According to a recent paper published in the Journal of Benefit Cost Analysis, most policy analyses underestimate the costs of interventions by ignoring the effects they might have on secondary markets.
For context, policy often impacts secondary markets that are for goods that are substitutes or complements to the market we are making a direct policy change in. An example given by the researchers is how a tax on coffee consumption might impact the market for tea.
In most cost-benefit analyses, these effects are largely ignored. From a practical perspective, they are quite difficult to measure, and from a theoretical perspective this exclusion has been justified by the assumption that because these markets are linked, our initial primary market analysis will inherently account for secondary market changes.
Mathematically, we can think of the total effects of a policy as an equation:
Net Effects = Primary Market Effects + Non-Market Effects + Secondary Market Effects
In most cases, the primary market effects and the non-market effects have different signs (e.g. a tax on cigarettes that improves public health). If the assumption that the secondary market effects are accounted for in the primary market effects holds, then we can ignore that part of the equation.
Unfortunately, most cost-benefit analyses do not take the required steps to satisfy this assumption. Instead, they often underestimate the second market effects and miss a key cost.
Fortunately, the researchers have developed a solution to this problem. By using fairly simple to calculate elasticities, they found a way to estimate the impact of a primary market change in a secondary market.
If analysts incorporate these simple elasticity-driven analyses into their cost-benefit analyses, they will be able to determine an estimate for the impact of a policy change on a secondary market. This means that these future cost-benefit analyses will be more accurate and can better inform policy decisions.
One interesting note the authors highlight at the end of their paper is that these effects are often quite small in practice. In the examples they provide, secondary market effects are small enough that ignoring them entirely likely would not affect a policy maker's decision.
In fact, in both examples provided by the authors the inclusion of the secondary market effects failed to move the point estimate outside the confidence interval for the initial primary market effects. In other words, there was no statistically significant difference between the model that included secondary market effects and the model that did not.
Still, these small improvements on the margins add up to better policy analysis. Although the examples presented by the authors could have potentially ignored these secondary effects, their inclusion increases the accuracy of the model. Additionally, some policymakers may be interested in the impacts on secondary markets. For instance, a Congressmember who represents a district that produces tea would be interested in the impacts of a tax on coffee, even if they look small on a national scale.
I imagine that given the size of these effects, many analysts will continue to ignore them. In a world where limited resources can go into performing these analyses, spending those resources on such a small effect will often not be deemed valuable.
This advancement illustrates the value that academics researching policy analysis can add. As cost-benefit analysis becomes more widely used, these small marginal adjustments will become more and more important. Hopefully, policymakers will be able to make use of the information that comes from these changes to the process too.