Introduction. Shocks: Frictions: Importance for macro:

Size: px

Start display at page:

Download "Introduction. Shocks: Frictions: Importance for macro:"

Jacob Daniels
5 years ago
Views:

1 Basic Facts on Reallocation

2 Shocks: Introduction Business Cycle Shocks common to all sectors although not common response. Sectoral shocks: Industry and region Idiosyncratic shocks Frictions: Search/matching Adjustment costs (capital/labor) Technology adoption costs Entry/exit Importance for macro: Growth vs. Fluctuations (Emerging vs. Advanced) Nonlinear micro + Heterogeneity Non-representative agent approach

5 Job Creation and Destruction Job Creation: Employment gains from expanding and new businesses. Job Destruction: Employment losses from contracting and exiting i businesses. Job Flows are SUBSET of worker flows! Measuring job flows requires longitudinal l business data Data infrastructure projects at Census (LBD, LEHD) htm,

6 g ( E E ) 1 / X ACC a / X X.5(E E- 1 ) POS NEG skt skt ( X g 0, j s ( X g 0, j s / X / X skt skt ) g ) g SEP s / EXCESSW EXCACC X ACC ACC SEP g max( g,0) NET POS SUM skt skt EXCESSJ skt skt POS NEG skt SUM skt skt NEG skt NETskt EXCSEP j s SEP min( g k ACC ( X / X ) ACC skt SEP ( X / X ) SEP skt j s j s skt skt,0) EXCESSW ( X / X ) EXCESSW skt skt j=firm/establishment, k=worker type, s=sector

7 Basic Facts Magnitude Relationship to worker flows Productivity Dynamics? Concentration Persistence Idiosyncratic Shocks Cyclicality By Plant Characteristics Along the way: U.S. vs. Europe vs. Emerging vs. Transition

8 Job Creation Rate (left axis) Job Destruction Rate (left axis) Excess Reallocation MA(5) (right axis) Figure 5. Quarterly Job Flows in Manufacturing, Seasonally Adjusted,

9 0.096 Job Creation Rate Job Destruction Rate Quarterly Job Flows in the U.S. Private Sector, Seasonally Adjusted,

10 Magnitude Quarterly: Approximately 6% Annual: Approximately 11% Five-year: Approximately 35% Ten-year: Approximately 60% Varies SUBSTANTIALLY by industry and plant characteristic Important transitory and permanent components About 1/3 to 1/2 of Worker Flows

11 A. Job Flow Rates Data Source JOLTS, continuous units BED Time Period Sampling Interval Job Creation Job Destruction December 2000 January 2004 Monthly March 1990 June 2003 Quarterly B. Worker Flow Rates Data Source Time Period Sampling Interval Hires Separations JOLTS, continuous units LEHD, selected states, full-quarter cases LEHD, selected states, cumulative flows December 2000 January 2004 Monthly :2 to 2003:3, Quarterly :2 to 2003:3 Quarterly

12 Table: Gross Reallocation of Outputs and Inputs in U.S. Manufacturing and Retail Trade Measure Creation Share of Destruction Share of Fraction of (Expansion) Rate Creation (Expansion) Due to Entrants (Contraction) Rate Destruction (Contraction) Due to Exits Excess Reallocation Within 4-digit Industry Manufacturing, Real Gross Ot Outputt Employment Capital Equipment Capital Structures Retail Trade, Employment Real Output Source: Tabulations from the Census of Retail Trade and Census of Manufacturers

13 Country Period Coverage Employer Unit Job Creation Job Destruction Net Growth Australia Manufacturing Establishments Canada Manufacturing Establishments Canada All Employees Firms Chile Manufacturing Establishments Colombia Manufacturing Establishments Denmark Private Sector Establishments Denmark Manufacturing Establishments Estonia All Employees Firms Finland All Employees Establishments France Private Sector Establishments France Manufacturing Firms France Nonmanufacturing Firms Germany All Employees Establishments Germany(Lower Saxony) Manufacturing Establishments Italy Private Sector Firms Israel Manufacturing Establishments Morocco Manufacturing Firms Netherlands Manufacturing Firms New Zealand Private Sector Establishments Norway Manufacturing Establishments Sweden All Employees Establishments USA Manufacturing Establishments USA Private Sector Establishments USA c Manufacturing Establishments United Kingdom All Employees Firms

14 Colombia Brazil Rate Year pos 0.15 neg Rate Year pos neg Rate Chile Year Mexico pos neg Rat te Year pos neg

15 Average Quarterly Job Flow Rates in the BED, 1990:2 2003:2 Job Creation Job Destruction Net Growth Total Private Resources Construction Manufacturing Wholesale Trade Retail Trade Transportation & Utilities Information Financial Activities Professional & Business Services Education & Health Leisure & Hospitality Other Services

16 Study Coverage Employer Unit Between Reallocation Rate Within Reallocation Rate Hamermesh et al Netherlands, Firms (1996), Table 2 All Sectors, Largarde et al (1994), France, All Et Establishments blih Table 1 Sectors, Dunne et al (1997), Table 5 USA, Manufacturing Establishments

17 Slovenia. Total, Within Firm and Between Firm Job Reallocation (By Worker Type) Rate Total Between Within Age Educ Gender All Three Group Category

18 Quarterly Annual Country/Sta Coverage Sampling Accession Job Separation Job te Frequency Rate Creation Rate Rate Destruction Rate USA Private Sector Quarterly (Selected States) USA Manufacturin Quarterly (Selected States) g USA -- Private Maryland Sector 1 USA -- Manufacturin Quarterly Maryland g Denmark Manufacturin g Annual Netherlands Manufacturin g 2 Norway Manufacturin Annual g Norway Banking and Insurance Annual

19 Productivity Decomposition P s P it et et e i

20 Within, Between, Cross, Entry and Exit effects P it e C s et k P et e C s et ( P et k P it k ) P et s et set( Pet Pit k) e C e N e X s et k ( P et k P it k ) Key Issues: 1. Longitudinal links 2. Horizon 3. Measurement

21 Decomposition of TFP in U.S. Mfg Net Entry (26.00%) Within Plant (48.00%) Reallocation (26.00%)

22 Contribution of Continuing Establishments vs. Net Entry to U.S. Manufacturing Labor Productivity Growth, S h ares Continuing Establishments Net Entry

23 Contribution of Continuing Establishments vs. Net Entry to U.S. Retail Trade Labor Productivity Growth, es Shar Continuing Establishments Net Entry

24 Comments on Decomposition in Literature Some question about how to interpret aggregate g index defined in this manner Typical check (e.g., BHC and FHK) to see how this index performs relative to standard aggregate measures Common result magnitudes very similar and correlations high Not clear what correct aggregate index is Standard aggregate indexes not well-justified on theoretical grounds (e.g., Fisher conditions under which h aggregate production function exists are very stringent) Standard decomposition summarizes activity weighted micro distribution taking into account unbalanced panel

25 Concentration For Manufacturing, more than 2/3 of creation/destruction accounted for by large (greater than 25% changers). Births account for 20% of annual creation, Deaths account for 25% of annual destruction. For non-mfg, entry/exit play an even larger role. Contribution of entry/exit larger over longer horizons. Importance: Job flows can t be fully accommodated by attrition/worker flows Adjustment costs?

26 Job Flow Concentration Net Density Job Creation Mean: 6.4 percent Share with Rate > 10 pct: 0.63 Job Destruction Mean: 8.0 percent Share with Rate > 10 pct:

27 Country Percent of Job Creation or Destruction Accounted for by Plants with Growth Rates in the Indicated Interval [-2,-1) [-1,-.2) [-0.2,0) (0,0.2] (0.2,1] (1,2] United States Canada Denmark Israel

28 Persistence Measured as the fraction of the change that persists for 1 year. Roughly 50 percent of quarterly job flows persistent Roughly percent of annual ljob bflows persistent

29 Country USA Denmark Netherlands Norway Period Horizon One Year Two Years One Year Two Years One Year Two Years One Year Two Years Job Creation Job Destruction

30 Between vs. Within Sector Shifts Decomposition of Excess Reallocation into Between/Within Components. 4-digit industry accounts for 13% of Mfg.; less for other sectors. 2-digit, state accounts for 14% of Mfg; 2-digit, region, size, age, ownership (14,400 sectors in Mfg) accounts for only 39% of Mfg. Idiosyncratic Shocks Dominate!

31 Decomposition of excess job reallocation EXCESSJ( k) ( SUM NET ) ( ( X / X )( ( X / X ) g ) st st st j s jt st k jt g jt )

32 Country Period Unit Number Average Fraction of Anal ysis of Sectors Number of Workers per Sector (in thousands) Resulting from Shifts Between Sectors USA plant 448/ USA plant Denmark plant Finland plant Germany plant Italy firm Netherlands firm Sweden plant Norway plant France plant France firm France plant New Zealand plant Chile plant Colombia plant Morocco plant

33 Sectoral Differences in Flows Non-manufacturing greater than manufacturing. Small, young, single unit, low wage businesses more likely to create and destroy jobs. These factors are highly interrelated!

34 Country USA Canada Netherlands Norway Industry Period Food Tobacco Textiles Apparel Lumber Furniture Paper Printing Chemicals Petroleum Rubber Leather Stone, Clay, Glass Primary Metals Fabricated Metals Nonelectric Machinery Electric Machinery Transportation Instruments t Miscellaneous Total Manufaturing

36 Excess Job Reallocation Rates By Business Age CHL( ) 1999) COL( ) 1999) MEX( ) BRA(2000)

37 NET Employment GrowthbyBusinessAge CHL( ) COL( ) MEX( ) BRA (2000)

38 NET Employment GrowthbyBusinessAge CHL( ) COL( ) MEX( ) BRA (2000)

39 Cyclicality Recessions are times of reorganization especially in manufacturing and especially for large, mature businesses in manufacturing. Spikes in permanent job destruction are key characteristic of recessions in manufacturing Open question about most recent recession

40 Job Creation Rate (left axis) Job Destruction Rate (left axis) Excess Reallocation MA(5) (right axis)

41 0.040 Hires Rate Separations Rate

46 JobCreat JobDest EST RUS SLN UKR year Job Flows in Selected Transition Economies

47 Cyclical Link Between Job and Worker Flows Micro Nonlinearity + Heterogeneity = Complex Aggregation

48 Overview Caution about drawing inferences for representative worker/firm based upon aggregate/industry evidence Aggregate behavior may reflect micro nonlinearities + heterogeneity More than just compositional aggregation problems

49 12 14 Separation Rates as a Function of NET (Not Seasonally Adjusted Data) 0 2 Separat tion Rate NET = Hire - Separation Rate bandwidth =.8 Note: Loess Smoothing Plot using JOLTS Monthly Flows Dec 00 - Oct 04

50 12 14 Hire Rates as a Function of NET (Not Seasonally Adjusted Data) 0 2 Hire Rate NET = Hire - Separation Rate bandwidth =.8 Note: Loess Smoothing Plot using JOLTS Monthly Flows Dec 00 - Oct 04

51 Worker & Job Flow Relations Hires and Separations vs Net Growth Hiring Rate Separations Rate HUCl Mean = SUCl Mean = Net Growth Rate Strong nonlinear relations of H and S to net growth H rates rise sharply to right of 0, S rates rise sharply to left of 0 Considerable hires at contracting, separations at expanding estabs

52 Worker & Job Flow Relations Excess Worker Flows vs Net Growth Excess Worker Flow Rate VUCl Mean = Net Growth Rate Excess Flow H JC = S JD, highlights churning from last figure Excess flows increase with the magnitude of net growth Greater excess churning for expanding estabs (~ 4-5 pct vs 2-3 pct)

53 Worker & Job Flow Relations Quits and Layoffs vs Net Growth Quits Rate HUCl Mean = Layoffs Rate SUCl Mean = Net Growth Rate Quits, Layoffs also exhibit strong nonlinear patterns Quits dominate separations at expansions Quits also account for many separations at contractions Layoff/quit ratio rises as net growth rate becomes more negative

54 Actual Predicted from Growth Distribution Predicted from OLS Hires

55 Actual Predicted from Growth Distribution Predicted from OLS Separations

56 Nonlinear aggregation dn n f n h H ) ( ) ( d f S dn n f n h H ) ( ) ( ) ( ) ( dn n f n s S ) ( ) ( Cross sectional distribution shape and location matters Concentration matters! Concentration matters!

57 Taking stock Idiosyncratic shocks dominate Technology Taste Cost Time varying intensity of reallocation shocks vs. changes in incentives? Efficiency? Is it at least productivity enhancing? Measurement issues for productivity What market structure/institutions promote efficient reallocation? Cautions: Second Best and Creative Destruction

58 Taking stock Cont... Large scale job (and worker flows) driven by idiosyncratic shocks Varies by sector, cyclically, y, time period, country From here two areas of focus: Cyclicality: y Theory More structured empirical analysis Productivity it and reallocation: Theory More structured empirical analysis

Understanding Creative Destruction: Implications for Labor Markets. John Haltiwanger University of Maryland and NBER

Understanding Creative Destruction: Implications for Labor Markets John Haltiwanger University of Maryland and NBER 1 Overview Healthy, market economies are dynamic High pace of output and input reallocation