It baffles me that an area of economic research with very important policy applications has produced more than 45,000 articles and still comes to a general consensus that is completely incorrect.
I’m talking about the macro-economic effects of superannuation.
The incorrect consensus, one that is drummed into economics PhDs across the world every year, is that a fully-funded social security system has no effect on the relationship between the capital stock in successive periods, while pay-as-you-go systems reduce capital accumulation and leaves future generations worse off, with lower consumption possibilities.
Translated, that means that a pension system that requires individuals to save through their lifetimes, to fully fund their retirement, is better than a system where the current generation of workers simply pays the current generation of retirees a pension through the tax and transfer system (a pay-as-you-go model).
Why is it better? Because it encourages investment that would not have occurred if people were not forced to save. And in macro-economic models, saving equals investment and investment always means construction of new physical capital. Surprisingly you can’t simply buy an existing asset from someone else in the model. And that’s why it is wrong. Since almost all investments to fund pension schemes are transfers of ownership of existing assets – shares, bonds and property.
A simple analogy should suffice to demonstrate the point. Think of a family household with two generations – working aged children and their retired parents.
Under a pay-as-you-go system the working age children simply give the parents an allowance to spend out of their own income. Simple enough. Under the fully funded system, the working age children save for their own retirement by buying assets in the marketplace. In this case, the house (and importantly the land it resides on) is the only asset. Currently the retired parents own the house. So the children incrementally make small payments on the home to the parents, and the ownership transfers the children. The interesting macro-economic question is whether this changes the incentives for the children to work, and the owners of capital (the children and parents) to invest in improving their home.
Common sense would say that both systems offer similar incentives and have identical costs to the working generation. Under a fully funded system however, there may be less incentive to work because of the enforce delay on spending one’s earnings. Or possibly more, since people are earning less for a given amount of work, and labour supply curves are often downward sloping.
At a macro-level the cost is simply under utilisation of resources. If parents retire earlier they might be better off in terms of utility, but there is a cost to society (a very small one) of them not working. If they retire later, under both systems, then the overall cost is lower. To put it another way, output would be higher if everyone retired for a shorter period of their life. There is nothing surprising or interesting in that statement. If a fully funded system is poorly funded and social security is very low it may encourage later retirement.
What it interesting is what happens when we examine the effects of transitioning from pay-as-you-go to fully funded. The children in the household not only pay for their parents during the transition period, they must also buy the house from them. That means the parents get twice the income – once from the transfer, and once from the asset sale. Only after that generation has lived with less than their fair share does the system become fair again for the third generation.
The generation facing the transition needs to work more for the same income, and may retire later due to tightening of social security for unfunded pensioners.
The net effect of a transition between the systems is to transfer wealth from the working generation to the retiring generation. Policies such as this deserve the criticism they receive about intergenerational fairness. The same logic applies to transition from fully funded to pay-as-you-go university eduction. In this scenario the working age generation benefits by not having to fund the younger generations education.
I’m not the only one to say this. More rigorous and formalised mathematical analysis also reveals that same problem.
Pareto-improving transition to a funded system is not possible because any instrument applied to the financing of pensions in the transition phase involves intragenerational redistribution.
Please share this article. Tips, suggestions, comments and requests to [email protected] + follow me on Twitter @rumplestatskin