Steve MacFeely is, chief statistician at the Organisation for Economic Co-Operation and Development and adjunct professor at University College Cork. MacFeely joined OECD in August of last year as chief statistician and director of statistics and the data directorate. Before joining the organization, MacFeely served as Director of Data and Analytics at the World Health Organization and as Director of Statistics and Information at U-N Trade and Development. He has also served as the Deputy Director-General at the Central Statistics Office (CSO) in Ireland.
Episode Description
According to the U.S. State Department, three-fifths of global GDP, three-quarters of world trade, and 90 percent of official development assistance can be accounted for in 38 countries. Those countries are members of the Organisation for Economic Co-Operation and Development – or OECD. Founded in 1961 and headquartered in Paris, the OECD’s goals include stimulating global economic growth by providing a forum for intergovernmental collaboration. It’s also the focus of this episode of Stats and Stories with guest Steve MacFeely.
+Full Transcript
Rosemary Pennington
According to the US State Department, three fifths of global GDP, three quarters of world trade and 90% of official development assistance can be accounted to 38 countries. Those countries are members of the Organization for Economic Cooperation and Development, or OECD, founded in 1961 and headquartered in Paris. The OECD goals include stimulating global economic growth by providing a forum for intergovernmental collaboration. It's also a focus of this episode of stats and stories, where we explore the statistics behind the stories and the stories behind the statistics. I'm Rosemary Pennington, stats and stories is a production of the American Statistical Association in partnership with Miami University's departments of statistics and media, journalism and film. Joining me is regular panelist John Bailer, emeritus professor of statistics at Miami University. Our guest today is Steve McFeely, Chief statistician at the Organization for Economic Cooperation and Development and adjunct professor at University College. McFeely joined OECD in August of last year as chief statistician and director of statistics and the data Directorate. Before joining the organization, McFeely served as Director of data and analytics at the World Health Organization and as director of statistics and information at UN Trade and Development. He's also served as the Deputy Director General at the Central Statistics Office in Ireland. Steve, thank you so much for joining us today.
John Bailer
It's a great pleasure to be here. Steve, it's a delight. You know, this is a conversation I was hoping we'd have for years, and we finally made it. I mean, you kept changing jobs, so I couldn't find you, but eventually I caught up to you and you know, so it's great to have you here, you know. So you've had this, this, this great pool of experience, and it's gonna be fun to explore that in our conversation today, but, but let's just start with, you know, maybe the where you are now, what? What does a chief statistician do?
Steve McFeely
Okay, well, broadly speaking, I have two jobs at the OECD. The first one is director of statistics and data, and that's really looking after my own directorate. So that's really the production of most of the macroeconomic statistics that you would associate with OECD. But the OECD has a decentralized statistical system, so a lot of the statistics that you associate with OECD don't actually come from the statistics Directorate. They come from the economics directorate, the education directorate. So I have a second job, which is Chief statistician, and that's trying to oversee that the governance standards, the quality standards that the OECD has adopted are adhered to by the other production units around the house.
John Bailer
Quick follow up, you know. So could you talk about one of the the data products that OECD produces that that kind of people are just waiting for with anticipation?
Steve McFeely
Well, let me there's probably lots, but let me give you one, maybe where my job as Chief statistician has come in. So that's Pisa. So the the education Directorate of the OECD does this survey called Pisa, which is looking at educational attainment standards and actual ability, educational ability. The reason I've been involved in that recently is the PISA survey makes what's what's known as a public use file available. So this is a DNA This is an anonymized public use statistics file. But recently, a number of countries have become concerned that, because of the massive computing power now available, that the traditional method, methods of anonymization are no longer sufficient, and that somebody may be able to de anonymize the file and see where the data came from. So in that case, I've become involved, because member states approach me. So then, in cooperation with the education department, I'm working now with, in fact, we're getting a lot of cooperation from the United Nations Statistics Division. They have a unit that's really been working hard on what are known as pets or privacy enhancing technologies. So we're working with the unsd, getting advice from them on trying to find the best pet approach or pet methodology that we can apply to the PISA data so that we can continue to publish public use record files that are openly available for researchers. Because otherwise, if we can find a solution, we may have to pull the data in and make them available. It to researchers on an individual contractual basis, but then the administrative costs on both sides is just very, very high. So we're desperate to try and keep these public use files available, but then we may have to just treat the data a little bit differently. So that just gives you one example of something where a chief statistician comes in and helps one of the other units with the statistical problem for a file that's highly used and highly anticipated.
Rosemary Pennington
I imagine that there are a lot of people who don't know about the OECD, and so I wonder, you know, given that the mission is sort of help promote global economic growth and and collaboration. I wonder what kinds of economic data are being produced by OECD that maybe you know, people who are not as up on it might sort of be interacting with or sort of seeing in their lives. Well,
Steve McFeely
let me get let me start with a little bit of history, because the history of OECD is interesting in this context. It was set up as the oeec and it was set up really, to administer the Marshall funds at the end of World War Two. GDP gets invented around the same time as the the OEE see is established. And in fact, one of the metrics for getting funding and continued funding is proof of economic performance. And don't forget, there's a Cold War emerging at this stage, and GDP emerges, as well as one of the tools that the now OECD uses as a kind of a metric for getting Marshall funding. It's also and that's really important, because that metric, then is used across European OECD member states, but later on, this idea of economic development actually comes originally from the OECD, because it's really a post war concept. So a lot of the metrics that were used by the OECD, they then kind of disperse out into the wider un and that brings me to GDP. So that's why GDP has become so kind of globalized, and it's one of the metrics that I think people would most associate with the OECD, but the national accounts more generally, but OECD, or, sorry, GDP, I think, is one of the metrics that we publish every month or every quarter, and we publish it for the OECD and for the g20 and people are waiting for those numbers. Now, more recently, we might get on to discussions later on, beyond GDP, but for the moment, it's still one of the most anticipated numbers. But added to that, I think other ones where we really contribute a lot would be around international trade, but also purchasing power, parities, so we do a lot of inflation, a lot of the stuff that you might associate with the statistics office, but the purchasing power parities is one that really for the kind of the OECD member states. It's the only place we can get those numbers, and they are pivotal to, like, the denominator for so many other statistics.
So again, there's huge interest in that.
John Bailer
So you know, you anticipated one of the questions, the natural follow up questions, when you're talking about about GDP. So you have this metric that, this index, that was developed, you know, decades and decades ago, and and the economy has changed, or production within the economy has changed that. So as you think about this in light of of how do you how do you best tell the story of what's happening, happening in terms of countries economic outputs, what are some of the the the ways that you could see GDP evolving?
Steve McFeely
Let's start with just GDP itself, and then we can get on to beyond GDP. So as you say, GDP is now about roughly 80 years old, and it's been updated three times since it was formally adopted in 1953 by the United Nations. It's about to be adopted, we hope, or updated again this this March at the UN statistics commission. So the last time it was updated was 2008 that's not so long ago. And yet, in economic terms, it's, it's eons ago. And the last time we updated it, we didn't have cryptocurrencies. Just as one example, data weren't nearly as important as a as a mode of production as they are today, and concerns around environment climate have become much more pronounced in that time, so the national accounts are periodically updated to. Try and absorb
that the latest changes.
So as I said, in 2025
this March, we will hopefully be adopting the 2025 revision. That will probably mean it will be adopted in practice around 2029 because once countries adopted in theory, they then have to start preparing. And it's fair to say, quite a few of the changes this time will require some lead in in terms of implementation. To give you one example, data will now be included in GDP, or, sorry, in the system of national accounts as a mode, as a as a mode of production, or a method of production, so it will appear on the balance sheet. This raises enormous questions about now how to value data, because if you don't have a market transaction, what price do you put on it? So for the data that's actually sold, we have a market price, but most data are actually what are called own production. So they're produced by the company itself and used by the company itself, and there's no transaction. Equally, we know for big companies like Google, probably the biggest asset they have are their data, and yet those data don't appear on the balance sheet. So this is the big question, what? How do we value the data? And then once you put something on a value, once you put something on a balance sheet, then the question then is, how do you depreciate that that value? So do data appreciate, or do they appreciate, you know, like a good bottle of wine. I mean, some to a statistician, some data sets actually grow on value with time because the time series gets longer, but other data clearly don't. So this, this one change, raises so many challenging questions about valuation, about length, like life cycle, of data like depreciation, and that's only one. There's other areas that are challenging, both statistically and politically. So one is in the new SNA there will be more emphasis on net production rather than gross production. That's both technically challenging, because you need a lot more information, but it's fair to say that not all member states are equally enthusiastic about moving away from a gross measure for a variety of reasons. I mean, net measures will probably be less flattering, but also, some countries are more actively consuming kind of natural resources, and a net measure will be tougher on them, so countries that have adopted more green practices will be less concerned. Countries that really haven't embraced green technology are very heavily dependent on commodities, on extractive industries, then a net measure will be tougher on their bottom line. So there's a mixture of technical issues and just how challenging these things will be to measure. But then the states themselves will have different ideologies and different concerns about these measures and how it's going to affect the bottom line.
Rosemary Pennington
You're listening to stats and stories, and we're talking with OECD chief statistician Steve McFeely,
John Bailer
so so that really leads to a difficult question. It's, why do countries participate? Then, you know, yeah, so you're, you know, there's this, there's some clear resistance and some reluctance, particularly if, if more, well, I don't know if I was gonna say more realistic, but it's, but just, it's a different measure of production was being assumed, what's, what's the value for the countries in terms of partnering with, with OECD or or more generally, with any kind of international characteristic of of trade or economic growth or value? Yeah.
Steve McFeely
So look, member states may have different views on different issues from time to time, and they will change often as administrations change, but they all recognize the value of being able to benchmark themselves, both in time but also geopolitically. So. How am I doing against my competitor? How am I doing against a country that seems to be developing a niche in an area. Is that really fueling their economy or not? Am I doing better than I was? So if you take the temporal one, I mean one of the advantages of an international system, and one of the advantages a lot of people see in the international system is. Is that the country has to submit their data to scrutiny by OECD, by the United Nations, so there's an extra layer of validation. So a lot of people are happier maybe, to use data from an international source, because they feel it's gone through a more rigorous kind of quality assurance, because they don't always know whether they can trust those data or not. So they want to see. You know, different policies have been enacted at different times. What did the data say? Did that lead to success? Did it lead to failure? Can we tell? And then the other thing the country themselves will always want to know is, I'm doing something. I'm trading with this this country, or my neighbor. How am I doing? Visibly, my neighbors in the region and every country. I mean, every time we publish statistics where there's any kind of a chart, the questions we get are always about the neighbors in their next door neighbor in the kind of region. So for that reason, it provides them a lot of information on what other countries are doing. Because all the time, countries are watching. Governments are watching as go as different governments are trying different policies to see, well, okay, which ones are working. So even though there may be a reluctance from time to time, over a particular indicator, and there may be tensions because of border disputes, and we may decide that we're going to include one piece of land in one country or another. So there can often be tensions. Nevertheless, despite all of that reluctance, those tensions, countries will always want to use those data, and people in those countries will always want to use those data because they also want to. They want an independent measure as well. Of is what the government telling us. Is that true? You know, when a government says, you know, we're our economy is performing really well, we hear all the time, from member states, the media, everybody pays huge attention when the OECD says something about a country, because they take it as impartial, and that impartiality is absolutely critical, in the same way that it will be from a statistics office in the country too. But sometimes there's tensions in the country, so they will look to OECD as that extra source of validation and impartiality.
Rosemary Pennington
I'm glad you bring up this idea of impartiality and sort of the national statistics offices, we on stats and stories, have had lots of conversations with national statisticians, with others, sort of talking through, kind of the struggles of doing official statistics, their importance, especially at this moment where there seems to be, like a heightened distrust of data in some places. Why are official statistics offices in countries important? And I guess, what challenges do you see them facing? You've been in official statistics for a long time now, you must have some insight there.
Steve McFeely
Yeah. So if you go back 30 years, maybe since the birth, really, of the internet, that's been a company that digital revolution inadvertently created a data revolution. So now there's data sloshing out of pretty much everything, our mobile phones, satellites, all sorts of measurement systems, those data vary enormously in quality, and a lot of people are producing metrics all purporting to be authoritative. But in fact, it's very hard to tell as a consumer is that a good product or not, because there's no kind of standards really that apply. So one of the advantages the national Statistical Office has is, first of all, they don't have any skin in the game. They're not selling anything except they're, you know, that the best version of truth that they have. They're not trying to sell you a product, they're not trying to get anything from you. They're just trying to measure what's going on in society, in the economy, in that country. And they've been doing it for a long time. So they have long time series. So they've all, you know, for such a long time, they've been worried about things like bias, about quality that a lot of other people, frankly, aren't worried about, but the statisticians are, and thirdly, every national Statistical Office that pretty much at least are signatories to the UN fundamental principles of official statistics. And effectively, that's their, I was going to say their Bible. It's not their Bible. It's certain. It's like the North Star, yeah, you know. And that really is sacrosanct. And some of those principles, they really highlight things. Like impartiality, statistical rigor, in international standards, professional independence and those standards. They can sound trite, but in fact, they're not. They're of fundamental importance, because it's back to that simple expression. They have no skin in the game. Their only interest is to produce the best data that they can that will help to inform policy, but also will help to inform the general public about what's going on. So for democracies, at least national statistical offices are critical, because that's how that's where you get your information from, that's where the there's no one really can compete with them. I mean, there's lots of people producing data, but I know who I would trust now in terms of challenges, the challenges are many. I mean, one is political, because of the data deluge, because of all of this information, there's been a mistaken belief, I think, in some administrations, that data are now cheap. And in fact, the increase in volume has actually done the opposite. It's actually made statistical production more expensive, good quality statistical production. Secondly, we're seeing, I think, a slow erosion of the social contract in many countries, where we've become much more individualistic, we're unfortunately thinking less about society as a whole, and that's leading to diminishing response rates, and we see this globally, that that's a challenge. So what it means is national statistical offices are really becoming more reliant on a data integration model, which is using traditional surveys and census, but integrating them with new types of data. Those new types of data are often unstable. They're becoming expensive now and then finally, the last challenge, which is probably the most worrisome at the moment, is in a number of countries, we've seen a tilt away from supporting scientific evidence. So we've seen this more kind of, whatever you want to call it, post truth, bespoke truth. And that's worrying, because then people, they don't understand the value of their statistical office, and they think any old, you know, you hear this expression, any number is better than no number, and it's like, Well, not if it's not a good number, it's not so there's a misapprehension, I think, and that's a threat to the statistics offices are alert to. But it's not always clear how to
John Bailer
how to tackle it. You know, as you're talking about this, I keep finding myself wanting to say, Amen to many of the to many of the observations that you've made, the value of official statistics, I think, is just incredibly underappreciated, even though the decisions that are being made this, you talk about evidence based policy in other contexts as being, you know, having that evidence that's of good quality and good insight. But this also leads to another thing that I've noticed you've been very active in, and that is the idea of, how do you effectively communicate this? And part of that is, is preparing, preparing consumers, citizens, to be more statistically literate. So, so can you talk about why you know, kind of your your perspective as a as someone who has been deeply involved in official statistics has kind of also, in parallel, been very committed to this idea of International Statistical literacy.
Steve McFeely
Yeah. I mean, this is a labor of love for me, but it's not entirely altruistic. I mean, I see them as very much connected. When I worked in Ireland, I had a it took me a while to convince my Director General at the time that as a statistics office, that we had a responsibility to the kind of statistics, statistical literacy discussion I I'm not sure I convinced him, but at least he gave me some resources. And I think that's been really important, because, again, statistical offices are so well placed, not just in having the data, in having the skills, but institutionally, they're well placed. They have connections. They're not well resourced, but compared to many other schools and other people, they often are well resourced, and they may be resourced in kind, not necessarily financially, but they have other networks. So for me, that the reason I'm so interested in it is I've I've had a lot of meetings with officials that I would have. Consumed, understood what we were doing and knew how to consume our product. And in fact, what I've discovered when I go into Ministries is, in fact, they don't really know how to use our product. In a lot of cases, they're actually terrified of the product, and that's a worry, you know, so you kind of realize some of the people you assume are your your core consumers, really don't understand the product. So professionally, that's really important. We need to address that. And that's why, over the years, I've been involved in setting up this in Ireland, this professional diploma, it was really geared towards helping people in the administration. But then you start to think, Well, okay, look, these guys are already working. To some extent. The damage is done, because mentally, there's some kind of trigger has already been flicked. They're already scared
of the data, and
you can tell them there's no reason to be scared, but they are. So the other half of me has said, Well, okay, look, we need to get into schools. We need to catch kids early and teach them that this is actually a joyous experience. This is actually fun. What you're doing is you're answering questions and you're playing with data. And in doing that, I think, I hope what we're doing say with the islp is we're really encouraging kids to think. We're helping them to be curious, but I hope as well, in doing that, we're helping them to be a bit skeptical and kind of look at data with a slightly arched eyebrow, kind of really, you know, by presenting data that doesn't make sense and making them think. And because, to me, that's the critical issue. I mean, it's about judgment, and I get really worried about this expression about data driven decisions, but data don't drive any decisions. I mean, somebody has to make a decision, a politician has to make a decision. And during COVID Especially, I was worried that a lot of people were being let off the hook. They were kind of saying, Oh, well, you know, the the data told us to do that. I said they didn't tell you to do anything. I mean, so it's trying to encourage judgment to just, you know, use the data, but be wary of the data all the time, even if you, if nothing else, if you can, if you can, get across to kids, when you look at a piece of data and just think, where did that data come from? Who produced it, and, critically, why did they produce it? Like, who are they, you know, if they're a private sector company, or if they have a particular angle, if they're an NGO, maybe they want to find more litter on the street, because that's their business, you know. So just, just be wary. Just just be be skeptical and look for corroborating evidence, you know, like cross check, just even that simple level of of statistical literacy, which isn't very sophisticated, but even that, and also just trying to again guide them towards your statistics Office, you know that's these are trusted sources. When you're living in a society where there's a morass of stuff being thrown at you, there's a there's a beacon in the middle emitting this white light, you know, and go towards it,
Rosemary Pennington Stats and Stories is a partnership between the American Statistical Association and Miami University departments of statistics and media, journalism and film. You can follow us on Spotify, Apple podcast or other places where you find podcasts. If you'd like to share your thoughts on the program, Send your email to statsstories@amstat.org or check us out at statsandstories.net and be sure to listen for future editions of stats and stories where we discuss the statistics behind the stories and the stories behind the statistics.