July 25, 2006
Mid-year jobs report
Looking west from Ohio to Iowa and Minnesota, there is a distinct falloff in economic growth, at least according to recent reports on payroll employoment. With only a three-week lag, the Bureau of Labor Statistics reports their estimates of payroll employment monthly for individual states. The reported monthly figures for June 2006, now complete the second quarter of this year.
The table below displays year-over-year payroll job growth in the seven Midwest states and the U.S. Note that job growth in all states except Iowa and Minnesota fell short of the U.S. growth of 1.4 percent.
One reason that explains lagging job growth in many Midwest states is their heavy concentration in manufacturing industries. As the Chicago Fed’s Midwest Manufacturing Index suggests, real output growth in manufacturing has been growing strongly now for 3 years in both the nation and in the Midwest. In general, U.S. manufacturing growth has been buoyed by strong domestic demand for capital investment goods and by growth in U.S. exports. Some notable (and growing) Midwest capital goods sectors are mining and construction machinery, farm machinery and equipment, heavy trucks, and electrical equipment. However, strong output growth in manufacturing does not typically propel much payroll job growth because real output gains are generally being achieved through higher productivity rather than through more labor input.
With respect to total payroll employment, the three easternmost states of Ohio, Indiana, and Michigan show the weakest year-over-year growth. Further to the west, job growth in Illinois and Wisconsin have been stronger, with still stronger growth for Iowa and Minnesota.
For some states, such as Illinois, recent payroll job growth is especially encouraging since growth had been lagging since the last recession. Along with Indiana, Michigan, and Ohio, Illinois employment has not yet re-attained its previous peak which occurred in the year 2000.
Illinois' job gains are being led by growth in professional and business service industries even while manufacturing employment has been declining. The Chicago-area economy, which comprises the bulk of Illinois, has been shifting into business and financial services while moving away from manufacturing. Chicago’s business and financial services depend on customers in surrounding manufacturing-intensive states but they also serve some global and national markets.
At the other end of the spectrum, Michigan’s recent job performance remains very much in a league of its own, even when compared to other Midwest states. The chart below indexes total payroll jobs to the first quarter of 2001. While the rest of the region has almost re-attained its former employment peak, Michigan employment remains 6 percent to 7 percent below its previous peak.
The troubles of domestic automakers Ford and GM, and their automotive parts suppliers, have been weighing down growth in Michigan. Since the year 2000, their combined share of U.S. light vehicle sales has declined from 51.1 percent to an average 41.3 percent year-to-date in 2006.
These companies are highly concentrated in Michigan. In addition to their global headquarters and many research facilities and part suppliers, for example, Ford and General Motors together maintain 12 of their 34 U.S. assembly plants (35%) in Michigan. For this reason, Michigan residents are closely following the strategic plans of these companies as they attempt to restore growth and profitability.
July 19, 2006
Honda and tax incentives
Honda recently decided on a site in Indiana for its new North American auto assembly plant over sites in Ohio and Illinois. Indiana offered Honda generous incentives of EDGE tax credits, training assistance, and real and personal property tax abatements totaling up to $41.5 million. In addition, the state will provide infrastructure support for water, wastewater, and road improvements of approximately $44 million. This offer was generous relative to packages that have been offered lately by northern states to woo automotive plants. Did the incentives swing the deal for Indiana? And how can states hope to recoup these upfront costs and revenue losses? More importantly, is society well served by such raw-knuckled competition among states for production facilities? The answers are not definitive, but, though often condemned, the use of fiscal incentives may not be such a bad thing.
Large offers of this nature have become commonplace. Speaking at the Chicago Fed’s recent symposium on the automotive parts industry, Sean McAlinden of the Center for Automotive Research reported that the state of Georgia offered Korean carmaker KIA a package estimated to be worth $409 million. This was noticeably larger than the recent average offers of $57 million in tax incentives for automotive assembly plants for northern states and $44.2 million (plus free or subsidized infrastructure and job training) for southern states.
On completion of such deals, company representatives often proclaim that the incentives did not determine the choice of location, but were rather a sweetener or a comforting pledge of good faith. Professional site selection analysts tend to echo these sentiments. Taking such statements at face value, why do states offer such high stakes packages?
No doubt, there are benefits at the ballot box to those elected officials who can brag about bringing jobs and income to the state. It has been argued that these benefits, especially for investment projects that loom large in the media, result in overly generous offers and poor decision-making by state officials. This is one reason that some states enact legal requirements making the terms of such deals easily available to public scrutiny.
But how can states afford to make such offers? One reason is that the public service costs of hosting businesses are usually lower than the taxes paid by them; that is, there is typically a fiscal surplus inherent in state business tax systems that allows state officials to discount the public tax and service bills on new investment. When I examined the likely costs of public services provided to businesses in a 1996 study, I found that, across all U.S. regions, direct business taxes tended to exceed the value of direct service benefits provided to business by a ratio ranging from 1.5:1 to 2:1. This excess may allow room for governments to lower business tax bills through selective incentives.
Even so, opponents of the use of selective tax abatements may argue that incentives were unnecessary and that businesses have an information advantage in bargaining with states for incentives even when they will end up choosing the same location in any event. Certainly, the proclamations of businesses that afterwards contend that incentives were not a primary consideration in their location decision bear this out. If so, states are arguably better off refraining from incentives and instead spending the business tax bounty on public services or returning personal taxes to state residents.
In the case of auto plants, it is interesting to note that even if individual states “give away the store” in luring a particular auto assembly facility, the end result may ultimately benefit the state’s economy. The reason is that the assembly plants typically attract auto parts suppliers to the area. As Chicago Fed economist Thomas Klier has shown (below), a typical assembly plant can draw a significant nearby supplier base. For recently opened assembly plants in North America, an average of 19 to 20 direct suppliers have typically opened up within 60 miles of the plant. More generally, assembly plants tend to pull in many more supplier plants within several hundred miles, and supplier plant employment generally exceeds assembly plant employment by around 3.5:1.
It is true that in the case of the Honda assembly plant in Greensburg, IN, many of the supplier plants will be outside Indiana’s border and tax reach. However, if all or many adjacent states are successful in attracting assembly plants, the spillover benefits of taxation and income will accrue in roughly equal measure to the states. A so-called cluster of automotive production capability may be achieved for the multi-state region.
But are the incentives necessary to achieve or preserve the region’s cluster of automotive plants? At least for highly capital-intensive industrial activities such as manufacturing, the so-called business climate of the state is paramount. Placement of an expensive investment by a company in a state must be based on a strong conviction that future government leaders will not expropriate the facility’s value through regulation, over-taxation, or non-cooperation in future land use and public infrastructure needs. The situation is not unlike making investments in a foreign country. When the capital investment is fixed and not easily moved, confidence in local government is a key factor in assessing investment risk.
In this regard, Honda’s decision to locate in Indiana rather than Ohio is understandable. While proximity to its large suppliers in Ohio and vicinity was a compelling reason for considering Indiana, the desirability of diversification among government entities may have also been a factor. As for the incentive package, there is surely more to a favorable state business climate than a flashy offer of tax incentives. But at the same time, the offer of a fiscal incentive package may be a strong signal to the business that its presence will be valued. In addition, a sizable and highly visible tax incentive package may represent an implicit acknowledgment by the state that the investment is wanted, making it more difficult for future political leaders to renege on the state’s cooperative relationship with the company.
Of course, implicit tax incentive contracts of this sort work both ways. Companies that receive generous tax incentive packages, but later do not deliver on promised jobs and investments, are easy targets for retribution by state officials. In many instances today, “clawback” provisions are included upfront that eliminate favorable tax treatment if companies do not deliver.
Even so, the “gold standard” by which public policies must be judged is whether the state could possibly do better. Opponents of tax incentives for business argue that, because of such tax breaks, critical public services such as education remain underfunded. In particular, public education suffers, contributing to sub-par income growth and exacerbating social problems such as crime, poor electoral participation, and poor public health. If we accept this view, the economic returns to the practice of competitive business tax incentives are not optimal; the economic returns from any short-term job and income gains to the local economy are less than the foregone returns that greater education spending would bring locally and nationally.
In rebuttal, one might argue that business taxes are not the only possible source of revenue for highly valued public services such as education. An ideal of government is one in which citizens understand both the value of public services provided and the real costs of these benefits and, subsequently, make their choices known at the ballot box.
With the tendency among governments to over-tax business activity, the electorate may believe that they are getting a free ride for public services—that they are not in fact paying for these services. But they are usually mistaken. People and households end up (indirectly) paying for public services in any event. After all, business taxes are ultimately reflected in higher product prices paid by state residents or in lower wages and salaries paid to employees.
So why do many voters and even some policy analysts advocate the taxation of business activity to finance public services that primarily benefit households? Some argue that Americans like their taxes hidden and furthermore that this is a reasonable way for governments to finance high-payoff public services. But this approach has risks. If taxes and prices for public services are hidden, can the citizenry really make sensible decisions about what levels, types, and extent of services government should provide? What do you think?
July 11, 2006
Interstate Income Convergence and Development Policy
These days, there is some concern over rising income inequality among workers and households in the United States, especially slow wage growth at the bottom of the income distribution. In contrast, from a longer-term perspective, America's general experience with household well-being has been strongly positive—both across time and geographically. Over the 20th century, household incomes have risen many times over. Reports from the Federal Reserve Bank of Dallas document American progress in tangible living standards, such as gains in homeownership, rising income, shorter work weeks, and rising life expectancy.
In its 2005 Annual Report, the Federal Reserve Bank of Cleveland takes a geographic perspective on economic progress. Here again, a longer-term perspective is very positive. Against the backdrop of rising national standards of living, the report finds increasing geographic income equality rather than inequality. In particular, average incomes across U.S. multi-state regions, and among states in general, have been profoundly converging rather than diverging.
The causes and mechanisms of this income convergence are worth exploring in identifying possible lessons and directions for economic development policy today. What factors and policies can keep states and regions out in front of the race for economic well being? States in the Midwest are especially concerned that they are falling behind economic growth and well-being of some other states in the South and West.
In its methodological approach, the Cleveland Fed analysis attempts to explain the lack of full convergence of interstate per capita income since 1934. Neoclassical economic theory predicts that full economic convergence will take place in a flexible market economy, such as the U.S. economy. If, as we generally believe to be the case in the U.S., states share the same technologies in production, and if factors of production (labor and capital) are mobile across regions, then wages and household incomes should converge. Such convergence takes place over time as workers migrate toward better jobs and income or as capital investment follows greater returns in lower-cost regions. (There are other variants of the economy’s flexibility that achieve the same result).
As the Cleveland Fed’s chart below shows, in the U.S., state per capita income has mostly converged since 1930—it has converged from a standard deviation of around 0.4 to a mostly flat standard deviation of 0.15 since the late 1980s. Much of this convergence reflects rapid growth and economic progress in the formerly underdeveloped southern states. In the South, major public investments in education and roads, accompanied by private investments by manufacturing companies and more recently by services, have brought up incomes close to national norms and eliminated many areas in dire poverty.
States of the Seventh Federal Reserve District have historically enjoyed average incomes above the nation. However, in recent years, per capita income growth in the District has lagged the nation's average. (The Midwest web page allows visitors to build customized charts of state personal income). The chart below illustrates that District states' per capita income remained at more than 10 percent above the nation's average during the early 1950s. By the late 1970s, relative income had converged to levels only 5 percent above the nation. A more preciptious decline took place from the late 1970s to 1983 when relative income first fell below the national average, and remained there throughout the 1980s. The more prosperous times of the 1990s lifted the region's incomes above parity for awhile only to fall below once again in the current decade.
As the second chart below reveals, all five states of the Seventh District have experienced relative declines since 1969.
Since full convergence of state incomes has not taken place, it may be the case that public policies are feasible to push a state’s per capita income above or below average. The Cleveland Fed’s statistical model comes up with fairly strong evidence in identifying variables that explain or at least correlate well with why states fail to fully converge with the national average. The strongest explanatory variable is “utility patents”—a proxy or general stand-in for states’ innovation and entrepreneurial activity. Apparently, local innovation can keep incomes high in a region as new firms (and high paying jobs) are spawned or through some other mechanisms.
The second strongest explanatory variable explaining lack of full convergence is differences in educational attainment among the states’ work forces as measured by high school and college education attainment. Presumably, while U.S. workers are mobile in moving to other states in search of higher wages, this migration mechanism is imperfect or slow to adjust. Accordingly, some economic returns from public investment in college education may accrue to the students’ home region (and therefore income convergence is incomplete). Education is also considered to be complementary with “patents” or innovation in economic growth because a highly educated population can more easily learn and adapt new technologies.
The important third variable is “industry specialization.” Places with concentrations of manufacturing had higher incomes early on, but this has since tended to dampen income growth over time (and has tended to do so persistently). Presumably, a region’s workers cannot or do not adjust quickly (e.g., move away or retrain) to the negative shocks that have affected such industry sectors.
What can we take away from such an analysis? The Cleveland Fed intends this initial research to be directional rather than prescriptive. That is, it offers guidance and direction for further research that is needed to identify viable and specific public policies.
In some sense, the research findings are also corroborative. That is, the findings are very much in the mainstream of current economic development policy discussion. How to innovate? How to educate? Which policies will continue to enhance growth? And ultimately, which particular policies toward educational attainment and entrepreneurship are effective and cost-effective?
What do the findings have to say specifically about Seventh District states and other Midwest industrial states? The side-by-side charts below illustrate the findings for each state. The left-hand chart lists the actual per capita incomes of each state in 2004 in relation to the national average. The right-hand chart lines up this same listing of states with the Cleveland Fed’s model of predicted per capita income. The particular contribution of each explanatory factor to the prediction—innovation, education, and industry mix—are also illustrated.
The Cleveland Fed analysis predicts Ohio, Michigan, Illinois, and Wisconsin to have higher per capita incomes than they actually do have in 2004. (Iowa and Indiana are right on par.) Predictions that are stronger than actual for these Midwest industrial states derive from their high research and patenting activities.
Why, then, are Midwest states lagging in their incomes despite their strong innovative traditions? The possible explanations are thus far elusive. It may be that the model’s enumerated patents are mismeasuring (overcounting) actual innovation taking place in the Midwest region, since patents are sometimes assigned to the headquarters of a firm in a region, even though the innovative activity takes place elsewhere. In an increasingly global economy, with large multinational companies, this data problem may be worsening over time.
Another possibility is that patents in the Midwest’s particular industries have lower economic returns lately as compared with patents in those industries (e.g., microchips or software) that are more specialized in other states and regions.
Yet another possibility has been most intriguing to Midwest leaders in economic development thought and policy. Is it the case that some other feature of Midwest behavior or policy is failing to commercialize research innovations that are taking place or available here?
Again, the possibilities are myriad. Among them, some researchers have pointed to superior mechanisms that have been crafted in successful regions, including Massachusetts, North Carolina, and California, whose universities have been successful in transferring research and development from the laboratory to commercial enterprise. The Federal Reserve Bank of Cleveland will take a closer look at this proposition during its November conference on the university’s role in technology transfer. At its October 30 conference, the Chicago Fed will also be taking up this issue as it investigates several possible roles that the university might take in regional economic development.
July 6, 2006
Manufacturing exit tough on Midwest central cities
If current trends continue, manufacturing activity will soon become extinct as a part of central city economies. The reasons for this exodus are largely the result of shifts in the technology of many types of production activity. Central cities—especially in the Midwest and Northeast—are generally densely populated and somewhat congested. Such conditions are not ideal for production activity. Many central cities vigorously attempt public policies to preserve manufacturing jobs, but the opposing forces appear to be very strong.
At one time, many central cities were the preferred locale for manufacturers. The reasons can be boiled down to two, transportation of laborers and transportation of materials.
As for labor, factories were once teeming with laborers. But due to labor-saving productivity gains, today’s factories are sparsely populated even though they produce many times more output. Senior Business Economist Bill Strauss calculates that today it takes 200 U.S. manufacturing workers to produce the same amount of product as 1,000 workers in 1950. Accordingly, during those earlier labor-intensive times, the transportation of manufacturing workers to the job site figured much more heavily into the factory cost equation. Transportation efficiency once was served by factory neighborhoods in central cities where workers could more easily commute by walking, driving, or by public transportation. The higher living density of central cities also meant that public services such as education and sanitation could be delivered cheaply to workers. Of course, it is not only manufacturing technology that changed. Better highways and rising standards of living (translated into higher car ownership) have also contributed to the ability of factories to staff their factories with (fewer) workers who live farther away. In turn, this opens up factory sites in suburban and rural areas.
Better highways, road vehicles, and logistics technology have also made the transportation of production material to central cities less attractive in comparison to areas of lower population density. Economically, railroads once dominated long-haul truck transportation of materials and components used in manufacturing, as well as the shipment of finished goods to other final markets. The technology of rail favors convergence into a central location (i.e., central cities) rather than the dispersed locations that are served by the crisscross pattern of our now ubiquitous highways. Over time, construction of divided highways and the advent of trucks having features such as refrigeration, trailers, and easily transferred containers have facilitated factory sites served by roads rather than by rail. Accordingly, factory sites can now better take advantage of the low land costs of rural and suburban areas rather than being restricted to those of the central cities.
The City of Chicago exemplifies the central city experience with manufacturing jobs. The chart below shows that, by one reckoning, manufacturing jobs in the city have declined from 367,000 in 1976 to under 100,000 today—a loss of approximately 10,000 per year. In contrast, the employment experience of Chicago’s suburban areas has been much milder.
The experiences of other central cities has been somewhat similar to Chicago’s, even some of those cities located in the faster-growing Sunbelt regions. The table below, drawing on data from the Census of Manufactures, describes the manufacturing job changes from 1977–2002 of the 10 most populous U.S. cities (as of 1980). Over the period, manufacturing jobs in these 10 cities dropped by 62%, which is more than double the pace of manufacturing job loss in the overall U.S. Although the job gains of San Diego and the slight loss by Phoenix seem to be exceptions, they are not. Rather, their experience reflects the fact that the land area of those cities has expanded by 2.5 times and 4 times, respectively, through annexation since 1980. At the same time, Midwest city boundaries have remained essentially fixed.
So too, as shown by the table below, the city of Chicago’s experiences are mirrored closely by the central cities of the industrial Midwest. The high population density of places such as central city Milwaukee and Cleveland came about during a different era than the more recent growth of low-density (and expanding) cities of the West and Southwest.
Many central cities of the Midwest owe their original existence to manufacturing, so the steep loss of manufacturing jobs in central cities has typically been painful. In response, these cities often attempt to combat manufacturing decline through public policies. For example, some policy initiatives to make manufacturing activity more competitive in cities include clearing land, cleaning up environmental hazards, preserving or setting aside land exclusively for manufacturing purposes, or easing freight transportation congestion. However, so far, the allure of suburban and ex-urban manufacturing locales has been too strong to overcome.