Say’s law is ascribed to the French economist Jean Baptiste Say, who coined the phrase: ‘Supply creates its own demand’. This was seen as an important contribution to understanding the economy in the 18th century. This concept is however still used in modern economics, and has become especially important in healthcare economics where it has metamorphosed to become known as ‘supply induced demand’. In simple terms this is meant to describe a situation where an increase in healthcare facilities and services, subsequently cause an increase in demand. Importantly, causality runs from the supply of these services to demand, i.e. supply induced demand, or SID.

SID has now become a popular concept in the South African healthcare sector. In December 2016, the Competition Commission, within the context of the Health Market Inquiry referred to a forthcoming report in which they plan to measure SID: “In-hospital claims increases are being driven both by utilisation (increasing numbers of admissions) and cost per admission. Utilisation increases not explained by patient demographic and chronic illness changes might be due to supplier-induced demand for services. In order to assess this, we would need to demonstrate that rates are excessive relative to both past practices and to those in other countries. We would also need to show that rates of utilisation are higher where there are more suppliers of a given service, other factors being equal.”[1]

Since then, and following above inflation increases in medical scheme premiums in 2017, more parties have resorted to blaming SID for these premium increases. In the past year, both the Government Employees Medical Scheme (GEMS) and Discovery Health have made statements in the popular press, suggesting that (amongst other factors) an increase in the supply of private hospital beds is to blame.

There are some economic concepts which are helpful in unpacking these claims. Econex has done detailed research in the past on the drivers of private healthcare costs. It is important to consider what the actual drivers are, before ascribing increased utilisation and costs to supply side factors. In our previous research we have found that a large part of increased demand can be explained by benign factors such as an ageing population, increased burden of disease, technology, etc. In addition, we have considered the theoretical foundations of SID in terms of economic theory. From this research we find that a clear definition and explanation of how SID is hypothesised to operate in local markets are important when designing an econometric model to measure its effect on utilisation and costs. The economic literature further highlights the importance of correctly defining the market/sample used for the study of SID, as well as the importance of identifying and controlling for the initial state of the market (i.e. whether demand and supply are in equilibrium to begin with) before measuring SID. Finally, we find that proving causation and including the relevant statistical controls in such a measurement exercise would be crucial for the final results obtained.

Interestingly, both the comments by GEMS and Discovery Health focus on SID as potentially ‘induced’ by hospitals. This is contrary to the wealth of economic literature which focuses on SID as potentially ‘induced’ by doctors. Logically, the latter follows on from the reasoning that doctors are generally the agents in a hospital; the hospital itself cannot ‘act’ to ‘induce’ care. Economic theory is clear that one of the market failures in healthcare is the principal-agent problem. Patients are advised by doctors (agents) as to healthcare decisions, which are by definition complex. This is especially relevant in the South African healthcare sector where private hospitals are prohibited from employing doctors. It is therefore difficult to understand from a practical perspective how the hospitals will induce demand for their services. In simple language, what is the mechanism used by the hospital to induce the patient to stay an extra night, or to volunteer for an expensive procedure? The theoretical literature does ask how the doctor can potentially induce demand in a setting where there is asymmetric information (i.e. the doctor knows much more than the patient), but not how hospitals can engage in such behaviour. The forthcoming publication by the HMI has not been released yet, but it is hoped that this would be well informed by the underlying economic theory and previous attempts of SID measurement.

The Econex research note on this topic as available at:

[1] Competition Commission, 2016. Report on Analysis of Claims Dtaa and Initial Cost Attribution Analysis. Available at:, p.18.

Author/s: Mariné Erasmus and Helen Kean

Nothing in this publication should be construed as advice from any employee of Econex and should be seen as general summaries of developments or principles of interest that may not apply to specific circumstances.