Skip to main content
SpringerLink
Log in

Estimating the intergenerational elasticity of expected income with short-run income measures: a generalized error-in-variables model

  • Published:
Empirical Economics Aims and scope Submit manuscript

Abstract

It has recently been argued that the intergenerational income elasticity (IGE) ubiquitously estimated in the economic mobility literature should be replaced by the IGE of expected income. This article advances a generalized error-in-variables model for the estimation of the latter IGE with short-run proxy measures of income and the Poisson pseudo-maximum likelihood estimator of constant elasticities. Empirical analyses with data from the panel study of income dynamics offer clear support for the account of lifecycle and attenuation biases the model provides. Together, the model and the associated empirical evidence supply a methodological justification for the estimation of the IGE of the expectation with proxy income variables that satisfy some conditions. This eliminates the main obstacle for making this IGE the workhorse elasticity of the economic mobility field, which would in turn dissolve the selection bias problem generated by the current expedient of dropping children with zero income from samples in order to make estimation of the conventional IGE possible.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others

Notes

  1. These formal results first appeared in earlier working-paper versions of this article (Mitnik 2017b, 2019), which circulated with exactly the same title as this article.

  2. The parameter \({\beta }_{1}\) is (also) the IGE of the expectation only when the error term satisfies very special conditions (Santos Silva and Tenreyro 2006; Petersen 2017; Wooldridge 2002:17).

  3. These selection biases are with respect to the true IGE of the conditional geometric mean of children’s income.

  4. Zeros are much less of a problem for short-run parental measures, as (a) they typically are multiyear averages (more on this later), (b) parents with zero income for extended periods of time are unlikely to be able to raise their kids themselves, and (c) there are good reasons to use as parental income the income of the family in which the child was raised, which is unlikely to be zero (see Hertz 2007:35). In addition, dropping parents with zero income is in principle unproblematic as it does not involve selecting on the dependent variable.

  5. Note that children with zero income or earnings can be unproblematically kept in the sample when estimating the IGE of expected income. This is particularly apparent when Eq. (3) is rewritten as \(Y=\mathrm{exp}\left({\alpha }_{0}+{\alpha }_{1}\mathrm{ln}X\right)+{\Psi },\) where \(E\left({\Psi }|\mathrm{ln}x\right)=0\).

  6. In the general case, the measurement errors are equal to \({\lambda }_{0t}+({\lambda }_{1t}-1)\mathrm{ln}Y+{V}_{t}\) (children) and \({\eta }_{0k}+({\eta }_{1k}-1)\mathrm{ln}X+{Q}_{k}\) (parents).

  7. The motivation for these assumptions is that, if they obtain, the GEiV model reduces to the classical error-in-variables model when \({\lambda }_{1t}={\eta }_{1k}=1\), i.e., when there is no lifecycle bias (see Haider and Solon 2006; Nybom and Stuhler 2016).

  8. In some cases, a multiyear average of the logarithm of parental income has been used, which implies that the geometric mean of parental income over those years is the proxy measure S.

  9. Nybom and Stuhler (2016) also found that lifecycle bias is very close to zero around age 40.

  10. PML estimators are consistent regardless of the actual distribution of the error term, provided that the mean function is correctly specified (Gourieroux et al. 1984).

  11. Santos Silva and Tenreyro (2006, 2011, Forthcoming) have argued that the PPML estimator should be preferred over other consistent estimators of constant-elasticity models of expected outcomes, as both theoretical arguments and simulations indicate that the NLLS estimator is inefficient (often to the point of being useless in empirical applications) and highly sensitive to outliers, whereas the gamma PLM estimator is very sensitive to measurement error. By contrast, the PPML estimator “is reliable in a wide variety of situations” and therefore “has the essential characteristics needed to make it the new workhorse for the estimation of constant-elasticity models” (Santos Silva and Tenreyro 2006:649); in addition, it behaves well even when the share of zeros is very large (Santos Silva and Tenreyro 2011).

  12. The same is true of any other estimator that could be used to estimate the IGE of expected income.

  13. In the general case, the measurement error for children is \({\theta }_{0t}+({\theta }_{1t}-1) Y+{W}_{t}\). For parents, it is \({\pi }_{0k}+ {P}_{k}\) when \({\pi }_{1k}=1\) and \({\pi }_{0k}+\left({\pi }_{1k}-1\right)\mathrm{ln}X+{P}_{k}\) in the general case.

  14. This entails no loss of generality because it can always be achieved by changing the monetary units used to measure income, i.e., by dividing the children’s income variable by its mean and the parental income variable by the exponential of the mean of its logarithmic values minus 1.

  15. The PSID collects income information referring to the calendar year prior to the survey. Unless I indicate otherwise, I always refer to “income years” rather than “survey years.” The PSID switched to biannual data collection in survey year 1997. For cohorts that became 56 years old in years in which data are not available, I require that the children be present in the PSID at age 55.

  16. In order to be able to compute a reasonable approximation to long-run income and earnings, the analyses exclude children with fewer than 12 years of income or earnings (as relevant).

  17. For the 1953 and 1952 cohorts, parental information refers to when the children were 14-20 and 15–20 years old, respectively, rather than 13–20 years old.

  18. Ideally, I would use just one sample to study left-side and right-side biases jointly rather than separately. Unfortunately, no feasible PSID sample allows one to construct approximate long-run income measures for both parents and children. More generally, apart from the (restricted access) Swedish registry data used by Nybom and Stuhler (2016), which includes measures of family income but not of earnings, there is no other panel data that could be used with that purpose.

  19. With the statistical packages typically used by social scientists, it is as easy to estimate the IGE of the expectation by PPML using microdata as it is to estimate the conventional IGE by OLS using such data. See Online Appendix, G, for how to estimate the former IGE with the statistical package Stata.

  20. In the literature, controls other that ages are typically not included. The reason is that an IGE does not measure a causal effect. Rather, it is descriptive measure (e.g., Aronson and Mazumder 2008:146; Bjorklund and Jantii 2011, esp. secs. 4 and 6; Hertz 2007:26; Stuhler 2012:2), comparable in nature to, for instance, the Gini coefficient.

  21. The estimates underlying the curves (and standard errors) as well as additional estimates from models with controls for parental age can be found in the Online Appendix, F.

  22. A similar point was made by Nybom and Stuhler (2016: 264) for the conventional IGE. They suggested averaging income information across years (within children) when estimating the conventional IGE. This can be expected to have the same effect as pooling years of information into one sample (as I do here).

  23. I substituted the correlation for the covariance for scaling purposes.

  24. The estimates underlying the curves (and standard errors) as well as additional estimates generated with measures of parental income based on 1–25 years of information (at various parental ages) can be found in the Online Appendix, F.

  25. Reliance on an asymptotic value is implicit in arguments along the lines that the fact that the differences between IGE estimates based on n, n + 1 and n + 2 years of information are small and decreasing means that n + 2 years are enough to eliminate the bulk of attenuation bias.

References

  • Aaronson D, Mazumder B (2008) Intergenerational economic mobility in the US: 1940 to 2000. J Hum Resour 43(1):139–172

    Google Scholar 

  • Björklund A, Jäntti M (2000) Intergenerational mobility of socio-economic status in comparative perspective. Nord J Polit Econ 26:3–32

    Google Scholar 

  • Black S, Devereux P (2011) Recent developments in intergenerational mobility. In: Card D, Ashenfelter O (eds) Handbook of labor economics, vol 4b. Elsevier, Amsterdam

    Google Scholar 

  • Bloome D (2015) Income inequality and intergenerational income mobility in the United States. Soc Forces 93(3):1047–1080

    Article  Google Scholar 

  • Bloome D, Western B (2011) Cohort change and racial differences in educational and income mobility. Soc Forces 90(2):375–395

    Article  Google Scholar 

  • Böhlmark A, Lindquist M (2006) Life-cycle variations in the association between current and lifetime income: replication and extension for Sweden. J Law Econ 24(4):879–896

    Google Scholar 

  • Bratsberg B, Røed K, Raaum O, Naylor R, Jäntti M, Eriksson T, Österbacka E (2007) Nonlinearities in intergenerational earnings mobility. Consequences for cross-country comparisons. Econ J 117:C72–C92

    Article  Google Scholar 

  • Carroll R, Ruppert D, Stefanski L, Crainiceanu C (2006) Measurement error in nonlinear models: a modern perspective. Chapman, Boca Raton

    Book  Google Scholar 

  • Chadwick L, Solon G (2002) Intergenerational income mobility among daughters. Am Econ Rev 92(1):335–344

    Article  Google Scholar 

  • Chetty R, Hendren N, Kline P, Saez E (2014) Where is the land of opportunity? The geography of intergenerational mobility in the United States. Q J Econ 129(4):1553–1623

    Article  Google Scholar 

  • Corak M (2013) Income inequality, equality of opportunity, and intergenerational mobility. J Econ Perspect 27(3):79–102

    Article  Google Scholar 

  • Corak M (2006) Do poor children become poor adults? Lessons from a cross country comparison of generational earnings mobility. IZA Discussion Paper No. 1993

  • Deutscher N, Mazumder B (2021) Measuring intergenerational income mobility: a synthesis of approaches. Working Paper Series WP-2021-09, Federal Reserve Bank of Chicago

  • Fixler D, Johnson D (2014) Accounting for the distribution of income in the US National Accounts. In: Jorgenson D, Landefeld JS, Schreyer P (eds) Measuring economic sustainability and progress. Chicago University Press, Chicago

    Google Scholar 

  • Gourieroux C, Monfort A, Trognon A (1984) Pseudo maximum likelihood methods: theory. Econometrica 52(3):681–700

    Article  Google Scholar 

  • Grawe N (2003) Life cycle bias in the estimation of intergenerational earnings persistence. Analytical studies branch research paper series, statistics Canada

    Google Scholar 

  • Haider S, Solon G (2006) Life-cycle variation in the association between current and lifetime earnings. Am Econ Rev 96(4):1308–1320

    Article  Google Scholar 

  • Helsø A-L (2020) Intergenerational income mobility in Denmark and the United States. Scand J Econ 123(2):508–531

    Article  Google Scholar 

  • Hertz T (2007) Trends in the intergenerational elasticity of family income in the United States. Ind Relat 46(1):22–50

    Article  Google Scholar 

  • Ichino A, Karabarbounis L, Moretti E (2011) The political economy of intergenerational income mobility. Econ Inq 49(1):47–69

    Article  Google Scholar 

  • Landersø R, Heckman J (2017) The scandinavian fantasy: the sources of intergenerational mobility in Denmark and the U.S. Scand J Econ 119(1):178–230

    Article  Google Scholar 

  • Mayer SE, Lopoo L (2008) Government Spending and Intergenerational Mobility. J Public Econ 92:139–158

    Article  Google Scholar 

  • Mazumder B (2005) Fortunate sons: new estimates of intergenerational mobility in the united states using social security earnings data. Rev Econ Stat 87(2):235–255

    Article  Google Scholar 

  • Mazumder B (2016) Estimating the intergenerational elasticity and rank association in the United States: overcoming the current limitation of tax data. In: Cappellari L, Polacheck S, Tatsiramos K (eds) Inequality: causes and consequences. Bingley, Emerald

    Google Scholar 

  • Mazumder B (2001) The miss-measurement of permanent earnings: new evidence from social security earnings data. Federal Reserve Bank of Chicago Working Paper 2001–2024

  • Mitnik P (2020) Intergenerational income elasticities, instrumental variable estimation, and bracketing strategies. Sociol Methodol 50(1):1–46

    Article  Google Scholar 

  • Mitnik P, Grusky D (2020) The intergenerational elasticity of what? The case for redefining the workhorse measure of economic mobility. Sociol Methodol 50(1):47–95

    Article  Google Scholar 

  • Mitnik P, Bryant V, Weber M (2019) The Intergenerational transmission of family-income advantages in the United States. Sociol Sci 6(15):380–415

    Article  Google Scholar 

  • Mitnik P, Bryant V, Weber M, Grusky DB (2015) New estimates of intergenerational mobility using administrative data. Statistics of Income Division, Internal Revenue Service

  • Mitnik P, Bryant V, Grusky D (2022) A very uneven playing field: economic mobility in the United States. Stone center for inequality dynamics discussion paper 2022–3, University of Michigan-Ann Arbor

  • Mitnik P (2017a) Two-sample estimation of the intergenerational elasticity of expected income. Stanford center on poverty and inequality working paper

  • Mitnik P (2017b) Estimating the intergenerational elasticity of expected income with short-run income measures: a generalized error-in-variables model. Stanford center on poverty and inequality working paper

  • Mitnik P (2019) Estimating the intergenerational elasticity of expected income with short-run income measures: a generalized error-in-variables model. Stanford center on poverty and inequality working paper

  • Nybom M, Stuhler J (2016) heterogeneous income profiles and life-cycle bias in intergenerational mobility estimation. J Hum Resour 15(1):239–268

    Article  Google Scholar 

  • Petersen T (2017) Multiplicative models for continuous dependent variables: estimation on unlogged versus logged form. Sociol Methodol 47:113–164

    Article  Google Scholar 

  • Santos Silva JMC, Tenreyro D (Forthcoming) The log of gravity at 15. Port Econ J [Available as Discussion Paper 01/21, Discussion Papers in Economics, University of Surrey.]

  • Santos Silva JMC, Tenreyro S (2006) The log of gravity. Rev Econ Stat 88(4):641–658

    Article  Google Scholar 

  • Santos Silva JMC, Tenreyro S (2011) Further simulation evidence on the performance of the Poisson pseudo-maximum likelihood estimator. Econ Lett 112:220–222

    Article  Google Scholar 

  • Schennach S (2016) Recent advances in the measurement error literature. Ann Rev Econ 8:341–377

    Article  Google Scholar 

  • Solon G (1992) Intergenerational income mobility in the United States. Am Econ Rev 82(3):393–408

    Google Scholar 

  • Solon G (1999) Intergenerational mobility in the labor market. In: Ashenfelter OC, Card D (eds) Handbook of labor economics, vol 3A. Elsevier, Amsterdam

    Google Scholar 

  • Solon G (2004) A model of intergenerational mobility variation over time and place. In: Corak M (ed) Generational income mobility in North America in Europa. Cambridge University Press, Cambridge

    Google Scholar 

  • Stuhler J (2012) Mobility across multiple generations: The iterated regression fallacy. IZA discussion paper No. 7072

    Google Scholar 

  • Wooldridge J (2002) Econometric analysis of cross section and panel data. The MIT Press, Cambridge

    Google Scholar 

Download references

Funding

This study was partially funded by The Pew Charitable Trusts and The Russell Sage Foundation. “Measuring Economic Mobility with Tax-Return Data: Toward an IRS Platform” (Principal Investigators: Pablo Mitnik and David B. Grusky), 2011–2013, $176,736.

Author information

Authors and Affiliations

  1. Stone Center for Inequality Dynamics, University of Michigan, 426 Thompson St. Room 2074, Ann Arbor, MI, 48109, USA

    Pablo A. Mitnik

Authors
  1. Pablo A. Mitnik
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Pablo A. Mitnik.

Ethics declarations

Conflict of interest

Author Pablo Mitnik declares that he has no conflict of interest.

Human and animal rights

This article does not contain any studies with human participants or animals performed by the author.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (PDF 1286 KB)

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Mitnik, P.A. Estimating the intergenerational elasticity of expected income with short-run income measures: a generalized error-in-variables model. Empir Econ 65, 2779–2803 (2023). https://doi.org/10.1007/s00181-023-02442-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00181-023-02442-6

Keywords

JEL Classification

Search

Navigation