AdobeStock_331832345.jpeg
VoxEU Column Migration

The indirect fiscal benefits of low-skilled immigration

There is a widespread perception that low-skilled immigration is a fiscal burden for society. This column incorporates indirect fiscal effects of immigration that arise in general equilibrium into various models that have been emphasised in the empirical immigration literature. It finds that the indirect fiscal effect is in fact positive, with one low-skilled immigrant in the US adding between $700 to $2,100 to the public finances through this channel each year.

What is the fiscal impact of low-skilled immigration?  A widespread perception is that it is a fiscal burden for society. In his widely read blog, Nobel Laureate Paul Krugman (2006) concluded that “the fiscal burden of low-wage immigrants is also pretty clear... I think that you'd be hard pressed to find any set of assumptions under which Mexican immigrants are a net fiscal plus.”

More recently, a report by the National Academy of Sciences (NAS) also studied the fiscal impact of immigration in the US (National Academy of Sciences 2017). The report focuses on immigrants’ direct fiscal effects – taxes paid by the immigrants minus costs for benefits and services they receive. The authors consider various scenarios and for most of them find that low-skilled immigrants indeed have negative effects on public finances. George Borjas, a member of the NAS panel, summarises the report’s findings on low-skilled immigration thus: “(r)egardless of which scenario, it is obvious that low-skill immigrants impose a fiscal burden in the long run...” (Borjas 2016a: 14). The NAS report was also politically influential and cited by President Trump in his first address to Congress in 2017, where he stated: “(a)ccording to the National Academy of Sciences, our current immigration system costs America’s taxpayers many billions of dollars a year.”

In a new paper (Colas and Sachs 2020), we challenge this perception of low-skilled immigrants as a fiscal burden. We enrich the debate by considering indirect fiscal effects that occur in general equilibrium through changes in native wages and native labour supply induced by low-skilled immigration. We show that one low-skilled immigrant in the US adds between $700 to $2,100 to public finances each year through this indirect fiscal effect. This outweighs the direct fiscal costs for the more optimistic scenarios of the NAS report and significantly reduces the burden in the other scenarios. It should therefore be accounted for when calculating the fiscal effects of immigration.

Theory

To reach this conclusion, we derive formulas for this indirect fiscal effect for various models that have been emphasised in the empirical immigration literature. The benchmark model we consider is the so-called  ‘canonical model’ of the labour market (Acemoglu and Autor 2011) where high-school (low-skilled) and college-educated (high-skilled) labour are imperfectly substitutable inputs in production and individuals with different productivity levels are perfect substitutes within these skill levels. We model labour supply responses of natives along both the extensive participation margin and the intensive effort or hours of work margin and allow for labour supply elasticities to differ with income, gender, and family status.

Formalising and quantifying the fiscal effect of a low-skilled immigrant in such a model is complex: the low-skilled immigrant will increase the share of low-skilled workers, which will trigger an increase in high-skilled wages and a decrease in low-skilled wages. These wage changes, in turn, affect labour supply decisions of natives which, in turn, creates another round of wage changes and so on and so forth, creating a fixed-point problem. We follow Sachs et al. (2020) and formalise this fixed-point problem in terms of integral equations. We then obtain a closed-form expression for the indirect fiscal effect of a low-skilled immigrant that consists of the following estimable statistics:

1. Own-wage elasticities for both skill groups

2. Income-weighted average of labour supply elasticities for both skill groups

3. Income-weighted averages of marginal tax rates for both skills groups

4. Income-weighted averages of the products of marginal tax rates (participation tax rates) and intensive (extensive) margin elasticities for both skill groups. 

Intuitively, the own-wage elasticities provide a measure for how much relative wages of natives change due to immigration. Second, labour supply elasticities matter since the labour supply responses of natives mitigate the initial wage shock – for example, the decrease in low-skilled wages due to low-skilled immigration is mitigated if low-skilled natives lower their labour supply as a response to the immigration influx. The third and the fourth component then measure how these changes in wages and labour supply translate into tax revenue. The third element captures the extent to which wages, and thus tax payments, of high-skilled natives increase and those of low-skilled workers decrease as a result. If the tax system is progressive in the sense that the income-weighted marginal tax rate of high-skilled individuals is larger than that of low-skilled individuals, tax revenue increases through this mechanism. The fourth term captures that changes in labour supply of natives also affect tax revenue. We show that whether the increase in labour supply of high-skilled individuals fiscally outweighs the decrease in labour supply of low-skilled boils down to the products of marginal (participation) tax rates and intensive (extensive) margin elasticities. While these formulas give useful intuition from a theory perspective, their relevance can only be assessed through quantification. 

Quantification

To quantify the model, we primarily need information on (1) cross-sectional distributions of earnings conditional on skill, (2) the distribution of intensive and extensive margin elasticities in the population of natives, (3) own-wage elasticities and — perhaps most challengingly —  (4) we need a very careful calibration of the US tax-transfer system. Our baseline dataset is the American Community Survey, from which we can calibrate (1). With respect to the elasticities, we build on a large empirical literature. We consider a scenario with uniform elasticities based on Chetty (2012) and on a scenario with elasticities that differ by income, family status and gender that is based on Bargain et al. (2012). Regarding (3), we make assumptions on the elasticity of substitution based on Card (2009), which can then be transformed into own-wage elasticities. 

For (4), we conduct our own empirical quantification of the US tax-transfer system. We first use    the NBER's TAXSIM to assign tax rates to all individuals in our American Community Survey data. However, TAXSIM does not account for the effective tax rates that are implied by welfare-transfer programmes. Programmes like the Supplementary Nutrition Assistance Program or Temporary Assistance for Needy Families imply an increase in effective tax rates since transfers are phased out as income increases.  To account for this, we use data from the Survey of Income and Program Participation to estimate effective transfer phase-out rates of these welfare programs. Another important detail that is not captured in TAXSIM is that higher earnings imply higher Social Security benefits after retirement.  For this, we combine Social Security formulas with estimates of life-cycle earnings paths that we estimate from the NLSY79. Our quantification of how average marginal and participation tax rates vary across the income distribution of displayed in Figure 1.

Figure 1 Marginal and participation tax rates by individual earnings

Notes: The top panel gives the marginal effective tax rates implied by income taxes, the social security system, and transfer programs. The bottom panel reports the corresponding participation tax rates. Income taxes are given by the sum of state and federal income taxes, social security is defined as payroll taxes minus the discounted sum of future social security benefits, and transfer payments are the sum of TANF and SNAP phase outs.

Results

We combine our closed-form expressions with this quantification to evaluate the indirect fiscal effects of low-skilled immigration under the variety of assumptions on parameters of the production function and magnitude of labour supply elasticities. The results are summarised in Table 1. Across the scenarios we find that the indirect fiscal effect of one low-skilled immigrant is between $770 and $1,470 per year.

Table 1 Indirect fiscal effects

Notes: The three columns show the indirect fiscal effect under different assumptions of the elasticity of substitution, ranging from 1.5 to 2.5. Each row displays the indirect fiscal effect for different assumptions about the labour supply elasticity.

To compare these indirect fiscal effects to the direct fiscal effect, we calculate an annualised direct fiscal cost associated with low-skilled immigrants using results from the 2017 NAS report. These annualised direct fiscal costs under different scenarios which vary the marginal cost of public goods and the education of an immigrant are displayed in Table 2. In almost all cases, the direct fiscal effect is negative and of a similar magnitude to the indirect fiscal effects we calculated. In some of the scenarios that we consider, accounting for the indirect fiscal costs of an immigrant turns the total fiscal effect from a fiscal burden to a fiscal surplus and significantly reduces the fiscal burden in the other scenarios. 

Table 2 Annuitised direct fiscal effect of a low-skilled immigrant

Notes: This gives the direct fiscal effect of an immigrant that arrives at age 23 and dies at age 79. We use a discount rate of 1%. Only direct fiscal contributions are accounted for and rely on Figure 8-21 of National Academy of Sciences (2017). We calculate the annuity value for the period of 23 until 65 (age of retirement).

Robustness

There is some controversy in the literature over the appropriate model to analyse and estimate the wage effects of immigration. A natural concern is that the indirect fiscal effects are also sensitive to these modelling choices.  As such, we extend our model to allow for a variety of different production functions and labour supply responses that have been emphasised in the immigration literature. These extensions and the associated indirect fiscal effects are summarised in Table 3. Whether we allow for (i) imperfectly substitutable workers within skill levels and a finer stratification of skill (Borjas 2003), (ii) immigrant-native complementarity in production (Ottaviano and Peri 2012), (iii) alternative definitions of skills (Dustmann et al. 2013), or (iv) endogenous occupation choice (Peri and Sparber 2009), results are rather similar.  We find indirect fiscal effects of low-skilled immigrants in the range of $1,000 to $2,100.

Table 3 Estimates of annual indirect fiscal effect of one low-skilled immigrant under different model specifications

Notes: Estimates of annual indirect fiscal effect of one low-skilled immigrant under different model specifications. For the  ‘Simple Textbook Model‘ and the ‘Canonical Model‘ we use our results associated with an elasticity of substitution between high-skilled and low-skilled workers of 2, the central value we use in our quantification. For the 'Canonical Model’ with labour supply adjustments, we display our results with common labour supply elasticities. For all specifications, we show the indirect effect for the average low-skilled immigrant.

Conclusion

Whether or not low-skilled immigrants 'pay their fair share’ has become a hot-button issue in US politics. Our analysis challenges the commonly held belief of low-skilled immigrants as a fiscal burden and highlights the importance of accounting for indirect fiscal effects when considering the fiscal impacts of immigration.

References

Acemoglu, D and D Autor (2011), “Skills, Tasks and Technologies: Implications for Employment and Earnings”, Handbook of Labour Economics 4: 1043-1171.

Bargain, O, K Orsini and A Peichl (2014), “Comparing Labour Supply Elasticities in Europe and the United States: New Results”, Journal of Human Resources 49: 723-838.

Borjas, G J (2003), “The Labour Demand Curve Is Downward Sloping: Reexamining the Impact of Immigration on the Labour Market”, The Quarterly Journal of Economics 118: 1335-1374.

Borjas, G J (2016), “A User's Guide to the 2016 National Academy Report on ‘The Economic and Fiscal Consequences of Immigration’”, Working paper.  

Card, D (2009), “Immigration and Inequality”, American Economic Review 99: 1-21.

Chetty, R (2012), “Bounds on Elasticities with Optimization Frictions: A Synthesis of Micro and Macro Evidence on Labour Supply”, Econometrica 80: 969-1018.

Colas, M and D Sachs (2020), “The Indirect Fiscal Benefits of Low-Skilled Immigration”, CEPR Discussion Paper No. 15325.

Dustmann, C, T Frattini and I P Preston (2013), “The Effect of Immigration Along the Distribution of Wages”, The Review of Economic Studies 80: 145-173.

Krugman, P (2006) “Notes on Immigration”, New York Times op-ed. 

National Academy of Sciences, U S (2017), The Economic and Fiscal Consequences of Immigration, National Academies Press.

Ottaviano, G I and G Peri (2012), “Rethinking the Effect of Immigration on Wages”, Journal of the European Economic Association 10: 152-197.

Peri, G and C Sparber (2009), “Task Specialization, Immigration, and Wages”, American Economic Journal: Applied Economics 1: 135-69.

Sachs, D, A Tsyvinski and N Werquin (2020), “Nonlinear Tax Incidence and Optimal Taxation in General Equilibrium”, Econometrica 88: 469-493.

1,995 Reads