On Tuesday, the U.S. Treasury sold a staggering $179 billion of debt to the public in a single day, a huge trade even by Treasury market standards.
The large issuance was made necessary by the fact that the U.S. had recently butted up against the debt ceiling and therefore cash on hand had been running low. However, this one-off huge issuance is only the beginning. With the U.S. deficit set to balloon due to both tax cuts and spending increases, some are seeing this as a prelude. Supply of Treasury bonds is about to increase in a big way.
How much will this matter for interest rates?
I have long argued that investors should ignore bond supply completely. As a trader, my observation has been that supply only seems to matter for a single day. After that, demand seems to be in control. That view is backed up by historical data.
The chart below shows the change in 10-year Treasury rates vs. the change in net Treasury supply. For this chart, "supply" is net of the Fed's bond ownership, so we can think of it as a change in Treasury float. The data is from 1990-now, with all changes being rolling 12-month periods.
Source: Federal Reserve
Looking at the plot visually, it doesn't seem to have much of a pattern. Indeed, if we throw a trend line in there (dotted blue line), it has the wrong slope, i.e., the fact that it heads downward as we move from left to right suggests that more bond supply means lower interest rates, which is non-sensical.
If you look closely you'll notice that periods of very high supply are actually dominated by interest rates falling. Specifically, of 23 periods where supply increased by 10% or more, in 18 of those rates actually fell.
A big part of the problem is that the periods of rapidly rising supply were also periods where the economy was weak. Yields are dropping because of the economy and that is overwhelming any impact supply is having. However, even trying to correct for that, you have a hard time finding a meaningful effect from bond supply.
I've used statistics techniques to try to isolate the supply effect from the economy growth effect using GDP, stock prices, measures of Fed policy (including anticipated Fed policy), etc. You still get either an insignificant supply effect or the equation tells you that the slope falls the wrong direction.
I've also tried using forward supply, in effect assuming that the market knows big supply is coming before it actually hits. No dice. You still don't get a meaningful impact. I won't bore you here with a bunch of regression equations, but if you are curious, feel free to e-mail me or hit me up on Twitter (@tdgraff).
Still, I don't think it makes sense to conclude from this that supply doesn't matter at all. Rather, the effect is so small that it gets swamped by other things and therefore the effect is very hard to see. To put it in microeconomic terms, the demand curve is probably pretty flat, thus a big shift in supply doesn't matter much for price. It is probably also true that the demand curve shifts around a lot, which seems to dominate the price action.
Could this time be different? We don't have a good example in recent U.S. history of a big fiscal boost happening while the economy was already strong.
In 2018 we should be getting a huge increase in debt at the same time unemployment is already at 4%. That could certainly have an effect on inflation (or inflation pressure) and thus the path of Fed policy. However, I think we should be careful about the causation. If this fiscal boost causes higher rates, it will be because of how it impacts the economy, not because the Treasury needs higher yields to entice buyers.
Therefore, I believe we can continue to safely ignore bond supply. I will continue to focus on the path of future Fed policy, which in turn means we're watching employment gains, wage growth, and inflation pressure.
Tom Graff is a regular contributor to Real Money Pro. Click here to learn about this dynamic market information service for active traders.