Thursday, November 20, 2014

Macroprudential policy and distribution of risk

There is very little doubt that housing prices and leverage played a strong role in the global financial crisis that started in 2008. As the effects of the crisis disappear many countries still struggle with the fear that the dynamics of household debt and leverage resemble those of the pre-crisis period (e.g. Sweden).

While I am sympathetic to the idea that increased leverage and debt increases risk, I am less convinced by the theoretical justifications that are commonly used. Typically, there is an assumption that leverage and debt are associated to the notion of "living beyond our means", which makes this behavior and unsustainable. This is not correct and there are plenty of subtleties that should not be ignored that are related to distributional issues.

Let's keep things simple. Imagine a world where the stock of housing is not going to change (we have enough houses for everyone). And imagine no population growth either (we do not need more houses). In this world, someone needs to own the assets, either those who live there or those who rent them to others. Imagine there is some utility that we all derive from owning our place. So if you ask me to choose between owning and renting assuming the financial cost of both is the same, I would rather own a place.

Initially, there are some people who rent their house because they do not have access to enough credit to buy a property. This requires that someone out there owns property and is happy renting it to others. Let's now change the access to credit of some of the constrained households. Either because changes in financial regulation, interest rates or simply because of economic growth, more households have access to mortgage financing. We expect to see an increase in the purchases of houses by tenants, an increase in the price of housing and an increase in the amount of leverage. Realize that the households that are buying houses are not living beyond their means. They are simply buying a good (the ownership of the house) that they did not have access to before. Their wealth (i.e. "equity") has not gone down, only their leverage has increased. So they are not living beyond their means, they simply hold more wealth in assets that is backed by the debt they now have access to. [There will be as well other general equilibrium effects that I am ignoring it].

How much macroeconomic risk does it represent the increase in debt and leverage? It all depends on how close these new households are to being financially constrained. If now there is a shock that reduces their income and housing prices, they might find themselves financially constrained and possibly bankrupt. This risk could be related to the increase in the level of debt prior to the shock but it does not need to be. What really matters is not the aggregate level of debt but the % of households that are now in a riskier position because they hold an illiquid asset whose price might change. To assess this risk we need to know something about the distribution of income, wealth and risk. In the aggregate, if there is a shock, the change in wealth associated to a decrease in housing prices is always the same. But what matters is who owns those assets. Is it those who have access to funding or those who are at the edge of being financially constrained? It might not be unreasonable to assume that as debt and leverage increase, it is mostly due to increases happening at the margin (those with lower financial resources) but it does not always need to be the case.

There are plenty of other examples one could make to reach a similar conclusion (housing prices grow because of changes in population, move to cities, increases in ownership of a second house, reduction in the size of household,...).

In summary, this is what I learn from this and other similar examples;

1. Debt should not always be equated to living beyond our means. It could be if the borrowing goes straight to consumption spending but when it goes to purchases of assets, in particular housing, the real issue is one of leverage and potential risk.

2. Leverage through acquisitions of assets can add risk to a household as long as we believe that the prices behind those assets are risky (and more so if we already know they overvalued). Understanding the fundamental driving forces behind the changes in asset prices is a must.

3. While leverage can increase macroeconomic risk, it can also be a sign of economic development. As countries develop leverage tends to increase because of increased sophistication and size of financial markets (this is what economists called financial deepening). This is normally seen as a positive development.

4. The macroeconomic risk of increased debt and leverage depends on the distribution of income, wealth and asset purchases. While the overall risk could be related to the aggregate amount of debt and leverage, the relationship is a lot more complex than that. Macro prudential policies need to be sophisticated and incorporate distributional issues in their assessments. One more reason why understanding the distribution of income and wealth is important.

Antonio Fatás