Income Differences Across Countries
advertisement
Income Differences Across Countries Pete Klenow Stanford University Society for Economic Dynamics July 6th, 2006 Vancouver, Canada 1 2000 PPP Income per capita 90th/10th 25.6 75th/25th 8.8 S.D. of logs 1.16 Source: Penn World Table 6.1 (86 countries) 2 Source: Penn World Table Tanzania Ethiopia Nigeria Malawi Pakistan India Indonesia China Egypt Thailand Brazil Russia Mexico Malaysia Argentina South Korea Israel Italy Great Britain France Germany Sweden Japan Australia Switzerland Hong Kong Canada Norway United States Relative Income Per Capita 2000 100 90 80 70 60 50 40 30 20 10 0 3 Production function α 1−α Y = K ( AhL) Y = PPP GDP pop = population L = hours worked K = PPP physical capital h = human capital per worker A = a residual 4 A useful decomposition α ⎤1−α Y L ⎡K = ⎢ ⎥ pop pop ⎣ Y ⎦ Ah 5 Point 1: Quality and Variety Quality: Perhaps 30% higher in rich vs. poor. ICP: services are “comparison resistant”. ⇒ half is missed in PPP calculations. Source: Hummels and Klenow (2005) 6 Price of Medical Care 4 2 CHE BMU NOR SWE DNK AUTUSA FIN GER ATG JAM FRA BEL YEM ESP NLD TTO LBN ISR JPN IRL NZL BHR PRTBHS AUS ITA GBR GAB QAT NGA ISL CAN SYR GRC OMN BRB BRA HRV SVN HKG ARG URY BWA TURBLZ KOR SGP EGY KNA GRD CHL VCT FJI JOR POL MEX BEN CIV VEN LCA DMA COG ECU MLI PER MAR EST SEN SVK HUN CMR MDG SWZPAN TUN RUS LVA CZE BOL KEN THA KAZLTU PHL BGR IDN MUS SLE IRN UKR ROM ZWE UZB ZMB ALB BLR MDA VNM GEO BGD MWI KGZPAK GIN ARM AZE LKA NPL TJK TKM Relative to the U.S. = 1 MKD 1 1/2 1/4 1/8 TZA 1/16 LUX 1/32 1/64 1/64 1/32 1/16 1/8 1/4 1/2 PPP GDP per worker (U.S. = 1) Source: Penn World Table (114 countries in 1996) 1 2 7 Price of Education 4 NGA 2 CHE Relative to the U.S. = 1 LBN 1 LUX MKD MWI TZA 1/2 ZMBCOG BEN SLE SEN KEN MLI MDG 1/4 GAB BWA CMR CIVGIN ZWE YEM 1/8 VNM 1/16 MDA NPL 1/32 SYR BHR MUS TUN GRC OMN KOR SVN PRTBHS TTO FJI KNA HRVTHA BLZ LCA HUN DMA JAM TURPOL SVK PER ATG ARG CZE MEX CHL VCT GRD IRNBRA URY PAK EST LVA LTURUSPAN VEN ECU IDN BOL LKA PHL ALB UKR KAZ ROM BLR BGR UZB BGD AZE KGZ 1/64 1/64 MARJOR EGY SWZ GER SWE HKG DNKUSA AUT JPNFRA NOR NLDBEL FIN QAT BMU GBR ISL ITA CANIRL NZL ESP ISR AUS SGP BRB GEO ARM 1/32 1/16 1/8 1/4 1/2 PPP GDP per worker (U.S. = 1) Source: Penn World Table (114 countries in 1996) 1 2 8 Point 1: Quality and Variety Variety: Perhaps 15% higher in rich vs. poor. Suppose 2/3 (consumer portion) missed. Taken together, factor of 30 rather than 24! Source: Hummels and Klenow (2002) 9 Point 2: L/pop Prescott, Rogerson: Explains income gap b/w U.S., Western Europe Alwyn Young: Explains 20% of growth in Asian Tigers, China Parente, Rogerson & Wright: Poor do more home work, less market work 10 Labor Force / Population 70% JPN CHN 60% SGP ZAR 50% TZA 40% 30% ISLCHE DNK THA SWE CAN USA HKG NOR JAM ROM GBR GER CMR FIN AUS COG RWA KEN MOZ BRB UGA BENGHA ZWE NER MLI SEN CAF NZL NLD PNG GNB GMB LSO CYP AUT COL MUS PRT FRA URY PRYTURPOL TWN PER MWI HTI NPL BEL ARG KOR LKA GRC TGO HUN ITA PHL ESP IDN CRI IND BOL TTO IRL CHL ISR PAN GUY BRA SLE VEN MYS FJIBWAMEX ZAF ZMB TUN DOMDZA SLV HND ECU IRN EGY GTM PAK SYR JOR BGD 20% 5 6 7 8 9 10 11 log income per worker Source: ILO via Penn World Table (97 countries in 1996) 11 Source: WDI via Caselli (2005) 12 Source: ILO via Caselli (2005), 41 countries in 1996 13 Development Accounting α ⎡ K ⎤1−α Y = L ⎢ ⎥ pop pop ⎢Y ⎥ ⎣ ⎦ 24 hA 1 14 L/pop: Open Questions ILO data suspect? 10% of Asian growth, vs. Young’s 20% Hours worked per agricultural worker? Alternatively, count home production. 15 Point 3: K/Y Richer countries have higher PPP I/Y (corr 0.6). Estimate K/Y using perpetual inventory and an initial stock guess. ⇒ richer countries have higher K/Y (corr 0.7). 16 K/Y 4 JPN CHE 3 ROM TZA 2 1 THA POL HUN LSO FIN NOR SGP FRA AUT KOR GER BEL GRC ESP SWE ISL CAN ITA DNK NLD AUS NZL ISR PRT HKG USA CYP GBR JAM PER MEX ECU DZAPAN BRA MYS GUY GNB IRN VEN ARG CRI TUR ZWE IRL ZMB PHL JOR TWN BWA CHN CHL FJI IDN TUN NPL COGHND URY DOM BRB COL ZAF TTO PRY PNG LKA MUS PAK MWI SYR IND BOL CMR MLI CAF KENBGD TGO SLV NER GMB GTM SEN BENGHA RWA EGY SLE MOZ HTI UGA ZAR 0 5 6 7 8 9 10 11 log income per worker Source: Penn World Table (97 countries in 1996) 17 Development Accounting α ⎡ K ⎤1−α Y L = hA ⎢ ⎥ pop pop ⎣ Y ⎦ N N 24 1 2 18 Forces driving K/Y variation 19 Forces driving K/Y variation Saving rates? 20 Investment Rates at Domestic Prices 1996 % Investment Rate at Domestic Prices 60 COG TKM 50 THA 40 KNA JOR JAM 30 MLI 20 KEN ZMB BEN TZA 10 NGA MWI MDG SGP CZE ATG HKG GRD BLZ JPN SWZ VNM EST HUN VCTDMA BWA TUN AZE TUR ECU CHL LCA PAN RUS LKA PHL PRT AUT ROM GAB MUS LTU NPL MDA ISR NOR IRN MEX SVN BHR BLR KGZ UKR LBN NZL HRV ZWE PER QAT GER UZB BGD MAR POL BRA ARG CHE NLD MKD GRC ESP AUS ARM PAK LVA GIN SEN BRB OMN CAN ISL IRL BEL USA ITA KAZ DNK FINFRA GBR YEM VEN SWE CMR CIV BOL EGY TTO BHS BMU URY GEO FJI ALB TJK SYR KOR SVK IDN LUX BGR SLE 0 1/64 1/32 1/16 1/8 1/4 1/2 1 2 1996 PPP GDP per worker (U.S. = 1) Source: Penn World Table 21 Forces driving K/Y variation Saving rates? NO 22 Forces driving K/Y variation Saving rates? NO Investment prices? 23 1996 Price of Equipment 8 MKD SYR Relative to the U.S. = 1 4 GAB NGA COG YEM 2 CIV EGY CHE FIN DNK BMU JOR JPN SWE ATG GER SEN AUS QAT LBN BEN GRC FRA IRLNOR HRV PRT NZL AUT UKR RUS GBR NLD ITA GIN ISL HUN MDA SVN THA SVK CZE BOL ALB PERLTUEST ESP OMN BEL ECU LVA BGR VEN IRN USA URY BWA TUR FJI ARG BRA POL ZMB CHL BLR CAN KEN VNM MEX PAN ISR AZE KAZ MWI KNA MDG CMR ROM BHR LCA GEO SGP TUN MUSBHS BRB JAM DMA GRD MAR SLE TTO KGZ TJK KOR HKG MLI ARM IDN VCT SWZ BLZ PHL LKA NPL ZWE PAK BGD UZB 1 TZA 1/2 LUX TKM 1/4 1/64 1/32 1/16 1/8 1/4 1/2 1 2 1996 PPP GDP per worker (U.S. = 1) Source: Penn World Table 24 Forces driving K/Y variation Saving rates? NO Investment prices? NO 25 Forces driving K/Y variation Saving rates? NO Investment prices? NO Consumption prices? 26 1996 Price of Consumption 4 MKD Relative to the U.S. = 1 2 1 1/2 TZA 1/4 CHE JPN SWE NOR DNK BMU FIN FRA GER AUT BEL ISLNLDSGP NGA SYR IRL ISR NZL GBRAUS ITA ESP USA GRC CAN HKG PRT KOR ATG BHS YEM SVN ARG BRA URY BHR LBN LCA HRV JAM KNA DMA CHL FJI PER VEN TTO ECUVCT GRD BLZ THA PAN JOR MEX QAT TURPOL COG HUN ZMB OMN EST CZE RUS GAB BOL LVA IRN MDGBEN SVK LTU SEN PHL CIV BWA BRB IDN ALB MAR MLI PAK MWI CMR EGY LKA KAZ ROM GEO TUN MUS BGD KEN BGR AZE UZB UKR SWZ ARM BLR ZWE KGZ VNM GIN SLE MDA NPL TJK LUX 1/8 TKM 1/16 1/64 1/32 Source: Penn World Table 1/16 1/8 1/4 1/2 1996 PPP GDP per worker (U.S. = 1) 1 2 27 Forces driving K/Y variation Saving rates? NO Investment prices? NO Consumption prices? YES 28 K/Y: Open Questions Why are richer countries better at making K? Just a reflection of h differences? Eaton & Kortum: Quality differences mask import barriers? 29 Point 4: MPK Lucas: why doesn’t capital flow from rich to poor? Lately it has: S/Y is more correlated with Y/L (0.5) than is I/Y (0.05 at domestic prices). Why doesn’t capital flow to equalize MPK’s? Caselli & Feyrer: It does! 30 Caselli & Feyrer MPK’s α Y Naive MPK ≡ K Corrected MPK ≡ α PY Y − land rents PK K 31 Naive MPK 50% SLV PRY CIV 40% EGY COG 30% BOL BDI MUS BW A MAR ZAF URY COL ECU TUN PHL VEN JOR 20% LKA PER CHL TTO MEX DZA JAM MYS CRI ZMB IRL SGP PAN PRT 10% GRC KOR NZL ESPGBR JPNFIN SW E AUS NLD ISRCAN DNK AUT FRA NOR ITA BEL USA CHE 0% 0 10,000 20,000 30,000 40,000 50,000 60,000 PPP GDP per worker Source: Caselli & Feyrer (2006), 52 countries in 1996 32 MPK Corrected for Prices, Land Rents 50% 40% 30% 20% SLV BW A SGP MUS URY 10% 0% MEXCHL JOR PHL PERMAR PRY ZAF JAM MYS COL PAN TUN LKA EGY VEN BOL TTO CIV ECU CRIDZA COG ZMB BDI 0 10,000 20,000 KOR PRT GRC 30,000 ISR ESPGBR JPNFIN SW E IRL NLD DNK ITA BEL AUT AUS FRA NOR CAN CHE USA NZL 40,000 50,000 60,000 PPP GDP per worker Source: Caselli & Feyrer (2006), 52 countries in 1996 33 MPK: Open Questions Why does K’s share rise with development? Caselli & Coleman, Hansen & Prescott No single MPK? Banerjee & Duflo. 34 Point 5: Schooling and h Higher attainment in rich countries (corr 0.9). Can estimate h using Mincerian formulation (log of h is linear in schooling). 1 more year of schooling ≈ 10% higher h 35 Years of Schooling 15 12 9 6 ZAR 3 TZA NORUSA NZL CAN SWE AUS KOR CHE GER FIN POL ISR DNK ROM HKG JPN NLD BEL GBR CYPIRL HUN ARGGRC ISL TWN BRB PAN FJI AUT PHL FRA TTO PER URY CHL MEX ESP ITA SGP VEN MYS LKAJOR CHN ECU PRY THA GUY ZAF BWA CRI MUS PRT SYR ZMB BOL ZWE COG EGY JAM DZAIRN COLTUR SLV TUN DOM IND IDN HND BRA LSO KEN PAK GHA CMR UGA GTM TGO MWI HTI PNG CAF BGD RWA SLEBENNPLSEN GMB MOZ NER MLI GNB 0 5 6 7 8 log income per worker 9 Source: Penn World Table and Barro-Lee (97 countries in 1996) 10 11 36 Development Accounting Y L = pop pop N N 24 1 α ⎡ K ⎤1−α h ⎢ ⎥ N Y ⎣ ⎦ 2 A 2 37 Development Accounting Y L = pop pop N N 24 1 α ⎡ K ⎤1−α h A ⎢ ⎥ N N Y ⎣ ⎦ 2 6 2 38 Point 6: MPH Single Mincerian return in all countries? Psacharopoulos & collaborators: higher Mincerian return in poor countries Banerjee & Duflo: when well-measured, similar across countries 39 Mincerian returns vs. schooling 30 JAM 25 20 BWA MAR 15 TWN GTM CIV BRA 10 FRA NPL TUN LKA 5 ZAF COL PAN KOR PHL SGP CHL PRY ECU KEN ARG URY DOM GHA SDN PER IDN HKG VNM KWT MYS RUS ISR GBR EST HUN ETH 0 2 4 6 8 10 12 14 16 Years of attainment Source: Banerjee and Duflo (2005); 38 countries, various years 40 Mincerian returns vs. schooling (better data) 30 25 20 15 GTM PAK COL KEN 10 SDNNPL GHA UGA CMR 5 PAN KOR JPN SGP PHL CHN CHL ECU PRY THA BOL ZAF IND ZMB ARG USA FRA URY VEN MYS DOM CAN PRT CRI FIN PER AUS GER SLVEGY CHE AUTGRC LKA ESP IDN POL GBR NLD HKG ZWE NOR CYP SWE HUN DNK NICBRA ITA 0 2 4 6 8 10 12 14 16 Years of attainment Source: Banerjee and Duflo (2005); 59 countries, various years 41 Limits to the Mincerian Approach Manuelli & Seshadri / Erosa, Koreshkova, Restuccia: ln hi = f ( si , yi , ai ), y = inputs, a = ability d ln hi ∂f ∂f ∂yi ∂f ∂ai = + + ∂si ∂yi ∂si ∂ai ∂si dsi But x-country = x-individual? 42 x-country vs. x-individual Country TFP differences ∂yi ∂yi x-individual ⇒ x-country ∂si ∂si Even more so with public schools? ∂ai ∂ai x-individual Yet perhaps x-country . ∂si ∂si 43 h: Open Questions Production function for h? Accumulation at home, on the job? Externalities? Sources of h variation across countries? 44 Point 7: Immigrants and h Immigrants provide useful info on h. Different source countries, one market. Hendricks: U.S. Census data for 1990 45 Immigrants vs. Natives in the U.S. Earnings Relative to Natives (same age, education, sex) 160% 140% JPN ZAF 120% 100% 80% YUG PRT TUR SYR ARG GRC CZE HUN LKA TWN KEN INDROMIDN URY BRB BRA EGY MYS SOV POL TTO IRN JOR CHL PAN JAM VEN GUY IRQ CRI DMA THA ECUCOLBLZ PAK FJI BGD BOL DOM CHN PER KORMEX PHL GTM SLV ETH HTI HND GHA NGA NIC DNK GBR SWE CHE NORAUS FRA BEL NZL AUT CAN IRL ITA GER NLD ISR ESP HKG PRI 60% 40% 20% 45º 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Real GDP per worker relative to the U.S. in 1990 Source: Hendricks (2002); 74 countries in 1990 46 Source: Hendricks (2002) 47 Development Accounting α ⎡ K ⎤1−α Y L = h A ⎢ ⎥ N pop pop Y ⎣ ⎦ N N 2−4 24 1 2 48 Development Accounting α ⎡ K ⎤1−α Y L h A = ⎢ ⎥ N N pop pop Y ⎣ ⎦ N N 2−4 3−6 24 1 2 49 Point 8: Technology and A Same technology in all countries? TFP gaps exist even within countries. Firms may have to invest in adoption. Why can’t FDI eliminate any differences? See Chad Jones’ recent paper. 50 Source: Parente and Prescott (2005) 51 Technology and A: Open Questions Data on barriers and investments? Channels for international diffusion? Right model? Parente & Prescott Howitt Klenow & Rodriguez-Clare 52 Points 9 & 10: Misallocation and A Perhaps less efficient allocation of K and L within poor countries. Maybe no higher average MPK, but more dispersion of MPK’s within poor countries. Parente & Prescott Schmitz Restuccia & Rogerson 53 Point 9: Agriculture and A Within agriculture, K and land less efficiently allocated in poor countries? Between agriculture and rest of economy, L less efficiently allocated in poor countries? Restuccia, Yang, & Zhu Gollin, Parente & Rogerson Caselli Handbook 54 Source: Caselli (2005) 55 Source: Caselli (2005) 56 Productivity in Agriculture Ag. Non-ag. S.D. of log Y/L 1.47 0.57 90th/10th 45 4.2 Source: Caselli (2005), 80 countries in 1985 57 Counterfactual Calculation S.D. 90th/10th Actual log Y/L 1.1 22 With U.S. emp. shares 0.6 4 Source: Caselli (2005), 80 countries in 1985 58 Ag. vs. Non-ag.: Open Questions Model of transition with Y/L gaps? Lucas, Hansen & Prescott Gaps reflect h differences? Caselli & Coleman, Jeong & Kim Gaps reflect home production? Parente, Rogerson & Wright 59 Point 10: Manufacturing and A Big TFP gaps across plants within industries. TFP gaps may imply MPK and MPL gaps. If so, large potential gains from reallocation. 60 TFP Dispersion within 4-digit industries 90th/10th 75th/25th U.S. 1.9 1.3 China 5.6 2.5 India 5.7 2.4 Source: Syverson (2004) for the U.S., Hsieh and Klenow (2006) for China and India. 61 Potential A Gains from Reallocation Melitz model (monopolistic competition between firms with different productivity). Reallocate K and L to equalize MPK and MPL across plants. Result: China and India could double A. Source: Hsieh and Klenow (2006) 62 Mfg. Misallocation: Open Questions Right model to gauge gaps and gains? Gaps tied to observable distortions? More measurement error in poor countries? 63 Recap of Development Accounting α 1−α Y L ⎡K ⎤ h A = ⎢ ⎥ N N pop pop Y ⎣ ⎦ N N 2−4 3−6 24 1 2 64 Plenty of Open Questions Quality and Variety Magnitude and sources of h differences Externalities Extent and sources of misallocation 65
