When do Bubbles Burst? Or how do we know that we have hit the ceiling?

There has been a wide talk about the determinants of the financial crisis as well as our ability to forecast it. Many neoclassical economists argue that we could not have predicted anything; yet some of their colleagues did (with varying degrees of success). Keynesians, (or more likely Minskyans) such as Steve Keen are said to have predicted the credit crunch and the crash that followed. Austrian economists have also had their share of predicting as well, again with various degrees of success.

This paper by Dirk Bezemer, presents a list of people who have "anticipated" the financial crisis, as early as 2005 (page 9 of the said paper). Notable names include Robert Shiller (2006), Nouriel Roubini (2005 and 2006) and Steve Keen (2006). Some of the common reasons of analysts/economists behind their predictions are:

1. “a concern with financial assets as distinct from real-sector assets,
2. with the credit flows that finance both forms of wealth,
3. with the debt growth accompanying growth in financial wealth, and
4. with the accounting relation between the financial and real economy.”

Yet, the reasons behind their estimations do not seem so clear. For example have a look at the following graph, obtained by Keen's blog:

With no intention of being disrespectful to neither of the economists/analysts whose views were proven to be much better (to be exact here would be to use infinitely better) than their colleagues, I cannot help ask the following: why should it happen in 2008 and not earlier?

Have a look at reason 3. Debt growth accompanying growth in financial wealth. How can we know that a ceiling has been reached? We can know when we are exceeding the "normal" levels but why not in 2001 when the debt-to-GDP ration was higher than the one in the Great Depression of the 1930's? (the careful reader will see that at the initiation of the Great Depression the debt-to-GDP ratio was less than 175% of GDP, approximately the same as in 1995-1996)

The next graph indicates the level of home owners in the red line (i.e. the percentage of people who choose to buy a house instead of renting one) and the percentage of sub-prime loans compared to overall loans in the US economy.

Home ownership began its decrease in 2004; yet, sub-prime loans boomed during that period. One really has to wonder who the bankers were lending to. Less people willing to buy a house in addition to more sub-prime loans... This might just be the answer, yet why didn't the break happen earlier?

Prices in the US cannot point to the answer as they are not the ones causing the crunch, but the ones receiving it. The Case-Shiller index, however, may state otherwise:

Here, it appears that the index peaked in 2006, dropped lightly until 2007 and started a free-fall in 2007. Could we have known that 2006 was the peak year of the index, ex ante? I have no definite answer to that question. Surely, some people believed that then. And history proved them right. Yet, a prediction on its own is not enough. We have to understand the rationale behind these forecasts. Why wasn't the peak of the debt-to-GDP ratio in 2004 or 2005? Foreclosures were also increased after 2007. But not before it.

Correlations, co-movements or links are easy to find ex post. We know what happened, thus we can find what we are looking for. Yet, we are subject to what is called as Survivorship Bias. We concentrate on the people or things that "survived" some process and inadvertently overlooking those that didn't because of their lack of visibility. 

In other words, we can see more clearly looking back. Keen is right that most macroeconomic models do not account for the four factors mentioned above. Yet, even if they did could they have predicted when we would reach a ceiling? Not really. We are still unaware of many issues which help in lengthening a bubble's life. These are our known unknowns. Each case is somewhat unique in its determinants and evolution, not to mention length. We could have suspected that we were experiencing one when prices exploded for a large period of time. Nevertheless, we could not know when it would burst.

Could we have know whether we were experiencing a bubble as early as 2006? Probably yes. Could we have known when it would burst? Probably not.

To quote Keynes: "Markets can remain irrational much longer than you can remain solvent"