Supply variabilities in public workfares

Highlights

• Recent evidence shows that large public works have increased private wages in India.

• We show how variability in program provision can compress wage increases.

• This “compression” increases with variability in program provision.

• It is more severe in areas where inflation is low relative to where it is high.

• Reducing program variability can improve welfare without increasing expenditure.

Abstract

This paper presents a model of the labor market where public workfares increase private wages by reducing labor supply. In a dynamic setting, we show that when wages are downwardly rigid, forward-looking employers optimally compress wage increases in response to intertemporal variability in the level of program implementation. The model generates two key predictions: greater variability in program provision results in a larger compression of wage increases, and compression of wage increases is more severe under low inflation. We empirically verify these predictions using data from two large workfares from India.

Introduction

Public works have long been a standard response of the government in its role as an employer of last resort. Starting with the Poor Employment Act of 1817 in Britain (Blaug, 1963, 1964) and the New Deal Programs during the 1930s in the US (Kesselman, 1978; Bernstein, 1970), large public works have played an increasingly dominant role in providing income support for the poor in many countries. In recent times public workfare programs have been implemented in a number of developing countries in Latin America, Asia, and Africa (Drèze and Sen, 1991; Lipton, 1996; Subbarao, 1997; Keddeman, 1998). The workfare programs in Europe (Lødemel and Trickey, 2001), work-for-dole in Australia (Bessant, 2000), and the government response to Fukushima disaster in Japan (Nagamatsu, 2014) are examples of similar programs that have been implemented for specific groups in developed economies as well.

Apart from their role as safety-net programs, such labor market interventions from the government have been acknowledged to affect key labor market indicators in the private sector as well. For example, public workfare programs can decrease employment in the private sector under perfectly competitive labor markets. On the other hand, if employers have a significant degree of control in the wage determination process, then public workfare programs can potentially increase both employment and wages in the private sector by improving the bargaining power of workers or by increasing worker's reservation wages (Basu et al., 2009).¹ As monopsonistic labor markets are increasingly found to be pervasive (Ashenfelter et al., 2010; Manning, 2011; da Costa and Maestri, 2019), low wage equilibria could exist due to distortions in the labor market and not because of low worker productivity. In such a scenario, workfare programs can not only improve the welfare of program participants but also of non-participants through wage increases in the private sector (Muralidharan et al., 2018).

However, it is important to acknowledge that apart from employer market power, labor markets may also have other types of distortions like downward nominal wage rigidity that may have a strong influence on how public works affect the private labor market. Studies by Kahn (1997); Lebow et al. (1999); Knoppik and Beissinger (2003, 2009); Dickens et al. (2006) and Behr and Pötter (2010) provide evidence of downward nominal wage rigidity in labor markets under different contexts. More recently, Kaur (2019) provides evidence of downward rigidity in nominal wages using wage and employment responses to rainfall shocks in India. Hence, it becomes important to consider the effects of such rigidities on the labor market response to public workfare programs.

Finally, the impact of public workfares depends, to a large extent, on how well they are implemented. Successful implementation of public workfares is often hampered by the lack of institutional capacity and budgetary constraints that are most severe for countries most in need of such labor market interventions. The presence of such constraints leads to variability in the level of program implementation.²

This paper provides a theoretical framework to understand how variability in program implementation and the consequent uncertainty can have large negative effects on the gains from public workfare programs. In other words, we show how after controlling for the level of program provision today, uncertainty regarding the level of program provision in the future can dampen or “compress” wage increases. Hence, variability in public workfares can effectively reverse the welfare gains from such policy interventions.

We develop a model of the labor market with a general increasing and concave production function and risk neutral utility functions. These modeling choices lead to results that are general and robust to assuming specific production or utility functions. Our model accommodates three key features of (i) employer market power, (ii) downward nominal wage rigidity, and (iii) temporal variability in the implementation of public workfares. Using a static model of a private labor market, we first show that public workfare leads to a contraction of labor supply for private employment which increases private wages.³ The level of program provision determines the extent of contraction in private labor supply and the consequent increase in private wages. For example, if the public workfare absorbs (say) 5% of the work force then private wages rise by 2%. If the workfare employs a larger proportion of the workforce then private rise further (up to the marginal product of labor).

Next, to accommodate temporal variation in program provision we introduce a two-period model. Keeping the present level of program provision the same, we compare the first-period wage in the static and two-period models. Continuing with the earlier example, if the workfare employs only 1% of the labor force in the second period, the employer would optimally want to reduce wages. This, however, is not possible when wages are downwardly rigid. As a result, forward-looking employers compress wage increases in the first period to be less than the statically optimal wage increase of 2%. Therefore, the compression of wage increases can be attributed to the trade-off between employers’ current and future profits where the level of private wage today is dependent not only on the current but also future levels of program provision. Hence, if the provision of public workfare deteriorates in the second period with certainty, ceteris paribus, the first-period private wage in the two-period model is lower than in the one-period static model.

In reality, however, there is uncertainty regarding the level of program provision in the future. To incorporate this uncertainty, we assume that the level of program implementation or the rate at which government rations demand for work in the program is drawn from a given distribution. Here, greater uncertainty can be associated with an increase in the variance of the distribution that lies between the two extremes of no implementation and complete implementation. In this framework, the optimal wage in the first period is based on optimizing current and expected future profit. We find that greater uncertainty regarding the level of program provision in the future leads to higher compression of wage increases today.

The uncertainty resulting from variability in the program provision could affect wages through other channels as well. Employers could be risk averse and hence reduce investment levels, resulting in lower productivity and lower wages. Uncertainty could also affect choice of goods to manufacture, or affect migration patterns of workers, which could also affect wages. Here, we present a simple model that abstracts away from investment, risk aversion, and migration to show the simplest mechanism through which variability in the provision of public workfares can compress potential wage increases. Adding more complexity to the model may enrich it and demonstrate multiple channels that, under additional assumptions, may give a similar result. Therefore, our model is more general and robust as it relies on fewer assumptions.

Usually, increases in private wages due to large public workfares are expected to negatively affect private employers' profits. In other words, while a decrease in private wages due to a lower level of program implementation may reduce workers' utility, it increases employers’ profit. Counterintuitively, we show that a reduction of wages due to increased variability in program provision reduces both employer profit and worker utility. The compression of wage increases reduces employer profit as the marginal gain due to lower private wages is more than offset by the loss in marginal product of labor when wages are not competitively determined.⁴ An important consequence of this finding is that by simply reducing supply variabilities, without necessarily increasing the overall fiscal burden on the state, the government can unambiguously improve overall welfare that results from the accompanying decompression of wage increases.

Our model gives two key empirically verifiable predictions. First, our model predicts that the increase in wages due to a public workfare is lower when variation in program provision is high. Second, the compression of wage increases is more severe when inflation is low. This is because the downward nominal wage rigidity constraint binds less often when inflation is high. We use data from two large workfare programs in India to empirically verify these two key predictions. We find that the increase in real private wages due to the workfare is lower as the variability in program provision increases.⁵ In support of the second prediction, we find that the compression of wage increases is significantly higher for low inflation districts relative to high inflation districts.

Our results contribute to the literature in two important ways. First, we supplement the literature on the effects of downward rigidity in wages (Elsby, 2009; Stüber and Beissinger, 2010; Falvey and Kreickemeier, 2009). These studies model wage rigidity through a loss in worker productivity following a downward revision in nominal wages. In these models, firms compress wage increases today in anticipation of future wage cuts that may prove costly to implement. We share the same insight as these studies except our objective is to associate compression of wage increases with the need to cut wages in the future due to a probable decline in program provision; downward nominal wage rigidity is assumed in our model. A similar framework can be used to assess other policy interventions that affect labor market outcomes under downward nominal wage rigidity.

Second, we contribute to the recent theoretical literature on public works by explicitly incorporating temporal variability in the implementation of the public workfares. While recent studies like Basu et al. (2009), Imbert and Papp (2015), and Sukhtankar (2016) acknowledge incomplete implementation in public works and the consequent variable effect on key labor market indicators, our dynamic model allows us to explicitly parameterize the effects of temporal variability in program provision.⁶ On the other hand, while studies like Basu (2013) and Bahal (2016) discuss labor market outcomes under public workfares using models with more than one period, there is no role for variability in program provision as both these studies assume complete program implementation.

The next section discusses our model of public workfare and section 3 discusses some extensions of the model including inflation and perfect competition. Section 4 uses data on large public workfares in India as a case study to present some important empirical regularities regarding the programs and rural labor markets in India, justifying the key assumptions of our model. Section 5 tests the key empirical predictions of the model. Section 6 concludes.