they automatically move over and buy an electric car. And if we did that, over the next 30 years, we’d end up with every car being electric, and nobody having suffered and have got the new car that they wanted at each point in time, but switching over gradually. But that requires you to be providing more electricity all the time to meet this, which requires upgrading the grid and installing more wind farms or solar cells, and maybe more small nuclear reactors and perhaps investing more in fusion in the hopes that the current breakthroughs can be made useful for society before 2050, and so on. And the paper tries to spell out how all these steps interact all the way down, clean right down to farming, how we get rid of the massive amounts of nitrous oxide, methane and even CO2 to come out to farming. That’s a huge concern in New Zealand, your neighbouring country, poor farmers, they’re objecting to fart tax. I don’t blame them. I mean, so how can they deal with it? Right? It’s, it’s not like you can deal with the tax when cars were getting more efficient when they’re driving less or getting an electric car. They need to think of the technology that will reduce methane emissions from animals. I don’t know if you know that there’s an island off Orkney called North Rolandsay, where the sheep are not allowed off the Shore, there’s a wall around the island and all the sheep are kept on the shore, and they eat seaweed. But methane…
Tim Hughes 39:25
Yeah, I heard about this recently. And because I was going to say I agree with what you say about technology, making these changes. So you know, rather than forcing people’s habits to change or you know, doing something drastic with our food chain, etc, the technology will contribute towards those changes. And yes, I saw that the seaweed or additives made from the seaweed could be one of the solutions for for methane. So just by adding it to the food. Obviously, it’s early days to see if that may or may not work on scale. But it’s encouraging It is encouraging to see those breakthroughs.
David Hendry 40:02
I think the breakthrough that’s needed is to synthesise the chemical. that does it. Because I don’t think you can grow enough seaweed to feed all the world’s cattle, sheep, goats, etc. I think that’s not on. But knowing that asparagopsis taxiformis, which is the one that seems to be best for stopping the thermogenic reactions within animals, it could be synthesised in the way that aspirin was taken from willow trees, and then Bayer worked out how to synthesise it. And I think these these things are possible. So yeah, I mean, our paper suggests that all of it is possible. Some of it needs subsidies, I don’t think tax is the right way to do it. Because we saw the uprising in France from the yellow vests. And that’s happened in Sweden, people object to their lifestyle being disturbed. This doesn’t disturb their lifestyle. It just says, oh, you know, you’d be better off if you do this. And then you can keep manufacturing going making cars but electric cars and wind turbines and solar cells and heat pumps and so on. All of it’s out there. And the thing that I do emphasise when I meet sceptics is by the end of the 19th century, we had cars that were electric with rechargeable batteries that could go up to 50 miles between recharges. We had solar cells, on roofs, we had wind turbines that were being used on farms, we knew that climate change was caused by excess CO2. And everything was in place for an all electric society, we knew how to generate it from hydro power, from wind power from solar power. But then the Americans discovered oil and the internal combustion engine. And that
Tim Hughes 41:54
So that technology was there at the end of the 19th century. You’re saying?
David Hendry 41:57
At the end of the 19th century, all of it was there. And we trace who invented it, how they invented it, how it developed? Yep, it was all there. Not LED lighting, that’s an important, more recent development.
Gene Tunny 42:12
Yes. Can I ask about that? That paper? I’ll have to look it up. It sounds fascinating. So have you you’re you’ve done modelling, have you of this path to net zero for Britain? Is that what you’re saying?
David Hendry 42:20
That’s what we’re saying yes
Gene Tunny 42:24
Okay. And yeah, it’s feasible. If there’s this technology, some technology shifts, technological improvements, but also that there may need to be some subsidies for electric vehicles, I think, was that what you were…
David Hendry 42:37
For the electric vehicles, but also for the grid. You need a massively improved grid, both because there’s vastly more electricity, but it’s got to be more resilient to climate change, because climate change is going to happen. Irrespective, even if we managed to reduce everywhere, it’s still going to carry on for a long time, because the oceans and the air have got to calibrate the temperature. And that takes a long, long time to happen. So sea level rise will continue, the Earth will continue to warm but at a slower and slower rate, if we stop pumping out quite as much CO2. And obviously, if we can find ways of extracting it, to research that, that would help. One of the things that does extract it, believe it or not, is basalt. Stuff that volcanos erupt, right? Now, if you look at photographs of volcanoes, the land around them is very fertile. So you can actually replace artificial fertilisers by ground up basalt. And that will act as a fertiliser, because it’s got all the minerals in it, but it also absorbs CO2. So it actually helps reduce CO2 whereas artificial fertilisers in making it they produce CO2, they produce nitrous oxide, etc, etc. So one of our proposals is that we start switching quite rapidly to using ground up basalt which costs next to nothing. There’s 300,000 cubic kilometres of basalt in India. That’ll take a long time to use that up.
Gene Tunny 44:12
Right, I’ll definitely have to check this out. I mean, this is a big issue for Australia. How do we get to net zero? And I mean, Britain’s probably got some advantages over us, you, you don’t have as big an area. I mean, we’re gonna have to build all of this transmission to connect up the renewable energy. Like we don’t have nuclear energy here and the Opposition party is trying to push it, but then I think there’s going to be a lot of community resistance to that here in Australia.
David Hendry 44:37
Yeah, I can believe that. But do people understand small nuclear reactors? That’s the only ones we’re arguing for, not the big ones, the small ones. In Britain, lots of big ones. And they’ve produced a lot of transuranic waste, that’s going to be a huge problem for humanity. Now, there are two advantages to small nuclear reactors. One they can use that transuranic waste as their fuel and greatly reduce the amount of radioactivity that needs to be dealt with from it. And secondly, they’ve been used in nuclear submarines for 50 years, and there’s never been an accident. So they’re very safe and they don’t have any fissionable material that terrorists might want for bombs. I mean, the stuff they’re using is useless. Other than burning up the waste, it’s a problem anyway. So if the public knew that these are harmless, that they’re getting rid of a problem, you don’t have nuclear reactors, so it’s less of an argument there. But in Britain, people would jump at the chance to cut the amount of nuclear waste that needs to be disposed of, burying it or put it in deep caves, etc. And these guys can do it
Gene Tunny 45:52
Right Yeah. These are the small modular reactors, are they?
David Hendry 45:56
Yes,
Gene Tunny 45:58
I think that’s what Peter Dutton, who’s the Opposition leader here, what he’s talking about,
David Hendry 46:02
Oh good for him, I think they are actually an important component, but only one possible component of an electricity provision, that would give more energy security. And, and be something that can work in almost all circumstances.
Tim Hughes 46:18
This is an area that we’ve talked about a few times, and one of the things that comes up is that the most likely scenario would be to have a suite of different options as to where they get the power from. So for instance, we’re very lucky here in Australia, we have abundant supply of sunshine. And so that’s clearly one of the options open to us, which we currently use, and it will grow. But there’s also hydro, there’s wind, there’s other options and having the various different things available. So that for instance, I mean, I know in the UK, for instance, like to rely on solar isn’t something that you’d want to rely on fully. So it would be the same everywhere I imagine and that that those suite of options or those suites of available power supplies would be different around the world. But it does seem to be that a lot of this is driven by the market, which we’ve noticed here and it has come up in conversation, which is that’s that seems to have been a big change, where that it’s been widely accepted that climate change is real, and that most people do want to have clean oceans, clean atmosphere, clean fuel. And so that driving force from the market, seems to be also then instigating the technology from the suppliers of those options, you know, people like Elon Musk, or, you know, these, these people who can make things happen very quickly, much more quickly than governments can. So it seems to be accelerating and going in the right direction. And so the net zero target is 2050. I think, is that right for the UK?
David Hendry 47:51
That’s right, it’s too far in the future. But we’ve picked it because the costs of adjusting to it are near zero, and probably even positive benefits from doing it slowly, in terms of machinery running out, cars getting obsolescent, trucks needing replaced. Developments, I mean, in solar cells, for example, Perovskite cells are now able to produce 30% of the energy from the sun as against the standard solar cells 22. That’s an enormous improvement. And that technology will take a little while to get commercialised and applied. And then people will have much, much more efficient solar cells to put on their roofs.
Tim Hughes 48:32
And the infrastructure needed for electric vehicles is obviously going to be enormous, especially in the built up areas. So it’s going to take some time for it all to happen.
David Hendry 48:44
Absolutely, but if the market prices correctly, it can be profitable for them to instal all the connections, it doesn’t necessarily cost much in the same ways they built filling stations. I mean, they didn’t build them for fun. They built them to make a profit to build these guys to make a profit as well.
Gene Tunny 49:03
Yeah, there’s some big issues here. Tim, one thing I would say on the solar I mean, even though we’ve got abundant sunshine, the challenge is, it we’ve got to store that solar energy for when it’s actually because yeah, that’s one of the problems because you don’t have it at night and yet there’s a big peak in demand when everyone gets home from work. And yeah, that’s why we’re having to build hydro where the State government’s here is investing heavily in hydro and trying to progress some couple of hydroelectric plants quite rapidly, which is, which is what you need to do so
David Hendry 49:33
Yeah, Britain’s rethinking hydro again, taking the Great Glen and converting it to a massive lake to reservoir to bag more hydro and Norway has always produced most of its energy from hydro. The first ever house to be lit by electricity was driven by hydroelectric. Armstrong, the gun manufacturer, built a hydroelectric system for just providing his house with electric lighting. That’s the first in Britain. So that’s part of the 19th century that we could have got an all electric world. And storage is a big problem Gene really is. And we’ve recommended using nighttime much later nighttime surplus energy to produce hydrogen. There are several methods, let’s not go into them parallel assistant electrolysis and creating liquid hydrogen, right. And liquid hydrogen is a fantastic storage of power. Okay, and you can use it either for heat, or to provide the electricity that you don’t have otherwise, indirectly or as a high heat source for industry if we’re going to get rid of coal and oil, they’ll need a high heat source. Well, you just see NASA’s rockets taking off and you realise you can get a lot of heat using liquid hydrogen mixed with oxygen
Gene Tunny 50:56
And so do you think that could be commercially viable, we’re trying to build a hydrogen industry here, not me, but the state government and the industry. And I know the Japanese are very interested, Mitsubishi and companies like that they’re looking at, they’ve got all these exploratory projects up and down the coast here. I mean do you think it could be commercial?
David Hendry 51:16
Yes, definitely. I don’t know what the cost to steel makers is of their energy provision. But if the hydrogen is made from the surplus energy at night, from things that wind turbines, which often have to be switched off, because you’re producing electricity that can’t be consumed, but it will always be able to be consumed from making surplus hydrogen, that’s our surplus energy for making hydrogen. And the cooling will also require a lot of energy. So I think it could actually, they could actually end up paying people to make hydrogen. Right to stop the wind turbines being turned off when a large percentage of your electricity is coming from wind turbines. And it’s coming at night, when you know three in the morning, the demand is zero. So I think there’s strong possibilities of using that.
Tim Hughes 52:10
That’s good battery technology is really another area as well, of course that is is going ahead. I was just trying to remember the name of the technology, I think it’s single cell batteries. If that sounds right. But I know Toyota, for instance, have invested a lot of money in this next generation of batteries. And it’s been talked about in the realms of that there will be sufficient enough in a car that you’ll be able to power your house from your car. So it’s that kind of capability that is being expected. Remember John Atkins, mentioned this in one of the
Gene Tunny 52:42
Yes, we might have to go back to that Tim and have a look and put some links in the show notes.
Tim Hughes 52:49
It’s a thing, it’s a thing. I didn’t make it up, I promise.
David Hendry 52:51
Okay, so you have to remember that using that kind of technology, a glider went around the world. Right? Yeah. It was a glider, but they did do it, and Britain has several electric aircraft for short distance travel, which are all electric. And trains. I think both Germany and Britain have been developing trains that ran off fuel cells of the kind that are driven by hydrogen. And they do produce enough electricity. But at the moment, the machines that do it are enormous and very heavy. So they have to find some way of producing fuel cells that work from much less expensive and heavy technology. But why not have solar cells in the roof of your car? Well, at the moment people would rip them off, of course, thieves would just take them, but they become ubiquitous. That may be one of the routes that we could do. Another as you mentioned, Musk, at one point, Tesla put up a video on the car being the battery. And they used graphene tubes filled with electricity all the way around the car, and then it provided enough electricity to drive the electric engine for 1000 miles. They dropped the video very quickly. And we don’t know if they dropped it because it was giving away secrets of they didn’t want or it didn’t work to try to match it didn’t work. I think it should work. I can’t believe that I mean graphene is a super capacitor can store enormous amounts of electricity. And there must be a way of using it and making graphene has now become very straightforward. You can take waste plastic, it was a laser and you turn it because it’s a carbohydrate you turn it into carbon and the carbon can be turned into graphene. Just pick up the tiny bits and join them the way that they originally discovered graphene in Manchester.
Gene Tunny 54:57
Incredible. That’s just incredible what’s going on, David we’ll have to put, I’ll put a link in the show notes to your research group on climate at Oxford, isn’t it. So I’ll put a link in because there’s links to all sorts of great stuff you’ve done and all great articles. There’s one more thing I wanted to cover before we wrapped up, because I know we’re getting close to time. There was one thing that you mentioned in that talk that you gave in 2010. This was to the Institute for New Economic Thinking, if I remember correctly, I think was George Soros in the audience? It was incredible. It kept showing George Soros in the audience because I think it’s his, his Institute, or he funds it. But there’s a mention of the insurance company, you talked about this large US insurance company that you’ve done some modelling for or they were using your approach and 5000 variables, and but only a few 100 data points? And could you give us a flavour of what type of modelling that was? I mean, without revealing anything commercial, I was just…
David Hendry 55:55
Yeah, okay. So the 5000 variables are things like a 28 year old, single woman living in Texas with one car, and no children and owns their house. And that’s a variable. Right. So they then price their insurance for her house, knowing all these factors. They were finding that the sales were getting too difficult. And they wanted some simplification of what was actually driving their profit. So Jurgen Doornik who actually did this work, already had auto metrics, working with quite short time series on these sorts of various people. But you could then model all the things that might be taken into account and discover, actually, it didn’t matter that they were single, it probably didn’t matter, they’re female, it probably didn’t matter they’d no pets, right? What mattered was that they owned the house or didn’t own the house that did have a mortgage or didn’t have a mortgage and whether it was fixed term or so very few variables turned out to matter. And it allowed the company to dramatically simplify the number of people that they employed for sales, right, they came way down until when the financial crisis hit, their cost base was vastly lower. And they survived the financial crisis in the way that many insurance companies didn’t, because they, you know, they lost so much money in housing, for example. So that, although it says 5000 variables in most of them could never have mattered when being male would not apply taking female, for example. So males would not have been entered in the models for females and age that older people can’t be young females. So you can see immediately, although they talked about, we talked about 5000 variables, and there were most of them wouldn’t have been relevant in any given situation.
Gene Tunny 57:56
Right. And so this was the autometrics or autometrics is that part of
David Hendry 58:00
OxMetrics. That’s part of the OxMetric system. And it’s written in Ox, Ox is a computer language that Jurgen developed in the 1990s, early 2000s, it was the first attempt to get fully automatic modelling. And I have to say, our first attempts were really ridiculed by the profession, the idea that you can automatically model it didn’t require human intervention. Well human intervention is of course, thinking about what goes into the model. After that, itt’s pointless spending hours trying to see which is the matter when the programme can do that much more efficiently, much faster and much more generally. So it gives, we believe it gave much more time for thinking and much less time wasted in front of the computer desperately trying to find the model.
Gene Tunny 58:51
Yeah, it sounds to me like our central banks and treasuries and finance ministries should be, yeah they should, if they haven’t got a copy of your programmes, they should get a copy of them and start applying them because we certainly need to do better than the, than we have been in terms of forecasting. So…
David Hendry 59:09
RBS definitely has a copy, the sorry the RBA.
Gene Tunny 59:14
Okay, yeah. Yeah. It’s hard to know what they rely on some, I guess they they’ve got a model. I don’t know to what extent that it informs their policy actions. They got this Martin model, which is Yeah, yeah, it’s I don’t know. I don’t it doesn’t I don’t know whether it’s was developed using your
David Hendry 59:25
No, it wasn’t.
Gene Tunny 59:30
No I might have to come back to that because it’s a it’s a rather complicated model not not today, but just just good to Yeah, it’s good to good to find that out at least they’ve got your software if they and hopefully they’re making use of it to some degree. Okay, Tim anything more for Sir David?
Tim Hughes 59:50
No, I just think so it was really interesting. Thanks again for your time, and I think it just reinforces the, the way you talk about the modelling and the conversation we had about AI. Again, it’s something that’s come up in other areas where it’s a really powerful tool, but there’s a human discernment at some point to sort of like, bring it all together. And without that, it’s only so useful. And so I think it’s always encouraging to see that we need that human intervention to make sense of things. And sometimes humans get in the way of good modelling. I imagine, you know, that, but if we can give that amount of work over to the models and the AI to do the work for us as a tool, then, yeah, it’s very powerful.
Gene Tunny 1:00:35
Very good. Okay. Well, Sir David Henry, thanks so much for your time. I really enjoyed it really appreciate it really learned a lot. So thanks. Thanks so much.
Tim Hughes 1:00:40
Thanks for your time.
David Hendry 1:00:43
Thanks a lot. Thanks for having me interviewed.
Tim Hughes 1:0046
You’re welcome.
David Hendry 1:0050
Take care.
Gene Tunny 1:00:54
So Tim, that was an amazing conversation with Sir David Hendry, what did you think?
Tim Hughes 1:00:59
Yeah, that was fascinating. I really enjoyed it. I was a very happy audience member for most of that but yeah, I really lapped it up.
Gene Tunny 1:01:06
Ah, you’re more than just an audience member Tim! I mean, I think you are. You’re participating in that conversation. And you’re asking some good questions. So yeah, it was good to have you onboard. And what I found fascinating is I mean, I had the I had the line of questions about forecasting. And then we went broader than than just the forecast. And we started talking about climate econometrics. And you know, what he’s doing the modelling of getting to net zero for the UK, which I thought was absolutely fascinating.
Tim Hughes 1:01:40
Yeah. And he mentioned that the modelling hadn’t actually progressed in many ways, like, not necessarily with the climate econometrics, but with the other modelling that we were talking about in the first half of the conversation, which was surprising to hear.
Gene Tunny 1:01:53
Yeah, the forecasting. I mean, not what he’s doing, because he’s really a leader in forecasting
Tim Hughes 1:01:59
Yeah so his model he could back
Gene Tunny 1:02:03
Yeah, I mean, and of course, he’s going to back his own modelling, but I’d actually believe it because he’s one of the gurus of econometrics. So when I was studying econometrics, or first started studying back in the 90s, he was one of the big names. And yeah, the approach that he had this general to specific approach where you’re, you’ve got this very general specification, you’re trying to hone in on this more specific specification, you’re searching for the right functional form the right way to express the equation, the right variables, the right number of lags, and some clever things to get things back on track. If they’re shocked things like error, they call them error correction mechanisms. You remember, he was talking about the volcano modelling the global temperature after a volcanic eruption, I thought that was really interesting how he had this clever little functional form this V shape to get to get it back on track to model that I thought that was really clever. And I mean, he’s renowned for doing that sort of thing in his economic models. And, yeah, I was surprised that there hasn’t been a more widespread take up of that approach. And I think my takeaway from this is that, yeah, there needs to be more education or more outreach from from David’s group, I guess, at Oxford to really, you know, promote their their methodology, I guess they’re they’re doing it, they’re trying to do the best they can. And it looks like, you know, central banks or reserve banks got a copy of their software, the OxMetric software, which has PcGive and the other, the other parts of that software, but from what I’ve seen, it’s not as widely used as it probably should be
Tim Hughes 1:03:50
With any modelling, couldn’t that just be run in tandem as a hypothetical to see how it might have performed against the current models? Is that how it would work?
Gene Tunny 1:04:00
Yeah, yeah. And hopefully, they’re doing that. But
Tim Hughes 1:04:03
yeah, because like, you would think at some point, you know, if it’s outperforming, everybody’s interested in, you know, the best outcome s o it would be interesting to see, didn’t, didn’t actually ask that direct question. But that would be interesting to know if that might be feasible or possible, or if it’s been done?
Gene Tunny 1:04:20
Well, I think hopefully, the central banks and Treasuries are using this approach, or they, or I hope they they’re experimenting with it, they should use it more from what I can tell. What I found interesting about our conversations, he was talking about how I forget whether it was the Treasury in the UK or the Bank of England, they were overestimating productivity growth in the UK for a consistent period for like a 10 year period or something. And so that sort of mistake could have real consequences because if you, if you’re overestimating productivity growth, then you’re overestimating what you’re GDP growth is, what your economic growth is, your, in your forecasts, and therefore, what that would mean is that you may not have your policy settings, your monetary policy settings, right, because you think GDP growth is going to be faster than it actually will be. And so maybe you’re not giving, you don’t have the bank rate low enough to to help, you know, promote economic growth. So that sort of forecasting error can have material consequences, if you know what I mean. So it’s important to get these forecasts, right, because if you’ve got those forecasts of where the economy is going wrong, then that can affect what the Bank of England does, or what the government does with its budget.
Tim Hughes 1:05:41
Yeah. And it seems to be human. The human influence, which is the most unpredictable, or the, the element that is most likely to bring around an incorrect forecast.
Gene Tunny 1:05:53
Yeah, I mean, I guess the human element and yeah, I mean, all sorts of things. But the problem is that the economy Yeah, I mean, ultimately, it’s about humans, because they’re, they’re the the units in an economy. But the economy is so complex, and so many moving parts, it’s just very difficult.
Tim Hughes 1:06:13
There was some, also, with the second part of it, the climate, econometrics, which was fascinating, I didn’t realise the extent to which Sir David had been involved in that. And so he was very deeply involved, and it was really interesting, that part. And I know we’ve talked about climate and getting to net zero and the the challenges faced with that, and the changes in the technology that is driving us towards that, which is obviously an ongoing and very interesting subject. And I have to say, so I made a, I was trying to recall this information that our friend John Atkins initially mentioned to us about solid state batteries. So I was trying to remember what it was exactly. So this, these is the one of the new generation of batteries, which, you know, is still in development. So we’ll link in the show notes, I found something from the Guardian of July 2023 that explains a little bit more about solid state batteries.
Gene Tunny 1:07:00
Oh, good. So what does it say?
Tim Hughes 1:07:05
It says, just briefly, that basically Toyota has made a breakthrough that will allow it to halve the weight, size and cost of batteries, it would be that we are, da da da make batteries more durable, and believed it could now make a solid state battery with a range of 1200 kilometres or 745 miles that could charge in 10 minutes or less. And the company expects to be able to manufacture them as soon as 2027. So this is obviously not, it hasn’t happened yet. So this is a projection. But it seems that they’re quite close. So this is the sort of, I saw this a few months ago, it wasn’t the Guardian article, it was something through ABC, I believe in Australia. Along those lines of about Toyota has worked with solid state batteries. Clearly, there’s technology being, you know, pushed forward all around the world on different areas, and which ones come to the fore or make it to market like, you know, we can only speculate. But it certainly seems that there are changes coming and more efficient, effective ways of storing energy and producing energy, that are all moving towards that target of net zero by 2050.
Gene Tunny 1:08:16
So this is something that’s better than the lithium ion batteries. Yeah.
Tim Hughes 1:08:21
And like everything else, there were problems at the beginning that they’re trying to work out. Yeah. So yeah, it says that for benefits compared with liquid based batteries. I think one of them was I saw that it doesn’t, they don’t burst into flames as easily, which is a good thing, like so when they say there’s no spills, for instance, from a solid state battery, if they get in a prang, or whatever may happen. I mean, I’m sure that there’s pros and cons with most new technologies. And I’m sure there’s none. You know, it’s no different with solid state batteries. But it seems to be one of the ones that’s coming through, it’s been around for a little while, and it seems to be progressing. But it’s just one of those areas that by the end of this decade, it’s going to be very different. And of course, by 2050, there’ll be things we haven’t even heard of yet, that will be key parts of the whole target net zero target,
Gene Tunny 1:09:10
We might have to get someone on the show who can explain to us batteries, solid state versus other types of batteries, because I’d be fascinating to know about the technology and what minerals are required. Right? Because I mean, you had that conversation with Guillaume about the Dark Cloud, wasn’t it? And that’s, you know, the, the fact that we do need to mine all of these, these critical minerals and that’s that’s got consequences, of course. So yeah, we’ll be good to have that conversation. Tim, one of the other things I found fascinating, yeah a couple of other things in the conversation. I found David’s point about all the technology that was available at the end of the 19th century. I thought that was fascinating. That was fascinating.
Tim Hughes 1:09:55
I didn’t I didn’t know that at all. That surprised me. It surprised me big time.
Gene Tunny 1:09:59
Yeah. yeah. And also the point about nuclear energy I was, yeah, that was really surprised by that, that he was so positive about it and thought it could work here in Australia, and that perhaps some of the concerns that we have in Australia, about nuclear energy are misplaced.
Tim Hughes 1:10:20
So that was specifically modular.
Gene Tunny 1:10:23
Yeah, the small modular reactors, and he’s saying they’re a lot safer. And this is something I talked about Ben’s, I talked with Ben Scott, about a couple of years ago, I think, on this show, the potential of small modular reactors, I though that was good that David brought that up.
Tim Hughes 1:10:39
Yeah, that was interesting. I hadn’t actually heard of that before. And I know that we’ve, like spoken about it before with Josh Stabler, for instance, with the the likelihood that there’s going to be different solutions to the energy provision in the future. So it’s not just going to come from one main source it’s going to come from most likely several different ones. So if modular nuclear power stations are a part of that, that’s quite possible. But it’s clearly going to be not just one thing. And just on that subject as well. There was it was mentioned about graphene, David mentioned, oh, yeah, I know this, that’s come up in a conversation we’ve had before here in Brisbane. There’s GMG, who are involved in that, here in Brisbane,
Gene Tunny 1:11:23
with GMG. So G stands for graphene I guess, and there’s…
Tim Hughes 1:11:27
Graphene Manufacturing Group. And so it’s in the space of renewable energy, and this whole push towards clean energy.
Gene Tunny 1:11:35
But what’s graphene got to do with it? Because graphene is a material, isn’t it?
Tim Hughes 1:11:39
Very good question, Gene. And this is something that I will get back to you as soon as I know. Yeah, because this is actually my solid, solid state battery moment in the in the wrap up, I managed to get another one in. So we’ll put something in the show notes to do with the graphene, but I know it was, it came up in a conversation we had with
Gene Tunny 1:12:02
ah apparently it’s going to create products with a better efficiency than the existing ones.
Tim Hughes 1:12:07
Yeah so it’s part it’s part of the all of this emerging technology towards better energy storage. Like most other things, it’s happening in many different parts of the world at the same time but we do have this, this company in Brisbane,
Gene Tunny 1:12:19
it looks like Yeah, yeah. Sorry Tim I just noticed it looks like someone’s built a graphene solar farm. So it looks. Okay.
Tim Hughes 1:12:25
Yeah. So to be to be explored, like, because it’s not something I know a great deal about, or, you know, so I think that w’e’ll definitely will, will earmark that for the next conversation we have about clean energy and where that’s going.
Gene Tunny 1:12:36
Yeah, for sure. Because this is it’s an ongoing issue. And I mean, the conversation, this isn’t going away. And you look at this northern hemisphere summer, and yeah, I mean, that’s just going to intensify this conversation, I think. Yep. Yeah. Okay. The other thing I thought was good about the conversation with David, I liked your question about AI and what’s happening with AI? And David pointed out, well, what they do, their algorithm, their automatic model selection algorithm, their auto metrics, that’s a form of AI. I thought that was a good point that he, he made there. Yeah, yes. Yes. So yeah, it was good question. So all in all, what a terrific conversation. And yeah, I really thought Sir David was amazing. He’s someone I would love to have here in Australia participating in the Australian policy debate on energy in particular, I think he could be that he could provide that sage perspective. He’s someone you’d pay attention to, he’s someone who’s very thoughtful, you know, good communicator, and as well as being a real gentleman.
Tim Hughes 1:13:43
Yeah, I really enjoyed it. And I found it very eye opening. And yeah, I think there’s, like you say, it’s an ongoing conversation. So it’ll, it’ll keep evolving. And hopefully, if we can, maybe they’ll be a round two with Sir David to continue that conversation.
Gene Tunny 1:13:59
We can only hope, so Tim, anything else before we wrap up this, this debrief?
Tim Hughes 1:14:05
No, I think I think I should stop while we’ve still got enough room in the show notes for
Gene Tunny 1:14:13
you introduce a new concept – then I ask ‘Tim that’s fascinating can you tell me more?’ Nope!
Tim Hughes 1:14:20
They say a little knowledge is dangerous. On this one I was lethal, but no, it was fun. I enjoyed it. And yeah, it’s it’s something that affects us all. And it’s something that’s changing very quickly. And so yeah, we’ll we shall return to that conversation no doubt.
Gene Tunny 1:14:36
Absolutely. Tim Hughes thanks for joining me on this conversation.
Tim Hughes 1:14:40
Thanks Gene.
Gene Tunny 1:14:43
Righto, thanks for listening to this episode of Economics Explored. If you have any questions, comments or suggestions, please get in touch. I’d love to hear from you. You can send me an email via contact@economicsexplored.com or a voicemail via SpeakPipe. You can find the link in the show notes. If you’ve enjoyed the show, I’d be grateful if you could tell anyone you think would be interested about it? Word of mouth is one of the main ways that people learn about the show. Finally, if your podcasting app lets you then please write a review and leave a rating. Thanks for listening. I hope you can join me again next week.
1:15:27
Thank you for listening. We hope you enjoyed the episode. For more content like this or to begin your own podcasting journey head on over to obsidian-productions.com
Credits
Thanks to Obsidian Productions for mixing the episode and to the show’s sponsor, Gene’s consultancy business www.adepteconomics.com.au. Full transcripts are available a few days after the episode is first published at www.economicsexplored.com. Economics Explored is available via Apple Podcasts, Google Podcast, and other podcasting platforms.