Do Deficits Cause Crowding Out?

This is an age old question in economics and one that causes much debate. The basic idea in classical economic thought is that government deficits are not stimulative because they supplant private capital on a 1-1 or even worse basis. Now, this is a highly political debate as well in the current time period even though Democrats and Republicans are both heavily responsible for the current fiscal imbalances, but I digress on that point.

In any case, I am interested in studying the portfolio crowding out channel, which is simply the idea that more government paper on the market increases interest rates higher than they would have been otherwise. I have to confess that I have run a hideously simple regression to study the matter as a first approximation to see if anything is there at all. The basic design is this simple:

Dependent Variables: AAA 10-year corporate bond yields - 10 year treasury bond yields and BAA 10-year corporate bond yields - 10 year treasury bond yields
Independent Variables: Government deficit as % of GDP and a binary variable indicating financial crisis or recession

I am using these two credit spread measures because theoretical speaking the market for credit instruments is a layered one. Demand is highest for U.S. treasuries, next highest for AAA corporates, and lower for BAA corporates. As such, if crowding out exists, you would expect to see deficits cause spreads to widen for the corporate securities as more demand is satiated on the high end by the burgeoning supply of treasuries. Using credit spreads as opposed to nominal rates also deals with the problem of adjusting rates for inflation as it appears there is little effect on credit spreads from inflation. On a theoretical basis, we would expect to see rates on lower end corporates expand by more than on higher end corporates as they are the more marginal investment and would have to proportionately offer more in a situation where crowding out is occurring.

I ran this from 1954-2009 and got the following (I am putting the regression in plain English terms):

For AAA - 10 year treasuries:
For every 1% of GDP the deficit expands, AAA credit spreads expand 0.11%
The presence of a recession or financial crisis causes spreads to expand 0.52%
These two variables explain approximately 28% of the variation in credit spreads over the period of 1954-2009.

For BAA - 10 year treasuries:
For every 1% of GDP the deficit expands, BAA credit spreads expand 0.17%
The presence of a recession or financial crisis causes spreads to expand 0.84%
These two variables explain approximately 50% of the variation in credit spreads over the period of 1954-2009

All results were significant at the 95% and 99% significance levels.

Now, I should have probably used a GDP growth rate as opposed to a binary variable for recession or financial crisis, though that presented problems when attempting to model the presence of a financial crisis such as in 1998 or 1987. In the future I may use stock market performance and GDP growth. The problem with a great many of these variables is that they are heavily correlated with each other so that presents difficulties. For example, recessions and deficits are highly correlated, particularly in recent history.

What I think this shows is that deficits may, and I will emphasize may, have a small effect on corporate credit spreads, though there are likely other factors. For instance, the level of corporate bond issuance is not taken into account here and that would have a bearing on these measures. In the current market, corporate spreads have been narrowing to levels that would not be explained by the variables here and that could be because corporate issuance is so low that they are able to sell at low interest rates. Still, I was encouraged that the theoretical result did emerge here which is that we do see the stratified effect on corporate spreads.

As far as how to use this information, I think what it means is that you probably shouldn't worry too much about deficits' effect on corporate bond yields and by extension corporations' ability to borrow because deficits evidently explain quite little of the movement in credit spreads. I think the odds are considerable that if I had spent more than 15 minutes on this, the effects would have been even smaller.Any source