New Economist

UK’s double dip recession: your questions answered

This morning the Office for National Statistics published preliminary estimates of an -0.2% fall in March quarter 2012 GDP. This follows a -0.3% contraction in Q4, so pulling the UK into a ‘double dip’ recession (the first since the 1970s). Below are some frequently asked questions, and my response.

Was the fall a surprise?

Most forecasters were expecting very modest growth in Q1, with the median estimate (according to Bloomberg) for a +0.1% increase. These forecasts were in line with most leading indicators and partial indicators, which pointed to sub-trend growth last quarter.

Though disappointed, however, I doubt many commentators (including myself) were that surprised. Preliminary GDP numbers are notoriously volatile, and there are several variables which can unexpectedly throw the result (more about that later).

Why did growth fall in Q1?

The services sector, which accounts for more than three-quarters of output, is not in recession - it rose a modest 0.1% in Q1 (a lot lower than business surveys and retail sales suggested). Industrial production fell by -0.4%, but that alone would not have been enough to produce a negative number. What dragged it lower was a -3% fall in construction output, which detracted 0.2 percentage points from the headline number. Without that fall, GDP would have been flat last quarter.

Should we believe these estimates?

Not particularly. Preliminary GDP estimates in the UK are based on less than half the data that go into the final national account estimates. The rest are extrapolated or imputed. They are frequently subject to large downward (or more commonly, upward) revisions. In a year or two’s time we could well discover there was no ‘double dip’ recession after all.

Former MPC member Andrew Sentance gave a powerful example in a March 30 blog:

In the early 1990s, the recession appeared to continue through 1992 and into early 1993. At the time, it was believed to be the longest recession that the UK had experienced since the Second World War. The most recent data show the decline in GDP coming to an end in late 1991, exactly when Norman Lamont declared he could see the “green shoots of recovery”. He was right, though ridiculed at the time.

Most economic commentators are likely to react with a degree of scepticism to today’s numbers, for three reasons. First, past GDP estimates have experienced major revisions. Second, there are serious quality concerns about some of the inputs, especially construction. Third, and most important, because these estimates are at odds with most of the other main indicators of economic activity. The numbers simply don’t gel.

So we shouldn’t be too worried?

Yes we should. There is a real risk that the headline contraction, reinforced by ‘doom and gloom’ stories the media will pump out over coming days, will weaken already soft consumer confidence and spending, discourage British business from hiring or spending, and deter foreign investment. If Britons talk themselves into a funk we could see what (to me at least) looks like a temporary soft patch into a more serious and protracted downturn. It’s all about Keynes ‘animal spirits’, as Akerlof and Shiller and others have noted.

How will the Government respond?

Calls for the Chancellor, George Osborne, to abandon (or at least modify) his fiscal austerity plans will grow. However ministers are terrified at the prospect of losing their AAA rating, and there are few signs of a ‘Plan B’ in the works. Indeed on Monday the Treasury chief secretary said he had asked Government department’s to identify another £16bn of cuts.

With the Office for Budgetary Responsibility set to downgrade its growth forecasts, larger spending cuts will be required simply to maintain the Government’s debt-to-GDP targets. So should the government cut even more, maintain its existing spending plans, or rethink its approach?

We can expect a protracted debate in the UK about whether the pace of spending cuts is too high, and hence self-defeating. It is a view forcefully made by Jonathan Portes, and even the IMF in their January 2012 World Economic Outlook Update (PDF):

Overdoing fiscal adjustment in the short term to counter cyclical revenue losses will further undercut activity, diminish popular support for adjustment, and undermine market confidence.

Forecast combination

One for the econometricians and forecasters amongst my audience. Call me a nerd, but I was quite excited to come across this new IMF working paper, The Use of Encompassing Tests for Forecast Combinations, by Turgut Kisinbay:

The paper proposes an algorithm that uses forecast encompassing tests for combining forecasts. The algorithm excludes a forecast from the combination if it is encompassed by another forecast. To assess the usefulness of this approach, an extensive empirical analysis is undertaken using a U.S. macroecoomic data set. The results are encouraging as the algorithm forecasts outperform benchmark model forecasts, in a mean square error (MSE) sense, in a majority of cases.

Interaction models - don't get caught!

In my naivety I had assumed most social scientists understood how to model interaction effects. Then I read this post by Omar on orgtheory.net, and realised maybe I was wrong:

So we are agreed interaction models are awesome. However, as your stats 101 teacher told you, you have to be careful about two things: (1) never omit the main effects. Thus you don’t test hypothesis 1 using any of these specifications:

Y=a+b1X+b2XZ+e

Y=a+b1Z+b2XZ+e

or Alanis forbid:

Y=a+b1XZ+e

And (2) when interpreting b1 and b2 in the fully specified model, remember that those effects are conditional on the value of the other variables. b1 is now the effect of X on Y when Z=0 and b2 is now the effect of Z on Y when X=0. If your variables don’t have a meaningful zero point (like a racial attitudes scale), center them at their mean so that you can say “b2 is the effect of being Southern on voting republican for those who have average levels of racial animus towards blacks.”

Seems simple. Everybody knows this. Why am I even explaining this to you? Well, as noted by Brambor, Clark and Golder (2006) in a recent article in Political Analysis, a survey of 156 articles published in the major Political Science journals shows that only 10% of researchers specified their interaction models correctly. A large chunk of them outright omitted main effects, which can lead to incorrect significance tests of the interaction term. In some of these articles the entire contribution was riding on the interaction term. So things are not so simple. Consider the horror:

In an award-winning article in the American Political Science Review, Boix (1999) examines the factors that determine electoral system choice in advanced democracies. He makes two main conclusions. First, ethnic or religious fragmentation encourages the adoption of proportional representation in small and medium-sized countries (621). He draws this conclusion based on a model that includes an interaction term between ethnoreligious fragmentation and country size. However, he does not include either of the constitutive terms. When these terms are included, there is no longer any evidence that ethno-religious fragmentation ever affects the adoption of proportional representation (italics added).

You should read the article to see other horror stories. The lesson: if your dissertation/paper is riding on an interaction effect, don’t be a fool. Estimate a fully specified model.

A salutory lesson. Here is a non-gated version of the Brambor, Clark and Golder paper, Understanding Interaction Models: Improving Empirical Analyses (PDF). NYU's Matt Golder has more useful material on interaction models on his web pages.

Liberate prediction markets

The new Statement on Prediction Markets, published by the AEI-Brookings Joint Center for Regulatory Studies, could hardly have a more luminous bunch of authors: Kenneth Arrow, Thomas Schelling, Hal Varian, Daniel Kahnemann and Vernon Smith to name a few. Their plea is a simple one:

We argue that U.S. regulators should lower barriers to the creation and design of prediction markets by creating a safe harbor for certain types of small stakes markets. We believe our proposed change has the potential to stimulate innovation in the design and use of prediction markets throughout the economy, and in the process to provide information that will benefit the private sector and government alike.

Though Steve Levitt refused to sign up, ostensibly because he did not think "the letter went far enough", I am sure this is a cause most economists would endorse. You can add New Economist to the pile...

More statistical woes

The Office for National Statistics' woes may worsen, if today's FT report by Chris Giles is an indication: Welsh transfer a 'risk' to statistics, says Bank

The relocation of the majority of government statisticians to Newport in Wales poses a serious risk to the quality of economic data, according to the Bank of England. Papers released by the Treasury select committee yesterday revealed the extent of concern within the Bank about the performance at the Office for National Statistics.

Cuts in the quality of data collection, with across-the-board cuts in survey sample sizes, have alarmed its economists and the monetary policy committee, making sensitive interest rate decisions - such as the one the MPC will announce today - more difficult.

...So concerned is the Bank that it should make decisions on the best possible information, it has ensured its economists will in future be able to access computer programmes that use other independent information to improve the reliability of official data.

In its memorandum to the Treasury select committee, the Bank argued the planned relocation of the ONS statisticians to Newport "poses a serious risk to the maintenance of the quality of macroeconomic data".

"If substantial numbers of ONS staff are unwilling to relocate, the loss of skilled individuals could have a severe impact on a range of statistics."

It argued that the delays in the ONS's attempts to modernise the national accounts were "very unwelcome; so were the cuts to regular activities and specific ONS products such as the 2007 Blue Book".

...Yesterday the ONS insisted that it was doing everything possible to minimise the risks associated with its relocation. It said: "We are confident what we've got in place will do the job."

The data quality problems seem to have two main causes: survey cutbacks due to an ONS budget squeeze, and staff shortages because of departures by statisticians unwilling to move from London to Wales. Today's Independent reports Move to Wales 'risks data quality'

As part of Gordon Brown's plans to move civil service jobs out of London, the bulk of the ONS's Pimlico-based staff were to be relocated to Newport, South Wales. But about a third of those affected have resigned, retired early or taken redundancy. Karen Dunnell, who heads the ONS, admitted yesterday that she expects only around 15 per cent of those workers asked to move to comply.

An unintended and, it would seem, rather unfortunate consequence of the much-heralded 2004 Lyons Review.

UPDATE: Edmund Conway has further details in Monday's Telegraph: ONS revolt may delay relocation to Wales

The Office for National Statistics may delay the transfer of hundreds of its staff to Wales amid concerns that the troubled department is facing a full-scale revolt from its senior Whitehall officials. Insiders have warned that the department is in crisis, with nine out of 10 executives claiming the ONS is badly managed.

National statistician Karen Dunnell has indicated to staff that the department is prepared to push back the deadline for relocation if important statistics come under threat. It follows a warning from the Bank of England that the relocation poses a serious risk to the quality of key economic data.

The senior civil servants' union, the First Division Association (FDA), has warned Parliament that the changes are putting the ONS's ability to produce world-class statistics at risk. In a further blow, it also claims that the modernisation programme started by the previous director, Sir Len Cook, had "affected adversely both efficiency and quality" of statistics.

...Only 11pc of senior civil servants think the ONS is well-managed as a whole, or that change at the department was well-managed, according to an internal survey.

Doing research in China: constraints, conformity and self-censorship

Hong Kong-based economics professor Carsten A. Holz asks an important question in this month's Far Eastern Economic Review: Have China Scholars All Been Bought? Of course, you know the answer:

Academics who study China, which includes the author, habitually please the Chinese Communist Party, sometimes consciously, and often unconsciously. Our incentives are to conform, and we do so in numerous ways: through the research questions we ask or don’t ask, through the facts we report or ignore, through our use of language, and through what and how we teach.

Foreign academics must cooperate with academics in China to collect data and co-author research. Surveys are conducted in a manner that is acceptable to the Party, and their content is limited to politically acceptable questions. For academics in China, such choices come naturally. The Western side plays along.

China researchers are equally constrained in their solo research. Some Western China scholars have relatives in China. Others own apartments there. Those China scholars whose mother tongue is not Chinese have studied the language for years and have built their careers on this large and nontransferable investment. We benefit from our connections in China to obtain information and insights, and we protect these connections. Everybody is happy, Western readers for the up-to-date view from academia, we ourselves for prospering in our jobs, and the Party for getting us to do its advertising. China is fairly unique in that the incentives for academics all go one way: One does not upset the Party.

What happens when we don’t play along is all too obvious. We can’t attract Chinese collaborators. When we poke around in China to do research we run into trouble. ...China researchers across different disciplines may not all be equally affected. Economists and political scientists are likely to come up against the Party constraint frequently, and perhaps severely. But even sociologists or ethnographers can reach the forbidden zone when doing network studies or examining ethnic minority cultures.

Holz does not pull any punches, likening the Communist Party to the mafia. The examples he gives of 'doing research' are both fascinating and horrifying. Hat tip: China Challenges

The anglo-saxon union blues

Recent months have seen further evidence of union decline in the anglo-saxon countries. Rates of union membership have fallen further in 2006, as trade union recruitment efforts fails to keep pace with a growing workforce.

In Canada, union membership levels increased by 62,000 in 2006, while union density fell from 30.0% in 2005 to 29.7% in 2006. "Those in the public sector ...were four times as likely as their private-sector counterparts to belong to a union (71.0% versus 17.5%)", writes StatsCanada.

In the United States, the BLS reported in January that 12.0% percent of US employed wage and salary workers were union members in 2006, down from 12.5% a year earlier. "The union membership rate for government workers (36.2 percent) was substantially higher than for private industry workers (7.4 percent)." Less than one in twelve private sector workers are in a union!

The latest Australian union membership data, released earlier this month, reported that the proportion of employees who were trade union members slumped from 22.4% in August 2005 to 20.3% in August 2006. By my calculations (see Excel time series), this was the largest percentage point decline since 1999.

Yesterday saw a similar trend reported in the United Kingdom. According to the latest annual statistical report:

The rate of union membership (union density) for employees in the UK fell by 0.6 percentage points to 28.4 per cent in 2006, down from 29.0 per cent in 2005. This was the largest annual percentage point decline since 1998.

UK public sector union density (58.8) was again much higher than private sector (16.6%). Roland Gribben reports the numbers in today's Telegraph. His piece included this rather upbeat interpretation from British unions:

...but the TUC said the figures in a Trade and Industry Department publication represented a "success story" because membership had almost stabilised since the Thatcher years.

Margaret Thatcher left office in 1990, when union density amongst British employees was 37.8% It is now 28.1%, a decline of one-quarter. That's a rather sanguine interpretation of 'stability'.

Note: Government statistical estimates on union membership are not available from Statistics New Zealand, although the Industrial Relations Centre at Victoria University do publish annual estimates.

The globalisation of labour

Has globalisation hurt workers in the rich economies? The latest Economist discusses a new IMF study in its Economics Focus piece, Smaller shares, bigger slices. They conclude that globalisation has reduced labour's share of national output:

It finds that both technological change and the globalisation of labour markets have depressed labour's share in all four groups. For the 18 countries as a whole, reckons the IMF, technology has mattered more. However, there are marked differences among them.

...The effects of labour globalisation were most evident in Anglo-Saxon and small European countries. However ...In Anglo-Saxon and smaller European countries, labour-market policies have partially offset the depressing effects of technology and globalisation on labour's share, mainly by shaving the tax wedge between what workers take home and what they cost to employ. In large European countries, increases in the ratio of unemployment benefits to wages have hurt labour's prospects, probably against policymakers' intentions.

But globalisation has also boosted worker's living standards through lower inflation:

Although globalisation has reduced labour's share of the pie, it has made the whole pie bigger, raising output and productivity and lowering the prices of traded goods and services. So are workers getting smaller shares but larger slices? Yes, concludes the fund: trade has helped, largely by making imports cheaper. Across the 18 countries studied, changes in trade prices boosted average real pay by 0.24% a year.

The analysis comes from Chapter 5 of the semi-annual IMF World Economic Outlook, which has just been published. You can download the chapter, The Globalization of Labor (PDF), or watch a video of the briefing by IMF economist Florence Jaumotte, who along with Irina Tytell, was the main author.

Fitting data curves

One of the problems with standard economics training is that the econometrics focuses on linear regression and tends to skip over non-linear functions. Most economists would never consider a curve may be, say, a Poisson distribution. Many sem to think adding a squared or cubed term to their model is the height of sophistication!

Part of the problem is that few economists seem to eyeball the data before crunching it. Now there's a tool to help. In a new post on the Statistical Modeling, Causal Inference, and Social Science blog, Aleks Jakulin at Columbia University points us to a great online tool, ZunZun. It lets you use 2 and 3 dimensional 'Function Finders' to "help determine the best curve fit for your data".

I've tried it and it works. Of course, one needs to avoid over-fitting the data. But the developer of the software, James R. Phillips, has promised to add the AIC sooon; this will help select the more parsimonius function. 

Hail the new empiricism

How do economists know what they know? In a call for a new empiricism, Barbara Bergmann asserts that economists mainly make it up. Her column, Needed: A New Empiricism, is published in The Economists' Voice Volume 4, Issue 2. Nice to see her mention Truman Bewley's pathbreaking work on wages and recessions:

The most intensive interaction with business people so far seems to have been carried out as a solo venture by Truman Bewley who interviewed about 300 business managers, asking why they don’t lower wages in a depression. The failure to do so, which may (or may not) be an important reason for the failure of unemployment to dissipate rapidly, has long been a subject of speculation among economists. Bewley’s book lists no less than 25 published theories that economists had invented to explain thephenomenon, 24 of which were wrong. Bewley was the first who dared to go out and ask those making the decisions what they did and why. They told him that to lower wages would create severe morale problems, and so would interfere with the operations of their firm.

This study was indeed daring. As if business people could know better than economists why they behave. The presumption! (For more details see Bewley's book, Why Wages Don't Fall During a Recession, and discussion papers). Bergmann concludes:

While observation-less theorizing is still the most common way economists think to advance the science, the new interest being shown to behavioral and experimental methods at places like Princeton and Harvard is an encouraging sign. But actual observation of business behavior has barely started. Whether it will flourish depends on whether an ever increasing corps of economists willing and able to do the work of observation gets trained and then finds jobs.

A few years ago I had occasion to ask Truman Bewley whether he was training students at Yale to carry on research like that he did on wages. His answer, I am sorry to report, was, “No, that would ruin their careers.”