Structural Aspects of German Export Dependence: Part I

by Edward Hugh: Barcelona

According to the Federal Statistics Office German sales abroad fell 1.4% in March when compared with February. There is nothing in and of itself either deeply significant or earth shattering about this, it is possibly even a blip, although the most likely cause is the impact of the first quarter slowdown in the United States. It does however raise the rather interesting question of just what will happen to growth in the German economy as and when there is a slowdown in the rate of increase in global trade, if this means that German exports will be unable to maintain their recent blistering pace of expansion.

This post will treat certain "stylised facts" about Eurozone imbalances (and here) and the relationship of these to the global economic imbalances situation as by now reasonably well established , at least for the purposes of this weblog. The first of these the "stylised facts" relates to the current debate over the so-called decoupling thesis, about which Claus Vistesen had an excellent recent post. This thesis generally asserts that the current wave of global growth is virtually unprecedented in modern times,and that the structural drivers of this growth - which is characterised by very strong rates of growth in some emerging markets (and in particular in the eurozone context in the new East European EU accession countries) - are new, and fundamentally different from those which characterised global growth in the 1990s. Another "stylised fact" is that large parts of the global economy are still extremely export-dependent on growth in consumer demand in the United States, although the fact that growth globally is running at some 5% per annum while most of the economies in the developed world are growing at under 3% per annum also means that this level of dependence is now in secular decline. This latter point logically leads to the conclusion that growth in the developing world is providing a momentum to this upswing which is what marks the whole process as a novel one.

A third "stylised fact" would be that a number of countries are now revealing themselves as having export-dependent economies. This group can basically be broken down into two components, China and the oil producing states on the one hand, and the growing list of countries (Japan, Germany, Italy, Finland, Austria..) whose median ages are rising steadily upwards over the 40 year mark (and onwards in some cases towards the mid forties) due to the ongoing process of population ageing largely caused by a widespread and generalised demographic transition. Here in this post we are primarily concerned with trying to better understand what is happening in this latter group of "elderly" economies, and Germany will be treated as one very good example of these.

That these "elderly" countries should show themselves to be increasingly dependent on exports should be neither surprising in terms of conventional economic theory - in this case Franco Modigliani's Life Cycle Hypothesis of movements in consumption and saving through the age ranges as applied to populations - nor in terms of the "known facts", since the idea of rising median age being associated with export dependence is now finding increasing empirical support in the anticipated countries (Germany and Japan in particular), where a characteristic pathology of congenitally weak internal demand, comparatively high (even if now declining) saving, and consequent strong dependence on exports to obtain GDP growth seems to be the order of the day. (More argumentation on the median age issue can be found in this post, while Claus ably argues out the implications of all of this for Germany here and here).

So it could be argued that we have both theoretical justification and growing empirical evidence for advancing an "ageing structural characteristics" hypothesis, and I would say we have good grounds for asserting this hypothesis, although doubtless all of this needs more elaboration and validation as more countries age.

But even as we move forward to try and do just that it is probably worth trying to go one step further, and examine some of the implications of this structural dependence for the future growth possibilities of the economies concerned. Are they still able, as some would claim, to generate self-sustaining economic recoveries? Or are they from-now-on-and-henceforth condemned to ever ride the coat-tails of global growth and demand which is generated elsewhere. Much as Claus and I may have been surprised by the resilience shown by the German economy, and indeed by the strong current growth we are seeing across the eurozone as a whole, which as Claus has argued tends to be significantly dependent on Germany, a form of "coupling" which in some ways mirrors the global one in connection with the US, but which involves the very significant difference that while those who are directly coupled to the US economy are in some form or another dependent on the US consumer, those who are indirectly coupled through the mechanisms of the EU - I am thinking here in particular of the eastern accession countries - or through the zone itself are in fact coupled to Germany and thus coupled to Germany's ability to sell exports to consumers both inside the EU and elsewhere.

In Many cases, as we shall see, this generates a strong circularity in eurozone growth, since those in this chain who are also among Germany's major customers (often buying investment goods via cheap euro-denominated credit made available thanks to Germany's high savings rate, or thanks to a highly leveraged carry trade -in Swiss Francs or Japanese Yen - intermediated via the Austrian banking system), are thus sensitive to the strategic importance of the US consumer for the German economy due to their need to export back final product to Germany and other EU destinations, even though the US consumer obviously does not form part of the EU "inner circle".

Turning to the most recent data published by the German Federal Statistical Office we can see that in fact taking into account calendar and seasonal adjustments, German exports in April increased by 0.9% over March 2007. So to some extent the drop in exports in March may be thought of as a blip. Yet if we look at the entire time series for from January 2006 (see Chart below, and please click on image to read) we can see that exports increased very rapidly in the first months of 2006, then the rate of increase slowed a little, there was another burst of activity in September and October, and since that time the rate of increase has been steadily dropping (the far right column showing the adjusted figures month-on-month is perhaps the clearest).

(Please click over chart to read clearly)

In order to investigate further the real nature and dynamics of this export-driven growth phenomenon in Germany it is perhaps worthwhile putting the recent evolution of the German exports in some sort of context. Fortunately we are aided in this by the publication of a recent IMF working paper: What Explains Germany’s Rebounding Export Market Share? by Stephan Danninger and Fred Joutz.

As the authors note in their abstract:

Germany’s export market share increased since 2000, while most industrial countries experienced declines.

Now this statement does perhaps need a little clarification, since while it is true that the German market share in exports has increased in volume terms (as shown in the first of the following charts which compares the German performance with the Italian one), in value terms the German share has been declining slightly of late (as shown in the second chart). Nonetheless the German performance is still remarkable given the resilience of German exports even in value terms in the context of a global trade environment where growth has been very rapid indeed, and where several new contenders have made a dramatic entrance on the scene.





Again, as the IMF authors note, exports (and the investments that servicing these implies) form a very important part of the recent German growth phenomenon:

Germany’s export sector has become its main source of economic growth. Since 1999 about 80 percent of real GDP growth was generated from net exports (see the first chart below). Real exports have grown by more than 7 percent per annum since 2000 on the back of growing trade volumes with both traditional European partners and emerging economies. Since 2000 Germany has also begun to regain export market share, especially among industrial countries and the euro area (see second chart below, which is in volume terms).





The paper itself looks at four hypotheses which might help explain Germany's recent stellar export performance, and these are:

(i) improved cost competitiveness through moderate collective wage agreements since the mid 1990s;
(ii) ties to fast growing trading partners as a result of a desirable product mix or long-standing traderelationships;
(iii) increased export demand for capital goods as a response to a global rise in investment activity, and
iv) regionalized production patterns through off-shoring of production to lower cost countries, partly a result of European economic integration (Sinn 2006).

The authors then go on to assess the relative importance of each of these four hypotheses, and examine how their findings influence the prospects for continued German export growth and German economic activity in general. This focus is especially important since, as they note, export growth has been strong despite generally weak domestic demand and low consumption growth.

Their analysis in fact shows that recent export growth can be traced back to the ability of German exporters to meet global demand and to exploit new production and cost-cutting opportunities from offshoring activities. Their analysis also provides some empirical support for the claim that German exports increased as a result of a regional division of labor in the production of goods (the bazaar economy hypothesis, most notably associated with Hans Werner Sinn). They find that the first two factors (ie the ability of German exporters to increase output and offer products that fit in with the recent surge in global demand and their ability to exploit new production and cost cutting opportunities from offshoring activities) are able to explain about 60 percent of the faster increase of German exports since 2000 vis-a-vis the other industrial countries. Changes in relative prices, measured by the real effective exchange rate, on the other hand, seem to have contributed surprisingly little to export growth despite a prolonged period of wage moderation.

I say surprisingly, since this downward movement in relative unit labour costs has often been mentioned as an explanatory variable of some importance in the recent revival of economic activity in Germany.





As the above charts indicate German unification resulted in a steep increase in wage costs, mainly from pressures to close the wage gap between the new and the old Länder, and from tax increases to cover the cost of extending the welfare state. The resulting loss of cost competitiveness and economic restructuring led to a sustained period of high unemployment. By the late 90s a period of extended wage restraint followed and to a large extent the position was reversed.

Since the late 90s wages and salary growth in Germany has lagged behind productivity growth — the cost-neutral margin — in almost every year. Wage costs per unit of output began to decrease sharply in 1995 and have remained at a low level since 2000 despite a significant nominal effective appreciation of the euro. The main factor responsible for this adjustment has been muted wage growth in industry. Average hourly nominal wage growth has declined continuously and has been hovering since 2003 at somewhere around 1-2 % per annum rate. This has lead many observers to conclude that cost competitiveness has been a leading promoter of German export growth, and some have even argued that a return to more normal wage growth is now becoming possible, and that if this were to be realised it would help strengthen the ever-weak German domestic demand position (the so-called Goldilocks recovery hypothesis).

However the IMF authors find that improved internal wage cost competitiveness has played a comparatively minor role in explaining Germany's brisk export growth. They find that Germany's prolonged effort in containing costs through wage moderation has been significant, but this effect has in turn been diluted by the appreciation of the euro. Hence cost competitiveness may be considered to have improved Germany's position vis-à-vis the euro area countries - and consequently it does help explain the significant rise of Germany’s export market share within the eurozone - but this factor does not in and of itself serve to explain the export performance with non-eurozone economies.

Now if we look at recent data from the Federal Statistical Office we will find (as shown in the chart below) that despite the recent concerns over at the ECB about wage induced inflation, the actual movement in gross wages and salaries in Germany has been extremely weak both in 2006 and to date in 2007, despite the apparently favourable growth environment, and so far, for example, there is no real evidence of the "pass through" of the 3% VAT increase. Indeed in the first quarter of 2007 the costs of an hour worked in industry and services as a whole rose by only a calendar-adjusted 0.4% when compared with the same quarter a year earlier. Also upon adjustment for seasonal and calendar effects, labour costs per hour worked remained nearly unchanged (–0.1%) in the first quarter of 2007 against the previous quarter.This is significantly below the rate of inflation, thus it could be said that deflation in real German wages continues, despite a generally tighter labour market.


(Please click over chart to read clearly)


The German Statistical Office also make a Europe-wide comparison of the level and development of labour costs:

In 2006, labour costs in Germany rose by a calendar-adjusted 1.1% in industry and the market service branches – shortly referred to as the private sector – compared with the preceding year. The increase (as measured in euros) was markedly higher especially in the Eastern European member states Czech Republic (+11.7%), Slovakia (+12.1%), Estonia (+16.6%), Lithuania (+18.5%), Latvia (+23.3%) and Romania (+23.5%). Among the reasons for this are appreciations of the national currencies against the euro. In France the corresponding rise was 3.3%, in the United Kingdom 3.6%

The comparison with the UK and France here is most interesting, and it can be seen that the upward movement in German nominal wages is far lower than in these two major European economies.

So the question is really why is this happening?

Well, recent research from the Munich-based economist Dalia Marin may help us out a bit here. In a paper entitled "Is Human Capital Losing from Outsourcing? Evidence for Austria and Poland", Marin et al (see link below, as well as other related material, including a link to the PhD thesis of her student Alexander Raubold, on which much of the resulting research seems to be based) argue that:

"multinational firms in Austria and Germany are outsourcing the most skill intensive activities to Eastern Europe taking advantage of cheap abundant skilled labor in Eastern Europe. We find that the firms’ outsourcing activities to Eastern Europe are a response to a human capital scarcity in Austria and Germany which has become particularly severe in the 1990s."

Now this finding when I discovered it really rather surprised me, although perhaps, with hindsight it shouldn't have. I had imagined, like many others doubtless also do, that Germany and Austria were outsourcing primarily unskilled, or lower skilled work (this is, for example, the situation in Sweden, see the Becker et al paper referenced below). But then reflecting a little on simple economic theory, and taking on board the fact that both Austria and Germany are now suffering relative shortages of young people, and the huge differential in skilled wages which exist between East and West , then it may well make sound economic sense for German and Austrian companies to employ relatively more intensive quantities (in terms of labour-capital ratios) of relatively cheap skilled labour in the East, and this does seem to be what has been happening.

Corporations’ outsourcing of skill intensive firm activity to Eastern Europe has helped to ease the human capital crisis in both countries. We find that high skilled jobs transferred to Eastern Europe account for 10 percent of Germany’s and 48 percent of Austria’s supply of university graduates in the 1990s.

The Austrian number seems really enormous, the German one less so, but it is still important. In addition Marin produces data which shows that German affiliates in the accession countries were paying (prior to 2004) 17 percent of their German parent wages but were increasing their productivity to 60 percent of the parents’ productivity level. Simple mathematics therefore suggest that they were able to reduce the labor costs by 72 percent relative to their parent-firm cost in Germany.

On examining the pattern of multinational investment and outsourcing across sectors Marin found that German investment was predominantly engaged in manufacturing activity in Eastern Europe (almost 60 percent of total investment), of which manufactured goods and machinery and transport are the most important sectors. She also found that some 90 percent of German investment in machinery and transport were outsourcing (rather than local market oriented) investments.

Marin then went on to look at the levels of skill intensity for outsourced work and asked just how skill intensive the activity undertaken by German affiliates in Eastern Europe actually was when compared to their parent firm activity in Germany. She used two indicators to measure the skill intensity of German affiliates in Eastern Europe:

a) the share of workers with a university or college degree and
b) the share of personnel engaged in R&D or engineering activities in the manufacturing and service sector.

The data she studied suggest that the high-skill ratios of affiliates (the number of university or college workers in percent of total affiliate workers) are 2 to 3 times as large as that of German parent firms in the Eastern Accession Countries. The share of university or college graduates among affiliate workers in Eastern Europe was found to vary between a high of 86 percent (Czech Republic) and low of 8 percent (Slovenia).

The most skill intensive activity was being undertaken by affiliates in the Czech Republic (with a skill share of 86 percent), and Slovakia (skill share of 40 percent). This compares with an average share of university or college graduates of German parent firms of 18 percent only. Thus, measured by the number of university and college graduates, German affiliates in Bulgaria were found to be 12 times as skill intensive than their German parent firms, and affiliates in the Czech Republic 5.5 times as skill intensive. Only affiliates in Hungary were found to have a skill share below that of German parent firms.

A similar picture emerged when she looked at the skill intensity of German affiliates as measured by the share of workers engaged in R&D and engineering. The R&D personnel ratios of affiliates in Eastern Europe ranged between 4.0 percent (Slovakia) and 27.8 percent (Croatia and Russia). This compared with an average R&D personnel share of 13.6 percent of German parent firms. Thus, German affiliates in the Czech Republic are 1.7 times as R&D intensive as their German parent firms.


These are striking and puzzling numbers. German and Austrian multinationals tend to outsource
the most skill and R&D intensive activities to Eastern Europe. Why is this happening? Well according to Marin:

Economic theory guides us to look at the factor endowment of these countries for an answer. If countries outsource the most skill intensive activities to other countries, then these countries must be poorly endowed with skills relative to their trading partners.When Germany and Austria’s endowment with skills is compared to Eastern Europe.....we find.....the Baltic States, Russia, Hungary, and Bulgaria are the most skill rich countries as measured by the share of the labor force with a tertiary education level. Germany’s education level lies below the OECD average and roughly matches that of the accession countries average. In particular, Germany is less skill rich than the Baltic States, Russia, and Hungary.In this ranking of countries Austria turns out to be the most skill poor country.


Maquiladoras In Reverse?

Maquiladoras are affiliates of US multinationals in Mexico which specialize in the low skill intensive part of the value chain. Sruggling with the phenomena she found before her, Dalia Marin was lead to ask whether it might in fact not be the case that with the Eastern Enlargement process an inverse Maquiladoras effect was emerging in Germany, since German multinationals are evidently outsourcing to some extent the more skill intensive stages of production to Eastern Europe and retaining the more labor intensive stages of production in Germany. As a result, the relative demand for skilled labor declines in Germany and in this way, outsourcing of high skill intensive activities to Eastern Europe has helps to ease the (partly demographically driven) human capital crisis in Germany. So this may in fact offer us some part of the explanation as to why relative wage for skills in Germany (or skill-bias in wages) has not increased on the back of the information technology revolution of the 1990s since outsourcing activities have removed some of the demand pressure on skills from the German labor market. It may also help to offer some part of the explanation for why hourly wages on aggregate are not rising faster inside Germany itself.

It is important to note at this point that with the shift from an industrially driven economy, to a knowledge-based services one, there is a secular rise in the proportion of higher skilled work which is needed in an economy. So it is not a question of absolute numbers of skill intensive workers declining, but that in Germany these numbers have not been rising as rapidly as should have been the case. The following graph from Alexander Raubold (see link referenced below) perhaps helps make this clearer:

(Please click over graph to read clearly)

Now the above graph illustrates movements in the skill premium and the German labour market over the last 15 years or so. Essentially it shows the ratio of high-skilled to low-skilled wages and employment and the high skilled workers wage bill share in German manufacturing. What can be seen is that over the period from 1990 to 2004 the wages of high-skilled workers remain fairly constant relative to those of low-skilled workers. The right hand axis charts the evolution of the skill premium. What can be seen is that, following a slight peak in 1993, followed by a decline until 1995, there was a slow and steady increase from 1995 to 2004, from an index reading of 160 to 164, but that all in all these changes remain marginal. But when we come to look at employment, the employment of more highly skilled (or non-direct-production) workers increased by 27% over this period, moving from being 48% of the total workforce in 1991 to 64% in 2004.

So the real question is: how - given the relative shortages of new skilled labour inside Germany - has this subtantial increase in the skilled labour employment share been possible with such a slight change in the skill wage premium. I think the arguments Delia Marin offers may well provide us with some significant part of the answer.

Now aside from the interesting explanation insight that all this may offer into the dynamics of the German labour market, I would also like to add that the process Marin identifies represents decidedly bad news for those economies which face labour shortages and ageing workforces. Let me explain why. Essentially it has been argued - and standard neo-classical theory indeed anticipates - that tightening in ageing labour markets should lead to rising wages and relative capital deepening (as capital gets substituted for ever more expensive and scarce labour). Now this has always been thought to imply that the society which is subjected to this process would move up the value chain under such pressure, and as such domestic value-added and productivity would increase in a way that meant that living standards (and hence health and pensions systems) could be maintained. Now, if Marin's work proves to be backed by further research findings, I would say we might be seeing the beginnings of another possibility here, one where labour market tightening in some countries leads to skill outsourcing, and to an increase in the relative proportion of non-skilled activity in the ageing home country.

I say IF here, since we obviously need to see more evidence to reach any definitive conclusion. But if confirmed this skill outflow would account for another anomaly: the outflow of skilled workers and inflow of unskilled migrants - which just balance each other - in the German case. Basically it could be hypothesized that you need cohorts of a certain specific weight in your demographic profile - to maintain a natural downward movement on the skill premium across time - in order to ba able to make the knowledge economy transition based on largely domestic labour. Absent this, you seem to spring a leak, and the premium is kept reined-in by an outflow of work to emerging markets where there are bigger younger cohorts, and a more plentiful supply of young workers with the right skill profile. Ironically this then has the perverse effect that there is not enough employment generated for the young skilled workers you actually have, and you get an outflow. It is important to emphasise here that I am postulating that this perverse reverse-Maquiladora effect is posited on two structural characteristics of elderly economies:

a) growing relative shortages domestically in young skilled workers
b) increasing dependence on export driven growth due to weak growth in domestic demand.

This latter point (b) is important since it is the absence of strong growth in domestic demand (most noticeably typified by the absence of housing driven growth) which creates a deficit in certain sectors of knowledge economy services employment which have to be compensated for by dependence on export markets, and a strong tendency towards certain types of outsourcing. In comparison - for example - the US economy has seen a significant drift in outsourcing employment, but the labour market for skilled young workers has remained strong due to growth in the domestic services sector, and indeed the US is a net importer of skilled labour.

Since similar processes MAY also be at work in Japan and Italy, this would seem to be a line of enquiry which is well worth checking out to see if the German (and to some extent Austrian) impacts are replicated.

Macroeconomic Adjustment in the Euro Area II: Italy

by Edward Hugh: Barcelona

Early warning. This is a monster post. I hope it is worth the ride.

In fact this is the second post in a series I started here (and unfortunately it is also a monster one, so again watch out). Essentially I am taking Chapter 2 of the recent European Economic Advisory Group (EEAG) report on the European Economy 2007 - which is entitled Macroeconomic adjustment in the Euro Area: the Cases of Ireland and Italy - as a starting point for a much more general reflection on how the euro system works, and as a pretext for continuing the search for the structural drivers of the European economies (and here and here). The current post will largely deal with the Italian case, but will attempt to draw out parallels and comparisons. A third post will look at the role of exports in the German economy, while a fourth one will attempt to draw together the threads and examine what may be learnt.

So coming to Italy, as the report notes, GDP growth in Italy has been very sluggish over the last decade or so. As can be seen from the chart below, GDP growth since 2002 has been well below the EU average:



(Please click over all images to get a better view).

This is not, however, a recent phenomenon, since as the next chart shows, Italian GDP growth has been low since the early 1990s:





In fact Francesco Daveri and C. Jona-Lasinio - in a most interesting paper on Italy's relative economic decline entitled Italy's Economic Decine, Getting the Facts Right (of which more below) - say the following:

Italy’s per-capita GDP growth was 5.4% in the 1950s, 5.1% in the 1960s, 3.1% in the 1970s, 2.2% in the 1980s and 1.4% in the 1990s. A rough-and-ready extrapolation of this decade-long continued slowdown would lead to expect no more than 0.5% in the 2000s (so far we are at some 0.6% over 2000-05, if per-capita GDP stays constant in 2005). In any case, nowadays, the miracle years of the1950s-1960s seem quite far away in time.

They also produce this chart which puts Italian long term growth in context:



And the situation becomes even clearer if we look at this chart they produce, where it can be seen that per-capita GDP in Italy has only rosin more quickly than the combined rate of France, Germany, the UK and Spain in six years since 1980, and in none in the last decade (during which time, it should be remembered, German per capita GDP growth has itself been very low).




as Daveri and Jona-Lasinio comment:

"(The chart)...concisely shows that Italy, being much poorer than Europe (poorer than France, Germany and the UK, but richer than Spain) in 1950, has been catching up fast until – roughly - the early 1990s. This process of convergence has reversed its course since then, however. In the 1950s and the 1960s, Italy grew faster than Europe six times in each decade; in the 1970s and the 1980s this occurred four and five times respectively. In the 1990s, instead, this occurred only twice, in 1991 and 1995. Since 1995, then, Italy’s per-capita GDP has grown less than (the other big countries in) Europe’s GDP."

they also add this rather intriguing comment:

"Altogether, the long-run data suggest that the bad performance of the Italian economy is not the figment of the currently unfortunate business cycle contingency. This is why speaking of decline may not be totally unwarranted. With one caveat to add, though: given that the rest of Europe has been and is still growing at a positive pace, Italy’s alleged decline is of a relative, not an absolute type. Italy’s per-capita GDP has simply grown not as fast as Europe’s GDP, but has not diminished over time (yet)"

That "yet" word seems to be looming rather ominously, since with the growing age-related dependency issue just over the horizon, such a decline in the future can certainly NOT be ruled out given current trends. More importantly though, Daveri and Jona-Lasinio seem to give little credence to the idea that Italy's problems can be simply placed at the door of an immediate business cycle conjuncture, or shock (although such shocks may well be exacerbating the problem).

Going back now for a moment to the first chart it is clear that, despite a relatively relaxed monetary stance from the ECB, Italian inflation has been trending down in recent years. This is undoubtedly in part associated with the relatively sluggish rate of growth in internal demand from which Italy has been suffering - and it should also be noted that this drop in inflation has come despite a steady fall in Italian unemployment and a continuing rise in wage costs. The contractual hourly earnings index rose by 3.3% year on year in October, November and December 2006. Over this period the largest annual wage increase was recorded by public-sector workers, whose salaries rose by 5% y-o-y, continuing a longer-running trend. In general - and despite lower productivity rates, government workers have been getting higher pay increases than private workers during the last five years. Wage inflation was at its lowest in the services sector, with a rate of 1.9% y-o-y in October and 1.8% in November and December. In industry (excluding construction), wages rose by an average 3.7% y-o-y over the three months. (Overall hourly earnings rose at 3.3 % in 2004 and 3.0% in 2005).

Of course the recent drop in energy prices does form part of the inflation picture, and it will be interesting to watch what actually happens if energy costs resume their upward trend. It should be noted, however, that despite this downward trend in the rate of price increases, and despite the lackluster growth, the Italian inflation rate has remained stubbornly around, and often slightly above, the euro area average, as can be seen from this chart:


Now turning to the constituents of GDP, we can see from the chart below that, whilst domestic consumption did gain some momentum during 2006, the general picture has been one of very slow growth in domestic private demand (please click over chart to read). In 2004 private consumption grew by only 0.5%, and in 2005 it was absolutely flat (0% growth). This position is now increasingly being mirrored by Italian government consumption which is quite tightly constrained due to the growing deficit problem. Indeed what evidence we have to date from 2007 suggests that the underlying weakness in domestic consumption continues.



Now, as can be seen from the next chart, public consumption in Italy is relatively important, since private consumption only constitutes some 60% of GDP (as opposed to nearly 70% in the USA), so given the relatively weak export performance of the last decade (of which more below), and the weak growth in private consumption, government spending has been playing a significant role in maintaining what growth there has been. This will now have to change if Italy is to bring the deficit (which is currently around 107% of GDP) down.


Returning to the EEAG report, as the authors say:

Our analysis focuses on Italy as an example of slow adjustment in response to shocks reducing foreign demand. The creation of a common European currency coincided with a strong crisis in competitiveness and productivity in Italy, exacerbated by the appreciation of the euro since 2002.

In addition they also note:

The Italian export crisis has not erupted suddenly but has been developing since the mid-1990s. Between 1995 and 2005, the share of Italian exports in world exports at constant prices fell from 4.6 to 2.7 percent, a 40 percent drop. The comparison with Germany...is striking: Over the same period, the German export share grew by 15 percent. If exports shares are instead calculated at current prices, the share of Italian exports in world exports fell from 4.6 to 3.7 percent (see De Nardis and Traù 2005). Of course, Italy is not the only developed country to lose market shares over the period, as there is a trend shift in favour of the emerging market economies.

Well the comparison with Germany is here frankly interesting, since quite simply, in terms of their demographic profile, Germany and Italy have a lot in common, a lot more in common with each other, than either have with say the UK, or France, or Spain, or - indeed - Ireland. Both Germany and Italy have suffered from ongoing weaknesses in their domestic consumption (and neither of them, of course, have recently had housing 'booms', of which more later), and both need to address this structural issue in internal demand - which seems to be related to population ageing - by becoming more competitive in their export sectors. But what differentiates these two economies is that while Germany GDP growth has increasingly come to depend on a very efficient export sector, in Italy shortcomings in exports and domestic consumption have rather been compensated for by increasing public spending and growing government indebtedness.

The accompanying chart makes the relative export evolution clear:



So Germany has done immensely better than Italy in the export sector, and indeed, much better than many others internationally, since Germany's share in global exports started from a relatively high initial level. The comparison shown in this graph is revealing:




It becomes even more revealing when we look at how closely the Italian path mirrors the French one. Yet France has not had anything like the same low growth problems that Italy has had, basically due to the relatively better health of domestic consumption. And the question many might like to be asking themselves at this point is why should this be, why this difference in internal consumption between Italy and France?

And it is here that I find myself parting company somewhat from the general approach of the EEAG report, since they tend to focus on what they take to be a shock-induced external competitiveness problem which they argue may have developed in Italy since 2002, whilst I think it is important to try to situate Italy's recent problems in a much broader historical and evolutionary perspective:

Nonetheless, the Italian competitiveness crisis substantially worsened after 2002, coincident with the appreciation of the euro. It is apparent that the Italian export crisis became acute after 2002. The index of industrial production for the exporting sectors lost approximately 6 points relative to non-exporting sectors from 2003 on. A similar gap can be detected for capacity utilisation. In response to the large external shock to export demand, adjustment would require real depreciation.

Now certainly there is no doubting the fact that the Italian export sector has been suffering from a competitiveness problem, and while there were signs that export growth was once more picking up again at the end of 2005, growth once more turned negative in the 3rd quarter of 2006. An indication of the competitiveness problem can be found in this chart which compares real effective exchange rates (or if you prefer relative movements in unit labour costs):



Indeed the general poor productivity performance achieved in Italy in recent years is highlighted in the chart below (which comes from the OECD 2006 factbook) where Italy can readily be seen to have been the worst performer in the OECD in the years between 2002-2004.



So there seems to be no denying this part of the argument, Italy's productivity performance has been lamentable. Italian economists Francesco Daveri and C. Jona-Lasinio - in a paper entitled Italy's Economic Decine, Getting the Facts Right - examine the Italian productivity 'problem' and come to three conclusions:

We reach three main conclusions. First and foremost, most of Italy’s economic decline stems from decreasing labor productivity (not hours). Second, the standard decomposition of industry productivity trends shows that the bulk (80%) of Italy’s productivity slowdown originates from a generalized within-industry slowdown (or outright declines), mainly in durable and non-durable manufacturing. Diminished inter-industry reallocation from agriculture onto market services contributed the remaining 20% of the slowdown. Third, the labor productivity slowdown was mostly accounted for by a marked deceleration of TFP, which was not the result of an unfortunate cyclical contingency (the current slowdown is worse than in any former downturn in the last twenty years).

So, essentially, there has been a within-industry slowdown (ie TFP has not risen fast enough as industries have not transited to higher value activities) AND there has not been a rapid enough movement into the new market services areas.

In some ways this is very compatible with the story the EEAG highlight in the following chart:



What we can see is that activity has been maintained much more in the non-export than in the export-related sectors, and much of this domestically oriented activity may be in more traditional, low-value-added activity, increasingly staffed, possibly, by newly arrived migrants with low skills and education (see below).

But noting this, and leaving matters there is in many ways to remain on the surface, and not to get to the heart of the matter. As Daveri and Jona-Lasinio so cogently argue, Italy's decline is a much longer term phenomenon, and we need to get through to the underlying structural issues here. In this context we may think that comparing Italy and Ireland is rather like comparing apples and pears, the two economies are fundamentally different structurally, but why are they different?

If instead we compare Germany and Italy - which are from many points of view much more alike - we can see that the fact that Italy has been much slower - for internal political reasons - to bite the bullet of the reform process, means that it has been public spending and not the export sector which has been the growth driver on the margin, but this still leaves us with unsolved issues.

What is it that separates BOTH Germany and Italy from France, and why is it that the former cannot rely on dynamic internal consumption, while the latter can? The answer to this apparent conundrum may come in one word: housing, and in the different ways in which the different eurozone economies responded to one and the same nominal rate of interest in terms of the presence or absence of construction booms.

So what determines whether or not there is a construction boom in any given country? Well strangely enough the answer to this seems to be relatively (indeed perhaps deceptively) simple: median age. I have yet to find a society whose population has a median age which is substantially over 40 which has had a construction boom over the last five years which has been of anything like the magnitude of those which have been seen in those societies with median ages in the 35 to 40 bracket.

And why is this important? Well quite simply, developed economies are becoming more and more services driven, and an important component in this whole shift to services economies has been a growing importance for the construction share in economic activity.

Italy is, of course, like most other OECD economies, making the transition to becoming a services, rather than an industrial society, and the share of services in GDP is growing constantly. In the Italian case the services sector clearly constituted the main driver of growth in gross value added in the third quarter of 2006, despite slowing compared with the first half of the year. Services expanded by 0.3% in July-September compared with the previous quarter, down from quarter-on-quarter growth of 0.9% in the first and second quarters. Financial services and real estate, which had been flat in 2005 showed the sharpest rise, growing by 0.6% in the third quarter compared with the second.

So there is some dynamism in the real estate sector, but has there been a boom? And what exactly has been happening to the housing sector in Italy over recent years?

Well the first thing to note - looking at the comparative graphs below, which show trends in EU house prices - is that from the late 80s to the early 90s Italy had quite a significant housing boom, following which prices flattened out, subsequently remaining either static or even falling slightly until the early years of this century:



The graph below shows the position in even more detail, since it gives annual movements in real house prices for Italy from the mid 90s to date, and here we can see how, while at the end of the 90s prices were actually dropping, they did start to rise in the early years of the century, but they subsequently peaked again, and started and have since started to flatten out again, running at a rate somewhere around the annual shift in Italy's CPI, that is they are more or less flat in real terms.



The above graph has been produced by the Royal Institution of Chartered Surveyors as part of their European Housing Review 2007. In that review they have this to say about the current state of the Italian housing market:



The housing market continues to be fairly flat. Prices in 2006 again rose by around 4% in nominal terms and sales dipped somewhat. Prices, in fact, were flat in most of the major cities and it was in suburban and smaller town localities where the market was strongest. So, the revival seen in the economy does not seem to have filtered through to the housing market as yet. The impetus from it was probably offset by rising interest rates in a country where variable mortgage interest rates predominate. It is expected that the market might slow even further in 2007 with continued pressure from rising interest rates. However, no actual fall in prices is anticipated.

The following comments from the RICS report also seem interesting and relevant:


The mortgage market is fairly recent. Outstanding mortgages were only 18% of GDP in 2005, although this was up from 6% in 1990.10 Outstanding mortgage debt per capita is almost nine times less than in heavily mortgaged Denmark, for instance. Surveys suggest that only around half of purchasers use mortgage facilities, despite the apparent tax benefits of doing so.

At present, Italy has an exceptionally low level of personal borrowing for an affluent society. The overall ratio of longterm household debt to household disposable income was only 41% in 2004, less than in any other of the world’s major economies.

The introduction and growth of mortgage lending interestingly has paralleled similar developments in central and eastern Europe. The Italian experience highlights the fact that a country does not inevitably experience a major house price boom just because a new mortgage market rapidly develops.

This last point, that a country does not necessarily experience a housing boom simply because a new mortgage market rapidly develops seems especially important. The question is why not, what are the sensitivity factors? As I keep stressing willingness to borrow on aggregate, which is in part a function of the overall age structure, does seem to be one of the important indicators.

Of course, as the report also notes, demography not only influences the demand for credit, it also influences the long term demand for housing:

The total population is virtually static, because natural declines are being just offset by immigration. Household numbers are also static at 20.5 million. Overall, population is decreasing in the major cities and increasing in the smaller centres. Moreover, there has been a long-term trend, as elsewhere, for residents to move away from the city centre to the suburbs – although this trend is now associated with a counter inwards movement towards the city centres.

In the absence of increased immigration, the total population is expected to stabilise, and then decrease by around 4%, over the next 15 years. Moreover, the population is ageing. There was a very high birth rate in the 1960s but there has been a subsequent demographic transformation, which gives Italy the lowest birth rate amongst OECD countries. As a result, the population is expected eventually to fall significantly, by 6.6. million between 2020 and 2050; by that latter date the population will be a full 15% less than now.

Life expectancy is also high and rising. The result of these changes is that the country is experiencing one of the greatest demographic shocks of all advanced nations. People over 65 years old currently equal 10 million - about 18% of population - and the number will grow steadily to reach an estimated 16 million by 2040. By 2050, 31% of the population will be 65 or over. Such a transformation is likely to have profound effects on the housing market.

Both the decline in population and increasing share of the elderly in it are also stronger trends in the northern regions than elsewhere. In the absence of significant inter-regional or international migration, significant housing surpluses may begin to arise in Northern Italy a decade from now onwards.

So an ageing and potentially declining population (Italy's natural rate of population growth went negative back in 1992, and population has subsequently only risen due to inward migration) would seem to logically exert an influence on demand for the total housing stock, with the one wild card here really being immigration, to which topic we will now move on.

Now the most recent comprehensive study of migration into Italy to appear online is a paper presented by Giuseppe De Bartolo of the University of Calabria: Immigration in Italy, The Great Emergency. The paper makes a useful read, and this despite the fact that the tone and title of the paper is at times rather alarmist - and despite the fact that he clearly overestimates Italy's relative position in the migrants league when he states that "Italy with an annual migratory total of 300 thousand people is preceded only by the United States, whose total is of about a million people". Numerically speaking this latter statement is clearly false, since with an inward flow of somewhere between 500,000 - 600,000 annually Spain obviously has been receiving more migrants than Italy (as, possibly, has the UK in 2004 and 2005 with so many East Europeans arriving), whilst proportionately certainly Ireland (and probably Greece) have almost certainly had rates of inward migartion which have exceeded the Italian inflow. And this has been the case despite the fact that Italy is ageing much more rapidly than any of these other societies, and thus could arguably be more in need of younger migrants. So the fact that Italy has lagged behind in all of this is not without significance. The question is that even if you need migrants, they may not come, and what we need to look at here is what the actual drivers of migration are.

Well, according to the Italian National Statistics Office (ISTAT), on January 1st 2006 there were 2,670,514 resident foreigners legally in Italy, which constituted an increase of 268,357 (or 11.21%) in relation to the same population in 2005.




Now rather than focusing on the absolute numbers of foreign born population in Italy, what is more interesting for us to focus on here are the flows of migrants into Italy over the last decade or so. As the chart below clearly shows these "official" flows tend to rise and fall in relation to the various "regularisations" which have taken place. It should always be remembered that this data only gives a partial view of the actual rate of flow at any moment in time, since during a "regularisation" immigrants who are already in Italy suddenly "appear" as and when they become legal. This is rather different from the situation in Spain, for example, where immigrants - despite having no legal status in the country - normally register directly with their local town hall since subsequent legalisation often depends on the date of this initial registration. As a result we normally have reasonably reliable data on the migrant presence in Spain. Turning to the chart, we can clearly identify the regularisation peaks:



(Source: Blangiardo G. C. & Molina S. (2006). Immigrazione e presenza straniera. (In Fondazioni Giovanni Agnelli (Ed.) Generazioni, famiglie, migrazioni. Pensandoall’Italia di domani. Torino.)

These correspond to the four major regularisation processes in Italy, as Giuseppe De Bartolo explains:


Between 1990 and 2002 the Italian governments passed four regularisation acts: with the law n. 39 of 1990 (the so-called Martelli's Law) regularised 218 thousands of unauthorised migrants, most of them were Africans and Asiatics. In comparison to the following regularisations, with the Martelli law there was the greatest number of irregulars in comparison to the legal component (120.9 rectified for every 100 regular foreigners - limiting ourselves to the immigrants that originate from the countries of strong migratory pressure). This is to be attributed to the circumstance that the law imposed for the regularisation, which was only to show to have already been in Italy on the date of December 31st 1989. On the occasion of the regularisations favoured by the DL (Decreto Legge - Law Decree) 489/95 (Dini decree) and with the DPCM (Decreto del presidente del Consiglio dei Ministri – President of Council of Ministers decree) of October 16th 1998 the rate of irregularity appeared less because of the greater rigor of the norms. With the first provision (Dini decree) 244 thousand people were regularised, while with the second 217 thousand irregulars were regularised. With the two provisions the citizens of central eastern Europe profited the most because of the increase of the illegal flows coming from Albania and Romania. The law n.189 of 2002 (the so-called Bossi-Fini) can be considered the most important legislative measure in this matter. This law has allowed 647 thousands to be regularised; a number just less than the residence permits emitted altogether (680 thousand) on the occasion of the previous provisions since 1990 (Istat 2005)

The table below also provides a convenient summary of the flows across the years.





In this context it is perhaps worth noting that the most recent data on migration into Italy comes from the national statistical office (ISTAT) in the form of three relevant reports:

Popolazione residente e stranieri residenti nei comuni italiani (April, 2007), and

La popolazione straniera regolarmente presente in Italia (April, 2007), and

Indicatori demografici (March, 2007)

These reports are unfortunately only available in Italian.

What they reveal is that (and as previously noted) on 1st January 2006 there were 2.670.514 foreign nationals legally resident in Italy, and this was an increase over 2005 of 268.357 (+11.2%). The 2006 increase was, however, considerably below the increase experienced between 1st January 2005 and 1st January 2004 (+411.998, or +20.7%) and that of 1st January 2004 over 1st January 2003 (+440.786, or +28.4%).

Thus between 1 January 2006 and 1 January 2003 there was an increase of legally present foreign residents of 72% (or + about 1,120,000 people). This gives a flow roughly 350,000 people a year, and this may give us some idea of the real rate of flow. As a result on 1 January 2006 this population constituted some 4.5% of the total Italian population (which was 58,751,711) a level which was still well below the general "Old EU" average.

Again this increase can perhaps be seen in the chart below:





One of the most significant impacts of this immigration will have been to nudge the Italian median age slightly down (but only slightly) since the median age of the foreign born population is 30.8 while the median age of the is possibly around 43. ISTAT estimates the median age of the whole population (nationals and foreign born) at 42.6 on 1st January 2006.

The impact of all of this on the Italian population pyramid can be seen below (again remember to click if you want to see better).




So the arrival of migrants has had some small impact on the Italian age structure, and in particular in the large 30-40 age group, but again as can be seen from the pyramid cohort size is about to decline rapidly and to exert any notable impact on the ageing process these migrant inflows will need to increase dramatically. Otherwise, as is clear from the pyramid, Italy is likely to suffer a grave shortage of younger workers over the next decade or so, especially given that a significant percentage of the existing migrants are in the 30-40 age group, and a decade from now these will be in the 40 - 50 age group.

However inward migrant flows do not constitute the whole picture here. Italy (just like Germany) is also experiencing a significant outflow of qualified and educated young workers:

since the mid-1990’s the share of college graduates among emigrants from Italy has become larger than that share among residents of Italy. In the late nineties, between 3% and 5% of the new college graduates from Italy was dispersed abroad each year. Some preliminary international comparisons show that the nineties have only worsened a problem of ”brain drain”, that is unique to Italy, while other large economies in the European Union seem to experience a ”brain exchange”.

This state of affairs is a pretty significant one, especially in the light of the fact that Italy's population has not been replacing itself since the early 1990s (ISTAT, latest data, PDF link). Since that time there has only been a continuing population increase as a result of inward migration. But, as this Lavoce article stresses, the balance in human capital terms is hugely negative here. That is to say, the inward-migration that is currently taking place in Italy is extremely important in labour force terms, but this added man- and woman-power can only serve to make the path of the Italian economy a sustainable one if at the same time young educated Italians stay and enter the labour force in much more productive, higher-value activities. This position becomes especially important when we bear in mind the large cohort size reduction Italy is about to experience in the 30-40 age group.

Now, if we turn to the Spanish case, we can see from following chart the rapid increase in the foreign born population in Spain in recent years:



Source, INE, Spanish national statistical office.

Now what we can see is that since 1st January 2000 (and rounding out just a little) the foreign born population resident in Spain has increased from nearly 1 Million, to around 4,150,000 on 1st January 2006. This is a total of 3,150,000 in 6 years, or an average of a little over 500,000 a year.

Now it is important to keep in mind that this increase comes from 2 sources, the migrants themselves, and residents from other parts of the EU who have bought homes in Spain during the construction boom (the greying northern population phenomenon). While it may be important for analytic and compositional reasons to distinguish between these two populations (eg for growth accounting purposes), for our present concerns they are a by-product of one and the same phenomenon: Spain's construction boom (Italy could and should, after all, have been attracting these potential clients for her services).

So the question is, do we here have a measure (or rough proxy) for the differences between having and not having a construction boom at a critical moment in your history. As I estimate above, Italy has been attracting migrants at the rate of around 300,000 a year, while Spain has been attracting something more like 500,000, and it should be noted that the Spanish population is (or rather was) only approximately two thirds of the Italian one, so to have had the "Spanish disease" maybe Italy should have been attracting migrants at a rate of around 650,000 a year, or more than double what she was attracting. Since these two countries are quite comparable culturally (that is, it is in principle no more difficult for an undocumented migrant to arrive in Italy and stay than it is in Spain) the differences in the flows most probably relate to differences in the underlying dynamics (ie the creation of employment) and these differences can also be seen in the relative growth rates of Spain and Italy.

But again, Italy's problem doesn't end here. Having such a strong inflow of migrants has clearly revolutionized the whole Spanish employment situation, and creating work for unskilled migrants at the bottom of the ladder has lead to increasing employment of more qualified Spaniards higher up (and note here the OECD productivity chart above, since these different retention and employment creation profiles for the young and educated are probably one of the features in the Spanish productivity performance, which while not outstanding is certainly much better than Italy's).

One key data point here tells all. In the world there are approximately 3.5 million people who live outside Italy and hold Italian passports. Many of these live in Latin America and are the descendants of earlier Italian immigrants. Yet, what we have noted here in Spain is a constant stream of young people entering from LA using Italian passports (young educated Spanish people are not - other than for external experience reasons - leaving in any significant numbers). Strangely this is not happening in Italy, and the young and educated are not arriving in significant numbers, in fact precisely the opposite is happening, young Italians are leaving, and in growing numbers.

So I conclude with a question: is there yet another inflection point to be looked for here, the one between a level of migrant flows which helps you just keep going (and even produces only horizontal growth in activity terms) and one which accelerates the rate of increase in domestic consumption to such a rhythm that it eases a rapid transition to a services economy (ie vertical growth), and in doing so not only gives employment to the native young and educated but even attracts such people from outside in growing numbers (the Spanish phenomenon)?

What I am saying is that everyone would clearly like to attract educated migrants, but not everyone can. Is the difference a function of your overall growth rate, and is housing here a key factor? That, I think, is the question. To be, or not to be (built).