Japan: Selected Issues
Author:
International Monetary Fund. Asia and Pacific Dept
Search for other papers by International Monetary Fund. Asia and Pacific Dept in
Current site
Google Scholar Close
Publication Date:
31 Jul 2017
eISBN:
9781484313428
Language:
English
Keywords:
ISCR; CR; inflation expectation; tax; income; investment model; savings-investment balance; sectoral investment rate; firm-entry regulation; OECD-compiled measure; Competition; Currencies; Private investment; Global
Download PDF (548.6 KB)

Selected Issues

Abstract

Selected Issues

Why ISN’t Private Investment Higher in Japan?1

What is holding back private investment in Japan? Are market competition and regulation playing a role in depressing investment? A novel firm-level dataset with almost half a million Japanese firms is used to address these questions. Findings support the hypothesis that sectoral concentration (i.e. reduced competition) has had a significant negative impact on firm- and sector-level investment. Results point to potential benefits from decreasing barriers to entry, protection of incumbents, and market concentration in some sectors. OECD measures of product market regulation indicate that there is room for further reform in the gas and telecom sectors, and deregulation of professional services.

A. Introduction

1. Private business fixed investment is key for long-run growth. Even though private business fixed investment accounts for only about 15 percent of GDP, this investment is key for long-run growth as it drives capital accumulation and innovation. Moreover, investment is important to explain business cycles as it tends to be the most volatile component of GDP.

2. Japan’s private business fixed investment has yet to recover to the levels observed in the 1980s. While private business fixed investment contributed 0.5 percentage points to 2011-16 real GDP growth, this is low relative to the 1.4 contribution over 1981-90. The 1991-2000 and 2001-10 periods show negative average contributions from private business fixed investment, at −0.3 and −0.1 percentage points respectively.2

A01ufig1

Components of Real GDP Growth

(In percent)

Citation: IMF Staff Country Reports 2017, 243; 10.5089/9781484313428.002.A001

Source: IMF staff estimates.

3. Private investment not only appears low in a Japan context, but also seems to have underperformed relative to other advanced economies. In the period before the great recession, the average contribution to real GDP growth from private gross fixed investment in Japan was below that of other major advanced economies (Table 1). For the 2010-16 period, US, UK and Canada had stronger growth contributions from private gross fixed investment than Japan, while France and Italy had less.

Table 1.

Japan: Real Growth and Contribution from Private Gross Fixed Investment

(period average, in percent)

article image
Source: WEO Database.

B. What is Holding Back Investment? Hypotheses

4. A combination of factors is likely behind depressed investment in Japan, including production offshoring, population ageing, weak corporate governance, and policy uncertainty. Weak private investment contributes to Japan’s positive savings-investment balance and current account surplus. While firm profitability has increased since 2011, investment has remained weak against the predictions from a simple Tobin’s Q model.3 There is evidence that increases in offshore production by Japanese firms (which is associated with higher cash holdings) can contribute to explaining the sluggish investment (Kang and Piao 2015). Population ageing, with associated worker shortages and implied depressed growth outlook, is likely another contributor to the investment drag. Weak corporate governance has been found to increase corporate savings (Aoyagi and Ganelli 2014) contributing to relatively weak investment. Moreover, high policy uncertainty has been found to negatively impact investment (Arbatli et al 2017).

5. Relatively low competition, the focus of this study, could be another factor depressing investment. The measure of firm entry in Japan (number of new firms started in a given year) shows a significant decrease in the early 1990s, and a steady decline since the mid-2000s, pointing to weakened market competition. Japan’s decrease in firm entry since the great recession is mostly explained by Industrials and Consumer Discretionary sectors. Other advanced economies also show a decline in firm entry in this period, but Japan’s decrease is faster than in others: Japan’s firm entry in 2014 was only 24 percent of its level in 2005, similar to France, but below Italy, Canada, and the UK (where it ranged from 31 to 54 percent of 2005 firm entry).

A01ufig2

Firm Entry

Citation: IMF Staff Country Reports 2017, 243; 10.5089/9781484313428.002.A001

Note: Number of firms (public and private) founded per year.Source: Compustat.
A01ufig3

Firm Entry, Selected Industries

Citation: IMF Staff Country Reports 2017, 243; 10.5089/9781484313428.002.A001

Note: Number of firms (public and private) founded per year.Source: Compustat.

6. Market competition and investment have an ambiguous relationship from a theoretical perspective, but empirical evidence points to a positive connection. While the presumption is that a less concentrated market structure would be optimal, it may be that increasing returns to scale are important or that incentives are needed over the longer term to invest (incentives provided by expected market power/profits). Aghion et al (2005) propose that there is an inverted U-shape relationship between market competition and investment/innovation: for lower levels of market competition, the “escape competition” effect dominates and investment increases with competition; while for higher levels of market competition, investment decreases with additional competition (as lower profits decrease the incentives to invest/innovate). An empirical investigation by Alesina et al (2005) finds that, for non-manufacturing industries, more market competition (less regulation, lower barriers to entry, etc.) increases investment (i.e., dominance of the “escape competition” effect). Therefore, increasing market concentration or decreasing market competition in Japan could be a contributing factor to the observed weak investment.

7. The following analysis provides novel evidence of a positive relationship between market competition and investment in Japan. While the potential contributing factors to low investment in Japan are manifold, this analytical work aims to integrate them into the empirical analysis, conditional on data availability, and estimate the contributing role of market competition.

C. Empirical Analysis: Data Description and Summary Statistics

8. Data and sample: Firm-level data from ORBIS are used to study the drivers of firm-level and sectoral-level investment. ORBIS is a commercial database provided by Bureau Van Dijk, including information from business registries on over 130 million firms worldwide. The database goes back in time for some countries but for Japan a sizable sample is available starting in 2001. In terms of coverage, 99 percent of ORBIS data cover private companies. The ORBIS sample used for the estimation includes over three million firm-specific observations, covering the period 2001-14 for about half a million unique Japanese firms.4 The unbalanced panel includes around six years per firm on average. Close to 2,000 of the included firms are labeled as listed. Data cleaning done to deliver this sample included removal of public administration and financial firms, use of only unconsolidated accounts, and removal of firms labeled by ORBIS as having limited financial data. Due to the very significant heterogeneity in firm-level variables, outliers are excluded (by trimming top and bottom two percent of observations for the measure of investment).

9. Firm-Level Variables: Firm investment is defined as the annual change in the stock of physical capital (measured by tangible fixed assets), and the investment rate is measured relative to the stock of capital from the previous year. To include both listed and private firms, instead of a measure of Tobin’s Q (that relies on market valuation available only for listed firms), a measure of productivity is used. Productivity is proxied by profitability (profit before taxation over total assets), as frequently done in the literature.5 A measure of total firm assets (i.e. fixed plus current assets) is included to capture scale or firm size effects. Cash holdings by the firm measures cash at the bank and in hand. An indicator variable is used to identify the listed firms in the sample.

10. Sectoral-Level Variables: A time-variant Herfindahl Index (HHI) by sector is computed with data on operating revenue (turnover) to measure industry concentration and competition. HHI index is defined as follows:

H H I j t = Σ i = 1 N s i j t 2

where i represents a firm, j represents a sector, t represents the year, and sijt represents the market share of firm i in sector j in year t (between 0 and 100).6 By definition, HHIjt varies between 0 and 10,000, with 10,000 representing full concentration of the sector j in one firm. In addition, the abovementioned firm-level variables (investment, productivity, total assets, cash, and listed) are aggregated up to compute average sectoral measures (of investment, productivity, total assets, cash, and listed) using firms’ share of sectoral turnover as weights. Table 2 provides descriptive statistics for the defined variables, while Figure 1 shows HHI for selected sectors with increasing concentration over the period.7

Table 2.

Japan: Descriptive Statistics

article image
Source: Orbis.
Figure 1.
Figure 1.

Herfindahl Index for Selected Sectors

(from 0 to 10,000)

Citation: IMF Staff Country Reports 2017, 243; 10.5089/9781484313428.002.A001

Note: Sector 8 is “Lumber and Wood Products, Except Furniture”,9 is “Furniture and Fixtures”, and 18 is “Fabricated Metal Products, Except Machinery and Transportation Equipment.”

11. Additional Statistics on Cash Holdings: For the economy as a whole, cash holdings by firms presently stand at about 50 percent of GDP. However, the studied firms in the ORBIS sample held about 10 percent of GDP in cash in 2014, with total holdings standing at US$ 443 billion. Cash holdings for firms in the sample have doubled between 2001 and 2014. Each firm, on average, held US$ 1.6 million in cash in 2014. However, there are large discrepancies across firms and sectors, with three sectors amounting to over 60 percent of the cash: 27 percent of the cash held by firms in the services sector, 19 percent held by firms in wholesale trade, and 18 percent by firms in the construction sector.8 Regarding the highest cash holdings per firm, each listed firm holds US$ 23 million on average, while the average chemical firm holds US$ 8 million, and the average transportation equipment firm holds US$ 5 million. In relation to the average yearly turnover by sector, firms in mining, services, and leather rank highest, holding over 30 percent of their net sales in cash, as of 2014 (Table 3).

Table 3.

Japan: Firms Cash Holdings, 2014

article image
Source: Orbis
A01ufig4

Currency and Deposits by Non-Financial Sectors

(In percent of GDP)

Citation: IMF Staff Country Reports 2017, 243; 10.5089/9781484313428.002.A001

Source: Haver Analytics.
A01ufig5

Cash Holdings by Sector

(In percent of total cash, 2014)

Citation: IMF Staff Country Reports 2017, 243; 10.5089/9781484313428.002.A001

Source: Orbis firm-level data.

D. Empirical Analysis: Estimation and Regression Results

12. A firm-level investment model is augmented with a sectoral concentration measure to study the impact of market competition on firm-level investment. The linear regression model used is an augmented version of the Q-theory of investment, with a role for profitability/productivity together with other variables as specified in (1). HHI, a key variable for the analysis, measures the yearly degree of concentration in the sector (where higher HHI represents higher concentration or less competition). Firms’ cash holdings is another key measure given its relevance for Japan and its potential impact on investment. A measure of firm’s assets is included to control for the effect of firm size on investment, possibly via financing costs and governance differences. The indicator variable for listed firms captures investment idiosyncrasies of public firms. αt denotes time fixed effects that capture yearly economy-wide shocks, such as exchange rate fluctuations, economic uncertainty shocks, and population ageing. δi denotes firm fixed effects that capture time-invariant firm-specific factors that may affect their investment. εijt represents the residual and captures all unobserved yearly variation in firm investment. As usual, the identifying assumption needed for unbiased coefficient estimates is that the residual is uncorrelated with the included right-hand-side variables.

I i j t = β 1 ( Pr o d u c t i v i t y i j t − 1 ) + β 2 ( C a s h i j t − 1 ) + β 3 ( A s s e t s i j t ) + β 4 ( L i s t e d i j t ) + β 5 ( H H I j t ) + α t + δ i + ε i j t ( 1 )

13. A sectoral-level investment model complements the firm-level model. Using weighted firm averages by sector, sectoral yearly investment variation is studied with specification (2). In addition to the five right-hand side variables parallel to those in (1), time and sector fixed effects are included (αt and πj). Sector fixed effects capture time-invariant factors by sector. When estimating the impact of concentration on investment, an advantage of model (2) over model (1) is that it focuses on sectoral investment (instead of firm-level investment) while concentration is also measured at the sectoral level. In other words, firm-specific drivers of investment behavior are removed from the estimation in (2).

I − a v g j t = β 1 ( Pr o d u c t i v i t y − a v g j t − 1 ) + β 2 ( C a s h − a v g j t − 1 ) + β 3 ( A s s e t s − a v g j t ) + β 4 ( L i s t e d − a v g j t ) + β 5 ( H H I j t ) + α t + π i + ε j t ( 2 )

14. Results point to a negative impact of sectoral concentration on investment (Table 4). As expected, productivity shows a positive and significant impact on investment, both in the sectoral- and firm-level regressions. Firms’ size (measured by assets) also shows, for the most part, positive and significant effects on investment. The coefficient on cash holdings suggests a negative relationship with investment, in line with production offshoring behavior that has been found to boost cash holdings at the expense of domestic investment (2015 IMF WP/15/183). Importantly, sectoral concentration is estimated to have a negative and significant impact on sectoral- and firm-level investment, evidencing that a lack of a competitive environment within a sector adversely affects investment. Notably, the estimated coefficients on HHI indicate that a reduction in sectoral concentration by one standard deviation would increase a firm’s investment rate by about 0.1 percentage points (columns 3 and 4), and it would increase the sectoral investment rate by 1.0–1.8 percentage points (columns 1 and 2) which amounts to 30–50 percent of the average sectoral investment rate. An alternative way to quantify the results is to compute the hypothetical impact on sectoral investment from a decrease in sectoral concentration via a reduction in HHI by 800 (which takes place when a firm with 40 percent market share becomes two firms with 20 percent market share each) – in this case, results indicate that the sectoral investment rate would increase by 1.4–2.5 percentage points.

Table 4.

Japan: Regression Results on the Impact of Market Concentration on Investment

article image
Standard errors in brackets ** p<0.01, * p<0.05, ~~ p<0.10, ~ p<0.15

E. Priorities for Reform: Product Market Regulation

15. OECD data can help identify specific areas that may benefit from enhanced competition and product market reform to boost investment. OECD-compiled product market regulation (PMR) measures for non-manufacturing sectors cover 35 OECD countries and other non-OECD countries, with data up to 2013, and focus on network sectors, retail distribution, and professional services.9 The available OECD PMR measures apply to broadly three of the eight 1-digit SIC sector codes covered in the regression analysis above, accounting for almost 30 percent of the studied firms and 40 percent of the 2014 cash holdings.

16. The restrictiveness of Japan’s regulation stands out among OECD countries for protection of incumbent firms, energy, telecommunication, and professional services. While Japan places below OECD averages regarding overall restrictiveness of product market regulations, there are noticeable exceptions. Regulatory protection of incumbents stands out as an area with room to improve and potential to decrease barriers to entrepreneurship. Regarding specific sectoral regulation, gas (within the energy sector) and telecommunications show restrictiveness of regulation above OECD averages. Sectoral regulation for professional services also stands out as relatively restrictive.

A01ufig6

Restrictiveness of Economy-wide Product Market Regulation

(Index scale of 0-6 from least to most restrictive)

Citation: IMF Staff Country Reports 2017, 243; 10.5089/9781484313428.002.A001

Sources: OECD Product Market Regulation Database; and Koske (et al) 2013 update.
A01ufig7

Regulatory Protection of Incumbents

(Index scale of 0-6 from least to most restrictive)

Citation: IMF Staff Country Reports 2017, 243; 10.5089/9781484313428.002.A001

Sources: OECD Product Market Regulation Database; and Koske (et al) 2013 update.

17. Highlights of the main contributors to restrictiveness in product market regulation, when compared with G7 peers: OECD-compiled measures identify restrictiveness in Japan’s gas sector, mainly regarding firm-entry regulation and high market concentration. Regarding telecom, OECD-compiled measures identify relatively high public ownership and market concentration in Japan. Lastly, regarding the four professional services sectors analyzed by OECD, engineering and accounting stand out as the most restrictive in Japan relative to G7. Within engineering, relative restrictiveness in a G7 context is highest regarding conduct regulation (with relatively high regulations on prices and fees and regulations on advertising) and entry regulation (educational requirements). Within accounting, conduct regulation also shows restrictiveness among G7, with relatively high regulation on the form of business.

A01ufig8

Sectoral Regulation in Gas sector

(Index scale of 0-6 from least to most restrictive)

Citation: IMF Staff Country Reports 2017, 243; 10.5089/9781484313428.002.A001

Sources: OECD Product Market Regulation Database; and Koske (et al) 2013 update.
A01ufig9

Sectoral Regulation in Telecommunication

(Index scale of 0-6 from least to most restrictive)

Citation: IMF Staff Country Reports 2017, 243; 10.5089/9781484313428.002.A001

Sources: OECD Product Market Regulation Database; and Koske (et al) 2013 update.

18. Complementing the above regression results, previous research has found positive effects from product market reforms in areas where Japan has room for further reform, including network industries and professional services. Gal and Hijzen (IMF 2016) study product market reforms in 18 advanced economies for 1998–2013, finding that reforms have positive effects on capital, output and employment, and that their effects increase over time. They also find that product market reforms promote firm entry, particularly those that reduce entry barriers, and that credit constraints can weaken the positive impact on investment. Canton et al (EC 2014) finds that, for the EU for the period 2008–11, less strict regulation in the professional services sector (regarding entry barriers and the exercise of these professions) improved allocative efficiency and intensified business dynamics. Monteagudo et al (EC 2012) finds significant effects from the 2006–09 EU deregulation of the services sector, conservatively estimating that it lifted EU GDP by 0.8 percent.

A01ufig10

Sectoral Regulation in Professional Services

(Index scale of 0-6 from least to most restrictive)

Citation: IMF Staff Country Reports 2017, 243; 10.5089/9781484313428.002.A001

Sources: OECD Product Market Regulation Database; and Koske (et al) 2013 update.

F. Conclusions and Policy Discussion

19. Market concentration is found to have a significant negative effect on investment. Weak private investment contributes to Japan’s positive savings-investment balance and current account surplus. A combination of factors is likely at play in explaining the relatively depressed investment in Japan, including demographic changes, production offshoring, corporate governance, and policy uncertainty. While the role of most of those factors has been previously studied, this note focuses on the impact of sectoral concentration on investment using novel firm-level data for Japan. The estimated effect of market concentration on investment points to benefits from lower barriers to entry and lower regulation to boost investment and innovation.

20. Relative to OECD and G7 peers, there is room to improve product and services market regulation. OECD-compiled measures for non-manufacturing sectors indicate that there is room to reduce barriers to entry and protections to incumbents in some industries with high concentration (i.e. gas, telecom), and there is room for deregulation in professional services.

21. Other sectors (not covered by the OECD measures) may also benefit from deregulation and further reforms in Japan. These include agriculture, healthcare, childcare, and senior care. Deregulation efforts need to cover national and local regulation to foster competition and investment while avoiding regulatory arbitrage.

References

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

    • Search Google Scholar
    • Export Citation

  • Alesina, A., S. Ardagna, G. Nicoletti, and F. Schiantarelli, “Regulation and investment” Journal of the European Economic Association, 3(4), 791–825, 2005.

    • Search Google Scholar
    • Export Citation

  • Aoyagi, C., and G. Ganelli, “Unstash the Cash! Corporate Governance Reform in Japan” IMF WP/14/140, 2014.

  • Arbatli, E.C., S.J. Davis, A. Ito, N. Miake, and I. Saito, “Policy Uncertainty in Japan” IMF WP/17/128, 2017.

  • Canton, E., D. Ciriaci, and I. Solera, “The Economic Impact of Professional Services Liberalisation” EC Economics Papers 533, 2014.

  • Gal, P., and A. Hijzen, “The short-term impact of product market reforms: A cross-country firm-level analysis” IMF WP/16/116, 2016.

  • Kang, J.S., and S. Piao, “Production Offshoring and Investment by Japanese Firms” IMF WP/15/183, 2015.

  • Koske, I., I. Wanner, R. Bitetti and O. Barbiero, “The 2013 update of the OECD product market regulation indicators: policy insights for OECD and non-OECD countries” OECD Economics Department Working Papers, No. 1200, 2015.

    • Search Google Scholar
    • Export Citation

  • Monteagudo, J., A. Rutkowski and D. Lorenzani, “The economic impact of the Services Directive: A first assessment following implementation” EC Economic Papers 456, 2012.

    • Search Google Scholar
    • Export Citation
1

Prepared by Mariana Colacelli. Yihan Liu provided valuable research assistance.

2

The chart presents the contribution from private gross fixed investment (that combines residential and non-residential private investment). The contribution to real GDP growth from residential investment was 0.1 percentage points in 2011–16, and averaged close to 0 in the 1980s, 1990s and 2000s.

3

Tobin’s Q proposes that investment should increase when the market value of installed capital (which reflects profitability) is above the replacement cost of installed capital.

4

Due to the lag in reporting, most recent comprehensive data available correspond to 2014.

5

An alternative measure of productivity was used (operating revenue per employee) without altering main results.

6

Twenty-six sectors were defined as follows: 1-digit SIC codes were used to define seven of the sectors, and the other nineteen were defined with 2-digit SIC codes for manufacturing.

7

Out of the twenty-six studied sectors, highest average HHI over the period is recorded for leather, petroleum refining, mining, furniture, and apparel sectors.

8

The nineteen manufacturing sectors taken together amount to almost 25 percent of cash holdings in 2014.

9

Data cover seven network sectors (telecoms, electricity, gas, post, rail, air passenger transport, and road freight) and five services sectors (retail distribution, accounting services, legal services, engineering services, and architectural services). The four indicators of regulation in accounting, legal, engineering and architectural services are aggregated into one indicator of regulation in professional services. Koske et al (2015) provide further details on the dataset.

Same Series

Other IMF Content

Other Publishers

  • Collapse
  • Expand
Cover IMF Staff Country Reports
Japan: Selected Issues
Author:
International Monetary Fund. Asia and Pacific Dept
Volume/Issue:
Volume 2017: Issue 243
Publisher:
International Monetary Fund
ISBN:
9781484313428
ISSN:
1934-7685
Pages:
37
DOI:
https://doi.org/10.5089/9781484313428.002

Components of Real GDP Growth

(In percent)

Firm Entry

Firm Entry, Selected Industries
Figure 1.

Herfindahl Index for Selected Sectors

(from 0 to 10,000)

Currency and Deposits by Non-Financial Sectors

(In percent of GDP)

Cash Holdings by Sector

(In percent of total cash, 2014)

Restrictiveness of Economy-wide Product Market Regulation

(Index scale of 0-6 from least to most restrictive)

Regulatory Protection of Incumbents

(Index scale of 0-6 from least to most restrictive)

Sectoral Regulation in Gas sector

(Index scale of 0-6 from least to most restrictive)

Sectoral Regulation in Telecommunication

(Index scale of 0-6 from least to most restrictive)

Sectoral Regulation in Professional Services

(Index scale of 0-6 from least to most restrictive)