Treasuries: Are they Bubblicious?
One of the themes running around these days is that there may now be a bubble in US Treasury securities. Certainly there are good reasons to be concerned about this possibility. For one, the history of the last 10+ years seems to be one of assets hopping from one bubble to another (tech bubble, housing bubble, credit bubble, etc...) and some investors actually start to think of “identifying the next bubble” as a healthy strategy for growing an asset base.
At the level of economic fundamentals, it is obvious to even the casual observer that central banks all over the world and their government counterparts are expanding the supply of fiat currencies in order to 1) promote fiscal stimulus and stave off long-term depression, and 2) transfer risky assets from private balance sheets to public ones in order to ensure that banks can roll over the operating debt of solvent businesses, who will then continue to employ their workers. Although economic theory suggests that the market just needs to adjust asset prices low enough so that businesses can start growing, the real economy has nonlinear feedback loops where panic leads to hoarding and can carry over into business failure, etc.. It is true that nonlinear feedback loops can make it difficult to see how policy interventions will affect the economy, but it also means that the adjustment process is not always a smoothly functioning one, even if the economy is simply “left to its own devices.”
So an interesting question is: why would anyone buy US Treasuries (especially long-term treasuries)? Even Nassim Taleb has argued that “every human being [on the planet] should be shorting US Treasuries.”
One key investment principle is “beware crowded trades.” The main rationale is that crowded trades tend to be riskier than they appear, because small changes in expectations or news can lead to stampedes into and out of the trade. Generally, this risk is not incorporated in most participants’ analysis, and therefore they may be investing too much, anticipating a better risk-return than is actually warranted. When you see a crowded trade like this, it helps to think through how one might try to take the opposite side of the trade... i.e. what reasons might one have to buy Treasuries, particularly the long one.
So what are the reasons one might buy long Treasuries?
1) Buy Treasuries if you think that the long T-bond is offering historically high/attractive rates, and therefore it makes sense to lock those in. Obviously, this rationale doesn’t make much sense right now. With current 10Y rates at around 3.6%, there is not a great deal of room to move downward, there is fear that inflation may drive rates higher, and there is also fear that demand for UST will decline as foreign investors worry about currency risk and/or inflation. This, presumably, is why Nassim Taleb has been advocating the whole world going short (a stupid suggestion coming from him, since short positions are inherently more risky than long positions. Avoid, possibly, but I don’t want Mother Theresa and the like to be making shorting decisions with their nest eggs).
2) Buy T-bonds to immunize long term liabilities or guarantee income. This is more of a risk-control reason than a true attempt to make investment gains through Treasury exposure, but it is a perfectly legitimate reason to have long term treasury securities in a portfolio. The key here is that the investor would need to be extremely risk averse and just want to be certain of having the necessary liability or income needs covered. If they are concerned about inflation, then an allocation to TIPS or some allocation to diversified equities could be added to address that risk.
Indeed, after the current financial crisis, many pension funds may simply want to lock in much of their current liabilities and leave only a small portion of their portfolios in risky assets, given long term scenarios.
3) Buy T-bonds if you want convexity. One point often forgotten is that long-term bonds tend to have more convexity. As coupon payments come in, they may be reinvested at higher interest rates, so if inflation does come down the line, there will be some proportion of the investment that will be able to be reinvested at a higher rate. Similarly, if there is truly deflation, then there will be a capital gain available to the investor. When the future is highly uncertain, there is value to the bondholder in having positive convexity. If there is a sudden flight to quality for some reason, one will have a capital gain from the demand, plus a boost from the convexity. If inflation kicks up or treasury demand drops, convexity will cushion the drop somewhat, and the bondholder will still have positive carry from the underlying bond. Ideally, one might short out the duration risk and just keep convexity, perhaps by shorting long term treasury zero-coupon securities (which would presumably have to be engineered).
4) Buy T-bonds if you want to hedge short exposure to credit risk. As bad as things look for the US Government, things could be worse for US Corporations. If global consumption fails to pick up, company performance may turn out quite lackluster and the effect will be to widen credit spreads. In this case, the strategy would be to short corporate credit bonds and hedge interest rate risk by buying Treasuries. With a large number of mortgage resets coming down the pipeline, it may well make sense to short credit spreads this way, and the enormous pressure to sell treasuries may result in a slight premium for anyone willing to buy.
So there are reasons to buy US Treasuries, and if you find yourself in a situation where one of them applies to you, it may make sense to add treasuries to the portfolio. With so much opinion weighing against the US bond, it is helpful to think about what others may be missing. True, the US is on shakier ground than it once was, but it is still the largest and most liquid government bond market in the world by an order of magnitude or two, and whatever its weaknesses, the principal alternatives all have serious issues of their own.