Chicago's economic transformation: past and future.

Schindler, Graham ; Israilevich, Philip ; Hewings, Geoffrey 等

Hog butcher for the world, Tool maker, stacker of wheat, Player with railroads and the nation's freight handler; Stormy, husky, brawling, City of the big shoulders. --Carl Sandburg, "Chicago" (1916)

The charismatic city that Carl Sandburg depicted in 1916 has since undergone a dramatic transformation. By the late 1970s, deep recessions had all but eliminated much of the romanticism associated with Chicago's industrial economy. Indeed, the city and the entire Midwest seemed to produce almost daily reports of manufacturing plant closings and layoffs. The dire straits of those times appeared to portend only more gloom for the future. But during the same period, Chicago was fostering a growing service sector. By maintaining the growth of the industries in this sector, Chicago has reconstructed itself into a strong, thriving economy.

As several recent studies have shown, Chicago's economic structure changed dramatically during the 1970s and 1980s, with services coming to dominate both output and employment. In fact, by the early 1980s, service employment had surpassed manufacturing employment in the metropolitan area, foreshadowing a similar transformation nationally a few years later. Not only did Chicago lead the nation in this trend, but the relative degree to which services took over from manufacturing was more dramatic in Chicago than in the nation as a whole. Meanwhile, with manufacturers retiring old inefficient factories, the relics of what some would call Chicago's glory days, and replacing them with establishments featuring new highly efficient capital, manufacturing output exhibited little or no growth in real terms. With more efficient factories now in use, labor requirements were not as high as in the past; accordingly, between 1970 and 1992 manufacturing firms within Chicago cut over 280,000 jobs.

Chicago's new economic character has arisen because of two fundamental changes, each of which may have different policy implications. First, through the difficult transitional period of the 1970s and early 1980s, firms survived because they became more efficient and were able to compete in the increasingly globalized marketplace. Second, the growing dominance of the service sector has fundamentally altered the composition of Chicago's exports. These changes combined to reduce the importance of the pro-cyclical manufacturing exports in favor of those from the less cyclically sensitive service industries.

At the same time, Chicago's interaction with other midwestern or national economies has also been affected. Despite the fall of manufacturing's share of total exports, Chicago's linkages with other economies have strengthened. This is due to three key developments: increased dependence on external markets for inputs in manufacturing, increased service sector trade, and altered consumption patterns as a result of increased residential mobility. In sum, Chicago's economy has become less cyclical and more dependent on the rest of the nation.

Studying Chicago

A sketch of structural changes such as these prompts further questions: Which changes are likely in the future? Which industries were most important in engendering change, and will they continue to be prominent in future transformations? How will the recent changes affect the economy's reaction to exogenous shocks? Is it reacting differently than in the past, and what might we expect in the future? How might these changes affect other midwestern economies? These are the types of questions that the Chicago Region Econometric Input/Output Model (CREIM) was designed to answer.(1) Created by the Regional Economics Applications Laboratory, CREIM allows us to examine changes that have taken place in the past and to forecast those that might occur in the future. Additionally, it can estimate the direct and indirect impacts of these changes on various sectors of the economy.

CREIM combines traditional input/output analysis with time-series analysis.(2) The input/ output component enables a detailed analysis of purchases and sales between industries, while the time-series component allows for the analysis of intertemporal change in the transaction flows of goods and services. Together, these two components produce a detailed analysis of structural change over time at the sectoral level.(3) By taking into account transaction flows between industries, we are able to estimate the spillover or indirect effects within the economy that direct analysis cannot capture because it examines each sector irrespective of its effect on other sectors. For example, fabricated metals production requires machine tools and steel as inputs; machine tool companies require steel to build the tools. An increase in the demand for fabricated metals products will impact the steel industry in at least two ways: by increased direct consumption (from fabricated metals) and increased indirect consumption (from machine tools). In this way, each sector of the economy is linked to all others, either directly or indirectly.

The experiment

To analyze the impact of Chicago's structural change over time, we designed an experiment to answer the question, how would a national recession affect the Chicago economy, and would the magnitude of this impact change over time? For ease of exposition, we decided to take the recession of 1982 as an example. Using CREIM, we estimated total exports from Chicago to the rest of the nation for both 1981 and 1982. We calculated that 1982 exports were 5.5 percent less than 1981 levels; we then used the magnitude of this export shock to simulate a national recession.

The impact on exports from the region during the 1982 recession was largely concentrated in manufacturing industries, with durable goods hit particularly hard. We estimate that between 1981 and 1982, exports of primary metals industries dropped by approximately 33 percent, exports of fabricated metals by about 20 percent. Despite the large export losses in some sectors, other sectors actually saw exports rise. Of the 36 sectors in the model, nine of them, including food and kindred products, petroleum products, and business services, exhibited export growth between 1981 and 1982. Of these nine, all the resource and manufacturing sectors saw exports subsequently fall from those levels by 1983.

In order to ease interpretation of our results, we did not use the estimated sectoral export changes, which had varying signs. Rather, we distributed the total export change from 1981 to 1982--5.9 percent of 1982's total exports--across all sectors, weighted by the size of their estimated exports in 1982. This gave all export changes the same sign and facilitated interpretation of the final results. We then used this new vector of export change to represent an economy-wide export recession, with the overall magnitude equal to that of the 1982 recession, to shock CREIM. We introduced the recessions separately into CREIM as a fall in exports of 5.9 percent for the years 1975, 1985, 1995, and 2005. These four observations provided vectors of changes in output, employment, and so on, which would determine Chicago's reaction to national downturns over time. For the simulated recessions we assumed that only exports were shocked, not technology; thus, the estimated changes in productivity and wage levels within Chicago occurred only in response to the reduction in exports, not in some exogenous form due to the recession.

The recessions introduced into the model are constant through time in relative terms even though exports as a share of total output are increasing over time. From 1975 to 2005, exports' share of total output is forecasted to increase from approximately 30 percent to 37 percent. Thus far, total manufacturing output--traditionally considered the source of most exports--has remained relatively constant, while service-oriented output has grown and is expected to increase even further (see figure 1). In fact, manufacturing exports were between 29 percent and 34 percent of total output during this time, while service exports grew from 35 percent of total output in 1975 to 43 percent in 2005 (see figure 2). As a result of this forecasted rise in total exports, the recession vector introduced into the model as a change in export demand relative to total output increased in size from 1.8 percent in 1975 to 2.2 percent in 2005. This slight increase in the size of the shock will impact its nominal effect but not the results, because each sector's share of the total impact is fixed.

[Figures 1 & 2 ILLUSTRATION OMITTED]

The results of the four simulations are presented in tables 1 and 2. Table 1 shows the percent change, due to a shock, in sectoral output and employment relative to their respective values from a base case simulation where there is no external shock.(4) Table 2 shows each sector's contribution to total output and employment change caused by the shock.(5)

TABLE 1

Sectoral effect of shock relative to a base case (percent change) 1975 1985 1995 2005 CGRP(a) -6.33 -5.80 - 5.31 -5.21 Employment Total -5.37 -5.09 -4.67 -4.56 Resource -3.72 -3.98 -3.11 -3.55 Agriculture -5.69 -5.96 -5.87 -5.89 Mining -5.39 -5.10 -4.70 -4.73 Construction -3.25 -3.56 -2.59 -3.02 Manufacturing -5.68 -5.61 -5.38 -5.42 Nondurables -5.36 -5.59 -5.48 -5.48 Durables -5.87 -5.62 -5.30 -5.37 Total services -5.22 -4.98 -4.64 -4.50 TCU(b) -5.65 -5.39 -5.13 -5.16 Trade -5.64 -5.51 -5.29 -5.32 FIRE(c) -5.59 -4.73 -4.66 -4.76 Services -4.64 -4.67 -4.32 -4.14 Government -5.50 -5.02 -4.71 -4.73 Output Total -5.45 -5.29 -4.99 -5.08 Resource -3.54 -3.71 -2.76 -3.17 Agriculture -5.70 -5.96 -5.85 -5.85 Mining -5.39 -5.10 -4.70 -4.73 Construction -3.25 -3.56 -2.59 -3.02 Manufacturing -5.96 -5.96 -5.76 -5.93 Nondurables -6.09 -6.31 -6.12 -6.34 Durables -5.88 -5.64 -5.38 -5.42 Total services -5.29 -5.09 -4.90 -4.92 TCU(b) -5.64 -5.37 -5.09 -5.12 Trade -5.64 -5.51 -5.29 -5.32 FIRE(c) -5.56 -5.47 -5.63 -6.01 Services -4.64 -4.37 -4.12 -4.02 Government -5.50 -5.13 -4.83 -4.85

(a) Chicago gross regional product.

(b) Transportation, communications, and utilities.

(c) Finance, insurance, and real estate.

TABLE 2

Sectoral share of total change (percentages) 1975 1985 1995 2005 Employment Resource 3.94 4.62 5.02 5.37 Agriculture 1.00 1.12 1.41 1.58 Mining 0.17 0.14 0.13 0.13 Construction 2.76 3.36 3.49 3.67 Manufacturing 29.20 22.13 16.20 13.04 Nondurables 9.95 8.84 7.42 6.43 Durables 19.25 13.30 8.77 6.61 Total services 66.87 73.24 78.79 81.59 TCU(a) 7.12 6.33 6.60 6.68 Trade 22.92 22.40 20.09 17.41 FIRE(b) 10.98 9.85 9.69 8.47 Services 24.71 33.47 41.16 47.64 Government 1.14 1.19 1.25 1.38 Output Resource 4.74 5.43 4.95 4.84 Agriculture 0.28 0.32 0.38 0.40 Mining 0.68 0.30 0.18 0.11 Construction 3.78 4.81 4.39 4.33 Manufacturing 46.90 39.02 37.63 34.59 Nondurables 19.15 19.90 20.49 20.48 Durables 27.75 19.12 17.14 14.11 Total services 48.37 55.55 57.41 60.57 TCU(a) 9.09 9.00 9.40 10.25 Trade 12.81 13.61 13.85 14.56 FIRE(b) 11.30 16.17 15.40 15.39 Services 14.11 15.73 17.71 19.25 Government 1.06 1.04 1.05 1.12

(a) Transportation, communications, and utilities.

(b) Finance, insurance, and real estate.

Recession impacts

The two tables suggest many interesting observations. First, in table 1, for several sectors, the changes between one column and the next are not always in the same direction. This is because of the nonlinear movements of multipliers and input/output coefficients revealed by CREIM over time.(6) For example, although the impact on total output exhibits a downward trend, it shows increased sensitivity in 2005 relative to 1995.

The relative impact on total service employment decreases during each period. Similarly, from beginning to end, total service output change, manufacturing employment change, and manufacturing output change all exhibit modest decreases in their respective sizes, although in each of these cases the change is not monotonic over time. Within manufacturing, nondurable output seems to be more sensitive to change than durable output in each time period. Nonetheless, employment sensitivity early on was higher for durable sectors; in the later forecasted years, it is higher for nondurables.

The impact on total value added for the Chicago economy, as measured by gross regional product. exhibits a continuous reduction over time. In the 1975 simulated recession the economy lost 6.3 percent of gross regional product; however, when we apply the recession impact in other years, that level drops continuously to 5.2 percent in 2005, a reduction of nearly 18 percent.(7) The size and direction of this fall are not so much an indication that Chicago's economy is becoming more interrelated over time, but more a result of the increased diversification of exports away from the cyclically sensitive manufacturing sectors.

Table 2 shows that over time, the contribution of manufacturing to total output change decreases quite steadily from almost 47 percent in 1975 to less than 35 percent in 2005. Within manufacturing, nondurables seem to account for more of the change, trading places with durables after the first time period. Not surprisingly, durable output was greater than nondurable prior to the early 1980s. During the 1980s nondurable output levels began to match those of durables; by the mid- 1990s we expect nondurable output to surpass durable output. This large increase of nondurable output within the Chicago area results from strong nominal growth in its three largest sectors: food and kindred products, petroleum and coal products, and chemicals and allied products. Additionally, the rubber and plastic products sector, although only the fifth largest nondurable manufacturing sector, exhibits the most dramatic growth in percentage terms.

Despite the increased contribution of the nondurable sectors to output change, their contribution to employment change diminishes throughout the period. However, the gap between nondurables' and durables' contributions to employment change shrinks, so that by 2005 they make relatively equal contributions to total employment change. This is largely due to the fall of durables' share of total manufacturing employment relative to that of nondurables. This switch is forecasted to continue so that eventually, nondurable industries will constitute a majority of manufacturing employment (see figure 3). The gap between durable and nondurable shares of manufacturing employment is shrinking despite the larger productivity gains within nondurable industries, specifically in the petroleum and coal products and rubber and plastic products sectors. At the aggregate level, manufacturing's contribution to total employment change falls from 29 percent in 1975 to 13 percent in 2005 (table 2), undoubtedly a result of manufacturing's declining share of total employment (see figure 4). As the Chicago region's exports are shocked by national economic movements, the transmission of the impact from the nation to the region involves simultaneous changes of both output and employment. Such effects can easily be analyzed with the elasticity of relative productivities:

[Figures 3 & 4 ILLUSTRATION OMITTED]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where [x.sub.i] and [n.sub.i] are respectively the output and employment of sector i. The variable [[epsilon].sub.i] is the ratio of sector i's share of total output change divided by its share of total employment change. If [[epsilon].sub.i] [is greater than] 1, which indicates transmission reduction, then a change in output of sector i will translate to a relatively smaller change in employment; the transmission mechanism will dampen the impact of an output shock on employment. Conversely, if [[epsilon].sub.i] [is less than] 1, which is a transmission magnification, then a change in output of sector i will translate to a relatively larger change in employment; that is, the transmission mechanism will amplify the impact of an output shock on employment. Table 3 contains the elasticities of relative productivities for all the sectors through time. Table 3 Elasticity of relative productivities 1975 1985 1995 2005 Resource 1.203 1.175 0.988 0.900 Agriculture 0.275 0.289 0.273 0.251 Mining 3.931 2.177 1.462 0.901 Construction 1.368 1.430 1.260 1.178 Manufacturing 1.606 1.763 2.323 2.653 Nondurables 1.925 2.252 2.760 3.186 Durables 1.442 1.438 1.953 2.135 Total services 0.723 0.758 0.729 0.742 TCU(a) 1.277 1.420 1.425 1.535 Trade 0.559 0.608 0.689 0.837 FIRE(b) 1.029 1.641 1.590 1.816 Services 0.571 0.470 0.430 0.404 Government 0.933 0.877 0.843 0.807

(a) Transportation, communications, and utilities.

(b) Finance, insurance, and real estate.

For total services, the elasticity of relative productivities rises from 0.72 to 0.74; for manufacturing [[epsilon].sub.i] grows from 1.61 to 2.65. Thus while manufacturing is becoming less sensitive to output changes, so are services; in that sector, employment is becoming less sensitive to changes in out put, although it still exhibits transmission magnification. Since manufacturing is cyclical, the reduced sensitivity of manufacturing employment to output means that the Chicago economy will be less sensitive to cyclical changes. Furthermore, although the services sector is less cyclical, its reduced employment-to-output sensitivity will further reduce Chicago's overall vulnerability to cyclical fluctuations. This transformation is largely due to the significant productivity advances in manufacturing, which have yielded such large gains that the economy's total productivity is growing faster than that of services.

Similar analysis can be made of changes within manufacturing. Both the durable and nondurable sectors show significant productivity growth, but that of nondurables is 30 percent to 50 percent higher, accounting for large output gains despite the continued reduction in overall employment. The overall effect becomes obvious in [[Epsilon].sub.i]. For nondurables, [[Epsilon].sub.i] grows from 1.93 to 3.19, while for durables it increases from 1.44 to 2.14. Thus while Chicago will become less sensitive to exogenous manufacturing shocks, durable employment will still remain more susceptible to output variations than will nondurable employment.

Thus, it appears that the thinning out of underperforming manufacturing firms in the 1970s and 1980s and the overall increased competitiveness of Chicago firms have set the economy in a much stronger position. Within manufacturing, significant productivity advances seem to account for most of the decreased sensitivity to the economic downturns of the early 1990s. This trend is expected to continue into the near future, with nondurables eventually dominating both manufacturing output and employment. In total, the service transmission magnification should be strongly outweighed by the transmission reduction of the manufacturing industries, thus reducing Chicago's vulnerability to exogenous macroeconomic shocks.

Interregional linkages

As older inefficient manufacturing establishments have been either closed or re-tooled, firms have lost the deep roots that formerly held them to the relatively centralized industrial areas of Chicago. Thanks to advances in computers and communications, service industries no longer need to remain in the central business district. The allure of lower taxes and the opportunity to reduce other overhead costs has led many firms to relocate to suburban areas in order to cut costs and increase competitiveness As firms move people move. In 1990, 62 percent of the six-county region's total population lived outside Chicago proper; in 1970 the number was 52 percent. During this time, the city's total population fell by almost 590,000 while its suburbs gained 875,000 people. As the population and work force of Chicago's six-county region have grown and spread out, a large portion of economic activity has moved into the suburbs and surrounding regions. This occurred as geographic consumer spending zones were pushed across regional boundaries. Accompanying this has come an increase in the number of people who work within the Chicago region but do not live within its borders.

Between 1980 and 1993 the number of people who held jobs within the six-county area but who lived outside the area increased by approximately 32,000. We estimate that by 2010 this figure will increase by an additional 30,000. This means that the regions surrounding Chicago, including northern Indiana and southern Wisconsin, are becoming more dependent on Chicago for their own economic growth as people earn money in Chicago but spend most of it in their home communities. Such a development affects the Chicago economy in several ways. Our estimates suggest that between 1969 and 1992, those people who lived outside the Chicago area but held jobs within it had average annual wages and salaries of approximately $2 billion (in 1982 real dollars), most of which would be spent outside Chicago. This has obvious tax consequences for the six-county region's local governments, especially when multiplier effects are considered. Additionally, the growing number of nonresident employed raises the pressure on Chicago firms to increase their relative competitiveness as they are forced to compete for quality labor. As a result, local issues such as taxes or transportation will have magnified impacts on the region as a whole.

The links between Chicago, these communities, and other parts of the country have increased in another way as well. As mentioned, both exports from and imports into Chicago, as a percent of total output, have increased. Thus Chicago has become more reliant on those regions to purchase its goods, while those regions rely on Chicago's products to produce their own goods, a portion of which will then be purchased in Chicago. For example, a survey of fabricated metals establishments revealed that in 1992, 23 percent of intermediate goods purchases came from other midwestern states, and 21.7 percent of sales were made to the same region. A further 28 percent of sales were made to the rest of the United States.(8) This two-pronged strengthening of interregional linkages becomes important because Chicago's economic transformation has fundamentally changed the way the city reacts to economic stimuli, and this change will now affect other communities much more than ever before.

Conclusions

Chicago's economy is not glamorous. It is not dominated by high-tech businesses or manufacturing behemoths. Its recent economic performance has not been spectacular; economic growth in the 1980s lagged that of the United States, although Chicago did see less of a downturn in the early 1990s. But during the past 20 years Chicago's economy has exhibited significant structural change, the implications of which are now beginning to surface. These structural changes appear to have yielded benefits for both Chicago and the Midwest.

We have shown that Chicago has been transformed into an economy highly dependent on the service sector for its growth and vitality. Service exports have grown relative to service output as well as to total output. This relative movement away from manufacturing exports helps reduce exposure to manufacturing fluctuations on the macro level, thus cushioning Chicago from some macro shocks. But the overall increase of exports as a percentage of total output means that Chicago is now potentially more dependent on the rest of the Midwest and nation to provide a demand base for locally produced goods. This implies an increased exposure to macro fluctuations. But within manufacturing, productivity increases have been dramatic, results of a newly updated capital stock and a possible shift of intermediate purchases from other firms within the region to more competitive lower-cost providers outside Chicago. This effect has strengthened manufacturing's transmission reduction effect, the decreased sensitivity of employment to changes in output, thus decreasing Chicago's exposure to the highly cyclical manufacturing sectors.

The economic connections between Chicago, the Midwest, and other regions have been intensifying, with local firms increasingly relying on external markets to supply intermediate goods and to purchase final goods and services. Other changes have come about on the consumer side. As population has grown in the suburbs and other outlying areas disproportionately to Chicago's central business district, more people living outside Chicago have come to depend on Chicago to supply their jobs. This implies that neighboring regions now depend more heavily on Chicago to supply the funds for consumer demand.

In sum, these findings suggest that Chicago's economy is becoming less sensitive to exogenous shocks. The combined health of services and manufacturing seems to be the largest contributing factor, even though interregional linkages are now obviously much more important than they once were. This should indicate that although Chicago's economy seems strong, the economic health of other midwestern economies now becomes much more significant to our own.

NOTES

(1) We define the Chicago region as the six-county metropolitan area containing Cook, DuPage, Kane, Lake, McHenry, and Will counties.

(2) CREIM, defined in real 1982 dollars, forecasts output, employment, and income for 36 roughly two-digit SIC sectors, as well as many other final demand and demographic variables up to 25 years into the future.

(3) See Israilevich et al. (1994) and Schindler et al. (1994) for further details.

(4) Let x denote output, b denote the base case, and s denote the shocked case. Then the impact of the shock on output in percentage terms for sector i, i = 1... N, is [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. The impact is similar on employment, n.

(5) Sector i's share of total output change is [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] The impact is similar on employment, n.

(6) See Schindler et al. (1994) for a complete discussion.

(7) Although we present results for only four years, the general results do not change if we examine the data on an annual basis. Annually, though not continuously, the same downward trend is clear. Additionally, for the sectoral results, the tabular information is representative of annual trends.

(8) See Carter (1994) for further details.

REFERENCES

Carter, K.L., "The fabricated metals industry in Chicago," University of Illinois at Urbana-Champaign, manuscript, 1994.

Israilevich, P.R., and R. Mahidhara, "Chicago's economy: Twenty years of structural change," Economic Perspectives, Federal Reserve Bank of Chicago, Vol. 14, No. 2, March/April 1990, pp. 15-23.

--, "Hog butchers no longer: 20 years of employment change in metropolitan Chicago," Economic Perspectives, Federal Reserve Bank of Chicago, Vol. 15, No. 2, March/April 1991, pp. 2-13.

Israilevich, P.R., G.J.D. Hewings, M. Sonis, and G.R. Schindler, "Forecasting structural change with a regional econometric input-output model," Regional Economics Applications Laboratory, University of Illinois at Urbana-Champaign, discussion paper no. 94-T-1, 1994.

Schindler, G.R., P.R. Israilevich, and G.J.D. Hewings, "Three-dimensional analysis of economic performance," Regional Economics Applications Laboratory, University of Illinois at Urbana-Champaign, discussion paper no. 94-P-6, 1994.

U.S. Department of Commerce, Bureau of Economic Analysis, Regional Economic Measuring Division, "Regional economic information systems report," CD-ROM, May 1995.

Graham Schindler is a research associate, Philip Israilevich is the associate director, and Geoffrey Hewings is the director of the Regional Economics Applications Laboratory (REAL). REAL is a cooperative venture between the University of Illinois and the Federal Reserve Bank of Chicago. Philip Israilevich is also a senior economist and research officer at the Federal Reserve Bank of Chicago. The authors thank David Allardice, Kathy Moran, and Janice Weiss for their comments and suggestions.