A Re-estimation of Powell & Skarbek (2006)
Introduction
Powell and Skarbek’s 2006 paper, “Sweatshops and Third World Living Standards: Are the Jobs Worth the Sweat?” was the “first economic study to compare systematically sweatshop wages with average local wages”, and thus received large amounts of reception by a number of economic institutes and reporters. Powell and Skarbek first examined reported sweatshop wages in ten Asian, Latin American, and Caribbean countries vis-à-vis the average standard of living in each of these countries, measured as the average national income. Sweatshop wages were then compared with the average income per worker, and wages were then compared to two international poverty lines. Similar analyses were done with a larger set of wages, compiled from US news reports.
The findings of the study confirmed what the economic orthodoxy has proclaimed for many years; multinationals contracting in foreign countries provide substantially higher pay than domestic firms, and therefore assist in economic and human development. This thesis has also been confirmed by a number of past and more recent texts, including but not limited to Greene et al. (2007), Powell (2014), Krugman (1997), and Brown et al. (2003). Not surprisingly, the economic consensus has been in conflict with the anti-sweatshop movement, based especially in the US.
As it has been 15 years since the publishing of this study, it would be wise to recalculate and reevaluate Powell & Skarbek’s findings. I will begin by defining the methodology used, and from there cover my results.
Methodology
The general methodological approach used here will be cross-country examinations, using data provided by national reports, wage reports conducted by NGOs, and the World Development Indicators (WDI).
The standard of living is a measure of the socioeconomic status of people, the latter of which is heavily influenced by the amount of wealth, assets, income they have accumulated or will continue to accumulate. The standard of living is also dependent on the environmental and political conditions of a country. Multiple measurements of the average standard of living will be used, as living standards are highly complex and therefore no single measurement will be sufficient. Measurements of the average standard of living will consist of gross national disposable income per capita and average wages, and poverty lines may be used to provide benchmarks for a “life-sustaining” standard of living.
Gross national disposable income (GNDI) is preferred over gross national income because the former accounts for the full consumer capabilities of people. GNDI, as opposed to GNI, includes worker remittances in it’s calculation, which, in many countries, serves as a significant source of income. As a component of the standard of living, GNDI helps us approximate the spending power of consumers in their own national economy. GNDI is calculated as follows:
There are a number of limitations when comparing GNDI per capita with average sweatshop wages. The primary concern is that GNDI per capita is an approximation based upon the total population, while wages are approximated much more accurately with factory surveys or wage bill estimations. Due to this problem, GNDI per capita is likely understated relative to wages. To solve this problem, per worker GNDI will also be used as a comparative tool.
The average wage is sourced from national reports and surveys conducted by the statistics bureaus of the countries in this study. As national reports provide wages in local currencies, these wages are converted to international dollars using 2015 PPP exchange rates. For certain countries, wages are expressed in dollars, or in hourly and yearly terms. If expressed in dollars, the wage is converted to the local currency using 2015 market exchange rates and then reconverted to 2015 international dollars. If in hourly terms, the wage is multiplied by the typical hours worked per day (this number is determined by taking the mandated workweek hours and dividing them by the number of mandated work days) to create the daily wage rate. This wage rate is multiplied by the average monthly work days to find the average monthly wage. Average monthly work days are calculated by taking the number of mandated off days per week multiplied by 52, representing total yearly off days. 365, the total number of days a year, is subtracted by the total yearly off days and the difference is divided by 12. Yearly wages are divided by average monthly workdays.
13 countries are sampled, including the 11 countries examined in Powell & Skarbek (excluding Nicaragua, Haiti, and Costa Rica). These countries are Bangladesh, Indonesia, China, Vietnam, Sri Lanka, El Salvador, Honduras, Dominican Republic, India, the Philippines, Mexico, Thailand, and Cambodia. Sweatshop wages in these countries are sourced from the Fair Labor Association’s (FLA) 2016 factory pay report. The FLA calculates average compensation in factories as basic wages plus benefits minus taxes. Wages are provided for the year 2015 and therefore living standard indicators will be set in 2015.
Equivalent poverty lines for each country sampled are created based off of the general $1.90/a day 2011 PPP international poverty line (IPL), the $5.50/a day line and the upper estimate of the “Ethical Poverty Line” (EPL) created by Peter Edward in 2006. While the $1.90 a day line is the most used threshold when measuring poverty, it does not represent any significant movement towards a decent livelihood, and thus is not sufficient in serving as a living standard benchmark. The $5.50 IPL was formulated by the World Bank in 2017 as the upper maximum of what was considered to be “moderate poverty”. While it still suffers from a number of technical and methodological problems as the previous line does, it’s higher value ensures that those living at or above the line are living more adequate lives. For a more accurate estimate of the level of consumption needed for adequate living, Peter Edward calculated a set of poverty lines based upon the kink points on the Preston Curve, at and after which life expectancy would not be significantly impacted by income. At a consumption rate of $3.3/a day in 2002 international dollars, the maximum life-expectancy is around 75 years. Poverty lines are set in country-equivalent terms in 2015 values using the following formula.
The primary issue with comparing wages with IPLs is that IPL values are typically represented as daily consumption rates, while most workers are paid weekly or monthly, and therefore converting the monthly average wage into daily terms would not be entirely accurate. For this reason, IPL values are turned to monthly figures and doubled, to account for both the worker and one other family member, possibly a child of the worker.
Results
Figure 1 depicts the relationship between average sweatshop wages and GNDI, a measure of final disposable income. In most sweatshop-exporting countries, the wage-to-GNDI ratio remains at 2 or below, with Sri Lanka having the lowest at 0.72, and the highest ratio held by Honduras, at 4.62. The degrees of the ratios in this relationship varies immensely from other wage relationships for several reasons. In countries like the Dominican Republic and Sri Lanka, where sweatshop wages are above average compared to other countries, net secondary income is abnormally high due to large worker diasporas providing remittances to domestic family members. This makes their per capita GNDI relatively high, resulting in an undervalued wage.
In all but 2 countries, the average monthly sweatshop wage exceeds the average monthly GNDI. It should be noted that GNDI per capita is taken from the total population in a country, which may lead to skewed data as the countries sampled here have large youth populations which generally do not have the capacity to work. In Figure 2, sweatshop wages are examined in relation to per worker GNDI in 2015, which is a better comparative tool in this context.
9 countries have sweatshop wages falling below per worker GNDI, however, interestingly enough, wages in 4 of these 9 countries are almost equal to GNDI, with China having an essentially 1:1 relationship between sweatshop wages and worker GNDI. Sri Lanka and the Dominican Republic still have a relatively low wage in Figure 2, and as explained previously, this has to do with the high GNDI resulting from significant worker remittance inflows. Figure 3 graphs the relationship between average sweatshop wages and average national wages.
In 8 out of 13 countries, average sweatshop wages are equal to or lower than national wages. A number of countries, made up of Sri Lanka, El Salvador, the Dominican Republic, the Philippines, Mexico, Thailand, and Cambodia have sweatshop wages valued at 64–92% of the national wage, a differential of just 28% between the highest paying and lowest paying countries in this group. We can only see such similarities in wage-wage relationships, as opposed to wage-income relationships.
For now, wages have been examined in relation to income and other forms of earnings, with little evidence showing that, generally, sweatshops provide an above average standard of living for their workers. Figure 4 takes an alternative comparison, measuring average sweatshop wages vis-à-vis three poverty lines; the country equivalent $1.90 and $5.50/a day 2011PPP$ (recalculated in 2015 values) lines, and the country equivalent $3.30/ a day 2002PPP$ ethical poverty line, also set in 2015. It should be noted that these comparisons are not to be taken as poverty headcounts, but rather as to show how well off the average sweatshop worker may be using a living standard “benchmark”.
For all countries, the average sweatshop worker earns more than enough to surpass the $1.90 line, excluding Bangladesh where the sweatshop wage is a slim 15% greater than the extreme poverty threshold, while Thailand has a wage nearly 7 times the $1.90 threshold. Sweatshop wages, on average, provide pay satisfying the EPL and $5.50 (“moderate poverty”) thresholds, however Bangladesh and Cambodia do not provide such level of pay for sweatshop workers.
While sweatshops provide wages that are generally above some poverty values, this does not necessarily mean that many workers are not in poverty or facing poverty-like conditions. These findings point the simple fact that the average sweatshop laborer faces some level of poverty, highly variable across different countries, but “extreme poverty” as defined by the World Bank may not be much of a concern among factory workforces.
Conclusion
The results of this article are ultimately insufficient in determining how well sweatshop wages hold up against average living standards or if sweatshop wages provide an above average standard of living. A comparison of wages with GNDI per capita may prove the Powell-Skarbek hypothesis, however a comparison of sweatshop wages with per worker GNDI or with national average wages does not prove the theory. Perhaps what is more important is clearly defining how the “standard of living” is to be measured, as the above or below average positions of sweatshop wages are determined almost entirely with what the researcher considers to be the average standard of living. While sweatshops do provide important employment opportunities for the poor in developing nations, it’s unlikely that they provide an above-average standard of living by any significant margin.
Tables (For Referencing)
References
Brown, Drusilla, et al. “The Effects of Multinational Production on Wages and
Working Conditions in Developing Countries.” National Bureau of Economic Research, 2003, https://www.nber.org/system/files/working_papers/w9669/w9669.pdf. Working Paper №9669
Capelli, Clara, and Gianni Vaggi. “Why Gross National Disposable Income Should Replace Gross National Income.” Development and Change, vol. 47, no. 2, 2016, pp. 223–39. Crossref, doi:10.1111/dech.12225.
Edward, Peter. “The Ethical Poverty Line: A Moral Quantification of Absolute Poverty.” Third World Quarterly, vol. 27, no. 2, 2006, pp. 377–93. Crossref, doi:10.1080/01436590500432739.
FLA. “Towards Fair Compensation in Global Supply Chains: Factory Pay Assessments in 21 Countries.” Fair Labor Association, 2016.
Greene, Zoë, et al. “Negative Impacts of Minimum Wage and Anti Sweatshop Legislation.” Humanomics, vol. 23, no. 2, 2007, pp. 83–92. Crossref, doi:10.1108/08288660710751335.
Kakwani, Nanak, and Hyun H. Son. “Global Poverty Estimates Based on 2011 Purchasing Power Parity: Where Should the New Poverty Line Be Drawn?” The Journal of Economic Inequality, vol. 14, no. 2, 2016, pp. 173–84. Crossref, doi:10.1007/s10888–016–9322-x.
Krugman, Paul. “In Praise of Cheap Labor.” Slate Magazine, 21 Mar. 1997, slate.com/business/1997/03/in-praise-of-cheap-labor.html.
Powell, Benjamin, and David Skarbek. “Sweatshops and Third World Living Standards: Are the Jobs Worth the Sweat?” Journal of Labor Research, vol. 27, no. 2, 2006, pp. 263–74. Crossref, doi:10.1007/s12122–006–1006-z.
Powell, Benjamin. Out of Poverty: Sweatshops in the Global Economy (Cambridge Studies in Economics, Choice, and Society). Cambridge University Press, 2014.
