How to spot fraudulent economic arguments: an example from Lexit

When I last talked about Lexit, I said Lexiter arguments had “many structural similarities to right wing arguments for Brexit.” In this post I want to go further and suggest ways you can spot when economic arguments are fraudulent, using an article by Thomas Fazi and William Mitchell advocating Lexit as an example. To show how generic the tactics are, I will demonstrate how exactly the same techniques were used by those who argued for austerity.

The first sign that the wool is being pulled over your eyes is wildly exaggerating your opponents case. It is so much easier to attack a straw man. This particular article talks about the “Left’s anti-Brexit hysteria” i.e anyone who disagrees with us is acting hysterically. They talk about those opposing Brexit predicting “Brexit Armageddon”. The tactic works particularly well if you can find an example of someone who did exaggerate their case, and focus on how their fears have not come to pass.

I am very familiar with this tactic from the UK in 2013. From 2010 to 2012 growth was very low, and the expected recovery from the Great Recession did not materialise. People like myself blamed austerity, and we were quite right to do so. But some in 2012 made the foolish claim that the economy will never start growing again unless austerity came to an end. So when growth returned to trend in 2013, proponents of austerity used those claims to announce the critics were wrong and austerity had been vindicated. The example that Lexiters and Brexiters alike use is of course the Treasury short term post Brexit forecast.

That forecast was not about the longer term impact of Brexit, but uncertainty about the prospect of Brexit. It overestimated the immediate impact of Brexit. But the real question is not whether forecasts were wrong, but whether the Brexit announcement and subsequent uncertainty has damaged the UK economy. The article says the economy is still growing, and quotes some other selected statistics, implying all is fine. They do not mention that we have gone from the top to the bottom of the international growth league. Of course they cannot avoid acknowledging that sterling fell sharply on announcement. However they say but this has not destroyed the British economy: another straw man (who ever seriously said the depreciation would destroy the economy?!), and no mention of what the depreciation did do, which is cut the real incomes of workers. The debate over whether Brexit uncertainty has harmed the economy is over, as no one seriously thinks it has not. 

A second sign of a fraudulent argument is to focus on a single study that supports what you want to say, and ignore all the rest. The Brexiters have Minford of course, but Lexit have the study by Coutts, Gudgin and Buchanan (CGB) published by the Centre for Business Research (CBR) at Cambridge. Those advocating austerity had two well known papers: one suggested there was empirical evidence that high government debt reduced growth, and the other that austerity was expansionary. In all these cases the studies are outliers, so you need some reason why the majority of academic work can be discredited.

This leads to a third tactic: tar academic work that goes against what you say with some broad assertions that have only a grain of truth. If your audience has a political bias, play to it. With austerity it was that those against austerity were old fashioned Keynesians who just wanted a larger state. (Some were, but most were just using a much more modern Keynesian framework and its macroeconomics implications.) The way this paper does the same with the many studies that have suggested Brexit will be harmful in the longer term is so laughable I have to quote it:

“The neoliberal biases built into these models include the assertion that markets are self-regulating and capable of delivering optimal outcomes so long as they are unhindered by government intervention; that “free trade” is unambiguously positive; that governments are financially constrained; that supply-side factors are much more important than demand-side ones; and that individuals base their decision on “rational expectations” about economic variables, among others. Many of the key assumptions used to construct these exercises bear no relation to reality.”

It will be news to most trade economists whose work is the basis of the long term case against Brexit that they are assuming things like governments being financially constrained and rational expectations. A key part of why Brexit is a bad idea is the gravity equation, which says trade is more likely to take place with neighbouring countries. It is a robust empirical relationship that makes no assumptions about governments or markets at all. [1] The article pours scorn on the ‘neoliberal’ idea that openness to trade is good for the economy, and again neglects to say that the relationship linking openness to productivity growth is also empirical, not theoretical.

Another tactic that if you see being employed you should start to worry is to impugn the motives of your opponents. That may be a common enough tactic in political discourse, but not if you are trying to make a serious economic argument. It is notable how academic economists who have criticised the Brexit work of Minford for example do not try and suggest he is ideologically motivated. Being good academics they instead question the model he uses (e.g. no gravity) and how he uses it. That is what should be done. But this article tries to discredit the current government’s analysis of the impact of Brexit by implying the results have been fiddled to produce large numbers for the long term costs of Brexit. They write:

“The models are notoriously unreliable and easily manipulated to achieve whatever outcome one desires. The British government has refused to release the technical aspects of their modeling, which suggests they do not want independent analysts examining their “black box” assumptions.”

Apparently the current UK government do not want us to see their workings because if we did we could see how they had fiddled results to make Brexit look bad. The attempt in this case is again laughable, because this government is pro-Brexit so any fiddling would go the other way.

All these four signs are easy to spot. The fifth is less obvious to the non-expert, which is to inconsistently use lots of broad brush statistics that do not get the heart of the issue, or which are problematic for other reasons. For example in arguing that Brexit has had as yet little impact on the economy they quote unemployment data, just as the government did from 2013 to argue that the economy was strong as a result of austerity. It is problematic because strong employment growth coupled with weak output growth means poor productivity. As I noted here, current productivity growth performance in the UK is worse than it has been for centuries.

In terms of broad brush statistics, they argue that GDP growth has been weaker rather than stronger than post-war trends as a result of EU membership, as if nothing else had been going on in all these years. The reality is that we joined the EU in part because our post-war growth and particularly productivity was worse than most other countries, and from the 1980s onwards it has if anything been better than other major countries. But why look at GDP rather than exports, where you would expect the impact of EU membership to be clearer. In fact the CGB study they quote from extensively does exactly this, and the following chart shows the results.

The benefits for UK exports of being in the single market are clear from this data. (To see how the CGB study erroneously avoids this conclusion, see here.)

There are of course plenty of Brexit specific points as well (the article does not even mention the Irish border), but this post is long enough already. I wrote it because it struck me when reading the Fazi and Mitchell article how similar its rhetorical devices were to those of many who tried to sell austerity. So if you read anything that wildly exaggerates the opposing position, focuses on a single academic study and ignores the rest, particularly if it suggests that there is some generic problem with these other studies which may include the motives of those who did them, and if it uses broad brush statistics when it could have used data that was more relevant, beware that you may be reading a piece of political persuasion rather than serious analysis.

[1] The confusion is compounded when they quote Paul Romer’s criticism of macroeconomics, and then drop the macro bit to pretend that Romer was talking about the whole of economics.

The journalist as amateur scientist

Paul Romer has talked about two types of discourse, one political and one scientific. He uses that distinction to critique aspects of current practice among economists. I want to do the same for journalism.

Political discourse involves taking sides, and promoting things that your side favours. It is like a school debate: you consider only evidence that favours the point of view you want to promote. Scientific discourse involves considering each piece of evidence on its merits. You do not aim to promote, but assess and come to a conclusion based on the evidence. That does not prevent the scientist arguing a case, but their argument is based on considering all the relevant evidence. There are no sides that are always right or invariably wrong.

Of course, any scientist makes choices about what evidence is relevant, and this will be influenced by existing theories. Ideally the theory you prefer can be changed by new evidence, but scientists being only humans can sometimes be reluctant to accept evidence that contradicts long held theories. But there are always younger scientists looking for new ideas to make their name. The scientific method works in time, which is why we are where we are today.

My argument is that journalists should be like amateur scientists. Amateur because part of their work will involve seeking out expertise rather than starting from scratch, and they do not have the time or resources to investigate each story as a scientist might. A term frequently used is ‘investigative journalist’, but that normally means someone who has weeks to work on one story. Instead I’m talking about journalists who only have a day. The key point is that they should not search for evidence that fits the story they wanted to write before doing any research, but allow the evidence to shape the story.

For example, suppose the story is about EU immigrants and benefits. What a journalist should note is that unemployment among EU immigrants is lower than natives. What a journalist who wants to write a story that makes immigrants look bad might do is say that the number of EU immigrants without a job make up a city the size of Bristol. This combines selection of evidence (where is the equivalent figure for natives is not reported) with simple deception: most people conflate ‘without a job’ with ‘unemployed’, rather than being people happy looking after children, for example.

If this all strikes you as obvious, at least to journalists working in broadsheet newspapers, the example above is taken from the Telegraph, and the post in which I discuss it contains a tweet from a Times economics editor saying that all journalists (and yours truly) take a stance and select facts that supports this stance.

There is actually a third type of journalism, which you could call acrobatic discourse, because it is always looking for balance. It is sometimes called ‘shape of the earth: sides differ’ journalism. Its merit is that it appears not to take sides, but as this extended name is meant to demonstrate, it is certainly not scientific. It is the kind of journalism that says the claim that £350 million a week goes to Brussels and could be spent on the NHS is ‘contested’, rather than simply untrue. In that sense, it can be uninformative and misleading, whereas scientific reporting is informative and is not misleading. Here is a twitter thread from Eric Umansky on a particularly bad example from the New York Times. Of course acrobatic journalism is easier and keeps the journalist out of trouble.

One of the side effects of acrobatic journalism is that it typically defines the two sides it wishes to balance. It therefore tends to be consensus journalism, where the consensus is defined by the politicians on either side. To see why this is problematic you just need to look at how Brexit is discussed and reported by the BBC since the referendum.

I began writing this post during the debate surrounding Nick Robinson’s Steve Hewlett Memorial Lecture. It is certainly strange for that debate to focus on outfits like The Canary, rather than the elephants in the room that produce political journalism to millions every day, who also tend to criticise the BBC whenever they get the opportunity. Yet the copy from these newspapers, and not The Canary, is regularly discussed by the broadcast media. The emergence of left social media journalism is a result of the consensus defining by-product of acrobatic journalism, which for a year or more defined the other side as the PLP rather than the Labour leadership.

I suspect many journalists would say that my idea of them being an amateur scientist is just impractical in this day and age, when they have so little time and resources. But what I have in mind (journalism as amateur scientists) is not very different from what journalists on the Financial Times do day in and day out. Chris Cook is an example of a journalist working in the broadcast media who does the same. But it is wrong to blame individual journalists for being more acrobatic than scientific, because the institutions they work for often demand it.

Nick Robinson’s lecture is much more nuanced and interesting that the subsequent media discussion would suggest. For example he identifies the problem with the way Facebook selects news that is discussed in more detail by Zeynep Tufekci in this TED talk. But there are two elephants in the room that he fails to discuss: the role of the increasingly politicised right wing press I have already mentioned, and the conflict between scientific and acrobatic journalism, both of which he praises without addressing the conflicts between them. [1]

[1] There is a clear example of this in the comments he recalls making on the Brexit debate just before the vote. He proudly says he called the £350 million claim untrue, but he then adds

“I did, incidentally, also say that the Remain claim that every household in Britain would be £4,300 a year better off was misleading and impossible to verify.”

This is acrobatic journalism at its worse. Yes, the BBC did think the £4,300 figure was ‘misleading’, but only because they did not talk to an economist who would have told you it was not. It shows a failure to be a good amateur scientist. But worse that that, this clumsy attempt at balance puts the central claim of the Remain campaign in the same bracket as £350 million a week lie, which it certainly is not.

Avoiding political discourse

When PoliticsHome interviewed me about what the Blanchflower review of monetary policy might come up with (which will not report for some time), I couldn’t help but end with a plea.

“My biggest fear is that some people may try and put their political spin on anything we recommend, even when those recommendations come from an analysis of the technical academic literature. When I accepted an invitation to be on Labour's Economic Advisory Council, one or two people did suggest that this would damage my reputation as an economist. Given that being a member placed no restrictions on what I could do or say in public, or any obligation to support party policy, I thought it was an extraordinary accusation to make. Can you imagine a medic being told that they had damaged their reputation by advising policy makers about medical research?”

I used the example of medical research for reasons outlined here, and because I like comparing economists to medics.

As the economist Paul Romer has noted (see also here), political discourse is not like scientific discourse, and it is a problem when academics sometimes adopt a political way of thinking in their day job. In political discourse it is critical whose side you are on. If you are on one side anything that favours or helps the other side is presumed wrong. So when I agreed to be on Labour's Economic Advisory Council (EAC), many of those that were opposed to Corbyn’s leadership naturally assumed that I must be a Corbyn supporter. When I said that if George Osborne set up an equivalent of the EAC and invited me I would agree, and that I had indeed given advice to his Treasury and his advisors, the accusations changed from being the enemy to aiding and abetting the enemy. [1]

We were aiding the enemy because I and other EAC members are ‘providing legitimacy’ to the new leadership: we are being used for our reputations. And sure enough, here is John McDonnell doing exactly that in a recent speech:

“We’ve enshrined these commitments in our Fiscal Credibility Rule, drawn up with help from the world-leading economists on our Economic Advisory Council.”

Except that is exactly what happened: I gave a paper at the first EAC on what I thought the fiscal rule should be, it was discussed there, McDonnell’s team then developed it internally, discussed their version of it with Labour’s shadow cabinet and the rule was made public.

Labour’s new fiscal rule, as far as I know, is the first fiscal rule that sensibly responds to the possibility that monetary policy can run out of reliable ammunition (QE is much less reliable than fiscal policy), and I hope that sets an example that others will follow. [2] I also think politicians should get credit for listening to economists and adopting sensible rules or creating useful institutions. That is exactly what Gordon Brown did in 1997/8, and what George Osborne did in 2010 in establishing the OBR.

In my view the EAC is a useful innovation in economic policy making, because the link between academic economics and policy has become weaker in recent years for various reasons. The EAC is unusual in part because it makes the politician who set it up vulnerable. Any member of the EAC can, if they wish, publicly criticise Labour policy and get additional media coverage as a result of EAC membership. And academics being ideas people rather than political people might be particularly prone to do that. That is a level of vulnerability that many politicians would avoid. (See this for example.) That is one reason why it is misguided to suggest that we should have given our advice in private rather than through a public body like the EAC.

Some who are enmeshed in political discourse say that policies are two a penny and can be put in place in an instant: it is only leadership and winning that matters. I suspect much the same is true for many (not all) political commentators, who often see policies as simply weapons to attract voters. As Gaby Hinsliff remarks, many political commentators have a bias to those they see as winners. But I don’t want a world where academics only give policy advice to politicians who they support or who they think will win. It politicises economics, and that can damage the credibility of the subject.

Because academic economists, or academics of any kind, give advice on their policy area to politicians should never have to mean that those academics support those politicians. People enmeshed in political discourse find it hard to understand that, but I refuse to accommodate them. My job is economic policy, and I while I will do what I can to try and get politicians to adopt policies based on sound economics, those means do not include adopting a political discourse.

[1] For the record, no one in the Labour Party from 2010 onwards asked me for my advice on any occasion. This was despite (or maybe because of) everything that I wrote opposing George Osborne’s austerity policies and defending the Labour government's fiscal policy on my blog (which started at the end of 2011). Had they done so, I would have gladly given it.

[2] Those who might be tempted to see the zero lower bound knockout as just a 'loophole', you just betray your mediamacro mindset. Those tempted to say anyone on the left would have adopted this type of rule, go read some MMT.