Yesterday, I noted the U.S. bankruptcy filing rate of 2.38 per 1,000 persons is at historic lows. The next question is always why. In this post, I am going to try to walk through an explanation in four graphs. The upshot is that consumer debt is low but rising. As I like to say, it takes years of study to come to the conclusion that people file bankruptcy because they are in debt. This is not to say that other factors are not contributors -- unemployment, general economic conditions -- but the primary macroeconomic driver of bankruptcy filings is the amount of debt on household balance sheets.
The first graph shows total U.S. consumer credit. The consumer credit data come from the Federal Reserve's G.19 data series on consumer credit. The Fed's definition of consumer credit includes nonrevolving debt like car loans and revolving debt like credit cards. These figures do not home mortgage debt, which the Fed tracks elsewhere but also that I have found not to be predictive of bankruptcy filing rates. Even over the time periods in the graph, it is important to account for population increases and inflation. I use population data from the U.S. Census and the Consumer Price Index from the Bureau of Labor Statistics. (And, for those who did not know that, there is your fun fact for the day: the BLS is the government entity that calculates the inflation rate.)
Thus, this first graph shows per capita consumer debt adjusted for inflation, not including mortgages. That figure currently stands at $11,471 per person -- every man, woman, and child -- in the United States. (It would be better to use the adult population as the base, but that additional complication makes data-gathering more difficult without increasing the accuracy of our conclusions.) As the graph clearly shows, after a small dip in the Great Recession total consumer credit has been climbing. That would suggest the opposite of what we have been seeing -- increasing bankruptcy filing rates.
But, the consumer credit figure includes student loans. Bankruptcy generally is much less effective in dealing with student-loan debt. Absent a difficult showing of "undue hardship," student loan debt is nondischargeable in bankruptcy. Student loan debt is now the largest component of consumer debt, more than credit cards, and has been for some time. Therefore, to get an accurate picture of how consumer debt levels are interacting with the bankruptcy system, we have to back out student loan debt from the amount of consumer credit. The second graph shows consumer credit without student loans, again on a per-capita, inflation-adjusted basis. Now, the dip in consumer borrowing becomes more pronounced. But, still since 2013, consumer credit minus student loans has been rising. Bankruptcy rates have been falling.
The third graph adds the bankruptcy filing rate to the second graph. The left-hand axis is the same as before with consumer credit minus student loans. On the right-hand side is the daily bankruptcy filing rate adjusted for population, as shown by the red line. This third graph demonstrates what we already suspected. The consumer credit graphs and the bankruptcy filing rate graphs do not appear to be following the same trends when laid on top of each other.
But, we also know that people do not file bankruptcy the moment they incur new household debt. If a borrower is on the brink of imminent bankruptcy, many avenues of consumer borrowing are foreclosed. In any event, only in the fevered dreams of neoclassical economists do massive numbers of people borrow money with the intention of strategically defaulting immediately.
How long does it take for the effects of household borrowing to show up in the bankruptcy courts. My previous research suggests two or three years. The last graph takes the midpoint -- ten calendar quarters or two-and-a-half years--and lags bankruptcy bankruptcy filings. It matches up the household borrowing of, for example, early 2006 with the bankruptcy filing rate in mid 2008. Now, the trendlines match up more closely. The scales are not the same on the both sides as the graphs are not intended to show the exact point estimates for the relationships but the general trends in the data.
I have said it many times on this blog: bankruptcy filing rates are about the amount of consumer debt. These graphs are yet another way to understand that relationship. And, because consumer debt has been rising, we should expect bankruptcy rates to follow in the coming years.
