Norway: Selected Issues

Abstract

Norway: Selected Issues

A Firm-Level Analysis of Productivity in Norway1

As Norway undergoes a transition away from oil and gas, boosting external competitiveness and in particular productivity is crucial to create a dynamic non-oil tradable sector that could potentially replace oil and gas. Meanwhile, productivity growth in Norway has faltered for the past decade compared to peers, reflecting both cyclical and structural factors. This paper explores two possible explanations for the lagging productivity performance, namely product market regulation (PMR) and the low level of research and innovation. An extensive dataset of mainland Norwegian firms is used to empirically assess the potential productivity gains from product market reforms as well as increasing research and development (R&D) spending.

A. Introduction

1. Norway has experienced a sharp fall in productivity growth since the mid-2000s. Norway’s average labor productivity—as measured by real output per hour worked in the mainland economy—grew rapidly during the first half of the 2000s, but started losing ground around 2005. Productivity growth in the private sector of mainland Norway has dropped from about 3 percent per annum in the 1996–2005 period to 0.8 percent during 2006–14. The slowdown in productivity trends also occurred in other advanced economies. In Norway, the relatively sharp slowdown was in part due to structural factors such as growth in labor migration that resulted in employment expansion in low skilled-intensive sectors (Productivity Commission, 2016). In addition, resource reallocation from the traditional to the oil-related and nontradable sectors during the oil boom in the 2000s—a symptom of the “Dutch disease” effects—appears to have also contributed to declining aggregate productivity growth (see first chapter of the Selected Issues).2

A02ufig1

Labor Productivity in Selected Countries

(Index, 2000=100)

Citation: IMF Staff Country Reports 2016, 215; 10.5089/9781498345705.002.A002

Sources: OECD and Fund staff calculations.Note: Labor productivity is GDP per hour worked in 2005 international US$.

2. Meanwhile, there is scope to further ease product market regulation and enhance innovation. Norway ranks favorably compared with peers in many indicators of business environment and entrepreneurship such as access to finance, bankruptcy legislation, and firm birth rates (Nordic Innovation, 2012). However, the OECD’s indicator of product market regulation (PMR) at the sector level indicate that several sectors (such as electricity, gas, rail, postal services, and retail) remain more regulated compared with best practice in peer countries.3 Barriers to entrepreneurship in general have declined more slowly than elsewhere (OECD, 2016). Also, research and innovation activity as measured by gross domestic expenditure on R&D lags behind other advanced economies at similar income levels. These factors could be constraining productivity growth and dampening private sector dynamics in the Norwegian mainland economy.

A02ufig2

Norway: Product Market Regulation, 2013

Citation: IMF Staff Country Reports 2016, 215; 10.5089/9781498345705.002.A002

Sources: OECD and Fund staff calculations.Note: PMR indicators range from 0 to 6, increasing with restrictiveness. Comparator group include Australia, Austria Belgium, Canada, Czech Republic Denmark, Finland, France, Germany, Island, Ireland, Italy, Japan, Korea, Luxemburg, Netherlands, New Zealand, Portugal, Spain, Sweden, Switzerland, and United Kingdom. Frontier refers to average best three countries in the comparator group.
A02ufig3

Gross Domestic Expenditure on Research and Experimental Development

(percent of GDP)

Citation: IMF Staff Country Reports 2016, 215; 10.5089/9781498345705.002.A002

Sources: OECD and Fund staff calculations.

3. The paper is organized as follows. Section B briefly discusses the factors constraining business dynamics in Norway, including product market regulation and the low level of research and innovation. Section C offers a quantitative perspective, using an extensive firm-level dataset to estimate the potential productivity payoffs from relaxing regulatory burden or increasing R&D investment. Section D concludes.

B. Obstacles to Private Sector Growth in Norway

Product market regulation

4. Some sectors in Norway including network industries and retail trade present scope for further deregulation.4 State ownership has diminished but remains extensive—companies with partial or complete state ownership (e.g., Statoil, Telenor, Norsk Hydro) account for about 11 percent of total employment (IMF, 2014). The government is intending further partial or complete sell-offs in a number of companies, including reduction in the state’s holding of Telenor to 34 percent (OECD, 2016). While competitive market models are operating reasonably well in telecoms and electricity, in part due to participation in an integrated open market with other Nordic countries, rail and postal services have only been partially liberalized. Recent policy initiatives aimed at improving competition in network industries include reorganizing the railway sector, establishing a new road development enterprise, and reducing Norges Posten’s monopoly on postal services (National Budget, 2016; OECD, 2016). In the retail sector, exemptions from the Competition Act still apply in book retailing, and barriers to entry remain high particularly in the grocery market (Revised National Budget, 2016). The government has proposed significant liberalization of shop opening hours, including allowing all shops to open on Sundays (OECD, 2016), but this proposal has yet to be passed by the Parliament. On the other hand, professional services (e.g., legal, accounting, engineering) appear comparatively liberalized according to the OECD indicators.

5. The relationship between product market reforms and firm productivity enjoys theoretical and empirical support. Regulations that prevent firm entry and exit can restrict competition and reduce information available to consumers. Relaxing barriers to entry (such as regulated prices or licensing requirements) would allow new productive firms to enter the market and increase competitive pressures, thereby encouraging incumbent firms to cut costs and/or improve product quality, and ultimately improve productivity. Such reforms could generate productivity gains that go beyond firms in the regulated markets by affecting downstream producers who rely on inputs from the regulated upstream sectors. For example, the deregulation of network industries could result in cheaper and better quality of network services, producing ripple effects throughout the economy. Indeed, a growing body of literature shows that benefits from reducing anti-competitive regulation extend beyond the immediate sectors being liberalized.5

6. The adverse impact of product market regulation on productivity may well be more pronounced for high-tech and knowledge-intensive sectors. A number of studies have used the framework of Aghion and Howitt (2005) to document that anti-competitive regulation hinders productivity growth in high-tech and knowledge-intensive sectors, which make intensive use of high skilled labor and ICT capital inputs. Since regulation of services hampers the efficient and dynamic allocation of resources among firms, it also slows down growth in ICT-using sectors, which use intermediate service inputs more intensively than other sectors. Moreover, an important channel through which restrictive regulations limit productivity growth is by hindering the process of convergence to best practice productivity. Such adverse effects are stronger for firms that are closer to the technology frontier and international best practices because they rely on innovation rather than imitation (Nicoletti and Scarpetta, 2003; Conway and Nicoletti, 2006; Arnold, Nicoletti, and Scarpetta, 2008).

Research and innovation

7. Despite having expanded in scope and quality over the last 20 years, Norway’s level of research and innovation remains low compared to peers. Not only does Norway spend less than advanced neighboring economies on R&D, it is also less efficient in translating R&D spending into innovation results.6 In its second-phase report, the Productivity Commission pointed out several reasons for this poor performance, including priorities and criteria influencing research funding allocation and research institutions’ management and adaptability. The Commission recommended strengthening professional strategic management at the research institutions, better cooperation between research and industry, and implementation of measures for scientific quality in funding decisions, among others (Productivity Commission, 2016). In addition, competition-enhancing product market reforms may also boost innovation activity, given that competition up to a certain level tends to induce firms to innovate (Aghion and others, 2005).7

8. Work to enhance the efficiency of Norway’s research sector is underway. In particular, the 2016 budget proposed several measures to support innovation activity, such as increasing the maximum deductibility basis under the SkatteFUNN (i.e. the Research Council of Norway) R&D tax incentive scheme and expanding appropriations for Innovation Norway’s entrepreneurship grant scheme and pre-seed capital fund (National Budget, 2016). The government is also reviewing the system of funding allocation by the Research Council of Norway to ensure quality of awarded projects and reduce administrative costs (Revised National Budget, 2016).

9. A large literature has found evidence for a positive association between R&D and productivity. While empirical estimates of the impact of R&D spending on productivity growth range from zero to substantial, a general consensus that R&D has productivity-enhancing effects appears to have emerged (see e.g., Congressional Budget Office, 2005 for a review). The rate of return on R&D has been found to be of about the same size or slightly larger than that for conventional investments.

C. A Quantitative Perspective

10. In this section, we attempt to quantify the productivity gains from relaxing product market regulation and improving innovation. While the relationships between product market reforms or innovation and productivity have been widely explored in a cross-country context, to our knowledge it has not been done specifically for Norway, at least in the recent literature. In addition, our contribution is to utilize the rich information available in firm-level data to investigate this question.

Firm-level data

11. An extensive firm-level dataset is employed to estimate the productivity payoffs of reforms. The Orbis database compiled by Bureau Van Dijk provides financial data at the firm level on value added, number of employees, and fixed assets, among other variables, allowing for the computation of firm-level productivity and other measures of firm performance. We focus on firms in the non-financial, non-oil private sector, and apply an extensive procedure to prepare the data for the analysis, including removing firms with missing key information or extreme values of financial ratios.8 The final (post-cleaning) Norway sample consists of 80,474 public and private firms for the period between 2005 and 2014, resulting in over 125,000 firm-year observations.9

12. We calculate different measures of firm productivity for the analysis. Specifically, we compute both labor productivity (i.e., real value added per worker) and three measures of total factor productivity (TFP) for each firm using three different methodologies (Box 1).

Measures of Firm-Level TFP

Three measures of firm TFP are computed for the analysis. First, an index number-based TFP measure is calculated as the Solow residual from a Cobb-Douglas production function with labor and capital as factors of production. For each 1-digit NACE sector, the labor and capital shares are obtained from the OECD STAN database. The Cobb-Douglas production function has the general form:

Aist=Yist/[ListαKist1αs]

Where Aist denotes TFP of firm i in sector s in year t, Yist is real value added, List is the number of employees, Kist is the firm’s value of real fixed assets, and αs denotes labor share in sector s. Thus, the assumption of constant returns to scale in every sector is made.

Second, a production function of the following form is estimated using OLS for each NACE sector:

lnYist=βs+αsLlnList+αsKlnKist+γt+ɛist

Year fixed effects are included to capture time-varying common shocks to all sectors. We obtain the labor and capital shares from the regressions (no longer assuming constant returns to scale), and use them to compute firm TFP as before.

Third, we estimate the same production function but using the Levinsohn-Petrin (LP) methodology of instrumenting for the unobserved productivity shock (Levinsohn and Petrin, 2003). The idea is that more productive firms tend to hire more inputs, thus rendering input use correlated with productivity and causing the OLS coefficients to be inconsistent and biased. In line with the literature, we use as instrument the firm’s working capital (defined as the difference between current assets and current liabilities), in the absence of good data on intermediate inputs. The three measures of firm TFP are highly and significantly correlated with each other. The simple correlations range from 0.45 to 0.76.

Impact of product market regulation

13. We measure the burden from PMR for all sectors in the Norwegian economy using input-output linkages between regulated and downstream sectors. As a measure of regulation, we use the OECD’s indicators for seven network sectors, retail and professional services. Regulation in those industries can affect firms in other sectors of the economy (i.e. the downstream sectors) through their use of upstream inputs. For example, a manufacturer who relies more extensively on the use of rail and postal services would bear a heavier burden from regulation in the rail and postal services sectors, either through paying higher prices or enduring lack of or sub-optimal quality of services. We call this indirect burden from regulation upstream PMR and measure it by combining the PMR indicator with the intensity of upstream input usage calculated from Norway’s input-output table for the year 2013 (Box 2).

14. The following empirical specification is used to investigate the correlation between upstream PMR and firm productivity:

Yist=β*UpstreamPMRst+γXist+Zt+Ds+Dr+ɛist

Where Yist refers to the natural logarithm of firm productivity (either labor productivity or TFP), UpstreamPMRst denotes the indicator of upstream regulation in the downstream sector s, Xist is a vector of firm-level control variables (e.g., leverage defined as the ratio of total debt to total assets and company age10), Zt is the output gap to capture the economy’s cyclical condition, and Ds and Dr are sector and region fixed effects. The β coefficient is expected to be negative, that is, more restrictive regulation is expected to correlate with lower firm productivity. We run the regressions by firm size class (i.e. micro, small, medium, and large) to allow for the impact of deregulation to vary across firms of different sizes.11

15. We also test the hypothesis that PMR has differential impacts across sectors based on their innovation intensity. We use Eurostat’s taxonomy of high- and medium-technology manufacturing sectors and knowledge intensive services at the NACE 3-digit level to classify firms into two categories—‘HTKIS’ (high-tech and knowledge-intensive sector) firms and ‘non-HTKIS’ firms.12 The idea is that a higher level of product market competition would be expected to spur innovation particularly for firms in technology or knowledge intensive sectors, thereby generating larger productivity gains. To test this hypothesis, we augment the baseline specification by interacting the Upstream PMR variable with the HTKIS indicator (which takes value of one for HTKIS firms and 0 for others). Thus, the δ coefficient would give the additional productivity impact of PMR on HTKIS firms over and above that for non-HTKIS firms.

Yist=β*UpstreamPMRst+δ*UpstreamPMRst*HTKIS+θ*HTKIS+γXist+Zt+Ds+Dr+ɛist

Measuring Indirect Regulatory Burdens

The OECD indicators of PMR are used to measure regulatory provisions in seven network sectors, retail trade and professional services covered in the analysis over the sample period. The seven network sectors include air transport, electricity, gas, post, rail, road transport, and telecom, and professional services comprise of accounting, legal, architect, and engineer). The evaluation of the network sector-specific PMRs follows a bottom up approach, aggregating data on entry regulation, public ownership, vertical integration, market structure, and price controls. Sector regulation of retail trade is assessed by compiling evaluations of six dimensions, i.e. entry regulation, restrictions on shop size, protection of existing firms, regulation of shop opening hours, price controls, and promotions or discounts. Similarly, professional services regulation is examined in two main areas of entry and conduct regulation. The scale of the PMR indicators ranges from 0 to 6, with higher values indicating more regulation. The indicators are provided on a yearly basis for network industries but they are only available every five years in 2003, 2008, and 2013 for retail trade. The regulated network sectors, retail trade, and professional services account for about 26 percent of total output in the economy.

From the Norwegian input-output table for 2013, we extract information on the use of inputs for each of the NACE Revision 2 sectors as well as their output. The variation in input usage across industries called input intensity allows us to extend the regulatory burden on network sectors, retail trade, and professional services to the entire economy, thereby capturing the indirect regulatory burden from upstream sectors on all firms. Using both the PMR indicators and input intensities, we follow Bourles and others (2013) to measure the indirect regulatory burden from regulation in upstream industries on downstream sectors. More specifically, we aggregate PMRs and input intensities (from upstream regulated sectors) for each downstream two-digit level sector as follows:

UpstreamPMRdt=Σu=17PMRut*Intensitydu

PMRut is the direct regulatory burden for regulated sector u at time t, and Intensitydu refers to sector-specific input intensities of downstream sector d from upstream regulated sector u, measured as the units of regulated product u that are needed to produce one unit of final output in sector d. Thus, UpstreamPMRdt measures the indirect regulatory burden that the downstream sector d is subject to at time t, calculated as the weighted average of the direct regulatory burden in regulated sectors and the sector-specific input intensities. The text figure below illustrates the level of upstream PMR from the seven network sectors, retail trade, and professional services for selected two-digit-level downstream sectors in Norwegian economy. With varying input dependency on product in regulated sectors, the downstream sectors are subject to upstream product market regulation from the seven network sectors, retail trade, and professional services that ranges from 0.002 to 0.23.

A02ufig4

Upstream PMR for selected sectors 1/

Citation: IMF Staff Country Reports 2016, 215; 10.5089/9781498345705.002.A002

Sources: Statistics Norway, OECD and Fund staff calculation.1/ Upstream PMR takes into account regulations in the seven net work industries, retail trade, and professional services.

16. Estimation results indicate that regulation in upstream sectors significantly affects firm productivity in downstream sectors. The results point to a negative and significant correlation between upstream PMR and firm productivity in downstream sectors, and are robust to multiple specifications and different productivity measures (Tables 1 and 2).13 Firms operating in sectors that rely more heavily on inputs from the regulated industries are likely to be less productive than others. Our results also suggest that the impact of PMR on firm productivity varies by firm size: it is most pronounced for medium firms and least pronounced for large firms, with the impact on micro and small firms being somewhere in the middle. For example, a one standard deviation reduction in PMR is associated with higher TFP by over 15 percent for medium-sized firms, but only by 6 percent for large firms.14 The magnitude of the estimated impact is similar to that in other comparable studies on productivity and PMR (see e.g. Lanau and Topalova, 2016 for Italy; Geng, Ho and Turk, 2016 for Denmark). Finally, it is worth noting that the size of the coefficients is similar for labor productivity and TFP, but the explanatory power of the regressions is higher using TFP than labor productivity as dependent variable.

Table 1.

Effect of Upstream PMR on Downstream Labor Productivity

article image
Robust standard errors in brackets. *** p<0.01, ** p<0.05, * p<0.1.
Table 2.

Effect of Upstream PMR on Downstream TFP

article image
Robust standard errors in brackets. *** p<0.01, ** p<0.05, * p<0.1.

17. We also find evidence that PMR affects innovation intensive firms disproportionately. The coefficient on the interaction between Upstream PMR and HTKIS dummy is negative and significant (for some size classes), indicating that HTKIS firms tend to bear a relatively heavier burden from anti-competitive regulation. The differential impact is again largest for the medium size class, about three times as large for HTKIS firms as for non-HTKIS. These findings are consistent with those reported for OECD countries (Nicoletti and Scarpetta, 2003; Conway and Nicoletti, 2006; Arnold, Nicoletti, and Scarpetta, 2008; Moreno-Badia, 2009).

18. Narrowing the gap between PMR in Norway and the frontier would generate sizable windfall productivity gains. In a stylized policy experiment, we use the estimated coefficients from Table 2 to calculate the average change in steady-state firm TFP from reducing Norway’s upstream PMR indicator such that a quarter of the distance between Norway and the frontier is closed. This would mean deregulation in all upstream sectors, including the seven network industries, retail, and professional services. Our calculations suggest that such deregulation would increase average firm TFP in Norway by roughly 40 percent, with greater benefits accruing to small and medium-sized firms relative to larger ones. In addition, HTKIS firms would record much larger productivity gains compared to non-HTKIS firms. Since the regulatory gaps between Norway and the best practice are particularly large in the gas and postal services sectors, these industries present more scope for deregulation than others.

A02ufig5

Impact of Partially Closing PMR Gap between Norway and Frontier on Average Firm TFP, by Size

(Percent change)

Citation: IMF Staff Country Reports 2016, 215; 10.5089/9781498345705.002.A002

Sources: Fund staff estimates.Note: Assuming a quarter of the gap is closed.

19. These results should be interpreted with the usual caveats. They can only be indicative of potential productivity gains from deregulation. As is well known, it is an empirical challenge to isolate the impact of any structural reform from that of other reforms that may be implemented at or around the same time. In addition, the OECD’s PMR indicators—although widely used in the empirical literature—are only crude proxies for the state of regulation in any country, which makes cross-country comparison problematic. In any case, some degree of regulation in certain sectors may be justified by other policy considerations or societal preferences, which arguably makes simply lowering regulation to the level of the “best practice” somewhat of a stylized policy experiment.

Impact of R&D investment

20. The empirical specification to test the effect of R&D spending takes the following form:

Yist=β*RDst+γXist+Zt+Ds+Dr+ɛist

where RDst is the logarithm of R&D expenditure at the sector level provided by the OECD, and Yist is the log of either firm value added or productivity (TFP) (other notations remain as in the previous section). This specification can be easily derived from a Cobb-Douglas production function with R&D capital as one of the factor input. The coefficient β gives the elasticity of output or productivity with respect to R&D investment. As R&D expenditure is measured at the sector level, this elasticity could be interpreted as also capturing the positive spillover effect of the R&D spending undertaken by other firms in the same sector.15

21. Results indicate that R&D investment has a positive effect on firm performance (Table 3). The elasticity of value added with respect to R&D spending is estimated to be about 0.24−0.3 depending on firm size. That is, a 10 percent increase in the sector’s R&D expenditure is associated with a 2.4−3 percent improvement in firm’s value added. This magnitude falls in the mid-range of elasticity estimates in the literature, which range from close to 0 to about 0.5 depending on the sample and the methodology (see e.g. Congressional Budget Office, 2005 for a summary). The value added elasticity is remarkably stable across firm size, whereas the TFP elasticity is highest for medium-sized firms. Neither elasticity is statistically significant for large firms.

Table 3.

Effect of R&D Spending on Firm Value Added and TFP

article image
Robust standard errors in brackets. *** p<0.01, ** p<0.05, * p<0.1.

D. Conclusion

22. There is ample scope for improving Norway’s productivity performance. The deterioration in productivity growth over the past decade has several structural components (e.g., Dutch disease, immigration) that can only be gradually unwound. Moreover, Norway’s well-developed policy and institutional framework implies that low-hanging fruits are limited. Nevertheless, our analysis, which focuses on product market regulation and research/innovation, find some evidence of the potential for productivity gains in those areas. In particular, relaxing the existing constraints in the product market (e.g., state ownership and other barriers to entry in certain sectors) and boosting (both the quantity and quality of) R&D spending are found to associate with higher firm productivity, with larger impact on high-tech and knowledge-intensive firms—the building blocks of the “new economy.” Our quantitative perspective supports the recommendations by the Productivity Commission, and highlights the urgency and importance of building a dynamic and productive private sector in the Norwegian mainland economy that will ultimately need to replace natural resources as the main engine of growth.

Appendix I. Data Sample and Cleaning Procedure

Our sample includes all firms for which key variables are provided, including value added and the number of employees. Our data comes from the commercial Orbis database provided by Bureau Van Dijk. We retrieve the universe of firm-level data available over the period 2005-14, resulting in a total of 727,669 firm-year observations. We select unconsolidated financial statements of companies where available and consolidated statements otherwise, excluding subsidiaries to avoid double-counting. The firms are distributed geographically across 436 regions in Norway, including Nord-NorgeNordlandBodo, Nord-NorgeTromsTromso, Nord-OstlandetAkershusAsker, Nord-OstlandetAkershusBarum, OstlandetAkershusSkedsmo, OstlandetOsloOslo, OstlandetOstfoldFredrikstad, SorlandetVest-AgderKristiansand, TrondelagSor-TrondelagTrondheim, VestlandetHordalandBergen, VestlandetMore og RomsdalAlesund, VestlandetRogalandSandnes, VestlandetRogalandStavanger.

A number of filtering rules are applied to the original sample. Following the literature, we exclude all firms in the mining and quarrying industry (to focus on the non-oil economy), financial services industry (where high leverage is not an indication of distress and liquidity is held to meet regulatory requirements and not to undertake positive net present value investment projects) and in public administration and defense (Fama and French, 1992; Bates and others, 2009). We also delete observations with negative values for key variables of interest- such as current assets, fixed assets, total assets, leverage, shareholder funds, sales, and cost of employees; we drop the bottom and top 5 percent of the distribution of return on assets and return on equity. Our final sample includes 80,474 firms distributed across 17 major sectors employing close to 1.4 million workers. The majority of firms belongs to wholesale and retail trade, followed by construction, manufacturing, professional services, and information and communication sectors.

The majority of firms in Norway are very small privately-held firms. We group firms in different size categories using the number of employees. Micro firm employ less than 10 employees (69 percent of the sample), firms with employees less than 50 but more than 10 are labeled as small (26 percent of total), medium firm have between 50 and 250 employees (4 percent of total), and above that are large firms (1 percent). The overwhelming majority of firms (99.7 percent), suggesting that focusing on large or listed firms only is likely to provide an incomplete picture of economic activity in Norway. Also, 84 percent of firms are active, and the rest is either dissolved or in liquidation. We keep both active and inactive firms in our sample to capture the dynamics of the market in terms of not just entry but also exit.

Firms of different size have different asset composition and funding structure. In Norway, small firms invest much less in fixed assets than medium and large firms and a higher fraction of their assets is kept liquid. On the funding sources, small firms rely more on equity than debt financing. Noteworthy is that profitability indicators are greater for small than for large firms. Finally, 17 percent of firms in our sample are start-ups (established less than 5 years ago), 40 percent are young (between 5 and 10 years of operations), 42 percent are mature (between 10 and 35 years of age), and the remaining 1 percent have been in the market for more than 35 years.

Table A1.1.

Firm Distribution, Value Added, and Employment across Sectors

article image
Table A1.2.

Asset Composition, Funding Structure, and Profitability across Firm Size

article image

Appendix II. Variables Definition and Key Descriptive Statistics

Description and sources of all variables entering the regressions appear in Table A2.1.

Table A2.1.

Variables Definition and Sources

article image

Summary statistics on the key variables entering the empirical specification appear in Table A2.2. Since we keep both active and inactive or dissolved firms, the latter typically may have negative equity and hence the debt-to-assets ratio that exceeds 100 percent.

Table A2.2.

Descriptive Statistics for Variables Entering the Regressions

article image

Labor Poductivity, TFP, and R&D variables are in logs; Upstream PMR, Firm Leverage, and Output Gap variables are in percent.

Appendix III. Additional Results

The results for upstream PMR are robust to using alternative productivity measures. In addition to the results for labor productivity and the Levinsohn-Petrin measure of TFP reported in the text, we test the sensitivity of our results to using two alternative TFP measures (described in Box 2). The baseline results hold in both robustness checks (Tables A3.1 and A3.2).

Table A3.1.

Effect of Upstream PMR on TFP—Solow Residual

article image
Robust standard errors in brackets. *** p<0.01, ** p<0.05, * p<0.1.
Table A3.2.

Effect of Upstream PMR on TFP—OLS

article image
Robust standard errors in brackets. *** p<0.01, ** p<0.05, * p<0.1.

References

  • Aghion P., N. Bloom, R. Blundell, R. Griffith, and P. Howitt, 2005, “Competition and Innovation: An Inverted-U Relationship,The Quarterly Journal of Economics, Vol. 120, No. 2, pp. 701728.

    • Search Google Scholar
    • Export Citation
  • Aghion, P., and P. Howitt, 2005, “Appropriate Growth Policy: A Unifying Framework,2005 Joseph Schumpeter Lecture, 20th Annual Congress of the European Economic Association, Amsterdam, August 25.

    • Search Google Scholar
    • Export Citation
  • Arnold, J., G. Nicoletti, and S. Scarpetta, 2008, “Regulation, Allocative Efficiency, and Productivity in OECD Countries: Industry and Firm-level Evidence,OECD Economics Department Working Papers No. 616, (Paris: Organization for Economic Cooperation and Development).

    • Search Google Scholar
    • Export Citation
  • Arnold, J., B. Javorcik, Lipscomb, and A. Mattoo, 2011, “Does service liberalization benefit manufacturing firms? Evidence from the Czech Republic,Journal of International Economics 85(1)

    • Search Google Scholar
    • Export Citation
  • Barone, G. and F. Cingano, 2011, “Service Regulation and Growth: Evidence from OECD Countries,The Economic Journal 121(555), 931957.

    • Search Google Scholar
    • Export Citation
  • Bourlès, R., G. Cette, J. Lopez, J. Mairesse, and G. Nicoletti, 2013, “Do product market regulations in upstream sectors curb productivity growth? Panel data evidence from OECD countries,The Review of Economics and Statistics 95(5).

    • Search Google Scholar
    • Export Citation
  • Conway, P., and G. Nicoletti, 2006, “Product Market Regulation in the Non-Manufacturing Sectors of OECD Countries: Measurement and Highlights,OECD Economics Department Working Paper No. 530 (Paris: Organization for Economic Cooperation and Development).

    • Search Google Scholar
    • Export Citation
  • Correa-López, M. and R. Doménech, 2014, “Does anti-competitive service sector regulation harm exporters? Evidence from manufacturing firms in Spain,BBVA Research, Working Paper, N° 14/13, Spain.

    • Search Google Scholar
    • Export Citation
  • Congressional Budget Office, 2005, “R&D and Productivity Growth,Background Paper, June 2005.

  • Cornell University, INSEAD, and World International Property Organization, 2015, The Global Innovation Index 2015: Effective Innovation Policies for Development, Fontainebleau, Ithaca, and Geneva.

    • Search Google Scholar
    • Export Citation
  • Forlani, E., 2012, “Competition in Services and Efficiency of Manufacturing Firms: Does ‘Liberalization’ Matter?,Katholieke Universiteit Leuven, LICOS Discussion Paper No. 311/2012.

    • Search Google Scholar
    • Export Citation
  • Geng, N., G. Ho, and R. Turk, 2016, “Product Market Reform and Productivity in Denmark,IMF Selected Issues Papers, Chapter 4, forthcoming, Washington DC.

    • Search Google Scholar
    • Export Citation
  • Lanau, S. and P. Topalova, 2016, “Product Market Reforms and Firm Performance: Evidence from Italy,IMF WP 16/XX, International Monetary Fund, Washington DC, forthcoming.

    • Search Google Scholar
    • Export Citation
  • Moreno-Badia, Marialuz, 2009, “Productivity Growth and Structural Reforms”, IMF Selected Issues Papers, Chapter 4, Washington DC.

  • Ministry of Finance, 2016, “National Budget 2016.

  • Ministry of Finance, 2016, “Revised National Budget 2016.

  • Nicoletti, G., and S. Scarpetta, 2003, “Regulation, Productivity, and Growth: OECD Evidence,Economy Policy, Vol. 18 (April), pp. 972.

    • Search Google Scholar
    • Export Citation
  • Productivity Commission, 2016, Second-phase Report (in Norwegian).

1

Prepared by Nan Geng, Giang Ho, and Rima Turk.

2

The traditional sector typically consists of all non-oil tradable activities (e.g., non-oil manufacturing and agriculture/fishing).

3

The “best practice” or “frontier” is calculated as the average of the three best performing countries in the comparator group. For example, the frontier for network industries as a whole consists of the UK, Germany, and Australia, while for retail industry Sweden, Australia, and New Zealand.

4

Network sectors include air transport, electricity, gas, post, rail, road transport, and telecom.

5

A number of papers have documented the presence of adverse effects from upstream inefficiencies using input-output linkages in a single country context (Arnold and others, 2011; Forlani, 2012; Correa-López and Doménech, 2014; Lanau and Topalova, 2016) and across OECD countries (Barone and Cingano, 2011; Bourlès and others, 2013).

6

For example, Norway’s innovation efficiency score—which captures the ratio of innovation output to input—is low at around 0.7, ranking at the 56th percentile among 141 economies (Cornell University, INSEAD, and World International Property Organization, 2015). The number of patents per capita also considerably lags other advanced economies. In addition, Norway appears to have significantly fewer “unicorns”, i.e. start-up companies with a value of over $1bn, and employment in high-growth companies is also lower than in comparable countries (Productivity Commission, 2016).

7

Aghion and others (2005) hypothesize that the relationship between competition and innovation follows an inverted U-shape, with higher competition initially increasing then decreasing the rate of innovation.

8

See Appendix I for a description of the sample and the procedure we implement to prepare the Orbis data for analysis.

9

There are considerably more observations for the recent years (2013 and 2014) due to missing number of employees in earlier years. The focus on non-resources part of the economy is due to the fact that productivity is a slightly different concept for oil companies, given the time-to-build between investment phase and production phase; it also depends in large part on remaining reserves.

10

We classify firms across four age classes: start-ups, young, mature, and well-established (Appendix I)

11

We classify firms into four size classes: Micro = 10 employees or fewer, Small = 11 to 50 employees, Medium = 51 to 250 employees, and Large = more than 250. Instead of running regressions by firm size class, we also try controlling for the logarithm of total assets; the results are qualitatively unchanged.

12

Eurostat classifies manufacturing industries according to their technology intensity (based on the ratio of R&D expenditures to value added) and services according to their degree of knowledge intensity (based on the share of people with tertiary education in the activity).

13

The definitions and summary statistics for the variables entering the regressions are reported in Appendix II. Tables 1 and 2 report the results for labor productivity and the Levinsohn-Petrin measure of TFP (our preferred measure), respectively. Additional results using other TFP measures are presented in Appendix III.

14

To calculate the average effect on firm productivity from reducing Upstream PMR, we keep input use intensity across all sectors constant at the average level.

15

The Orbis database also collects information on the firm’s R&D spending; however, this variable has many missing values and thus cannot be used in the analysis.

Norway: Selected Issues
Author: International Monetary Fund. European Dept.