Q&A with Andy Haldane, Bank of England

Andy Haldane
By Tim Phillips
Fri 19th July 2019 | Blog

Andy Haldane, Chief Economist and Executive Director, Monetary Analysis & Statistics at the Bank of England has been spearheading the efforts of the Bank to engage more fully with the public, and to encourage greater understanding of its role, and of the data the Bank publishes. He has also been an enthusiastic supporter of CORE, and spoke at our launch in November 2013.

As we near completion of the next release of our Doing Economics practical projects, we asked Andy why he believes that what he calls the “tsunami of data” offers a huge opportunity for all citizens, and how CORE’s material can help develop these skills for economists and non-economists alike.

In 2019, how important are data skills for people who are not economists?

There are some simple but important reasons why that’s the case. Firstly, there’s more data around. In 2020, 1.7MB of data will be created every second for every person on earth[1]. We are all being bombarded with numbers ever greater regularity, so it is important to be able to make sense of them—and just as importantly, to avoid making big mistakes when we use these numbers.

Why should everyone be engaged with this?

People are the economy. The things they say, how much they spend, when and how much they work: these are the building blocks of our economy. And so we all benefit if we can make sense of data. It helps us understand the trends in society, the decisions we take day-to-day, and the choices we make when we vote.

I’ve been aware of the CORE Project since the get-go … It has been fantastic to see the vision turn into something real

What do we lose if we leave the decision-making to an expert class?

That was the way it was done for many years. But the world has changed, and that expert class is no longer as trusted as it was.

Interrogating the numbers is one of the important mechanisms through which society can hold the executive to account. It’s also a mechanism by which trust can be built. If more people believe that our judgement is sound, because they can see the data that underpins it, it helps to build understanding of our role.

In the Bank of England we’ve made a big push over the last few years to improve the degree of understanding and trust in us, and that means pushing out more information and opening up more data, for example in the Bank of England’s KnowledgeBank initiative.

Learning how to apply principles rigorously in the real world has traditionally been challenging for many students.

True. But “rigorous” does not mean “making it boring and technical and complex”. It means things like being able to make judgements soundly based on the data and—very important—being able to spot where data has not been used appropriately.

One of the most important roles of the Bank of England is to be able to point out situations when non-rigorous judgements are being made by people on the basis of a poor understanding of the data, sometimes our own data.

Do you have an example?

There is a lot of focus right now in the UK on the impact of our businesses on the economy, some of which are making the decision to delay investment. They are doing this for a number of reasons, one of which is obviously Brexit. But it’s very important that the conversation about our economy is not only about what businesses are doing, because there are also these things called ‘consumers’, and it turns out that consumers make a contribution to aggregate GDP that is seven or eight times more important than the contribution of firms. And so, we need to ensure that the conversation about what goes on in the economy gives at least equal weight to the activities of firms and consumers.

For creating understanding of statistics, how important is it to work with real-world data, such as the projects in our Doing Economics ebook?

That’s one of the most crucial things about real data. At the Bank, we bring a lot of our economists straight from university. An important part of learning the job is to understand what types of data exist, and which data does not exist, which types of data are solid, and which are less reliable. Also, where data comes from, when we can believe it, and when we should not believe it.

When I joined the Bank I spent many years learning this. To get under the skin of data, what it means and its limitations, it is essential to get your hands dirty by working with real datasets.

Should our teaching emphasise that working with data is a detective story, rather than a question with one correct answer?

Preparing for the Bank’s Monetary Policy Committee meeting is like doing the world’s most complicated jigsaw puzzle. Some of the pieces of data are missing. Some of the other pieces of the jigsaw are damaged, because the data isn’t as robust as you want it to. We’ve also lost the cover of the box, and so we don’t even know what the jigsaw puzzle would look like if we finished it.

But we still have to try to do the jigsaw, and that’s exactly what makes working with data fascinating. Sometimes we have to go out and look for new pieces for an entirely new jigsaw, because suddenly we have new questions, and putting our existing pieces together can’t answer them.

And so are more pieces becoming available for you, or anyone who wants to try to solve these puzzles?

Yes! For example, at the Bank we now have the opportunity to make use of very granular microdata. This is data on things like VAT returns, or the benefits people get paid, or the taxes that they pay, or even how much traffic is on the roads, the number of ships coming through our ports, or satellite data. This data is just beginning to be used to make sense of our world.

Interrogating the numbers is one of the important mechanisms through which society can hold the executive to account

But do we still need those practical data skills you learned?

More than ever, yes. This data is not neat, like the data you get from the Office for National Statistics. It has ragged edges, and noise, and the samples may not be consistent. Big Data is so big that sometimes you need to know how to filter it to extract information that you can work with.

Some of this new data isn’t even made up of numbers. In economics there has recently been a very interesting push to make use of the words that people speak and write and analyse the frequency and meaning of them as well.

Does this require an interdisciplinary approach?

One of the great virtues of bringing in all this new data is it means that we can develop techniques that draw on a number of disciplines. Bringing big data into the equation, literally and metaphorically, means we need insights from data science, machine learning, physics, sociology, anthropology, history, and psychology among others.

How does CORE contribute?

I’ve been aware of the CORE Project since the get-go. I was at the launch in 2013 because at that time it absolutely was needed. It has fulfilled that promise. It has been fantastic to see the vision turn into something real.

You couldn’t wish for more from the project as a way of resetting the clock on the curriculum. It has engaged students with the principles that make the economy interesting, tackling issues that we know matter most, like financial crises, inequality, the climate crisis. These are front and centre in CORE’s work. This type of thinking is also front and centre of what we are trying to do.

Finally it is interested in reaching wider audiences, not just those studying economics. I think it’s absolutely the right direction of travel, and the desire to develop those skills for non-economists and well as economists is absolutely the direction that we’re also taking at the Bank. All of that counts as a huge success story.

[1] Domo (2019), Data never sleeps, sixth edition.