Shift-share analysis

Last updated

A shift-share analysis, used in regional science, political economy, and urban studies, determines what portions of regional economic growth or decline can be attributed to national, economic industry, and regional factors. The analysis helps identify industries where a regional economy has competitive advantages over the larger economy. A shift-share analysis takes the change over time of an economic variable, such as employment, within industries of a regional economy, and divides that change into various components. A traditional shift-share analysis splits regional changes into just three components, but other models have evolved that expand the decomposition into additional components.

Contents

Overview

A shift-share analysis attempts to identify the sources of regional economic changes. The region can be a town, city, country, statistical area, state, or any other region of the country. The analysis examines changes in an economic variable, such as migration, a demographic statistic, firm growth, or firm formations, although employment is most commonly used. [1] [2] The shift-share analysis is performed on a set of economic industries, like those defined by the North American Industry Classification System (NAICS). The analysis separates the regional economic changes within each industry into different categories. Although there are different versions of a shift-share analysis, they all identify national, industry, and regional factors that influence the variable changes.

Traditional model

The traditional form of the shift-share analysis was developed by Daniel Creamer in the early 1940s, and was later formalized by Edgar S. Dunn in 1960. [2] Also known as the comparative static model, it examines changes in the economic variable between two years. Changes are calculated for each industry in the analysis, both regionally and nationally. Each regional change is decomposed into three components. [3]

  1. National growth effect is the portion of the change attributed to the total growth of the national economy. It equals the theoretical change in the regional variable had it increased by the same percentage as the national economy.
  2. Industry mix effect is the portion of the change attributed to the performance of the specific economic industry. It equals the theoretical change in the regional variable had it increased by the same percentage as the industry nationwide, minus the national growth effect.
  3. Local share effect is the portion of the change attributed to regional influences, and is the component of primary concern to regional analysts. [3] It equals the actual change in the regional variable, minus the previous two effects.

Formula

The regional change in the variable e within industry i between the two years t and t+n is defined as the sum of the three shift-share effects: national growth effect (NSi), industry mix effect (IMi), and local share effect (RSi). [4]

The beginning and ending values of the economic variable within a particular industry are eit and eit+n, respectively. Each of the three effects is defined as a percentage of the beginning value of the economic variable. [4]

The total percent change in the economic variable nationwide for all industries combined is G, while the national and regional industry-specific percent changes are Gi and gi, respectively.

These three equations substituted into the first equation yield the following expression (from where the decomposition starts), which simply says that the regional economic variable (for industry i) grows at the speed of the regional industry-specific percent change. Note that usually (in case of slow growth) 0 < gi < 1 and that gi refers to the whole period from t to t+n.

Example

As an example, a shift-share analysis might be utilized to examine changes in the construction industry of a state's economy over the past decade, using employment as the economic variable studied. Total national employment may have increased 5% over the decade, while national construction employment increased 8%. However, state construction employment decreased 2%, from 100,000 to 98,000 employees, for a net loss of 2,000 employees.

The national growth effect is equal to the beginning 100,000 employees, times the total national growth rate of 5%, for an increase in 5,000 employees. The shift-share analysis implies that state construction would have increased by 5,000 employees, had it followed the same trend as the overall national economy.

The industry mix effect is equal to the original 100,000 employees times the growth in the industry nationwide, which was 8%, minus the total national growth of 5%. This results in an increase in 3,000 employees (100,000 employees times 3%, which is the 8% industry growth minus the 5% total growth). The analysis implies that the state construction would have increased by another 3,000 employees had it followed the industry trends, because the construction industry nationwide performed better than the national economy overall.

The local share effect in this example is equal to the beginning 100,000 employees times the state construction employment growth rate of −2% (it is negative because of the loss of employees), minus the national construction growth rate of 8%. This results in 100,000 employees times -10%, for a loss of 10,000 employees. However, the actual employment loss was only 2,000 employees, but that equals the sum of the three effects (5,000 gain + 3,000 gain + 10,000 loss). The analysis implies that local factors lead to a decrease in 10,000 employees in the state construction industry, because the growth in both the national economy and the construction industry should have increased state construction employment by 8,000 employees (the 5,000 national share effect plus the 3,000 industry mix effect).

Names and regions

Shift-share analysts sometimes use different labels for the three effects, although the calculations are the same. National growth effect may be referred to as national share. [4] [5] Industry mix effect may be referred to as proportional shift. [5] Local share effect may be referred to as differential shift, [3] regional shift, [4] or competitive share. [6]

In most shift-share analyses, the regional economy is compared to the national economy. However, the techniques may be used to compare any two regions (e.g., comparing a county to its state). [7]

Dynamic model

In 1988, Richard Barff and Prentice Knight, III, published the dynamic model shift-share analysis. [8] In contrast to the comparative static model, which only considers two years in its analysis (the beginning and ending years), the dynamic model utilizes every year in the study period. Although it requires much more data to perform the calculations, the dynamic model takes into account continuous changes in the three shift-share effects, so the results are less affected by the choice of starting and ending years. [8] The dynamic model is most useful when there are large differences between regional and national growth rates, or large changes in the regional industrial mix. [8]

The dynamic model uses the same techniques as the comparative static model, including the same three shift-share effects. However, in the dynamic model, a time-series of traditional shift-share calculations are performed, comparing each year to the previous year. The annual shift-share effects are then totaled together for the entire study period, resulting in the dynamic model's shift-share effects. [8]

Formula

The regional change in the variable e within industry i between the two years t and t+n is defined as the sum of the three shift-share effects: national growth effect (NSi), industry mix effect (IMi), and local share effect (RSi). [8]

If the study period ranges from year t to year t+n, then traditional shift-share effects are calculated for every year k, where k spans from t+1 to t+n. [8] The dynamic model shift-share effects are then calculated as the sum of the annual effects. [8]

The growth rates used in the calculations are annual rates, not growth from the beginning year in the study period, so the percent change from year k-1 to k in the economic variable nationwide for all industries combined is Gk, while the national and regional industry-specific percent changes are Gik and gik, respectively. [8]

Esteban-Marquillas Model

In 1972, J.M. Esteban-Marquillas extended the traditional model to address criticism that the regional share effect is correlated to the regional industrial mix. [9] In the Esteban-Marquillas model, the regional share effect itself is decomposed into two components, isolating a regional shift component that is not correlated to the industrial mix. [9] The model introduced a then-new concept to shift-share analyses, a homothetic level of the economic variable within an industry. This is the theoretical value of the variable within an industry assuming the region has the same industrial mix as the nation. [9]

In the Esteban-Marquillas model, the calculations of the national share and industrial mix effects are unchanged. However, the regional share effect in the traditional model is separated into two effects: a new regional share effect that is not dependent on the industrial mix, and an allocation effect that is. The allocation effect indicates the extent to which the region is specialized in those industries where it enjoys a competitive advantage. [9]

Formula

The regional change in the variable e within industry i between the two years t and t+n is defined as the sum of the four shift-share effects: national growth effect (NSi), industry mix effect (IMi), regional share effect (RSi), and allocation effect (ALi).

The beginning and ending values of the economic variable within a particular industry are eit and eit+n, respectively. The beginning value of the regional homothetic variable within a particular industry is hit. [9] It is based on the regional and national values of the economic variable across all industries, et and Et respectively, and the industry-specific national value Eit.

Each of the four shift-share effects is defined as a percentage of either the beginning value of the economic variable, the homothetic variable, or the difference of the two. [9]

The total percent change in the economic variable nationwide for all industries combined is G, while the national and regional industry-specific percent changes are Gi and gi, respectively.

Arcelus Model

In 1984, Francisco Arcelus built upon Esteban-Marquillas' use of the homothetic variables and extended the traditional model even further. [10] He used this method to decompose the national share and industrial mix effects into expected and differential components. The expected component is based on the homothetic level of the variable, and is the effect not attributed to the regional specializations. The differential component is the remaining effect, which is attributable to the regional industrial mix. [10]

Arcelus claimed that, even with the Esteban-Marquillas extension, the regional share effect is still related to the regional industry mix, and that the static model assumes all regional industries operate on a national market basis, focusing too heavily on the export markets and ignoring the local markets. [10] In order to address these issues, Arcelus used a different method for separating the regional share effect, resulting in a regional growth effect and a regional industry mix effect. Both of these are decomposed into expected and differential components using the homothetic variable. [10]

Formula

The regional change in the variable e within industry i between the two years t and t+n is defined as the sum of the eight shift-share effects: expected national growth effect (NSEi), differential national growth effect (NSDi), expected industry mix effect (IMEi), differential industry mix effect (IMDi), expected regional growth effect (RGEi), differential regional growth effect (RGDi), expected regional industry mix effect (RIEi), and differential regional industry mix effect (RIDi). [10]

The eight effects are related to the three traditional shift-share effects from the comparative static model. [10]

The homothetic variable is calculated the same as in the Esteban-Marquillas model. The beginning value of the regional homothetic variable within a particular industry is hit. It is based on the regional and national values of the economic variable across all industries, et and Et respectively, and the industry-specific national value Eit. [10]

Each of the eight shift-share effects is defined as a percentage of either the beginning value of the economic variable, the homothetic variable, or the difference of the two. [10]

The total percent changes in the economic variable nationally and regionally for all industries combined are G and g respectively, while the national and regional industry-specific percent changes are Gi and gi, respectively.

Related Research Articles

Analysis of variance (ANOVA) is a collection of statistical models and their associated estimation procedures used to analyze the differences among group means in a sample. ANOVA was developed by statistician and evolutionary biologist Ronald Fisher. The ANOVA is based on the law of total variance, where the observed variance in a particular variable is partitioned into components attributable to different sources of variation. In its simplest form, ANOVA provides a statistical test of whether two or more population means are equal, and therefore generalizes the t-test beyond two means.

Logistic function S-shaped curve

A logistic function or logistic curve is a common S-shaped curve with equation

Logistic regression Statistical model for a binary dependent variable

In statistics, the logistic model is used to model the probability of a certain class or event existing such as pass/fail, win/lose, alive/dead or healthy/sick. This can be extended to model several classes of events such as determining whether an image contains a cat, dog, lion, etc. Each object being detected in the image would be assigned a probability between 0 and 1 and the sum adding to one.

Analysis of covariance (ANCOVA) is a general linear model which blends ANOVA and regression. ANCOVA evaluates whether the means of a dependent variable (DV) are equal across levels of a categorical independent variable (IV) often called a treatment, while statistically controlling for the effects of other continuous variables that are not of primary interest, known as covariates (CV) or nuisance variables. Mathematically, ANCOVA decomposes the variance in the DV into variance explained by the CV(s), variance explained by the categorical IV, and residual variance. Intuitively, ANCOVA can be thought of as 'adjusting' the DV by the group means of the CV(s).

Production function physical output of a production process to physical inputs or factors of production

In economics, a production function gives the technological relation between quantities of physical inputs and quantities of output of goods. The production function is one of the key concepts of mainstream neoclassical theories, used to define marginal product and to distinguish allocative efficiency, a key focus of economics. One important purpose of the production function is to address allocative efficiency in the use of factor inputs in production and the resulting distribution of income to those factors, while abstracting away from the technological problems of achieving technical efficiency, as an engineer or professional manager might understand it.

Independent component analysis in signal processing, a computational method

In signal processing, independent component analysis (ICA) is a computational method for separating a multivariate signal into additive subcomponents. This is done by assuming that the subcomponents are non-Gaussian signals and that they are statistically independent from each other. ICA is a special case of blind source separation. A common example application is the "cocktail party problem" of listening in on one person's speech in a noisy room.

The accelerator effect in economics is a positive effect on private fixed investment of the growth of the market economy. Rising GDP implies that businesses in general see rising profits, increased sales and cash flow, and greater use of existing capacity. This usually implies that profit expectations and business confidence rise, encouraging businesses to build more factories and other buildings and to install more machinery. This may lead to further growth of the economy through the stimulation of consumer incomes and purchases, i.e., via the multiplier effect.

The Balassa–Samuelson effect, also known as Harrod–Balassa–Samuelson effect, the Ricardo–Viner–Harrod–Balassa–Samuelson–Penn–Bhagwati effect, or productivity biased purchasing power parity (PPP) is the tendency for consumer prices to be systematically higher in more developed countries than in less developed countries. This observation about the systematic differences in consumer prices is called the "Penn effect". The Balassa–Samuelson hypothesis is the proposition that this can be explained by the greater variation in productivity, between developed and less developed countries, in the traded goods' sectors than in the non-tradable sectors.

In economics, an input–output model is a quantitative economic model that represents the interdependencies between different sectors of a national economy or different regional economies. Wassily Leontief (1906–1999) is credited with developing this type of analysis and earned the Nobel Prize in Economics for his development of this model.

Land-use forecasting undertakes to project the distribution and intensity of trip generating activities in the urban area. In practice, land-use models are demand-driven, using as inputs the aggregate information on growth produced by an aggregate economic forecasting activity. Land-use estimates are inputs to the transportation planning process.

In statistics, multicollinearity is a phenomenon in which one predictor variable in a multiple regression model can be linearly predicted from the others with a substantial degree of accuracy. In this situation, the coefficient estimates of the multiple regression may change erratically in response to small changes in the model or the data. Multicollinearity does not reduce the predictive power or reliability of the model as a whole, at least within the sample data set; it only affects calculations regarding individual predictors. That is, a multivariate regression model with collinear predictors can indicate how well the entire bundle of predictors predicts the outcome variable, but it may not give valid results about any individual predictor, or about which predictors are redundant with respect to others.

The Solow residual is a number describing empirical productivity growth in an economy from year to year and decade to decade. Robert Solow, the Nobel Memorial Prize in Economic Sciences-winning economist, defined rising productivity as rising output with constant capital and labor input. It is a "residual" because it is the part of growth that is not accounted for by measures of capital accumulation or increased labor input. Increased physical throughput – i.e. environmental resources – is specifically excluded from the calculation; thus some portion of the residual can be ascribed to increased physical throughput. The example used is for the intracapital substitution of aluminium fixtures for steel during which the inputs do not alter. This differs in almost every other economic circumstance in which there are many other variables. The Solow Residual is procyclical and measures of it are now called the rate of growth of multifactor productivity or total factor productivity, though Solow (1957) did not use these terms.

The Solow–Swan model is an economic model of long-run economic growth set within the framework of neoclassical economics. It attempts to explain long-run economic growth by looking at capital accumulation, labor or population growth, and increases in productivity, commonly referred to as technological progress. At its core is a neoclassical (aggregate) production function, often specified to be of Cobb–Douglas type, which enables the model "to make contact with microeconomics". The model was developed independently by Robert Solow and Trevor Swan in 1956, and superseded the Keynesian Harrod–Domar model.

Articles in economics journals are usually classified according to the JEL classification codes, a system originated by the Journal of Economic Literature. The JEL is published quarterly by the American Economic Association (AEA) and contains survey articles and information on recently published books and dissertations. The AEA maintains EconLit, a searchable data base of citations for articles, books, reviews, dissertations, and working papers classified by JEL codes for the years from 1969. A recent addition to EconLit is indexing of economics-journal articles from 1886 to 1968 parallel to the print series Index of Economic Articles.

In statistics and econometrics, panel data or longitudinal data are multi-dimensional data involving measurements over time. Panel data contain observations of multiple phenomena obtained over multiple time periods for the same firms or individuals.

Fixed effects model statistical model that represents the observed quantities in terms of explanatory variables that are treated as if the quantities were non-random

In statistics, a fixed effects model is a statistical model in which the model parameters are fixed or non-random quantities. This is in contrast to random effects models and mixed models in which all or some of the model parameters are considered as random variables. In many applications including econometrics and biostatistics a fixed effects model refers to a regression model in which the group means are fixed (non-random) as opposed to a random effects model in which the group means are a random sample from a population. Generally, data can be grouped according to several observed factors. The group means could be modeled as fixed or random effects for each grouping. In a fixed effects model each group mean is a group-specific fixed quantity.

Mediation (statistics) mediation variable

In statistics, a mediation model seeks to identify and explain the mechanism or process that underlies an observed relationship between an independent variable and a dependent variable via the inclusion of a third hypothetical variable, known as a mediator variable. Rather than a direct causal relationship between the independent variable and the dependent variable, a mediation model proposes that the independent variable influences the (non-observable) mediator variable, which in turn influences the dependent variable. Thus, the mediator variable serves to clarify the nature of the relationship between the independent and dependent variables.

Marketing mix modeling (MMM) is statistical analysis such as multivariate regressions on sales and marketing time series data to estimate the impact of various marketing tactics on sales and then forecast the impact of future sets of tactics. It is often used to optimize advertising mix and promotional tactics with respect to sales revenue or profit.

Meta-regression is a tool used in meta-analysis to examine the impact of moderator variables on study effect size using regression-based techniques. Meta-regression is more effective at this task than are standard meta-analytic techniques.

Linear regression statistical approach for modeling the relationship between a scalar dependent variable and one or more explanatory variables

In statistics, linear regression is a linear approach to modeling the relationship between a scalar response and one or more explanatory variables. The case of one explanatory variable is called simple linear regression. For more than one explanatory variable, the process is called multiple linear regression. This term is distinct from multivariate linear regression, where multiple correlated dependent variables are predicted, rather than a single scalar variable.

References

  1. Cheng, Shaoming (2 February 2010). "Business cycle, industrial composition, or regional advantage? A decomposition analysis of new firm formation in the United States". The Annals of Regional Science. 47 (1): 147–167. doi:10.1007/s00168-009-0361-0.
  2. 1 2 Shi, Chun-Yun; Yang Yang (2008). "A Review of Shift-Share Analysis and its Application in Tourism". International Journal of Management Perspectives. 1 (1): 21–30.
  3. 1 2 3 Leigh, Nancey Green (2013). Planning Local Economic Development. Sage Publications. pp. 174–175. ISBN   9781452242590.
  4. 1 2 3 4 Stevens, Benjamin; Craig Moore (1980). "A critical review of the literature on shift-share as a forecasting technique". Journal of Regional Science. 20 (4): 419. doi:10.1111/j.1467-9787.1980.tb00660.x.
  5. 1 2 Knudesn, Daniel C. (2000). "Shift-share analysis: further examination of models for the description of economic change". Socio-Economic Planning Sciences. 34.
  6. "Georgia Statistics System". University of Georgia. Retrieved 24 October 2013.
  7. Michael LaFaive; James M. Hohman (31 August 2009). "The Michigan Economic Development Corporation: A Review and Analysis". Mackinac Center. Retrieved 5 December 2013.
  8. 1 2 3 4 5 6 7 8 Barff, Richard; Prentice L. Knight III (April 1988). "Dynamic Shift-Share Analysis". Growth and Change. 19 (2): 1–10. doi:10.1111/j.1468-2257.1988.tb00465.x.
  9. 1 2 3 4 5 6 Esteban-Marquillas, J.M. (1972). "A reinterpretation of shift-share analysis". Regional and Urban Economics. 2 (3): 249–261. doi:10.1016/0034-3331(72)90033-4.
  10. 1 2 3 4 5 6 7 8 Arcelus, Francisco (January 1984). "An extension of shift-share analysis". Growth and Change. 15 (1).