By James Mackintosh 

After spending $2 trillion on government bonds in an effort to stimulate the economy, the U.S. Federal Reserve can hardly admit that it doesn't know how, or even if, it worked.

Fed Chairwoman Janet Yellen on Wednesday came as close as she's ever likely to get to accepting that quantitative easing is still poorly understood even by the experts. Explaining why the central bank prefers to set short-term rates rather than buy or sell stuff, she said it was because "we believe we understand pretty well what the effects [of rate changes] are on the economy," and so do investors. Left unsaid: No one's really sure how, or if, QE works.

This matters enormously to investors as the Fed sets out on quantitative tightening. It's starting small, allowing a maximum of $10 billion of bonds a month to mature without the money being reinvested. But in a year, that will be up to $50 billion a month -- more than the Fed bought each month during the first phase of QE3 in 2013.

The basic investor belief about QE is simple: It makes bond yields go down, shares go up and the currency go down. All make intuitive sense. Buying more bonds pushes up their price (so reducing yield). Those who sold bonds have to redeploy their money and so buy shares, while the combination of lower yields and money creation weakens the currency. On top of this "portfolio balance" effect, there's a signaling effect, because QE suggests no rate increases for a long time.

One might then assume that reversing QE will mean the opposite: Bond yields rise, shares fall and the dollar strengthens. There are at least three reasons not to worry too much about this happening -- and one good reason to be concerned.

First, economists vary widely in their assessment of the size of the impact on bond yields, but it isn't nearly as big as one might expect. Studies suggest the effect of the $600 billion of Treasurys bought in QE2 was a drop in the 10-year Treasury yield of between 0.16 and 0.45 percentage point, with big margins of error. To put that in context, yields have risen 0.23 point in the past two weeks alone.

Second, common sense makes the measures doubtful. One way economists come up with their estimates is to look at price moves on days when news about QE was revealed, which is fair enough. But the same days that bond yields fell, stock prices often fell too, the opposite of the usual story about QE. To make matters worse, look at the time period as a whole, and during QE1, QE2 and QE3, bond yields went up, along with stock prices, while the dollar was mixed -- plausibly a sign that QE had worked and boosted expectations of growth and inflation. If the markets have a similar delayed reaction to the end of QE, can we be sure which direction bond yields will move?

Third, the effects of QE clearly depend on what's going on in markets at the time. Every study finds a bigger impact from crisis-era QE than from the later versions, as would be expected. Central bank intervention helped fix dysfunctional markets, so it played a role beyond merely buying bonds. Effects on investor sentiment are likely to be bigger when sentiment is deeply depressed, too. A study by Bank of England staff and Tomasz Wieladek of Barclays last year found QE had double the effect on U.S. economic growth during the panic period than later QE rounds.

At the moment, markets are functioning perfectly well and sentiment if anything is too positive, judging by high valuations for U.S. equities and junk bonds. Reversing QE is thus likely to have less effect now than it would if markets were in turmoil.

The natural assessment, then, is that the Fed reducing Treasury holdings will push up yields, but not by much. For one to think that yields will fall would require a belief that the economy won't be able to cope with the tightening effect -- at a time when it seems to be doing fine, if not spectacularly well.

Further, we can't be sure that QE had much, or any, effect on growth and inflation. Studies mostly find a noticeable impact, but teasing out how much of that was really down to QE requires too many assumptions to be reliable. As Claudio Borio and Anna Zabai of the Bank for International Settlements put it in a paper last year, "these results generally have to be taken with more than a pinch of salt."

Stock markets tend to tumble when bond yields fall because of economic concerns, and they tend to be less upset when yields go up, although usually the reason for a yield rise is faster growth, which shareholders like, not tighter monetary policy. So it's plausible that even if the reverse of QE does push up bond yields, shareholders won't be that bothered.

There's a good reason for investors to worry, however. At the moment, the one measure on which U.S. stocks don't look expensive is when they're compared with bond yields. If a rise in yields removes that advantage, it takes away the only valuation support for U.S. equities, making it even harder to justify buying at record highs.

(END) Dow Jones Newswires

September 21, 2017 13:07 ET (17:07 GMT)