Paolo Mengano, IV.3 Structural Reforms in Transportations: Dynamics and Sectorial Spillovers in:

Oliver Holtemöller (Ed.)

How Can We Boost Competition in the Services Sector?, page 134 - 169

1. Edition 2017, ISBN print: 978-3-8487-4676-7, ISBN online: 978-3-8452-8902-1,

Bibliographic information
Structural Reforms in Transportations: Dynamics and Sectorial Spillovers (Paolo Mengano) Paolo Mengano, Bocconi University Abstract This paper empirically investigates the short-run and long-run effects of implementing deregulation in the Transportation sector and the spillover effects in the Manufacturing sector. Exploiting the novel CompNet database and the OECD’s database on Product Market Regulations, this paper adopts a Panel SVAR methodology on a sample of 16 European countries over a period of 12 years. The transition dynamics in the sectori‐ al economy caused by a sector-specific change in regulation are very simi‐ lar to the ones generated by a modification of the national legislation: re‐ forms have recessionary effects on the economy in the short-run despite being expansionary in the long-run. The spillover effects in the Manufac‐ turing sector lead to contractions in the economy that are more moderate than the ones occurring in the Transportation sector, and take place only with a delay of one year. Introduction “Structural reforms are, in my view, best defined as policies that permanently and positively alter the supply-side of the economy. This means that they have two key effects. First, they lift the path of potential output, either by raising the inputs to production – the supply and quality of labour and the amount of capital per worker – or by ensuring that those inputs are used more efficient‐ ly, i.e. by raising total factor productivity (TFP). And second, they make economies more resilient to economic shocks by facilitating price and wage flexibility and the swift reallocation of resources within and across sectors.” [Mario Draghi, Sintra 2015] Structural reforms play a central role in governments’ tool-kit as they are complementary to monetary policies and represent an efficient approach to recover from recessionary periods. IV.3 1 IV Past Reforms in the Services Sector and their Effects 134 Indeed, policies aimed at deregulating market can increase the Gross Domestic Product by increasing investments and improving employment allocation across firms as well as by stimulating the competition on the market, thus enhancing innovation and productivity. Furthermore, they can amplify the mechanisms targeted by monetary policy by boosting the eco‐ nomic expansion and avoiding short-run costs. Therefore, after the recent financial crisis, these policies have gained much more attention. Although governments have undertaken important steps in this direction, there re‐ mains room for stronger convergence across European countries in the de‐ gree of market regulation. The main reason is that politicians are still re‐ luctant to liberalize the markets due to the temporary “costs” of structural reforms. According to the common perception, implementing reforms might temporary lead to a reduction in employment and to a decrease in the number of firms operating in the markets. Those may represent strong dis‐ incentives as politicians are usually more oriented towards the short-term. On contrary to the political world, the academic community has always recognized the importance of structural reforms and has investigated their effects extensively. Nonetheless, the focus of the analyses conducted was laid almost exclu‐ sively on the aggregate impacts on the national economy. Only few studies have instead investigated the micro effects of structural reforms. Against this background, this paper aims at further shedding light on the granular dynamics triggered by markets deregulation and thereby pro‐ vides a substantial contribution to the ongoing debate. More precisely, it addresses the following questions: What are the effects of sector-specific reforms and what are the spillover effects to other sectors? In particular, how do key variables such as employment and firms’ investments respond to deregulation shocks in the short-run and in the long-run? The micro effects of structural reforms are analysed using the Competi‐ tiveness Research Network (CompNet) firm-level-based database and the OECD’s database on Product Market Regulations in a Panel SVAR frame‐ work. Constrained by data availability, this paper investigates the effects of market deregulation in the Transportation sector (downstream sector) and the spillovers in the Manufacturing sector (upstream sector). The sam‐ ple considered covers 16 European countries for a period of 12 years. The results suggest that labour market rigidity and investments reduc‐ tion are the main drivers of the initial recessionary period faced by the re‐ formed sector. Moreover, this paper provides empirical evidences that IV.3 Structural Reforms in Transportations 135 deregulation in the downstream sector also affects the upstream sector. It finds that the latter is affected by “lagged spillover effects” that have con‐ tractionary effects on the sectorial economy. Yet, the recessionary period in the upstream market is less pronounced than in the reformed sector and it is followed by an expansionary period in the long-run. Literature Review Being the most decisive governmental instrument for intervening on the markets, structural reforms have always played a key role in the academic debate, as proved by the long series of researches published. Indeed, since the ’50s there has been a long and new strand of the literature that began dealing with the need and, most important, the effects of deregulation in both the product and the labour market. This included remarkable contri‐ butions both from an economic perspective61 and from a law perspec‐ tive62. This strand of literature analysed two main features of deregulation. First, it documented the effects of the reforms implemented during the 20th century and provided a theoretical framework to analyse their effects. The second feature was that the focus was laid mostly on the long-run ef‐ fect of reforms. With the contemporaneous success of the “theories of growth”, the main emphasis was on the drivers of the divergence during the expansion of modern economies. Policies, reforms and institutions were accounted as one of the main causes. Due to the scarceness of data, empirical studies were realized mostly on the US economy. Further remarkable analyses were conducted also on the Russian economy that, after the fall of the Soviet Union, experienced a broad economic revolution in terms of both the agents on the market and the structure of the market itself. Yet, the prominence of the empirical 2 61 For instance, interesting contributions were provided by Averch and Johnson (1962), Bailey et al. (1994), Blanchard and Wolfers (2000), Boeri, Nicoletti and Scarpetta (1999), Boyer (1987), Coe and Snower (1997), Fernandez and Rodrik (1991), Gersbach (2000 and 2003), Gilles (1993), Joskow and Rose (1989), Mc‐ Donald and Solow (1981), Neven, Roller and Zhang (1999), Nordhaus (1990), Winston (1993 and 1998). 62 For instance, significant contributions were provided by Levine (1975 and 1981), Marshaw (1979), Rutledge (1955) and Schwartz (1972). IV Past Reforms in the Services Sector and their Effects 136 works analysed the labour market institutions, which were easier to docu‐ ment in terms of data. Following the seminal contribution of Blanchard and Giavazzi (2003), it is state of the art that reforms lead to “short-term pain” and “long-run gains”. This work is still a cornerstone in the economic literature. In their work, Blanchard and Giavazzi confirm the beneficial effects of structural reforms on the economy in reducing and redistributing rents. But, most important, building an ad hoc macroeconomic model and divid‐ ing the period considered into short-run and medium-/long-run, they as‐ sess the existence and importance of distribution and transitionary effects after the implementation of reforms. Indeed, they are able to capture the short-run effects that translate either in the potential shrink of incumbent firms, if the deregulation affects the product market, or in lower unem‐ ployment, if the deregulation affects the labour market. Since this seminal contribution, a growing literature has studied the top‐ ic of structural reforms and market deregulation with specific emphasis on the short-term and long-term dynamics. It is possible to identify two dif‐ ferent streams in this literature according to the methods adopted to repre‐ sent and simulate structural reforms. The first trend relies on the assumption that the main channel in which structural reforms and deregulation affects the aggregate economy is via the composition and distribution of economic rents. Both in the product and in the labour market. This affects directly the price levels and the allo‐ cation of inputs and outputs as well as market efficiency and innovation. Along these lines, the most recognized way of measuring economic rents is through mark-ups or price-cost margins, i.e. the difference between prices and marginal costs – the distance from the perfect competition. As a consequence, in macroeconomic theories, structural reforms are modelled as an exogenous one-time cut in firms’ mark-ups. The main finding within the first strand is that the removal of structural imperfections to the functioning and to the free entrance in the markets would positively affect the economic growth. This should immediately re‐ sult in an efficient internal market. In addition, encouraging research and innovation, deregulation should lead to higher productivity and employ‐ ment levels in the national economy. Finally, structural reforms in the mar‐ kets are designed to increase competition and enhance firms’ productivity, triggering lower prices thus stimulating the internal demand and reducing the inflation. IV.3 Structural Reforms in Transportations 137 Adopting sharp decrease in mark-ups estimates as proxy for reforms, significant contributions have been provided, among others, by Coppens and Vivet (2004), Bayoumi et al. (2004), Griffith et al. (2006), Eggertsson et al. (2014), Gomes et al. (2013) and Varga et al. (2014), which imple‐ mented new theoretical and empirical approaches to investigate the effects of structural reforms. Coppens and Vivet (2004), for example, analysing the Belgian telecom‐ munications sector, which was largely deregulated in the late 90’s, find that liberalizing formerly regulated markets would result in lower prices and higher productivity levels. They identify as main driver for this phe‐ nomenon the declining importance of economies of scale in the supply chain. Although they had a crucial role in the industrialized economy of the 20th century, their role has been replaced with innovations thus the adoption of new technologies. Indeed, with the technology advancements, higher level of productivity became reachable independently from the size of the firm. Therefore, they affirm that a free market, which would guarantee tougher competition, would lead to levels of productivity unreachable with monopolistic competition or with the absence of continuous stimulus for innovation. Moreover, with lower barriers to entry, the number of firms operating in the markets would increase and this would have positive ef‐ fects in terms of job creation and employment. Along these lines, Bayou‐ mi et al. (2004) and Griffith and Harrison (2004) have very similar and in‐ teresting results. Empirically estimating firms’ mark-up to proxy econo‐ mic rents, they model structural reforms as a one-time reduction in the lat‐ ter. They find that deregulation in the market affects the national economy both in a direct and in an indirect way. The direct effects are due to the reduction in the production costs and the removal of barriers to entry, which reflect directly on the overall productivity level on the market. On the other side, indirect effects are driven by three main efficiencies: alloca‐ tive efficiency, productive efficiency and dynamic efficiency. The Alloca‐ tive efficiency effect consists in the erosion of incumbents’ market power as a result of tougher competition. This ultimately leads to lower firms’ mark-ups. In the resulting less concentrated economic environment, inputs are used more efficiently and, therefore, the allocation of goods and ser‐ vices is improved. Moreover, the least productive firms are then forced to exit the market causing an immediate increase in the average level of pro‐ ductivity. The productive efficiency effect entails that firms are incen‐ tivized to re-organize their work in a more efficient way in order to in‐ IV Past Reforms in the Services Sector and their Effects 138 crease their production and their productivity levels. As a consequence, wastes of resources and inefficiencies on the market should rapidly dimin‐ ish. The last dynamic efficiency effect guarantees that in the attempt of en‐ hancing the productivity levels, both research and innovation are boosted to move forward the modern technological frontier. Finally, they also af‐ firm that the mix of these positive effects might take time to materialize and this, in the short-run, would result in a contraction of the economy. In a more recent work, Eggertsson et al. (2014) explore the short-run link between structural reforms and the interest rate set by the European Central Bank (ECB). Given the relevance and the timing of their work, the analysis aims at providing a solid background for political action taken in order to recover from the international economic and financial crisis. To answer their research question, they build a standard dynamic stochastic general equilibrium model calibrated to match salient features of the Euro‐ pean Monetary Union economy. Surprisingly, they find that deregulating the market during a period of crisis – with the ECB setting the interest rate at the Zero Lower Bound – might even yield to a deepening of the preexisting recession. The two main drivers that they identify are the further decrease in the inflation and the increase in the real rates. Moreover, con‐ sidering the role of expectations in the model, the outcome may be even worse with people expecting the reforms to be temporary. Hence, they un‐ derline the necessity of having an accommodating monetary policy in or‐ der to avoid the negative effects of implementing structural reforms in a ZLB environment. In another work, Gomes et al. (2013) try to enlarge the target of the in‐ vestigation by performing an analysis on the effect of deregulation in product and labour markets both from a national and an international per‐ spective. Adopting a multi-country dynamic general equilibrium model for the Euro Area, they study the effects of reforms in Germany. To shed light on the international transmission mechanisms, they are able to disentangle the effects on the German domestic economy and the effects on the rest of the Euro Area. Furthermore, to complete their analysis, they present an al‐ ternative scenario simulating reforms in Portugal and, again, examining the effects in the national economy and in the rest of the Euro Area. Final‐ ly, they confirm that the tougher competition in the market resulting from structural reforms would lead to an increase in the long-run output. More‐ over, focusing on the divergence of the European economies, they assess that coordination between countries in the implementation of structural re‐ IV.3 Structural Reforms in Transportations 139 forms would homogenize their economic performances, in terms of both price competitiveness and real activity. Varga et al. (2014), always within this first strand of the literature and so designing structural reforms as a one-time reduction in firms’ markups, further investigate the intertemporal effects of structural reforms. Us‐ ing a semi-endogenous growth model they analyse deregulation in South‐ ern European countries (Italy, Spain, Portugal and Greece). They find that deregulation in the market leads to a larger demand for inputs of produc‐ tion and, thus, to a better allocation of resources on the market by theoreti‐ cally simulating reforms in the final goods sectors in their model. Al‐ though beneficial in the long-run, they confirm that deregulation, on aver‐ age, has a temporary negative impact on the national GDP, and more gen‐ erally on the national economy, in the first years in the aftermath of the reforms. On the other hand, the second stream adopts another way of analysing the deregulation of markets. This relies upon ad hoc indicators of product and labour market regulation. In the last years, as an answer to the grow‐ ing interest in studying the underlying structure of the market, several in‐ stitutions have focused on the production of indicators in order to give a trustful representation of the level of regulation. Those indicators collect records and information mainly from national laws and surveys filled out by governments, regulators and active individuals on the markets. Perhaps the most important contribution in this sense has been provided by the OECD, which created a series of indicators that is widely recognized and used in the recent economic literature. Furthermore, exploiting these novel data, the OECD has published a long series of methodological papers pre‐ senting interesting insights on the regulatory scenarios both in OECD countries and in non-OECD countries.63 In addition, relevant stylized facts and econometric-founded results emerge from their analyses. For instance, Boylaud and Nicoletti (2001) analysing specifically the Retail sectors find that a steady deregulation trend has undergone in the 80’s and 90’s in the OECD countries, mainly in the Food sector. Furthermore, they recognise a tendency to reintroduce stricter regulation for large outlets, mainly in European countries. According to their empirical findings, this should pre‐ vent efficient market dynamics, such as the exit of unproductive firms and 63 For the series of OECD’s papers please refer to: Nicoletti et al. (2000), Boylaud and Nicoletti (2001), Conway et al. (2005), Conway and Nicoletti (2006), Wölfl et al. (2009), Wölfl et al. (2010) and Koske et al. (2015). IV Past Reforms in the Services Sector and their Effects 140 the entrance of more productive ones thus preventing productivity en‐ hancement. In turn, this would result in a less competitive environment, lower employment growth and higher prices for the consumers. More recently, Wölfl et al. (2010) highlights a negative relation be‐ tween Product Market Regulation (PMR) and GDP per capita. Extending the time series of the PMR, that is computed every five years, by adopting Bayesian Model Averaging (BMA) techniques, they are able to identify the PMR trends as one of the main reason for the divergence in the econo‐ my performance of countries so similar for other characteristics. Also, they link the diverse effects of the market regulations on the economy growth to the level of economic development on the country. Having test‐ ed for the heterogeneity of the parameters, they calculate thresholds ac‐ cording to the methodology developed by Hansen (1999) to separate the level of economic development of the countries. Their empirical results suggest that a strict regulation hampers economic growth more in de‐ veloped countries than in developing countries. This is very much in lines with Chang (2011). A large number of researches adopting these indicators, both for direct empirical testing and for model calibrations, have been published in the recent past thereafter providing very interesting contributions to the eco‐ nomic knowledge of the effects of structural reforms. Among others, the main contributions might be found in Nicoletti and Scarpetta (2003), Cette et al. (2014) and the novel series of papers by Cacciatore, Duval, Fiori and Ghironi. In their preeminent work, Nicoletti and Scarpetta investigate the scope and depth of cross-country differences in terms of pro-competitive regu‐ lation and privatization policy in the OECD area. Despite a general dereg‐ ulatory trend, they find that convergence in regulatory setting is still lack‐ ing. The main reason is in the pace of reforms and the different policy ori‐ entation in the OECD countries. In addition, in a complementary analysis, they attempt to link this regulatory divergence to the cross-country hetero‐ geneity in terms of GDP growth investigating the correlation of reforms with output and economic growth. This enables them to assess that lower barriers to entry and state control lead to faster innovation and develop‐ ment of technologies. Moreover, they find strong empirical evidences sug‐ gesting that less regulated market brings productivity enhancements and efficiency gains to the national economy. On the contrary, limitations to market entry might cause scare adoption of new technologies and a reduc‐ tion in both the competitive pressure and in the technological spillovers. IV.3 Structural Reforms in Transportations 141 A novel approach has been performed by Cette et al. (2014) which, combining the OECD product market indicators with the EU KLEMS mi‐ cro-database, attempt to investigate the micro channels underlying the im‐ plementation of structural reforms. The novelty of their work relies on the analysis of the impact of deregulation directly on production prices and wages. Considering a multi-products productivity function, they show the country/sector inefficiency produced by heterogeneous levels of regu‐ lation. Exploiting the power of micro-data, they create a model which is able to catch the differences among service sectors and manufacturing sec‐ tors and among downstream industries and upstream industries. Moreover, the workers are also divided in high skilled and low skilled, as standard procedure in the literature. Introducing the price and wage channels, they show that reforms positively affect sector level and country level produc‐ tivity. Thereafter they also show that, in particular in the European coun‐ tries, the productivity enhancements deriving from a more accommodating regulatory practise might be very high. Nonetheless, to see the effects materialising, a minimum period of five years should be considered and effects would be heterogeneous between countries. Recently, Cacciatore, Duval, Fiori and Ghironi provided signifi‐ cant contributions to this literature with a novel series of papers.64 They offer very useful and interesting results to base our understanding of the effects and the transition dynamics generated by the implementation of structural reforms. To study the effects of product and labour market liberalization, Caccia‐ tore and Fiori (2016) first build a real business cycle model with endoge‐ nous product creation and labour market frictions. Within this new theo‐ retical framework, they find that the markets undergo important transition dynamics in the aftermath of deregulation. First they find that individual reforms – reforms targeting only at the product market or at the labour market – have recessionary effects on the national economy in the shortterm. Afterwards, they assess that joint deregulation in both the good and the labour market reduce significantly the temporary “costs” generated by individual reforms. Finally, they affirm that policymakers should exploit the interdependence of policies in order to increase national welfare and efficiency in implementing structural reforms, since joint deregulation 64 Please refer to Cacciatore, Fiori and Ghiorni (2013), Cacciatore, Duval, Fiori and Ghironi (2015) and Cacciatore and Fiori (2016). IV Past Reforms in the Services Sector and their Effects 142 moderates the level and the volatility of wedges and distortions in agents’ equilibrium decisions. Later, Cacciatore, Fiori and Ghiorni (2015) extend the theoretical model to a two- country monetary union framework. This is in order to include monetary policy analysis in the study and to provide a more complete pol‐ icy advise. Shifting the focus on the monetary policy, they discover that implementing the optimal monetary policy in a highly regulated market leads to different outcomes compared to a low regulated industry. They are able to show that, despite welfare gains from the Ramsey-optimal policy are sizable, the Ramsey allocation requires significant departures from price stability both in the long run and over the business cycle. Further‐ more, they assess that an expansionary policy would significantly reduce transition costs. Lastly, they prove that deregulation reduces static and dy‐ namic inefficiencies in the markets. Lastly, Cacciatore, Duval, Fiori and Ghironi (2015) transform the for‐ mer model into a New Keynesian small open economy framework. Doing so, they are able to enhance the understanding of the effects of structural reforms on trade activities and international competitiveness indicators. Their results suggest that reforms do not generate significant deflationary pressure. This way the inability of monetary policy to deliver large interest rate cuts in their aftermath either because of the zero bound on policy rates or because of membership in a monetary union may not be a relevant ob‐ stacle to reform. Therefore, concerns about the zero lower bound on inter‐ est rates (or inability to use independent monetary policy in a monetary union) should not be viewed as stumbling blocks on the way to increased market flexibility. Differently, they find that alternative simple monetary policy rules do not have a large effect on transition costs. Methodology We estimate a Panel Structural Vector Autoregressive (Panel SVAR) mod‐ el in order to capture the dynamics in the sector of Transportations and the spillover effects in the Manufacturing sector produced by structural re‐ forms. In this section, the data preparation, the model specification and the robustness check implemented are presented. 3 IV.3 Structural Reforms in Transportations 143 Data preparation All the data considered are taken in natural logarithms to homogenize and facilitate the analysis. Indeed, this transformation allows to reduce the noise deriving from the variables that are extremely heterogeneous in terms of levels (for instance the Real Value Added produced in the Ger‐ man Manufactory sector has a different scale compared to the Finnish or the Slovenian one). Preserving the relations among the variables, the loga‐ rithmic form also increases the cross-country comparability of the sample. In order to control and disregard potential trends in the series, the Ho‐ drick-Prescott filter (decomposition) has been adopted. This method com‐ prehends an algorithm that decomposes the time series into its trend plus a cyclical component. Initially proposed by Hodrick and Prescott (1997), the filter estimates the trend component, τt , and computes the cyclical part as the difference between the original series, yt , and its trend, i.e.: yt = τt + ct  (1) To calculate τt , the following equation has to be minimized, ∑t = 1 T yt − τt 2 + λ∑2 T − 1 τt + 1 − τt − τt − τt − 1 2  (2) The first term is the sum of the squares of the deviation of the series, yt  from its trend,τt . The second term is a penalization for the variation of the growth rates in the trend component. Indeed, depending on the value as‐ signed to the parameter λ , which multiplies the sum of the squares of the trend component's second differences, reflects the penalty of incorporating fluctuation into the trend. In order to have deviations from the steady state in the Impulse Responses, an infinity penalty has been assigned to the trend component. This way, with λ ∞ , the change in the trend is con‐ stant, so the resulting trend is linear. Panel VAR Following the methodology proposed by Cacciatore and Fiori (2016), we adopted a Panel SVAR model with two lags to perform the analysis. This enables to have the representation of the joint dynamics of the variables under analysis and to exploit the past behaviours of the data in order to obtain their responses to deregulation shocks. A Panel SVAR approach 3.1 3.2 IV Past Reforms in the Services Sector and their Effects 144 seemed particularly suited to address the research questions given that it captures the interdependencies among the variables of interest, does not impose unrealistic restrictions, can easily incorporate variation in coeffi‐ cients and shocks over time, and most important accounts for heteroge‐ neous dynamics across observation grouped by specific characteristics, i.e. country and sector. In formula, the Panel SVAR can be written as: B0yit = B1yit − 1 + B2yit − 2 + f i + gt + uit   (3) where yit is the n x 1  vector of all the endogenous variables with i  indicat‐ ing the country and t   the year; yit − 1 and yit − 2  are the n x 1  vectors of the lagged endogenous variables; B0 , B1  and B2  are the n x n  coefficient matri‐ ces of the present and past observations of yit ; f i  and gt represent, respec‐ tively, country fixed effects, accounting for unobserved, time-invariant cross-country heterogeneity, and year fixed effects, accounting for the presence of common shocks across countries; uit is the vector of error terms that follows a white noise process. It is possible then to rewrite equation (3) dividing each side by B0−1 : yit = Φ1yit − 1 + Φ2yit − 2 + ci + dt + ϵit   (4) where Φi = B0−1Bi with i =1,2 (5) ϵit = B0−1uit  (6) ci = B0−1 f i  (7) dt = B0−1gt  (8) in order to get the vector autoregressive representation of a dynamic struc‐ tural system. Therefore, the VAR can be considered as the reduced form of a general dynamic structural model. The coefficient estimation is performed by a simple OLS procedure. Due to the shortness of the time series, each coefficient is restricted to be the same across country. Moreover, the model is designed according to the assumption of legislative delays in the effective implementation of the structural reforms, i.e. the VAR is assumed to be recursive. Beside fulfill‐ ing the legislative delays assumption, this procedure identifies the struc‐ ture of the model by constructing an error term in each equation that is IV.3 Structural Reforms in Transportations 145 uncorrelated with the other errors. Hence, this provides a valid back‐ ground for employing a simple OLS estimation65. We adopted the Cholesky decomposition to orthogonalize the variancecovariance matrix, thus the shocks. This method consists in decomposing a positive definite matrix into the product of a lower triangular matrix and its conjugate transpose: Ω = PP′  (9) Now, considering Ω  – the variance-covariance matrix of u  – to be a real symmetric positive definite matrix, there exists a unique lower triangular matrix A , with 1s along the principal diagonal, and a unique diagonal ma‐ trix D  with positive entities along the principal diagonal such that: Ω = ADA′  (10) According to this decomposition, D  is a diagonal matrix whose element j, j   is the variance of u . Dividing it in D1/2 D1/2 , where D1/2  is the diago‐ nal matrix whose element j, j   is the standard deviation of u , we can write equation (10) as: Ω = AD1/2D1/2A′  (11) Adopting now the Cholesky decomposition, we can define P ≡ AD1/2  in order to get: Ω = PP′  (12) P  is a n x n  lower triangular matrix with the standard deviation of u . Following these methodologies, we can express the model in terms of the orthogonal shocks ηt = P−1uit  which is constant across country: B L yit = f i + gt + uit = f i + gt + Pηt  (13) Impulse responses to orthogonalized shocks can be then formulated as: ∂yit + s ∂η jt = Ψispj  (14) where, denoting pj  the j -th column of P  and Ψis  the n x n  matrix of coeffi‐ cients of the s -th error term in the MA ∞   representation of the Panel 65 See Stock and Watson (2001) and Kilian (2013) for more detail on recursive VARs. IV Past Reforms in the Services Sector and their Effects 146 VAR model66, the element in the row j  of Ψispj identifies the response of an increase in η j  in t   for the value of the j -th variable at time t + s  holding all other η− j  constant. Although the Cholesky decomposition has been criticized67, there is a solid economic explanation for its adoption in this exercise. Indeed, the lower triangular matrix allows to keep the first-ordered variable exoge‐ nous. This way, ordering the regulatory variable first, it is possible to have the regulation exogenous from the rest of the model, that, given the re‐ silience of the policy maker to act only driven by the current economic conditions, is a perfectly reasonable assumption. Furthermore, being inter‐ ested in simulating deregulation in one sector (Transportation) and identi‐ fying the spillover effects in another (Manufacturing), the lower triangular matrix allows to have direct effects from the Transportation sector to the Manufacturing one and not vice versa. Before computing the impulse responses, the bias-corrected bootstrap intervals as proposed by Kilian (1998) was implemented. Indeed, despite the panel dimension of the sample, the time series remain very short. It has been shown that VAR estimates might yield to inaccuracy due to short time series. Among the several corrections proposed, the “bootstrap-afterbootstrap” guarantees the minimum bias in the estimation. It consists in a 2-step technique where, in the first step, the bias term is approximated in first order, and, in the second step, the bootstrap is repeated including the previous estimates. The validity of this method holds for VAR models es‐ timated in levels, as deviation from a linear trend, and in first difference. Robustness In order to prove the robustness of the results obtained in the analysis, we have estimated all the results also with a first-difference Panel SVAR on the same set of variables. In formula: B0∆ yit = B1∆ yi, t − 1 + B2∆ yi, t − 2 + ∆ gt + ∆ uit  (15) 3.3 66 See Chapter 10 of Hamilton (1994) for the formal derivation of the MA(∞) repre‐ sentation of VAR models. 67 In particular, for imposing a “causal chain” rather than learning about causal rela‐ tionships from the data. See Kilian (2013) for an extensive discussion on draw‐ backs of applying the Cholesky decomposition to VAR models. IV.3 Structural Reforms in Transportations 147 Country fixed effects are not included given that ∆ f i = 0 . Although the magnitude of the responses might differ, the patters seem to be very consistent, therefore proving the validity of the results shown. Moreover, considering the shock on the regulatory variable in firstdifference provides a better approximation for a permanent shock, such as a structural reform, further improving the legitimacy of the exercise per‐ formed. All the impulse responses of the first-difference Panel SVAR are pre‐ sented in the Annex A. Data To perform the analysis, the data considered was taken from two different sources: the firm-level-based CompNet database and the OECD’s database on Product Market Regulation. The CompNet firm-level-based database68 The recent economic literature rejects the hypothesis of the representative firm in modelling the economy. Indeed, it is now well recognized that each firm is different from the others and might have very diverse reac‐ tions to specific events. Also, from a macro perspective, the analysis of aggregate figures and mechanisms might benefit from the micro founda‐ tion of theories and studies. In light of the firms’ heterogeneity, the role of micro-data is crucial in order to analyse and investigate the economy. However, micro-data are subject to two major constraints. First, they are usually confidential. Either this is due to the firms that do not want to share their balance sheets data or to the institutions collecting the data which are not allowed to disseminate this information. Second, due to a number of reasons, for instance different currency, purchasing powers or accounting standards, the micro-data often are not suitable with crosscountry comparisons. This, of course, represents a substantial limitation for any kind of analysis. 4 4.1 68 For further information on the CompNet database, please refer to Lopez-Garcia et al. (2015). IV Past Reforms in the Services Sector and their Effects 148 The Competitiveness Research Network (CompNet), established in 2012, represents the answer of the European System of Central Banks to the need of granular data. The Network investigates the topic of competi‐ tiveness at three different levels: micro, macro and inter-regional dimen‐ sion. One, and probably the key, result of CompNet is the firm-level-based database built assembling the firm-level data provided by almost 20 European countries. To solve the confidentiality issue regarding the firm-level data, CompNet adopted the so-called “distributed micro-data approach” developed by Bartelsman et al. (2004). Along these lines, every participating country69 receives a common computer program that extracts the relevant information from the population of firms in the national database and aggregates them in such a way to preserve confidentiality. This way, the common methodology harmonizes industry coverage, time span, raw variable definitions, estimation techniques and sampling proce‐ dures. The code then produces a number of indicators and variables, based on firm-level data, that are suitable to perform analysis across-industry and across- country. The final database includes two different samples: the full sample and the 20E sample. The full sample is computed creating the indicators based on all the data retrieved from the National Central Banks (or National Sta‐ tistical Institutes). Although, the data is harmonised and cleaned from out‐ liers, no further data cleaning procedure is performed. Differently, the 20E sample includes indicators computed on data on firms with at least 20 em‐ ployees. Given the different data coverage across countries and the exis‐ tence in some countries of reporting thresholds in terms of either number of employees or turnover, considering such firms allows to improve the homogeneity of the sample. Furthermore, a population re-weighting pro‐ cedure is implemented. This includes re-weighting the number of firms in the sample with the official number retrieved from Eurostat. With this pro‐ cedure, it is possible to have a more realistic representation of the firm population in every country and the data coverage results more accurate. For all the remaining limitations of the database in terms of coverage and data comparability, please refer to Lopez-Garcia et al. (2015). The data used in the analysis covers a sample of 16 European countries, namely Austria, Belgium, Croatia, Estonia, Finland, France, Germany, 69 Country’s participation might be determined by the National Central Bank (NCB), the National Statistical Institute (NSI) or both. IV.3 Structural Reforms in Transportations 149 Hungary, Italy, Lithuania, Poland, Portugal, Romania, Slovak Republic, Slovenia and Spain, for a period of 12 years (2001-2012). To study the sectorial effects of structural reforms, data for 2 1-digit sectors – Trans‐ portation and Manufacturing70 – are considered. The data used refers only to the so-called 20E sample. As explained above, this sample includes on‐ ly the firms with at least 20 employees and guarantees higher representa‐ tiveness and better comparability. The variables used are, in order of analysis: – Real Value Added: defined initially as nominal turnover – raw materi‐ als and then converted in real terms multiplying it for the ratio of a sector-specific value added deflators provided by Eurostat over a sector-specific Purchasing Power Parity indicator provided by the Groningen Growth and Development Centre. – Employment: defined as the number of employees engaged in each firm. – Number of firms: defined as the number of firms included in the database. – Labour costs: defined as the gross employee compensation. – Total Factor Productivity (TFP): estimated according to a revision of Wooldridge (2009) proposed by Galuscak and Lizal (2011). This esti‐ mation comprehends an implementation of Olley and Pakes (1996) and Levinsohn and Petrin (2003) methodologies in a Generalized Method of Moments (GMM) framework, controlling for capital measurement error. The regression used in order to obtain the firm-level TFP is: rvat = β0 + β1kit + β2ki t − 1 + β3mi t − 1 + β4ki t − 12 + β5mi t − 12 + β6ki t − 13   +β7mi t − 13 + β8ki t − 1 mi t − 1 + β9ki t − 1 mi t − 12 + β10ki t − 12 mi t − 1 + γYeart  +ωLi t − 1   (16) All variables are expressed in logarithmic form. Materials inputs are measured by mi t .  Since labour and TFP are simultaneously deter‐ mined, while capital takes time to build, labour is instrumented by its first lag. All the interaction terms between capital and materials are in‐ cluded to control for non-linearities. Moreover, a full set of year dum‐ mies is included to control for sector-specific trends. The estimation is 70 According to the NACE Rev. 2 classification, Manufacturing comprehends sectors 10-35 (without sector 19 which is not included in the CompNet database) while Transportation includes sectors 49-53. IV Past Reforms in the Services Sector and their Effects 150 performed via a GMM following Wooldridge (2009) clustering stan‐ dard errors at the firm-level. Once obtained the estimates, TFP is computed as the residual of the difference between the real value added and the fitted values for real capital, labour and a year trend (all in logs): TFPit = rvait − β0 + β1kit + γYeart + ωLi t − 1   (17) – OP gap: computed as initially proposed by Olley and Pakes (1996). It is an indicator of allocative efficiency measured by industry-level co‐ variance between productivity and size. It is retrieved by decomposing the productivity of an industry as follows: yst = ∑i ∊ Sθitωit = ω̅it + ∑i ∊ S θit − θit ωit − ω̅it ) (18) where yst  is the weighted average of firm-level productivity in industry s  at time t   with shares of industry size as weights. θit  and ωit  represent, respectively, size and productivity of firm i  at time t  . θit  and ω̅it  denote the average size and productivity of industry s  at time t  , respectively. S  is the set of firms in sector s . The decomposition splits the weighted average industry productivity in two components: the unweighted mean productivity and the covariance between productivity and size. The second term captures the allocative efficiency in industry s  at time t   since it reflects the extent to which firms with productivity higher than the average have larger market shares. In the literature, the most common specification for computing this indicator comprehends labour productivity as firm’s productivity and labour, measured as number of employees, as firm’s size. – Price-cost Margin: defined as (turnover – variable costs)/turnover. Turnover represents incoming revenue from goods and services and variable costs consist of wage bill (including other benefits) and cost of materials and services (e.g., subcontractors, electricity and fuels). In theory, this indicator should measure the difference of prices and marginal costs. Given the empirical difficulties of retrieving this infor‐ mation from balance sheets, turnover and variable costs are adopted as proxy. The price-cost margin (PCM) is a non- parametric measure of market power and can be used to represent the concertation in the mar‐ ket. A low value of the PCM indicates a very competitive market – close to perfect competition – while high levels of PCM represent a market with few firms and prices set very high from the marginal costs – closer to the idea of monopolistic competition. IV.3 Structural Reforms in Transportations 151 OECD’s database on Product Market Regulations71 It is now accepted by the economic community that competitive markets guarantee the best development for the national economy. As a conse‐ quence, pro-competition regulation in product markets and restrictions to monopolistic behaviours might help boosting the growth. Acknowledging these economic gains, governments have gradually removed strict regula‐ tions in product markets over the past decades, reducing state involvement in business sectors, making it easier for entrepreneurs to create firms and to expand them, and facilitating the entry of foreign products and firms. While in some cases regulation was totally abolished, in others it was re‐ placed by better designed legislations which might even result in competi‐ tion enhancements. To provide quantitative analysis and solid economic and econometric support the OECD developed an indicator of Product Market Regulation (PMR) in 1998 to measure national regulations and to track reforms progresses over time in a number of countries. This database has been regularly updated every 5 years, thus in 2003, in 2008 and in 2013. The main economy-wide indicator of Product Market Regulation is provided with a complementary set of indicators that measure regulation at the sector level. Since the beginning, the indicators have become crucial for the OECD’s work as they enrich the information of regulatory practices in OECD countries and ease the investigations on their link with economic perfor‐ mance. They have been integrated in the Going for Growth exercise and OECD Economic Surveys where the OECD uses them to formulate recom‐ mendations for policy reforms. The indicators are also widely used by na‐ tional governments, other international organizations, academia and inter‐ national forums such as the G20. OECD’s database on Product Market Regulation includes a wide set of information on regulatory structures and policies. To collect the data, a questionnaire with around 1400 questions is sent to the governments of the participating countries. The database covers all OECD countries and 21 non-OECD countries. For a number of selected indicators, the explicit questions are complemented with publicly available data in order to create annual time series beginning in the 70’s. As result, the indicators are based 4.2 71 For further information on the OECD’s database on Product Market Regulation, please refer to Koske et al.(2015). IV Past Reforms in the Services Sector and their Effects 152 on “objective” data about laws and regulations as opposed to “subjective” assessments by market participants in opinion surveys. Hence, they cap‐ ture the de jure policy settings. The general country-level indicator of Product Market Regulation is then complemented with a set of indicators that encompasses information by sector. Data are available for seven net‐ work sectors (telecom, electricity, gas post, air transport, rail transport, and road transport) and five services sectors (legal services, accounting ser‐ vices, engineering services, architecture services and retail distribution). The computation of these sector indicators follows a bottom-up approach as indicated in Figure 4.3.1 – only the indicator adopted in the analysis is reported as an example for the bottom-up approach. The majority of the underlying data is included in the computation of the economy-wide indi‐ cator of Product Market Regulation. The component of market structure on the electricity, gas, rail transport, post and telecom sectors, is the only information not incorporated in the economy-wide indicator since the lat‐ ter has a more political nature. The seven indicators of regulation in network sectors are aggregated in‐ to one indicator of energy, transport and communications regulation (ETCR). The four indicators of regulation in accounting, legal, engineer‐ ing and architectural services are aggregated into one indicator of regu‐ lation in professional services. Composition of indicator of regulations Source: Koske et al. (2015) Figure 4.3.1: IV.3 Structural Reforms in Transportations 153 To perform the analysis, we used the indicator of Product Market Regula‐ tion in the Transportation sector for the same countries and period of the data in the CompNet database. Results In this section, the results of the empirical exercise are presented. Every result comprehends an ad hoc set of variables in order to investigate a spe‐ cific driver of the post-reform economy. For parsimonious reasons, the maximum number of variables included in each simulation is seven (plus the controls, i.e. country and year fixed effects). This is in order to prevent a risky decrease in the degrees of freedom of the Panel SVAR model, which would decrease geometrically with every extra variable. Moreover, a so-called “marginal approach” was adopted: meaningful variables such as employment and number of firms are held in all the simulations as a measure of robustness and each simulation includes only an extra variable to further enlarge the scope of the analysis. Baseline specification The first analysis conducted aims at investigating the dynamics of inputs and output. This specification comprehends the two main inputs for the production chain, i.e. employment and number of firms (where number of firms is used as a proxy for investments), and the sectorial output, i.e. real value added. It can be seen that, since the first year after the implementation of struc‐ tural reforms in the Transportation sector, there is a recessionary effect on the whole economy of the sector. Indeed, there is an immediate contrac‐ tion in both inputs and output. Following the theoretical models developed in the novel series of papers by Cacciatore, Duval, Fiori and Ghironi, it can be assessed that these results reflect the combination of the dynamics in employment and investments. From the demand side, the reduction is associated to lower households’ consumption and producers’ investments in order to finance the entry of new firms and new products on the market. From the supply side, with new firms entering the market and the expecta‐ tions of future tougher competition, incumbents downsize investments and employment in order to consolidate their positioning in the market and to 5 5.1 IV Past Reforms in the Services Sector and their Effects 154 further enhance their productivity. In addition, since the new jobs created by the firms entering the market take time to materialize, there will be a temporarily mismatch between labour demand and labour supply resulting in lower unemployment until the workers’ reallocation towards new en‐ trants and growing firms occurs. Furthermore, Figure 4.3.2 shows that, in the aftermath of the deregula‐ tion in the Transportation sector, very interesting dynamics undergo in the Manufacturing sector. Indeed, the latter seems to be affected by “lagged spillover effects” deriving from the implementation of structural reforms in Transportation. The recessionary trend in the Manufacturing sector seems to be very similar to the one in the Transportation sector, but it ap‐ pears deferred by one year and affects the economy of the sector for a shorter period. Figure 4.3.7, in the Annex, shows the robustness check with the same simulation on first- difference variables. The dynamics of inputs and output Notes: Panel SVAR, impulse responses to deregulation shocks. All variables are in per‐ centage deviations from trend; “_T” and “_M” denote variables in the Transportation sector and in the Manufacturing sector, respectively. The solid line represents the me‐ dian responses and the two dashed lines denote the 68% confidence intervals. Figure 4.3.2: IV.3 Structural Reforms in Transportations 155 Labour costs To enhance the understanding of the effects of structural reforms, labour costs have been included in the analysis. The dynamics of labour costs Notes: Panel SVAR, impulse responses to deregulation shocks. All variables are in per‐ centage deviations from trend; “_T” and “_M” denote variables in the Transportation sector and in the Manufacturing sector, respectively. The solid line represents the me‐ dian responses and the two dashed lines denote the 68% confidence intervals. Labour costs follow the same trend as the employment. As a consequence of the mismatch in the labour market, labour costs, and so wages, experi‐ ence a sharp decrease in the aftermath of the structural reforms. Indeed, the exit of the least competitive and efficient firms from the market and the consequential gradual reallocation of the labour force, leads to a severe contraction of labour costs for the firms in the Transportation sector. Very interesting here is also the dynamic in the Manufacturing sector. Laying the focus on this set of variables, it is possible to see that the Manufactur‐ ing sector is again facing “lagged spillover effects” due to the deregulation in the downstream sector. With a delay of one year, the Manufacturing sector is distressed by a contraction in both employment and labour costs, but in a more moderate scale than in the sector where the structural re‐ 5.2 Figure 4.3.3: IV Past Reforms in the Services Sector and their Effects 156 forms have been implemented. Although, the beginnings of the contrac‐ tions in the two sectors are misaligned, the recovery and further thrive of the economy takes place with the same time horizon in both sectors. This, of course, underlines the existing integration of the two sectors within the national supply chain, and thereafter, the dependence between both. Figure 4.3.8, in the Annex, shows the robustness check with the same simulation on first- difference variables. Productivity To further understand the micro drivers of structural reforms, in this sec‐ tion the attention is laid on the effects on the levels of productivity in both the Transportation and Manufacturing sectors. With the contemporaneous contraction of the two main inputs of the production chain, employment and investments – number of firms –, also the productivity level in the Transportation sector suffers from a tempora‐ ry drop. Despite the fact that the firms forced to exit the market due to the fiercer competition are the least competitive ones, the average productivity decreases in the aftermath of the deregulation. It seems reasonable that, in an economic environment with lower investments and an inadequate labour market, the productivity growth falls. Nonetheless, this decline is less tough than the one in the two inputs. Indeed, the productivity drop is less than proportional to the decrease in employment and in the number of firms. This should be due to the fact that the firms exiting the market are the least productive so their exit should have only very moderate repercus‐ sions on the overall productivity. Therefore, the main cause for the fall in productivity might be identified in the mismatch in the labour market be‐ tween demand and supply created by the slow reallocation of the labour force towards the new firms entering the market. On the other side, in the Manufacturing sector the productivity level has a marked increase in the first year despite experiencing the same contrac‐ tion as in the Transportation sector. This immediate increase in the first year reflects the ability of the upstream sector to benefit instantaneously from the removal of the inefficiencies caused by the high regulation in the downstream sector. Nevertheless, the simultaneous deterioration in both employment and investment leads to a contraction in the average produc‐ tivity level in the Manufacturing sector. 5.3 IV.3 Structural Reforms in Transportations 157 The dynamics of productivity Notes: Panel SVAR, impulse responses to deregulation shocks. All variables are in per‐ centage deviations from trend; “_T” and “_M” denote variables in the Transportation sector and in the Manufacturing sector, respectively. The solid line represents the me‐ dian responses and the two dashed lines denote the 68% confidence intervals. From the fourth year after the implementation of the structural reforms, both sectors have an average productivity level higher than before the deregulation. This is a clear signal that lowering barriers to entry leads to a more efficient utilization of the resources available in the market and thereafter to an increase in the national productivity after a temporary re‐ cessionary period. Figure 4.3.9, in the Annex, shows the robustness check with the same simulation on first- difference variables. Resources allocation Having observed that structural reforms have negative effects in the shortterm on the two main inputs of the production chain, in this section a mea‐ sure of allocation of resources among firms is included in the analysis. This is in order to shad light on the employment of resources in a more liberalised market and to assess whether they are allocated towards the most productive firms. Figure 4.3.4: 5.4 IV Past Reforms in the Services Sector and their Effects 158 The dynamics of resources allocation Notes: Panel SVAR, impulse responses to deregulation shocks. All variables are in per‐ centage deviations from trend; “_T” and “_M” denote variables in the Transportation sector and in the Manufacturing sector, respectively. The solid line represents the me‐ dian responses and the two dashed lines denote the 68% confidence intervals. The OP gap measures the covariance between inputs and productivity sig‐ nalling whether resources and productivity have the same path on the mar‐ ket. It can be seen from the figure above that, following the deregulation, there is an initial sharp increase in this index in the Transportation sector. This is most likely due to the immediate exit of the least productive firms from the market. Indeed, a less regulated market “forces” the least produc‐ tive firms to exit because of the tougher competition generated by the re‐ moval of barriers to entry. Thereafter, their exit makes more resources available for the firms still active on the market. Despite the reallocation of the labour force being a gradual process, the fact that scarce resources are no longer allocated to non-productive firms leads to an increase in the indicator of resource allocation on the market. Nevertheless, with new firms entering the market and the employment increasing again, the allo‐ cation deteriorates in the fourth year. This is a clear signal that there are more firms operating in the market and investing in inputs so the resources are no longer available only for the most productive firms. The transition dynamics in the Manufacturing sector generated by the spillovers of the deregulation in the Transportation sector lead to a com‐ Figure 4.3.5: IV.3 Structural Reforms in Transportations 159 plete opposite outcome. Figure 4.3.5 shows that in the first years in the aftermath of the reforms there is a large decrease of the OP gap and that it turns out to be positive only starting from the fourth year. Differently from the Transportation sector, in the Manufacturing sector the mechanisms ac‐ cording to which the resources are at disposal of the firms that can use them in the most efficient way weaken, hence the indicator of resources allocation on the market deteriorates. When the economy recovers from the initial contractionary period and benefits from the removal of ineffi‐ ciencies in the access to the market, also the allocation of resources results improved compared to the pre- reform level. Yet, these transitionary effects are not in line with the economic intuitions and might represent the starting point for further researches. Figure 4.3.10, in the Annex, shows the robustness check with the same simulation on first-difference variables. Competition and concentration In this last section, the analysis focuses on the impact of structural reforms on the competition and concentration in the markets. To disentangle these dynamics, the non-parametric measure of price-cost margin has been in‐ cluded in the SVAR model. Given the difficulties of having an accurate empirical measure of market competition, an index determining the differ‐ ence of the price from the firms’ marginal cost is a solid proxy. This indi‐ cator can be interpreted as a measure of the distance of the economy from perfect competition or from monopolistic competition. From Figure 4.3.6, it can be seen that also the effect on concentration is dynamic over the time. Anyway, differently from the dynamics in inputs and output, in labour costs or in productivity presented in the previous sec‐ tions, concentration experienced a three-fold path in the aftermath of re‐ forms. In the first year, the initial reaction to deregulation is very positive for the concentration and leads to a noticeable decrease in firms’ mark-up. This is mainly due to the expectation of fiercer competition in the postreform markets that brings firms to reduce their margins in order to pre‐ serve their competitiveness. The outcome will be the same also in the long-run. Indeed, after the temporary contractions in the economy, the re‐ sulting markets will be characterized by lower firms’ mark-ups compared to the pre-reform environment. 5.5 IV Past Reforms in the Services Sector and their Effects 160 However, it seems that firms are willing to fully exploit the short-term inefficiencies of the market before lowering permanently their margins. Therefore, for a few years after the implementation of structural reforms, firms would set their mark-up very high. Doing so, they take advantage of the temporary contraction in the production of that specific sector. Indi‐ rectly they are exploiting the transitory inadequateness of the labour mar‐ ket and the slow dynamics of the firms entering and exiting the market setting high mark-ups as long as the beneficial effects of the reforms oc‐ cur. On the other side, mark-ups in the Manufacturing sector are less affect‐ ed by the deregulation in the Transportation sector. It is evident that when the recessionary period finishes also the Manufacturing sector will be characterized by lower firms’ margins, but there are no statistically signifi‐ cant temporary effects in the short-run. Figure 4.3.11, in the Annex, shows the robustness check with the same simulation on first-difference variables. The dynamics of competition Notes: Panel SVAR, impulse responses to deregulation shocks. All variables are in per‐ centage deviations from trend; “_T” and “_M” denote variables in the Transportation sector and in the Manufacturing sector, respectively. The solid line represents the me‐ dian responses and the two dashed lines denote the 68% confidence intervals. Figure 4.3.6: IV.3 Structural Reforms in Transportations 161 Conclusions The role of structural reforms in liberalizing product markets has recently gained more attention. The financial crisis has highlighted the explicit need for reforms in order to recover from the recessionary period and to boost economic activities. The lack of more traditional supply-side policy instruments suggests that structural reforms in product market are essential for policymakers to raise competitiveness and generate sustained econo‐ mic growth. Exploiting the novel CompNet firm-level-based database and the OECD’s database on Product Market Regulations, this paper analyses the effect of sector-specific structural reforms in a Panel SVAR framework. Indeed, the use of these two database allows to link the granular aspects of the national economy to the sectorial product market regulation in a wide range of European countries. Adopting the methodology proposed by Cac‐ ciatore and Fiori (2016), this analysis attempts to identify the transitionary dynamics triggered by deregulation in the Transportation sector in the short-run and in the long-run. In addition, the focus is also laid on analysing the spillover effects in the Manufacturing sector. The empirical results confirm the conclusion of both Blanchard and Gi‐ avazzi (2003) and Cacciatore and Fiori (2016): structural reforms in the product market lead to a recessionary period despite being expansionary in the long-run. The contribution of this analysis is the empirical proof that this result holds also in a more granular dimension, and, more specifically, in the case of a sector-specific reforms. Indeed, deregulation in the Transporta‐ tion sector brings to misalignments in the labour market and to contrac‐ tions in investments. Moreover, the combination of these two effects gen‐ erates a reduction in the sectorial output, thus in the real value added. In fact, in the long-run, the re-alignment of labour demand and supply, the growing investments and the entrance of new firms brings to a large in‐ crease in the value added that, exceeding the pre-reform levels, leads the economy to thrive. In addition, empirical results show that structural re‐ forms by stimulating the competition in the market enhances the produc‐ tivity growth and the resources allocation among firms. The second main finding is that structural reforms in the Transportation sector generate spillover effects in the Manufacturing sector. Hence, the upstream sector is affected by transition dynamics very similar to the ones in the downstream sector, even without any change in the relative legisla‐ 6 IV Past Reforms in the Services Sector and their Effects 162 tion. Nonetheless, these spillover effects occur only with a delay of one year and they are more moderate than in the reformed sectors. By re-assessing the linkages of each sector of the economy, this analy‐ sis finds that the long-run beneficial effects of the sector-specific reforms interest the whole economy. Furthermore, it proves that these benefits ma‐ terialize approximatively only in the fourth year after the deregulation of the market. Taking into consideration the duration of the mandate of the policymakers in each country in the sample, as reported in Annex B, it seems obvious that they do not have incentives in implementing reforms that would generate beneficial effects on average only after the expiration of the mandate. Further research should be conducted in order to identify how to design or to complement such reforms in order to minimize the short-term costs. Annex Annex A The dynamics of inputs and output Notes: First-difference Panel SVAR, impulse responses to deregulation shocks. All variables are in first-difference and in percentage deviations from trend; “_T” and “_M” denote variables in the Transportation sector and in the Manufacturing sector, respectively. The solid line represents the median responses and the two dashed lines denote the 68% confidence intervals. Figure 4.3.7: IV.3 Structural Reforms in Transportations 163 The dynamics of labour costs Notes: First-difference Panel SVAR, impulse responses to deregulation shocks. All variables are in first-difference and in percentage deviations from trend; “_T” and “_M” denote variables in the Transportation sector and in the Manufacturing sector, respectively. The solid line represents the median responses and the two dashed lines denote the 68% confidence intervals. The dynamics of productivity Notes: First-difference Panel SVAR, impulse responses to deregulation shocks. All variables are in first-difference and in percentage deviations from trend; “_T” and “_M” denote variables in the Transportation sector and in the Manufacturing sector, respectively. The solid line represents the median responses and the two dashed lines denote the 68% confidence intervals. Figure 4.3.8: Figure 4.3.9: IV Past Reforms in the Services Sector and their Effects 164 The dynamics of resources allocation Notes: First-difference Panel SVAR, impulse responses to deregulation shocks. All variables are in first-difference and in percentage deviations from trend; “_T” and “_M” denote variables in the Transportation sector and in the Manufacturing sector, respectively. The solid line represents the median responses and the two dashed lines denote the 68% confidence intervals. The dynamics of concentration Notes: First-difference Panel SVAR, impulse responses to deregulation shocks. All variables are in first-difference and in percentage deviations from trend; “_T” and “_M” denote variables in the Transportation sector and in the Manufacturing sector, respectively. The solid line represents the median responses and the two dashed lines denote the 68% confidence intervals. Figure 4.3.10: Figure 4.3.11: IV.3 Structural Reforms in Transportations 165 Annex B Duration of policymakers’ mandates References Averch, H.; Johnson, L.L. (1962): Behavior of the Firm Under Regulatory Constraint. The American Economic Review, 52(5), 1052–1069 Bailey, E.E.; Joskow, P.L.; Niskanen, W.; Noll, R. (1994): Economic Regulation. American Economic Policy in the 1980s, Chapter 6, 367-452 Bartelsman E.; Haltiwanger, J.; Scarpetta, S. (2004): Microeconomic evidence of cre‐ ative destruction in industrial and developing countries. The World Bank, Policy Research Working Paper No. 3464 Figure 4.3.12: IV Past Reforms in the Services Sector and their Effects 166 Bayoumi, T.; Laxton, D.; Pesenti, P. (2004): Benefits and spillovers of greater competi‐ tion in Europe: a macroeconomic assessment. ECB Working Paper Series 0341 Blanchard, O.; Wolfers, J. (2000): The Role of Shocks and Institutions in the Rise of European Unemployment: The Aggregate Evidence. Economic Journal, Vol. 110, 1-33 Blanchard, O.; Giavazzi, F. (2003): Macroeconomic Effects of Regulation and Deregu‐ lation in Goods and Labor Markets. The Quarterly Journal of Economics. Vol. 118, issue 3, 879-907 Boeri, T.; Nicoletti, G.; Scarpetta, S. (1999): Regulation and Labour Market Perfor‐ mance. IGIER Working Paper No. 158 Boyer, K.D. (1987): The Costs of Price Regulation: Lessons from Railroad Deregula‐ tion. The RAND Journal of Economics, 18(3), 408–416 Boylaud, O.; Nicoletti, G. (2001): Regulatory reform in retail distribution. OECD Eco‐ nomic Studies, OECD Publishing, vol. 2001(1), 253-274 Cacciatore, M.; Fiori, G. (2016): The macroeconomic effects of goods and labor mar‐ kets deregulation. Review of Economic Dynamics. Vol. 20, 1–24 Cacciatore, M.; Fiori, G.; Ghironi, F. (2013): Market deregulation and optimal mone‐ tary policy in a monetary union. NBER Working Papers 19025 Cacciatore, M.; Duval, R.; Fiori, G.; Ghironi, F. (2015): Short-Term Pain for Long- Term Gain: Market Deregulation and Monetary Policy in Small Open Economies. NBER Working Papers 21784 Cette, G.; Lopez, J.; Mairesse, J. (2014): Product and Labor Market Regulations, Pro‐ duction Prices, Wages and Productivity. NBER Working Papers 20563 Chang, H. (2011): Institutions and economic development: theory, policy and history. Journal of Institutional Economics, vol. 7: 4, 473–498 Coe, D.; Snower, D. (1997): Policy Complementarities: The Case for Fundamental Labour Market Reform. IMF Staff Papers, vol. 44, issue 1, 1- 35 Conway, P.; Janod, V.; Nicoletti, G. (2005): Product Market Regulation in OECD Countries: 1998 to 2003. OECD Economics Department Working Papers 419, OECD Publishing Conway, P.; Nicoletti, G. (2006): Product Market Regulation in the Non-Manufactur‐ ing Sectors of OECD Countries: Measurement and Highlights. OECD Economics Department Working Papers 530, OECD Publishing Coppens, F.; Vivet, D. (2004): Liberalisation of Network Industries: Is Electricity an Exception to the Rule? NBB Working Paper No. 59 Eggertsson, G.; Ferrero, A.; Raffo, A. (2014): Can structural reforms help Europe? Journal of Monetary Economics, vol. 61(C), 2-22 Fernandez, R.; Rodrik, D. (1991): Resistance to Reform: Status Quo Bias in the Pres‐ ence of Individual-Specific Uncertainty. American Economic Review. American Economic Association, vol. 81(5), December, 1146-55 Galuscak, K.; Lizal, L. (2011): The Impact of Capital Measurement Error Correction on Firm-Level Production Function Estimation. Working Papers 2011/09, Czech National Bank, Research Department IV.3 Structural Reforms in Transportations 167 Gersbach, H. (2000): Promoting Product Market Competition to Reduce Unemploy‐ ment in Europe: An Alternative Approach. Wiley Blackwell, vol. 53(2), 117-33 Gersbach, H. (2003): Structural Reforms and the Macroeconomy: The Role of General Equilibrium Effects. CEPR Discussion Papers, C.E.P.R. Discussion Papers, 4043 Gilles, S. (1993): On the Political Economy of Labour Market Flexibility. NBER Macroeconomics Annual 1993, vol. 8, 151-196 Gomes, S.; Jacquinot, P.; Mohr, M.; Pisani, M. (2013) : Structural Reforms and Macroeconomic Performance in the Euro Area Countries: A Model-Based Assess‐ ment. International Finance. Wiley Blackwell, vol. 16(1), 02, 23-44 Griffith, R.; Harrison, R.; Simpson, H. (2006): The link between product market re‐ form, innovation and EU macroeconomic performance. European Economy – Eco‐ nomic Papers 243. Directorate General Economic and Financial Affairs (DG ECFIN), European Commission Hamilton, J.D. (1994): Time Series Analysis. Princeton University Press.Princeton, N.J. Hansen, B.E. (1999): Threshold effects in non-dynamic panels: Estimation, testing, and inference. Journal of Econometrics, Elsevier, vol. 93(2), 345- 368 Hodrick, R.J.; Prescott, E.C. (1997): Postwar U.S. Business Cycles: An Empirical In‐ vestigation. Journal of Money, Credit and Banking, vol. 29, no. 1, 1-16 Joskow, P.L.; Rose, N.L. (1987): The Effects of Economic Regulation. The Handbook of Industrial Organization Kilian, L. (1998): Small-sample confidence intervals for impulse response functions. Review of Economics and Statistics, vol. 80 (2), 218–230 Elgar, E. (2013): Structural Vector Autoregressions. Chapters, in: Handbook of Re‐ search Methods and Applications in Empirical Macroeconomics, chapter 22, 515-554 Koske I.; Wanner, I.; Bitetti, R.; Barbiero, O.(2015): The 2013 update of the OECD's database on product market regulation: Policy insights for OECD and non-OECD countries. OECD Economics Department Working Papers 1200, OECD Publishing Levine, M.E. (1975): Regulating Airmail Transportation. The Journal of Law & Eco‐ nomics, 18(2), 317–359 Levine, M.E. (1981): Revisionism Revised? Airline Deregulation and the Public Inter‐ est. Law and Contemporary Problems, Vol 44, 179-195 Levinsohn, J.; Petrin, A. (2003): Estimating Production Functions Using Inputs to Control for Unobservables. Review of Economic Studies. Vol. 70 (2), 317-342 Lopez-Garcia, P.; di Mauro, F.; the CompNet Task Force (2015): Assessing European competitiveness: the new CompNet micro-based database. ECB Working Paper No. 1764. European Central Bank McDonald, I.M.; Solow, R.M. (1981): Wage Bargaining and Employment. The Ameri‐ can Economic Review, 71(5), 896–908 Neven, D. J.; Röller, L.; Zhang, Z. (1998): Union Power and Product Market Competi‐ tion: Evidence from the Airline Industry. CEPR Discussion Papers, 1912, C.E.P.R. Discussion Papers IV Past Reforms in the Services Sector and their Effects 168 Nicoletti, G. and Scarpetta, S. (2003): Regulation, productivity and growth: OECD ev‐ idence. Economic Policy, CEPR;CES;MSH, vol. 18(36), 04.[41], 9-72 Nicoletti, G.; Scarpetta, S.; Boylaud, O. (2000): Summary Indicators of Product Mar‐ ket Regulation with an Extension to Employment Protection Legislation. OECD Economics Department Working Papers 226, OECD Publishing Nordhaus, W.D. (1990): Soviet Economic Reform: The Longest Road. Brookings Pa‐ pers on Economic Activity, Economic Studies Program, The Brookings Institution, vol. 21(1), 287-318 Olley, G.S.; Pakes, A. (1996): The dynamics of productivity in the telecommunications equipment industry. Econometrica, vol. 64(6), 1263- 1297 Rutledge, I.C. (1955): Regulating Business by Independent Commission, by Marver H. Bernstein. Indiana Law Journal, vol. 31, iss. 1, article 10 Schwartz, L.B. (1972): The Economics of Regulation: Principles and Institutions. Uni‐ versity of Pennsylvania Law Review, vol. 120, no. 3, 584-596 Stock, J.H.; Watson, M.W. (2001): Vector Autoregressions. Journal of Economic Per‐ spectives. American Economic Association, vol. 15(4), 101- 115 Varga, J.; Roeger, W.; in ‘t Veld, J. (2014): Growth effects of structural reforms in Southern Europe: the case of Greece, Italy, Spain and Portugal. Empirica. Springer, vol. 41(2), 323-363 Winston, C. (1993): Economic Deregulation: Days of Reckoning for Microeconomists. Journal of Economic Literature (Impact Factor: 9.24). 02/1993; vol. 31(3), 1263-89 Winston, C. (1998): U.S. Industry Adjustment to Economic Deregulation. Journal of Economic Perspectives, vol. 12(3), Summer, 92 Wölfl, A.; Wanner; I., Röhn, O.; Nicoletti, G. (2010): Product Market Regulation: Ex‐ tending the Analysis Beyond OECD Countries. OECD Economics Department Working Papers 799, OECD Publishing Wölfl, A.; Wanner, I.; Kozluk, T.; Nicoletti, G. (2009): Ten Years of Product Market Reform in OECD Countries: Insights from a Revised PMR Indicator. OECD Eco‐ nomics Department Working Papers 695, OECD Publishing Wooldridge, J.M. (2009): On estimating firm-level production functions using proxy variables to control for unobservables. Economics Letters, vol. 104(3), 112-114 IV.3 Structural Reforms in Transportations 169

Chapter Preview



‘How Can We Boost Competition in the Services Sector?’ is a key question in the process of creating a more effi-cient economic environment in Germany. This book contains a collection of conference contributions on service sector reforms from members of academic institutions, ministries, the EU Commission and other organisations. The conference consisted of a keynote on the importance and implementation of structural reforms in Europe and two panels that dealt with the evaluation of past reforms in the services sector and the potential scope and effects of further reforms.

Since the 1990s, productivity growth in Germany and other Member States of the European Union has been significantly lower than in the US. The development of productivity growth in the services sector is estimated to account for two thirds of this widening gap. The European Commission advocated reforms in the services sector in its country-specific recommendations for Germany. At a conference in Berlin in July 2016, experts from various fields presented and discussed studies on service sector reforms.

With contributions by

Oliver Holtemöller, Brigitte Zypries, Joaquim Nunes de Almeida, Dirk Palige, Henrik Enderlein, Stefan Profit, Davud Rostam-Afschar, Paolo Mengano, Oliver Arentz, Erik Canton, Jochen Andritzky


„Wie können wir den Wettbewerb im Dienstleistungssektor stärken?“ Dies ist eine Schlüsselfrage für eine größere Leistungsfähigkeit des ökonomischen Umfelds in Deutschland. Dieses Buch versammelt Konferenzbeiträge von Mitgliedern wissenschaftlicher Einrichtungen, von Ministerien, der EU-Kommission und anderen Organisationen zu Reformen im Dienstleistungssektor. Die Konferenz umfasste einen Eröffnungsvortrag zur Bedeutung und Durchführung von Strukturreformen in Europa und zwei Gesprächsforen zur Bewertung vergangener Reformen im Dienstleistungssektor und zur möglichen Reichweite sowie zu den möglichen Auswirkungen weiterer Reformen.

Die Zunahme der Produktivität ist seit den 1990er Jahren sowohl in Deutschland als auch in anderen Ländern der Europäischen Union deutlich geringer als in den USA. Es wird geschätzt, dass die Entwicklung des Produktivitätszuwachses im Dienstleistungssektor für zwei Drittel dieses zunehmenden Abstandes verantwortlich ist. Die Europäische Kommission spricht sich in ihren länderspezifischen Empfehlungen zu Deutschland für Reformen in diesem Sektor aus. Auf einer Konferenz im Juli 2016 in Berlin stellten Experten aus unterschiedlichen Bereichen Studien zu solchen Reformen vor und diskutierten deren Ergebnisse.

Mit Beiträgen von

Oliver Holtemöller, Brigitte Zypries, Joaquim Nunes de Almeida, Dirk Palige, Henrik Enderlein, Stefan Profit, Davud Rostam-Afschar, Paolo Mengano, Oliver Arentz, Erik Canton, Jochen Andritzky