http://eiop.or.at/eiop/texte/2004-005a.htm
: 12.6.2004| This paper's comments page | Send your comment! to this paper |
| The European Union benchmarking experience. From euphoria to fatigue? |
| Michael Zängle |
| European Integration online Papers (EIoP) Vol. 8
(2004) N° 5; http://eiop.or.at/eiop/texte/2004-005a.htm |
| Date of Publication in
: 12.6.2004 |
| | Abstract |
Back to homepage |
PDF | | This paper's comments page | Send your comment! to this paper | |
1
“We are all benchmarkers now!” – this enthusiastic assertion by the then President of the European Commission, Jacques Santer, was echoed some years later in question form: “All benchmarkers now?” (Sisson, Arrowsmith and Marginson 2002). The overall answer may be summarised as “yes, nearly all, indeed, surprisingly many”. The speed and spread of acceptance of the EU benchmarking exercise are impressive, indeed.
Santer’s statement is a political statement, which is a reminder to understand “we” and “all” as an admonition and moral obligation. Since its beginnings the promotion of the benchmarking approach has been wrapped up in terms of co-operation, consensus, dialogue, and involvement of as many partners as possible, of as many stakeholders as possible, of as many businesses and corporations as possible, be they big or small, of civil society, of men and women, of conventional and non-conventional NGO’s, of actors at national and at sub-national levels, in short, of ‘all of us’.
Since the benchmarking adventure was kicked off by the particular interests of the European Round Table of Industrialists (ERT, Tidow 1999) there was some need for Santer’s all embracing, encompassing admonition. Turning ‘all of us’ into benchmarkers asks for the reconciliation of and mediation between conflicting interests of class, fractions, interest corporations, nations, regions and gender, to mention only the most important cleavage lines. The developmental process of EU benchmarking may be conceived of as successively broadening the approach from (big-)business sector benchmarking (ERT) to socially and politically inclusive benchmarking, inviting and admitting a wide spectrum of interests to participate. In this sense, the ‘We’ in Santer’s dictum relates to ‘all of us’ as ‘citoyens’.
2
Obviously, the social and political inclusiveness of benchmarking will impact on the design and philosophy of the benchmarking system. Both the definition of theoretical concepts and the concerning operational definitions of performance indicators take into account the necessity of consensus among conflicting interests and world-views. A great deal of discursive labour, ‘meaning-making’ (Jacobsson 2002) and selective ignorance (Straßheim 2001), has been spent in redefining e.g. the labour market in such a way as to keep to the Round Table of Industrialists’ premises on the one hand, and to allow as many particular interests as possible to become involved on the other hand. Judged by standards of ideology formation, this efforts are paying off. What was started as a vision of particular interests now asserts itself as a general consensual world view accepted and endorsed by ‘all of us’.
Evidently, the benchmarking exercise has been successful both in broadening its social inclusion and in enriching its vision, a vision which is now deeply rooted in political philosophy and the history of political thought and democratic theory. Take as evidence the OMC research forum of the European Union Centre at the University of Wisconsin-Madison (Overdevest 2002, Sabel and Zeitlin 2003) to mention only one of several possible examples.
A third dimension of success may be added, the dimension of implementation and technical feasibility. It is true there has been severe criticism as to the timely availability and to the reliability and validity of indicators. In addition, the econometric foundations of the strategy are occasionally deemed insufficient (Chalmers and Lodge 2003). But obviously, the benchmarking system works well enough to allow for the decision whether the overall process was kept on track or not.
While the implementation of the benchmarking approach is a success story, the implementation of the benchmarking targets is not. When it comes to delivering ‘Lisbon’ by 2010, failure is imminent. “The European Union is at risk of failing in its ambitious goal, set at Lisbon in 2000, of becoming by 2010 the most competitive and dynamic knowledge-based economy in the world capable of sustainable economic growth with more and better jobs and greater social cohesion” (Employment Taskforce 2003: 11).
Making this failure visible even now, years before the deadline, is proof of the technical quality of the benchmarking system, and the possible failure in delivering Lisbon is proof of the political difficulties of strategic target setting. Will this failure in goal achievement and target setting backfire on the benchmarking approach itself? Even though some benchmarking fatigue (Overdevest 2002: 23) might be expected, chances are high that “licking the wounds” of having failed the Lisbon targets will be done in terms of the dominant “embedded neo-liberal project” (Apeldoorn 2003).
This paper attempts to make a contribution to a research question that has been summarised as follows:
“Will there be a ‘benchmarking fatigue’? The long-term learning envisioned by social theorists implies a serious commitment to on-going policy revision and reconfiguration. Any number of possible interruptions -- from cyclical economic downturns to political crises -- could derail commitment to on-going policy learning. Because the success of the OMC is predicated in part on its iterative application, and because consistency in implementation is important, systematic research attention to whether there is a tendency for benchmarking fatigue would be useful contribution to the OMC literature” (Overdevest 2002: 23).
This problem definition has been reconceived and broadened, however. Rather than dealing with any exceptional circumstances that could derail the benchmarking process, the focus is on the possibly detrimental consequences of the regular business-as-usual iterations of the benchmarking cycle. The very functioning and technical performance of the benchmarking exercise may increasingly intensify the discrepancy between what ought to achieved and what actually can be accomplished, both at Community level as well as at the level of (most of) the Member states (EU 15). Disillusionment and fatigue are the likely outcomes of such a development. Since the benchmarking process is part of a ‘meaning-making’ process, which is shaping and re-shaping the world-view of the European project, “fatigue” could acquire the deeper meaning of an ideological crisis. The solution of this crisis may well consist in weakening the very instruments and new-governance techniques, that have been burdened with the high hopes of establishing and defending a deliberative Social Europe. If things turn out this way, benchmarking and policy learning will get weaker, and the neo-liberal branch of the European project will be invigorated.
The present paper elaborates these basic arguments.
3
Benchmarking is expected to support the achievement of the overall goals and objectives of the Lisbon strategy (3.1). Being part of the Open Method of Co-ordination, benchmarking is a special case of “soft” regulation meant to protect against the possibly disrupting forces of the ongoing re-establishment and corroboration of the Single Market (3.2).
The decisions of the Lisbon Summit (2000) may be summarised as the establishment of an integrated strategy for Europe to become the most competitive and dynamic knowledge-based economy in the world, capable of sustainable economic growth with more and better jobs and greater social cohesion (Box 1).
|
Box 1: The overall Lisbon goals and objectives |
|
“5. The Union has today set itself a new strategic
goal for the next decade: to become the most competitive and dynamic
knowledge-based economy in the world, capable of sustainable economic growth
with more and better jobs and greater social cohesion. Achieving this goal
requires an
(EC 2000) |
It is important to keep in mind that the overarching goals, increasing competitiveness and labour productivity, increasing employment, increasing the quality of jobs, and increasing social cohesion, are all meant to be achieved by 2010. By implication, the overall objectives are deemed to be mutually reinforcing, bound together in a virtuous cycle of synergetic upward development. Thus GDP growth and the growth of productivity are planned to be predominantly knowledge-based, which in turn calls for the improvement of human capital and the quality of jobs. A higher quality of jobs will enhance job satisfaction, which in turn will speed up productivity and GDP growth. The European social model, after modernisation, will be less expensive and at the same time protective enough to increase social cohesion, even if the restructuring of the economy as a whole and the fettered deregulation the labour markets may give rise to some risks of social exclusion (Box 1). The challenge of the Lisbon process as a strategy amounts to the task of orchestrating this virtuous cycle of upward development. The Open Method of Co-ordination is expected to support this exercise.
Lisbon 2000 celebrated the baptism of the Open Method of Co-ordination (OMC, e.g.: Hardes 1999; Deppe, Felder and Tidow 2000; Goetschy 2001; Hodson and Maher 2001; Overdevest 2002; Schäfer 2002; Scharpf 2002; Jacobsson and Schmid 2003; Wessels 2003). It was only the baptism, not the begetting, since the OMC as a new governance approach had been developed avant la lettre within the framework of the European Employment Strategy (EES).
“The Open Method was most fully specified for the European Employment Strategy (EES), which came to be known as the ‘Luxembourg Process.’ Its core is an iterative procedure, beginning with an annual joint report to the European Council which is followed by guidelines of the Council based on proposals from the Commission. In response to these guidelines, member governments present annual ‘National Action Plans’, whose effects will then be evaluated in the light of comparative benchmarks by the Commission and a permanent committee of senior civil servants. These evaluations will feed into the next iteration of joint annual reports and guidelines, but they may also lead to the adoption of specific recommendations of the Council addressed to individual member states. In any case, however, "the harmonization of the laws and regulations of Member States" is explicitly excluded from the measures the Council could adopt (Art. 129 TEC). In other policy areas, procedures may be less formalized and less demanding, but the essential characteristics are the same” (Scharpf 2002).
4
Monitoring social cohesion is meant to protect against the downward pressure of the market-enhancing measures of deregulation and liberalisation. If we conceive of the European Employment Strategy, or the Lisbon strategy in general, as an ‘embedded neo-liberal approach’ with guilty conscience (for embeddedness cf. Apeldoorn 2003, for particular EU modifications of the neo-liberal paradigm, cf. Noaksson and Jacobsson 2003, for contextualised and path-dependent learning vs. neo-liberal one-size-fits-all, cf. Hemerijck and Visser 2003), this guilty conscience – far from being merely symbolic –, consists institutionally of soft governance protection mechanisms against ‘social dumping’, downward tax competition and ‘races to the bottom’ (Genschel 2000).
The ongoing implementation and continuous corroboration of the Single European Market and the pressures of tax competition and ‘regime-shopping’ (Lütz 2003: 21) call for compensatory action. The increasingly strict insistence on individual responsibility, the requirement that individuals be prepared to ensure their own employability and to deal with their occupation and health risks on their own might produce an increasing proportion of the work force at risk of exclusion and poverty.
It is true that the concern about a possible ‘social dumping’ or ‘race to the bottom’ has been attenuated by the research of the last ten years. Empirically, there is no evidence of an overall downward trend of social protection in Europe. The same holds true with tax competition. The race to the bottom – as a clear-cut trend – simply did not occur (Gittermann 2002). We should keep in mind, though, that what we observe empirically is the net result of conflicting tendencies (Genschel 2000). After all, there may be downward pressures. They may be offset, however, by counter-forces and counter-trends. For example, as has been argued for tax competition, constant and even increasing tax rates are no proof against competitive downward pressure on taxation. Given the increasing demand of social expenditure in aging societies ridden with unemployment, taxation should have risen proportionally. But this is not what happened. Instead, states preferred to partially shift the burden to public debts. Thus, slower-than-proportional tax increases may quite well be associated with inter-state tax competition (Hemerijck 2002). This type of analysis is pertinent to other aspects of social protection as well.
Thus, Social Europe is always under threat, vigilance against downward pressures is never futile, and the ‘guilty conscience’ of the predominantly neo-liberal Lisbon project needs to be continually stirred.
Benchmarking is expected to play an important role as a watchdog and by inducing upward-directed learning cycles, but measured against the former leftist hopes to establish a Single Social Europe, this soft governance approach will appear as a second best solution (Sabel and Zeitlin 2003: 9). But even as a second best solution, the institutional and procedural establishment of upward-directed learning is deemed to be an important stronghold against neo-liberalism of the Thatcher type. There are many risks, however, that this possible stronghold may turn into a defective device, entailing non-intentional detrimental consequences. One of these risks is connected with too large a discrepancy between aspirations and capacity (section 4).
In this section 4, I will present an overall description of what has been achieved till now, compared to the targets. This description will be done for the European Union as whole (EU-15), and, additionally, for Ireland in particular (4.1.). Considering Ireland separately may prove helpful for the following reasons. Ireland is an impressive performer in regard to many of the Lisbon objectives. So even if Europe (EU-15) as a whole is lagging behind the targets, Ireland seems to be proof of the possibility of the expected dynamism, albeit under exceptional conditions. Unsurprisingly, Ireland has become a kind of model, earning considerable praise, e.g. in the latest “Recommendations” (CEC 2004b). Thus, looking merely at the mean performance of the EU-15 as a whole might leave us with too pessimistic an impression. There are important and striking exceptions, and certainly Ireland is one of them. Contrary to the Scandinavian good performers, Ireland did not enter the EES and Lisbon process from a high level of performance, but had to catch up from a difficult starting point.
Section 4.2. deals with the question of a possible upward convergence of the member states’ employment rates. For even if Europe misses the targets, there may be some development in the right direction, which is at least partially attributable to the benchmarking exercise. A second question will be dealt with in section 4.2.: how stable are the performance rankings over time? Is there a danger of subliminally labelling member states as ‘heroes’ or ‘villains’ after several iterations of the benchmarking cycle, thus inducing frustration and withdrawal rather than a positive learning climate for those who are most in need of progress? In section 4.3. I will return to the Irish case, but from the opposite perspective. Rather than showing why Ireland is a best case, focus is now on possible deficiencies of this model as they are detectable by the benchmarking indicators. Selectivity in the development and use of performance indicators is one of the mechanisms of ‘model-making’ in particular and ‘meaning-making’ in general.
5
The Lisbon strategy was born at what has often been called the dot-com summit (Lisbon 2000). Years of European growth had preceded this famous ‘Spring’, and the IT-driven long-lasting boom of the US economy gave a good reason for speculations about a new forthcoming growth paradigm. Once again, dreams were dreamt about economic growth free of cyclical depressions, and Europe’s decision to engage in a knowledge-based strategy to be monitored and re-focused at annual ‘Spring meetings’ was deemed to be the spring of a more dynamic European economy. This optimistic mood may partially explain the impressive boldness of the Lisbon targets.
Meanwhile however, the new growth paradigm partially disclosed itself as a bubble, and terrorism, or the war on terrorism, dealt a second blow to the overall optimism of the year 2000. Against this background, the London-based Centre for European Reform (CER, http://www.cer.org.uk/about/index.html ), basically supportive of the Lisbon agenda, announced its Lisbon Scorecard IV in a mood of disillusion and disappointment:
“With cynicism, even derision – this is how many Europeans look at the EU’s key economic target, namely to become the ‘most competitive and dynamic, knowledge-based economy in the world’ by 2010. The ‘Lisbon’ reform process, launched by EU leaders in 2000, has not lived up to expectations. The EU is not catching up with the US. Rather, it is falling further behind. As the Lisbon agenda approaches its half-term review, key reform targets, such as creating jobs or improving Europe’s science and technology performance, look out of reach. EU governments are at a loss about how to kick-start the Lisbon agenda.” (CER March 2004).
Indeed, the European Commission themselves admit that Europe will fall behind the target to outperform the United States by 2010 (CEC 2004a). Even worse, far from outperforming Europe’s main benchmarking comparator in GDP growth and competitiveness, the gap seems to be widening (OECD 2003). EUROSTAT data (Structural Indicators) show a high risk of the Lisbon targets rapidly moving out of reach.
Figure 1 reports the EU15 growth performance compared to Europe’s main comparator US (Sisson et al. 2002: 18), drawing on EUROSTAT structural indicators (ecobac.csv). There are certainly no clues whatsoever in figure 1 that European growth rates might become threatening to the United States. Clearly, the strategic target of an average annual economic growth of 3 %, necessary to create 20 million jobs by 2010, seems to be beyond reach.
Being only an exceptional example, Ireland as the European tiger shows a GDP growth development that would have been necessary for EU15 as a whole if Europe were to beat the United States by 2010. Enthusiasm rather than sound reasoning seems to be necessary to continue to believe in a possible success. This is true even after the application of more sophisticated methods than the mere inspection of Figure 1 (CEC 2003, Employment in Europe 2002, chapter 2; Employment Taskforce 2003, Chalmers and Lodge 2003).
While it seems really difficult to fully believe in the Lisbon endeavour, its falling short of the targets is not equivalent to full-fledged failure. After all, things might be even worse without the Lisbon efforts. Indeed, these efforts may have created considerable added value, even if not enough has been achieved to really measure up to the targets. This added value may consist, for example, in increased economic competitiveness or improved capacities in governance and meta-governance, or even both. Of course, causal proof in the sense of an econometric separation of a benchmarking effect is virtually impossible (Barbier 2004: 21) given the difficulty of convincingly constructing the counterfactual situation (what would have happened without the benchmarking exercise?). Thus the following sections are confined to a merely descriptive approach to data analysis, taking the benchmarking data at face value.
6
Benchmarking the Lisbon Agenda induced a rapid development of indicators, performance criteria and targets, supporting several OMCs that have been introduced into additional policy fields and into the Lisbon process as a whole (Hodson and Maher 2001, Sabel and Zeitlin 2003). Burgeoning indicator construction has even aroused some concern the benchmarking process might sprawl into a self-contained, autistic bureaucracy, lacking focus and purpose (Chalmers and Lodge 2003). This danger has lately been averted, however, by the streamlining efforts of the Commission (CEC 2003). The short-list of structural indicators now consists of no more than 14 indicators supported by a much richer publicly accessible database. The present paper draws on this database http://www.forum.europa.eu.int/irc/dsis/structind/info/data/index.htm , table 1).
The overall goal of the Lisbon strategy is sometimes summarised as closing the gap in GDP per capita (indicator 1, table 1, above) from the most dynamic competitors in the world, with the US being the main comparator in the European benchmarking exercise (Figure 1, above). As mentioned above, eliminating the gap is less than what ‘Lisbon’ actually aspires to (Box 1). Indeed, Lisbon calls for outperforming the world’s leading geo-economic regions, trying to become the leading growth pole among the competing triads.
The indicators 2 to 14 (table 1) are deemed to be the driving forces of GDP growth: “Other indicators in the list cover the most important factors driving GDP growth” (EC 2003). While this assumption may be valid for “labour productivity” (indicator 2 of the short list), this is not as evident for the rest of the indicators. This streamlined picture of mutually re-enforcing policies, actions and activities depicted by respective indicators, holds true only under certain specific conditions. Absent these conditions, collisions and inconsistencies among core objectives may prevail (cf. section 6.).
One of the expected outcomes of OMC benchmarking is upward convergence of (economic) performance among member states. Being deeply rooted in approaches of policy learning (Hermerijck and Visser 2003), EU benchmarking is trying to achieve more than mere monitoring (Kissling-Näf and Knoepfel 1998). Apart from monitoring, benchmarking supports striving for excellence by learning from best cases (CEC 1996; Löffler 1996; Schütz, Speckesser and Schmid 1998; Tidow 1999, Deppe et al. 2000; Mosley 2000; Straßheim 2001, Mosley, Schütz and Schmid 2002, Sisson, Arrowsmith and Marginson 2002). Peer comparisons among member states at variable territorial levels are meant to induce self-reflexive learning cycles, thereby speeding up the laggards and encouraging them to catch up with ongoing improvements. The expected outcome is an overall upward trend with increasing inter-state homogeneity of performance (upward convergence).
To analyse this point in more detail, I will focus on the employment performance (indicator 3 in table 1). In monitoring terms, some upward convergence in employment performance may indeed have happened in Europe during the last decade (table 2). Measuring best performance in employment promotion by calculating the unweighted mean of the employment rate of the three best performing member states, and measuring poor performance by calculating the unweighted employment rate mean of the three poorest performing countries, we may operationally define convergence as a percentage level of poor performance compared to best performance.
In table 2, dividing the values in column 2 (poor performers’ employment rate) by the values in column 1 (best performers’ employment rate) leaves us with the convergence indicator expressed in percentage terms in column 4. For monitoring and benchmarking purposes, this rather crude procedure will be sufficient, whereas causal analyses would require more sophisticated econometric methodology.
For example, poor performance by 1990 with an employment rate of 52,3 % (un-weighted mean of the three poorest performing countries) makes up for 68 percent of best employment performance (employment rate of 76,8 %, unweighted mean of three best performing member states). From 1990 to 2002 there is a convergence of 8 points, with three years of stagnation included during the period of economic downturn.
An optimistic interpretation of table 2 will point to column 4, which at least at a descriptive level of analysis shows increasing convergence during the last decade. Pessimists will focus on column 3, which shows rather poor progress in employment promotion. It is true, there is some convergence, but the suggested upward trend, after visual correction for cyclical effects, is far from striking.
7
We may expect the best-performing trio to change its country composition over time. As to the first two ranks, compositional change is moderate, however (table 3). Denmark has taken the lead since 1993, replacing Sweden as first ranking member state. Throughout the complete period under investigation, Denmark never held a rank lower than two (descendent ranking), whereas Sweden alternated between the first three ranks. Thus, Denmark is clearly the prime model when taking a high proportion of the working age population into employment is the goal. The employment rates of Denmark, the Netherlands, Sweden and the UK are above the Lisbon targets, adding some bright spots to the predominantly dismal painting. At the same time we should keep in mind that these bright spots could be there anyway, quite independent of any OMC embedded benchmarking efforts. After all, these countries were leading in terms of employment performance long before the OMC discourse and benchmarking business came to the forefront.
Within the period observed, the group of three poor performers comprises only four elements, with the exceptional case of Ireland included. Ireland almost moved from bottom to top within a decade. Disregarding this exception, the composition of the poorly performing trio is stable, consisting of Spain, Italy and Greece. Italy seems to be the member state actually having most difficulties in establishing and maintaining a sufficiently high employment rate, whereas Spain has managed to initiate improvements in recent years.
Due to their dependence on self-referential learning, efforts of stable poor performers would gain in effectiveness if rates of change rather than sequential cross-sectional measurement played a more front-stage role. OMC feedback cycles and country-specific learning cycles have to be synchronised if learning is to actually occur. In other words, the “sensitivity” (i.e. the temporal variance of performance indicators, Rossi, Lipsey and Freeman 2004: 232), and the temporal and territorial manoeuvrability of improvements have to be consistently related.
To poor performers – if performance is predominantly measured by sequential cross-sectional designs – several iterations of the benchmarking cycle will be stigmatising, attracting the “naming and shaming” (Sisson, Arrowsmith and Marginson 2002: 19) of a scapegoat, whereas manageable improvement rates might ameliorate the benchmarking incitement structure. In terms of sequential cross-sectional performance measurement, for years to come the ‘heroes’ will be of the Scandinavian blend, and the ‘villains’ will be of South European origin (Murray 2004, for ‘heroes’ and ‘villains’ in the Lisbon scorecard). Thus, there is a considerable risk of “provocation”, a type of “fatal remedy” associated with “perverse” benchmarking effects. Instead of compliance, adversarialism and hostility will prevail, making naming-and-shaming a source of distrust (Chalmers and Lodge 2003: 15).
Having dealt with the overall performance of EU15 and some aspects of interstate convergence and variability, I now turn to Ireland as an exceptional case, which is especially appropriate to deal with the complexity of the Lisbon Agenda. This complexity arises from the high aspirations calling for the reconciliation and simultaneous mutual re-enforcement of potentially conflicting goals.
The Irish growth and employment performance during the last decade has been impressive indeed, being one of those exceptional cases in Europe showing competitive dynamism (even if we take into account the rather sharp decline in the most recent years not yet covered by the published EUROSTAT structural indicators). The Irish list of high achievements is long. Raising employment, improving productivity, fighting unemployment, attracting foreign capital, “…social partnership, its tax system, a good regulatory environment and investment in human capital are also major factors” (CEC 2004b: 17). Conversely, the list of “recommendations” is short and basically affirmative of the Irish way (attracting more people to the labour market, making work a real option for all, investing more and more effectively in human capital and lifelong learning; CEC 2004b: 18).
However, if we systematically look for some limitations of the Irish model in EUROSTAT’s database of structural indicators, we come up with the finding that some of the Irish dynamism seems to have been bought with poverty risks. That is why Ireland is not fully living up to the Lisbon ideal of Social Europe. Ireland exhibits signs of prioritising growth over social cohesion (figure 2), thus failing the requirement of the Lisbon Triangle to even increase social cohesion while at the same time promoting growth and employment (box 1).
8
In order to depict as clearly as possible the Irish trade-off between GDP growth and social cohesion, the risk-of-poverty rate after social transfers (cf. section 6.) has been transformed into a performance indicator. To this purpose, the standardised risk-of-poverty rate has been mirrored, i.e. sign-reverted (multiplied by minus 1). Thus, lower z-scores of this indicator signal lower performance. Ireland exhibits declining protection against the risk of poverty during the period covered by the respective EUROSTAT-indicators (figure 2). Extremely high relative growth performance (standardised GDP growth in figure 2) has been “paid for” by high relative risks-of-poverty rates. Moreover, within the period of observation this “trade-off” seems even to be increasing. Unfortunately, the development of social cohesion indicators is somewhat sluggish, with 2001 being the latest available year of measurement.
The above depiction of a divergent development of relative GDP growth on the one hand and the relative decline of protection against poverty risks on the other hand cannot claim to be proof of a “trade-off” in a strict sense of the term. Indeed, the EUROSTAT database of structural indicators presumably cannot support such an analysis. What has been depicted, however, is an increasingly less favourable relative performance in avoiding poverty risks, contrasting with the overall impressive relative performance in regard to most other performance indicators. Beyond this substantial result, the above illustration shows the opportunities of secondary analyses in detecting how the selective use of indicators can be turned into a means of ‘model-making’ and what might be called discursive ‘best-case production’.
To summarise, even if there are some exceptional cases of sufficiently high dynamism, they do not necessarily prove the feasibility of the Lisbon aspirations to install a virtuous cycle of mutually reinforcing processes in implementing possibly conflicting objectives. This is true even of Ireland, which is falling short of the Lisbon requirement of reconciliation and simultaneous achievement of both raising economic growth and increasing social cohesion. If we take the Lisbon Agenda in its full complexity, at best two or three countries, if any, will live up to its main objectives jointly. And these member states happen to be those which had entered the process from a high level of performance, in fact having already accomplished what others had and still have to achieve. For the rest of the member states, among them the larger economies, the Lisbon aspirations appear to be over-demanding.
Now, if the performers persistently wrong the targets, the targets may be wrong and not the performers. This leads to the problem of effective target setting (section 5).
Target setting in benchmarking and management by objective exercises is a highly precarious process, constantly in danger of turning vicious or counterproductive (Schütz, 2001, Zängle and Görtler 2000). Over-ambitious targets will prevent the commitment of realistic players, reducing their involvement to cynical pro forma rituals. In this case, there are high transaction costs just for keeping up appearances and paying lip service. On the other hand, less demanding targets may lead into the same difficulty. Setting fake targets that would be achieved by ongoing processes anyway will produce transaction costs without any added benefits at all, delivering nothing but deadweight and windfall outcomes. In principal-agent negotiations, the principal will try to make the targets as demanding as possible, whereas the agent will try to sell what appears to be demanding but may be brought about just by “chance or nature’s changing course untrimmed” (Shakespeare, Shall I compare thee to a summer’s day).
Targets may be too bold, targets may be too timid, and targets may be fake, with all three alternatives producing detrimental consequences. So why not be cautious and discard political target setting altogether? Discarding political targets amounts to benchmarking by ‘empirical’ benchmarks (Schütz, Speckesser and Schmid 1998, Mosley 2000), as distinct from ‘theoretically derived’ or ‘institutionally or politically agreed upon’ benchmarks. To clarify this point, a look at how the Commission went about it in their Communication to the Spring European Council in Barcelona may be helpful (table 5).
9
Table 5 extracts relevant examples from table 2 in the Commission’s Communication (COM (2002) 14 final). Row 1a in table 5 may be understood as a case of full-fledged benchmarking measurement. Calculating the ratio of EU performance against US performance in percentage terms will produce the indicators in row 1a. Thus, the performance of the European Union in producing wealth amounts to only 65.1 percent of US performance (average of 1999 to 2001), and, even if this truth may be painful to Old Europe, there was no reduction of this performance gap in the period between the Lisbon 2000 and the Barcelona Council 2002. Even if we look at the unweighted mean of the three best performing EU member states, wealth as measured by GDP per capita is under-performing (96.0). Given the overall Lisbon objective of outperforming the best competitors of the world, and given the common practice of selecting the US as the main comparator, the definition of this Lisbon target is rather precise, certainly it is far from being vague: By 2010, EU-Europe’s benchmarking indicator measuring wealth performance has to be above the 100 % reference line.
We now turn to the employment rate in row 2a (table5). Here, benchmarking measurement is less well developed. In benchmarking the employment performance, the explicit calculation of the performance ratio against a specific competitor is omitted, but there is still a clearly specified target, this target being derived from comparison with the United States. Thus benchmarking measurement in row 2a is somewhat less concise than in row 1a, but it still relies on the same informational components (performance value of unit to be compared, target value of performance derived from main competitor’s performance value).
By contrast, benchmarking measurement is of a different kind when public spending on education is concerned (table 5, row 4a). In this case, there is no target specified, and the benchmark is truly and exclusively derived empirically. To find the benchmark, just sort the benchmarking units by their indicators, next select the value of the best case, or the average value of a few best cases. The danger of this targetless benchmarking is obvious. It works even if the overall trend is misdirected. There is no telling where the train is bound, but everybody wants to be the first to climb it. Or, to cite two well-known similes: the simile of sheep, competitively speeding up behind their bellwether, heading for disaster, or, the cases of self-destructive lemming-like learning by imitation. (Sisson, Arrowsmith and Marginson 2002: 16). Both as a protection mechanism against downward pressures of deregulatory market enhancement efforts and as an incitement mechanism to improve an overall trend, this targetless benchmarking is toothless, even if comparisons with the best performers will support some relative upward-directed improvements.
Up to now, I have illustrated 1) full-fledged benchmarking measurement that draws on both performance ratios and explicit political target setting; 2) best-performance benchmarks combined with explicit political target setting; and 3) best performance benchmarks without any explicit political target setting.
All these three alternatives are benchmarking procedures, as distinct from monitoring procedures (Auer and Kruppe 1996). Whereas benchmarking is comparison with best cases to induce competition, comparisons in monitoring are not explicitly learning-oriented. Thus, as far as fostering upward convergence is concerned, monitoring is even less effective than establishing empirical best performance benchmarks.
Whilst, as a rule, benchmarking is meant to support upward convergence and is methodologically equipped accordingly, some monitoring techniques may be designed as protection devices against downward pressures by watching minimum standards (Scharpf 1999). In sharp contrast to monitoring minimum standards, OMC benchmarking is geared towards comparison with best performance, supported by political or institutional target setting to improve the overall trends. Rather than pointing to a downward levelling room-of-manoeuvre (Scharpf 1999) the EU benchmarking procedures pressure for upward convergence.
All in all, the OMC benchmarking system seems to be sufficiently equipped with clearly defined targets and statistical data to check for progress in goal achievement. It is the very technical performance of the system that poses the real threat of making obvious the discrepancy between what is aspired and what can actually be achieved.
To involve even best performing member states, common targets have to be demanding, which entails the risk of over-burdening the laggards and even the mediocre performers. Fixing the level of aspiration in target setting is a choice between a rock and a hard place, however. With targets too ambitious to stay within the realm of manoeuvrability, the benchmarking exercise will expose itself to mockery and derision; with less demanding targets windfall outcomes will reduce benchmarking effectiveness. Obviously, OMC benchmarking depends on experimental trial-and-error double-loop learning about the feasibility and effectiveness of target setting.
10
Unfortunately, here again a dilemma raises its two horns. Double-loop learning, as it increases in intensity, lowers commitment and purpose. On the other hand, without double-loop learning, action will degenerate into purposeless commitment. At any rate, as things stand at interim to 2010, a political and publicly visible revision of the targets as a result of double-loop learning seems to be out of the question.
How did the benchmarking exercise reach this impasse? Why are the targets so unrealistically ambitious? A possible answer refers to the interests of the initiators and driving social forces of the benchmarking process. Representatives of the most competitive ‘capital fractions’ had not much to fear, but could hope to gain a great deal by moving from a protectionist neo-mercantilist approach toward a more neo-liberal strategy, further opening up the European economy to comparison and competition against the world’s leading players (Apeldoorn 2003). In addition, there is a second influence by high performers who have nothing to fear but much to gain, the influence of the Scandinavian social democratic forces. As a thorough investigation into the genesis of the European Employment Strategy (Barbier 2004) shows beyond doubt, the overall neo-liberalist thrust of the European project has been thoroughly modified by an adaptive integration of the Swedish model.
Thus, to answer the question why the Lisbon benchmarking targets are so ambitious, it may be enough to carefully listen to Larsson, one of the main architects of the EES:
“For instance, the 70% target of employment rates in Lisbon is realistic. We had made comparisons with the US. We were challenging the US. It is already achieved for countries like Denmark, Austria, etc., for countries like Greece, they have to go this way, because employment is a key” (Larsson, interview transcript by Barbier 2004: 51).
From this point of view, the European Employment Strategy, and, more generally, the Lisbon strategy, can be summarised as follows: Challenging the United States by turning the Swedish model into targets for the laggards (Barbier 2004, 50-59). However, the main assumption of this approach is that countries ‘like Greece … go this way’ fast enough.
The performance indicators or the benchmarking indicators in general gain additional meaning from their institutional, procedural and ideological contexts. The analysis of these contextual connotations will be carried out in two steps, with step one referring to the narrower context of the OMC system, and step two referring to the European project as a unique crystallisation of a basically neo-liberal world-view.
Benchmarking is only one of the several elements making up the OMC system of governance (Jacobsson 2002). The fact that it is an integral element of a system of governance adds distinctive and particular features to the benchmarking exercise (OMC embedded benchmarking).
The OMC system of governance may be placed above intergovernmentalism and below supranationalism. Mutual benchmarking of member states within such a “Third Way” system (Jacobsson 2002: 7; Jacobsson and Fifell 2003) amounts to respecting national authority and responsibility for the fields of performance benchmarked, while at the same time admitting high interdependence both of problems and of problem solution. Member states are held responsible for achieving the targets within their territory irrespective of their size (in terms of GDP, population or spatial extension) and power, and irrespective of the territorial origins of the problems to be solved.
Obviously, however, economic interpenetration among member states is highly asymmetric, resulting in a strongly skewed distribution of leverage. “Peer reviews” among sovereign peers is a matter of peer pressure, with some of the peers exercising more pressure than others. Thus, “naming and shaming” among sovereign states is quite different from naming and shaming in the classroom. National pride, the protection of sovereignty and the dignity of the state are at play, with the players bargaining both over benchmarking targets, performance measurement, and benchmarking results. Strong member states will even try to change the rules of the game, whereas the weakest ones will have to take the blame by established standards even though their factual capacity to impact ‘their’ economy may be rather negligible.
11
Cyclical benchmarking deliberation means discursively corroborating and redefining a common imagination of what causes what, implying some agreement about which problems are of common concern (Jacobsson 2002), and how far they are “home-made” or may be classified as “imported” or “historically unique” (German reunification). At the same time, however, performance indicators usually deal with the member states as independent, equally weighted units of analysis, holding them separately accountable for what they can account for only jointly and inseparably. Thus OMC benchmarking learning cycles may be thought of as ongoing redefinitions of the remits of domestic social policy, thereby asserting or denying the territorial responsibilities of a given state. OMC benchmarking starts from the political definition and recognition of areas of common concern, and by its procedures of reviews and QMV recommendations it discursively construes distinctions between a homemade portion and an imported portion of the respective problem. These distinctions translate into agreed upon limits of the factual sovereignty or factual steering capacity of the respective state. By doing all of this in the form of benchmarking and mutual learning, member states retain their dignity, while at the same time allowing for mutual pressuring or even blackmailing.
One major feature of indicator development consists of imagining common problems and of deciding about the sharing of responsibilities among member states. What holds true for the imitation of ‘best practice’ among sovereign states, may be valid for indicator development as well: Since bargaining about indicators translates into bargaining about possibly admitting domestic failure (e.g. ‘home-made’ poverty risks, or the controversy over indicators for the identification of the “working poor” between the UK and some other member states, Barbier 2004: 43), indicator development and mutual benchmarking among states is a rather sensitive exercise (for ‘best practice’ imitation cf. Hemerijck and Visser 2003: 27). Persistently having to learn from other states is equivalent to making inferiority visible, which seems rather hard to bear for representatives of states and elected politicians.
We now turn to the second step in analysing the contextual meaning of the benchmarking indicators, taking regard of a wider context.
Beyond the continual redefinition of ‘home-made’ problems, and beyond the reframing of what is the ‘domestic’ and the ‘foreign’, negotiating about the OMC benchmarking results impacts on the ongoing discursive reconstruction of the boundaries between state, society and market (Jessop 1996, Jessop 2002, Jessop 2003), influencing the rescaling of state activities (Brenner, Jessop, Jones and MacLeod 2003). The gateway to this discourse is opened up by the ‘topoi’ of territorial and functional subsidiarity, the Lisbon promotion of a ‘fully decentralised and integrated approach’, and the high importance given to the involvement of the civil society, social partners, public-private partnerships and networks (EC 2000: § 38).
Currently, considerable efforts are being made to establish databases at sub-national level, supporting benchmarking exercises at regional or even at local level (Dahan, Duell, Marin and Singer 2003). These efforts take place within the broader context of a definition of territorial areas of policy intervention, the institutionalisation of sub-national levels of government (Ireland) and even suggestions for implementing OMCs for multilevel governance within the member states (Employment Taskforce 2003). These efforts lend support to the overall neo-liberal tendency towards regional and even local differentiation of wages and wage-bargaining. The Social Dialogue is expected to involve the social partners in this ongoing reframing of industrial relations.
The shortlist of the Structural Indicators (table 1, above) is meant to signal priorities among the indicators of Lisbon’s three main domains of action. The overall objective relates to GDP growth as an indicator of living standards. Next follows “labour productivity”, measuring global competitiveness. The remaining shortlist indicators are said to represent driving forces of these two core target indicators. Being conceived of as instrumental and contributing forces, the theoretical concepts behind these indicators have to be reshaped, omitting or downplaying conflicting and inconsistent connotations, and bringing to the fore all those aspects of meaning that improve the overall compatibility of the indicators.
As an example, take the relationships among competitiveness (labour productivity), employment (employment rate, unemployment rate) and social cohesion (at-risk-of-poverty rate). Increasing labour productivity and achieving full employment at the same time is premised on certain rates of economic growth and assumptions about its job content. Absent these and additional conditions, increasing labour productivity may entail higher unemployment. To at least partially free itself from these contingencies, the Lisbon strategy redefines the employment task predominantly as labour force mobilisation rather than as the avoidance of unemployment, all of this being in line with the neo-liberal supply-side approach and the high priority of employment in the ‘Swedish model’. Employment policy in terms of the European Employment Strategy is to be taken as a mobilisation of human resources, whereas the former neo-Keynesian task of achieving full employment is reconceived as a contribution to social cohesion (table 1, indicator 10, preventing long-term unemployment, for a careful delineation of the Swedish model both against neo-liberalism and neo-Keynesianism cf. Barbier 2004, Deuxième Partie).
12
Fighting unemployment or achieving full employment in the neo-Keynesian neo-corporatist regime, besides fostering mass purchasing power, used to include highly centralised measures of labour-supply reduction (e.g. early retirement, extension of public sector, different modes of reduction of working time). By contrast, EES employment promotion precludes all these interventions at national level, while at the same time ‘activating’ both men and women, and both the younger and the older working population. Possibly, and under certain conditions, the EES (or the Lisbon process in general) itself induces additional unemployment, which subsequently has to be dealt with for reasons of cohesion.
Redefining the main task of employment policy in terms of the neo-liberal discourse is but one of several steps in improving the overall consistency of the EES and, more generally, of the Lisbon strategy. A second, equally important step deals with the re-conception of “Social”. No longer should “Social” be taken to denote collectivist social protection. Quite the contrary, social cohesion and social protection are paradigmatically different. In terms of the Lisbon strategy, systems of social protection have to be modernised under the general rationale of making work pay (cf. the indicators for the poverty trap, unemployment trap), or of making pension systems sustainable.
Thus, indicators of social cohesion are there for watching for possibly detrimental consequences of modernisation (e.g. at-risk-of-poverty rate). Rather than dealing with replacement rates or with generosity rates as indicators of social protection or social rights, indicators of social cohesion deal with possible losses of human resources (some of which may have been induced by undue speed in modernising social protection systems), while at the same time watching for pernicious consequences of social exclusion. This pervasiveness of the employment perspective, making high employment the key to the establishment of a virtuous cycle of synergies, even extends to the interpretation of the statistics on accidents and strikes (“If accidents are on the rise, it will decrease the employability, the same applies for strikes”, Larsson, interview transcript by Barbier 2004: 51).
The Lisbon strategy advocates striking a balance between (labour market) flexibility and social security. Re-conceiving ‘Social’ as social coherence is one way to reconcile these at first sight contradictory requirements. Making social security a matter of an individual’s flexibility to anticipate social risks is a second way to join apparently contradictory forces:
“While promoting flexibility on the labour market, it is also important to foster new forms of security. Security in today’s labour markets is not a matter of preserving a job for life. In a more dynamic perspective, security is about building and preserving people’s ability to remain and progress in the labour market” (Employment Taskforce 2003: 28).
Social security used to be a collectivist system of income replacement after the loss of earned income; in contrast, in terms of the Employment Taskforce, security now consists in an individual’s flexibility to anticipate and prevent the social risks. Indeed, after social security has been redefined in terms of an individual’s flexibility to prevent social risks, flexibility and security are now well balanced. While this re-conceptualisation of security, in its recognition of security as a value, is certainly beyond laissez-faire liberalism, it is nevertheless directed towards making individuals responsible for avoiding social risks, thereby corroborating the neo-liberal tendency of individualism.
In its efforts of harmonising conflicting objective by stressing synergies and compatibilities and by downplaying contradictions and dilemmas, the European project has developed a common language of roses without thorns, of consensus without conflict, “un code apparemment neutre” (Barbier 2004: 40), which helps us to understand the wide and rapid acceptance of the benchmarking exercise (cf. introduction).
13
The ‘embedded neo-liberal approach’ has succeeded in becoming the ‘dominant project’ (Jessop 2002) after a struggle among three rivalling projects, the neo-liberal, the neo-mercantilist and the social-democratic project (Apeldoorn 2003), with the social-democratic project to be subdivided into ‘Delorism’ (Apeldoorn 2003) and the ‘Swedish model’ (Barbier 2004). This struggle has not come to a standstill, and indeed, it cannot be halted. There is no ‘end of history’, nor is there any teleology, making a given worldview as inevitable as its advocates may wish it to be. For now, however, as things stand at interim to 2010, not only the EES seems to have enough “cruising speed” (Goetschy 2001), but the dominant project as a whole seems to proudly rule the waves. To return once again to the shortlist of structural indicators, given the predominance of the neo-liberal approach, competitiveness will be measured first of all by labour productivity (Table 1, indicator 2). Winds may change, however, reshuffling the historical contingencies and re-arranging the parallelogram of forces, thus permitting rivalling conceptions of competitiveness (Jessop 2003: 121) to step to the front stage.
For now however, given its embeddedness within an encompassing dominant project, the viability of the benchmarking exercise seems to be certain enough. What if the rather probable scenario comes true? What if the Lisbon adventure falls short of its targets? One of the possible outcomes is an ideological crisis of “new governance”. There has been much ado about “new governance”, as defined by management by objectives and benchmarking, by involving civil society and social partners, and by burdening MOC’s or, generally, ‘soft law’ with legitimising and participatory functions like “deliberative polyarchy” (Sabel and Zeitin 2003) and “democratic experimentalism”, (Eberlein and Kerwer 2002). Albeit not very likely, the crisis may reach even deeper, penetrating the mere ideological layers and endangering Europe’s already staggering political and economic integration. After all, if the Europe of Fifteen did not manage to initiate and co-ordinate additional efforts to reduce the competitiveness gap, why should the Europe of Twenty-Five be better equipped to succeed?
As it is, Europe will lose face. “Shaming and naming”, one of the rationales of EU benchmarking, will, ironically enough, turn against Europe’s strategy itself. Europe, having started with bold objectives and highly ambitious targets will have to shame herself strategically, thus tainting the sublime connotations of “reflexivity” and “self-referential learning”. It is but a step from the sublime to the ridiculous, and the Lisbon process is in danger of earning nothing but mock and derision: “Europe's leaders have a stated objective of making Europe the most competitive economy in the world by 2010. If you believe that can be achieved, you'll believe anything” (Ash 2004). Yes, there will be mockery, yes, there will be fatigue, yes, there will be disillusionment, and, not to be forgotten, benchmarking will show its ugly face of ‘perverse effects’ and ‘fatal remedies’, especially those in the wake of over-commitment: “Grandstanding and hype depletes resources and leads to disillusionment and abandonment” (Chalmers and Lodge 2003:16).
While it might be sensible enough to worry about the dangers of failure, talking about the dangers of success seems to be rather weird. But, at second sight, there are dangers of success as well. To become the most competitive knowledge-based geo-political region of the world is difficult enough, but to maintain this position against overwhelmingly powerful competitors will entail additional social costs. Processes of re-commodification (Esping-Andersen 1990; Schönig and Theisen 2002) will speed up, working life will be even more demanding, indeed over-demanding, for an increasing proportion of the population. One outcome seems probable enough: the latent tensions within the Lisbon triangle (making productivity grow, increasing employment, increasing social cohesion) will become manifest. Social protection will and must be the first to step down, fading out into social cohesion, with collectivist social protection systems then discursively reconstructed as lost “social costs” rather than “assets of competitiveness”.
Apeldoorn, B. van, 2003: The Struggle over European Order: Trans-national Class Agency in the Making of "Embedded Neo-Liberalism". In: Brenner et al. (eds.) 2003: State/Space. A Reader, Oxford, 147–164.
Ash, T.G., 2004: Comment: Who rules Europe?, in: The Guardian, February 19.
Auer, P., Kruppe, Th., 1996: Monitoring of Labour Market Policy in EU Member States. In: Schmid, G. O'Reilly, J., Schömann, K. (eds.) 1996: International Handbook of Labour Market Policy and Evaluation, Cheltenham UK, Brookfield US, 899-922.
Barbier, J-C., 2004: La stratégie européenne pour l’emploi : genèse, coordination communitaire et diversité nationale. Rapport pour la DARES (Ministère du travail), http://eucenter.wisc.edu/OMC.
Brenner, N., Jessop, B., Jones, M., MacLeod, G., 2003: State/Space. A Reader. Oxford.
CEC Commission of the European Communities, 1996: Benchmarking the Competitiveness of European Industry. COM (1996) 453 final.
CEC Commission of the European Communities, 2002: Delivering Lisbon. Making change happen. COM (2002) 18 final.
CEC Commission of the European Communities, 2003: Structural indicators. COM (2003) 585 final.
CEC Commission of the European Communities, 2004a: Report from the Commission to the Spring European Council: Delivering Lisbon–150Reforms for the enlarged Union. COM (2004) 29 final.
CEC Commission of the European Communities, 2004b: Strengthening the implementation of the European Employment Strategy. Proposal for a Council decision On Guidelines for the employment policies of the Member States. Recommendation for a Council recommendation On the implementation of Member States’ employment policies, COM (2004) 239 final.
CER Centre for European Reform (CER, http://www.cer.org.uk/about/index.html)
Chalmers, D., Lodge, M., 2003: The Open Method of Coordination and the European Welfare State. In: carr Ddscussion paper No. 11. June 2003, ESRC Centre for Analysis of Risk and Regulation, London.
Dahan, J., Duell, N., Marin, F., Singer, V., 2003: Renforcer la dimension locale de la stratégie européenne pour l’emploi. Étude de la faisabilité sur les indicateurs destinés aux niveaux régional et local et à l’économie sociale. Étude financée par la Commission Européenne – DG Emploi et Affaires Sociales. Groupe Bernard Brunhes Consultants Paris, ECONOMIX München.
Deppe, F., Felder, M., Tidow, S., 2000: Strukturierung des Staates. Das Beispiel europäischer Beschäftigungspolitik. In: Deppe, F., Tidow, S. (eds.) (2000): Europäische Beschäftigungspolitik, Forschungsgruppe Europäische Gemeinschaften (FEG) Studie Nr. 15, Marburg.
Eberlein, B., Kerwer, D., 2002: Theorising the New Modes of European Union Governance. In: European Integration online Papers (EIoP) 6(5). http://eiop.or.at/eiop/texte/2002–005.htm
EC European Council (2000), Presidency conclusions. Lisbon European Council 23 and 24 March 2000, http://europa.eu.int.
Employment Taskforce, 2003: Jobs, Jobs, Jobs. Creating more employment in Europe. Brussels, http://europa.eu.int/comm/employmentsocial/employmentstrategy/ tasken.htm.
Esping-Anderson, G., 1990: The Three Worlds of Welfare Capitalism. Cambridge.
Genschel, P., 2000: Der Wohlfahrststaat im Steuerwettbewerb. In: MPIfG Working Papers 00/5. http://www.mpi-fg-koeln.mpg.de/pu/workpap/wp00-5/wp00-5.html.
Gitterman, D., P., 2002: A Race to the Bottom, a Race to the Top or the March to a Minimum Floor? Economic Integration and Labor Standards in Comparative Perspective. In: Vogel, D., Kagan, R., (eds.), UCIAS Edited Volume 1, Dynamics of Regulatory Change: How Globalization Affects National Regulatory Policies, Berkely, 1–41, http://repositories.cdlib.org/uciaspubs/editedvolumes/1/10/.
Goetschy, J., 2001: The European employment strategy from Amsterdam to Stockholm: Has it reached its cruising speed? In: Industrial Relations Journal, No. 5, 401–418.
Hardes, H.-D., 1999: Zur Frage der Notwendigkeit einer strategischen Koordinierung der Beschäftigungspolitik in Europa. In: Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, No. 2, 203 –218.
Hemerijck, A., 2002: The Self-Transformation of the European Social Model(s). In: Internationale Politik und Gesellschaft / International Politics and Society , No. 4.
Hemerijck, A., Visser, J., 2003: Policy Leraning in European Welfare States. Unpublished manuscript. Universities of Leyden and Amsterdam.
Hodson, D., Maher, I., (2001): The Open Method as a New Mode of Governance. Journal of Common Market Studies, Vol. 39, No. 4, pp. 719-746.
Jacobsson, K.,2002: Soft Regulation and the Subtle Transformation of States: The Case of EU Employment Policy, http://www.govecor.org.
Jacobsson, K., Schmid, H., 2003: The European Employment Strategy at the Crossroads: Contribution to the Evaluation. In: Foden, D., Magnusson, L., (eds.), 2003, Five Years' Experience of the Luxemburg Employment Strategy. Brussels. http://www.govecor.org.
Jacobsson, K., Vifell, A., 2003: Integration by Deliberation? On the Role of Committees in the Open Method of Coordination. Prepared for the workshop on 'The Forging of Deliberative Supranationalism in the EU', Florence, 7-8 February 2003, Stockholm.
Jessop, B., 1996: Veränderte Staatlichkeit. In: Grimm, D., (Hrsg.): Staatsaufgaben. Baden-Baden, 43-74.
Jessop, B, 2002: The future of the capitalist state. Oxford.
Jessop, B., 2003: The Future of the State in an Era of Globalisation. In: IPG Internationale Politik und Gesellschaft / International Politics and Society, No. 3.
Kissling-Näf, I., Knoepfel, P., 1998: L'évaluation et le monitoring. In: Bussmann, W., Klöti, U., Knoepfel, P. (Hrsg.), 1998: Politiques publiques. Évaluation, Paris, 139-150.
Löffler, E., 1996: A Survey on Public Sector Benchmarking Concepts. In: Hill, H., Klages, H., Löffler, E., (eds.), 1996: Quality, Innovation and Measurement in the Public Sector. Frankfurt a.M.,137 –160.
Lütz, S.,2003: Governance in der politischen Ökonomie. In: MPfG Discussion Paper, No. 03/3. http://www.mpi-fg-koeln.mpg.de/pu/mpifg_dp/dp03-7.pdf
Mosley, H., 2000: Benchmarking National Labour Market Performance in the European Union: 1999 Report. Report prepared for European Commission, Directorate-General V, Employment and Social Affairs (DG V/A2) 2000.
Mosley. H., Schütz, H., Schmid, G., 2002: Effizienzmobilisierung der Arbeitsverwaltung: Leistungsvergleich und Lernen von guten Praktiken (Benchmarking). In: WZB Discussion paper FS 02-209.
Murray, A., 2004: The Lisbon Scorecard IV .The status of economic reform in the enlarging EU: http://www.cer.org.uk/pdf/wp505lisboniv.pdf
Noaksson, N., Jacobsson, K., 2003: The Production of Ideas and Expert Knowledge in OECD. The OECD Jobs Strategy in contrast with the EU employment strategy. In: SCORE Rapportserie 03:7, Stockholm.
OECD, 2003: Economic Outlook 74. Paris.
Overdevest, C., 2002: The Open Method of Coordination, NewGovernance, & Learning: Towards a Research Agenda. New Governance Project Working Paper.
Rossi, P.H., Lipsey, M. W., Freeman, H. E, 2004: Evaluation. A Systematic Approach. London, New Deli.
Sabel, C. F., Zeitlin, J., 2003: Active Welfare, Experimental Governance, Pragmatic Constitutionalism: The New Transformation of Europe
Schäfer, A.,2002: Vier Perspektiven zur Entstehung und Entwicklung der "Europäischen Beschäftigungspolitik". MPIfG Discussion Paper 02/9.
Scharpf, F.W., (1999): Regieren in Europa. Effektiv und demokratisch?, Frankfurt/Main, New York
Scharpf, F. W., 2002: The European Social Model: Coping with the Challenges of Diversity. In: MPIfG Working Paper 02/8. Köln: http://www.mpi-fg-koeln.mpg.de/pu/workpap/wp02-8/wp02-8.html
Schönig, W., Theisen, H., 2002: Arbeitsmarkt- und Sozialpolitik in Deutschland und Europa zwischen neuer Mitte und neuem Populismus. In: Aus Politik und Zeitgeschichte, No. B 46-47.
Schütz, H., Speckesser, S., Schmid, G., 1998: Benchmarking Labour Market Performance and Labour Market Policies: Theoretical Foundations and Applications. In: WZB discussion paper, No. FS I 98 –205.
Schütz, H., 2001: Zielsteuerung in europäischen Arbeitsverwaltungen. In: Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, No. 2, 207-225
Sisson, K., Arrowsmith, J., Marginson, P., 2002: All Benchmarkers Now? Benchmarking and the ‘Europeanisation’ of Industrial Relations. ESRC “One Europe or Several?”-Programme. Working Paper 41/02. Website: http://www.one-europe.ac.uk/pdf/w41marginson.pdf.
Straßheim, H., 2001: Der Ruf der Sirenen–150Zur Dynamik politischen Benchmarkings. Eine Analyse anhand der US-Sozialreformen. In: WZB discussion paper, No. FS II 01-201.
Tidow, S., 1999: Benchmarking als Leitidee. Zum Verlust des Politischen in der europäischen Perspektive. In: Blätter für deutsche und Internationale Politik, No. , 301–309.
Wessels, W., 2003: (Open) Methods of Self Coordination New Modes of Governance as a typical step in the Evolution of the EU System? Mainz (unpublished manuscript).
Zängle, M., Görtler, E, 2000: Vom Benchmarking zum Benchlearning. Zu den Eingliederungsbilanzen der Bundesanstalt für Arbeit. In: Planung und Analyse, No. 5.