Choosing Skilled Foreign‐Born Workers: Evaluating Alternative Methods for Allocating H‐1B Work Permits

• Published date: 01 January 2018
• DOI: http://doi.org/10.1111/irel.12203
• Author: Chad Sparber
• Date: 01 January 2018

Choosing Skilled Foreign-Born Workers:

Evaluating Alternative Methods for Allocating

H-1B Work Permits*

CHAD SPARBER

The H-1B program allows highly educated foreign-born labor to temporarily work

in the United States. Quotas restrict the number of H-1B recipients. In many

years, all available work permits were allocated by random lottery. This paper

argues that an alternative distribution method based upon ability would increase

output, output per worker, and wages paid to less-educated workers. Baseline esti-

mates suggest that a change in allocation policy could result in a $26.5 billion

gain for the economy over a 6-year period. This estimate grows when H-1B

demand rises.

Introduction

Economists agree that highly educated workers are scarce and productive

inputs in the creation of macroeconomic output. The H-1B program attempts

to increase the supply of highly educated workers, and therefore output, by

providing temporary work permits to foreign-born individuals in specialty

occupations seeking employment in the United States. Current policy restricts

the number of new H-1B permits distributed to prospective employees of most

firms to 65,000 per year, plus an additional 20,000 for workers who have

obtained a master’s degree or higher education in the United States.

The distributional consequences of this program are widely debated in the

academic literature. Basic supply-and-demand models argue that the increased

supply of educated foreign labor should reduce compensation paid to similar

native-born workers. Though some empirical studies support this view, alterna-

tive work argues that immigrants instead complement native-born labor and

expand employment opportunities. Policy implications of these studies inform

opposing views about whether access to skilled worker permits should be

expanded or contracted. This paper takes no stance on that debate: We assume

∗The author’s address is Colgate University, Hamilton, New York. E-mail: csparber@colgate.edu. The

author is grateful for funding provided by the National Science Foundation.

JEL Codes: J61, F22

INDUSTRIAL RELATIONS, Vol. 57, No. 1 (January 2018). ©2017 Regents of the University of California

Published by Wiley Periodicals, Inc., 350 Main Street, Malden, MA 02148, USA, and 9600 Garsington

Road, Oxford, OX4 2DQ, UK.

3

that the current quota is given and fixed, and we recognize that empirical

assessments of how an H-1B program expansion would affect labor-market

opportunities for native-born workers have been well covered by other studies

including Kerr, Kerr, and Lincoln (2015); Doran, Gelber, and Isen (2014);

Ghosh, Mayda, and Ortega (2014); and Peri, Shih, and Sparber (2015a).

This paper instead focuses on how H-1B work permits are allocated. Two

options are considered. The first operates according to recent practice in which

U.S. Citizenship and Immigration Services (USCIS) has distributed permits

through a random lottery. The second is a hypothetical alternative that would

award permits according to ability. Because this method is hypothetical,

regression analysis cannot identify the costs and benefits of each system.

Instead, this paper performs a calibration exercise built upon a simple theory,

observed data, and prior work. The simulations demonstrate that an allocation

method assigning permits according to ability—as measured by the wage and

marginal product of labor associated with individual applicants—would

increase macroeconomic production and output per worker regardless of the

elasticities of substitution across education and nativity groups. Wages paid to

workers with little educational attainment would also rise. Whether highly edu-

cated native-born workers benefit or suffer from the alternative policy, in con-

trast, does depend upon relative elasticities.

A priori, some of these implications might appear obvious: If policy is

designed to allow the most productive individuals to work in the economy,

then productivity will increase. Nonetheless, this calibration exercise is worth

conducting for at least two reasons. First, while many people continue to

debate broad immigration issues such as the expansion or contraction of the

H-1B program, little attention has been paid to the effect that narrower immi-

gration policy changes could have on the economy. Policymakers should have

a sense of the economic ramifications of the current H-1B allocation method.

Although one might expect that allocation favoring ability would increase out-

put, this paper assesses how large potential gains might be. It is useful to

examine potentially beneficial changes to the U.S. immigration system beyond

altering the size of flows entering the country. Second, the academic literature

has developed highly contested estimates of the elasticity of substitution

between native and foreign-born workers. These estimates form the core of

many debates on the consequences of immigration. This paper, in contrast,

illustrates that this parameter is inconsequential for determining the average

macroeconomic effects of H-1B permit allocation.

The analysis proceeds as follows: the next section develops the theory. It

takes a restrictive view of the benefits of immigration. Recent work has argued

that the H-1B program and highly educated foreign-born workers are vital for

the science, technology, engineering, and mathematics (STEM) workforce of

4/ C

HAD SPARBER

the economy, which is in turn responsible for much of the country’s techno-

logical and productivity growth. Two-thirds of H-1B workers are employed in

computer-related occupations, for example. This paper’s model instead restricts

foreign inputs to the production of finished goods. Those workers might or

might not complement native-born and less-educated workers in the production

process, but the model does not permit them to generate technology spillovers.

As noted above, the degree of complementarity does not affect the main pro-

ductivity implications of the model, though it will influence wage effects. The

theory section closes by outlining the mathematics behind the two allocation

policies considered, and by acknowledging the limitations and simplifying

assumptions of the model.

In the third section I describe the data and assumed parameters used to cali-

brate the model. The exercise uses data found in the U.S. Census and Ameri-

can Community Survey (ACS), USCIS information acquired from a Freedom

of Information Act (FOIA) request, and parameter estimates produced by past

studies. The simulations in the fourth section show that changing the allocation

scheme from the current lottery method to one favoring the most productive

workers can raise output (and output per worker) by 0.15 percent over a 6-year

period (the maximum length of time an individual can work on H-1B status,

with limited exception). On the one hand, this figure appears small. On the

other hand, the annual flow of 85,000 H-1B workers represents just 1.3 per-

cent of foreign-born skilled employment in 2014, so the magnitude of the

response is naturally limited in scope relative to the total size of the U.S. econ-

omy. Given the context of the program’s size, its magnitude is large but rea-

sonable. Moreover, a 0.15 percent rise in U.S. income represented roughly

$26.5 billion in 2014. This amounts to a level of output comparable to the

entire gross domestic product (GDP) of Jamaica in purchasing power parity

terms, and it exceeds the GDP of nearly 100 nations of the world. Estimates

for the potential GDP gain are particularly sensitive to H-1B demand and rela-

tive permit scarcity. Using different data and parameter assumptions, figures

range from $8.3 billion when many workers who desire a permit are able to

secure one, to $43.3 billion in recent years when at most one-third of prospec-

tive H-1B applicants receive a permit.

Theoretical Model of Production

Recent studies have produced ample evidence that highly educated foreign-

born workers (and H-1B workers more specifically) generate technological

gains. For example, Hunt (2011) argues that immigrants are more entrepre-

neurial and innovative than native-born workers. Hunt and Gauthier-Loiselle

Allocating H-1B Work Permits /5