I’ve been meaning for a long time to collect my thoughts about US interest rates, and where they are and should be going. As is often the case, I’m largely in agreement with Paul Krugman, at least as far as long-term rates are concerned. On the other hand, I’m a bit more hawkish in relation to short-term rates than Brad DeLong, with whom I agree on a lot of things.
I’m planning on reworking this piece as I have new thoughts, and in response to comments. so please treat it as a work in progress.
Warning: long and boring (but maybe scary) post over the fold.
h4. Interest rates
Much of the discussion has the same confused character as debates about the desirability of budget deficits. The essential problems are similar. In the short run, both interest rates and budget deficits can be controlled by governments (central banks count as part of government for this purpose). Other things being equal, low interest rates and budget deficits tend to stimulate economic activity, and are therefore appropriate when the economy is in recession[1].
In the long run, however, government budgets must balance[2]. Similarly, interest rates must be determined by the intertemporal consumption plans of consumers and the available opportunities for investment. The problem is that the long run can be a very long time coming and no-one knows when it begins. Even the 10-year bond rate is clearly affected by judgements about the policy stance of the central bank.
The link between short-term rates and long-term rates can be seen by considering arbitrage or, as it’s sometimes called, the ‘carry trade’. If a central bank is committed to keeping short-term rates at, say 1 per cent, but market forces dictate a long-term rate of 5 per cent, speculators can make as much money as they want by borrowing short and lending long.
Another way to look at is to note ten years is just (about) 40 terms of 90 days. So, on average, the annualised rate of interest on 90-day loans has to be about the same as the 10-year bond rate, sometimes higher and sometimes lower. Because of transactions costs and risk premiums, arbitrage isn’t costless, so there isn’t an exact equality. In general, short term rates are a bit lower than long-term rates, but a large gap can’t be sustained indefinitely.
The interest rate problem is therefore really two problems. First, what is a reasonable value for the rate of interest in the long run. Second, given that the short-term rate is currently below the long-term rate, how soon should it be increased.
The first question itself is in two parts. The face-value or nominal interest rate is in part a compensation for future inflation, and in part a real interest rate, reflecting the existence of profitable investment opportunities, and the impatience of consumers. The real interest rate has generally been somewhere between 2 and 4 per cent. Given that savings rates are exceptionally low at present in the US and elsewhere (denoting high levels of impatience) the rate ought to be at the high end of the range, especially if you believe that technological progress has opened up lots of investment opportunities.
As regards inflation, the combination of low short-term rates and exploding budget deficits is bound to produce an acceleration if it persists long enough. Given that there’s a significant chance of a rapid acceleration, and that the bogey of deflation has now largely disappeared, it seems reasonable to pick an average rate of around 3 per cent.
Combining the two suggests that the long-term nominal rate of interest for the US ought to be between 6 and 7 per cent. The ten-year bond rate currently just below 5 per cent and has been below 4 per cent until quite recently, reflecting the influence of very low short-term rates. But the same reasoning implies that, at some point, the ten-year bond rate is likely to overshoot the equilibrium range.
Coming to the short-term rate, there is a trade-off between the need for stimulus now and the inevitable price of higher rates in the future. There’s been a big dispute between those, like The Economist who want to put rates up immediately and those like Brad de Long who want to keep them low while employment remains depressed (one reason may be disagreement about how far the economy is from its ‘natural’ equilibrium).
What would be the consequences of an increase in short-term and long-term interest rates. Higher short-term rates would depress consumption, particularly things like purchases of new cars. This could be problematic for Ford and GM, which are essentially finance companies with a manufacturing arm these days.
But the real puzzle relates to long-term rates and mortgages. Most US homeowners are in the enviable position of having fixed-rate loans which they are free to refinance if they wish. This is an amazingly generous one-way bet, but it’s not clear who is on the other side of it. The securitization and hedging of mortgages has become so complex that no-one knows who really holds them. Some mortgages are even more favorable than this, being assumable (that is, they can be passed on to a new buyer).
If interest rates rose a lot, refinancing would stop pretty quickly, though there may be a last-minute flurry as people try to lock in rates in anticipation of an increase. Moreover, homeowners with non-assumable mortgages would be forced to stay put, since moving would entail taking on a new mortgage at a much higher rate. The big problems, though, would be on the other side of the market, where the mortgagees would have assets that, on standard analysis might have halved in value. I discuss this a bit more here.
There are a lot of other scary possibilities relating to derivatives markets. These haven’t been seriously tested since the big expansion of the 1990s. Most people seem to think everything will be OK, but no one can be sure.
fn1. Some economists (for example, supporters of the new classical model) dispute this, but I don’t intend to debate this point here.
fn2. Under standard accounting conventions, governments can run deficits forever, but in economic terms, either the budget must balance, or public debt will follow an explosive path leading inevitably to repudiation. Roughly speaking, the appropriate economic definition of a balanced budget is one consistent with a stable ratio of public debt (net of income-earning assets) to national income.
{ 9 comments }
arb 05.11.04 at 12:48 pm
From one of the linked articles: “This securitisation is sound only if the credit rating agencies have got their risk assessments right, which in turn requires that the accounts on which those assessments are based should be valid. ”
Sometimes the securitisation is made specifically because the rating agencies haven’t got their assessments right, since the securitisation is in fact an attempt to arbitrate the ratings. For instance, if you know a deficiency in the ratings systems, you can construct a security (e.g. a basket of junk bonds) which will be rated better than it should be – e.g. because the ratings agencies are bad at judging default correlations – and therefore you are able to sell it to investors at a higher rate.
Elaine Supkis 05.11.04 at 2:10 pm
We all know what is going to happen next. It is painfully obvious.
If zero interest or 1% interest was good despite inflation, we would have seen this forever and forever. Why did rates rise so fast in the seventies? Obviously. Inflation thanks to…oil.
Walt Pohl 05.11.04 at 3:01 pm
Mortgage rates are already going up, which is having the paradoxical effect (here in Seattle) of increasing the amount of refinancing and driving up housing prices. Everyone who was thinking about refinancing or buying a house is expecting interest rates to keep going up, so they’re trying to lock in the lowest rates they can.
P O'Neill 05.11.04 at 3:22 pm
I was expecting you to reinforce your policy recommendation at the end of the post — which I assume is to raise short-term rates sooner rather than later. Since many of these asset market distortions flow from a gap between short & long, and we all seem agreed that long rates are headed upwards, then it should be close the gap ASAP, right? One problem though is that the 1994 experience shows how ugly this process of raising short rates can be, even if the final outcome is OK.
wcw 05.11.04 at 3:47 pm
A couple notes. One, tighten up the writing. The fewer words you use to put across a piece like this, the better. Two, there is a lot of good analysis about mortgage rates, including the initial burst of refinancing one commenter noted in Seattle, and eventual ‘burnout’ as rates keep rising. This subject is sufficiently hoary to have made it into textbooks. Three, your statement that ‘homeowners would be forced to stay put, since moving would entail taking on a new mortgage at a much higher rate’ makes absolutely no sense. A mortgage holder at a low rate implicitly makes money as rates go up; he can sell his asset by negotiating to exchange it with the prospective buyer in return for some agreed-upon premium — as indeed was quite common way back in the late seventies, the last long rising-rate environment I remember.
Gotta run — I’ll check back to see if you update this piece.
mproust 05.11.04 at 4:22 pm
1) Why do deficits need to be balanced in the long run to avoid explosive growth? You seem to be saying that a situation in which public debt is a constant proportion of GDP, or of total private sector assets, that this situation is unstable. Do you really mean this?
2) I thought Ford and GM were social welfare agencies, in particular pension funds and health care organizations, with mfg arms.
3) What exactly happened when rates rose in 1994? I’ve heard reference to this a lot in the last couple of months. At the time I was part of a 2 career family with 2 small kids (in contrast to now where the situation has changed since my kids are now teenagers), so I was paying little attention to such things.
John Quiggin 05.11.04 at 9:20 pm
wcw – could you explain your argument a bit further? Are you saying that the mortgage can be transferred with the house? If not, the asset is lost at the moment of sale, which is my point.
mproust
(1) I would define long-run budget balance as a situation where the debt-GDP ratio is constant.
(2) I like it! And high interest rates are good for the Ford and GM pension arms
(3) the main concern is that indebtedness is greater now than 94.
wcw 05.11.04 at 9:54 pm
Yes, mortgages can be trasferred. The process is so common in the US that even after twenty years of declining rates, everyone still knows what you mean when you “assume” a mortgage.
First Google hit on “assume a mortgage” is http://www.ehow.com/how_7228_mortgage.html
However, while rising rates do not make US homeowners stay put, falling property values can dry up the market. Many people seem unwilling to ‘take the loss’ when the true value of their property drops. I observed both situation growing up in California. During the ’70s, as rates climbed ever-higher but property values did, too, the market was highly liquid. During the early nineties, in the brief respite between the ’80s boom and the late-’90s one, prices troughed and settled — and turnover dropped.
Y'81 05.11.04 at 10:08 pm
What is the evidence for your claim that annualized 90-day rates must over time be the same as the 10-year rate? I believe that, since the yield curve is generally positively sloped, short-term rates will be lower than long-term on average over any period you choose.
As to the claim that speculators can make “as much money as they choose” by borrowing short and lending long, I would say, “only if they have an infinite liquidity facility to draw.”
Comments on this entry are closed.