Germany’s Rocky Road to Full Employment

The Austrian economist Joseph Schumpeter once wrote on Marx’ economic theory that “…Marx’ economic ideas are embedded into a variety of very hot (in German: “dampfende“) phrases which gives them a temperature which is far from the own temperature” (from “Kapitalismus, Sozialismus und Demokratie“, own translation). The same can be said about the structural reform debate in Germany.

FT Deutschland’s chief economist Thomas Fricke collected a number of quotations, which are worthwile to be translated (own translation, German version in brackets):

“Without Hartz IV the job market development would have developed somewhat more weakly” (Ohne Hartz IV wäre die Arbeitsmarktentwicklung etwas schwächer verlaufen) – IAB institute Nuremberg

“At least one third of the new jobs are to be due to structural reforms” (Mindestens ein Drittel der neuen Jobs sind auf Strukturreformen zurückzuführen) – Holger Schmieding, Economist

“The improvement on the job market is above all demand-driven, not supply-driven” (Die Verbesserung am Arbeitsmarkt ist vor allem nachfrage-, nicht angebotsgetrieben) – David Soskice, Economist

“A majority of the decrease in unemployment goes back on improvements of supply” (Ein Großteil des Falls der Arbeitslosigkeit geht auf Angebotsverbesserungen zurück) – Michael Burda, Economist

“The effects of Agenda 2010 can not yet be quantified” (Man kann die Effekte der Agenda 2010 noch nicht quantifizieren) – Hans-Werner Sinn, Ifo-Institute

“The growth rate of potential output might be higher today around a quarter percentage point.” (Das Potenzialwachstum dürfte heute um einen viertel Prozentpunkt höher sein.) – German Bundesbank

“Germany’s growth potential did not increase” (Deutschlands Wachstumspotenzial ist nicht gestiegen) – Adam Posen, IIE

“The government of Mrs. Merkel introduced the fourth reform-political phase in November 2005” (Die Regierung von Frau Merkel hat im November 2005 die vierte reformpolitische Phase eingeleitet.) – IW institute (close to employers’ association)

“Trend growth rose since 2003 around 0,8 percentage points to 1,7 per cent.” (Das Trendwachstum ist seit 2003 um 0,8 Prozentpunkte auf 1,7 Prozent gestiegen.) – IW institute (close to employers’ association)

“Der Aufschwung seit 2005 gehört Frau Merkel. Diese Hypothese ist offenkundig falsch” – Hans-Werner Sinn, Ifo-Institute

Tu sum up: Economists have no common idea about: (i) the sources of the German unemployment problem, (ii) the growth effect of supply-side reforms, (iii) the identification of policy actions “as reforms”, (iv) who owns the “fruits” of reforms if there are any (see point (ii)). We could stop here by using the famous quote by Keynes on how politicians use ideas of economists.

However, different opinions in an open society are not bad, as Hayek and even Schumpeter once argued. Michael Burda — affiliated with the Schumpeter institute at the Humboldt University Berlin — calls for a calm and rational discussion, a position I would completely agree with. The high unemployment rate in Germany is a problem which can not be taken seriously enough. And I completeley agree with his argument, that one has to look at demand and supply schedules simultanously to investigate any economic issue. [1]

In the following I will argue, that the confusion that might arise reading the quotes is due to (i) the usage of conventional wisdom too often as a “one-size-fits-all-approach” without having a careful look at the empirical evidence, (ii) the interactions between supply and demand that are very important but far from being a one-way street. Under certain instances it can be argued that supply might create its own demand but the inverse relationship might hold as well (as Larry Ball once argued, the position of the NAIRU might shift over the cycle and the shift is not independent from the demand effects ….).

Defining conventional wisdom on the nature and aims of structural reforms in Europe

I collected several quotations from this blog which characterize the important aspects [2] of conventional wisdom in European economics debates nicely:

“It is at this point well recognized by economists that the low economic performance of Europe, and the euro area in particular, has to be imputed to low productivity growth associated with lack of structural reforms.” (posted by Ester Faia)

“Labor markets have been at the core of Europe’s policy agenda for many years, particularly after the introduction of the euro.” (posted by Ester Faia)

“In many respects the new laws cut inframarginal benefits but left marginal work disincentives little changed – hardly what an economist would have recommended. Nevertheless, they applied the medicine where it was needed; the cut in the duration of benefit in the insurance program (Arbeitslosengeld) and more important a reduction in the benefit levels and qualification for the the follow-up needs-based program (Arbeitslosenhilfe) were revolutionary. And contrary to what my critics say, these medicines have been recognized by prestigious analysts like Richard Layard and Steve Nickell as a central explanation for the long-term reduction of unemployment in OECD countries.” (posted by Michael Burda)

It is not unfair to characterize important features of conventional wisdom positions in Europe by three propositions:

  • The main economic problem in Europe is a lack of structural reforms.
  • Labor market reforms are of highest priority. [3]
  • Labor market reforms are the central explanation for a reduction in long-term unemployment in OECD countries.

The problem with “one-size-fits-all” and statistical significance

Quite often “one-size-fits-all” clothes do not fit. Let’s illustrate the point. It is great to find a common ground for arguments in the work by esteemed economists as Richard Layard and Steve Nickell to show the relevance of our first point. Both (together with Richard Jackman) are well-known as the authors of one of the best-written (and possibly best-selling) books in this field, which gives an excellent synthesis of New Keynesian arguments for labor market analysis. Quite recently, the latter of the mentioned authors, Steve Nickell, together with co-authors (NNO, 2005) presented estimates of the effects of institutional variables on unemployment behaviour. Another recent and often-cited paper is the OECD study of Bassanini and Duval (BD, 2006). The empirical finding of both studies that can be summarized in the words of Wendy Carlin and David Soskice (2007):

“… employment protection legislation is not significant in either study, unemployment benefit variables are positive and significant, coordination is negative and significant, and the tax wedge is positive and significant. NNO find a positive and significant effect on unemployment of the change in union density, but the level is not significant in BD. The index of product market regulation is positive and significant in BD, as is the real interest rate in NNO.” (Carlin and Soskice, 2007, p.23)

However, statistical significance does not imply, that the factor at work is of great importance or at least of equal importance empirically in all OECD countries. Carlin and Soskice (2007) take the estimated coefficients of the Nickell study and present dynamic simulation for unemployment in Germany and UK.

Figure 1: Dynamic simulation of the model in Nickell et al. 2005


The upper panel shows dynamic simulations under changing institutional settings and demand control variables compared with the actual unemployment rate, the lower panel performs the same simulation but keeps the institutional variables constant at 1960 levels. The result is striking. The change in unemployment in UK is to a large extent be explained by changes in institutional variables; in Germany, the institutional explanation simply fails — even if the study revealed a significant effect in the panel study. [4]

What can we learn for this story? Not more or less than that the structural unemployment story is not a “one-size-fits-all” approach for all OECD or European countries. A story that might fit convincingly in the UK case does not necessarily hold for Germany. The long-term unemployment problem in Germany seems to be driven largely by demand factors, an interpretation which was always shared by Larry Ball, the late Franco Modigliani, Ronald Schettkat, and Robert Solow to name just a few. [5]

Demand and supply side interactions revisited

Even if the “structural” explanation of the evolution of unemployment is not completely convincing for the German case, it could well be the case that the German labor market needs “structural reforms” to get rid of the problem of long-term unemployment. This is an issue which has been debated since decades and can be theoretically justified i.e. by decreasing search efforts, lost skills and incentive problems. The longer the problem lasts, the more difficult might it be to reduce long-term unemployment effectively. Some people even argue, that a convincingly strong regime shift in the reform policy will initiate a consumption boom which might even “kick-start” the economy.

There are several channels why these lock-in effects might arise. Hysteresis, insider-outsider models, efficiency wage considerations… Economic theory offers a lot of explanations. Asymmetries in financial markets, credit and balance sheet channel views, … All these theories have in common that a link connects both of Marshall’s scissors.

Another channel originates from the open economy characteristics in a currency union. The New Keynesian textbook workhorse model allowing for multiple unemployment equilibria in a open economy with a dual labor market (export-oriented versus domestically orientied) and liquidity constrained consumers tells such a story. In that view the long-lasting unemployment performance, at least to a large extent, can best be explained by demand shocks — as in Carlin and Soskice (2007).

What does this mean for Burda’s arguments regarding Germany’s upturn? In short, it means that structural reforms might go hand in hand with negative demand repercussions and the effects not negligible. Under certain circumstances, the reforms might even produce perverse effects in such a model. Let’s briefly repeat what Burda said:

All anecdotes aside: the Hartz reforms were designed to increase labor supply. Increasing labor supply generally implies lower wages, increasing labor demand, and more output, especially for an open economy like Germany with an export exposure now well over 40%.

This is a very interesting issue — both theoretically as well as empirically. Indeed, it is the core of the argument for the confusion in the German debate. In our interpretation, the data simply do not fit into this story. As we outlined before, Germany faced a decade of wage restraint and real depreciation — and until recently no sign of recovery. Consumption and investment were anemic for a couple of years (and consumption is yet far from being very strong). On the other hand, export exposure was high already in the 1990s, so what is the missing link to explain that long slump?

As indicated above, the German story fits well into a New Keynesian framework with an economy having an export-oriented and a domestic sector and a high proportion of liquidity-constrained consumers (a fact which might be highly relevant for Germany), where a wage restraint policy might have perverse effects. As Carlin and Soskice (2007) put it:

A cut in the real wage has two opposite effects on aggregate demand: on the one hand, it improves international competitiveness by lowering the real exchange rate, increasing exports and reducing imports and hence raising aggregate demand and lowering unemployment. If this was the only effect, the aggregate demand curve in the diagram would be upward sloping. But a cut in the real wage also lowers consumption assuming that liquidity constraints prevent workers from smoothing consumption in response to the fall in the real wage. In a small and therefore more open economy we assume the first effect dominates so that a lower real wage leads to an economy wide improvement in unemployment (…); and vice versa in a large economy (…); (in both cases employment in the export sectors should improve).(Carlin and Soskice, 2007, p. 10).

The logic of the argument can best be seen in the following figure.

Figure 2: The “perverse” demand effect and unemployment in the Carlin and Soskice (2007) model


Since the debate between Keynes and Pigou, Keynesian economics in the tradition of Leijonhuvhud, Tobin, Blinder, Akerlof and Yellen was always quite sceptical that supply creates it own demand in a smoothless way. Therefore economists “invented” monetary and fiscal stabilization policy and in an open economy setting external demand can of course be helpful in stabilizing the economy. However, under certain circumstances and even under stable aggregate demand schedules, the demand curve can be very steep in the unemployment/ real wage space — even negative under reasonable assumptions. That would call for a further stabilization policy action. The question is: Did Germany turn the “perverse” effect into a “normal” effect and what is the cause for it? In our view it is the accumulation of three demand effects: a remarkable increase in competitiveness over a decade (which might dominate the negative repercussions at the end), the end of the slump in construction and a turn in the fiscal policy stimulus.


To summarize, structural reforms and their interaction with the real world are a tricky issue. Economists can neglect certain possible negative repercussions in theory but in policy-making we have to consider both of Marshall’s scissors and all possible interactions between them.

[1] Sebastian Dullien and I outlined our scepticism regarding the Hartz IV explanation for the recent upturn in a number of posts (here, here, and here) — briefly: (i) the upswing has strong cyclical characteristics, (ii) the recent stock market boom is mainly driven by a fundamental but not a German factor, (iii) wage restraint is certainly a factor at work but the beginning of this development dates back much longer than the famous Schröder agenda.[2] Of course, there are other aspects as well.[3] To be fair, Faia’s statement is about a smooth functioning of the currency union — an argument to be distinguished from the reduction in long-term unemployment.[4] See for an early statement: Logeay, Camille (2003): “Arbeitsmarktinstitutionen und Arbeitslosigkeit: Stand der wissenschaftlichen Diskussion”, DIW Wochenbericht , 22/2003.[5] This list can be extended. It would include Heiner Flassbeck, Gustav Horn, Wolfgang Scheremet, Friederike Spiecker, most of DIW team before 2004 and the IMK team after 2005. Of course, the recent minority vote of Peter Bofinger, member of the Council of Economic Experts fits. See also Fritsche, Ulrich; Logeay, Camille (2002): Structural Unemployment and Output Gap in Germany: Evidence from an SVAR Analysis within a Hysteresis Framework. Berlin (= DIW Discussion Paper. 312) and Gottschalk, Jan; Fritsche, Ulrich (2005): The New Keynesian Model and the long-run vertical Phillips Curve: Does it hold for Germany? (= DIW Discussion Paper. 521) . Also: DEP Discussion Papers. Macroeconomics and Finance Series 1/2006.

2 Responses to "Germany’s Rocky Road to Full Employment"

  1. David F. Milleker   November 10, 2007 at 8:17 am

    Ulrich, Thanks for this excellent piece. It’s not only good reading, it very much tunes in with my own line of thinking that supply-side reforms should be matched by appropriate demand side-policies in order to achieve effects. I’m especially grateful for the distinction between small and large open economies. As long as we don’t shatter the belief of Germany belonging in the first category, Germans will not be making any headway.

  2. Guest   November 11, 2007 at 5:24 pm

    The comparison between the UK and the German experiences is striking. Since both countries are comparable in size and openness the main difference between the higher and lower unemployment (partial) equilibrium may indeed reflect liquidity constraints and consumption smoothing as mentioned in your report. The main issue then is if access to credit markets leads to a higher permanent income path in the long run. Current industrialized country experiences are worth watching closely.