Over the past quarter-century, Kansas has had an average annual economic growth rate that ranks 35th among the states; and ranks 4th among the seven states in the Plains region. Relatively slow labor productivity growth in Kansas helps to explain the relatively slow growth rate of the Kansas economy. Professor Peter Orazem presented the evidence for this viewpoint in the Fall 2004 issue of this journal.1 The following analysis strives to further illuminate Orazem’s findings by evaluating economic growth and productivity trends among the Plains states and the different regions of Kansas.
Economic growth is defined as a sustained increase, over a period of time, in the material goods and services produced within a specified geographic region. This definition allows for two basic sources of growth: (1) the number of people that produce and (2) the efficiency with which the people produce over a given time period; that is, their labor productivity. The two elements can change at different rates.
The labor productivity component of economic growth drives labor compensation levels, and thereby the average standard of living within an economy. Businesses cannot, on a sustained basis, pay workers more than the value of what they produce. Therefore, compensation levels should closely track increases in the average monetary value of output per worker. For the state of Kansas, over the past quarter-century, the relationship between the dollar value of output per worker and compensation per worker is nearly exact, having a statistical correlation of 98 percent.2
From the perspective of public policy analysis, people involved with the policy making process should distinguish between the terms “economic growth” and “economic development,” because people mistakenly use the terms interchangeably. Economic growth has a concise, measurable definition. Economic development has a more amorphous meaning. Simply stated, economic development constitutes the many interrelated economic processes that culminate in economic growth—particularly the component of growth driven by improved labor productivity.
A process known as “capital deepening” defines the core aspect of economic development. Capital deepening simply refers to the capital intensity of the production processes within an economy. However, many complex economic phenomena underlie that simple meaning—phenomena associated with increasing rates of technological innovation and technological diffusion, increasing degrees of production specialization (including the manufacture of production capital itself), increasing organizational complexity, and increasing levels of relevant know-how within the workforce. These phenomena must come together in a mutually-reinforcing way on the front lines of individual businesses—usually through a risky process of trial and error—before economic development manifests itself as productivity-driven economic growth.
Table 1 reports estimates of the employment and productivity components of economic (output) growth for the Plains states and the United States, using two different data sets for comparison. The comparison reveals the differences between the best available data to use for economic growth and productivity analysis (gross state product) and that which this study must use to approximate economic growth and productivity at the sub-state level (total compensation).3 Note that output growth is composed of the sum of its employment and productivity components.
The gross-state-product panel of Table 1 shows that Kansas has the 4th highest growth of output among the seven Plains states (a region defined by the U.S. Bureau of Economic Analysis). A more in-depth study of the data reveals that Kansas ranked 6th among the Plains states in terms of the share of output growth attributable to labor productivity growth. That implies, conversely, that Kansas ranks second among the Plains states in terms of the share of output attributable to employment growth. (Note that the share of output growth attributable to productivity is a different metric than productivity growth itself. From 1977 to 2003, labor productivity in Kansas grew 26 percent; that growth rate ranked 6th among the Plains states. Among the Plains states as a group, productivity grew 30 percent. South Dakota, the top-ranking state in the Plains, experienced productivity growth of 38 percent.)
The total-compensation panel of Table 1 shows that employee-related compensation offers a reasonable, yet imperfect, approximation for measuring economic growth. Using this set of metrics, Kansas ranks 5th in terms of both total output growth and the share of output growth attributable to productivity growth. Orazem has shown that productivity growth and per-worker compensation growth deviate from one another, despite their strong statistical correlation. For example, in Kansas, compensation has grown $0.75 for every $1.00 that productivity has grown; for the Plains as a group, compensation has grown $0.66 for every $1.00 that productivity has grown.4
Any analysis of growth trends must contend with the influence that the start- and end-points have on the growth calculation. Table 1 begins in 1977 because that is the earliest date in which consistent gross state product data is available. Using total compensation data, which goes back to 1969, Kansas would have a much higher aggregate growth rate— greater than both the Plains and the United States—but it would rank 4th among the Plains states. Minnesota and South Dakota would remain the growth leaders in the Plains. (Table A, in the Appendix, reports for the 1993-to-2003 time period, the same data reported in Table 1.)
Figure 1, which uses per-worker total compensation as a proxy for productivity, provides more detail related to the productivity component of Table 1, for the selected regions. The juxtaposition of the curves on Figure 1 (post-1977) offer an extremely close approximation to the picture that would result from using gross state product data, reinforcing the near-perfect statistical correlation between per-worker compensation and per-worker productivity. The different levels of the curves represent approximations of the market-determined dollar-value difference of the per-worker output in the different regions. The slopes of the curves represent approximations of the growth rate of labor productivity in the different regions. (The statistics reported in parenthesis in the legend of Figure 1 indicate, in percentage terms, the total and average annual growth of productivity, as measured by per-worker compensation. For example, from 1969 to 2003, Kansas experienced estimated productivity growth of 55 percent, which translates into an average annual growth rate of 1.3 percent.)
Figure 1 illustrates three noteworthy features of Kansas’ economic history. First, Kansas experienced relatively strong productivity growth during the 1970s. Good times in the oil extraction and marketing business drove a lot of this growth. (The same forces also help explain the 1970s productivity growth in the Dakotas.) Second, the economic dominance of Johnson County began to manifest itself in 1976—and then in earnest following the national economic recessions of the early 1980s. Third, something significant appears to have happened in 1986 that arrested productivity growth in Kansas, relative to other Plains states, for the next decade. Only in 1996, does Kansas productivity (as measured here) begin to grow at rates greater than some of the other Plains states. A relative productivity growth lag of this duration has had material consequences for the competitiveness of Kansas businesses and the level of compensation provided to Kansas workers.
Figure 1 also helps frame an interesting question related to the history of economic growth in the Plains states: Why have Minnesota and South Dakota experienced productivity growth rates above the national average while all other Plains states have experienced growth rates below—except for Nebraska, significantly below—the national average?
This question has particular relevance in the context of a tenet of economic growth theory known as “convergence.” Economists distinguish between “absolute convergence” and “conditional convergence.”
Absolute convergence predicts that per-worker incomes—and per-worker income growth rates— among regions will converge to similar levels. In particular, absolute convergence predicts that economies with lower per-worker incomes will grow faster than economies with higher per-worker incomes. In a world (or country) that allows for the free mobility of capital and labor, incomes should converge to a common trend as investors deploy capital to areas that have the highest rates of investment return (which is, in part, a function of wage rates) and people move to areas that offer the highest compensation (which is, in part, a function of capital deployment).
Conditional convergence contemplates situations, both natural and man-made, that may not allow absolute convergence to work. It predicts that economies will grow faster the farther away they are from their “natural” level of economic activity—that is, they quickly catch up to where the “should be,” given their combination of natural, man-made, and human resources, once transitory or institutional impediments to growth are removed (or mitigated). Public policies (or other types of shocks, like natural disasters) offer an array of forces within an economy that may make it different from otherwise similarly situated economies. Correcting economic policies that deter the economic development process can allow an economy to accelerate its economic growth to a level that is closer to its “natural” level.
Economic research provides support for absolute convergence, both internationally and among the U.S. states. Absolute convergence tends to show up empirically among the more similarly situated economies, like the U.S. states or the developed economies of Europe. Dissimilar economies—like the industrialized economies and the underdeveloped economies of the world— tend not to demonstrate empirical patterns of absolute convergence. Instead, dissimilar economies tend to show economic growth patterns more consistent with conditional convergence.5
Among the U.S. states, there existed a strong tendency toward absolute convergence until the 1970s. Since then, disparity in per-worker incomes (and growth rates) has tended to persist.6 The productivity growth patterns of Minnesota and South Dakota, illustrated in Figure 1, underscore this modern tendency. The concept of absolute convergence would predict the relatively fast productivity growth rate of South Dakota, since it began the 1970s with the lowest per-worker compensation in the Plains region (and 3rd lowest in the nation); it would not necessarily predict the strong growth of Minnesota, the state with the highest per worker compensation in the Plains (and 25th highest in the nation).
Both structural issues and policy regimes may have made the notion of conditional convergence more relevant to the growth patterns of the U.S. states. A recent academic study investigated the influence of state and local policies on economic growth by pulling together three isolated strands of research into a unified framework. The study evaluated growth patterns across the states from 1979 to 1997. Recalling that capital deepening (as discussed above) is a key economic development process that influences productivity growth, the relevant finding of the study for this discussion is that “state and local policies have a more profound influence on the private capital-to-labor ratio in a region than on private output.”7 In addition, the notion of conditional convergence may help explain why another well-crafted study focused on state tax policy, covering the period 1960-1992, concluded that “it appears that state and local taxes have temporary growth effects that are stronger over shorter intervals and a permanent growth effect that does not die out over time, at least for the sample considered.”8
Two major tax policy events occurred in Kansas in the mid-1980s, one federal and one state-specific. One (or both) of these events may have provided a shock to the Kansas economy that helps explain the decade-long stagnation of productivity growth, which started in 1986 (see Figure 1). The federal event was the Tax Reform Act of 1986, the largest change to the federal tax code since 1954. The Kansas event was the 1985 legislation that ordered reappraisal of all property for property taxes purposes (effective 1989) and presented to Kansas voters an amendment to the state constitution that created a brand new property classification system; the amendment passed in November of 1986. Major tax policy changes like these tend to postpone investment activity while taxpayers assess the implications and wait for certainty on the outcomes.9
The extent to which these tax policy changes had an affect on capital deepening in Kansas requires more in-depth research. However, Figure 2 provides readily-accessible information that offers some clues. Figure 2 illustrates the rate of new business starts (sole proprietorships and partnerships) for the regions shown in Figure 1. New business starts offer one proxy for assessing the relative attractiveness of Kansas, relative to other states, as a place to invest and take business risks.
The cursory evidence provided by Figures 1 and 2 suggest that (1) the federal Tax Reform Act of 1986 may have had a transitory effect on investment-related activity in Kansas and (2) the Kansas-specific property tax changes may have had a more enduring effect on investment related activity in Kansas. Similar to the patterns shown in Figure 1, Figure 2 indicates that Kansas experienced significant discontinuities in the rate of new business formation about 1986. Related to the 1985 Kansas legislation, only three other states besides Kansas showed a dip in new business formation between 1985 and 1986. Relating to the federal legislation, notable discontinuities (both positive and negative) occurred across many states between 1986 and 1987; most of the positive spikes occurred in Midwestern states. Related to the possible enduring effect of the 1985 Kansas legislation, from 1988 to 1994 Kansas ranked 49th among the states (just ahead of Oklahoma) in terms of the average annual growth rate of new businesses; from 1995 to 2003, Kansas ranks 45th. These rankings represent a substantial drop from pre-1985 growth rates (34th from 1969 to 1985; 22nd from 1980 to 1985).
The Kansas Department of Commerce uses the regions delineated in Figure 3 to administer and track economic development-related initiatives. These regions will inform the regional analysis.
From an economic development perspective, the boundaries of the map in Figure 3 have a somewhat arbitrary demarcation (e.g., the separation of Salina- McPherson and Topeka-Lawrence). It makes more sense to think of economic development in terms of concentric rings around population centers. Population density tends to promote productivity growth.10 People commonly remark that cities have higher wages because cities have a higher cost of living. However, the economic causality runs in the opposite direction; cities tend to have a higher cost of living because they nurture the productivity gains that allow for higher pay, which, in turn, allows people to bid up the price of real estate and other amenities close to the center of economic activity. The East Central (EC) region tends to most closely approximate the ring around a population center—the Kansas half of Kansas City.
The East Central region drives the economic growth of Kansas, and Johnson County drives the East Central region.11 Table 1 and Figures 1 and 2 provide Kansas economic growth statistics with and without Johnson County included. Over the past three decades, among peer counties (those with a population of 225,000 or more in 1973), Johnson County ranks 3rd in terms of employment growth and just inside the top quartile in terms of (estimated) productivity growth, which has made it competitive with fast-growing counties like Fairfax, Virginia (Washington, D.C.); Travis, Texas (Austin); DuPage, Illinois (Chicago); and Cobb, Georgia (Atlanta). (Fairfax and Travis also rank in the top-10 with regard to estimated productivity growth.)
The fact that a major metropolitan area like Kansas City straddles a state border raises questions about how the Kansas side of the border compares to the Missouri side in terms of economic performance.12 Table 2 provides some cursory answers. It lists estimated economic growth components for Jackson County, Missouri (home of downtown Kansas City) and the counties contiguous to both Jackson County and the Kansas border (plus Leavenworth County, Kansas). In terms of aggregate economic growth, Johnson County, Kansas is the clear leader, from 1969 to 2003. Cass County, Missouri, which is contiguous to Johnson County, took the lead in the last decade. However, Cass County has a small economic base, so relatively small amounts of absolute growth register as relatively high percentage changes.
The more pointed analysis in Table 2 comes from evaluating how each county has grown. The most striking aspect of the data is how little employment growth has contributed to overall economic growth in Wyandotte County, Kansas and Jackson County, Missouri. The experience in Leavenworth County, Kansas is only moderately better. Productivity growth has driven most of the economic growth experienced in these three counties. Platte County, Missouri and Johnson County, Kansas have had the opposite experience: employment growth accounts for substantially more than half of their overall economic growth.
The phenomenon of Johnson County is that it has experienced strong productivity growth in combination with stellar employment growth. The data in Table 3 augments the data in Table 2 to show the estimated productivity growth rates driving the productivity share of growth. Leavenworth experienced better productivity growth than Johnson over the 1969-to-2003 period. Clay and Cass experienced better productivity growth than Johnson over the 1993-to-2003 period. Note, however, that Johnson started both time periods with a significantly larger economic base than every county in Tables 2 and 3 except Jackson.
Table 4 presents estimates of economic growth-component data for select Kansas counties. It considers two sets of counties. The first set includes the most populous county in each Kansas Department of Commerce Economic Development region. The second set includes the counties that experienced the aggregate economic growth in percentage terms, as measured by total compensation, from 1969 to 2003. Some regions have only one county listed (those counties listed in bold type), because the county has both the largest population and the highest estimated aggregate output growth.
The top portion of Table 4 reports the components of growth for the most populous county in each region. Given the discussions above about the importance of population centers as growth of Sedgwick (Wichita) and Shawnee (Topeka) creates the headline story for Table 4. These are the second and third most populous counties in Kansas, yet their estimated output and employment growth fall significantly below the state average. Thirty years ago, Sedgwick had about 87,000 more wage-and-salary jobs than Johnson. With a 30-year employment growth rate of more than five times that of Sedgwick, Johnson took the top rank for number of jobs in 1996.
Crawford and Ellis help tell another important Kansas economic story, one hidden underneath the statistics in Table 4. The South East region and the North West region, respectively, represent the areas of Kansas with the slowest growth of estimated aggregate output. Both regions have depopulated over the past three decades. Crawford has experienced population growth of only one percent from 1969 to 2003; Ellis experienced 11 percent population growth. Yet Crawford experienced wage-and-salary job growth of 69 percent over the same time period and Ellis experienced job growth of 141 percent. These counties have extremely high labor force participation rates. Both the nation and the Plains region have also experienced increasing labor force participation rates. However, Kansas has among the highest (and historically fastest growing) labor force participation rates among the 50 states. The combination of a saturated labor market and lagging productivity growth may impose an obstacle to future economic growth in Kansas.
As the top panel of Table 4 clearly shows, over the 1969-to-2003 time period, Shawnee and Sedgwick, respectively, have had the highest amount of growth attributable to productivity growth (or, conversely, the least amount of economic growth attributable to employment growth). However, over the 1993-to-2003 decade, Finney and Riley took the top-two spots, with Shawnee and Sedgwick dropping to 3rd and 4th; and Johnson dropping from 4th to 7th.
The bottom panel of Table 4, combined with the counties in bold text in the top panel, lists the county in each region that has experienced the greatest aggregate output growth over the 1969-to-2003 period. Among these counties, Coffey and McPherson, respectively, experienced the largest share of economic growth attributable to productivity growth. The construction and subsequent operation of the Wolf Creek nuclear power plant drove the growth in Coffey. Over the 1993-to-2003 period, aggregate growth in Coffey fell behind all other counties except Finney. Additionally, Finney and Coffey had the highest share of economic growth attributable to productivity growth over the past decade.
As discussed above, the share of output growth attributable to productivity growth is not the same thing as productivity growth itself. Figure 4 charts the relative trends in estimated productivity growth for the most populous counties in each region. (The numerals in parenthesis listed in the legend of Figure 4 represent, respectively, the total percentage growth and the average annual percentage growth from 1969 to 2003.) Johnson, Sedgwick, and Shawnee had relatively similar levels of labor productivity in 1969, higher than the average level in the Plains states. However, Shawnee and Sedgwick have experienced slower productivity growth than the Plains average, while Johnson has experienced productivity growth substantially greater than the Plains average. In fact, Johnson County surpassed the average U.S. level of productivity (as measured here) in 2000.
As with Table 4, the most significant features of Figure 4 concern the trends in Sedgwick and Shawnee. Among the most populous counties over the 1969-to-2003 period, Sedgwick experienced greater productivity growth than all counties listed but Johnson. However, it fell to 6th place over the 1993-to-2003 period. Shawnee fell from 4th place to 7th place. Somehow two of the most populous counties in Kansas failed to catch the productivity wave that swept across the U.S. during the 1990s.
Fortunately, a few Kansas counties caught the 1990s productivity wave. Riley experienced better productivity growth than Johnson over the 1993-to-2003 period. Among the fast-growing counties, Jackson (KS) and Coffey experienced better productivity growth than the Plains average during the 1990s. Over the 1969-to-2003 period, in addition to Johnson, Coffey, McPherson and Pottawatomie experienced greater productivity growth than the Plains average (see Figure A in the Appendix).
The Kansas economy has experienced significant productivity-driven economic growth in a few of its regions. Overall, however, most of the regions have lagged behind the Plains average. Professor Orazem has suggested that the relatively low population density of many Kansas regions may create a natural disadvantage to the quest for productivity-driven economic growth. Yet two of Kansas’ most densely populated regions have also experienced relatively poor productivity growth. It is worth exploring whether the overall policy mix in Kansas is reinforcing or counteracting the natural growth disadvantage associated with low population density. It is also worth exploring the degree to which the overall policy mix in Kansas unnecessarily inhibits the economic development process of capital deepening.
Table A - Components of Economic Growth in the Plains States and United States, 1993-2003
Figure A - Estimated Productivity Trends Among the Counties in Each Kansas Region with the Greatest Percentage Growth in Output, 1969-2003
1. Peter F. Orazem, “Slow Growth and the Kansas Productivity Puzzle,” Policy Research Institute, University of Kansas, Kansas Policy Review, Vol. 26 (2), Fall 2004. Note that Orazem ranked Kansas 37th in economic growth, using data available at the time. (http://www.ku.edu/pri/publicat/kpr/kprV26N2/kprv26n2.pdf)
2. Ibid., p. 3.
3. The most appropriate data to use for measuring economic growth and productivity is gross state product (the state equivalent of gross domestic product), because that metric strives to allocate corporate profits, and other measures of business value-added, to their proper geographic location. In the left-hand panel of Table 1, productivity is measured by dividing gross state product by the total number of workers (including self-employed people). In 2003, compensation of employees equaled about 57.5 percent of gross state product. The other components of gross state product relate to business profits and additional measures of business value added—important aspects of productivity measurement.
No analog to gross state product currently exists for counties, and county-level data is required for the regional analysis goals of this inquiry. For purposes of measuring productivity, using wage and salary disbursements offers the next best metric to gross state product. The estimates in the second panel of Table 1 use total wage and salary compensation (including voluntary and government-mandated employer-paid benefits) as a proxy for total output, wage and salary jobs as a proxy for employment, and total wage and salary compensation per wage and salary job as a proxy for labor productivity.
4. Orazem, “Slow Growth and the Kansas Productivity Puzzle,” p. 3.
5. Robert J. Barro and Xavier Sala-i-Martin, Economic Growth(Cambridge, MA: MIT Press, 1999.), Chapters 1, 10, and 11.
6. W. Mark Crain, Volatile States: Institutions, Policy, and the Performance of American State Economies (Ann Arbor, MI: University of Michigan Press, 2003), Chapter 2.
7. Stephen P.A. Brown, Kathy J. Hayes and Lori L. Taylor, “State and Local Policy, Factor Markets, and Regional Growth,” The Review of Regional Studies, Vol. 33(1), 2003, p. 41. (http://economy.okstate.edu/rrs/issue.asp?volume=33&issue=1)
8. Zolt Becsi, “Do State and Local Taxes Affect Relative State Growth?” Federal Reserve Bank of Atlanta, Economic Review, Vol. 81 (2), March/April 1996, p. 34). (http://www.frbatlanta.org/filelegacydocs/ACFD5.pdf)
9. Arthur P. Hall, “The Cost of Unstable Tax Laws,” Tax Foundation Special Report No. 41, October 1, 1994. (http://www.taxfoundation.org/files/ 79985a8649fa8a5d501dec544dc13960.pdf)
10. Peter F. Orazem, “The Growth of Cities and Rural Economic Development,” Center for Applied Economics, University of Kansas School of Business, Technical Brief 04-1119, November 2004. (http://cae.business.ku.edu/gen/cae_generated_bin/documents/basic_module/CAE-GrowthCities%20-%20final.pdf)
11. For an analysis of the KDOC regions, see Arthur P. Hall and Peter F. Orazem, “Long-Term Economic Trends in the Regions of Kansas, 1969-2003,” Kansas Inc. Research Report, August 2005. (http://www.kansasinc.org/pubs/working/KS%20Region%20Trends.pdf)
12. For a more in-depth analysis of economic trends along the Kansas- Missouri border, see Arthur P. Hall and Peter F. Orazem, “Economic Trends Along the Kansas-Missouri Border, 1969-2003,” Kansas Inc. Research Report, August 2005. (http://www.kansasinc.org/pubs/working/KS-MO%20Border.pdf)
Arthur Hall is the executive director, the Center for Applied Economics, School of Business, the University of Kansas.