Latvia versus Iceland revisited

 

Those who regularly read Paul Krugman’s blog at the New York Times will be familiar with his strong support for Iceland’s way to tackle its enormous financial crisis (a sharp devaluation, capital controls, letting banks go under) and his, well, lack of support for the path that Latvia chose and his post of 28 May is no exception but using what I think is the wrong argument; and given the spectacular data for Latvia in this context this is worth looking at.

He produces a graph like the one below. Latvia’s economy peaked in the 3rd quarter of 2007, Iceland’s in the 4th quarter of 2007. Letting, in both cases, GDP = 100 for the peak the graph produces the development in GDP after the crisis hit. We see the familiar story of Latvia taking a huge hit, in particular in 2009, and then recovering but GDP today (4th quarter 2014) is still some 6% below its peak in Q3 2007 while Iceland has returned to its GDP level of Q4 2007.

Figure 1: Development in GDP in Latvia and Iceland, 2004 – 2014, peak = 100

Source: Eurostat and own calculations.

This development makes Krugman conclude that “Truly, nothing much here to justify all that triumphalism”. That may be so (although I don’t consider myself “triumphant” or, for that matter, the opposite) but the lagging behind in terms of catch-up deserves a closer look.

One argument that has been used is that Q3 2007 is poor for comparisons since Latvia was seriously overheated at that time – don’t compare with outliers. Krugman also mentions this in his September 2013 blog on Latvia. I strongly believe in the Latvian economy being overheated in 2007 – I was here and I saw it but be that as it may.

A much better argument for why Latvia hasn’t caught up as much as Iceland is in employment figures. (Krugman, to be fair, alludes to this in a January 2013 blog post but why he doesn’t include it here I don’t get). I have used these numbers before but I repeat them here:

Figure 2: Latvian Labour force and employment, 1000s, 2004 – Q1 2015

Source: Central Statistical Bureau

When the economy peaked in 2007 some 1,080,000 people were employed with a historically low unemployment rate of about 6%. Today some 884,000 persons are employed but from a labour force of just 985,000. The labour force today is thus 100,000 persons lower than employment in 2007 and Latvia will thus never again see a level of employment as in 2007. So many fewer people obviously have a problem producing as much GDP even if their productivity has increased (and it has and quite a lot). So, the closer look at the data produce what we are quite familiar with here: The lack of catch-up compared to Iceland (where employment today is about the same as in 2007) is due to migration (blame it on low wages compared to the EU average, blame it on the crisis and its high level of unemployment) and poor demographics since the 1980s. In short, there are simply not enough people in Latvia to manage this catch-up yet – it will be done eventually but due to further productivity growth as Latvia (hopefully) achieves more convergence with the EU average income level.

This decline in the labour force is worth looking at more closely. In Figure 3 I have used Eurostat data to calculate the size of the labour forces in the EU countries and I then compare its size today (2014) with what it was at the time when most economies peaked, just before the financial crisis really hit (2008). I could have compared to 2007 or, as in Figure 1, have looked at where the individual countries peaked but it should not produce such different results (most countries in the EU went into recession in the first two quarters of 2008, i.e. very close to each other in terms of time).

Figure 3: Percentage change in the size of labour force, 2008 – 2014, EU28 (minus Luxembourg, Malta and Cyprus)

Source: Eurostat and own calculations Note: Luxembourg, Malta and Cyprus have been excluded given their small sizes where a change of relatively few thousand people can produce big percentages. They all have sizeable increases in their labour forces.

Given the Latvian data from Figure 2 we are not at all surprised that Latvia’s labour force is down but I must admit that I was a bit shocked that it was so much more than elsewhere.

And since poor demographics from the past 30 years keeps affecting the labour force negatively for the next 20 years (few children today cause labour force trouble some 20 years later) it emphasizes the need to see whatever can be done to increase the labour force. The increase in the retirement age is of course a very good idea but more has to be done against this relentless squeeze of those potentially in work.

PS. It is obvious that the labour force should shrink in places like Latvia, Romania, Portugal, Greece etc. – low GDP per capita and/or high unemployment and it is obvious why it should increase in Sweden and the UK – recipients of immigrants. But if anyone can tell me the story behind Hungary they are most welcome.

Morten Hansen is Head of Economics Department at Stockholm School of Economics in Riga and a member of the Fiscal Discipline Council of Latvia