Tag Archives: statistics

On not skipping over facts

Posted on by

I read recently Sarah Bakewell’s delightful book, [amazon_link id=”009948515X” target=”_blank” ]How to Live: A Life of Montaigne in one question and twenty attempts at an answer[/amazon_link]. Then by chance today (possibly via @Storythings?) came across this quotation from his [amazon_link id=”0140446044″ target=”_blank” ]Complete Essays[/amazon_link]:

“I realize that if you ask people to account for “facts”, they usually  spend more time finding reasons for them than finding out whether they  are true. … They skip over the facts but carefully deduce inferences.  They normally begin thus: “How does this come about?” But does it do so?  That is what they ought to be asking.”

Quite so. The more time I spend thinking about it – and it’s quite a lot already – the more puzzling economic facts seem to be.

[amazon_image id=”009948515X” link=”true” target=”_blank” size=”medium” ]How to Live: A Life of Montaigne in one question and twenty attempts at an answer[/amazon_image]  [amazon_image id=”B002RI92VQ” link=”true” target=”_blank” size=”medium” ]The Complete Essays[/amazon_image]

Accounts and holding to account

Posted on by

This is the last of my posts drafted on holiday last week, and is particularly timely despite being retrospective because my review of Jane Gleeson-White’s [amazon_link id=”0393246671″ target=”_blank” ]Six Capitals, or Can Accountants Save the Planet[/amazon_link] is published in the new issue of Foreign Affairs today.

I *loved* Jacob Soll’s [amazon_link id=”0718193628″ target=”_blank” ]The Reckoning: Financial Accountability and the Making and Breaking of Nations[/amazon_link]. It gives a long historical perspective on accountancy – no, wait – and particularly its importance in literally holding leaders and politicians to account. It is also extremely well written and lively. I got much joy from learning of the 1604 book Accounting for Princes, and that there was a real Musketeer called Captain D’Artagnan, he isn’t just a [amazon_link id=”1853260401″ target=”_blank” ]creature of fiction[/amazon_link]. That Jacques Necker’s [amazon_link id=”1246639963″ target=”_blank” ]Compte Rendu[/amazon_link] sold more than 100,000 copies in one year, 1781, alone. And this description of Josiah Wedgwood’s legacy: “Sound industrial management and tableware for the middling sort.”

[amazon_image id=”0718193628″ link=”true” target=”_blank” size=”medium” ]The Reckoning: Financial Accountability and the Making and Breaking of Nations[/amazon_image]

The book’s message is that there is a constant tension between the increasing sophistication of methods of accounting to hold to account, literally, kings or powerful companies or political leaders and the scope that sophistication creates for new ways of defrauding the people or the shareholders. So the methods of accountancy need to be embedded in a culture of honest dealing. Soll concludes that we lack that now: “Accountants have become separated from everyday culture. … All countries, rich and poor, hide the true costs of their pension benefits and health care as well as of infrastructure, off their balance sheets.” But there is no public outcry about bad public accounting, while deceitful bankers and financiers have not been held to account for the crisis. (One of my arguments in [amazon_link id=”B004NNUWY4″ target=”_blank” ]The Economics of Enough[/amazon_link] was that this situation is seriously unsustainable.)

Soll cites the proposal by Timothy Irwin of the IMF that governments should publish 50 year ahead balance sheets, but queries whether it is possible. “By separating finance into its own sphere, we have lowered our financial and political aspirations.” The numbers should be an integral part of society but are not: “If there is any historical lesson to be learned here, it is that those societies that managed to harness accounting as part of their general cultures flourished.” These have been rare – Renaissance Italian city states, the early American republic, Britain of the Glorious Revolution or for a time in the Industrial Revolution. They have been brief: Charles Davenant, one of the earliest national income accountants, tried to calculate the public finances after 1689 but by the time his Discourses on the Publick Revenues was published complained that information was being withheld.

My sole regret (as the author of [amazon_link id=”0691156794″ target=”_blank” ]GDP[/amazon_link]) about Soll’s book is that he does not cover economic statistics more generally, but rather public finances. They too matter for political accountability and are no longer doing that job (as I argue in my recent working paper). Still, it’s good to want more of a book rather than less. I highly commend [amazon_link id=”B00JK2H1ME” target=”_blank” ]The Reckoning[/amazon_link].

Final couple of days of holiday now. I have Sarah Bakewell’s [amazon_link id=”B00SLSG08O” target=”_blank” ]How To Live: A Life of Montaigne in one question and twenty attempts at an answer[/amazon_link]; and [amazon_link id=”077103850X” target=”_blank” ]Sapiens: A brief history of humankind[/amazon_link] left to read.

Generalizing about Africa (& why it’s a bad idea)

Posted on by

This morning I attended a breakfast at the Centre for Global Development at which Morten Jerven spoke about his new book, [amazon_link id=”1783601329″ target=”_blank” ]Africa: How Economists Get It Wrong[/amazon_link],” which I’m now looking forward to reading – especially after enjoying his previous book, [amazon_link id=”080147860X” target=”_blank” ]Poor Numbers: How We Are Misled by African Development Statistics[/amazon_link]. The new book is published today.

[amazon_image id=”1783601329″ link=”true” target=”_blank” size=”medium” ]Africa: Why Economists Get it Wrong (African Arguments)[/amazon_image]  [amazon_image id=”080147860X” link=”true” target=”_blank” size=”medium” ]Poor Numbers: How We are Misled by African Development Statistics and What to Do About it (Cornell Studies in Political Economy)[/amazon_image]

His talk kicked off with a critique of two successive approaches to ‘Africa’ by economists – a category refined in discussion to macroeconomists, and not absolutely all of those. The first approach was cross-country regression analysis, given its definitive shape by [amazon_link id=”0262522543″ target=”_blank” ]Robert Barro[/amazon_link], in which a dummy variable for African countries would have a negative coefficient. (Charles Kenny has an excellent critical review of growth regressions, and Dani Rodrik has a terrific paper on the limitations of trying to get policy prescriptions from this approach.) The second approach takes low growth in African countries as a given and tries to pin that to institutional failures – corruption, clientilism, lack of transparency etc – and is perhaps symbolized by Acemoglu and Robinson in their ambitious [amazon_link id=”1846684307″ target=”_blank” ]Why Nations Fail[/amazon_link]. Morten Jerven said, “There are economists who have written non-modest books.” He argued that the latter approach leads to policy prescriptions of the kind that ask, “Why aren’t you like Denmark?”

[amazon_image id=”0262522543″ link=”true” target=”_blank” size=”medium” ]Determinants of Economic Growth: A Cross-country Empirical Study (Lionel Robbins Lectures)[/amazon_image]  [amazon_image id=”B007HLIUN4″ link=”true” target=”_blank” size=”medium” ]Why Nations Fail: The Origins of Power, Prosperity and Poverty[/amazon_image]

Instead, he suggests focusing on trajectories from today’s specific situation, rather than seeking to explain deviations from a rich world ‘norm’. He argued too for, “Fingertip knowledge of each country’s data.” Yes!! The standard international datasets lead researchers to think history started in 1960, and also are misleading because they interpolate or otherwise guess to fill in the many gaps “They should leave gaps if there are no data – or at least put them in a different colour,” Jerven says.

There was some challenge from the attendees. One said that there clearly is something distinctive to explain about most African economies, where people are still poor on average. Another argued that most economists in the development field do recognise the need for a rich description of specific economies, so the book attacks a straw man. However, I think Jerven is right to highlight the cavalier approach far too many economists have to the data, and their tendency to generalise. Surely he is right when he says, “A paper about ‘Africa’ gets more citations than a paper about ‘Tanzania’.”?

And how valid are those generalisations, when you consider the following Economist covers from 2000 and 2011. Was there really so much change in just over a decade?

Hopeless?

Hopeless?

Or Rising?

Or Rising?

Angry statisticians and fiddled figures

Posted on by

Merijn Knibbe (@MerijnKnibbe) alerted Twitter yesterday to an extraordinary statement on the website of ELSTAT, the Greek official statistics agency. It was issued by the Members of the European Statistical System – the professional group of official statisticians in Europe  – and includes this statement:

“Therefore, we confirm our concern with regard to the situation in Greece, where the statistical institute, ELSTAT, as well as some of its staff members, including the current President of ELSTAT, continue to be questioned in their professional capacity. There are ongoing political debates and investigatory and judicial proceedings related to actions taken by ELSTAT and to statistics which have repeatedly passed the quality checks applied by Eurostat to ensure full compliance with Union legislation.”

The story – told in the opening pages of my book [amazon_link id=”0691156794″ target=”_blank” ]GDP: A Brief But Affectionate History[/amazon_link] – is that at the start of the Greek crisis, one of the most benign conditions required by the IMF was that the Greek government stop fabricating its GDP statistics, which it had been doing for some years in order to keep the loans flowing. The EU statistics body had refused to approve the statistics but international lenders (hello Goldman Sachs) didn’t seem to mind.

So the Greek statistical agency was dissolved, the new one (ELSTAT) created, and a former IMF economist, Andreas Georgiou, was appointed to lead it. One of his graduate school friends told me Mr Georgiou is one of the most honourable people he has ever known. Yet, almost immediately after his appointment to the job, some of the sacked former board members accused him of treason for cleaning up the Greek statistics and brought legal proceedings. “I am being prosecuted for not cooking the books,” Mr Georgiou said at the time. The continuing legal and political shenanigans are what the new statement refers to.

The independence and integrity of official statistics really matters. We take economic statistics far too seriously in one sense, often ignoring the margins of error and the judgements involved in their calculation (so it’s encouraging to see a vigorous debate about these issues), not to mention the fact that the categories we define are social constructs. Yet independent and freely available official statistics are a vital part of the fabric of a democracy, one of the key tools for holding governments to account – see my new working paper on this. The only OECD country moving away from independence for its official statisticians has been, bizarrely, Canada; all others have moved in the opposite direction. The statisticians’ statement this week about Greece does not augur well for how things there will turn out.

[amazon_image id=”0691156794″ link=”true” target=”_blank” size=”medium” ]GDP: A Brief but Affectionate History[/amazon_image]

Not all numbers are equal

Posted on by

I’ve started reading Tony Atkinson’s new book, I[amazon_link id=”0674504763″ target=”_blank” ]nequality: What can be done?[/amazon_link] and already think it a much better book than the famous [amazon_link id=”067443000X” target=”_blank” ]Capital in the 21st Century[/amazon_link] (which for me was marred by the half-baked r and g business – see for example the Jaume Ventura slides here – as well as the lack of any practical policy suggestions).

[amazon_image id=”0674504763″ link=”true” target=”_blank” size=”medium” ]Inequality[/amazon_image]

Although not far into [amazon_link id=”0674504763″ target=”_blank” ]Inequality[/amazon_link], I completely and utterly agree with the following, in a chapter describing carefully the sources and character of the data (something else on which Piketty is actually rather weak – hence the challenge much reported this week from a graduate student at MIT):

“In seeking to draw lessons from the statistics on inequality, we have to be confident in the quality of the data we are using. This is why I begin this chapter by describing and evaluation the sources of evidence on which scholars of inequality can draw. Such scrutiny is essential. All too often economists race ahead, drawing conclusions from figures that happen to be there, without asking why the data are suitable.”

Serendipitously, while reading this I’ve also been thinking about a keynote I’m giving soon at an OECD conference on [amazon_link id=”0691156794″ target=”_blank” ]GDP[/amazon_link] and the national accounts statistics in a couple of weeks’ time. All the thousands of studies and political claims resting on GDP growth figures are based on shifting sands, and we economists need to think far more carefully about what they take to be evidence for strong claims. There are some powerful examples in a paper presented by Samuel Williamson and Enrico Berkes recently at the Economic History Society conference.