Offshoring bias in manufacturing growth across advanced and emerging economies

Transcription

1 Groningen Growth and Development Centre Offshoring bias in manufacturing growth across advanced and emerging economies Robert Inklaar University of Groningen

2 Mo>va>on Material inputs are increasingly offshored i.e. outsourced to foreign manufacturers Prime reason: cost advantages But are official sta>s>cs equipped to measure this? Answers by Houseman et al. (2011) and Reinsdorf & Yuskavage (2011) suggest no Problem: price indexes understate subs>tutability

3 Stylized example t t+1 Price change Canadian supplier % Chinese supplier 5 5 0% China share 25% 50% Input price Op>on A 0% Op>on B % A: suppliers provide different products, so total price change is (weighted) average of individual price changes (official sta>s>cs approach) B: products are subs,tutes, so prices should be averaged (rather than price changes)

4 Houseman et al. Their data: US micro data on import prices Data underlying BLS import price index Their finding: price discounts of 20 to 80% of imports from emerging economies And not due to quality differences (Mandel, 2010) Import price infla>on is too high, so MFP growth is too high as well %p in (value- added) based MFP growth (up to 50% of MFP growth)

5 This paper Focus on subs>tu>on between source countries Not domes>c vs. foreign suppliers Use data on trade unit values Cover many more countries Compare MFP growth depending on aggrega>on Op>on A (average price change) as current prac>ce Sizable bias for advanced economies (±0.3%p, 20% of measured MFP growth)

6 Observa>on 1 France Germany United.Kingdom United.States EU15 EU12 Other.advanced Other.emerging Source: WIOD RoW The share of emerging economies in material imports by manufacturing firms has grown

7 Observa>on 2 Median price difference rela>ve to EU15 and percentage of products that have lower prices than EU15, by country group, % 72% 74% 38% 60% 69% year Other5advanced Other5emerging EU12

8 Observa>on 2 Import unit values from emerging economies are ogen lower No correc>on for quality differences SITC 4- digit level, so likely considerable Average import prices of Feenstra/Romalis vary less (but quality adjustment depends on the theore>cal model) Emerging economy discount is modest compared with Houseman et al. And decreasing over >me

9 Bias in import prices of product j Op>on A: Op>on B: Δp jt MA = i Δp MB jt = Δ v ij Δuv ijt i i A: imports from a different country i is a different product B: imports from different countries are perfect subs>tutes V ijt V ijt 1 Q ijt i Q ijt 1 i

10 Which products show a bias?! Differen8ated Exchange!traded Reference!priced Advanced.economies Emerging.economies Bias es>mates by type of product, using the conserva>ve Rauch (1999) classifica>on

11 Dual MFP: Bias in MFP growth Δa = Δp X Δp Y Δp X = w j Δp D j + j M w j Δp j j Input prices are a weighted average of domes>c inputs (labour, capital, intermediates) and imported materials Assump>on: current price of imported materials = op>on A (different products) Bias is given by difference between B and A for each product j Input structure from WIOD, input growth from Socio- Economic Accounts of WIOD

12 Considerable bias in advanced countries Italy Australia Spain Luxembourg Greece Ireland Denmark Japan Canada Portugal Belgium France United7Kingdom Netherlands Germany United7States Austria Sweden Finland Korea Average: 0.3%p! Corrected7MFP7growth Bias

13 Mixed picture in emerging economies Brazil Bulgaria India Cyprus Indonesia Romania Turkey Mexico Latvia Lithuania Russia Slovenia Hungary Estonia Slovakia Poland China Czech+Republic! Corrected+MFP+growth Bias

14 Discussion (1) Bias for US is smaller than that of Houseman Bias from switching foreign suppliers < domes>c=>foreign (in US) Nega>ve bias for emerging economies: move to higher quality imports? Possibly related to integra>on into global value chains Cross- country comparisons of MFP growth are barely affected (correla>on of 0.96)

15 Discussion (2) Argument for surveying input prices Seems a preferable approach => what is the firm paying for its inputs? Ask the firm! Compare US wholesale & retail trade Old situa>on: based on sales price (adjusted for margin rates) New situa>on: BLS survey of margin prices => Result: much lower growth, perhaps due to similar switching to cheaper suppliers