Quick viewing(Text Mode)

A Tale of Three Supply Shocks, National Inflation and the Region's

A Tale of Three Supply Shocks, National Inflation and the Region's

I SSUE 2MARCH/A PRIL 1997 thwe u s o tt ss

e c y onom

bank of dallas A TALE OF THREE SUPPLY SHOCKS, NATIONAL AND THE REGION’S

N RECENT YEARS,s several supply shocks —unusual shifts in costs —have kept U.S. inflation low by putting downward pressure on for certain , espe- cially computers, health care and, until 1996, energy.1 Because the sectors producing these and services are important to the Eleventh Federal Reserve District, these shocks Ihave had an impact on its economy. After examining these shocks’ effect on U.S. inflation, this article analyzes their impact on the Dis- trict and assesses the outlook for computers, energy and health care.

Supply Shocks and U.S. Inflation INSIDE U.S. inflation has remained low through early 1997, even though, Does Electronic since 1995, the has been below 5.75 to 6 percent, Mean the Death of ? a range below which inflation had previously tended to rise. There are three plausible explanations for this change in behavior. The Mexican Economy One is that job has held down .2 Another is that the Snaps Back competitive pricings environment of the 1990s has enabled the econ- Chart 1 As shown in the upper panel of Overall and Core Inflation Chart 2, medical inflation resumed Supply Shocks and the Region Diverge in 1996 … falling in late 1995 after leveling off

Year-over-year percent change over mid-1994 to early 1995. Indeed, Each of these supply shocks has 4 medical inflation has declined more affected Eleventh District employment

3.6 than overall inflation has in the mid- trends. With respect to computers, ad- Overall 3.2 1990s, reflecting the shift toward man- vances in technology have spawned

2.8 aged health care and the adoption of an increased for information

2.4 Core other cost- practices. equipment and, until late 1995, a con-

2 In the mid-1990s, the pace at which comitant rise in production and capac- 1993 1994 1995 1996 computer prices have fallen () ity. Indeed, as semiconductor orders … As Energy Swings has swung substantially. After slowing continued to exceed shipments, re- Drive Overall U.S. Inflation sharply in 1994, deflation in the “elec- flected in a domestic book-to-bill ratio 7 Year-over-year percent change tronic computers” category of the pro- above 1, high-tech manufacturing job 10 ducer price accelerated sharply growth was very strong in District states 8 (middle panel, Chart 2 ), partly reflect- (Texas, Louisiana and New Mexico), as 6 8 4 ing technological advances in computer shown in the upper panel of Chart 4. 2 chips as well as excess plant capacity However, as demand growth for 0 and inventory buildup prompted by computer equipment unexpectedly –2 –4 overly optimistic expectations. These slowed in 1996, computer chip plant 1993 1994 1995 1996 wholesale price movements influence SOURCE: Bureau of Labor Statistics. retail consumer prices in the CPI’s home furnishings category, which comprises Chart 2 furniture, computers, other electronic After Pausing in 1994, omy to operate at higher capacity levels goods and home appliances. Comparing Health Care without a pickup in inflation, or —put the middle and lower panels of Chart 2, Resumes … another way—has slowed the pace at one can see how wholesale computer Year-over-year percent change which excess demand pressures induce price deflation has influenced consumer 7.5 rises in inflation, as some of my home furnishings’ inflation in the mid- 6.5 research under way suggests.3 A third 1990s. Indeed, on a year-over-year basis, 5.5 explanation—the focus of this article— prices for home furnishings actually had 4.5 is that inflation has been low because dipped 0.1 percent as of December 3.5 several supply shocks have put down- 1996, while producer prices for elec- 2.5 ward pressure on prices for three key tronic computers had fallen 21 percent. 1993 1994 1995 1996 commodities: computers, energy and Excluding its home furnishings com- Computer Price Deflation health care. ponent, core inflation barely deceler- Accelerates … Between 1993 and early 1994, infla- ated in 1996. Year-over-year percent change tion fell, according to both the con- Thus, the pickup in overall infla- 0 sumer (CPI) and the core tion and the decline in core inflation –5 CPI, which excludes food and energy during 1996 largely reflect the two ex- –10 prices (top panel, Chart 1 ). The overall tremes of an acceleration of energy –15 CPI has drifted upward since early 1994, price inflation (overall) and an accel- –20 picking up from mid-1994 to mid-1995, eration of computer price deflation –25 slowing in late 1995 and then picking (core). This pattern suggests that iso- 1993 1994 1995 1996 up again in 1996. In contrast, core CPI lated price developments may be dis- NOTE: Prices based on the producer price index for the inflation was stable over 1994 –95 be- torting the inflation picture. One way to computer electronic category. fore slowing in 1996. A comparison of filter out the disproportionate influence … Driving Down the upper and lower panels of Chart 1 of unusual price factors is to measure Home Furnishings Prices suggests that most of the wiggles in the inflation by excluding the highest 10 Year-over-year percent change overall CPI reflect swings in consumer percent inflation components (by ex- 3 energy prices. On the surface, the slow- penditure weights) and the lowest 10 2 ing of core inflation last year seems percent inflation components.5 This puzzling in the face of rising energy “trimmed mean” CPI measure (Chart 3) 1 prices, which had tended to bolster shows an upward drift in inflation since 0 4 core inflation in the past. However, in- late 1995 — consistent with the view –1 novations in both health care delivery that the economy has been operating at 1993 1994 1995 1996 and computers have played an impor- levels of the unemployment rate pre- SOURCE: Bureau of Labor Statistics. tant role in holding down core inflation. viously associated with rising inflation.6

Page 2 Southwest Economy March/April 1997 Chart 3 and, within this sector, is less focused wane over 1997–98 if some analysts are Filtering Reveals a Modest on computer chip production. correct in predicting that microproces- Pickup in U.S. Inflation In contrast to computer prices, en- sors are nearing the end of the current

Year-over-year percent change ergy prices rose in 1996, reflecting a wave of innovation, as suggested by

3 sudden weather-related rise in demand the slower pace of microprocessor price Trimmed mean coupled with low inventories. This deflation since early 1996. Perhaps for 2.8 price rise spurred the oil industry to similar reasons, prices for integrated expand exploration and hiring.11 In circuits—another key computer com- 2.6 addition, new exploration technology ponent—have declined at a slower pace made it more profitable to search for since spring 1996. 2.4 oil under waters where an ocean- Nevertheless, computer price defla- 2.2 bottom salt layer had previously tion actually picked up to more typical obscured reserves— especially true of levels last year, as reflected in the GDP 2 1993 1994 1995 1996 the Gulf of Mexico. Both high energy chain-weight price index for computers

SOURCES: Bureau of Labor Statistics; author’s calculations. prices and a technological innovation and peripheral equipment favoring oil exploration in local waters (lower panel, Chart 5 ). This index ex- have boosted drilling in the gulf and cludes typewriters and noncomputer energy employment in the Eleventh equipment, whose share of GDP’s capacity outstripped demand. The tem- District’s states (middle panel, Chart 4 ). broader “information-processing and porary imbalance between production Increased worldwide demand for drill- related equipment” component of busi- and orders was reflected in a decline in ing equipment has also bolstered manu- the domestic book-to-bill ratio to levels facturing employment in areas such as below 1, generally indicating that do- Houston. Chart 4 mestic shipments exceeded orders.9 Finally, the restructuring of health District High-Tech and overexpansion of care has slowed job growth in that sec- Manufacturing Jobs Reflect Semiconductor capacity led to a buildup of inventories tor, despite the continued increasing Industry’s Fortunes and an unexpected drop in memory demand for health care associated with chip prices, which in turn slowed job a general aging of the population. Year-over-year percent change Book-to-bill ratio growth in the overall high-tech manu- Health care restructuring likely con- 8 1.2 facturing industry in District states. tributed to a recent slowing of the pace 6 1.1 This deceleration in high-tech job at which the share of private health care 4 1 High-tech Book-to-bill growth helps explain why nonfarm job employees in the District’s three states 2 job growth ratio .9 growth in Texas slowed toward the U.S. has risen (lower panel, Chart 4 ).12 0 .8 –2 .7 average in 1996, after several years in 1993 1994 1995 1996 which it substantially exceeded the Oil Drilling and District national average. As demonstrated by The Outlook for Computers, Energy Jobs Recover two Dallas Fed researchers, the high- tech sector—which includes high-tech Energy and Health Care Year-over-year percent change Number 2 320 manufacturing along with communica- District energy job growth 300 tions and computer-related services— Computers. Some analysts attribute 0 280 was a major contributor to Texas job last year’s speedup in computer price –2 growth over 1988 –94.10 Roughly half of deflation to two factors that reduced Texas 260 –4 rig count the broadly defined high-tech jobs in prices for semiconductors (memory and 240 –6 220 Texas are in the Dallas–Fort Worth metro- microprocessor chips), which are im- 1993 1994 1995 1996 plex, where there is a high concentra- portant computer components. First, Restructuring Slows Growth tion of telecommunications firms, while prices for memory chips (such as Of District Health Care Jobs roughly one-fifth are in Austin, where DRAMs) plummeted in early 1996, computer chip production expanded largely reflecting the overexpansion of (Health Care Share of Total District Employment) rapidly in the first half of the 1990s. capacity as growth in the demand for Year-over-year percent change Within Texas, the weakening of high-tech equipment unexpectedly 8 the computer chip was most moderated. Recently, memory chip 7.9 apparent in Austin, where overall job prices appear to have nearly stabilized 7.8 growth decelerated from a rapid to and may have bottomed out (upper 7.7 a moderate pace. By contrast, job panel, Chart 5 ). Second, computer 7.6 growth maintained a strong pace in the prices fell in 1996, partly because cost- 1993 1994 1995 1996

Dallas –Fort Worth metroplex, whose saving innovations to microprocessors, SOURCES: Bureau of Labor Statistics; Semiconductor economy, relative to that of Austin, is such as the Pentium chip, were more Industry Association; Baker Hughes. less dependent on the high-tech sector widely adopted. This second factor may

Federal Reserve Bank of Dallas Page 3 ness investment has fallen over time. ploration costs, oil production and costs, and as a result, margins Nevertheless, even though computer production of oil equipment may not have been squeezed either too thin or price deflation was near its long-run slip as much if a modest price decline to a minimum sustainable level. If average last year, the pace of deflation materializes. either possibility proves correct and if has been very volatile.13 The unpredicta- Health Care. There is some concern cost inflation for providers does not bility of past technological break- that much of the recent health care dis- decelerate enough, the pace of HMO throughs makes it difficult to anticipate inflation largely reflects the transition and insurance premium inflation could future computer prices. from traditional insurance to managed very well pick up.15 Whether health care Within the District, several factors care plans. Indeed, some studies main- inflation will rise or stabilize is uncer- suggest renewed but moderate output tain that employer costs for health care tain because further innovations in growth in this industry. On the plus side, coverage typically fall sharply within health care delivery may enhance the more realistic expectations and the re- the first or second year following a ability of managed care to reduce costs, turn of balance between chip orders switch to a health maintenance organi- and future technological gains could and shipments may be setting the stage zation (HMO) but thereafter increase at accelerate cost . for a modest near-term expansion. In the general pace of medical inflation.14 With respect to the Eleventh District addition, retail demand for computers Put another way, health care disinfla- economy, it is unclear whether the pace may be bolstered in coming quarters tion accelerates when the pace at which at which people with medical coverage by the introduction of new micro- people shift into managed care picks shift into managed care will slow. On processors, such as the MMX chip, up. But when the transition is largely the one hand, because Texas has lagged which enhances the audiovisual and completed, the trend of falling medical other states in moving to HMOs (HMO multimedia capabilities of new PCs. inflation could slow or even end. penetration in Texas was well below Furthermore, export demand could pick Some analysts believe that, in addi- the U.S. average as of 1994), the shift up if the of Western Europe tion to completing the transition to may continue after the transition in the and Japan begin to experience a strong managed care, HMOs and insurers will rest of the country is over. Thus, the recovery from their recent slowdowns. need to step up the pace of price in- shift toward managed care could On the downside, unless product im- creases in coming years to rebuild profit arguably continue to slow health care provements such as the MMX boost margins. Some of these analysts argue employment growth in the region. the demand to replace old PCs, some that HMO and insurance price hikes On the other hand, because many analysts are concerned that growth in have not kept pace with health care Mexicans come to Texas for health care, the domestic PC market will slow as the recovery of the Mexican economy the share of with personal could bolster health care employment computers rises at a slower pace than in 1997–98. How these forces will bal- Chart 5 in the early 1990s. Computer Component Prices ance out is unclear. Energy. Based on energy futures Stabilize After Plummeting markets (markets in which people buy Index, January 1991 = 100 energy today for delivery at a future 130 Conclusion date), oil prices are expected to be near Integrated circuits Memory chips 110 Microprocessors Computers (PPI*) $20 a barrel in mid-1997, down from Changes in the supply conditions for 90 $25 a barrel at year-end 1996. Factors computers, energy and health care have 70 behind this expected fallback include substantially affected inflation in the 50 an end to the severe winter weather and the composition of 30 in Europe that had helped drive up ’91 ’92 ’93 ’94 ’95 ’96 job growth within the Eleventh District. prices, some rebuilding of inventories * Producer price index. Because these industries have changed and an increased supply of oil. If these SOURCE: Bureau of Labor Statistics. markedly in recent years and may con- expectations prove correct, energy Computer Deflation Moves tinue to do so in unpredictable ways, prices will fall and help push down CPI Toward Its Long-Run Average supply conditions in these sectors can inflation from the 3.3 percent pace Year-over-year percent change be viewed as a major wild card for fu- posted between December 1995 and 10 ture U.S. inflation and District employ- December 1996. 0 ment patterns. However uncertain the In addition, if futures markets prove –10 outlook for computer, energy and –20 correct, energy-related job growth will Average = –11.8 health care prices, the way they actually –30 slow. However, producers have viewed Actual evolve will almost certainly affect the energy prices as being temporarily high –40 direction of U.S. inflation and the re- ’68 ’72 ’76 ’80 ’84 ’88 ’92 ’96 and have cautiously expanded produc- gional economy. NOTE: Prices based on the GDP chain-weight price tion and hiring. This prudence will index for computers and peripheral equipment likely temper any price-driven slowing investment. — John Duca of energy job growth in 1997. In addi- SOURCE: Bureau of Economic Analysis. tion, because of the decline in ex- (See “Notes” on page 8)

Page 4 Southwest Economy March/April 1997 mate the importance of merchant centives to induce consumers to use tion, the reputation of the issuing insti- acceptance, recall the advent of E-money rather than cash. tution will be vital to the acceptance of cards. The BankAmericard and Master- its E-money. Should consumers and Charge card were introduced in the merchants doubt the safety and sound- United States in the mid-1960s. But Free Enterprise and E-Money ness of the institutions issuing E-money, according to an article that appeared in they always have a near-perfect substi- Life magazine in 1970, “bank cards still In a free enterprise system, innova- tute to fall back on: . encounter areas of resistance. Most tions survive and flourish if the big department stores refuse to honor benefit to users from a new product or —Marci Rossell them.…Restaurants in many places will is greater than what existing have no part of them.” Although credit substitutes offer. E-money is no excep- cards were very attractive to consumers tion. Should consumers and merchants from the outset, the widespread use of fail to find the merits of electronic sReferences credit cards was delayed by a lack of money sufficient to overcome any costs acceptance by merchants. If E-money is associated with its use, E-money could Blinder, Alan S. (1995), Statement Before the Subcommittee on Domestic and International of the Committee on Bank- to succeed, it must prove its merits not very well go the way of the Edsel. ing and Financial Services, U.S. House of Representatives, October 11. only to the consumer but also to the The Federal Reserve to date has re- retail community. frained from imposing regulations on Gleik, James (1996), “Dead as a Dollar,” New York Times Magazine, June 16. From a merchant point of view, the electronic money (aside from the most promising aspect of E-money is the boundaries established by Regulation E) Levy, Steven (1995), “The End of Money?” Newsweek, October 30, potential for substantial cost savings. It in favor of allowing the innovation to 62–65. has been estimated that approximately develop in a relatively unfettered market O’Neil, Paul (1970), “A Little Gift from Your Friendly Banker,” Life, 4 percent of the total of a trans- environment. The issuers of E-money March 27, 48–58. action made with currency is tied up in do not expect individuals to hold rela- Rosenblum, Harvey (1996), “Electronic Money: Hype and Reality,” the counting, storing and protecting of tively large balances on stored-value Banking Strategies, November/December, 6–16. that cash. Merchants are likely to be cards. So long as individual balances Wenninger, John, and David Laster (1995), “The Electronic Purse,” charged a fee for E-money transactions, remain small, the potential failure of Federal Reserve Bank of New York Current Issues in Economics and as they are with credit cards, but elec- institutions that issue E-money poses no Finance, April. tronic money may be slightly cheaper significant risk to consumers. Govern- Zaretsky, Adam M. (1996), “Will That Be … Cash, Check, Charge and easier for merchants to handle than ment intervention, therefore, appears or Smart Card?” Federal Reserve Bank of St. Louis Regional Econo- cash. If so, merchants could offer in- unwarranted. In the absence of regula- mist, April. ATALEOFTHREESUPPLYSHOCKS, NATIONALINFLATIONANDTHEREGION’SECONOMY (continued from page 4) Notes

I thank Jeremy Nalewaik and Sheila Dolmas for research assistance, and Lori Taylor, D’Ann Petersen, ratio is based on nominal data, a spurt in computer price deflation will tend to lower this ratio because new Michelle Burchfiel, Fiona Sigalla and Mine Yücel for comments and suggestions. orders reflect more recent and thus lower prices than shipments. 1 Supply shocks are changes in technology (for example, computers), industrial structure (for example, health 10 See D’Ann M. Petersen and Michelle Thomas, “From Crude Oil to Computer Chips: How Technology Is care) or world resource prices (for example, energy) that alter an industry’s cost schedule and thereby cause Changing the Texas Economy,” Southwest Economy, Issue 6, 1995, 1–5. substantial changes in its . 11 For an analysis of District energy jobs and oil prices, see Stephen P. A. Brown and Mine K. Yücel, “The En- 2 See John V. Duca, “Inflation, Unemployment, and Duration,” Economics Letters 52 (September 1996): ergy Industry: Past, Present and Future,” Southwest Economy, Issue 4, 1995, 1–5. 293–98. 12 Nevertheless, this ratio could overstate the impact of health care restructuring because it excludes health care 3 For example, see Felix G. Rohatyn, “Recipe for Growth,” Wall Street Journal, April 11, 1996, A21. workers in the public sector and because health care workers in the private sector likely have been more af- 4 For example, see Jeffrey C. Fuhrer, “The Is Alive and Well,” New England Economic Review, fected by cost-cutting and mergers. March/April 1995, 41–56. 13 This measure is the cleanest aggregate measure of final computer goods prices that covers at least two 5 For an analysis of trimmed mean CPI measures, see Stephen G. Cecchetti, “Measuring Short-Run Inflation decades. By contrast, the producer price index for the “electronic computers” category begins in 1990, and for Central Bankers,” Federal Reserve Bank of St. Louis Review (forthcoming). the CPI’s home furnishings component blends computers with other items. 6 The trimmed mean CPI can still be affected by changing supply conditions. For example, medical inflation 14 See Elizabeth Kilbreth and Alan B. Cohen, “Strategic Choices for Cost Containment under a Reformed U.S. shifted from being a high outlier excluded from the trimmed mean during much of the 1980s and early 1990s Health Care System,” Inquiry 30 (Winter 1993): 372– 88; J. P. Newhouse, W. B. Schwartz, A. P. Williams and to being more in line with the pace of price rises in other items. C. Witsberger, “Are Fee-for-Service Costs Increasing Faster than HMO Costs?” Medical Care 23 (August 7 For further discussion, see Sheila Dolmas and Mine Yücel, “The Texas Economy: An Overview of ’96 and 1985): 960–66; and Linda Radey and Richard Fullenbaum, “Are Employers’ Health Benefit Costs Finally Outlook for ’97,” Southwest Economy, Issue 1, January/February 1997, 1–4. Under Control?” Review of the U.S. Economy: Ten Year Projections (Lexington, Mass.: DRI McGraw-Hill, 8 Data are based on the work of Dolmas and Yücel (1997). Note that because the book-to-bill ratio in Chart 4 1995), 51–3. is measured quarterly, whereas job growth is measured on a year-over-year basis, by construction the plot- 15 For details, see Milt Freudenheim, “Health Care Costs Edging Up and a Bigger Surge Is Feared,” New York ted job growth series will tend to lag the book-to-bill ratio. Times, January 21, 1997, national edition, A1 and C20; and Ron Winslow, “Health-Care Costs May Be Head- 9 Note that the domestic book-to-bill ratio does not reflect the role of foreign demand. In addition, because the ing Up Again,” Wall Street Journal, January 21, 1997, B1 and B6.

Page 8 Southwest Economy March/April 1997

Top View