vineviz wrote: ↑
Sat May 23, 2020 6:11 am
Beensabu wrote: ↑
Fri May 22, 2020 8:54 pm
vineviz wrote: ↑
Wed May 20, 2020 7:53 am
rockstar wrote: ↑
Tue May 19, 2020 7:34 pm
Now, here's the tricky part: how do you rebalance without taking interest rate risk if you buy a 30 year?
You probably wouldn't be rebalancing from long-term bonds to short-term bonds, so rebalancing doesn't typically affect your average portfolio duration.
So you would not be rebalancing from long-term bonds to equities at any point? That seemed to be what rockstar was asking. How would you be able to do so without taking interest rate risk?
You have said that the high volatility of long-term treasury bonds and their general non-correlation to equities leads to lower volatility for a portfolio, and that long-term bonds are thus better diversifiers than short-term or intermediate-term bonds due to their higher volatility. How does that work without rebalancing?
- If you have the time and inclination, could you please explain this?
Well I'm not sure I understand why people think rebalancing is a problem, or even complication, for duration matching so I'm not quite sure from which angle to attempt an explanation for why it's not.
Maybe people are getting stuck on the fact that rebalancing might require selling a bond before it matures?
Duration matching doesn't actually require holding a bond to maturity, though, in order to achieve it. All that matters is keeping the weighted average duration of the bonds equal (or approximately equal) to the weighted average time to the investor's future expenditures (i.e. their time horizon).
Rebalancing from bonds to stocks or vice versa won't effect the average
duration of the bond holdings (unless the investor intentionally tries to do that).
In practice, I'd counsel investors not to structure their portfolio in such a way that they'd be likely to use individudal
bonds as a potential source of funds for rebalancing for a host of practical reasons. One of those reasons is that it could be difficult to reduce each individual bond holding proportionately to effect a rebalancing.
And the reason an investor might set up their entire bond portfolio in individual bonds is to do something more like cash-flow matching (which is like duration matching on steroids), in which case the appropriate policy would be to NOT rebalance between stocks and bonds anyway.
Generally an investor rebalances between stocks and bonds either to buy stocks after large stock losses, or to buy bonds after large stock gains. The latter isn't relevant since it doesn't involve the risk of bonds themselves. The former can be relevant. An investor is generally intending to take more risk on the stock side and less risk on the bond side, so when stocks drop the investor is hoping that the bond side A) remains a ballast and B) can possibly be used for rebalancing to purchase stocks when they are low. What the investor in long-term bonds will find, some of the time, is that this has not occurred. Even though the investor would normally be indifferent to the market value of the long-term bonds in themselves since the investor has duration matched, for the portfolio as a whole the investor may find, some of the time, that long-term bonds have declined in value along with stocks. This means that the investor will experience greater psychological remorse than they would otherwise, and the investor will have lost at least some opportunity to rebalance, purchasing stocks at a lower value with higher expected reward.
So, here is the point. The value of duration matching is premised on immunizing against interest rate risk, an important determinant of volatility in bond prices. But volatility, and interest rate risk, can still impact the investor by creating time periods where the portfolio as a whole declines more than the investor would like.
One response the boosters of long-term bonds might make is that long-term bonds generally don't decline in value at the same time as stocks. I take it that isn't your point, vineviz, and at any rate it is plagued by recency bias and insecure economic theory. Just because treasury bonds have been a safe haven during many recent declines doesn't mean it's impossible for long-term bonds to decline during bad periods, to which the 1968-1981 period can attest. There might still be something here that long-term bonds have a low correlation with stocks, so they provide diversification--the flipside of which is, yes, the long-term bonds can sometimes go in the same direction as stocks. That just tees up the question whether the asset itself is worth including in a diversified portfolio regardless of duration issues since that point is true of any number of low correlation assets, including gold.
Another response might be that long-term bonds may still serve as a ballast in the portfolio since they probably would not decline as much as stocks and, however far they fall, they also themselves have a higher expected return afterward. The direction this response would go is that the investor should be less concerned about volatility in the portfolio right now because they have plenty of time to earn returns and reach their goals.
But that leads to another question that undermines the whole point of long-term bonds. If the investor has a need for the funds far enough in the future to have a long duration, then the investor also presumably has little near-term need for the stock portion of the portfolio. The same principle that immunizes the investor from interest rate risk by giving them the duration to earn the now higher expected return should also apply to stocks. So, the question becomes, again, whether the asset (long-term bonds) is worth including in a diversified portfolio.
In other words, if rebalancing is valuable, long-term bonds undermine it; if rebalancing isn't valuable, the investor's comfort with volatility and a long duration would presumably also apply to stocks. So long-term bonds really need to be worth it to own them. At these returns, and in light of inflation risk, that is a serious problem. I really don't get why, five pages in, people are still arguing about interest rate risk instead of these other aspects of the asset.