American Economic Association - Northwestern University
American Economic AssociationU.S. Economic Growth since 1870: One Big Wave?Author(s): Robert J. GordonReviewed work(s):Source: The American Economic Review, Vol. 89, No. 2, Papers and Proceedings of the OneHundred Eleventh Annual Meeting of the American Economic Association (May, 1999), pp.123-128Published by: American Economic AssociationStable URL: http://www.jstor.org/stable/117092 .Accessed: 02/08/2012 15:03Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at ms.jsp.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact support@jstor.org.American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to TheAmerican Economic Review.http://www.jstor.org
U.S. Economic GrowthSince 1870: One Big Wave?By ROBERTJ. GORDON*It is now 25 years since the growth rate oflabor productivity and of multi-factorproductivity (MFP) decelerated sharply both in theUnited States and in most other industrializednations. This "productivity slowdown" haseluded many attempts to provide single-causeexplanations. Slow productivity growth in thepast 25 years echoes slow productivitygrowthin the late 19th century. Perhaps both werenormal, and what needs to be explained is notthe post-1972 slowdown, but rather the post1913 "speedup" that ushered in the glorious60 years between World War I and the early1970's in which U.S. productivity growth wasmuch faster than before or after.This paper makes a sharp distinction between MFP growth calculated from inputs thatcombine simple measures of labor hours andthe capital stock and growth based on measures that adjust for the changing compositionof labor and capital. The first step toward anunderstandingof long-term trends is to compare like with like, splicing MFP data basedon unadjusted inputs prior to 1950 with post1950 data based also on unadjustedinputs, ascontrasted to the composition-adjustedinputsthat are now desirably incorporated into ourofficial MFP measures.The MFP record prior to 1929 still restslargely on the monumental work of JohnKendrick (1961) which, however, is based almost entirely on input quantities that lack anyadjustment for changes in composition.Edward Denison (1962, 1985) and ZviGriliches (1960) pioneered the developmentof composition adjustments for labor input.Dale Jorgenson and Zvi Griliches (1967) introduced a framework that treats the problemof composition adjustment in both labor and* Departmentof Economics, NorthwesternUniversity,Evanston IL 60208-2600. This research was supportedbythe National Science Foundation. I am grateful to ZviGriliches, Dale Jorgenson, and Jack Triplett for manyhelpful discussions.capital input in an elegant and symmetricfashion. After decades of fruitful research andconstructive advocacy by Jorgenson (e.g.,Jorgenson, 1990), in 1994 the U.S. Bureau ofLabor Statistics (BLS) adopted the Jorgensonframework for composition adjustmentfor allits publications on MFP growth over the period since 1948.L. StandardComposition-UnadjustedData:One Big WaveThe point of departure is the "standard"history of MFP growth since 1870 based onunadjusted inputs, linking Kendrick (1961)and BEA data to post- 1948 BLS data"stripped" of the usual composition adjustments. Column (ii) in Table 1 displays thetime-series behavior of standard MFP thatmight be called "one big wave."'1Annual percentage growth rates are computed over intervals that connect "normal" years and leapover such aberrationsas the Great Depressionand the two world wars and their aftermaths.This wave-shaped history shows symmetricslow MFP growth at the beginning (18711891) and end (1988-1996) of about 0.6percent per annum, faster growth at the"shoulders" in 1891-1913 and 1972-1988of 1.0- 1.2 percent, near-peak growth rates in1928-1950 and 1964-1972 of 1.9-2.1 percent, and the "winner and still champion" interval in 1950-1964 with MFP growth of 2.35percent per annum. World War I and its aftermath (1913-1928) spoil the symmetry somewhat, as there is no postwar parallel to theimpressive 1.43-percentMFP growth rate registered in that interval.The big-wave image raises at least two bigquestions: (i) "is it real?" and (ii) "whatcaused it?" This brief paper provides an introduction to both questions. The treatmentof "is' Source notes for the tables are available from the author upon request.123
AEA PAPERS AND PROCEEDINGS124TABLE 1-OUTPUTAND MULTIFACTOR PRODUCTIVITYGROWTH, 1871-1996 (ANNUAL GROWTH RATESPERCENTAGES)MAY 1999TABLE 2-LABORINPUT, WITH AND WITHOUTADJUSTMENTS FOR CHANGES IN COMPOSITIONComposition 96Slowdown, 1972-1996vs. 0.79Notes: Data refer to the nonfarm, nonhousing, privatebusiness sector.a The three columns labelled "MFP" refer to (ii) unadjusted MFP, (iii) MFP adjusted for changes in composition, and (iv) MFP adjusted for composition togetherwith the Gordon capital-quantityadjustment.it real?" combines a broad-brushsummaryofwhat is known about composition changes ininputs and some questions about what additional insight has been contributedin the past,or may be contributed in the future, by research on price-measurement biases in bothoutput and inputs. What is known about composition changes in inputs, and compositionadjustedgrowth in MFP, is summarizedin thetables.IL CompositionAdjustmentsfor LaborInputThe BLS composition adjustments in column (iv) of Table 2 combine subcomponentsof labor input stratifiedby age, sex, and educational attainment with weights based onearnings in each cell. The period of zero composition change during 1964-1979 combinesa positive contribution of increased educational attainment with an offsetting negativecontributionof the age-sex component, whichreflects the increasing share of teenagers andfemales in the labor force during that interval.Two time series of growth rates covering1913 -1979 have been constructed fromDenison's pioneering estimates of labor n,1972-1996vs. 1913-1972(v)(i) Denison, Denison, (iv) LaborHours original adjusted (BLS) .060.501.410.911.050.720.480.40 1.450.401.640.25 -0.03 1.610.420.372.180.00 2.181.850.54 2.391.160.54 2.390.54-0.060.49Note: The final column refers to labor input, taking account of the Denison composition adjustmentbefore 1950and the BLS adjustmentthereafter.The numbersin parentheses are the author's extrapolationsof the adjustedDenison method.position changes. The "original" Denison series combines his earliest (1962) estimates for1913- 1928 with his final ( 1985) estimates for1928-1979. The "original" series combinesstandardadjustmentsfor age-sex compositionand increasing years of educational attainmentwith three controversial procedures that havenot been adopted in subsequent work by theBLS or others.First, Denison made a radical assumptionthat, at the typical 50-60 hour work week observed in 1929 and prior years, workers wereso exhausted that their marginal hours wereunproductive. Hence Denison (1962 pp. 3543) assumed that at 1929 levels of hours perweek, any reduction in hours per week createda unit-elastic increase in "efficiency," thusmaking employment rather than hours therelevant measure for MFP. He reduced thenegative response of efficiency to hours reductions from unity at 1929 hours levels to-0.4 in the late 1950's, and lower responsesthereafter.Second, Denison assumed in his early(1962) writing thatfully 40 percentof the contribution to eamings of increased educational
VOL. 89 NO. 2PRODUCTIVITYGROWTHattainmentwas due to ability or other factorsinherent in individuals and should not betreated as a source of growth. In later writing(e.g., Denison, 1985) the set of factors wasbroadened to go beyond ability to family educational attainment,and the offset factor wasreduced to roughly 20 percent. Third,Denisontreated a given percentage increase in days ofeducation per year as contributingto the composition of labor input as the same percentageincrease in years of educational attainment.Since the BLS measures of postwar laborcomposition changes ignore all three ofDenison's factors, I have gone back toDenison's original tables to extract alternativemeasures of labor composition which are conceptually identical to the current BLS senes.The annual growth rates of the original andadjustedDenison series are shown in columns(ii) and (iii) of Table 2, and they differ mostduring 1913-1950 (due to the importance ofDenison's hours-efficiency adjustment) andby an amount diminishing almost to zero by1972-1979.III. Compositionand QuantityAdjustmentsfor CapitalInputThe BLS adjustments for the compositionof capital are based on the Jorgenson framework, which reweights components of the capital stock by the user cost of capital.Components with short lifetimes and correspondingly rapid depreciation rates receivehigher weights in the composition-adjustedcapital input measures ("J") than in capitalstock measures ( "K" ). Over the postwar period there has been a continuous substitutionof equipment for structuresand of short-livedequipment like computers for long-livedequipment like furniture.The BLS, which aggregates across categories of capital (equipment, structures,residential rental capital, andinventories), attributes a major compositionchange to the increased quantityof J-weightedcapital relative to the K-stock, and the growthrate of this composition effect is recorded incolumn (iii) of Table 3.Sufficient data to extend the BLS techniquebackward exist only to 1925, and as yet thereis no parallel composition-adjustment setiesavailable. In my research I have adopted the125TABLE 3-CAPITALINPUT, WITH AND WITHOUTADJUSTMENTS FOR CHANGES IN COMPOSITIONAND wdown,1972-1996vs. 1913-1972Composition(i)(iv)(v)Capital (ii)(iii) Gordon Capitalstock Gordon BLS quantity 1.38Note: The final column refers to capital input adjusted for composition effects, but not for the Gordon quantity adjustment.Jorgenson technique to reweight structuresand equipment, but not the individual components thereof, with minimal effects prior to1928, as seen in column (ii) of Table 3. Thesharp increase in the composition adjustmentfor capital after 1928 reflects the simple factthat there was a steady shift in the share ofequipment relative to structuresin the capitalstock after 1935 but not before. Indeed, the"structures-intensiveness"of the capital stockpeaked in 1919. The shift from structures toequipment, and within equipment to shorterlived equipment, suggests deep and difficultissues which are beyond the scope of this paper, particularlywhether computers are making a contribution to "true" productivity inproportion to the quantity of their attributes(speed, memory, etc.) now embedded in ournational income accounts.Otherimportantmeasurementissues involving the quantityratherthan the composition ofcapital are explored by Gordon (1998) and aresummarized in column (iv) of Table 3. Themost imnportantadjustmentis to allow for variable retirement ages that respond to net investment rates. Trucks, tractors,and structureswere not discarded at some arbitrarilyfixedlifetime during the Great Depression andWorld War II, when private investment wasnil. Instead they were "still there" to contribute to America's production miracle duringWorld War II. The most important impact is
126AEA PAPERS AND PROCEEDINGSto increase greatly the growth rate of capitalinput during 1928-1950 and to reduce itthereafter. Additional adjustments are madefor government-owned and privately operatedcapital (e.g., the Ford plant that made B-24'sat Willow Run in World War II), which boostscapital growth in 1928-1950 and reduces itthereafter,and also to include parts of government capital to avoid a bias in capital measurementas the nation shifted during the early20th century from private railway to publichighway and airway capital.IV. Alternative Measures of MFP Growthand the Questionsthat AriseI have already examined the standardMFPmeasures that are not adjusted for changes inthe composition of inputs, as displayed incolumn (ii) of Table 1. An alternative basedon the BLS input composition adjustmentsafter 1950 and the adjusted Denison measures for 1913-1950 (and an arbitrary 0.50labor-composition adjustment for 18701913) is shown in column (iii) of Table 1.As shown in the bottom line, where the slowdown (1972-1996 vs. 1913-1972) is displayed, there is virtually no contribution ofthe labor- and capital-composition measuresto explaining the slowdown. If anything,the composition adjustments deepen thepuzzle, as they average 0.72 points during 1979-1996 but a smaller 0.52 pointsduring 1913-1972. In contrast to the composition adjustments, which explain none ofthe slowdown, the Gordon capital-quantityadjustments appear to explain about 30 percent, as shown in column (iv) of Table 3.V. WhatNeedsTo Be KnownThere are many questions to raise aboutthese tables and many blank cells to be filledin. However, until now economists have ignored totally a major issue that is perhaps themost elusive blank cell of all, and this is priceindex bias. The recent Boskin Commission report offered an estimate that the ConsumerPrice Index was biased upwardby 1.1 percentin 1995- 1996, but this is of little relevance forthe study of historical changes in the trend ofMFP growth.MAY 1999A bias in capital-input price indexes contaminatesboth output and input measures, andin the Gordon (1990 p. 557) version it reducedMFP growth by only 0.17 percent over the entire 1947-1983 period, with only a very smalldifference in the reduction in the rapid-MFPperiod (1947-1973, 0.19 percent) and theslowdown period (1973-1983, 0.09 percent).One can surmise that improved estimates ofprice-index errors in earlier periods wouldlargely wash out, if they reveal thatmost errorsinvolve prices of capital input, both structuresand equipment.Considerationof data gaps points to the history of price changes from 1914 to 1947 as theblack hole where little is known. Nothing tomatch Albert Rees's (1961) seminal breakthrough in measuring prices for the 18901914 interval has yet been achieved for1914-1947. A few building blocks have recently been put in place, with Daniel Raffand Manuel Trajtenberg (1997) demonstrating dramatic declines in the price of automobiles for 1906-1941, and with WilliamWhite (1998) demonstrating similar declines in the price of tractors over much ofthe 1910- 1955 period. Set against this, however, is a set of my previous findings whichsuggest that the price index bias could belower prior to 1947 than after. In extensionsof my work on durable equipment prices(Gordon, 1996), I find much smaller ratesof bias before 1947. And in more recentwork, I find fragmentary evidence that official price indexes may greatly understateprice changes for housing, shelter, and apparel over the 1930-1970 period. To the extent that there is an upward bias in priceindexes for final goods in the 1990's, and adownward bias (or a smaller upward bias)in the 1930's through 1960's, some of the"big wave" of MFP growth might potentially be explained.VI. Explaliningthe "Wave"The research on input composition andquantity summarized here replaces the "bigwave" symmetry with a flatterprofile of corrected and adjusted MFP growth which proceeds at roughly 1 percent per annum duringthe periods 1891-1950 and 1972-1979, in
VOL. 89 NO. 2PRODUCTIVITYGROWTHcontrastto a much higher 1.7- 1.8 percent during 1950-1972 and a much lower 0.2-0.3percent during 1870-1891 and 1979-1996. Ifthat is the best that can be done with measurement, at least for now, what remains to explainthis huge difference in MFP growth ratesacross decades and epochs?Pending furtherresearch on price measurement errors, my basic explanation is perhapsthe most obvious but also the most neglected.I believe that the inventions of the late 19thcentury and early 20th centurywere more fundamentalcreatorsof productivitythan the electronic/internet era of today. I classify thoseearlier inventions into four clusters, startingwith electricity (including electric motors, theelectric light, and consumer appliances),internal-combustionengines (motor andsuburbs), "rearranging molecules" (petrochemicals, plastics, and elephone, radio, movies, and television).2The "big four" were much more profoundcreators of productivity growth than anythingthat has happened recently. Much of what weare seeing now is "second order," for example the VCR which combines TV and moviesbut does not have the fundamental impact ofthe invention of either, and much of the use ofthe internet which substitutes one form of entertainmentfor another. Enthusiasts of the internetmight consider thatthe computerhas notcreated the paperless society, but rather a duplication of electronic activities, all of whichgenerate paper, including the increasing pressure on academic societies to produce alternative paper and electronic versions of theirjournals and membership lists.Puzzles in the evolutionof long-runeconomicgrowtharethe ties thatbindeconomic historians,macroeconomists,and microeconomic expertson hedonic price equations and product composition. Yet the more researchthatemerges onspecific questions, the less is learned about theunderlyingstructureof the "big wave." The ex-2 A broaderconsiderationof consumer welfare, as contrasted with productivity itself, would add a fifth clusterincluding indoor plumbing and public infrastructure,providing running water and sanitarywaste disposal.127ercises summarizedin this paperplace more ofthe peak of MFP growthduring1950-1972 andless during1913-1950; yet the underlyingquestion remainsintact:why did the fundamentaldeterminantsof Americaneconomic growthcreatesuch a surgebetween 1913 and 1972, butneitherbefore nor after?I deeply believe that this wasa unique event that will not be replicatedin thelifetimes of our generationor thatwhich followsus, and I hope the challenge of proving mewrong stimulatesa new era of growth researchworthy of the pioneering efforts of Kendrick,Denison, Griliches,and Jorgenson.REFERENCESDenison,Edward F. The sources of economicgrowth in the United States and the alternatives before us. New York: Committeefor EJconomicDevelopment, 1962. Trends in American economicgrowth, 1929-82. Washington: BrookingsInstitution, 1985.Gordon,RobertJ. The measurementof durablegoods prices. Chicago: University of Chicago Press, 1990."The SearsCatalogRevisited:Appareland Durable goods." Unpublished manuscript,NorthwesternUniversity,April 1996. "CurrentProductivityPuzzles from aLong-Term Perspective." Unpublishedmanuscript, Northwestern University, September 1998.Griliches,Zvi. "Measuring Inputs in Agriculture: A Critical Survey." Journal of FarmEconomics, December 1960, 42(5), pp.1411-33.Jorgenson, Dale W. "Productivity and Economic Growth," in Ernst R. Berndt andJack E. Triplett, eds., Fifty years of economic measurement: The Jubilee of theConference on Research in Income andWealth. Chicago: University of ChicagoPress, 1990
American Economic Association U.S. Economic Growth since 1870: One Big Wave? Author(s): Robert J. Gordon Reviewed work(s): Source: The American Economic Review, Vol. 89, No. 2, Papers and Proceedings of the One Hundred Eleventh Annual Meeting of the American Economic Association (May, 1999), pp. 123-128 Published by: American Economic Association
American Economic Association The Invisible Hand and the Grabbing Hand Author(s): Timothy Frye and Andrei Shleifer Source: The American Economic Review, Vol. 87, No. 2, Papers and Proceedings of the Hundred and Fourth Annual Meeting of the American Economic Association (May, 1997), pp. 354-358 Published by: American Economic Association
American Economic Association The Peculiar Economics of Bureaucracy Author(s): William A. Niskanen Source: The American Economic Review, Vol. 58, No. 2, Papers and Proceedings of the Eightieth Annual Meeting of the American Economic Association (May, 1968), pp. 293-305 Published by: American Economic Association
American Economic Association Why Doesn't Capital Flow from Rich to Poor Countries? Author(s): Robert E. Lucas, Jr. Source: The American Economic Review, Vol. 80, No. 2, Papers and Proceedings of the Hundred and Second Annual Meeting of the American Economic Association (May, 1990), pp. 92-96 Published by: American Economic Association
American Economic Association Negative Time Preference Author(s): George Loewenstein and Drazen Prelec Source: The American Economic Review, Vol. 81, No. 2, Papers and Proceedings of the Hundred and Third Annual Meeting of the American Economic Association (May, 1991), pp. 347-352 Published by: American Economic Association
Fall 2019 What: Northwestern State University Robotics Competition and Smart Structures Show (RC&S 3, Fall 2019) When: Wednesday, December 4, 2019 8:00 AM – 1:00 PM Where: Northwestern State University, Student Ballroom Contact: Ms. Erin Bates (batese@nsula.edu) or Dr. Jafar F. Al-Sharab (jafar@nsula.edu) - Northwestern State University
1 Fraser Stoddart 12-Page Curriculum Vitae Dr J Fraser Stoddart / Board of Trustees Professor of Chemistry / Northwestern University Born May 24, 1942, Edinburgh, Scotland Nationality US Email stoddart@northwestern.edu Telephone 847-491-3793 Fax 847-491-1009 Group Website stoddart.northwestern.edu Address Department of Chemistry, Northwestern University, 2145 Sheridan
American Economic Association Beyond Markets and States: Polycentric Governance of Complex Economic Systems Author(s): Elinor Ostrom Source: The American Economic Review, Vol. 100, No. 3 (JUNE 2010), pp. 641-672 Published by: American Economic Association
accounting items are presumed in law to give a true and fair view. 8 There is no explicit requirement in the Companies Act 2006 or FRS 102 for companies entitled to prepare accounts in accordance with the small companies regime to report on the going concern basis of accounting and material uncertainties. However, directors of small companies are required to make such disclosures that are .
