The Economist puzzles over the--surprisingly low, IMHO anyway--yields on ten-year U.S. Treasury bonds:
Wrong-Footed: Why Have Treasury Bond Yields Fallen?: ...in the past week the price of bonds has jumped, taking the yield on the benchmark ten-year bond to a sniff under 4%—and lower than it was at the beginning of last year. This is odd. The ten-year bond yield is, in essence, America's and the world's long-term risk-free interest rate. A year ago, economists and investors were fretting about anaemic growth and deflation in America. Since then, growth and the appetite for risk have soared. So, at first, did yields on ten-year bonds. Now they are falling again sharply. Why?
The latest fall in yields was sparked by a report on January 9th which showed that far fewer jobs are being created—only 1,000 in December—than economists had expected, and raised renewed fears about an unsustainable “jobless” recovery. A parade of Fed officials said again that interest rates might have to stay low for a “considerable period”.
Although pessimism on jobs may be overdone, markets seem to believe the talk of interest rates remaining low. At the peak last September, the futures market was predicting that short-term interest rates would rise to over 3% next December. Now it thinks that they will be less than 2% because heady growth has not sparked inflation; indeed, it continues to fall. Consumer prices rose by only 1.9% in the year to December. The falling dollar should push up the price of imports, but there is no sign of that yet.
Plenty of investors need to buy Treasuries for reasons only loosely connected with growth worries. Asian central banks have been buying billions of dollars to stem the rise of their own currencies and have invested the money mostly in Treasury bonds and bills. The Bank of Japan has been the most aggressive. Last year, it intervened more heavily than any central bank in history, buying $187 billion and bringing its total foreign-exchange reserves (mostly held in dollars) to $674 billion. So great has been the desire of Asian central banks to keep a lid on their currencies that they now have reserves of $1.8 trillion, much of it parked in Treasuries.
Quite how much more they will want is a moot point. A recent rule change means that the Bank of Japan is allowed to buy another $600 billion or thereabouts of foreign currency. But even if Asian demand wanes, the appetite of another group of investors appears to be waxing: America's banks.
There are many good reasons why banks are adding to their stock of Treasuries. The first is that the yield curve—the difference between short and long-term rates—is historically very steep, which makes it much more profitable to borrow short and buy long-dated securities.
There is, moreover, a dearth of such assets. Lending to companies is shrinking. At the beginning of last year, there were $954 billion of corporate loans outstanding; now there are $884 billion. Eager to repair tattered balance sheets, companies have been investing from profits rather than borrowing the money from banks. The supply of mortgage-backed bonds is also drying up because fewer people are refinancing mortgages. Banks' holdings of mortgage bonds have fallen from almost $400 billion to just over $350 billion.
And then, of course, there are the fund managers with fewer long-dated Treasuries in their portfolios than their benchmarks require. This would have been splendid if bond prices had fallen. But they have gone up. Given how many investors want to buy them, the wait for a sharp fall could be a long one.
In the standard macroeconomic models, the long Treasury bond rate is expected inflation--currently running at 1 percent or so per year--plus the real interest rate that clears the expected flow-of-funds through financial markets. The ten-year lifespan, the 8.3-year duration of a ten-year bond at 4% means that the expectations of the flow-of-funds of rational investors pricing such a bond are taken over a long period into the future, a period long enough that we can make the simplifying assumption that business-cycle fluctuations in employment and investment are small enough to be ignored in the modeling exercise. So we look at the supply and demand for funds that households (and others) wish to save and that businesses (and others) wish to draw on to finance investment (and deficits).
And when we do look at such supply-and-demand factors, things look pretty bleak for interest rates. American households are saving little, and most of what they are saving is being vacuumed up by the federal deficit--a deficit which looks as though it will be a growing elephant in the bond market for at least a decade to come. Assuming the economy returns to full employment in the next couple of years, investment demand should be strong. Strong investment demand coupled with low national savings should produce high interest rates to balance supply and demand in the flow-of-funds--and the magic of rational expectations should translate that low supply of and high demand for finance over the next decade into high ten-year Treasury bond interest rates now.
The Economist appears to propose three answers:
Me? I reject all three of the Economist's explanations as inadequate. Instead, I see four live possibilities:
It is going to be interesting to watch...
Posted by DeLong at January 15, 2004 05:38 PM | TrackBack
Also, how would you explain why the stocks are expensive? The New Economy of 1996-2000 promised risk-free and almost limitless growth of earnings and got Dow to 10000. Today we laugh at it but Dow is above 10000 again. US stock market is a Ponzi scheme, true, but the 'Ponzi multiplier' (don't know if there is an official name for it) varies with time and I am at a loss to explain why it is so high now.
Posted by: Leopold on January 15, 2004 06:58 PMThe present value calculation (NPV) of the stock market is a function of both expectations of earnings and long term rates. It's true that the expectations of earnings are lower than they were in 2000, but it is possible to argue that this is offset by the decline in interest rates. Investing is a matter of selecting alternatives, and investments in long bonds look poor compared with 2000. So people buy stocks instead.
Posted by: Keith on January 15, 2004 08:16 PMKeith:
1. I heard 'bonds look poor so people buy stocks' before and I could never understand it. Right now the yields are low - that means the bonds are expensive, in other words people do buy them, right?
2. Maybe you are right for the interest rates. The combination of low interest rates and improved earnings could probably do it. I am still amazed at the magnitude though.
Posted by: Leopold on January 15, 2004 09:03 PMThere is a fifth possible explanation. Brad is in comunion with the cosmic all but investors aren't. If investors buy the supply side tax cuts cause only temporary deficits line, then they might predict a return to a balanced budget without spending cuts or tax increases. In Brad's view and mine, this would be irrational.
Now I think the existence of price level indexed bonds should have contributed more to the discussion (I am complaining about the maturities selected for such bonds not the discussion). I'd like to dream on about other assets. Wouldn't there be a market for budget deficit futures (asset pays 1 1 billionth of the budget deficit for fy 2008 on Jan 1 2009 say) ?
How about on the top marginal tax rate ? Total government spending etc.
Why do we, as we try to guess what investors think about all of predictions of inflation, investment, budget deficits, the capital account surplus and the cosmic all in order to figure out what they think about any of them ? That is why do investors have to guess about (long list) in order to decide what to buy ?
Posted by: Robert Waldmann on January 15, 2004 11:28 PMThis investor sold all of his bonds last June and is 100% invested in stocks. I am 20% over where I was at the top of the bubble. I update my stop losses every week (sometimes more often) and will move back into bonds when I can no longer find stocks to buy.
Of course, this is what I did last time. I missed a good part of the drop from the top bubble and still got back in too soon.
I think this behavior will be more common than in the last cycle and will moderate interest rate increases. I also think that, when the stock market turns next time, it will be very abrupt.
Posted by: Robert on January 16, 2004 12:07 AMJust to point out in the politest way that BDL knocked Treasury Secretary Snow rather severely a few months back for talking down the dollar, not because he wanted a lower dollar (because BDL wanted that as well), but because the manner in which he was doing it would increase U.S. interest rates. Well interest rates haven't increased, and so BDL is left scratching his head.
Same thing with Krugman. He predicted interest rates were going to explode because of forecasted U.S. budget deficits and - they haven't gone anywhere.
All this to say that when the economist is wrong about the market, they blame the market. It's no longer efficient, but has prices determined by "people who matter."
Really? Or is "people who matter" simply shorthand for "everyone"?
Or the market is irrational because it's a "bubble"...
Anyway, my two cents is that it's not expected deficits but realized deficits that matter, since as BDL points out, things could well change, and people could be anticipating that change. Realized deficits are not, for the moment, so large. If and when real deficits accumulate, market reaction will follow. (Which is not to say there may not be a break beforehand, e.g. because of the dollar - caused by the Fed's loose money policy - or from the housing market or any other source.)
The Economist seems right about China and Japan buying Treasuries now. BDL's comment about the enormous sums needed in the future are not pertinent to assess the current price - again because no one really knows a toss about what's going to happen in the future, either about deficits or anything else.
Posted by: Andrew Boucher on January 16, 2004 12:25 AMRereading my comment, I'm more definitive than I should be. Markets aren't always efficient, maybe this is one of those times, and maybe bonds will tank because suddenly someone will begin selling and everyone else will say, "What a wonderful idea!"
Posted by: Andrew Boucher on January 16, 2004 01:34 AMAndrew's last comment takes us back to the bubble hypothesis. The thing about bond, as opposed to equity bubbles, is that participants are more likely to understand that current high valuations are temporary.
Look at the Economist's explanation #3: this is fleshed out in the article with the explanation that for American banks, bonds are actually fairly attractive RIGHT NOW. Even at 4%, the 10-year bond offers a healthy spread over banks' cost of funds, and there aren't that many alternatives for the banks, RIGHT NOW (explanation #1). Their comfort is all the greater because of the Fed's frantic signalling that ST rates will remain this low for an extended period.
Brad's analysis of the LT flow-of-funds picture looks right, but the banks are planning to be bond traders, not bond investors. The Fed's first tightening should be interesting.
(1) A Democratic victory in 2004 is entirely possible.
I do not see George W Bush ever signing a tax increase of any sort. He remembers what happened to his dad. Praise from the professional classes and the budget hawks but defeat at the polls in November.
Also, Norquist and the professional tax cutters will never let that passt. The lessons from Clinton in 1993 will be forgotten.
The deficit is one of those things the electorate claim to care about to sound like they "care a lot". I don't think people really care terribly much about deficits until they start to affect them. People like tax cuts, and people like the goodies government provides.
(2) There is much more downward room in bonds than upward room.
(3) This is entirely possible. Free trade and productivity gains will leach jobs from the economy. The theory indicates that new technologies will generate jobs, which will surely happen, but when? I don't see nanotech becoming a successful profit centre until 2010 or beyond. So structural unemployment will worsen.
(4) I come here because I'm tired of arguing with goldbugs and wish to further my economic (self-)education.
Posted by: Shawn Pickrell on January 16, 2004 04:20 AMOK, way to many observations, but here they are.
Investors are not omniscient. Banks, in particular, have a reputation for doing what makes money now, and then having to swallow hard when conditions (in Latin America, Russia, or the Treasury curve) change. So the assessment Dave L offers of banks’ intentions may be correct, but so is his comment regarding the Fed’s first tightening. We have been through this borrow funds, buy bonds pattern before, and when the Fed tightened 25 bps, the 30-year rate rose 75 bps in short order. This is right in line with Brad’s bubble notion, but then, you can’t see a bubble, you can only guess. It is entirely likely that there is no bubble (remember the bubble talk in regard to the mid-June to early August rise in rates? – no disaster has followed).
Are we insisting too much that reality fit our theories? The Economists’ point 3 takes a pretty mechanical view of things. Long rates are high relative to Fed funds, low relative to most of living memory. The Fed is dumping a ton of liquidity into the system and it has to go somewhere. So here we are.
Brad’s points 1 and 3require that there be a fairly sharp disagreement about the medium term economic future between “people who matter in financial markets” and the economists who advise them - the Blue Chip guys include lots of bank economists. This seems an odd assumption. I am also not sure who the “people who matter in financial markets” are if not those making savings and investment decisions. They are us, I suspect, so Brad’s view requires that we think one thing when we make savings and investment decisions, another when we make spending decisions, another when we answer confidence and voter preference questions.
As regards Andrew B’s assertion regarding market efficiency, it is worth noting that markets can be wrong and still be efficient. I don’t think we can really know much about efficiency (which is mostly observed in the abstract anyhow) by simply observing prices that we think are wrong.
Brad DeLong also likes to point out that the tax cuts were a rather inefficient stimulus. If he is correct about this, why should it be so surprising that the proceeds of the deficit generating tax cuts are available to fund the deficit?
I would not expect any huge rise in real yields until some party which is currently refusing to spend their earnings changes their mind, and this I don't see yet. It could come as a result of a policy change in Japan or China, or it might come gradually as workers nearing retirement in Japan and the US switch from saving to spending down their savings.
As for point one of the Economist, it is not so hard to find companies which have 20% or so net profit margins and are using a nice chunk of earnings to grow at unsustainable double digit rates. The big box retailers (discount and specialty) are great examples of this. Not only won't they be needing funds from investors anytime soon, but before long they will have to start returning their earnings to the investors. Either that or they can compete to the point where margins are wrecked.
Snsterling - interesting train of thought that we can take further. If the issue is the effect on consumption ( or better put reduction in national savings), then consider Ricardian Equivalence which says there would be no stimulus at all. Hence, no effect on interest rates. Hey wait - - some of those defending the Bush tax cuts as not raising interest rates appeal to this theorem even as they say it will raise consumption. Of course that is absurd. But so far, consumption has been strong. Maybe not enough to make up for weak investment and net exports, but now we're getting back to the original issue.
Posted by: Harold McClure on January 16, 2004 06:32 AMThe S&P 500 PE on trailing operating earnings is about 20 and has been roughly at that level since Sept 2002 except for a small drop and rebound during the Iraq war. In my model this is fairly valued to cheap. But with rates, inflation & the risk premium so low the risk-reward ratio on stocks is highly skewed to the downside. It is almost impossible to rationally see a significantly higher PE, but it could easily fall.
The fact that this entire bull market has been driven by earnings growth rather than a PE rise
is extremely unusual. In all prior post ww II
first year of a bull mkt, earnings were little changed and the PE rose 30% on average.
My model also says bond yields should be around 5% not 4%. Maybe the economy is not going to be as strong as we all believe in 2004. Part of the reason my model has yields so high is the weak dollar-- has it gone far enough? I doubt it.
Right now we are seeing the European Central Bank
raising the possibility of intervention. but gold and metal stocks are also correcting. I think this is a counter-trend rally, not a trend change.
A 100 basis point difference between my bond model and actual yields is unusuall but it has happened before -- the model has a standard error of 50 basis points. actual yields outside
one std error band usually imply a trend reversal, ie, the recent bond rally is based on a false premis and will reverse.
I continue to believe that the risk of the dollar fall turning "disorderly" and forcing up rates significantly is very, very high.
Posted by: spencer on January 16, 2004 07:00 AMI'll put this up over on Angry Bear, too, but here are my two cents:
My two favorite explanations:
1. Vast quantities of foreign demand for US bonds is keeping bond prices high. As the Treasury report mentioned in the post below indicates, foreigners (especially central banks) are plunking down $20 to $30bn every month to buy US government bonds. I think that should be plenty to keep prices significantly higher than they would be otherwise. In addition, some of those buyers (again, especially central banks) have other priorities (such as exchange rate goals) that supercede their beliefs about the future of US inflation and interest rates.
2. Short-term rates affect long-term rates. The fact that the Fed is keeping short-term rates so low for so long does seem to cause long-term rates to fall. Yes, the Fed is supposed to only be able to affect short-term rates, but as one of my colleagues at the Board of Governors says, “I don’t know how or why, but we DO have control over long-term rates.”
Lawrence:
". I heard 'bonds look poor so people buy stocks' before and I could never understand it. Right now the yields are low - that means the bonds are expensive, in other words people do buy them, right?"
Remember that there is a third asset class - cash - which can be the marginal source of funds for both kinds of investment.
Posted by: dsquared on January 16, 2004 07:53 AMIf foreign central bank demand is keeping rates low -- and it is obviously part of the story and has been for 20 years -- why is dollar suddenly so weak. I believe primary driving force behind dollar is interest rate spreads and the spread between US-- German, UK rates is too low to attract enough capital to finance the deficit.
Consequently dollar falls.
this explains Euro, pound, etc, but not chineses other asian currencies tied to dollar where dollar is not weak.
sometime this year chinese overall trade deficit will turn negative even though will still have a big surplus with the US. At that point will china start selling t bills to finance its trade deficit with rest of the world?
Posted by: spencer on January 16, 2004 08:48 AMdsquared:
Yes, there is cash - and other alternatives. The question is however whether it is correct to say that people invest in stocks because bonds look poor when in fact bonds are expensive. You could make the same argument in reverse: people buy bonds because stocks are expensive.
Posted by: Leopold on January 16, 2004 09:45 AMBrad,
How do you mean by "This has more traction: unwinding the yield curve using the expectations hypothesis of the term structure can produce estimates of the overnight Federal Funds rate at 5% by 2008."?
Posted by: Stephen J Fromm on January 16, 2004 11:04 AMFinally, some talk about Bonds. I don't have time to participate fully but a great score card for non economist can be found at:
http://www.smartmoney.com/onebond/index.cfm?story=yieldcurve&nav=DropTabs
Check out the movie of the yeild curve from 1977 to now.
"Check out the movie of the yeild curve ..."
Nifty utility.
Inverse yield curve during the last half of 2000. Recession starts nine months after it went inverse. Classic.
Alan did it!
Posted by: Jim Glass on January 16, 2004 03:09 PMLeopold, sorry for my unclear wording. I meant that people buy stocks when bond yields look poor (ie, after bond prices have already risen). To the orginal point, I think that it is very possible to justify the Dow at 10,000 now IF you believe long term rates will stay this low.
Posted by: Keith on January 16, 2004 09:26 PMKeith,
Yes, I understand your original point, that low interest rates that persisted for a while together with the good current earnings news might account for a high stock prices (I am still amazed at the magintude).
As for the yields that have already risen. My problem is really with that argument. If bonds are more expensive than stocks, you would expect the risk/reward to converge. I do not understand why this convergence happens entirely through the stock prices rising (and the bond prices stay pretty much the same).
Posted by: Leopold on January 16, 2004 11:10 PMLeopold, you raise an intersting point on convergence. I am also often mystified by the way different asset markets work thru their relative risk/reward ratios. We have heard so much about "efficient" markets, but I often question this theory. Sometimes there are seemingly huge lags before prices get in line with each other. I thought that the dollar correction would have started long before now, and I'm still surprised that long bond yields have stayed this low for this long.
Posted by: Keith on January 17, 2004 12:56 PMFor those interested in profiting from their economic convictions and the collapse of the bond bubble, there are long dated (Jan 06) put options availible on an exchange traded fund whose value is linked to the Twenty year Government Bond. The symbol is TLT. Indeed given the apocalyptic tone of many of the postings it might be seen as a prudent hedge.
Posted by: satch on January 17, 2004 02:45 PMbanks aren't holding on to syndicated loans anymore becuase they are unprofitable. it has nothing whatsoever to do with demand for credit.
you would have to look at bank debt and bond debt oustanding to have any sense of a decent picture of credit demand since the bank and bond markets are fungible
similar story with mortgage debt. the gross amount of mortgages outstanding continues to rise. refi activity doesn't decrease/increase the amount of debt unless its cash out refi in which case it increases. home purchases continue to rise so total mortgage debt oustanding is still going up. banks don't hold on to their mortgages made anymore, the securitize it and sell it off - e.g. Fannie Mae and Freddie Mac.
at least start your hypothesizing with correct facts