Cross-Country Income Differences Revisited: Accounting for the Role of Intangible Capital

Transcription

1 Cross-Country Income Differences Revisited: Accounting for the Role of Intangible Capital Presented at the Fourth World KLEMS Conference, Madrid, Spain Wen Chen University of Groningen, The Netherlands May 23-24, 2016 World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

2 What is intangible capital? - Soft-side of assets Broad Category Type of Investment Computerised Information Software Databases Innovative Property R&D Mineral exploration Entertainment and artistic originals Design and other new product development costs Economic Competencies Branding (market research and long-lived advertising) Firm-specific human capital (training) Organisation capital (business process development) Source: Corrado, Hulten and Sichel (2005) World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

3 Motivation Existing literature: Increasingly recognized importance of intangible capital Country-specific growth accounting studies (e.g. CHS, 2009; Fukao et.al, 2009) Econometric analysis on intangibles and labor productivity (Roth and Thum, 2013) Yet missing: Incorporating intangibles in a development accounting exercise (intangibles data for a larger sample of countries needed) World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

4 Literature 1. The rapidly growing literature on intangibles contribution by providing intangible investment estimates for a wider range of countries 2. The vast and still expanding literature on international income differences contribution by explicitly accounting for investments in intangible capital World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

5 Coverage: Countries 60 economies (sum of GDP over 90% of the world total, covering countries at all stages of development), (17 years). World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

6 Coverage: Three Major Assets R&D Organization capital Brand equity Unconvered 26% R&D (13%) Organisation capital (25%) NA Intangibles (16%) Brand equity (20%) Data source: INTAN-Invest World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

7 Coverage: Market-sectors Exclude public sectors (NACE Rev.1): Public Adminstration (L) Education (M) Health and Social Work (N) That means: y = (sy Y + N) s L EMP k = sk K s L EMP World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

8 Measurements Expenditure-based approach Nc,t RD : Business expenditures on R&D N OC c,t : Fraction of managers wage number of managers Nc,t BE : Advertising spending + Market Research expenditures World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

9 Total intangible investment: N N= N RD +N OC +N BE Y = s M GDP + N; - where GDP (and I) are based on SNA 1993 (retrieved from UN NA) World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

10 Data Feature 1 Intangible investment positively associated with per capita income: Intangible Investment as % of MGDP' 0% 1% 2% 3% 4% 5% 6% 7% IND VNM HND PHL CHN UKR IDNPER EGY ZAFBRA BGR CRITHA ARG URY CHL TUR EST LVAVEN LTU RUS SVK HUN HRV MYS POL ROU MEX CZE ISR KOR PRT SVN MLT GRC CYP BEL DNK NLD DEUAUS FRA AUT CANIRL NZL TWN ESP Log of MGDP' per capita in 2011 PPP$ Notes: Author's calculation. The line shown in the figure is OLS regression line. The shares of intangible investment are averaged over time. GBR JPN FIN SWE ITA HKG USA CHE NOR SGP World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

11 Data Feature 2 Total Tangible Investment as % of MGDP' 21% 22% 23% 24% Tangible investment share Intangible investment share % 3.2% 3.0% 2.8% 2.6% Total Intangible Investment as % of MGDP' Notes: Author's calculation. The shares of tangible and intangible investments are averaged across countries. World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

12 Basic setup of Development accounting Benchmark production function (Hall and Jones, 1999): Y = A K α (Lh) γ per worker & CTRS: y = A k α (h) 1 α Rewrite as follows: y = A y KH ; y KH k α h 1 α World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

13 Basic setup of Development accounting Variance decomposition (y = A y KH ): var[log(y)] = var[log(a)] + var[log(y KH )] + 2cov[log(A), log(y KH )] Following Caselli (2005): VAF = var[log(y KH)] var[log(y)] World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

14 Extended Model Adding intangible capital: Y = A K α R β (Lh) 1 α β y = A k α r β (h) 1 α β Rewrite as follows: y A y KRH Using variance decomposition: VAF = var[log(y KRH)] var[log(y )] World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

15 Data Construction VAF = kα h 1 α y VAF = kα r β h 1 α β y k: PIM (δ K = 0.06; ; P I c,t from UN NA) r: PIM (δ R j ; ; P N j,c,t imputed) h: standard procedure as function of the average years of schooling s α: 1/3 (e.g. Caselli, 2005) α = 0.25; β = 0.15, following CHS (2009) World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

16 Results: Basic Model VAF = kα h 1 α y Table: Variance Accounted For of the Basic Model for 2011 Coverage var[log(y)] var[log(y KH )] VAF Own data Total Economy (60) % Own data (excl. F.USSR) Total Economy (51) % Data from PWT 8.1 Total Economy (60) % Own data Market Economy (60) % World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

17 Sensitivity to changing the exponents.5 Market economy (own data) Total economy (own data) Total economy (PWT).4 Variance Accounted For Physical capital share (%) World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

18 Results: Extended Model VAF = kα r β h 1 α β y Table: VAF of the Augmented Model for 2011 (Market Economy) Output elasticities var[log(y )] var[log(y KRH )] VAF Lower-bound α =.33 & β = % +5%-points Mid-range α =.33 & β = % +14%-points Upper-bound α =.25 & β = % +16%-points (Baseline) : denotes the difference in the explanatory power of the augmented model as compared to the basic model (i.e. VAF -VAF) in percentage points. World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

19 Varying Intangible Capital Share Variance Accounted For (Augmented Model) VAF' Upperbound specification VAF' VAF Upperbound improvement +20%-points +18%-points +16%-points +14%-points +12%-points +10%-points +8%-points +6%-points +4%-points +2%-points Additional Variance Accounted For by Intangibles %-points Intangible capital share (%) World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

20 Robustness of the main result var[log(y )] var[log(y KRH )] VAF Baseline result % +16%-points (1) Alternative OC % +15%-points (2) Dropping GRC&ESP % +16%-points (3) Dropping sample % +16%-points (4) Alternative δ j % +18%-points (5) Alternative K 0 & R % +16%-points (6) Alternative price P BS % +15%-points (7) Alternative price P GDP % +15%-points (8) Alternative price P I % +15%-points World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

21 Conclusions Intangible investment has become increasingly important over time and it is positively associated with income levels. In all variants of the model considered, differences in intangible capital systematically increases the VAF. In the baseline specification, it helps to account for another 16 percentage points of income variation, significantly diminishing the role of TFP. The explanatory potential of intangibles could be greater, given that only a subset of intangibles are captured in the current study. World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

22 Caveats One-size-fits-all output elasticities (e.g. α = 1 3 & β = 0.15). Analysis based on capital stocks rather capital services, which would have been more appropriate as short-lived assets should have a larger return in production as it would be indicated by its user cost. Results are comparable to previous studies (e.g. Caselli, 2005; Mutreja, 2014), thus it is still a useful step forward. World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

23 Thank you for your attention! World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

24 Price deflator for N Price indices of the industries that produce (in part) intangible assets, such as management consulting industry for OC, and advertising and marketing research industry for BE. R N j,t,us = P N j,t,us/p I t,us P N j,c,t = P I c,t/r N j,t,us Back World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22

25 Human capital h=e φ(s) s if s 4 φ(s) = (s 4) if 4< s (s 8) if s> 8 Back World KLEMS, Madrid Cross-Country Income Differences May 23-24, / 22