Feb 06
Don’t mention the (Franco-Prussian) war… or what French payments to Germany in the 1870s tell us about the euro crisis!

Michael Pettis, the Beijing-based economic theorist, has written a magisterial blog post that Matthew C. Klein, the American historian/ financial commentator, regards as 'literally the best analysis of the euro area's problems we've ever read'. Likewise, I was blown away by the power of Pettis' argument, but I must pass on this issue until I have finished Philippe Legrain's 'European Spring: Why our Economies and Politics are in a Mess and How to Put Them Right' and at least started Yanis Varoufakis' Europe after the Minotaur: Greece and the Future of the Global Economy.


Pettis' piece is a blog post but, at some 25 pages long, not as we know them.  It was crying out for a summary. Happily Matthew C. Klein has provided an excellent one in Alphaville in the FT: Michael Pettis explains the euro crisis (and a lot of other things too). Of course, I urge you to read Pettis or, if you don't have the time, to read Klein extensive summary of the Pettis post (accessing the Financial Times is less troublesome than you might think, but as I have a sub, I don't know what precise hurdles you might face).


Those as yet unconvinced that either is worth their time should take note of the following particular nuggets to be found among all the gold-standard writing.


1. The European debt crisis is not a conflict among nations…the imbalances that led eventually to the current crisis had their roots in hidden transfers between different economic sectors within Europe and not between countries…Resolving a debt crisis involves nothing more than assigning the losses. In the current crisis these costs have to be assigned to different economic sectors within Europe, but to the extent that the assignation of costs can be characterised as exercises in national cost allocation, it is easy to turn an economic conflict into a national conflict.


2. Until now, an awful lot of Europeans have understood the crisis primarily in terms of differences in national character, economic virtue, and as a moral struggle between prudence and irresponsibility. This interpretation is intuitively appealing but it is almost wholly incorrect, and because the cost of saving Europe is debt forgiveness, and Europe must decide if this is a cost worth paying (I think it is), to the extent that the European crisis is seen as a struggle between the prudent countries and the irresponsible countries, it is extremely unlikely that Europeans will be willing to pay the cost. (My emphasis)


3. Except in Greece, in Europe the main political parties on both sides of the political spectrum have until now chosen to maintain the value of the currency and protect the interests of the creditors. It has been extremist parties, either on the right or the left, who have attacked the currency union and the interests of the creditors. In many cases these parties are extreme nationalists and oppose the existence of the European Union. (Not surprisingly he prophesizes doom if these take control of the debate.)


I appreciate, these points might come over as assertions that you either accept or do not. But they are accompanied by serious historical analysis linking the deleterious effect on German industry of the reparations paid by France after the Franco-Prussian War to the calamitous effect on Spain of recent resource transfers from Germany.  The former he claims were the largest reparations in history and equivalent to 23% of France's annual GDP (20% of Germany's). But he goes on to point out that the resource transfer (much of it in the form of loans from German banks) to Spain between 2000 and 2009 (in the form of loans, of course) was greater, coming in at 31-32% of annual GDP. His crucial point is that 'The great beneficiary of French "largesse" turned out not to have benefited any more than Spain had benefited from German largesse 135 years later'.


This rather turns the table on the usual narrative. While he is careful not to assert that the problem was created by German outflows as opposed to Spanish inflows, he does point out (quite rightly to my mind): 'If it were the latter case, however, it would be an astonishing coincidence that so many countries decided to embark on consumption sprees at exactly the same time. It would be even more remarkable, had they done so, that they could have all sucked money out of a reluctant Germany while driving interest rates down. It is very hard to believe, in other words, that the enormous shift in the internal European balance of payments was driven by anything other than a domestic shift in the German economy that suddenly saw total savings soar relative to total investment.'


This helps scotch another false narrative – that of competitiveness. Here, cued up by Pettis, Michael C. Klein nails it.

For all of the suffering that has occurred in places such as Spain, Ireland, and Greece, we shouldn't forget that German workers have suffered from stagnant wages and decaying infrastructure.

One of the worst costs — for Germany — has been the lack productivity growth. For all the talk of Teutonic competitiveness, German labour productivity has grown at the meagre pace of just 0.6 per cent per year, on average, since 1998. Output per hour worked is actually lower now than it was in 2007. For perspective, this track record is worse than that of practically every other rich country — including Greece and Spain!

(Source: Organisation for Economic Co-operation and Development, author's calculations)


Happily, Pettis is not only in the business of deconstructing false narratives. He also suggests solutions. Along with the Greek Finance Minister Varoufakis, he does not see the issues in terms of zero-sum games. Both are enamoured with linking repayments to growth. The congruence of approach between the two is such that it is not surprising that Pettis states 'Varoufakis should really take the lead in designing an entirely new form of sovereign debt restructuring, not just for Greece but for the many countries, in Europe and elsewhere, that will soon follow it into default'.

If the Greek crisis were not so serious, for Greece and for Europe, the following quip might have been amusing:  'when someone like Yanis Varoufakis proposes that there are ways in which partial debt forgiveness increases overall economic value, instead of merely creating moral hazard, worried economists often recoil in horror, while finance or bankruptcy specialists (and an awful lot of hedge fund managers) shrug their shoulders at such an obvious statement'.

I hoped this has whetted your appetite for both pieces.​

Jun 04
105,000 jobs lost in Hertfordshire, 62,000 in Essex!

​In my last post I illustrated some shortcomings of employment forecasts at the district council level. I argued that the problem was poor data. I just happen to have compared the employment figures used by the forecasters (those from the Business Register and Employment Survey – BRES) with the more reliable figures recently released from the Census of Population. The results are truly astonishing.

First, though, a bit of history. The 2001 Census revealed that in Hertfordshire, in particular, the predecessor of the BRES  grossly exaggerated employment levels. Reliance on this flawed data by consultants working for Hertfordshire County Council (Bone Wells) and the East of England Regional Assembly (Roger Tym) caused mayhem until necessary corrections were made. The direction of net-commuting was thought to be out of London rather than into London! You can imagine the problems this caused for discussions of house building and employment targets in the East of England.

The shortcomings of the business survey data were forcefully brought to ONS' attention by Jill Tuffnell, who preceded me on their Central-Local Information Partnership Labour Market Statistics Group. Thanks to such campaigning, ONS began to show greater willingness to involve local authority experts, such as Jill and me, in 'cleaning' their data. But this ended with the introduction of the Business Register and Employment Survey in 2008. This 'new, improved' product meant there was no need for costly, time-consuming consultation!

While I did not believe ONS was right to have such faith in the BRES, I am really astonished that analysis of the 2011 Census versus the BRES reveals the same order of problem we had ten years ago. And in much the same places.

Averaging  employee numbers from the 2010 and 2011 BRES (taken in September) and comparing these with workplace employees from the March 2011 Census shows the BRES figures to be 25% higher than the Census ones in Hertfordshire, 14% in Essex, 10%-12% in Bedfordshire but only 6%-8% in the East Anglian counties. This is not exactly comparing like for like as the over-74s, armed forces and double-jobbing complicate matters. But not that much and certainly not so differently in the various counties in the region.

I wonder whether ONS will listen now. For their part Oxford Economics (see previous post) are admirably open to local authority help in cleaning the BRES data they rely on, but without ONS support little improvement can be expected. In the meantime, I would urge local authorities to follow Epping Forest's example and get hold of the Interdepartmental Business Register to check what is being fed into the BRES. Otherwise they are likely to be taken by surprise every ten years.  


Jun 04
An astonishing record of complete failure
​The title is from this​ excellent FT piece on economic forecasting  recently penned by Tim Harford. 
I have always been sceptical of the utility of economic forecasts at levels as local as English district councils. The poor record that Tim highlights relates, of course, to national economies. So it may not be surprising that I can show the local forecasts coming from the East of England Forecasting Model (EEFM) as seriously flawed (which I will do in a moment).   I don’t believe this is because the forecasters (Oxford Economics) are poor (I have, in fact, a lot of time for them).  Essentially, it boils down to poor data. 
There are no Gross Value Added statistics and total employment in businesses is often attributed to a pay point where few people work, messing up its geographical pattern. And this poor data is a year old before it is available for ‘forecasts’! It would be nice to think that the Office for National Statistics does all it can to improve the quality of the data underlying such forecasts. Alas, it does not. In this centralised State there is not enough interest in getting the local data right. 
The Census of Population might be thought an exception to this rule if it were conducted more frequently than every ten years and it took less than three years for the crucial employment figures to appear! However, as we have just got this workplace-based employment data, now is an opportune time to compare the EEFM forecasts of ‘workplace populations’ at district level with the Census results.
In Autumn 2010, but a few months away from the March 2011 Census, the EEFM produced ‘high growth’ and ‘low growth’ variants of its forecasts. So you would think that there was a good chance of the forecast ‘workplace population’ being within the bounds of these forecasts would you not? 
Think again. In 36 out of out of 47 of the East of England’s districts the actual figure exceeded the ‘high growth’ figure and in 10 districts it was lower than the ‘low growth’ figure. In only two cases did neither case apply. (Before anyone points out 36+2=48, I know. But all the above is logically possible. Think of it as a brain teaser. A prize for anyone who solves it.)
The two districts the forecast got right were South Norfolk and South Cambs. Now this is puzzling as these are both high growth districts next to under-bounded cities. Just the sorts of places you might think where the model’s source data might be particularly inaccurate. I cannot explain this. But funnily enough by 2013 the model got its ‘forecast’ for both these places very wrong, spectacularly in the case of South Norfoik. There growth between 2001 and 2011 was 36% (the highest in the region) but the model had it at 47%. 
The full picture for the East of England districts is below. Not a pretty sight! 
1406 04 chart.bmp


May 19
Be prepared for failure!

​On 23 May ONS is due to release 'workplace population'  figures from the 2011 Census. 10 years ago the equivalent figures from 2001 caused a stir, showing up flaws in the employment figures from business surveys (then the ABI now BRES) and revealed surprisingly low employment figures in some parts. Nowadays, nobody should be surprised by lack-lustre growth in Herts. But still there will be surprises. Or there will be if you don't contact me! I have estimated 2011 workplace populations from publicly available Census data and can reveal e.g. ​​​that Cambridge has been motoring on since 2001 while Herts, Beds and Norfolk have been stalling. Growth rates in the Greater South East (over 10 yrs mind) range from -5% (a fair few new towns continue to suffer) to 37% (in a special 'suburban' case)  or 26% (in a more successful new town)  if you exclude the special 'under-bounded' case.

Sep 24
Just how many people do small businesses employ in UK?

​Chuka Umunna roused me from my slumbers ​​this morning with a startling claim: two thirds of private sector jobs are due to small businesses. ​Really? It turns out the figures are startling but not as startling as he claims. I suspect  the Shadow BIS Secretary​'s  source is the October 2012 BIS Statistical Release on Business Population Estimates for the UK and Regions 2012. Specifically t​his​ table: ​​​​​


Apparently he translated the oft quoted 'nearly two-thirds' to 'two-thirds' (the percentage is 59% - I leave you to judge) and 'small and medium enterprises' or (more accurately) 'businesses employing fewer than 250' to ​'small businesses'. The more conventional definition yields a 47% share.  So I would have said nearly half of private sector employment is due to small businesses. But maybe that's why I am not Shadow Business Secretary!

Given his evident  interest in the subject I hope that he will press for parallel figures to be made available for local authority areas. The UK figures include estimates of non-registered businesses e.g. your window cleaner. No local figures include this adjustment. It is a moot point whether the growth in such self-employment should be factored into any assessment of the importance of small businesses (for instance very few of these will pay business rates which is Chuka's issue today). But what, in my view is essential, is that when such estimates are bandied about we should have the opportunity to see how our localities stack up. 

Aug 23
ONS breakthrough - 3 cheers for them

​The absence of any data on the role played by different sizes of business in local employment has long been a bugbear of mine. I have been berating anybody who would listen, especially ONS managers, about this for years. Imagine my joy when I heard the news below. And, to be honest, I also got some satisfaction when I discovered that in most areas most employment is due to larger companies, contrary to what a lot of people have been saying ​on the basis of completely inappropriate data!​

New BRES datasets split by size of business
From 14 August 2013, employee and employment information from the Business Register and Employment
Survey (BRES) split by size of business was made available on Nomis.

The information is available in two datasets one based on the size (in terms of employment) of the workplace
(site/local unit) where the employee works, and the other based on the size of the overall enterprise that the
employee works for.

Apr 01
Correlation between small businesses and employment growth

​Sometimes in the course of exploring reasons why a particular locality has performed well I will run regressions to get a feel for what the drivers might be.  Often these are inconclusive or not conclusive enough to stake one's house or reputation on them, but interesting nonetheless. This blog provides an outlet for such work. An example was when I found that while it looked like having lots of small (1-4 in employment) businesses seemed to help, as did being on the M25 (a simple dummy variable that) there was a lot going on that could not be explained​​. Why for instance have Epping Forest and Dacorum performed so differently? Any answers? Other than the Buncefield fire that is. ​​​​​​


Mar 20
Welcome to my blog!

This is where I'll be sharing my thoughts on topics that matter to me. Who knows... I might even share pictures, videos and links to other interesting stuff.

If I catch your interest, let me hear from you.