Public Works Don't

This Comment draws heavily on the excellent, comprehensive and more nuanced paper by Anna McCord Win-win or Lose? An Examination of the Use of Public Works as a Social Protection Instrument in Situations of Chronic Poverty (2005). First presented to the Conference on Social Protection for Chronic Poverty, IDPM; published as ‘The Social Protection Function of Short Term Public Works Programmes in the Context of Chronic Poverty’ in Barrientos A and Hulme D (eds) Social Protection for the Poor and Poorest: Risks, Needs and Rights, Palgrave, United Kingdom.

Public works are a popular social protection instrument throughout southern Africa. They can be effective as a response to transient or cyclical crises, by smoothing consumption and preventing the distress selling of assets. But programmes offering short term employment opportunities may not work so well in situations of chronic poverty. Even their main advocates, such as the World Bank, admit that “public works are essentially a temporary safety net and should never be used as a permanent escape route from poverty”[i]. Yet donors, including the World Bank and DFID, continue to advocate and support them, claiming unattainable goals such as “enhanced livelihoods for the poor people of Malawi[ii]”. Why?

There are four attractions of public works programmes to donors and governments, which override even their own appraisals of the actual benefits. These are that public works: 

• are consistent with the ideology of not giving beneficiaries “something for nothing” thus inducing “dependency”;
• provide a “win-win” combination of welfare transfers and the creation of productive assets;
• appeal politically because they “create jobs”;
• solve the complex problem of targeting by being “self-targeted”.

But each of these four apparent attractions can be disputed and invalidated.

The first misconception is that there is an intrinsic benefit in making the poor work in return for social protection. This is perceived to prevent “benefit scroungers”, and to reduce “dependency”. But it is hard to apply the term “benefit scrounger” when the values of transfers made through public works programmes are deliberately set so low that they leave participants well below the poverty line. And, even if the concept of “dependency” is accepted (which many would debate), it is difficult to understand how a dependency on the state to provide poorly-paid employment is any less dangerous than a dependency on the state to provide direct transfers. There is also an argument that public works confer “dignity” through employment, and that they impart skills which may be useful in finding future work opportunities. But there is an equally strong counter-argument that participation in public works may stigmatise rather than empower; and that most programmes do not last long enough to impart skills significantly greater than those that the participant already had. In addition, the work requirement may have negative impacts in terms of the opportunity cost it implies, by diverting labour away from domestic production. In general, because poor people cannot afford to be idle, they necessarily have to give up some other form of income in order to join a public works programme, which reduces still further the net value of the transfer.

The second questionable attraction of public works is that they create assets for the beneficiaries and for the community. But, by and large, the beneficial economic and developmental value of the assets created is assumed rather than empirically established. Where these assets are designed to reduce environmental threats, such as flooding in Bangladesh, the benefit is more readily apparent; but where they are intended to promote general economic growth, the value is less certain. In some cases, the construction of a road or community structure may have positive direct and indirect benefits to the community, but this must be critically evaluated … and in reality very seldom is. At worst, assets created through public works may be white elephants, unstrategically selected and uncoordinated with broader development initiatives. Because of the emphasis on labour-intensive approaches and lack of knowledge on the part of programme managers of how to carry out labour intensive engineering projects,  there is a risk that the quality of the assets created may be low; and frequently little provision is made for their maintenance and repair. To be of long-term benefit to the poor, the impact of improved road access or rehabilitated irrigation would require the public works programme (or complementary support) to provide a sufficient asset base and risk insurance for the beneficiaries to engage in new or diversified economic activities. On top of this, public works are a very expensive way to make transfers – the cost of transferring one unit to the poorest in Malawi is calculated to be 13.9 units through public works, compared with 1.7 for a direct cash transfer[iii]: so the economic value of the assets created has to be significant in order to justify such a cost premium.

Third, public works are popular with politicians because the government can claim to be creating jobs (and it is no coincidence that public works programmes are frequently initiated in the run-up to elections!). However, such a claim may well be disingenuous. The nature of public works in southern Africa is short term, usually lasting no more than two to four months. Whilst participants do work in return for payment, and thereby qualify as employees, the short term nature of their employment can only make a limited contribution to their longer term livelihoods. Offering a temporary consumption smoothing intervention in the context of chronic un(der)employment and poverty represents a serious mismatch between problem and policy response. Only one-third of participants on a DFID and ILO supported programme in South Africa considered that the programme would have any sustained impact on their livelihoods[iv].

Finally, public works programmes are touted as offering a simple alternative to other more complex and costly targeting approaches. This is based on the premise that the poorest will self-select themselves because of the work requirement and low wages, which are conventionally set at or below prevailing rates, and which the less poor will therefore find unattractive. But this assumes a perfectly functioning labour market. In reality, there is plenty of evidence of significant inclusion and exclusion errors, because the marginal value of labour varies within and between households. Thus public works may be attractive to surplus labour in less poor households, and unattractive to poorer households with limited labour. Moreover, the adoption of extremely low wages is often in tension with the social protection objectives of public works programmes, since such low wages are unlikely to have any significant impact on chronic poverty. And where the scale of public works employment is trivial in a context of mass un(der)employment (eg in South Africa, where the ratio is 200,000 to 4 million), demand will always outstrip supply, and there is a significant risk of the leakage of employment opportunities to those who are not among the poorest.

For public works to be used effectively in situations of chronic poverty, the literature is broadly consistent on key considerations, highlighting the need for:

• Long-term, sustained public works employment;
• Integration of public works programme with other developmental initiatives;
• Linkages with micro-finance, and micro-enterprise activities;
• Creation of assets which directly impact on reducing vulnerability and promoting livelihoods;
• Flexible or piece based employment, enabling participants to combine public works employment with other responsibilities and income earning opportunities;
• Higher wages; and
• Independent poverty targeting measures.

Very few, if any, public works programmes currently operating in southern Africa conform to all of these requirements. Even were they to do so, the alternative of straight cash grants may well be more efficient and desirable. Transfers can then be better targeted at those who need them, rather than just at those who are able to work for them. And any public work requirement will always have an opportunity cost: it will take people away from other work they could be doing that would more directly benefit their households.

Yet donors and governments continue to endorse public works over straight cash transfers, in the face of both theoretical and empirical evidence, on the basis essentially of ideology rather than evidence. The kind of public works programmes often advocated may not, in the context of chronic poverty, (i) avoid ‘dependency’, (ii) build assets which will promote livelihoods or growth, (iii) create jobs or (iv) target the poorest. This represents a victory of political over developmental considerations in the sphere of social protection.

---

[i] Subbarao, K., Bonnerjee, A., Braithwaite, J., Carvalho, S., Ezemenari, K., Graham, C., and Thompson, A. (1997). Safety Net Programs and Poverty Reduction, Lessons from Cross Country Experience. Directions in Development.  The World Bank, Washington, D.C.

[ii] Department for International Development (DFID) (2004).  Sustainable Livelihoods Through Inputs for Assets Output to Purpose Review, DFID Malawi, September 2004, unpublished.

[iii] Smith, W. J. (2001).  Spending on Safety Nets for the Poor: how Much, For How Many? The Case of Malawi.  Africa Region Working Paper Series, Number 11. The World Bank, Washington, D.C.

[iv] McCord, A. (2004a) Policy Expectations and Programme Reality: The Poverty Reduction and Employment Performance of Two Public Works Programmes in South Africa. Economics and Statistics Analysis Unit & Public Works Research Project, SALDRU, School of Economics, University of Cape Town.  ESAU Working Paper, Overseas Development Institute, London.