The Treasury White Paper on the economic impact of leaving the EU is based on unsound economics and does not withstand close analysis
Noble lies have been part of political activity since the time of Plato. In the debate ahead of the European Union referendum, it may have been inevitable that the pro-EU Cameron government would issue statements claiming that leaving the EU would make people worse-off. It may even have been inevitable that such statements would include high numbers for the alleged cost of Brexit. After all, the higher the numbers, the more likely it is that under-informed and undecided voters will become frightened, and that Project Fear will work by terrifying people so that they opt to stay in. Perhaps no surprise should have greeted the publication of alarmist numbers in the Treasury White Paper on The long-term economic impact of EU membership and the alternatives. At the related press conference on April 18, Osborne quoted from the command paper to claim that, on purportedly plausible assumptions, Brexit would cost the average British household £4,300 a year by 2030.
But what is surprising — and was not inevitable — is that Osborne and the Treasury should have done the job so ineptly. Plato may have been right that governments must sometimes be mendacious, to maintain unity in wartime, to increase alertness ahead of possible terrorism or whatever. But lies need to have some nobility, with enough slickness in presentation and credibility in substance, if they are to mould the public debate. The White Paper was soon trashed in quality Conservative-inclined publications, exactly those organs of opinion that a Tory Chancellor ought to be able to influence. Fraser Nelson, editor of the Spectator, described himself as a Europhile, but said that Osborne’s dishonesty was “simply breathtaking” and was such that he might vote “out”. Allister Heath in the Daily Telegraph was even more damning. He recalled the deception and trickery that took Britain into the Iraq War, and said that the Treasury’s “dodgy dossier” was “beneath contempt”.
The sheer badness of the document raises questions about the role of the civil service in the government of modern Britain, specifically about the ability of a permanent and supposedly non-partisan civil service to produce trustworthy documents in a politically-charged environment. Britain does not have an explicit written constitution, but it certainly has a number of implicit constitutional understandings.
In the past one of these was that command papers (presented to Parliament by Her Majesty’s command, don’t forget) met certain standards of factual objectivity and reliability. Quite simply, this command paper is not up to those standards. Arrangements need to be put in place to ensure that they are restored. The paper does not have a consecutive and easy-to-follow argument that connects the facts of the real world, and the Treasury’s interpretation of those facts, with Osborne’s £4,300 figure. Much of the material is so impenetrable that many readers might think the Treasury is deliberately trying to put them off. Nevertheless, the White Paper does have a logical argument running through all the padding and obfuscation.
An obvious truth is that since the Industrial Revolution living standards have improved as regions have become more interconnected within nations, and as nations have become more interconnected with each other through global trade and capital flows. Living standards depend on how much each person produces or “output per head”, also known as “productivity”. By implication, productivity is related to “openness”, the degree to which regions and nations are interconnected with each other. Further, a fair conjecture is that trade agreements between nations increase openness. Such agreements include the World Trade Organisation and also, more fundamentally in the Treasury’s view, the European Union.
It follows that the EU makes the UK better connected with its European neighbours, which boosts productivity, which helps living standards. It also follows, according to the paper, that the UK’s withdrawal from the EU would undermine trade with the remaining EU members, which would depress productivity, which would damage living standards. In the White Paper’s words, “the key transmission channel through to the economy in the long term comes from the impact of reduced openness, both from trade and foreign direct investment, on productivity. This is the main driver in the estimates of the long-term effects of EU membership on the economy.”
The estimates need of course to be linked to facts and data, and organised in a model which is able to identify the relative power of different forces. The analytical tool at hand is what the Treasury calls “the widely adopted gravity modelling approach”, pioneered in a 1962 paper by Jan Tinbergen. (Tinbergen, a Dutchman, was one of the first two Nobel economics laureates. In 1969 he shared the prize with the Norwegian, Ragnar Frisch.) In the gravity modelling approach, openness — measured by the ratio of trade between two or more countries to their national outputs — is related to certain relevant variables, such as the distance between the countries and the extent of any colonial ties. Of course, the trade-output ratio is likely to be higher the closer are two countries geographically and politically. That is just common sense. The subtlety of Tinbergen’s method is that it enables analysts to obtain numbers on the size and reliability of the effects.
How, then, do economists assess the importance of trade agreements (read: the WTO, the EU, the European Free Trade Area) relative to such factors as distance and colonial ties? The answer is that they use a statistical technique in which a major trade agreement is treated as “a dummy variable” in the dataset from which the equations are estimated. As the Treasury explains in the White Paper (page 159 for those who are interested), the relevant dummy variable here is set equal to one if both countries in the estimation exercise are EU members and at nil otherwise. By this device, the period before 1973 in which the UK was not an EU (or European Economic Community) member can be distinguished from the period after it. The results of the exercise measure by various well-known statistical tests whether any change in the UK economy’s openness after 1973 was significant, and so — presumably — worthwhile for productivity and living standards.
I realise that I may have lost readers in the last few paragraphs. Let it be said straightaway that the introduction of dummy variables is a commonplace method in statistics and is perfectly legitimate, even if it sounds like gobbledygook. Let it also be said that — superficially — the Treasury has been careful and rigorous in its work, with the pages bedizened by regression coefficients, references to “omitted variable effects” and such like, and numerous citations of academic articles.
Now let me explain why, despite all the effort invested in this project by possibly a dozen or so officials over many months, it remains a contemptible dodgy dossier. First, if the Treasury genuinely believes that openness in the sense of EU membership affects productivity, it is not necessary to involve Tinbergen’s gravity model at all. For a few decades the UK’s main statistical agency, now known as the Office for National Statistics, has prepared data on the productivity of our economy. All that needs to be done is to check whether EU/EEC membership had any positive effect on productivity growth, going directly to the time series in official publications.
Output per person employed in UK manufacturing industries
I set out the key numbers in the accompanying table (above), where the figures are taken from Economic Trends: 1994 Annual Supplement, prepared by the Government Statistical Service. I have chosen productivity in manufacturing rather than the economy as a whole, as manufacturing production was in the relevant period (and remains today) highly traded between nations. Its trade intensity implies that it ought to have been that part of the economy where the argument of the White Paper was most applicable. So productivity growth in manufacturing ought — if the Treasury’s core thesis were right — to have been higher in the years from 1973 than before. On the face of it the very first year of EU membership, 1973, fits beautifully with the Treasury thesis. It enjoyed the highest productivity growth (8.6 per cent) in any of the 22 years included in the table.
However, that one number is not conclusive. A well-known cyclical pattern in all economies is for productivity growth to be related to the growth of output. Productivity moves ahead most briskly when companies are full to the brim with orders, and sales and output are expanding so vigorously that a given labour force can be worked to the hilt. 1973 was a boom year for the world and UK economies. Productivity would have surged whether the UK had joined the EU or not.
A more secure conclusion comes from looking over periods of five or ten years. As the table brings out, productivity growth in the second half of the 1970s was poor. In the decade to 1972, output per head in UK manufacturing increased by 45.3 per cent (or at a compound annual rate of 3.8 per cent); in the first decade of EU membership, which includes all the years from 1973 to 1982, output per head in UK manufacturing went up by 19 per cent (or at a compound annual rate of 1.8 per cent). In short, productivity growth in the area of the economy that should have benefited most from EU accession decelerated afterwards. In the first decade after joining the EU it ran at less than half the rate seen in the previous decade.
The point I have just made is so crushing — and so easy to derive from official data and simple enough to explain — that the rather convoluted argument in the White Paper (that productivity growth depends on openness, that openness can be ascribed to membership of trading arrangements and measured in the Tinbergen gravity approach by the insertion of dummy variables in datasets, and so on) deserves to be binned. That begs the question of how apparently intelligent and well-intentioned officials could have thought that their circuitous and indirect method would be worthwhile, and of how the outcome of their method is so at variance with the obvious central facts of the case. I am not sure what has happened, but we must hope that they have not tweaked the assumptions and massaged the calculations in order to support Osborne’s £4,300 number. (Dear reader, the world is a wicked place.) Part of the trouble may stem from the Treasury’s use of an inappropriate and misleading measure of openness. This is my second point.
The UK’s openness: Total trade as a share of GDP. Both series show the sum of exports and imports as a percentage of GDP
The first chart in the White Paper (on page 19 of the introduction) shows the ratio of trade to the UK’s GDP, from 1940 to today. It shows a dramatic surge, from less than 15 per cent during the Second World War to over 60 per cent at present, and seems to give credibility to the claim in the text that openness is good for productivity and growth. The chart is sourced to the ONS “using Bank of England calculations”, and a footnote mentions a dataset, “Three centuries of data”, that anyone can download from the Bank’s website. I have downloaded the data and have had no difficulty in replicating the White Paper chart, which in this minimal sense is “the truth”. The footnote spells out the meaning of the trade-to-GDP ratio, saying it is “the sum of real exports and imports divided by real GDP in 2012 reference prices”. I have italicised the word “real” and the phrase “in 2012 reference prices” because — as we shall soon see — it is these technicalities that give the game away.
A short digression on the nature of economic progress is needed to understand the Treasury’s howler. Production is of two main kinds, goods and services. Economic history is to a large extent about the application of human skill and ingenuity in the production of goods, so that in advanced nations output per person in goods production (much of it manufacturing) has risen astonishingly, often by more than 50 times, since the start of the Industrial Revolution. The production of services has also seen remarkable efficiency gains, but to a lesser degree. Certain types of medical care are much the same today as in the 18th century, but the most extreme case might be hairdressing. With all due respect to hair stylists and the inventors of electrically-powered clippers, a haircut in 2016 is not that different from a haircut in 1816.
But the labour market is competitive, so pay (after adjusting for experience, qualifications, physical strain and so on) must be the same in both goods production and services. If wages per worker are the same in goods and services production while productivity growth is much more rapid in the manufacture of goods than in the supply of services, the price of goods must over the decades fall relative to the price of services. This pattern has in fact been observed in all dynamic economies over the long run. The price of a cotton shirt or a television set has risen much less since the Second World War than the price of nursing care or a haircut.
Now we come to the nub of the matter. The overwhelming majority of goods can be traded between regions and nations, whereas that is true of only a proportion of services. For practically everyone their doctors, nurses and hairdressers are local, whereas they can buy a Japanese car, a shirt made in Bangladesh or a television set from Turkey. Manufactured goods are both heavily traded compared with services, and fall in price relative to services over the long run. An inevitable accompaniment of economic growth is therefore that the volume of trade rises relative to the volume of national output, because output as a whole must have a big non-traded services component. The apparent rise in the ratio of trade to output in the White Paper is largely attributable to differential rates of technological change and the resulting movement in the relative prices of tradable goods and non-tradable services; it is not necessarily due to trade liberalisation and economies’ increased openness at all.
Let me put the matter another way. To repeat, the first chart in the White Paper acknowledges in a footnote that its figures are in real terms and in 2012 reference prices. In other words, the actual money values of exports and imports in, say, 1963 are not being compared with the actual money value of GDP in 1963. No, the Treasury is using a data series from the ONS where the ONS has adjusted the money values of 1963 for the change in prices between 1963 and 2012, and then divided the resulting volume figures for the sum of exports and imports by the volume figure for GDP. Because of relative price shifts over time, because the price of tradable motor vehicles, textiles and electronic goods has collapsed compared with the price of non-tradable nursing care and haircuts, the Treasury can offer a chart in which the trade-to-GDP ratio has soared.
But the Treasury’s chart does not correspond in the slightest to the summed values of money transactions in real-world exporting and importing. It is instead an artefact of statistical manipulations carried out by officials inside both the ONS and the Treasury. Contrary to the message the Tresury wants to give, the increase in the trade-to-GDP ratio (as they have presented it) may owe nothing to the UK’s economy’s greater openness.
I have obtained numbers for the values of the UK’s exports, imports and GDP, going back to 1948. In the accompanying chart (above) I compare the sum of the values of exports and imports divided by the value of GDP with the sum of the volume of exports and imports divided by the volume of GDP. The blue line is the same as that in the White Paper, which is an almost unbroken upward curve. The red line is the one I have prepared from actual values of imports, exports and GDP in the prices of the day. My line does show that the value of trade has increased, relative to the value of GDP, since the late 1940s and — let it be admitted — it does share this characteristic with the line taken from the White Paper.
However, anyone can see that the change in the UK’s openness over the six-and-a-half decades is very different in extent and timing on the two approaches. According to the value series, the UK’s openness in 1951 — after the 1949 devaluation and the Attlee government’s export drive, and before the spread of protectionism in the UK’s former colonial empire — was greater than in the early 1990s, 20 years after EU/EEC accession. At any rate, if Treasury officials plugged the volume numbers into their gravity model (as they seem to have done), they would have overestimated — in a grotesque and absurd way — the importance of EU/EEC accession to the UK’s economic dynamism.
My final point is that the White Paper takes it for granted that the UK’s continued membership of the EU is less protectionist (and hence more open, with all the benefits arising from openness) than any alternative. That is just not so. The EU’s ambition may be to have complete free trade in goods and services between its member states, and much has been achieved in that direction. But the EU’s supporters and functionaries are far from agreed that the EU should be a champion of free trade in the global context. On the contrary, the EU operates a “common external tariff” (i.e. the same tariff rate in all EU member states on imports from outside the EU) to put a commercial divide between it and the rest of the world. The tariffs vary from product to product, with the European Commission saying on its website that, “The rates depend on [their] economic sensitivity.”
The advantages of real free trade. How Singapore and Hong Kong outperform the EU and the UK
“Sensitivity” is a tease of a word. Is an import into the EU “sensitive” if foreign production costs and hence price are lower than the EU’s production costs and prices, so that tariff-free imports would damage EU producers? If so, the motivation for the common external tariff is protectionism, pure and simple. On this basis EU membership is the antithesis of international openness. As Gerard Lyons, economic adviser to Boris Johnson, said in a recent article in The Times, genuine free trade — the acme of openness — means the unilateral dismantling of all restrictions on imports. In other words, if the UK really wants to pursue openness in its international economic relations, it should quit the EU, and put itself in the same position as its former colonies, Singapore and Hong Kong. These two city-states, which have no barriers on imports at all, are now not only richer than the EU, but are also growing more rapidly.
George Osborne is honest in one respect. He may in some sense be supposed “to run the economy”, but he has never pretended to have a deep knowledge of economic theory. Instead, he is proud to be “political to the fingertips”. The justification for his Cabinet position is his political nous, his reputed skill in the tactics of rhetoric and persuasion (spinning press releases, lunching journalists, and the like) that will keep himself and his colleagues “at the top”.
Maybe so, but the soundbite at his April 18 press conference — a cost to every household of £4,300 by 2030 — has been justly derided as well as widely quoted. Cynics noted that someone who could not forecast his budget deficits more than six months into the future is hardly credible in projections about 2030. But the Treasury’s analysis of the cost of the Brexit is much more disgraceful than missing a deficit forecast. A politician has used his department to concoct a Big Number/Lie and relied on headline-seeking newspapers to repeat the Big Number/Lie so that it acquires the status in public debate of “a factoid”. For those new to the word “factoid”, one dictionary definition is “something fictitious or unsubstantiated that is presented as fact, devised especially to gain publicity and accepted because of constant repetition”. Politicians must have the power to govern of course, but both the Treasury White Paper and the dodgy dossier (“weapons of mass destruction”) behind the Iraq War argue that restraints are needed over a government’s ability to issue misleading or false statements. Lies may be noble in intention but ignoble in practice, and too much economy with the truth destroys respect for officialdom in all its forms. It is possible that every senior Treasury mandarin believes every word in the White Paper, but I doubt it.
The Treasury’s reputation for objectivity and rigour as the UK’s premier department of state is at stake here. Top civil servants should have a recognised procedure to enable them to distance their own views from politicised rubbish put our in their ministries’ names, particularly when such rubbish appears in a command paper.