Trade unionists in parliament and macroeconomic performance: evidence from Germany.

Berlemann, Michael ; Zimmermann, Klaus W.

1. Introduction

The primary motive of trade unions is to achieve the common goals of their members, such as job security, good working conditions and high salaries. Unions try to reach these goals by negotiating wages, worker benefits, working rules and complaint procedures, as well as workplace safety issues, collectively on the plant, firm, industry or national level. Obviously, these negotiations are the primary instrument unions employ to reach their goals. However, workers are not solely interested in nominal wages or working conditions; they are also concerned with inflation and real wages, net wages after taxation, or the benefits of the welfare state (Hyman and Gumbrell-McCormick 2010: 317). In many countries, unions therefore do not restrain their activities to negotiating nominal wages or working conditions, but also take part in general discussion on economic and social issues. Since the rules which apply to the labour market as a whole and the role of unions in this market are shaped at the legislative level, unionists also have an incentive to run for political office. In many countries around the globe, unionists are members of parliament or hold high political positions, such as Norbert Blum (German Minister of Labour 1982-98), Rudolf Hundstorfer (Austrian Minister of Labour since 2008), Robert Hawke (Australian Prime Minister 1983-1991), Greg Combet (Australian Minister for Climate Change and Energy Efficiency since 2007) and Bill Shorten (Australian Minister for Superannuation and Financial Services since 2010). Political lobbying might thus be considered as an additional channel of influence of trade unions.

It is an intriguing but also unresolved question whether, or under which circumstances, the behaviour of trade unions has an influence on macroeconomic performance. Basically, the answer to this question depends on the unions' wage policy, which is itself dependent on a multitude of factors, such as, for example, the degree of unionisation (Nickell, Nunziata and Ochel 2005), the centralisation of the wage negotiation process (Bruno and Sachs 1985; Calmfors and Drifill 1988), the extent of product market competitiveness (McHugh 2002), or the relative bargaining power of centralised and decentralised unions (McHugh 2002).

Various empirical studies have analysed whether union power, typically measured by the degree of unionisation (union density), has an effect on macroeconomic outcomes, thereby focusing on the direct effect through the bargaining process. As an example, Lye and McDonald (2006) find union power, measured by the degree of unionisation of workers, to have had a significant effect on unemployment in Australia in the 1970s. Similarly, Nickell, Nunziata and Ochel (2005) find unemployment in 20 OECD countries to depend on labour market institutions, among them the degree of unionisation. On the other hand, Bassanini and Duval (2009) do not find union power to influence unemployment in a sample of 20 OECD countries for the period 1982-2003. The relevance of the indirect political channel through which unionists might try to reach their goals has up to now primarily been discussed on the theoretical and anecdotal level (see, for example, Hyman and Gumbrell-McCormick 2010). In a recent and quite extensive study for Germany, Honigsberger (2008) concludes that the influence of German trade unions in parliament is quite limited. He therefore denies that unionised German members of parliament could be seen as a parliamentary arm of the unions. However, even Honigsberger's study delivers no convincing empirical evidence on this issue.

This article aims at contributing to fill this gap in the empirical literature using the example of Germany. After discussing the role of unions in general along the paradigmatic lines developed by Freeman and Medoff (1984), we summarise and discuss the study by Honigsberger (2008). Employing a newly constructed dataset for Germany, we then study whether the percentage of unionised members of parliament has an influence on macroeconomic performance. Differing from Honigsberger (2008), we find robust empirical evidence in favour of the hypothesis that the behaviour of unionised members of parliament matters for macroeconomic performance.

The article is organised as follows. In section 2, we study trade unions as institutions and discuss which actions, functions and effects are generally ascribed to them. In section 3, we summarise and discuss the study by Honigsberger (2008). Section 4 presents the results of our empirical analysis. The article ends with a summary of the main findings.

2. Trade Unions at Work

In 1973, a British folk rock band called The Strawbs became famous with the song 'Part of the union'. One of the most striking stanzas of the song is this one: 'So though I'm a working man / I can ruin the government's plan / Though I'm not too hard, the sight of my card / Makes me some kind of superman.' And every stanza is followed by the impressive chorus: 'You don't get me I'm part of the union / Till the day I die, till the day I die.' Those were the days when trade unions in England could promote the idea that former stokers on steam locomotives could travel around on diesel locomotives doing nothing, and were even paid for that as before.

Since that time, the prevailing opinion about trade unions has been quite unequivocal and might be summarised like this: Trade unions are the lobby of those with a job, and they compete with other 'special interest groups' (Olson 1965) for the attention of the political class. Indeed, it is rather different from country to country whether trade unions are perceived as 'normal' lobbies. In countries with corporatist structures, such as Germany, they traditionally have a very strong position (Hassel 2005b: 1). However, trade unions are also contracting partners. In societies with free collective bargaining, certain rituals determine the size of the wages which are valid industry-wide for a certain region and cannot be fallen short of. Since these contracts may also produce external costs, especially in the form of unemployment, trade unions (as lobbies) try to influence government and parliament to shift these external costs to third parties, primarily to the taxpayer. Union lobbying efforts can come in quite different forms, ranging from power demonstrations in public to noiseless cooperation with the government and its institutions.

'Most, if not all, unions have monopoly power, which they can use to raise prices above the competitive level' (Freeman and Medoff 1984: 6). (1) In neoclassical economics, trade unions have a bad image anyway, and they are said to pose a lot of problems (Tisdell and Hartley 2008: 313). First, it is generally assumed that they exert negative effects on the productivity and efficiency of firms and impede the mobility of labour. Second, due to their monopoly power, they can push through higher wages for their members, thereby causing a wage differential to the nonmembers. Third, their monopoly power is ensured by different practices, such as access restrictions, closed shops, strike ability and lobbying towards government, parliament and bureaucracy. From the members' perspective, trade unions are a special kind of multi-product club (Buchanan 1965; Cornes and Sandler 1999), offering various goods in exchange for a membership fee depending on income: wage negotiations on their behalf, supply of information about job markets, legal advice and representation in labour conflicts, negotiations about terms of employment in organisations, general insecurity reduction ('you don't get me, I'm part of the union'), and, of course, lobbying government and parliament.

This perspective on trade unions changed to a certain degree when Freeman and Medoff published their pathbreaking book What Do Unions Do? in 1984. The leading idea of the book was not to concentrate on trade unions as a textbook monopoly, but to treat them as social institutions which represent the 'collective voice'. Communication and, as a precondition, being able to raise one's voice are central for this approach, as Freeman (2005: 642) points out: 'For employee voice to be effective at the workplace, management must listen. For employee voice to be effective more broadly, the state and society must listen.' However, because 'voice without power [is] too often ignored' (ibid.: 643), a certain degree of power may be justified if there is an economic benefit which can be traced back to the formula of trade unions as social institutions. Even after two decades, Freeman is convinced that these advantages--lower fluctuation, better balancing of wages and fringe benefits, reduced dispersion in the income distribution, and the general political success of the trade unions in improving the conditions of the working class--more than outweigh the disadvantages of the textbook monopoly trade unions. At least with respect to the factual political power of unions in the United States, Freeman and Medoff (1984) have been quite sceptical. However, as Freeman notes (2005: 645), employees in many countries are organised in far stronger trade unions than in the United States. In Europe, for example, trade unions play a much bigger role in macroeconomic politics and overall political developments. Sometimes this may have led to the support of 'disastrous populist macroeconomic policies' (Freeman 2005: 648-649), which is also stressed by Pencavel (2005: 81-82). It is thus interesting to study how far the behaviour of unions has influenced macroeconomic outcomes.

Which power to influence the political process trade unions factually have strongly depends on the institutional systems in the relevant countries. There are various options trade unions have in almost each country: they can make propaganda to choose a certain party, they can support union-friendly candidates in the election, and they can also use their members for such political activity. However, three conditions must be met as a precondition for union power to be exercised. First, supported candidates must account for a substantial portion of the members of parliament. Second, these union-friendly representatives must--at least with respect to the important subjects--vote in a way the trade unions have in mind. Third, at least some of the unions' legislation plans must also be pushed through by majority vote.

Obviously, trade unions have no interest in uncovering their factual power and political connections. In order to avoid any suspicions that the parliament has been 'colonised' by the trade unions, they have a strong incentive to understate their influence. An obvious strategy to do so is to argue that a possible conflict between the obligation towards the union and the political mandate is always solved in favour of the latter. Interestingly enough, a recently published study by Honigsberger (2008) focuses exactly on this aspect. Before turning to an empirical analysis of the factual macroeconomic influence of unionised members of parliament, we will summarise and discuss Honigsberger's main results.

3. Unionised Members of Parliament as Parliamentary Arm of the Trade Unions?

Lobby groups are nowadays an integral part of the political process. The public is typically well aware of the fact that lobbies strive to attain rents in the form of cash transfers, tax cuts or competition-limiting regulation from the taxpayer and/ or the government. In parliamentary group states such as the Federal Republic of Germany, lobbies are more or less accepted. However, the advancement of special interests is seen as somewhat delicate when lobbying activities are undertaken by elected representatives who are members of such lobbies or even stand on their payrolls. Article 38(1) of the German Constitution demands from the representatives in the German Bundestag that they be representatives of the whole people, not bound by orders or instructions, and responsible only to their conscience. Of course, one cannot deprive elected representatives of the right of having a biography. It thus became customary, especially in the left party spectrum, to be a trade union member. But, at the same time, neither governments nor governing parties have an interest in being perceived as the 'parliamentary arm' of the trade unions. Similarly, the trade unions have no interest in being suspected of sending some kind of underground forces into parliament, eventually forming an all-party 'trade union block'.

Honigsberger (2008: 26) admits the parliamentary space to be an 'interventional field sui generis. However, he claims that there would be an essential difference between normal and trade union lobbyism. He argues that the special interests of trade unions are much closer to a society's goals than are those of other lobbyists (Honigsberger 2008: 29). Thus, although trade unions are some sort of lobby, their legitimisation would be based on mass and often majority interests. In spite of this obvious prepossession, Honigsberger (2008: 29) emphasises the different actional logics of trade unions and politics quite clearly: trade unions (as well as other lobby groups) try to direct political power in their interests to fulfil the wishes of their members and to increase their membership. However, different from political parties, they do not compete with opponents for members or votes. Furthermore, the concept of solidarity differentiates trade unions from politics in Germany. In the last decades, the median voter has clearly moved to the left. German politics is strongly cohesion-oriented and appeals to the social solidarity of the strong with the weak. In contrast, trade unions solely aim at the solidarity of employees. This should make them powerful to exploit social solidarity to their favour (Honigsberger 2008: 30).

Honigsberger's (2008) study nevertheless concludes that the influence of German trade unions in parliament is quite limited. His conclusion is based on two lines of argument. First, he argues that unionised members of parliament never had a majority in German parliaments. Second, he reports a number of interviews he conducted with unionised members of parliament, letting him conclude that 'the more interviews were conducted, the clearer it became how absurd the insinuation of a trade union block is' (Honigsberger 2008: 13). (2) We shall discuss both arguments in more detail.

Honigsberger (2008) tries to substantiate his first argument at the numerical example of the 16th legislative period (2005-09). In his numerical example, he focuses on the members of the Confederation of German Trade Unions (Deutscher Gewerkschaftsbund, DGB).

As Table 1 shows, the degree of unionisation differs remarkably between the various parties. Most unionists can be found in the left part of the political spectrum (Left Party, Social Democrats). Only 221 of the 614 members of the German parliament (36 per cent) turned out to be members of the DGB throughout the 16th electoral period. Thus, Honigsberger correctly states that there was no majority of DGB trade unionists in the Bundestag, not even in the legendary seventh electoral period from 1972 to 1976 where 49 per cent of the representatives were DGB members. According to Honigsberger (2008: 39), these numbers falsify the explicit suspicion or the hidden surmise that there is not only a quantitatively oversized but also a centrally controlled trade union block in the Bundestag. However, at least the 'oversize' argument is questionable. In the 2002 general elections, 48.3 million voters took part in the elections. The DGB members of parliament would have been correctly represented if 36 per cent of all voters were also members of the DGB. However, the DGB trade unions never had more than 11.8 million members (in 1991, an effect of German unification), and afterwards their memberships have decreased rapidly to 6.2 million (2008), which led Hassel (2005a) to speak of the erosion of the union lobby power. Thus, without any doubt, the DGB trade unions are overrepresented in the German parliament.

However, the group of organised employees in fact is considerable larger than the group of DGB members, as Honigsberger (2008: 42) admits. Adding the 10 members of parliament belonging to the Christian Trade Union Confederation (CGB), the 16 members of the Confederation of Civil Servants (DBB), the 89 members of the Christian-Democratic and Christian-Social Arbeitnehmerschaft (CDA, CSA) and the 10 members of other labour associations leads to a total of 334 representatives (54.4 per cent), (3) after correcting for double counts, and thus leads to a clear numerical majority. (4)

In his conclusion, Honigsberger presents four arguments for why a trade union block in Germany is neither existent nor relevant.

First, he argues that in spite of similarities in the process of political socialisation, the interests of unionists do not necessarily conform with those of the parties they represent. In order to avoid possible conflicts and to guarantee a functioning political leadership, the parties try to avoid giving mandates to powerful union members. This argument is in line with Honigsberger's observation that nowadays, in contrast to earlier years, trade union bosses cannot be found among the members of the German parliament. So, even if a trade union block would exist in parliament, it would be a complicated task to mobilise this group systematically (Honigsberger 2008: 64). It should be added that the visions trade unions have on broader social issues will necessarily become muddled to some extent if the favourite party of the unions leads the government. But that can be at least partly considered as a side-effect of the strategy of unions not to send their bosses to parliament, and of the complementary strategy of the parties to keep union bosses out. The mutual strategy to boil possible conflicts down obviously results in a deficit of strategic political leadership of the unions.

Second, from a sociological perspective there is the already-mentioned role conflict of unionised members of parliament. Trade unionists, in particular if they are employed full-time by a union, have a mandate of their interest group but are at the same time representatives of a party in parliament (Honigsberger 2008: 105). Honigsberger concludes from his interviews (5) that elected trade unionists solve this possible conflict always by ascribing considerably more weight to the political than to the interest group logic. (6)

Third, from a historical perspective, Honigsberger (2008: 89) argues that a trade union block in the parliament could not arise because of the often heavily differing positions on how to reach common goals. Due to the specific debating culture of the left political spectrum in Germany, a forceful combination of political and union interests could not evolve.

Finally, Honigsberger mentions systemic reasons for the alleged non-existence of a trade union block. He (2008: 94) claims that there has been a long-term process in which the power shifted from the legislative to the executive level. He argues that the German parliament has given up its role as a decision centre, at least as far as the representatives of the governing parties are concerned. As a consequence, the single representative would be much less important than in earlier times.

While agreeing in principle that a unionised member of parliament might face a loyalty conflict, the severity of this conflict seems to be somewhat overstated by Honigsberger, from our point of view. One might suspect this to be a consequence of the fresh experiences with the Agenda 2010 of Chancellor Schroder's second term in office, which caused a strong alienation process between the unions and the Social Democrats. Anyway, we have shown that the group of German employees had at least temporarily comfortable majorities in parliament--even though there never was a numerical majority of DGB members in the German parliament. Moreover, even strong minorities might have an influence on factual policies. How far unionised members of the German parliament have in fact influenced macroeconomic outcomes is essentially a question which should be answered on empirical grounds.

4. Empirical Results

Estimation Approach

In order to judge whether the share of unionised members in the German parliament has an effect on macroeconomic outcomes, we employ a simple econometric approach concentrating on those outcome variables which are in the centre of public interest: economic growth, unemployment and inflation. Instead of constructing a complex structural model for these variables, we follow the example of Eckstein and Tsiddon (2004) and employ a Vector-Auto-Regression approach (VAR). In VAR estimations, every endogenous variable is estimated on its own lags and the lags of all other variables in the model. Thus, a VAR estimation has the structure

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where X denotes a vector of endogenous variables (in our case, economic growth, unemployment and inflation), t is a time index, [epsilon] denotes the unexplained residual and n is the lag order of the endogenous variables. The optimal lag structure can be determined on the basis of the well-known information criteria (Akaike, Schwarz or Hannan-Quinn). The VAR approach has the advantage of capturing the dynamic features between economic growth, unemployment and inflation without the necessity to set up an explicit structural model of the German economy.

In order to study the influence of the degree of unionisation of the members of the German parliament, we follow the Eckstein-Tsiddon approach and add this variable as an exogenous regressor to our VAR estimation. Since German Unification likely has caused a structural break, we constructed a dummy variable U, which is defined as

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

and add it to the VAR system. Altogether, we end up with the following estimation model:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where Y denotes the exogenous variable--that is, the degree of unionisation of German members of parliament--and k is the relevant time lag with which the exogenous is affecting the endogenous variables.

Data

For our empirical analysis, we are in need of time series data on the inflation, unemployment and real growth rates. Moreover, data on the degree of unionisation of German parliamentary members are necessary. Due to data availability reasons, we make use of annual frequency data.

Since there is no official data on the degree of unionisation of German members of parliament, the data had to be extracted from various sources. In order to construct an annual time series, we employed the documentations by Hirche (1961, 1965, 1969, 1973), Honigsberger (2008), Pege (1996, 1999, 2002, 2003) and Richter (1983, 1987). To be able to calculate the degree of unionisation, a proper definition of the term 'union' is necessary. A wide definition covers all organisations which are concerned with negotiating working conditions and wages for their members. A narrower definition covers only those organisations which have the right to enforce their claims with strikes. For the case of Germany, the latter definition excludes certain public service unions, such as Gewerkschaft der Polizei (police) or Beamtenbund (civil servants). With respect to the focus of this study, the wider definition seems to be more appropriate. However, to control for the sensitivity of the presented results, we repeat all estimations for the narrower definition of unions.

The abovementioned sources allow constructing a time series of the degree of unionisation among German members of parliament for the period 1949 to 2007. Figure 1 shows the constructed time series over the whole sample period.

[FIGURE 1 OMITTED]

The time series tor unemployment rates, inflation and real GDP growth were extracted from the time series database of German Bundesbank. Originally, the unemployment rates come from the Federal Employment Agency (Bundesagentur fur Arbeit). The employed time series (7) is not seasonally adjusted and was initially only available in monthly frequency. We transformed the time series in annual frequency by calculating average values per year. The time series of inflation originates from the German Federal Statistical Office (Statistisches Bundesamt) and was available on an annual basis. (8) The same holds true for the time series of annual real GDP growth rates. (9)

Employing the Augmented Dickey-Fuller test, we first studied the stationarity properties of the time series (see Table 2). The null hypothesis of the existence of a unit root could be rejected on a 90 per cent confidence level for the inflation and growth time series when including a constant. With respect to the unemployment series, the inclusion of a time trend was necessary to be able to reject the null hypothesis of a unit root. Thus, in order to avoid problems of spurious regression, we decided to detrend the time series of unemployment rates before using it in the VAR estimations.

Estimation Results

In a first step, we detected the optimal lag-structure of the VAR model employing the well-known information criteria. As Table 3 shows, the Akaike, the Schwarz and the Hannan-Quinn criteria propose the inclusion of one lag of the endogenous variables. In light of the fact that we deal with annual data here, this results seems to be reasonable.

The results of the VAR estimations are summarised in Table 4. The dummy variable covering German Reunification is significant for all three performance indicators of the German economy. Unsurprisingly, the unemployment rate and inflation were on a lower level before Reunification, while the opposite is true for real GDP growth. However, in the focus of our interest are the results of the coefficients of the variable measuring the degree of unionisation of members of parliament. We find no significant effect for the unemployment rate. While the coefficient has a negative sign, the effect is not significantly different from zero on conventional confidence levels. However, for the remaining two performance indicators, we find significant effects. While the inflation rate is positively affected by the degree of unionisation, the opposite holds true for real GDP growth. Both effects are significant on a 99 per cent confidence level. The estimation results remain almost unchanged when using the narrower definition of trade unions outlined earlier. (10) Altogether, the empirical evidence points in the direction that unionised members of parliament are quite ineffective in organising higher employment, but at the same time contribute to higher inflation and lower economic growth.

Partisan Effects

In Germany, the degree of unionisation varies considerably from party to party. One might therefore suspect that the unionisation variable is not more than a proxy of the political orientation of the governing party. If this line of argument would hold, the presented results would be misleading and in fact would indicate the existence of some sort of partisan effects. As Hibbs (1977) and Alesina (1987) argue, the political orientation of the governing party might have an influence on macroeconomic outcomes whenever monetary policy is under the control of the government. For example, rational partisan theory predicts inflation rates to be significantly higher under left-wing governments, while unemployment rates or economic growth are the same over most of the terms of office. Only in the aftermath of unexpected election results might the real economy be temporarily affected by the government's political orientation.

In order to test the relevance of this argument, we expanded our VAR estimations by an additional exogenous variable, measuring the political orientation of the current government. We therefore constructed a partisan dummy P and added it to the estimation equation:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

The partisan dummy was coded as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] for governments headed by a Social Democrat (SPD) for governments headed by a Christian Democrat (CDU/CSU)

Whenever the political orientation of the governments changed within a year, the coding of the dummy variable referred to the party which headed the government most of the respective year.

The estimation results are summarised in Table 5. While we find no significant partisan effects for economic growth or inflation, detrended unemployment was larger under governments headed by Christian Democrats. However, the estimation results with respect to the unionisation variables remain almost unchanged. The same holds true when using the narrow definition of trade unions. (11)

Endogenous Unionisation

In our empirical analysis, we treat the degree of unionisation of members of parliament as an exogenous variable. However, the theoretical and empirical literature on trade union membership has revealed various possible factors determining the actual degree of unionisation (see, for example, Riley 1997) or Calmfors et al. 2001 for a review of the related literature). Various studies (for example, Bain and Elsheikh 1976; Bodman 1998; Carruth and Disney 1988; Carruth and Schnabel 1990; Pedersen 1978; van Ours 1992) found the share of workers being part of a trade union to depend on the business cycle. Thus, the actual rate of unemployment, economic growth and inflation possibly has influenced the degree of unionisation in Germany, too. In the light of this reasoning, one might argue that the degree of unionisation should be treated as an endogenous variable in our VAR estimation (see also Checci and Nunziata 2011). In order to take this argument into account, we estimated the VAR again, this time with the degree of unionisation as an endogenous variable. However, the degree of unionisation shows no patterns of endogeneity. It depends neither on the unemployment rate, economic growth nor inflation, but only on its own past history. Since we deal with the degree of unionisation of members of parliament here, and not with the degree of unionisation of the whole labour force, this result is not too surprising.

5. Summary

This article focuses on the role of unionised members of parliament. These parliamentarians are likely experiencing a role conflict: on the one hand, they have to fulfil their legal duties as independent parliamentarians; on the other hand, they are expected to support policies in favour of unions. Various researchers, such as Freeman and Medoff (1984), Freeman (2005) and Honigsberger (2008), came to the result that unionised parliamentarians failed to support the special interest goals of trade unions. However, none of these studies presented convincing empirical evidence in favour of this hypothesis. Based on a newly constructed dataset for Germany, we find unionised members of parliament to exert a detrimental effect on macroeconomic performance. While unionised members of parliament have failed to contribute to a higher level of employment, their behaviour has resulted in excessive inflation and suboptimal economic growth. Interestingly enough, this finding is in line with the typical conclusions of insider-outsider theory. Altogether, we might conclude that the political channel, through which unions might try to reach their goals, is not as unimportant as unionists (and Honigsberger) claim.

References

Alesina, A. (1987) 'Macroeconomic policy in a two-party system as a repeated game, Quarterly Journal of Economics, 102(3), pp. 651-678.

Bain, G. S. and Elsheikh, F. (1976) Union Growth and the Business Cycle: An Econometric Analysis, Blackwell, Oxford.

Bassanini, A. and Duval, R. (2009) 'Unemployment, institutions, and reform complementaries: Re-assessing the aggregate evidence for OECD countries, Oxford Review of Economic Policy, 25(1), pp. 40-59.

Bodman, P. M. (1998) 'Trade union amalgamations, openness and the decline in Australian trade union membership, Australian Bulletin of Labour, 24(1), pp. 18-45.

Bruno, M. and Sachs, J. (1985) Economics of World Stagflation, Blackwell, Oxford.

Buchanan, J. M. (1965) 'An economic theory of clubs, Economica, 32(125), pp. 1-14.

Calmfors, L., Booth, A., Burda, M., Checci, M., Naylor, R. and Visser, J. (2001) 'The future of collective bargaining in Europe' in T. Boeri, A. Brugiavini and L. Calmfors (eds) The Role of Unions in the Twenty-First Century, Oxford University Press, Oxford, pp. 1-155.

Calmfors, L. and Drifill, J. (1988) 'Bargaining structure, corporatism, and macroeconomic performance, Economic Policy, 6, pp. 16-61.

Carruth, A. and Disney, R. (1988) 'Where have two million trade union members gone?', Economica, 55(1), pp. 1-19.

Carruth, A. and Schnabel, C. (1990) 'Empirical modelling of trade union growth in Germany, 1956-1986: Traditional versus cointegration and error correction methods', Weltwirtschaftliches Archiv, 126(2), pp. 326-346.

Checci, D. and Nunziata, L. (2011) 'Models of unionism and unemployment, European Journal of Industrial Relations, 17(2), pp. 141-152.

Cornes, R. and Sandler, T. (1999) The Theory of Externalities, Public Goods and Club Goods, Second edition, Cambridge University Press, New York.

Eckstein, Z. and Tsiddon, D. (2004) 'Macroeconomic consequences of terror: Theory and the case of Israel, Journal of Monetary Economics, 51(5), pp. 971-1002.

Engels, F. (1887) The Condition of the Working Class in England, J. W. Lovell, New York.

Flanagan, R. J. (2005) 'Has management strangled U.S. unions?, Journal of Labor Research, 26(1), pp. 33-63.

Freeman, R. B. (2005) 'What Do Unions Do? The 2004 M-Brane Stringtwister Edition, Journal of Labor Research, 26(4), pp. 641-668.

Freeman, R. B. and Medoff, J. L. (1984) What Do Unions Do?, Basis Books, New York.

Hassel, A. (2005a) 'Die Erosion der gewerkschaftlichen Lobbymacht' in T. Leif and R. Spath (eds) Die funfte Kolonne. Lobbyismus in Deutschland, VS Verlag fur Sozialwissenschaften, Wiesbaden, pp. 188-198.

Hassel, A. (2005b) 'Gewerkschaften' in T. von Winter and U. Willems (eds) Interessenverbande in Deutschland, VS Verlag fur Sozialwissenschaften, Wiesbaden, pp. 172-196.

Hayek, F. A. (1960) The Constitution of Liberty, University of Chicago Press, Chicago.

Hibbs, D. (1977) 'Political parties and macroeconomic policy, American Political Science Review, 71(4), pp. 1467-1487.

Hirche, K. (1961) 'Gewerkschafter im Deutschen Bundestag, Gewerkschaftliche Monatshefte, 12, pp. 641-651.

Hirche, K. (1965) 'Gewerkschafter im 5. Deutschen Bundestag, Gewerkschaftliche Monatshefte, 16, pp. 705-712.

Hirche, K. (1969) 'Gewerkschafter im 6. Deutschen Bundestag, Gewerkschaftliche Monatshefte, 20, pp. 716-723.

Hirche, K. (1973) 'Gewerkschafter im 7. Deutschen Bundestag, Gewerkschaftliche Monatshefte, 24, pp. 83-90.

Hirsch, B. T. (2004) 'What do unions do for economic performance?, Journal of Labor Research, 25(3), pp. 415-455.

Hirschman, A. (1970) Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States, Harvard University Press, Cambridge.

Honigsberger, H. (2008) Der parlamentarische Arm. Gewerkschafter im Bundestag zwischen politischer Logik und Interessenvertretung, Edition Sigma, Berlin.

Hyman, R. and Gumbrell-McCormick, R. (2010) 'Trade unions, politics and parties: Is a new configuration possible?, Transfer: European Review of Labour and Research, 16(3), pp. 315-331.

Lye, J. and McDonald, I. M. (2006) 'Union power and Australia's inflation barrier, 1965:4 to 2003:3', Australian Journal of Labour Economics, 9(3), pp. 287-304.

McHugh, J. (2002) Wage Centralization, Union Bargaining, and Macroeconomic Performance, IMF Working Paper No. 02/143, IMF, Washington, DC.

Nickell, S., Nunziata, L. and Ochel, W. (2005) 'Unemployment in the OECD since the 1960s: What do we know?', Economic Journal, 115(500), pp 1-27.

Olson, M. (1965) The Logic of Collective Action, Harvard University Press, Cambridge.

Pedersen, P. J. (1978) 'Union growth and the business cycle: A note on the BainElsheikh model', British Journal of Industrial Relations, 16(3), pp. 373-377.

Pege, W. (1996) 'Wachstum am linken Rand, IW-Gewerkschaftsreport, 30, pp. 72-79.

Pege, W. (1999) 'Weiter Wachstum im linken Spektrum, IW-Gewerkschaftsreport, 33, pp. 3-9.

Pege, W. (2002) 'Jeder dritte Bundestagsabgeordnete in der Gewerkschaft', Unternehmen und Gesellschaft: Magazin fur Unternehmer und Fuhrungskrafte, 4, p. 20.

Pege, W. (2003): 'Gewerkschaften im Bundestag, Unternehmen und Gesellschaft: Magazin fur Unternehmer und Fuhrungskrafte, 3, pp. 28-29.

Pencavel, J. (2005) 'Unionism viewed internationally', Journal of Labor Research, 26, pp. 65-97.

Richter, K. (1983) 'Gewerkschafter im 10. Deutschen Bundestag, Gewerkschaftliche Monatshefte, 34, pp. 449-453.

Richter, K. (1987) 'Gewerkschafter im 11. Deutschen Bundestag, Gewerkschaftliche Monatshefte, 38, pp. 182-185.

Riley, N.-M. (1997) 'Determinants of union membership: A review, Labor, 11(2), pp. 265-301.

Tisdell, C. and Hartley, K. (2008) Microeconomic Policy: A New Perspective, Elgar, Cheltenham.

van Ours, J. C. (1992) 'Union growth in The Netherlands 1961-1989, Applied Economics, 24(9), pp. 1059-1066.

Notes

(1.) A quarter of a century before, von Hayek (1960: 267) had already supposed that 'we have now reached a state where [unions] have become uniquely privileged institutions to which the general rules of law do not apply'.

(2.) Quotations originally in German were translated into English by the authors.

(3.) Inexplicably, Honigsberger (2008: 42) counts only 324 representatives (52 per cent) for this case.

(4.) This finding also holds true for parliamentary committees (Honigsberger 2008: 68-69). Here, DGB unionists are especially overrepresented if the committee centres on questions concerning the labour force. Examples are the Committees for Labour and Social Affairs (44.4 per cent) and for Economy and Technology (41.7 per cent). When adding again the members of other labour organisations, these committees are likely to be dominated by employee interests.

(5.) One might also question the reliability of the interviews Honigsberger conducted among German unionised members of parliament. When interviewing them on their primary motivation and goals, one might hardly expect them admitting a violation of Art. 38 of the German Constitution. One might also not expect them to accuse colleagues of doing so.

(6.) Honigsberger (2008: 80) argues that for exactly this reason, German trade unions did not have the illusion that the unionised representatives of the parliamentary parties in government would fulfil the role of 'combat troops' against the Agenda 2010, an amazingly radical political program of Chancellor Schroder's government to increase incentives to take over even low-paid jobs. The program was highly controversial among the members of the governing Social Democrats and was heavily attacked by the German trade unions.

(7.) Time series code: UUCC02.

(8.) Time series code: UKFB99.

(9.) Time series code: JJ5000.

(10.) The estimation results are available from the authors on request.

(11.) The estimation results are available from the authors on request.

About the Authors

>> Michael Berlemann is a Professor at Helmut Schmidt University in Hamburg. He is also a Fellow of the CESifo Research Network and a Research Professor at the Ifo Institute of Economic Research, Munich. His primary fields of interest are macroeconomics, political economy and empirical economics. He can be contacted at [email protected].

>> Klaus W. Zimmermann is Professor Emeritus at Helmut Schmidt University in Hamburg. He is also a Research Affiliate at Hamburg Institute of World Economics (HWWI) and Guest Professor at the European University Viadrina in Frankfurt. His primary fields of interest are public choice and public finance. He can be contacted at [email protected].

Michael Berlemann * Klaus W. Zimmermann *

* Helmut Schmidt University, Hamburg

Table 1: Unionised members of German parliament

 Number of Number Percentage
 members of of DGB Percentage (2):
 parliament (1) members (2) (2):(1) [SIGMA](2)

Christian 226 10 4,4 4,5
Democrats
(CDU/CSU)

Social 222 161 72,5 72,9
Democrats
(SPD)

Free 61 1 1,6 0,5
Democrats
(FDP)

Left Party 54 35 64,8 15,8

Green Party 51 14 27,5 6,3

Total 614 221 36 100,0

Source: Honigsberger (2008:41, Table 1).

Table 2: Stationarity properties (Augmented Dickey-Fuller test)

 Critical
Endogenous Exogenous ADF- value Level of
variable variables statistic (p = 0,1) significance

Inflation Constant -3,708656 -2,595565 0,0065 ***
Real GDP growth Constant -3,594155 -2,595033 0,0089 ***
Unemployment Constant, -3,305526 -3,173943 0,0757 *
 trend

Source: Own calculations.

Note: * denotes significance at 10%; ** denotes significance at 5%;
*** denotes significance at 1%.

Table 3: Determination of the optimal lag-structure of the VAR model

 Sequential Final
 modified LR forecast Akaike
Lag statistic error criterium

0 NA 1,85e-07 -6,987275
1 189,9588 * 3,56e-09 * -10,94452 *
2 6.382109 4,40e-09 -10,74018
3 9,447201 5,00e-09 -10,62879
4 8,297578 5,82e-09 -10,50587
5 2,755558 8,00e-09 -10,23198

 Schwarz Hannan-Quinn
Lag criterium criterium

0 -6,643111 -6,856215
1 -10,25619 * -10,68240 *
2 -9,707690 -10,34700
3 -9,252136 -10,10455
4 -8,785046 -9,850568
5 -8,166993 -9,445619

Source: Own calculations.

Note: Endogenous variables: detrended unemployment rate, inflation,
real GDP growth.
Exogenous variables: degree of unionisation of parliamentary members,
Reunification dummy.

* denotes significance at 10%; ** denotes significance at 5%;
*** denotes significance at 1%.

Table 4: VAR estimation results

 Detrended
 unemployment Real GDP
 rate Inflation growth

Constant 0,140975 -0,047276 *** 0,105360 ***
 (1,11865) (0,01439) (0,03214)
 [0,12602] [-3,28421] [3,27791]

Detrended 0,909545 *** -0,002375 *** 0,001181
unemployment (0,03934) (0,00051) (0,00113)
rate (-1) [23,1195] [-4,69080] [1,04511]

Inflation (-1) 19,26968 *** 0,551428 *** -238976
 (4,97303) (0,06399) (0,14289)
 [3,87484] [8,61694] [-1,67244]

Real GDP growth (-1) -10,93391 *** 0,207048 *** 0,286882 ***
 (4,34137) (0,05587) (0,12474)
 [-2,51854] [3,70622] [2,29980]

Reunification dummy 0,471044 * 0.011263 *** -0,031858 ***
 (0,27048) (0,00348) (0,00777)
 [1,74154] [3,23602] [-4,09930]

Degree of -1,144421 0,104963 *** -0,143418 ***
unionisation of (2,10944) (0,02714) (0,06061)
members of [-0.54252] [3,86881] [-2,36619]
parliament

Adj. R-square 0,943917 0,796796 0,600496

F-statistic 179,4070 *** 42,56437 *** 16,93287 ***

Source: Own calculations.

Note: Standard errors in round brackets, t-values in square brackets.
* denotes significance at 10%; ** denotes significance at 5%;
*** denotes significance at 1%.

Table 5: VAR estimation results with partisan dummy

 Detrended
 unemployment Real GDP
 rate Inflation growth

Constant -1,096587 -0,045282 *** 0,122377 ***
 (1,31525) (0,01743) (0,03868)
 [-0,83375] [-2,59752] [3,16358]

Detrended 0,908257 *** -0,002373 *** 0,001199
unemployment (0,03859) (0,00051) (0,00113)
rate (-1) [23,5370] [-4,63885] [1,05654]

Inflation (-1) 20,80067 *** 0,548961 *** -0,260030 *
 (4,95887) (0,06573) (0,14585)
 [4,19464] [8,35218] [-1,78291]

Real GDP growth (-1) -9,013789 ** 0,203955 *** 0,260480 **
 (4,40381) (0,05837) (0,12952)
 [-2,04682] [3,49420] [2,01111]

Reunification dummy 0,698658 *** 0.010896 *** -0,034988 ***
 (0,29693) (0,00394) (0,00873)
 [2,35298] [2,76868] [-4,00648]

Partisan dummy (-1) 0,373006 * -0,000601 -0,005129
 (0,21868) (0,00290) (0,00643)
 [1,70573] [-0,20732] [-0,79745]

Degree of 0.655169 0,102063 *** -0,168163 ***
unionisation of (2,32219) (0,03078) (0,06830)
members of [0.28213] [3,31599] [-2,46218]
parliament

Adj. R-square 0,946063 0,792662 0,597442

F-statistic 155,9383 34,77027 14,10969

Source: Own calculations.

Note: Standard errors in round brackets, t-values in square brackets.
* denotes significance at 10%; ** denotes significance at 5%;
*** denotes significance at 1%.