Environment and economic activity have been commonly thought as notions withcomplicated trade-off between each other. Their interaction over decades was a subjectof the heated debates, thus no surprise this reflected in a various theoreticaland empirical studies. Especially they have escalated after the appearance of socalled “Porter Hypothesis” (PH), by Porter(1991) and Porter and van der Linde(1995), suggesting that stringent environmental regulation induce innovationswhich in turn increase the overall competitiveness at the industry levelthrough superior productivity. Authors advocate that accurately constructedenvironmental policies can bring benefits through process balancing –substituting input resources, reducing production disruptions, usage of lesscostly materials or better utilization of them, and product balancing – throughimprovement in its performance or average quality, reducing costs by eliminatingexpensive materials and shrinking in disposal costs. The central message isthat strict environmental policies can enhance productivity by triggeringinnovations.Figure 1.
Schematicrepresentation of the Porter HypothesisSource: Ambec et al. (2013) after Porter (1991) As it confronted heavily the traditional view that environmental regulationsare those instruments that negatively affect productivity (Jaffe et al. (1995), Gray (1987), Barbera and McConnell (1990)) itis also criticized in sense of undermining competitive abilities of the firmsto seek their profit-maximizing behaviour, thus to act rationally (Palmer et al. (1995)).
The main argumentagainst is summarized in the economist’s well-known saying: “There are no freelunches” (Sinclair-Desgagné 1991,2). Means that, innovation itself are financially demanding, and whilecomputing opportunity costs with including stronger environmental regulationson firms, it obviously leads to rise in production costs, which prevail theirrevenue (eg. Sinclair-Desgagné (1991), Ambec et al. (2013)). To theseconclusions considerably contributed US industrial slowdown in 1970s.
Forinstance, Jaffe et al. (1995) suggestedthat it is large abatement costs responsible for US firms productivity downturn, decline in competition and also incharge for pushing companies to reallocate their manufacturing processes toanother countries. On the other hand, as stated by Ko?luk and Zipperer (2014), Porterand van der Linde (1995) these older studies have considerableidentification issues, as are concentrated mainly on domestic effects andundervalued environmentally friendly innovations, neglecting industry andcountry-specific effects.This literature review concentrates mainly on more recent empiricalworks which are grouped according to the regions or industries covered, as resultsare very controversial. For instance, the assumption that environmentalstringency might increase productivity received a lot of critics fromneoclassical viewers as it is difficult to assimilate all the relevant factors ofits arguments in the theoretical models (Broberget al.
, 2013). According to Smithand Walsh (2000) there is a problem of methods used to identify the forcesto productivity adjustment, thus in practice it is complicated to reject PorterHypothesis. Suggesting that there are “no painless environmental policies” Smith and Walsh (2000,74) claim thatproductivity is rather harmed by costly environmental regulations. Prevailing empirical studies to some extentare concentrated on one of the three variants of Porter Hypothesis definedby Jaffe and Palmer, 1997: the “narrow”, “weak” or “strong”. The”narrow” version suggests, that under specific types of environmental policiesinnovation and productivity benefit.
“Weak” says that regulations will triggera certain type of innovation as firms likely to invest in other activitiescompared to those they could have done without new policy constraints. While”strong” states that with a “kick” of a new tighter regulation, companies broadenand optimize their decisions and processes what leads to increase in productivity.The authors own empirical results go along with a “weak” form, also statingthat coefficients greatly differ across industries.While analysing most studies in the field, Ko?luk and Zipperer (2014), found that consequences vary a lot,however, with traditional measures at the plant-level, among regulated andnon-regulated plants, outcome is negative, but not very robust. The effectdepends significantly on the particular plant characteristics.
The results fromthe industry level show that early studies commonly view environmentalstringency policies as a negative burden for productivity, while more recentones have evidence of no or positive linkage between them. As a cross-country analysis suffers from the lack of reliable andcomparable method of estimation, Bottaand Ko?luk (2014) constructed the proxy – environmental policy stringencyindex (EPS), that includes aggregated and scored instruments, related toclimate and air pollution and allows for quantitative and qualitative measureof environmental regulation tightness over a long-time period. They assumestringency as the “cost” on an activity, which is damaging the environment.
Itrises the opportunity costs of polluting, therefore, providing an incentive forenvironment-friendly activity. The index ranges from 0 (not stringent) to 6(the highest degree of stringency) and covers most OECD countries from 1990 to2012. The indicators are divided into market-based (e.
g. taxes on pollutants orother environmentally harmful activities) and non-market-based instruments(e.g. subsidies for environmental favourable activities). According to theresults, indicators showed positive and significant correlation with GDP aswell as with Environmental Performance Index (EPI) and negatively associatedwith emissions per unit of GDP in nominal and PPPvalues.
Besides, market-based instruments are significantly correlated withGreen Patent Index, what suggests that this type of instruments positivelyaffect the “green innovations”. With a help of the same EPS index, Albrizioet al. (2017) empirically tested the “strong” version of Porter Hypothesis –if the stringent environmental policy induces productivity, using a panel ofOECD countries on multi-factor productivity (MFP) growth. For the estimation, authorscombined the industry and firm-level impacts and policy instruments, accordingto their price mechanisms – market-based and non-market ones. The panel includes11 OECD countries and 22 manufacturing sectors over the time period 2000-2009. Totrack the effect, a three-year moving average has been selected for lagging EPSin time for both levels – industry and firm.
This moving average is also used asan interaction term with the distance to the global frontier. At the industrylevel, findings report that tightening in environmental policy have a positiveshort-term effect on productivity growth in countries, where industries aretechnologically developed. This impact fading with the increasing distance tothe global frontier and becomes not significant far from it. The overallestimated marginal effect from the industry analysis depends on thetechnological stage of the country-industry pair and the global frontier.
Thefirm analysis results only partially confirm these positive findings – onlyone-fifth of the firms benefit, while the least productive in the sampleexperience a negative effect. The authors suggest that this difference betweenfirm and industry level may be due to the sample composition in terms ofcountries or years included or to the entry-exit dynamics of the firms, inwhich the least efficient firms will exit the market, what will increaseoverall productivity of the industry. However, the empirical result explains only 16,5% of overallvariation, yet the most results are significant at 1% level. Lanoie et. al (2011) argue that theirstudy is the first to empirically detect the impact of all three channels ofPorter Hypothesis. As for its “strong” variation, the proxy for businessperformance and environmental policy instruments – command-and-controlregulation, environmental related taxes were used. The database consists of4,200 facilities in 7 OECD countries – USA, Canada, Japan, Germany, France,Hungary, and Norway.
Output presents that the negative impact of directstringency on productivity (-0.078) is detected. However,indirect effect (through environmental R) is positive.Jaraite and Maria (2012) studied the effect of environmental policy in terms of”enhancing performance of the European Union’s Emissions Trading Scheme” on efficiency andproductivity of power generation across EU states over the time period from1996 till 2007. It also contains as unwilling output and consists of emissions,generated from public electricity, combined public heat, power generation andpublic heat plants. Types of inputs such as labor, fuel inputs and netinstalled electrical capacity are included to the estimation. The key resultshows that the price of emissions positively effects the efficiency, however,there are no significant effect observable for productivity.
Similar resultsreceived Rubashkina et al. (2015), analysing policy stringency on the productivitygrowth in 17 European countries from 1997 till 2009, with a focus onmanufacturing sectors by the instruments variable approach. Pollution abatementand control expenditures (PACE) were used as a proxy for the environmentalregulation stringency, while total factor productivity (TFP) for the sectoraleconomic performance. Productivity equations were estimated in both – level andgrowth rates. The results showed any significant effect of policy stringency onthe factor productivity, across different specifications and regardlesscontrols used.
Interestingly, the model outcome represents that higher Rinvestments do not contribute to the productivity of the certain country-sector,even more – additional patents can even decline its productivity. However, Franco and Marin (2017) checkedfor the environmental tax stringency on innovation and productivity for 13manufacturing sectors with a panel of 8 European countries between 2001 and2007 not only in within-sectors but also in upstream and downstream sectors.The main results state that downstreampolicy stringency is the most relevant to productivity and innovation growth, however,within-sector regulations have positive the impact only on productivity.Besides, upstream regulations are negatively correlated with productivity. Thepossible reason might be that higher taxes imposed on downstream sectors forcestheir connecting upstream sectors to innovate and generate new technologieswhich boost the performance of downstreamers.
Aiken et.al(2009) specified regulatedand unregulated production boundaries to determine relation between pollutionabatement expenditures and productivity changes across manufacturing sectors inGermany, Japan, Netherlands and United States from 1987 through 2001 by”assigned input” model. The evidence display that pollution abatement expensesdoes not have significant negative influence on productivity growth. The survey by Rexhäuserand Rammer (2014) attempts to track the effect for Germany. Moreprecisely, how is the profitability affected when innovations are induced by voluntaryapplied environment regulation. The research is based on the collected firm-levelinformation on environmental innovation in Germany for different pollutants andwhether the innovation was induced by governmental environment regulation ornot.
The findings state that innovations that do not enhance resourceefficiency do not positively influence the productivity and vice versa. Thiseffect applies for both types of innovations – regulation-induced andvoluntarily implemented, with a larger effect for regulative ones. However, thepaper states that results represent only the resource efficiency but not the totalefficiency (productivity), so according to authors this is the argument againstthe Porter Hypothesis.
Lundgren,Marklund (2015) analysed how the firms environmental performance affect theeconomic performance (measured as profit efficiency) in Swedish manufacturingindustry over the period from 1990 till 2001, have found that if environmentalperformance is the result of the environmental policy then “it is not adeterminant for the profit efficiency”. And vice versa, when it is voluntarilyimplemented, then it affects positively and significant, so the Porterhypothesis is not supported. At the same time, Manello (2017) examines this aspect at international level, viafirm-level data of Italian andGerman firms operating in the chemical sector during the period 2004–2007. Forthe estimation DDF (difference-in-difference) framework is used to neutralizethe potential difference between economies in order to test the “win-win”opportunities, means if company is subjected to more stringent environmentalpolicy, its investment in innovation able to lift up the productivity andsimultaneously cut emission quantities.
The result states that the averagedistance to the frontier decreased over the years. This demonstrates thatplants which are implementing the best technologies available, overcome those, adoptingtechnologies with less strict environmental requirements. Generally, thedistance between the industries in two countries started to decrease afterinitial shock of the European Pollution Release and Transfer Register (E-PRTR)established in 2001.
The overall result, estimated by SequentialMalmquist–Luenberger indexes (SML) confirmed a support to the “win-win”opportunities for both Italian and German firms and also demonstrated significantcorrelation between policy stringency and TFP (Total factor productivity)growth indexes. Chatzistamoulou et al.(2017) estimates the changes of productivity output in Greek manufacturingindustries between 1993 and 2006 after the implementation of the Kyoto’sprotocol, devoted to balance the operating expenditures to provide pollutionabatement initiatives. The study uses industry-level balanced panel with 4700plant’s abatement expenditures (pollution abatement index – PAI, following Aiken, (2009)) as a proxy for thepolicy stringency. The empirical outcome shows only insignificant result on theproductivity growth with a considerable variation across industries.
On the other hand, Managi et al.(2005) question the relationship between environmental regulations,technological innovation and productivity growth in the offshore oil and gasindustry through a unique micro-level data set from the Gulf of Mexico. Analysis was done through implementingstandard statistical causality tests to detect relationship between differentproductivity indexes and regulations with DEA (Data Envelopment Analysis) tomeasure changes in productivity from 1968 to 1998. Authors argue thatparticular model have been chosen as it helps to decompose productivity partand gives a possibility to measure dynamics of different components over time.Output shows that despite increase in stringency of regulations, productivityin the market enlarged considerably. Survey on productivity growth andenvironmental regulation in Mexican and U.
S. food manufacturing industry conducted by Alpay et. al (2002), implies dual profit model assuming that profitchanges can be enforced by technological development, price adjustment orequilibrium attaining.
The data series on profits, prices, capital stocks, andenvironmental regulatory activity in Mexico and the United States were used forthe estimating model from 1971 to 1994 for Mexico and from 1962 to 1994 for theUS. The results report that regulatory inspections (as proxy for environmentalpolicy stringency) in Mexico have increased in average 2.8% of primalproductivity growth, however no clear impact of pollution abatement regulationon manufacturing productivity in the US. Berman and Bui (2001) investigatedthe effect of air quality regulation on productivity of oil refiners.Interestingly, they used a direct measure of a local pollution regulator on themost heavily regulated oil refiner in in the US – Air Basin (South Coast), LosAngeles and compared to other US regions in the second step. The totalproductivity was obtained as a sum of the data on physical amount from detailedproducts and materials in the Census of Manufacturers. The fixed-effect modelused in order to allow for the heterogeneity across plants and also to allowfor regulation differences that influence an abatement.
Inputs, which arelimited by the policy are quasi-fixed: pollution abatement capital andabatement operating costs (which include costs of labor, materials andservices). Labor, material and capital are the chosen variables. The finalestimated results cover the time of sharp stringency in regulations between1979 and 1992 and demonstrate that productivity in Air Basin considerablyincreased. Even during the period of the most stringent regulation – from 1987to 1992, refineries in South Coast still experienced growth, compared tofalling production in other US regions covered by less strict policies. At the same time, Greenstoneet al. (2012) conducted the research on how the air quality regulationsinfluence the manufacturing plants productivity (TFP) levels in US.
Usingdetailed production data of around 1.2 observations from Annual Survey ofManufacturers from 1973 till 1993, found that stringency in the policy leads tothe around 2.6 percent decline in TFP among surviving plants. Lanoie et al. (2008) in their empiricalanalysis, using GLS model to track the impact of environmental regulationstringency on the total factor productivity (TFP) of Quebeck manufacturingsector, stated and tested three types of assumption: 1. Dynamic assumption ofthe Porter hypothesis through the lagged variables; 2. Authors consider thateffect is more observable in the more polluting industries; 3. Impact isgreater in internationally competitive sectors.
When all sample is used, resultssupport the hypothesis that more stringent environmental policy leads topositive outcome for productivity, however only in dynamic case. This result isgreater for the industries that are more involved in international competition.With regard to more polluting industries, opposite have been revealed.Empirical research by Hamamoto (2006) devoted to study,whether the stringency of environmental policy in five manufacturing industriesin Japan have an effect on R activity and therefore is responsible forthe productivity growth in 1960s and 1970s. The empirical result implies, thatrelationship between pollution control and investment on innovations, using datafrom 1996 to 1976, are significant and positive. Afterwards, findingsdemonstrate that pollution control expenditures decrease average age of capitalstock and have a positive impact on modernization at 10% significance level.Further, it is found that increase in R investment, induced by increasedpolicy regulations contribute to the total factor productivity growth in thestated period.
Similar results were received by Yanget al. (2012) which examineTaiwanese manufacturing industries between 1997 and 2003 to find ifenvironmental policies (pollution abatement costs as a proxy) induce Rand productivity. The findings show that stricter environmental protection havepositive correlation with R what in turn has a conclusive influence onproductivity growth.
Referring to the industries, Rassier and Earnhart (2010) test the”strong” version of the Porter hypothesis by analysing the impact of the waste waterpolicy, measured by “waste water discharge limits” on the profits of the firms(proxied by the return of scales) operating in the chemical manufacturingindustry. Authors used a panel data analysis with a sample of quarterly data,which consists of 926 observations, including 59 chemical manufacturing firms and annual data with 337observations, consisting of 73 firms over ten years (1993-2003). The modelincludes several controls: salesgrowth, capital intensity, ageof assets, size, market share, industry concentration. Empirical results demonstratethat stringent clean water regulation serves as a negative factor for the firmsprofitability, showing with 90% confidence, that 10% increase in tighteningdischarge limit, declines return on sales by 1.7%.
Nearly the sameresults receive Gray and Shadbegian(2003), assuming that higher pollution abatement costs havea negative influence on productivity, analysing 116 pulp and paper millsbetween 1979 to 1990, with a substantial difference in integrated andnon-integrated ones.However, Sadeghzadeh (2014),developed a model to look over not only on productivity, but also on competitiveness.He suggests, that environmental regulation contributes to the productivitygrowth by reallocating input from less productive to more productive firms,making former to leave the market. Thus, market becomes more productivecompared to that it was before, however less competitive. This enablesremaining firms to increase the prices what in turn harms welfare.
The mainfindings also indicate that tighter environmental regulations deliver strongincentives to adopt cleaner abatement technologies, thus more stringent policyleads to increases in average productivity and environmental quality or a”win-win” situation. This conclusion is partly supported by Xepapadeas and de Zeeuw (1999) results, which indicate thatstringency in the environmental rules will not provide a “win-win” situation insense of reducing emissions and increasing profitability, but productivity issupposed to increase due to induced modernization of the capital stock by thestricter policy.One possible reasonfor such a mixed evidence in the literature might be, according to Ambec et al.
(2013), that the prevailingnumber of previous studies have reckoned only the static dimension, wheretechnology, processes, products and consumer preferences are fixed. While theactual dynamic competition reflects the reality with changing technological opportunitiescombined with incomplete information and complications in adjusting individual,group or corporate incentives. This means, if policies trigger innovations,which in turn reduce inefficiency, costs and stimulate technology developmentand therefore growth, they simply need time to occur by adapting optimally tothe new regulations. However, the most number of works regress proxiedstringency of regulation at time 0 on productivity at the same time 0, eventuallyit says nothing.
For instance, the studies which used the lags of three or fouryears between regulation stringency and changes in productivity allows fordynamic effect, as in Lanoie et al. (2008)or Managi et al. (2005) lead to thepositive outcome. The second misunderstandingthat not all policies, but only well-designed ones (eg. stringency,flexibility, predictability and competition friendliness) can contribute toproductivity growth (Ko?luk andZipperer, 2014. Ambec et al. (2013)).
Besides, market-based andflexible instruments like tradable allowances or performance standards are morefavourable for innovation than technological ones as they give more variationto find the best suited technological solutions to minimise the expendituresdue to compliance. For example, well-defined property rights for innovationsand R activities can benefit innovating firms and slow down diffusion.This view is also supported bySadegzadeh (2014), stating thatIf the encouraged technological change is a principal source of productivity enforcement,then the environmental regulations lead to explained by Porter, Paretoimprovement or a “win-win” situation by not only protecting the environment,but also stimulating aggregate competitiveness and productivity.
