The US High Tech Economy

8/3/2019 The US High Tech Economy

1/132

AmericasHigh-Tech

Economy

Growth, Development,

and Risks for

Metropolitan Areas

By Ross C. DeVolSenior Contributor: Perry Wong

Contributors: John Catapano and Greg Robitshek

July 13, 1999


2/132

Copyright 1999 by the Milken Institute

Milken Institute

1250 Fourth Street

Santa Monica, California 90401-1353

Add itional copies may be ordered online or by sending paym ent

(check, bank draft, or credit card information) to the above address.

For complete ordering information for this and all Milken Institute publications,

please see our Web site at ww w.m ilken-inst.org or contact us by ema il

(pu blications@milken -inst.org),

telephone (310.998.2600), or in writing.

Price: US$15.00 (includ es shipp ing and han dling)


3/132

Contents

Executive Su mmary ........................................................................................1

Introduction....................................................................................................17

Section 1 - Techn ology an d U.S. Growth Poten tial ................................19

Technology and Grow th Theory ..................................................................21

Historical Evidence of Techn ology and Grow th........................................22

IT Pessimists....................................................................................................23

IT Op tim ists ....................................................................................................24

Wh at to Con clud e? ........................................................................................28

Section 2 - Na tion al H igh -Tech In du stry Perfo rman ce..........................31

The Imp act of Technology on Economic Grow th......................................31

High-Tech Ind ustr ies Inclu ded ....................................................................34

Section 3 - Techn ology and M etropolitan Econom ic Performance ....38

Econom ic Geograp hy: Theoretical Underp inn ings ..................................39

Applications to High-Tech Ind ustries ........................................................43

High Technology Prod uction Spatial Patterns ..........................................47

Technologys Importance in Metro Growth Patterns................................48High-Techn ology Spatial Concentration ....................................................54

High-Techn ology Spatial Grow th ................................................................68

The Forces of Concentration an d Dispersion: Whos Winning? ............73

Section 4 - High -Tech In du stries an d Econ omic Risk s..........................79

Is the Business Cycle Still Relevant?............................................................81

High-Tech Ind ustries and the Business Cycle............................................83

Recession Risk: High -Tech Ind ustries and Metros....................................88

Section 5 - Key Econom ic Development and Business Plannin g

Im plications....................................................................................................95

Economic Policy and High-Tech Clusterin g ..............................................97

Imp lications for Business Planning ..........................................................104

Appendix - High-Tech Industry Profiles................................................106

References ....................................................................................................124

iii


4/132


5/132

1

Executive Summary

Section 1: Technology and U.S. Growth Potential

Technology is ubiquitous, directly or indirectly invading nearly all sectors

of the U.S. economy. Moreover, a comp elling case can be mad e that th e

high-tech sector is boosting th e long-term potential growth path of the U.S.

economy an d determ ining the relative economic success of metrop olitan

areas arou nd the country. Information technology (IT) describes the

infrastructure an d kn owledge that is necessary to make information

available rap idly. IT is increasingly th e software and commu nicationservices that pa tch equipm ent together. Advan ces in technology in recent

years have created entirely new grow th indu stries including e-commerce,

online information services, mobile commun ications, and greater ad vances

in med ical research. IT has spu rred rapid , seemingly continu ous

innovation by creating networks that generate value through p rodu ctive

interactive relationships or collaboration.

Firms have been investing in IT in an effort to boost prod uction

efficiencies, improve communication flows, and enhance overall business

operations. Firms believe that th ese investmen ts yield rewards su ch as

Advan ces in

techn ology in

recen t years havecreated en tirely new

growth indu stries,

includin g e-comm erc

online information

services, mob ile

comm un ications, and

greater advan ces in

medical research.

Figure 1

High Tech Surges in the 1990s

High-Tech vs. Total Real GDP


6/132

higher p rofitability by increasing prod uctivity and hence redu cing the

growth of unit labor costs. Our stand ard of living is d irectly tied to

prod uctivity growth . The Milken Institute has constructed an end -use

dem and measu re of high-technology prod ucts and services in an attemp t

to calibrate their rising contribution to economic growth. Its result isdisp layed in Figure 1. Since the 1990-91 recession, growth in th e high-

technology (high-tech) sector has been four times as large as in the

aggregate economy. During th e 1980s, the high -tech sector grew

app roximately twice as fast as the economy. Over the p ast three years,

growth in high-tech p rodu cts and services averaged over 20 percent,

directly elevating real GDP growth by 1.5 percentage point.

Information technology and it application has spawned much debate

amon g economists in r ecent years. IT pessimists point out th at d espite a

massive investment in these technologies, prod uctivity grow th has fallen

over the p ast 20 years. Furth ermore, they assert that IT investments could

not have imp roved p rodu ctivity growth because they comprise too small a

share of the capital stock and have m erely been substituted for more

expensive inputs, such as labor and other forms of capital. IT optimists,

how ever, make a comp elling case that outpu t grow th in services is being

understated, and that the reported productivity growth is biased

dow nw ard . IT optimists also claim tha t new technological innovations

require a substantial time p eriod before being absorbed th rough out the

economy. As eviden ce that something is indeed happ ening, they argue

that prod uctivity growth in the d urable goods indu stries has risen in the

1990s, and that over th e past th ree years, prod uctivity grow th in th e entire

economy h as accelerated.

What can w e conclude concerning the role of technology in p romoting

long-term economic growth? Some N ew Economy prop onents are

overzealous in the ad vocacy of their position. Conclud ing, based on on ly

three years of sup porting d ata, that sustainable prod uctivity growth h as

improved from 1 percent of the past 25 years to 2 percent wou ld be

prem ature. We consider ourselves to be cautious op timists. Modern

technology is altering prod uction processes through out the economy and

enabling truly globally integrated firms. With the exponential growth of

the Internet, many efficiency gains and lower costs will be realized.

Technological innovation has boosted productivity growth to the 1.5-2.0

percent ran ge, with m ore grow th p ossible. This w ill increase long-term

economic growth to between 2.5 and 3.0 percent in ou r view.

Section 2: National High-Tech IndustryPerformance

One of the m ost omnipresent influences of high technology has been in

business investment p atterns. IT hardw are has accoun ted for a growing

share of total business investment across a broad a rray of indu stries. In

mericas High-Tech Economy Milken Institute - July 13, 1999

2

Growth in the

h igh-tech sector

has averaged

four times that

of the overalleconomy du ring

the 1990s.

Technological

innovation h as

improved

prod uctivity growth

and b oosted U.S.

long-term econom ic

growth.


7/132

1970, informa tion processing equ ipment and related sectors represented 7

percent of real business equipm ent investment; however, last year it was

responsible for over 50 percent of all capital spending.

One of the most difficult tasks in analyzing high-technology industries isdefining which set of industries to include in the definition. The definition

will vary depend ing up on the research interests and d ata availability

across a nu mber of dimensions. Our primary research interests are in

determ ining the individu al contributions of high-tech industries to the

relative economic performan ce of metropolitan areas. For these reasons,

we are focusing on the va lue of outp ut for indu stries that may be

considered high technology. Manu facturing ind ustries such as dru gs,

computers and equipment, communications equipment, and electronic

componen ts are included , as are service ind ustries such as

commu nications services, comp uter and data processing services, and

research and testing services.

These ind ustries are among the fastest growing in th e United States. The

three indu stries with the most rapid grow th electronic comp onents and

accessories, compu ter and office equipment, and comp uter an d d ata

processing services are all vital information technology industries. Over

the past 20 years, high-tech ind ustries have almost d oubled their share of

indu stry outp ut in the United States to nearly 11 percent. Technology

services, at 5.8 percent of national output, are larger than technology

manufacturing.

Section 3: Technology and MetropolitanEconomic Performance

Technology is having a pervasive influence on the spatial distribution of

economic activity and , more importan tly, the relative rate of growth

amon g metrop olitan areas w ithin the United States. Prowess in

technological innovation and assimilation will likely determine the relative

success of nations in the futu re; it already is having profoun d impacts on

the regional economic landscape of the United States. Ironically, just w hen

globalization seemed to be forcing convergence among national economies

and cheap, versatile comm un ications seemed to be un derm ining the

inherent ad vantages of doing business in one p lace rather than another,

localities are emerging as important factors to the economic success ofindividu al nations. Perhaps the best ind icator of the ascendance of

regionalism is that p olicymakers from Kuala Lum pu r to Jerusalem are

busy trying to clone Silicon Valley. Geograp hic cluster ing is becoming

central to the creation and un derstand ing of what economists call

compara tive advantage in trade even in an information-age economy.

It is plain that economic geograph y a d iscipline that stud ies where

prod uction occurs has been neglected by the m ainstream. But since the

Milken Institute - July 13, 1999 Americas High-Tech Econom

3

Information

processing equ ipmen

accounts for over 50

percen t of all cap ital

spending.

High tech has

dou bled its share of

the U.S. economy ove

the p ast 20 years.

Policy m akers

from Kuala Lum par

to Jerusalem

are busy trying

to clone

Silicon Valley.


8/132

late 1980s, there h as been an intellectua l boomlet in the su bject. This

resurgence has been dubbed new economic geography and provides the

foundation for analyzing highly specialized economic activity within

regions. In essence, new economic geograph y attemp ts to provide th e

theoretical und erpinnings to explain the agglomeration processes thatresult in highly specialized economic activity within regions. The central

componen t of this theory is based u pon th e Marshallian notion that

increasing return s as in, the more th ats mad e, the cheaper it is to

make lead to competitive advantages.

Economic geograph y p osits tw o m ain opp osing forces: concentration an d

dispersion. Agglomeration forces, because of increasing returns, cause

economic activity to cluster, whereas centrifugal (or congestion) forces

pu sh economic activity outw ard . The op posing forces are in constant

competition a s they seek a spatial equilibrium. The r elative d ominance of

one over the other creates a landscape of economic activity, but it

ultimately leads the countervailing force to exert a stronger influence,

causing the geograp hic landscape to chan ge. Such externalities play a

par ticularly imp ortant role in the case of high-technology indu stries.

Much of standard location theory is applicable to stud ying wh ere high-

tech industries locate and whether sufficient critical mass is achieved to

develop clustering. Many factors interact in a comp lex, dyn amic

environmen t that ultimately determine the geograp hic evolution of

technology industries. The relative importance of the factors alters over

the stages of developm ent. The relative weights placed u pon the array of

location factors w ill differ accord ing to the specific high-tech indu stry

the major d ivide being w hether it is a man ufacturing or service indu stry.

A historical accident or exogenous force can provide a critical advan tage

for one location. Many of the traditional location factors that have always

attracted ind ustries also are imp ortant to high-tech firms. These factors are

generally referred to as cost-of-doing-business measures: tax rates or

incentives, compensation costs, land and office space costs, energy costs,

capital costs, and firms perception of the general business climate.

How ever, other factors app ear to contribute the most to high-tech firms

location d ecisions. They includ e: access to a trained/ edu cated w orkforce,

close proximity to excellent educationa l facilities and r esearch institut ions,

an existing network of suppliers, availability of venture capital, climate

and other quality-of-life factors, and the general cost of living.

We app lied a series of econometric app roaches to examine the imp ortance

of high-tech indu stries in explaining the relative economic growth of

metros. The strength of the explanatory p ower of high-tech indu stries in

determ ining the relative economic growth of metros is high, and the

relationship is robu st across most d imensions in these regressions. In one

app roach, we created a series of metro growth indices relative to the

national growth pattern. In the case of high-tech ou tpu t, the relative metro


4

Economic

geography posits

two main

opp osing forces:

concentration

and d ispersion.

High -tech firms

want access to atrained workforce,

close p roximity

to research

institutions, a

network of

supp liers, access

to ventu re capital, and

a good q uality

of life.


9/132

Milken Institute - July 13, 1999 Americas High-Tech Economy

growth index is created by comp aring grow th in metros relative to the

national average from 1990 to 1998. We chose the 1990s as the p eriod of

analysis because this is the p eriod d uring wh ich h igh-tech became vital in

determining performance between metros.

The variable w e were a ttemp ting to explain w as the relative total real-

outp ut grow th ind ex for each of the 315 metros based up on tw o factors:

movem ents in the relative metro real high-tech outp ut grow th index and

an index of concentrat ion (location qu otient) in high-tech activity in 1990.

Stand ard izing th e relationship by the concentration of high-tech ind ustries

in a m etro is imp ortant because if a metro h ad a low initial high-tech

density, a given percentage gain in high-tech grow th wou ld not provide

the same incremental stimu lus as in a m etro with a large initial high-tech

den sity. Based u pon this equation, we found that 65 percent of the total

outp ut grow th d ifferential between metros could be explained on the basis

of their relative growth in high-tech and the initial high-tech density (see

Figure 2).

High-tech ind ustries have a large d irect economic imp act on metro

economies, but the indirect and induced effects are critical to

un derstand ing their role in p romoting grow th. Because of the high-value

added production in high-tech industries, and the greater demand for

high-skilled labor, these indu stries comp ensate their em ployees well. For

example, extensive use of stock options in high-tech industries total

compensation m ix is a p owerful incentive sometimes enabling even

clerical personnel to become millionaires. The ind irect effect (or the

5

Figure 2

Metro Growth Explained b y High Tech

Actual v s. Pred icted (Cross-sectional)

High- tech activity

can exp lain

65 percen t of the

growth differential

between m etros in th

1990s.

The high -tech

sector

stimulates th e

non-high-tech

sectors of a metro

economy

...the multiplier

impact.


10/132


incremen tal stimulus to non high-tech indu stries) from h igh-technology

industries on metro economies is substantial. The induced effects stem

from nonhigh-tech firms and their emp loyees pu rchasing more local

goods and services du e to high sales by businesses and greater personal

income increases.

We tried several approaches and w eighting schemes to d evelop a

composite measure of high-tech sp atial concentration. A straightforward

app roach seemed to be the most ap prop riate. Combining the location

quotient with th e share of national high-tech ou tpu t in a multiplicative

fashion yields our composite measure of technology produ ction centers, or

as we term them , Tech-Poles. They ar e Tech-Poles in the sense of the

relative technology gravitational pu ll that they exert and are exhibited in

Figure 3.

The dominance of Silicon Valley (San Jose metro) as a high-tech indu stry

center is well documented. But our composite index of 23.7 is more than

three times the size of the second -ranked m etro, which is barely larger

than the third-place metro. As a Tech-Pole, the gravitational pull of the San

Jose metro area, hom e to H ewlett-Packard, Ap plied Materials, Sun

Microsystems, Intel, Cisco Systems, Oracle, and Silicon Grap hics, is

un paralleled. Silicon Valley created the p ersonal-compu ter ind ustry and is

developing another indu stry electronic networking that has the

potential to exceed the size of the PC.

Dallass position, second on the index, might cause some surprise. But

with a diversified high-tech base seven industries out of a possible 14

6

The relative

technology

gravitational pu ll

of metros is

measured in our

Tech-Poles.

As a Tech-Pole,

the gravitational

pu ll of the

San Jose

metro is

unparalleled.

Figure 3

Milken Institute Tech-Poles


11/132

are more concentrated than th e national average it deserves to be high

on th e list. The Dallas metro area includ es six of the nat ions 20 largest

telecommunication services companies. GTEs global headquarters is

based in Dallas, wh ile Nortel, Ericsson, Fujitsu, an d Alcatel have U.S.

head quar ters there. Austin is though t by many to be the center of Texaselectronic comp onents indu stry, but Dallas produ ces over 20 percent more

in terms of value of outpu t and emp loys 4,200 more p eople.

Despite th e loss of defense-related high-tech firms in the early 1990s, Los

Angeles ranks third on the index. The ran king is somew hat inflated,

perh aps, by the inclusion of non-high-tech p ortions of motion picture

prod uction and services. But even if the entire entertainment ind ustry

were exclud ed, Los Angeles wou ld rank seventh .

Boston places fourth, with an above-average concentration in 11 high-tech

indu stries. It is home to some of the leading u niversities and research

centers in th e nation. Although th e joint impact of defense dow nsizing and

heavy dependence on the lagging mainframe computer industry has been

sizable, it is enjoying renew ed vigor du e to its Internet-related p rowess.

Internet portal giant Lycos is headqua rtered in Waltham , while many other

startup s are emerging on Route 128.

Thanks to Boeing, aircraft is still a major ind ustry in Seattle. But softwares

rapid growth has made computer and data processing services a vital part

of the regions econom y. Seattle ran ks fifth on the ind ex. Microsoft, of

course, is at the n ucleus of the Seattle software cluster.

From here on dow n, the rankings are full of surp rises. On a comp ositeranking of high-tech services, Washington , DC, places first in the coun try,

and sixth overall in h igh technology. Virtually u nnoticed, Washington has

become a comm un ications hu b. Its Virginia su burb of Fairfax County is

home to America Online, UUNET, and PSINet. Over one-half of the

nations Internet traffic passes through local firms. Software and data

processing are major sectors of the local economy as well.

Albuquerqu es high r anking d erives from its success in attracting

electronic componen t manu facturers. Chicago is an importan t center of

commu nications equipm ent, courtesy of Motorola. The w indy city also has

an above-average concentration in d rugs an d in research and testing

services. New Yorks ninth-place rank is in part due to its major presencein telecommun ications serv ices. The city has very little high-tech

man ufacturing, which brings into question how mu ch silicon is really in

Silicon Alley.

Atlanta is the und isputed high-tech capital of the Southeast, with a

foothold in telecommu nication services, comp uter and da ta processing

services. Oaklands 14th position places it just behind Orange County,

California a fact that may surp rise those accustomed to thinking of it as


7

Dallas, thou gh

it might surp rise

some, is the

second most-

powerful

Tech-Pole.

L.A., Boston,Seattle,

Washington D.C.,

Albuquerque,

Chicago,

New York, and

Atlanta roun d

out the top 10

Tech-Pole list.


12/132

an aging seap ort and home to a struggling un derclass. San Diego, with it

highly d iversified high-tech economy, ranks 17th. Raleigh-Durham -Chapel

Hill, Denver, Austin, San Francisco, Hou ston, and Boise round out th e top

25.

Spatial concentration is importan t, but it d oes not gu arantee continuing

high-tech grow th. And here, with smaller clusters in the midst of rapidexpansion, the rankings a re qu ite different (see Figure 4). Albuquerqu e is

at the top, recording a comp oun d an nu al growth r ate of 12.7 percent in

high-tech output during the 1990s. This is almost entirely attributable to

Intels investments in electronic components and accessories production.

Pocatello and Boise are second and third , respectively. (Boise is hom e to

Hewlett-Packards laser-jet printer division as well as to Micron

Technologies.) Ceda r Rapids and Har risburg, rapid ly growing centers of

electronic comp onent m anu facturing, are fourth and fifth. Colum bus, GA,

ranks sixth d ue to comp uter an d data processing services. Merced, CA,

recorded strong grow th in p harm aceuticals, albeit on a very small high-

tech base. The Richland, WA, metro area w itnessed increases in

engineering and architectural services. Austin, a center of compu ter and

electronics produ ction, ranked 10th on the ind ex.

Eugene-Springfield, OR, has witnessed large percentage gains in

compu ters and office equipment, comm un ications equipm ent, and

compu ter and data processing services. Albany, GA, experienced solid

increases in pharmaceuticals as well as aircraft. Flint, MI, ranks high in the


8

Figure 4

Relative High -Tech Real O utp ut Grow th, 19901998

Smaller high-tech

clusters are in the

midst of rapid

expansion.

Albuquerque

recorded th e

fastest high-tech

growth in

the 1990s...

...followed b y

Pocatello, Boise,

Cedar Rapids,

Harrisburg,

Colum bu s, GA,

Merced, Richlan d,and Austin.


13/132

1990s du e to strong growth in compu ter and data processing services,

wh ile Portland, OR, is becoming a center of semicondu ctor chip an d wafer

prod uction. Hou ston was bu ilt on energy, but its futu re will have a high-

tech comp onent. Indeed , the city has the distinction of being th e largest

metro in the top 50. Compaq , the largest supp lier of PCs to the U.S.market, is based there.

Is high-tech p rodu ction becoming more spatially concentrated or more

dispersed? This is a complex issue and mu st be analyzed from several

perspectives. There is evidence that high-tech man ufacturing is becoming

less spatially concentrated, but that high-tech services seem to exhibit

strong agglomeration processes. Even thou gh m anu facturing seems to be

dispersing to p eriphery r egions, it remains h ighly geographically

concentrated and clearly shows th at agglomeration forces are exerting a

continuing influence.

Despite some evidence of rising spatial dispersion in high-tech

man ufacturing, it is remarkable how concentrated it rema ins.

Agglomeration forces have an amazing ability to sustain them selves, only

to be thw arted at some p oint by very h igh congestion-related costs. A

combination of high costs and either obsolete technology or firms is almost

required before high-tech activity begins to decline.

As we enter th e age of hum an capital, wh ere firms merely lease

know ledge-assets, firm s location d ecisions will be increasingly based

up on qu ality-of-life factors that are im portan t to attracting and retaining

this most vital econom ic asset. In h igh-tech services, strict bu siness-cost

measu res will be less importan t to growing an d su staining technologyclusters in metro economies. Locations that are attractive to kn owledge-

assets will play a vital role in determining the economic success of regions.

Section 4: High-Tech Industries andEconomic Risk

On balance, the benefits to the economy from technology far exceed the

less-noticed negative aspects of technology-driven economic development.

How ever, there are risks emanating from th e technology indu strys inherit

volatility, its growing importance in the overall economy, and the closer

relationship betw een it and the bu siness cycle of the U.S. economy. Willthese risks prove to be severe for metropolitan areas that hav e developed

high-tech clusters that are prone either to technology cycles or to

fluctuations eman ating from the broader economy, or is high tech imm un e

to the business cycle?

Because high-tech industries account for so large a share of national

outp ut tod ay, the economy is more vu lnerable to a high-tech contraction

than ever before. A synchronous shock spread across a number of related


9

High-tech

manu facturing is

becoming less

spatially concentrated

bu t h igh-tech service

may be b ecoming

more concentrated.

As we en ter the age o

hu man capital, firms

location d ecisions w iincreasingly be b ased

on qu ality-of-life

factors.

Is high-tech im mu ne

to the b usin ess cycle?


14/132

technology indu stries, such as comp uters and semicond uctors, in

combination w ith some other inausp icious developmen t could cause a

recession in the en tire economy.

Application of information technology likely will dampen business cyclesin the future. IT is improving businesses ability to manage inventories

and redu ce a key source of contractions in the broad economy. Chan ges in

inventory investment a re responsible for a very large prop ortion of the

change in total bu siness investment over the business cycle. Futu re

business cycles will be less severe, bu t not extinct.

We analyzed the behavior of high-tech industries over the business cycle,

app lying several quan titative ap proaches. We isolated the cyclical

component from the trend and irregular (random) component of each

high-tech ind ustry ou tpu t series, comp ared th e standardized cyclical

componen t of each ind ustry to the economy overall, and r an regressions to

determ ine their relative cyclicality. Lastly, we tested to see if the

relationship w ith the bu siness cycle is becoming stronger over time.

We determined that high-tech manu facturing ind ustries are among the

most volatile in the economy. A few high-tech ind ustries lagged the

cyclical movem ent in the overall economy, but m any m oved coincident

with it. Several high-tech indu stries that were not h ighly correlated w ith

the overall economy between 1965 and 1985 did show a relationship after

1985. Compu ters and office equipm ent has been cycling m uch m ore

closely w ith th e overall economy since 1985 (see Figure 5). The cyclical

movement in computers approximately matched that of the overall


10

Figure 5

Compu ters Align w ith the Business Cycle

Cyclical Component Divided by its Standard Deviation

Future b usin ess cycles

will b e less severe -bu t not extinct.

High-tech

manufacturing

indu stries are among

the m ost volatile in

the economy.


15/132

economy in the m ost recent business cycle. Ind ustries such as

commu nications equ ipmen t, electronic componen ts and accessories, and

compu ter and data processing services exhibit this pattern as w ell.

What d oes the closer co-movem ent between these indu stries and thebusiness cycle mean for high-tech clusters arou nd the country? In an

attemp t to add ress this question, we hypoth esized tha t a simulation of a

recession at some p oint in the futu re could show u s how ind ividu al high-

tech industr ies might beh ave. We chose to simu late a 1990-91 style

recession, w hich was mild by historical stand ard s w ith just a 1.0 percent-

peak-to-trough decline.

Based u pon our estimated relationships, high-tech manu facturing

indu stries wou ld experience a substantially more severe decline than the

overall economy and could even exacerbate the overall decline in th e

economy. For examp le, the estimated peak-to-trough decline in comp uters

and office equipm ent is 11.9 percent, closely followed by aerospace (Table

1). Output of electronic components also falls by more than the overall


11

Table 1

High -Tech Sen sitivity to Recession

Ranked by Peak to Trough

Compu ters have

been cycling m uch

more closely with the

overall economy sin c

1985, nearly m atchingit over the m ost recen

bu siness cycles.

Communications

equ ipm ent, electronic

comp onen ts, and

compu ter and data

processing services

exhib it this

pattern as well.

Percent

Percen t Decline,

Decline, Cycle

Peak to Relative

SIC Trough to Trend

1 357 Computer & Office Equipment -11.9 -21.2

2 372 Aircraft & Parts -10.7 -15.7

3 376 Guided Missiles, Space Vehicles, & Parts -10.7 -15.7

4 871 Engineering and Architectural Services -8.5 -4.7

5 381 Search & Navigation Equipment -3.2 -8.2

6 367 Electronic Components & Accessories -2.6 -14.5

7 382 Measuring & Controlling Devices -2.4 -8.6

8 366 Communications Equipment -1.9 -10.7

9 Total Economy -1.0 -4.9

10 481 Telephone Comunications Services 0.8 -3.7

11 781 Motion Pictures 1.3 -3.7

12 737 Computer & Data Processing Services 1.9 -7.5

13 283 Drugs 3.9 -1.0

14 384 Medical Equipment, Instruments, & Supplies 3.9 -0.5

15 873 Research & Testing Services 8.4 -3.6

Source: Milken Institute


16/132


economy. Drugs and med ical equipm ent are the only man ufacturing

indu stries that continue to grow du ring the recession. Engineering and

architectural services experience a peak-to-trough decline of 8.5 percent in

this simu lation. Outp ut continues to grow in the other four h igh-tech

service industries, but growth slows substantially.

The most severely imp acted metros have d ense concentrations of

compu ters and office equipm ent, aircraft, comm un ications equipment, and

electronic components. Metros with a heavy reliance on high-tech services

are generally not affected, but because of high-tech services growing

susceptibility to the business cycle, they will be more exposed in the

future.

This analysis suggests that m etros that have experienced a high d egree of

success in developing clusters of comp uter, semicondu ctor, and other

high-tech m anu facturing ind ustries should be more aw are of the potential

risks that a recession could pose to their local economies. Most economic

developm ent and governm ent officials have encouraged these ind ustries

to locate w ithin th eir borders, hoping th at this might h elp insolate them

from fu tu re bu siness cycles. While high-tech ind ustr ies will assist cities

long-term relative p erformance, they are u nlikely to shield them from

fluctuations in th e aggregate economy.

Section 5: Key Economic Development andBusiness Planning Implications

In light of the tremendou s oppor tunity for employment creation andincome gains, as well as the p otential risks of national and regional

economic dow ntu rns, a technology-based economic development strategy

mu st be carefully formulated.

Table 2 lists a set of variables that influence the development of regional

high-tech ind ustries. We divide those factors into three grou ps: pu blic

policy, comparative location benchm arking, and social infrastructure

developm ent. We rated each factor based up on its effectiveness in h elping

to establish a regional high-tech cluster in the different stages of regional

economic development. All factors in the table are interrelated; their

integral nature lends imp ortance to the role of local governmen t in thedevelopment process.

State and local governm ents, pu blic policies, and the interaction between

private an d pu blic sectors are cru cial for the genesis, expan sion, and

fortification p hases of high-tech d evelopm ent. Non etheless, du e to th e

unique characteristics of high-tech industries, governments role also is

limited. Overly active governm ent intervention and pu blic policy may be

counterproductive and even harmful to the long-term development of

high-tech indu stries.

12

In fu ture recessions,

metros with d ense

concentrations of

computers , aircraft,

communications

equip ment, and

electronic compon ents

will be the most

impacted.

A techn ology-based

economic

developm ent strategy

mu st be carefully

formulated.


17/132

Research centers and institutions are un dispu tedly the most imp ortant

factor in incubating high-tech ind ustr ies. A side effect of the technical

capab ility and scientific research activities of these institutions is the

training and education of the skilled labor that will be critical to the

expansion and reinforcement of regional high-tech indu stries. The federal

government had an unintended impact on the formation of high-tech

clusters arou nd the country th rough its location of research centers and

allocation of gran ts. Formation of p ublic-private ventur es that aim to

establish and maintain lead ing-edge regional research centers and

educational institutes is a critical long-term economic growth strategy.

Raleigh-Durh am-Chap el Hill, NC, rep resents one su ch success.

In the initial stages of high-tech m anu facturing d evelopm ent, with other

factors being equal, low-cost regions have a distinct advantage. The

dispersion of high-tech man ufacturing and processing from Tech-Pole

regions such as Boston has intensified. As technology application is

broadly ad apted , more standard ized forms of high-tech manu facturing can


13

Table 2

High-Tech Development Factors

State and local

governmen ts, pu blic

policies, and their

interaction b etween

the pu blic and

private sectors are

crucial for all stages

of high -techdevelopment.

Research cen ters

and in stitutions

are un dispu tedly

the most important

factor in incubating

high-tech

industries.

In cep tion G row th Fortification

Public Policy

Tax Incentives

Public Investment

Commercialization of Ideas

Comp arative Location Benchmark ing

Cost Factors

Research Institutions

Skilled or Educated Labor Force

Transportation Center Proximity to Supplies & Markets

Social Infrastructure Developmen ts

Attend ing Changing Needs

Re-education & Training Facilities

Establishing Trade Groups, & Affiliations

Housing, Zoning, & Quality of Life

Critical

Very Important Important


18/132

be mov ed to low-cost locations. Proximity to sup pliers and markets is

becoming less relevant tod ay as commu nication and shipping costs fall. If

initial low-cost regions cannot establish agglomeration as their location

costs rise, they can easily be superceded by other locations. So, while

importan t initially, low cost is not a sustainable comp arative adv antage inhigh-tech industries. Since the new high-tech economy is globally based

and hence has great mobility, a high-tech comp any can move from one

region or country to an other in a relatively short p eriod of time.

The San Francisco Bay Area is the most expensive location in the count ry,

yet the Bay Area remains the most concentrated Tech-Pole in the country.

One cannot rule out the possibility that the success of high-profile

metrop olitan ar eas is based at least partially on the high cost of doing

business in those regions. High costs can become the catalyst for the

existing firms in the region to pu rsue higher r evenues throu gh inventing

new concepts and prod ucts. High costs also force out marginal or older

prod ucts, an evolutionary step that is both necessary and essential for the

high tech cluster to remain at the leading ed ge of technology innovation.

Tax rebates and incentives can be a good tool in laying the found ation for

corporate placement and par ticularly can help sm aller entrepren eurs set

up basic operations. Government entities should be cautious in

distinguishing and recognizing the orientation of such p olicy, how ever.

Governmen ts function shou ld be, at m ost, to jum p-start the p rocess.

Providing a readily available labor pool is probably th e best investment

that state and local governments can m ake.

The process of establishing a high-tech economy is complex andmu ltifaceted. Its evolution is totally dynam ic and in m any aspects self-

guiding. Developing a regional culture th at is amenable to change and

growth and building a society that is open to new ideas are probably the

best strategies and principles governm ent can have to both attract and

expand high-tech indu stry.

Just as w e have changed our view about the contents of technology from a

relatively pu re form of prod ucts to a more complex combination of ideas,

creativity, and entrepren eurial activities, economic developm ent policy

should adjust to being about the bu ilding of cultural and social

environment as well as physical infrastructure. Establishing local public

and private trad e group s and affiliations is a sound policy in prom otingthe exchange of ideas, trade information, and pu blic awareness of the

developm ent. Attending to the need s of local firms and newcomers alike

will help the region in attracting the desired skilled labor.

The opportu nity is tremend ous for growth-oriented firms and bu siness to

tap into the n ext rising tech-cluster regions, wh ere the custom er base is

wid ening and th eir needs are customized . More impor tant, operating in

the high-tech region and amon g firms that gain the first experience and


14

Although initially

imp ortant, low cost is

not a su stainable

comparative

advantage in h igh-

tech ind ustries.

Governments

fun ction should

be to jump -start

the developmen t

process.

Growth-oriented

firms should tap into

the n ext rising tech-

cluster regions.


19/132


know ledge of the next wave of change is a sup reme comp etitive

advantage.

There is nothing more d ramatic than the imp acts of the high-tech service

boom u pon th e commercial and office construction markets. Mushroom ingInternet Web-service firms and grap hic design stud ios change the concept

of space planning and building. A growing high-tech region rapidly

transforms the local comm un ity. New tech clustering and rapid

accum ulation of wealth in a regional market creates business opportu nity

ranging from home building to retail. Professional service providers

should be equally vigilant about the high-tech frontier developm ent. These

new and relatively smaller enterpr ises tend to u tilize their limited and

precious capital to create new id eas and prod ucts.

Conclusions

The app lication of IT is improving p rodu ctivity growth and boosting the

long-term growth path of the U.S. economy. Advances in technology have

created entirely new growth indu stries, such as e-comm erce, and enabled

truly globally integrated firms. Although r emn ants of the old economy

remain, the N ew Economy is here. Core information technology indu stries

electronic comp onents an d accessories, comp uter and office equ ipmen t,

and compu ter and data processing services are the fastest-growing in

the U.S economy.

High-tech industries are critical in gauging the health of the U.S. economy.

They are determ ining wh ich m etropolitan areas are succeeding or failing.

Without growth in high-tech sectors, metros will be left behind. In ord er tofoster high-tech growth, metropolitan areas must understand what

location factors are most important to high-tech firms. That high-tech

clusters are perceived as imp ortant is clear from th e worldw ide attemp ts

to repeat the success of Silicon Valley. Cloning Silicon Valley will be

impossible, however, because the proper DN A sequence is locked away

somew here on Sand Hill Road. Those regions that come closest to

dup licating Silicon Valley, how ever, will be the lead ing technology centers

in the early stages of the 21st centu ry.

The high-tech economy d oes pose risks for m etropolitan ar eas.

Technology-driven economic development is causing a wid ening ofincome disparity along ed ucational attainm ent levels, redu cing job

security and job tenure, and resu lting in greater risk of unemp loyment

among w orkers in their 50s. Furth er, high-tech man ufacturing indu stries

are becoming more sensitive to the business cycle. Nevertheless, it is

impera tive for local governmen t an d economic development officials to

prom ote high-tech expansion, or they risk substand ard economic growth

in the future. Although high tech is not the only developmen t strategy to

pu rsue, it will be the key d istingu ishing feature of metropolitan vitality as

we enter th e new century.

15

New tech

clustering and rapid

accum ulation of

wealth creates

business

opportunity

ranging from hom e

bu ilding to retail.

Cloning Silicon

Valley will be

impossible because

the proper DNA

sequ ence is locked

away somewh ere on

Sand H ill Road.

It is imp erative for

local governmen t

and economicdevelopm ent official

to prom ote high -tech

expansion.


20/132


21/132

Introduction

For some y ears observers have assum ed th at it is technologicalinnovation that is pow ering the Great American Prosperity Machine and

sustaining Americas p reeminence in ind ustries ranging from

ph armaceuticals to entertainment to software. Only very r ecently have

economists been able to piece together evidence to qu antify that

technology really is translating into higher p rodu ctivity nationwide. And

only now can w e see how it is driving success (and failure) in local

economies.

Technology and high-technology (high-tech) industries are central to

un derstand ing the economic well-being of n ations, regions, firms, an dindividua ls as we enter th e 21st century. Many terms are used to d escribe

technologys impact on the economy an d future. Some r efer to the

app lication of technology as the N ew Economy ; others believe that w e

are entering the age of knowledge-based capital, w hether p hysical or

hu man . The term high-tech is probably oversubscribed, but we believe

that the U.S. economy is curr ently in the m idst of the most dr ama tic

technology-driven d evelopm ent in its history.

This studys main focus is to examine technologys importance in

determ ining the relative economic performance of metropolitan areas and

wh at lessons can be app lied to the futu re. To d o so, we first establishtechnologys role in p romoting economic growth in the aggregate U.S.

economy. We also examine the sensitivity of high-tech ind ustries to futur e

business cycles and the likely impact on metropolitan economies. Lastly,

we offer some implications and suggestions for economic development

officials and business firms. The study is divided into five sections. An

app end ix describes each of the industries that are includ ed in ou r

definition of high technology and discusses developm ents in them.

In Section 1, Technology and U.S. Growth Potential, we establish what

role the ap plication of technology is playing in prom oting long-term

economic growth. Theory and evidence of how information technology

aids m icro an d macro p erformance of the U.S. economy is examined . Inwh at ways should investment in technology boost long-term economic

growth ? Are information technology and related know ledge-based

services improving produ ctivity growth in the U.S. economy? Is there

evidence to sup port the p roponen ts of a New Economy?

Section 2, National High-Tech Industry Performance, catalogs many

measu res of technologys grow ing assimilation throu ghou t the economy.

Further, it defines and reviews the indu stries in our stud y. How d epend ent

17

Now we can see how

techn ology is driving

success (and failu re)in local economies.

This studys main

focus is to examin e

technologys

imp ortance in

determining the

relative economic

performan ce of

metropolitan areas.


22/132


is the U.S. economy on information technology in its investment patterns?

How have technology workers wages p erformed r elative to those of non-

technology w orkers is the gap grow ing larger? How large is high tech

in the United States and h ow importan t is it in determining overall

growth ? Which high-tech ind ustries are experiencing the m ost growth?

Section 3 is the focal poin t of our stu dy, entitled, Technology and

Metropolitan Economic Performance. It examines many spatial

dimensions of high-tech economic activity and highlights the theoretical

un derp innings of economic geography an d, more p articularly, the

agglomeration processes as they pertain to technology indu stries. How

importan t is geographic clustering of high-tech ind ustries in determ ining

the relative economic growth of metropolitan areas (metros) in the 1990s

and can this be measured quan titatively? Which metros lead in technology

indu stry vertical density and horizontal depth , and are true technology-

prod uction centers? What factors determ ine where h igh tech is

concentrated and wh ere the greatest growth is occurring? How can high-

tech ind ustries promote grow th in local non-high-tech sectors? Are

agglomeration (centripetal) forces causing high tech to be more

concentrated or are countervailing centrifugal forces causing a dispersion

of high-tech industries?

Section 4, H igh-Tech Ind ustr ies and Econom ic Risks, analyzes risks

emanating from high tech as they relate to the industrys inherit volatility,

its growing imp ortance in the overall economy, an d the closer relationship

betw een it and the bu siness cycle of the U.S. economy. Will the

assimilation of technology into virtually all economic sectors leave high-

tech industries more exposed to fluctuations in the broad economy thanever before? Have high-tech ind ustries become more sensitive to the

business cycle over time? What m ight be the imp act for m etros with a

large concentration of high-tech industries?

Lastly, in Section 5, Key Economic Developm ent and Business Plan ning

Implications, w e highlight economic development issues as they p ertain

to high-technology ind ustries. Furth er, we translate the find ings of the

stud y into implications for established firms an d n ew bu siness formations

in terms of expansion p lans or location analysis. What typ e of economic

environmen t helps create and foster superior d evelopm ent? Can economic

policy indu ce high-tech ind ustry formation and augm ent clustering in a

region? If they can be d up licated, what kind of p ublic policy and economicdevelopm ent strategies can exped ite and fortify developm ent in the

future? What non-high-tech businesses can benefit most by monitoring

new ly emerging technology centers?

The App endix, High-Tech Indu stry Profiles, contains a review of the

indu stry structure, changes in the indu stry, measures of indu stry size and

performan ce, and issues relating to future d evelopments of high-tech

industries as classified in this study.

18

Some of the qu estions

we ask are: How

important is the

geographic clustering

of high-tech

indu stries to the

relative economicgrowth of

metropolitan areas?

What type of

economic

environment

help s create and

foster superior

development?


23/132

SECTION 1

Technology and U.S.Growth Potential

High-technology ind ustries are critical in gauging the p erformance ofthe U.S. economy. They comp rise an ever-increasing p roportion of

economic outp ut and therefore are more imp ortant for monitoring

business-cycle developm ents. A compelling case also can be mad e that the

high-tech sector is boosting the long-term p otential growth path of the U.S.

economy. Technological advancement em bodied in both new and moreefficient trad itional capital goods, and innovative imp lementation of them

in the bu siness sector, are pr ime d eterminants of economic growth (Jarboe

and Atkinson 1998).

Due to acceleration in technological adv ances and innovation in

computers, communications equipment, other high-tech products and the

associated know ledge-based services, demand in these sectors du ring the

1990s has been rising at an increasing p ace. As stated by Joel Mokyr, an

economic historian a t N orthw estern University consulted by BusinessWeek,

Weve never had a period in w hich innovation has so perm eated our lives

as in the 1990s.

Informa tion technology is ubiquitous, d irectly or indirectly invad ing

near ly all sectors of the U.S. economy. Information technology is

transforming th e mix of economic activity aw ay from trad itional sources,

such as consumer d urables and business investment in stru ctures, wh ile at

the same time au gmen ting the level of aggregate outp ut. Informa tion

technology (IT) is the term used to d escribe the infrastructure an d

know ledge that is necessary to make information available rap idly. IT

increasingly comp rises the software and commu nication services that

patch equipm ent together. Advances in IT in recent years have created

entirely new growth indu stries including e-comm erce, online information

services, mobile commun ications, and r apid adv ances in med ical research.IT has spu rred rapid , seemingly continuous innovation by creating

networks th at generate value throu gh p rodu ctive interactive relationships

or collaboration. In the past, innovation transpired pred ominately by more

discrete ad vances in research and developm ent. We are still witnessing the

early stages of the imp act of IT and other new ly emerging technologies on

the potential growth of the U.S. economy.

19

Technological

advancement

embodied in b oth

new and m ore

efficient traditional

cap ital goods, and

innovative

implementation ofthem in the business

sector, are p rime

determinants of

econom ic growth .

IT has spurred rapid,

seemingly continuou

innovation b y

creating network s tha

generate value

through produ ctive

interactiverelationships.


24/132

A substantial portion of business costs involves the collection, exchan ge,

and use of information (Williamson 1990). Information exchange is at the

core of an efficient m arket-based economy. Markets function by

interactions amon g p rodu cers, supp liers, vendors, financial intermediaries,

brokers, trading p artners, and , ultimately, consumers. In a globalizedeconomy, informa tion is the m ost valuable resource. Firms h ave been

investing in information technology in an effort to boost prod uction

efficiencies, improve communication flows, and enhance overall business

operations. Firms believe that th ese investmen ts yield rewards, such as

higher p rofitability, by red ucing the grow th of u nit labor costs throu gh

greater ad vances in prod uctivity. Even Federal Reserve Chairman Alan

Greenspan, w ho is notorious for obfuscation, has noted the imp ortance of

IT in testimony before Congress, Our nation has been experiencing a

higher growth rate of prod uctivity outp ut p er hour w orked in recent

years. The dramatic improvement in computing power and

commu nications and information technology app ear to have been m ajor

forces behind this trend. A fund amental question is: How mu ch of the

increase in produ ctivity growth is sustainable?

While the recent positive economic evidence on th e imp act of investment

in high technology may be inconclusive in a statistical sense, there can be

no disagreement on h ow strong grow th has been in recent years. The

Milken Institute has constructed an end -use deman d m easure of high-

technology p rodu cts and services in an a ttemp t to calibrate their rising

contribution to economic growth. Figure 1.1 displays the surge in d eman d

for the high-tech sector during the 1990s. This measure includes business


20

Firms have been

investing in IT in an

effort to boost

production

efficiencies, imp rove

communication flows,and en han ce overall

bu siness operations.

Sin ce the 199091

recession, growth

in th e high -tech

sector has been

four times as large

as growth in the

aggregate economy.

Figure 1.1

High Tech Surges in the 1990s

High Tech vs. Total Real GDP


25/132

and consumer purchases of comp uters and related equipment,

commu nication equipm ent, scientific and engineering equipm ent, dru gs,

telecommunication services, pre-packaged software, cable TV, and other

related sectors. Since the 1990-91 recession, growth in the high -tech sector

has been four times as large as grow th in the aggregate economy. Dur ingthe 1980s, the high-tech sector grew approximately twice as fast as the

economy. Over the past three years, growth in high-tech p rodu cts and

services averaged over 20 percent, directly elevating real GDP growth by

1.5 percentage points.

Technology and Growth Theory

The contribution of technology to economic growth has been d ebated

since the time of the classical econom ists. Today, grow th th eorists

generally agree that technological change is a m ajor d eterminant of long-

run economic growth . However, they d isagree substantially on thecontribution of technology relative to other factor inpu ts and the am ount

of technological change that is due to ph ysical capital as opposed to

hu man capital. Technological advan ces improve the d egree to wh ich

inpu ts of ph ysical and h um an capital are translated into outp ut grow th. To

the degree that output grows more rapidly than the weighted-factor

inpu ts, prod uctivity growth is enhan ced. Productivity growth (the rate of

growth in outpu t per u nit of input, usually expressed in output p er man-

hour ) is a key determ inant of the long-run aggregate sup ply in the U.S.

economy. Our stand ard of living is directly tied to prod uctivity grow th.

The theory of an aggregate or economywid e prod uction fun ction has beenused for several decades to evaluate the p rodu ctivity of various factor

inputs su ch as cap ital, labor, and R&D expend itures. Basically, the

prod uction function p ostulates that differing combinations of inp uts can

be utilized to p rodu ce a given level of outp ut. Many h ave contributed to

the literature in th is area, but Robert Solow h as long been a major figure in

the ad vancement of growth theory, winning a N obel Prize for his efforts.

His pioneering work in the d evelopm ent of the neoclassical growth m odel

was the foundation for modern growth theory (Solow 1957). Solows

theoretical framework, w hich decomposed contributions to outp ut from

capital and labor on the basis of a constant-return s-to-scale p rodu ction

function, helped establish a temp orary consensus in th e 1970s on grow th

theory. The neoclassical model allowed the substitution of capital for labor.Solow found that a small fraction of economic growth could be assigned to

labor, and that capital formation accounted for approximately one-third of

growth. This leaves a large residual that is assigned to technological

progress. In Solows mod el, technological progress w as exogenou sly

(outside the system) determ ined, du bbed the Solow residu al.

Assigning such a large p ortion of outp ut grow th to exogenously

determ ined technological progress w as a trou bling concept for


21

Stand ard of living is

directly tied to

productivity growth.

The production

function postulates

that differing

comb inations of

inpu ts can be u sed to

prod uce a given level

of outpu t.


26/132


theoreticians, leading m any oth ers to attemp t to d evelop alternatives to

Solows mod el. In the late 1980s, a diverse set of theoretical and em pirical

work began to em erge as endogenous, or new, growth th eory. This body of

work differentiates itself from n eoclassical growth by em phasizing that

economic growth is an endogenou s outcome of a dynam ic economicsystem, not the w ork of some m ystical force emanating from outside.

Endogenous grow th theory postulates several chann els throu gh w hich

technology, hum an capital, and the creation of new ideas enable a virtuous

circle and feedback to economic grow th (Romer 1990 and Barro 1991). This

is critical in attempting to determine the contribution of information

technology to the growth process and, sp ecifically, how m uch it m ight

improve productivity growth. New growth theorists postulate that

know ledge has a separate and distinct impact on promoting technological

adv ance. Technology innovation, stemm ing from improved kn owledge, is

the key to prod uctivity improvements in their view.

What d oes this imp ly for the role that ad vances in information technology

and its diffusion play in long-term economic growth and prod uctivity?

App lying th e produ ction function, the elasticity of outpu t w ith respect to

IT investment shou ld be h igher than non-IT forms of investment. In other

word s, a buildup in the stock of information technology, both physical and

human, should add proportionally more to enhancing long-term economic

growth than an equal gain in the stock of other factor inpu ts. At the firm

level, as businesses substitute IT for other inputs, they will lower

prod uction costs for a given level of outpu t or increase prod uction wh ile

holding overall costs constant. As firms throu ghout the economy em ploy

IT more effectively over time, the aggregate potential growth path of the

U.S. is boosted.

Historical Evidence of Technology and Growth

Despite the intellectual ap peal of new growth theory, the historical

evidence over the past several decades is mixed. The central inconsistency

is that while there has been a massive investment in comp uters and other

information technology since the late 1970s, the rate of productivity

growth in the U.S. economy has slowed.

Labor produ ctivity (outp ut p er man -hour) growth averaged over 3.0

percent ann ually du ring the p ostwar p eriod throu gh 1973, but fell to asluggish 1.0 percent thereafter (see Figure 1.2). When measured on the

basis of multifactor produ ctivity (wh ich m easures outp ut grow th relative

to all factor inpu ts), which fell from arou nd 2.0 percent before 1973 to 0.3

percent ann ually through 1994, the produ ctivity record is even worse. This

led Solow himself to quip th at you can see the compu ter age everywhere

but in the p rodu ctivity statistics. Some economists have d ubbed this

inconsistency the p rodu ctivity paradox.

22

New growth

theorists hypothesize

that kn owledge

has a separate and

distinct imp act on

promoting

technologicaladvance.

Solow q uipped

that you can

see the

comp uter age

everywhere bu t

in the prod uctivity

statistics.


27/132

IT Pessimists

One of the main argu men ts from th e pessimists on information

technologys imp act on p rodu ctivity growth stems from the relative size of

IT compared to other factor inpu ts. Many contend tha t IT investmentsimp ly could n ot have an ap preciable impact on prod uctivity growth

because it is too small. Daniel Sichel, an economist at the Federal Reserve

Board , argues in a recent book that compu ters account for only about 2

percent of the United States total capital stock (Sichel 1997). He su ggests

that even if the return on compu ters is above that on other investments,

their overall contribution to economic growth will be mod est.

Furtherm ore, he argues that the high level of corporate sp ending on

compu ters exaggerates their importan ce because of the rapid obsolescence

of the investment. Consequently, the net investmen t of new comp uters is

relatively small.

Other IT pessimists argue tha t firms h ave reacted as they shou ld to

declining p rices of comp uters, nam ely by substituting them for other

relatively more-expensive inp uts su ch as labor or other typ es of capital.

Thus, the comp uter revolution is mainly a story of substitution rather than

prod uctivity grow th. This suggests that the frenetic pace of IT investment

is merely wheel spinning altering prod uction inputs w ithout increasing

outp ut r ather than creating real ad vances in p rodu ctivity. Some believe

that firms are investing in IT on blind faith. They argue th at produ ctivity


23

Many contend

that IT investmen t

simp ly could

not have

an app reciable

imp act onprod uctivity growth

because it

is too small.

Some believe that

firms are

investing in IT on

blind faith.

Figure 1.2

The Productivity Paradox

Outp ut per Man-Hou r (Nonfarm, Private)


28/132

gains were r ealized from the investmen ts mad e in the 1960s on large,

transaction-heavy m ainframe comp uters because a coherent, w ell-executed

plan was im plemented . By contrast, investment in todays decentralized

personal compu ter world h as not been thought ou t as well. The compu ter

systems that have been pu rchased w ere not subjected to the p roperinvestment ap pra isal process. It is even suggested by some th at these

fund s could h ave been better invested in research and d evelopm ent,

emp loyee training, and m anagerial advan ces. They point to stud ies

displaying no correlation between spend ing on information technology

and profitability by indu stry.

IT Optimists

The IT optimists offer counterarguments. The first one is the

measu remen t-error p roblem. Measuring outp ut is problematic, especially

in the service sector, which is not on ly a grow ing segment of the economybut an area w here IT should contribute significantly to prod uctivity

growth . Baily and Gordon (1988) attribute much of the d ram atic

slowd own in prod uctivity grow th in th e United States after 1973 to

improper measurement of quality-adjusted outp ut. They note that

prod uctivity in comp uter m anu facturing has increased dramatically in the

official data, bu t prod uctivity resulting from comp uter u se is not app arent.

The use of a hed onic price index (which measu res compu ter-processing

pow er in the compu ter indu stry) is a key reason for this. The hedonic

index ad justs the price to reflect the increase in the quality of comp uters

by focusing on the cost of performing a calculation in a constant time

period. They conclud e that IT is providing valuable customer services thatare not reflected in the official outp ut data. An increasing nu mber of

consumer services offers improved convenience as a result of IT

investment, but it is unlikely to be measu red accurately.

The best example of improved convenience is seen in the financial services

indu stry. Informa tion technology allows firms to p rovide sop histicated

cash management accounts, almost costless portfolio diversification

throu gh no-load mutu al fun ds, 24-hour m oney access machines, banking

by ph one, and, now , Internet ban king. Nevertheless, in th e banking, credit

agencies, and securities indu stries, the Bureau of Economic Analysis (BEA)

uses labor inpu t to extrapolate real outpu t changes. The only w ay the BEA

could report an increase in overall productivity in the financial servicesindu stry is throu gh a change in the comp osition of indu stry outpu t. When

other measu res of outp ut are u tilized, such as total trades per em ployee in

the stock brokerage ind ustry, or nu mber of checks processed per em ployee

in the banking sector, strong p rodu ctivity growth is witnessed.

The second major argum ent is the technology d iffusion argum ent. This

says that a n ew technological innovation such as IT may require a

substantial time p eriod before being absorbed throu ghout the economy. In


24

IT optimists coun ter

that measurement

error may be hid ing

ITs contributions.

When m easures other

than the BEAs

stand ards are used,

strong prod uctivity

growth often is

witnessed.


29/132


this view, IT has n ot yet been exploited fu lly. The optim ists say that there

are three p hases of technology imp rovement: invention, innovation, and

diffusion. The invention of a technology may occur far in ad vance of

developing a w ay to incorporate it into a new p rodu ction process as an

innovation. Even if this results in rap id p rodu ctivity improvem ents for the

firm that d eveloped the innovation, it may take years before other firms

observe and copy it. Optimists point to the innovation of the dynam o, the

fund amen tal electromechanical conversion device, which w as not

immed iately ap parent in th e p rodu ctivity statistics, either. Several decades

passed before indu stries began to app ly this technology. The slow

diffusion of the dynamo (electrification) delayed any significant impact on

productivity growth until the 1920s. Optimists believe that in the case of

IT, a combination of formal edu cation an d app lications-oriented h um an

capital will lead to the successful transformation of the IT revolution into

sustained p rodu ctivity gains. As more firms emp loy IT to innovate an d a re

copied, aggregate produ ctivity gains w ill become m ore app arent.

If the optimists are correct and diffusion is the key to realizing

prod uctivity grow th from IT investment, there should be some ind ication

of this in th e sectors of the economy tha t d eployed an innovation first. The

most encouraging evidence of IT boosting prod uctivity grow th is found in

du rable manu facturing. Output p er man-hour in the d urable

man ufacturing sector has grow n at an average ann ual rate of 5.0 percent

since 1990 (see Figure 1.3). During the 1980s, that sectors productivity

grew at an average rate of 3.2 percent. Outpu t growth in the

25

IT optimists observe

three ph ases of

technology

improvement

invention, innovation

and d iffusion - noneof wh ich m ay occur

simultaneously.

In the durable

man ufacturing sector

outpu t per man-hour

has grown at an

average ann ual rate o

5.0 percent since 1990

Figure 1.3

Productivity Increases Du rable G oods

Outp ut p er Hou r of all Persons


30/132


man ufacturing sector is easier to define and accurately measure than

growth in other sectors. The price deflators for most m anu facturing

indu stries better reflect the quality imp rovemen ts in the finished p rodu ct,

increasing the r eported real or inflation-adjusted ou tpu t. This evidence

sup ports the op timists contention that m easurement error is a problembecause much of IT investment has been concentrated in service sectors,

wh ere new forms of outpu t attributable to IT innovations are d ifficult to

quantify.

Another w ay to examine if IT investment is having a prod uctivity payoff is

to analyze micro evidence. Several recent studies at the firm level have

docum ented gains in p rodu ctivity from IT investmen t. Brynjolfsson and

Hitt (1996) performed one of the most compelling studies; they found

strong eviden ce of firm -specific retu rns to IT investment. They collected

data on firm investments in IT hardw are and IT labor and combined them

to create an overall measure of IT Stock. They then app lied a p rodu ction

function ap proach by relating th ree factor inpu ts (IT Stock, Non-Comp uter

Capital, and Labor) to firm Value Ad ded . Their results indicate that the

outp ut elasticity with respect to IT capital stock w as very h igh and that

prod uctivity w as much h igher in firms w ith larger IT capital stocks. They

estimated that the gross rate of return on IT investment was n ear 50

percent, comp ared to 15 to 20 percent for other investments. They

concluded that the prod uctivity parad ox at the firm level has disappeared

since 1991.

Other organization-level evidence su pp orts their find ings. Lichtenberg

(1995) analyzed data for hundreds of companies for the period 1988-92.

His results displayed th at the organ ization-level rate of return oninformation technology investment w as more than three times the rate of

return on investment in other non -IT equipment or stru ctures. Further, his

results ind icated tha t information systems emp loyment made a larger

contribution to outpu t growth th an any other typ e of emp loyment. There

is a substantial body of firm-level literature that strongly su pp orts higher

rates of return and prod uctivity from information technology investment.

This lends su pp ort to th e IT optimists position th at there is an

un derrep orting of outp ut grow th, causing a dow nw ard bias in the official

aggregate prod uctivity statistics.

Many economists are encouraged by the exemplary performance of the

U.S. economy since th e 1990-91 recession. By m any measures, the 1990shave exhibited u nanticipated prosp erity. The dynam ic natu re of IT

spend ing and its rising imp ortance in total investment in the economy

have caused some to herald the daw n of a New Economy, which wou ld

foster economic growth at a r ate closer to that wh ich p revailed d uring th e

1950s and 1960s. New Economy proponents believe that the U.S. economy

is able to sustain faster growth , with labor and capital more heavily

utilized, than was previously believed p ossible without tr iggering h igher

inflation. In essence, the sustainable, noninflationary growth potential of

the U.S. economy is higher.

26

Several recent stud ies

at the firm level have

docum ented gains in

prod uctivity from IT

investment.

New Econom y

propon ents believe

that the U.S. economycan su stain faster

growth , with labor

and capital more

heavily utilized, than

was previously

believed possible

without triggering

inflation.


31/132

A central piece of their sup porting evidence is the benign behavior of

inflation in the face of tight labor markets during the 1990s expansion.

Most economists believed that the nonaccelerating inflation rate of

unemployment (NAIRU), or the rate of unemployment that would cause

inflation neither to accelerate nor d ecelerate, was betw een 6 p ercent and

6.5 percent. With the u nemp loyment rate below 5.5 percent for over twoyears, core inflation has been benign (see Figure 1.4). Temp orary

ausp icious developmen ts such as plunging oil and commod ity prices and

mod erating health care costs have helped restrain inflationary p ressures.

Neverth eless, most economists acknowledge th at the N AIRU is lower than

previously anticipated. New Economy proponents contend that innovative

implemen tation of informa tion technology is playing a large role in this

improved inflation p erformance.

Another key piece of corroborating eviden ce from New Economy

prop onents is the improved prod uctivity performance over the past three

years. Outp ut p er man -hour in the nonfarm bu siness sector rose by an

average of 2.1 percent from 1996 through 1998. What is encouraging aboutthis developmen t is that it is occurring at a m ature stage of an expansion,

typically the time w hen p rodu ctivity growth begins to wane, hinting at a

possible long-term prod uctivity imp rovemen t. Drawing conclusions on the

basis of only three years of data is precarious; nevertheless, these

observations suggest that something real might be hap pening. Some N ew

Economy prop onents believe that the long-term real GDP growth ra te has

risen to 3 percent, based u pon 2 percent prod uctivity growth and labor

force grow th of 1 percent.


27

With the

un employment rate

below 5.5 percent for

over tw o years, core

inflation h as beenbenign.

Outpu t per man-hour

in the non farm

bu siness sector rose

by an average of

2.1 percent d ur ing

1996-98.

Figure 1.4

NAIRU Falling

Unemp loyment Rate and CPI


32/132

What to Conclude?

What shou ld be conclud ed concerning the role of technology in prom oting

long-term economic growth? On the basis of macroeconomic evidence as

officially reported by governm ent agencies, it is difficult to statisticallyprove th at the massive investment in information technology has caused a

rise in long-term economic growth in the Un ited States. This does n ot

mean , however, that IT investment is not improving economic

performan ce. Real outp ut grow th is und erstated in mu ch of the service

sector, partly because government statistical agencies have not been

adequately fund ed to d evelop ap propriate ways to measure an

information-age economy. Nontrad itional ways of measur ing outpu t in

service indu stries show th at these indu stries are d isplaying more rapid

growth and , therefore, greater advan ces in prod uctivity. If this

un derrep orting of outp ut grow th were aggregated across all indu stries, the

economywid e performan ce of the United States would look much better.

Another w ay to gau ge if some structura l change is occurr ing in the

economy is throu gh the estima tion of a potential GDP equation using

prod uction fun ction methodology. If a stand ard p rodu ction function

app roach, which does not differentiate between IT capital and nonIT

capital, severely und erstates the curren t level of economic activity, it

wou ld sup port the notion that there has been some structura l break in

long-term economic growth. We estimated a potential GDP equation for

the U.S. economy using a Cobb-Douglas production function. The size of


28

Real outpu t growth is

un derstated in mu ch

of the service sector,

partly because

governmen t agencies

have not been

adequ ately fun ded todevelop approp riate

ways to measure an

information-age

economy.

Figure 1.5

Actual GD P Above Potential

GDP Gap, Actual vs. Potential


33/132

the labor force is d etermined by u se of 10-year p opu lation cohorts and

labor force par ticipation rates by age cohort and gend er. The capital stock

is determined by ap plying dep recation rates by type to estima te the net

add ition to the capital stock from cur rent investment. The rate of

technological change w as introdu ced as a time trend . This equationexplains the long-term p ath of real GDP very accurately, but actual GDP is

currently above its estimated potential by the largest am oun t (5.7 percent)

in the sample period (see Figure 1.5). It is possible that this could occur

withou t an actual structural break in the relationship; however, it wou ld

be accompanied by accelerating inflation, certainly not decelerating

inflation. IT investment may plausibly account for some of this

discrepancy.

Some N ew Economy prop onents are overzealous in ad vocating their

position. It is prem ature to conclude th at sustainable prod uctivity grow th

has imp roved from 1 percent of the past 25 years to 2 percent on the basis

of three years of sup porting d ata. Nevertheless, we mu st ask whether

something is happ ening in the economy du e to app lication of information

technology. We consider ourselves to be cautious optimists. Modern

technology is altering prod uction processes through out the economy and

enabling truly globally integrated firms. With the expon ential growth of

the Internet, many efficiency gains and lower costs will be realized.

Technological innovation has boosted productivity growth to the 1.5 to 2.0

percent range and may p erhaps enable more improvement in the future.

This will increase long-term economic growth to between 2.5 and 3.0

percent in our view.


29

Actual GD P is

curren tly above its

estimated potential b

the largest amoun t

(5.7 percen t) in the

sample period.

Technological

innovation h as

boosted productivity

growth to the 1.5 to2.0 percent range and

may enable more

improvement in the

future.


34/132


35/132

SECTION 2

National High-TechIndustry Performance

The U.S. economy is a d ynam ic living and breathing organ ism, constantly

ad apting itself to changes in its environm ental conditions. Rapid ep ochs of

technological change have occurr ed in the p ast, the indu strial revolution

and electrification beginning in the 1990s to name two, but currently the

U.S. economy arguably is und ergoing the most d ramatic technology

developm ents in its history. New technology indu stries should be am ong

the top grow th performers as they are d iffused throu ghou t the economy.

History has d emonstrated this to be the case. High-tech indu stries,

how ever, are transforming the stru cture of the U.S. economy in an

un para lleled dimension. Whether or not high-technology investments are

substantially enh ancing long-run economic growth, high-technology

indu stries comp rise a mu ch larger share of the overall economy and , more

importan tly, account for a disprop ortionate share of economic growth in

the United States.

The Impact of Technology on Economic GrowthOne of the m ost omnip resent influences of high technology has been in

business investmen t patterns. Information technology hardw are has

accounted for a growing sh are of total business investment across a broad

array of indu stries. In 1970, information p rocessing equipm ent an d related

sectors represented 7 percent of real business equipment investment;

however, last year it was responsible for over 50 percent of all capital

spending. Most of this surge has occurred since 1980 as displayed in

Figure 2.1. From 1995 through 1998, growth in real business investm ent in

information-processing equipm ent av eraged 25 percent, d irectly

accounting for nearly 27 percent of total economic growth. This is a

narrow definition of information technology investment

The US High Tech Economy

Documents

Transcript of The US High Tech Economy