Inequality & Unemployment, Redistribution & Social Insurance, and Participation: A Theoretical Model and an Empirical System of Endogenous Equations

نویسندگان

  • Robert J. Franzese
  • Jude C. Hays
  • Kenichi Ariga
  • Nam Kyu Kim
چکیده

Conflicts of interest over the generosity and structure of redistribution and social insurance (call these jointly: social policy) include that between the relatively poor and wealthy—which theoretically produces the famous median-voter result that democratic demand for broad redistribution increases in the income skew—and that between the safely employed and the unemployed and precariously employed—which yields a different theoretical result, namely that inequality reduces median-voter demand for social insurance. In each case, the generosity and structure of social policy may itself affect simultaneously the efficiency of the labor market and the political participation of society’s less fortunate, which affects the identity and so the income and job-security status of the median voter. These considerations imply several endogenous relationships between economic performance (employment/income level and distribution), social policy (redistribution and social insurance), and political participation. This paper will elaborate the theoretically expected nature of these endogenous relationships, suggest identification conditions that derive from the theory and substance, and offer empirical estimates of the resulting system of equations. * Franzese acknowledges support from NSF grant #0340195. We both thank Carrie Steele at Illinois, and Kenichi Ariga, Nam Kyu Kim, and Joel Simmons at Michigan for invaluable research assistance in assembling the data and Chris Anderson, Pablo Beramendi, Sara Hobolt, Xiaobo Lu, and Stephanie Rickard for helpful comments on earlier drafts. Page 1 of 39 Inequality & Unemployment, Redistribution & Social Insurance, and Participation: A Theoretical Model and an Empirical System of Endogenous Equations ABSTRACT: Conflicts of interest over the generosity and structure of redistribution and social insurance (call these jointly: social policy) include that between the relatively poor and wealthy—which theoretically produces the famous median-voter result that democratic demand for broad redistribution increases in the income skew—and that between the safely employed and the unemployed and precariously employed—which yields a different theoretical result, namely that inequality reduces median-voter demand for social insurance. In each case, the generosity and structure of social policy may itself affect simultaneously the efficiency of the labor market and the political participation of society’s less fortunate, which affects the identity and so the income and job-security status of the median voter. These considerations imply several endogenous relationships between economic performance (employment/income level and distribution), social policy (redistribution and social insurance), and political participation. This paper will elaborate the theoretically expected nature of these endogenous relationships, suggest identification conditions that derive from the theory and substance, and offer empirical estimates of the resulting system of equations. Conflicts of interest over the generosity and structure of redistribution and social insurance (call these jointly: social policy) include that between the relatively poor and wealthy—which theoretically produces the famous median-voter result that democratic demand for broad redistribution increases in the income skew—and that between the safely employed and the unemployed and precariously employed—which yields a different theoretical result, namely that inequality reduces median-voter demand for social insurance. In each case, the generosity and structure of social policy may itself affect simultaneously the efficiency of the labor market and the political participation of society’s less fortunate, which affects the identity and so the income and job-security status of the median voter. These considerations imply several endogenous relationships between economic performance (employment/income level and distribution), social policy (redistribution and social insurance), and political participation. This paper will elaborate the theoretically expected nature of these endogenous relationships, suggest identification conditions that derive from the theory and substance, and offer empirical estimates of the resulting system of equations. I. Theoretical Models of Democratic Policy with Income, Employment, and Participation Inequality We begin with a reconsideration of the Moene-Wallerstein (2001) (M&W) model of unemployment, inequality, and the democratic demand for redistribution and social insurance, which offers a very useful formulation for considering jointly the redistributive and insurance motivations for social policies and the effects of inequality thereupon. First, we illustrate and discuss the implications of the M&W model for the expected relationships between inequality, unemployment, redistribution, and social insurance. Then, we extend the model discussion to consider median-preserving increases in income inequality (as explained below), continuous heterogeneity in income (or wages, or wealth), and, more crucially, heterogeneity in unemployment rates/risk correlated with income. Penultimately, we consider the effects of incomplete and heterogeneous political participation (voter abstention, weightier political input from some than others) that, critically, also correlates with economic status. Finally, we acknowledge and grapple with the endogeneity of all of the aspects of this expanded theoretical model: inequality (skew) and unemployment (risk), redistributive and social-insurance policies, and political participation. A. The Moene-Wallerstein Model of Redistribution & Social Insurance with Exogenous Unemployment & Inequality The M&W model includes population shares, σ0 of permanently unemployed, σH of high-income (wH) earners who face no appreciable employment risk, and σL of low-income (wL) workers who face appreciable risk, α, of losing their income source (job). Job-losers have probability β of regaining 1 All notation and equation numbering given here exactly follows M&W to facilitate comparison. 2 As the authors note (and as illustrated below), high-income earners can face unemployment risk without qualitative (but with quantitative) change to the conclusions, provided unemployment risk remains weakly negatively correlated with income. Page 2 of 39 employment, giving steady-state employed, e, and unemployed, u, population-shares, L H e σ σ β α β + + = and 0 1 L u e α α β σ σ + ≡ − = + , respectively. 3 Governments collect revenues by a flat tax, t, which generates revenues and expenditures (i.e., all revenues are spent and no borrowing) of [1] ( ) w e t T τ = , where [ ] L L H H e w w w σ σ β α+ + ≡ 1 is the average wage and ( ) t τ is revenue as a share of earnings (i.e., the average tax-rate), incorporating deadweight losses in that this revenue function is strictly concave (i.e., deadweight losses rise at increasing rates as tax rates rise) and with ( ) 0 0 τ ′ = and ( ) ( ) 0 1 0 τ τ = = (i.e., no deadweight cost if no tax, and no revenue if tax rates are zero or one). Finally, a share, γ, of revenues (cum expenditures), T, goes to current earners, with the remaining (1γ)T going to the unemployed. Individuals currently with and without jobs will thus have net income (cum consumption) of [2] ( ) ( ) e t T i E w t c γ + − = 1 and [3] ( ) ( ) e t T N c − − = 1 1 γ , respectively. M&W (2001: 862) assume utilities, u(c), are concave in consumption ( ( ) ( ) 0 , 0 < ′ ′ > ′ c u c u ), which implies some risk aversion, and satisfy other conditions assuring that insurance is a normal good (coefficient of relative risk aversion exceeds one: ( ) ( ) 1 > ′ ′ ′ − c u c u c ) and is demanded in some amount if risk is non-negligible ( 0 ) ( → ∞ → ′ c as c u ). Given all this, the lifetime utility of a currently employed low-income worker—the group which is, very importantly and highly plausibly, assumed also to be the median (i.e., to contain the median voter)—is a simple weighted average of that type’s employed and unemployed utility: ( ) ( ) [ ] ( ) [ ] N L E c u r w c u r r v ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ + + + ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ + + + = β α α β α β [6]. The weights simply and intuitively reflect the shares of life spent employed and unemployed, with the former weighted and the latter discounted by the time discount-rate, r, since employment is the current state. Inserting [2] and [3] explicitly into [6], yielding [6'], and then [1] explicitly into the result, yielding 3 The job-loss and job-finding rates are instantaneous. M&W solve the resulting continuous-time dynamic model for steadystate equilibria from which they derive their comparative statics. I will discuss only these steady states and so will skip their explicit derivation, expressing the model instead in simpler, static terms (without further loss of content or generality). 4 Empirical estimates generally suggest coefficients of relative risk-aversion of μ≥1. Log utility, which has constant relative risk-aversion μ=1, would not satisfy this, but other functions in the class of constant relative risk-aversion could. 5 ...and assuming infinitely-lived actors, but relaxation of this assumption adds only notational (and actuarial) complexity, requiring replacement of r with rH r − −exp 1 , with H the actor’s life expectancy, in the asset equation that produces [6] below. This would add little relevant substantive content, although it could possibly introduce a social life-insurance motivation. Page 3 of 39 [6"], will facilitate illustration and discussion of the model’s implications: ( ) ( ) { } [ ] ( ) ( ) ( ) [ ] e t T e t T L u r w t u r r v − − ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ + + + + − ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ + + + = 1 1 1 γ γ β α α β α β [6'] ( ) ( ) [ ] { } [ ] ( ) ( ) [ ] ( ) [ ] e w e t e w e t L u r w t u r r v − − ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ + + + + − ⎟⎟ ⎠ ⎞ ⎜⎜ ⎝ ⎛ + + + = 1 1 1 τ γ τ γ β α α β α β [6"] M&W designed this model to analyze the democratic choice of both the generosity or size of the system, captured in t, and the degree of welfare-targeting of its benefits, captured in the parameter (1-γ), which gives the share of spending on those without jobs. Consider the utility effects for the median voter, a currently employed worker of type L, of increasing t in this model. Starting with the first term in the first square brackets of [6"], this tax increase costs the worker some take-home wages, as reflected in the -twL term, but returns to her some public transfers insofar as (i) public spending accrues to the employed, γ; (ii) taxable income is not lost to the deadweight inefficiencies of taxation incorporated in ( ) t τ , and (iii) the source of revenue, average wages, w , are high, all combined as reflected in the ( )w t γτ + term. This is the usual motivation for and tradeoff in redistribution; the more skewed is the income (here: wage) distribution, i.e., the higher is w relative to wL, the more redistribution the median voter (an employed type L) would like by this motivation. However, as reflected in the r r + + + β α β term, these considerations matter in this model only in proportion to the share of time the worker expects to stay employed (the β α β + part of that term), and that she discounts the future (the +r parts). To the remaining proportion, r + +β α α , in which time the median expects to be unemployed, she will derive utility from the second half of [6"]. There, the increased taxes bring higher utility in those times when she is without a job in the future, which she expects to be β α+ of the time and discounts by r. In these times, to the degree public expenditures go to the unemployed, 1-γ, she will receive her equal share of total revenues, ( ) w e t τ , as divided among the unemployed, e − 1 , yielding the ( ) ( ) e w e t − − 1 1 τ γ in transfers represented in the second squarebracket term. This provides the insurance motivation and demand for welfare-targeted redistribution. 6 A few other implications surrounding the first, second, and cross (with w ) derivatives of τ(t) emerge as well (see Franzese 2002, ch. 2). For example, the greater the deadweight losses (the more concave is τ(t)), the less redistribution the median voter seeks by this redistributive motivation (or, indeed, by the insurance motivation also). Page 4 of 39 For our purposes, the most important results from the M&W model are these: • With exogenous targeting (γ fixed), a mean-preserving increase in income skew (i.e., a reduction in wL holding w fixed) increases the median voter’s preferred social-policy generosity (t*) if benefits go exclusively to the employed (γ=1) and reduces her preferred safety net if benefits go exclusively to the unemployed (γ=0). Thus, inequality increases (broad) redistribution whereas it decreases (targeted) social insurance. • With endogenous targeting (both γ and t* chosen democratically), a mean-preserving increase in income skew (i.e., wL declining with w fixed) increases the median voters’ preferred targeting of benefits to the employed (γ) and generally increases her preferred sum of insurance and redistributive spending, total social-spending, t*. Unconstrained desiredredistribution remains increasing in inequality, and unconstrained desired-insurance remains decreasing in inequality. With desired insurance spending rising with equality and desired overall spending declining with equality, however, at some point, all spending is welfaretargeted (to the unemployed: γ=0). At this point, wL=w0, these desires cross, and funding of the desired insurance becomes constraining. The results are: o A monotonic-positive relationship of equality to insurance spending, although with a kink at w0 and some flattening as equality continues to rise from there; o A weakly monotonic-negative relationship of equality to redistribution spending, being strictly negative through w0 but becoming zero at that and greater equality. o A non-monotonic relationship of equality to total social-spending (insurance + redistribution), with the sum declining as equality increases to w0, kinking there, and then rising as equality increases further. In the case where γi<1 so that the constraint is not binding (i.e., wL<w0), these implications of the full model are seen in the following first order conditions with respect to t and to γ, respectively: ( ) * L L w t w τ ′ = [14] 7 M&W consider three classes: permanently-unemployed, low-wage at-risk (L), and high-wage permanently-employed (H) workers. L is the median. Thus, we follow them in replacing subscript i with L in this section. 8 Median-voter equilibria do not generally obtain in more than one dimension: here, the choice of both t and γ; M&W also show, however, that if the policy choices are made sequentially (Shepsle’s SIE) or if the party system prevents coalitions of rich and poor v. middle, then the median voter (group), being median in both dimensions, remains determinant. 9 We have rewritten these two first-order conditions slightly to isolate further the implicit optimum choices on t and γ. Page 5 of 39 ( ) ( ) ⎠ ⎞ ⎜ ⎝ ⎛ + ⎟⎠ ⎞ ⎜⎝ ⎛ − = ′ ′ r e e c u c u

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

The Affect of Gender Inequality on Women Social Participation

This research was conducted to study the effect of gender inequality on social participation of women. The method used in this research was Ex-Post Facto and survey method. The statistical population included female students in different fields of faculty of humanities and faculty of training sciences and psychology (Islamic Azad University) Garmsar Branch in 2012-2013. Research instruments inc...

متن کامل

A Global Social Support System: What the International Community Could Learn From the United States’ National Basketball Association’s Scheme for Redistribution of New Talent

If global trade were fair, it is argued, then international aid would be unnecessary and inequalities inherent to the economic system would be justifiable. Here, we argue that while global trade is unfair, in part because richer countries set the rules, we believe that additional interventions must go beyond trade regulation and short-term aid to redress inequalities among countries that will p...

متن کامل

Verification of unemployment benefits’ claims using Classifier Combination method

Unemployment insurance is one of the most popular insurance types in the modern world. The Social Security Organization is responsible for checking the unemployment benefits of individuals supported by unemployment insurance. Hand-crafted evaluation of unemployment claims requires a big deal of time and money. Data mining and machine learning as two efficient tools for data analysis can assist ...

متن کامل

Analyzing the Causal Relationships between Economic Growth, Income Inequality, and Transmission Channels: New Empirical Evidences from Iran

This paper investigates causal relations between economic growth, income inequality, and transmission channels during the period 1972 to 2016. These channels include saving rate, investment rate, redistribution policies, human capital, and conspicuous consumption. There is no strong evidence that supports uni-directional or bi-directional causality. In addition, some of the transmission channel...

متن کامل

The Dynamics of Immigrant Welfare and Labor Market Behavior

The Dynamics of Immigrant Welfare and Labor Market Behavior This paper analyzes transitions into and out of 3 different labor market states, social assistance, unemployment and employment. We estimate a dynamic multinomial logit model, controlling for endogenous initial condition and unobserved heterogeneity, using a large representative Swedish panel data set, LINDA, for the years 1990 to 1996...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2007