Over at Equitable Growth: As I continue to try to worry--without great success--the question of just where the increases in financial instability produced by the prolonged period of past and expected future extremely low interest rates and by quantitative easing comes from...
Two sources of risk:
- Sudden downward revisions in the expected future cash flows of underlying real assets that back financial assets.
- Sudden upward revisions in the rate at which expected future cash flows are discounted.
To recap my thinking before now:
...that... the risks to stability... c[o]me from the... shadow banking system... hedge funds, money market funds and investment banks that do not take deposits from the public. José Viñals, the IMF’s financial counsellor, said:
Policymakers are facing a new global imbalance: not enough economic risk-taking in support of growth, but increasing excesses in financial risk-taking posing stability challenges.... Risks are shifting to the shadow banking system in the form of rising market and liquidity risks. If left unaddressed, these risks could compromise global financial stability.
The stability report said low interest rates were “critical” in supporting the economy because they encouraged consumers to spend, and businesses to hire and invest. But it noted that loose monetary policies also prompted investment in high-yield but risky assets and for investors to take bigger bets. One concern is that much of the high-risk investment has taken place in emerging markets, leaving them vulnerable to rising US interest rates.... The IMF said there was a trade-off between the upside economic benefits of low interest rates and the money creation process known as quantitative easing and the downside financial stability risks... developments in high-yielding corporate bonds were “worrisome”, that share prices in some western countries were high by historical norms, and that there were pockets of real estate over-valuation...
I have come to the conclusion that those who say that quantitative easing has increased systemic financial-market risks have simply not thought hard enough about what quantitative easing is. In quantitative easing the central bank takes duration risk off of the private sector's balance sheet and onto the governments, that is, the taxpayers'. The ratio of risk to be borne to private-sector risk-bearing capacity falls. The presumption is that this makes financial markets less, not more, vulnerable to systemic risk. You could tell some kind of complex contrarian story with demand and supply curves that slope in non-obvious ways. But none of those who say that quantitative easing increases systemic risk make such arguments--and if they understood quantitative easing properly, they would understand that they need to and feel impelled to do so.
The argument that ultra-low interest rates and the anticipated continuation of ultra-low interest rates for a considerable time period raises systemic financial risks is less mired in the, well, mire. But it, too, is not obvious. An economy sinks into depression when households, savers, and businesses in aggregate believe that they are short of the assets they need to hold to ensure liquidity--that after subtracting off assets they are holding as savings vehicles they do not have enough cash and enough safe nominal assets that could be pledged to immediately raise cash. As a result, the aggregate of households, savers, and businesses try to cut their spending below their income in order to build up their liquid cash and safe collateral balances; but since my income is nothing more than your spending, they fail and so production, income, and spending fall until the private sector finds itself so poor that it no longer seeks to build up its liquid cash and safe collateral balances.
In such a situation the government, by trading its cash and its safe collateral liabilities for risky financial assets and four currently-produced goods and services both:
creates more of what the private sector wants to hold, and so reduces or eliminates the gap between current and desired holdings of liquid cash and safe collateral.
lowers interest rates and so increases the value of the future relative to the present, providing a direct financial incentive both to spend now on the creation of long-duration real assets and to spend now out of what are now more valuable anticipations of future income.
The first-order effects are on aggregate demand--are on resolving the disequilibrium excess demand for liquid cash and safe collateral that creates the gap between planned spending and expected income. The effects on financial stability are second-order. They are:
On the one hand, higher asset valuations and higher levels of production and income greatly reduce the risks associated with financial assets backed by real wealth of one form or another.
On the other hand, the tilting of the intertemporal relative price structure greatly increases the incentive to create long-duration financial assets--which will inherently be speculative, and some of which will partake to some degree of the unhedged out-of-the-money put or even the Ponzi nature.
Which of these effects will be larger? For small reductions in interest rates, the first-order effect on the value of existing collateral assets in making finance safer will outweigh the second-order creation of new long-duration assets in making finance riskier.
Which of these effects will be larger? To the extent that prudential regulation is effective--or even exists--the range over which reductions in interest rates will improve stability is larger.
Which of these effects will be larger? To the extent that the economy is already flush with long-duration financial assets--which it is--the range over which reductions in interest rates will improve stability is larger.
The first-class study of this I know finds no evidence of the IMF's contention that policies of ultra-low interest rates have laid the foundations for increased risks of systemic financial instability in the United States. Outside the United States? Yes, times of low interest rates in the core are times of opportunity--cheap financing is available to finance economic development--but also times of danger--are their financial markets robust enough to control and manage the hot-money fluctuations?--in the periphery. But how much weight does the argument that prudential policy in the periphery may go wrong have in militating against policies that--correctly--aim at appropriate internal balance in the core?
I am now more inclined to view worries that ultra-low interest rate and quantitative easing policies raise risks of future financial instability as the last-gasp argument of the austerians--as one more attempt to find an argument, any argument, to justify universal bankruptcy and the war on the Keynesian Mammon of Unrighteousness.
So should I rethink this?
In conversation, the wise young whippersnapper Gabriel Chodorow-Reich assigns me to the camp of the "institutions" view: that the danger is that systemically-important financial institutions become severely distressed; that conventional stimulative open-market operations give them more reserves to back loans; that stimulative quantitative easing takes duration risk off the private sector aggregate balance sheet; and neither of those obviously raises the risk that systemically-important financial institutions become severely distressed--rather the reverse, because without truly perverse demand curves systemically-important financial institutions should hold some of those increased reserves, and part with some of the assets with duration risk that quantitative easing takes off the table.
I agree with this characterization. As I see it, first-order risk is risk to the underlying, and the more QE, the less risky underlying the private-sector holds, and the more forward guidance, the larger the expected pool of reserve assets to keep financial institutions out of distress. Second-order risk is the risk not of a fall in underlying cash flows but of a big upward move in the market discount applied to risky cash flows--a sudden fall in the risk-tolerance of the market that overwhelms VAR limits. Why should QE that reduces the pool of risky assets and forward guidance that raises the expected pool of reserves raise the variance of risk-tolerance innovations. Third-order risk is when institutions reaching for yield have demand curves that slope the wrong way--the lower the spread, the more risky assets they hold. But this seems to require that QE and extended guidance have triggered an enormous wave of risky-asset issue if institutions are going to end with riskier portfolios.
But, he points out, there is also a "spreads" view, which as I interpret it seems to be:
- Long-term risky interest rates mean revert to normal.
- Stimulative monetary policy--present and expected future--lowers long-term risky interest rates.
- Quantitative easing lowers long-term risky interest rates too.
- Thus when long-term risky interest rates revert to their normal, they must undertake a larger jump.
- And that larger jump upon normalization causes greater losses and so creates more of a risk of severe financial distress.
And on this I have three thoughts:
I. It is not obvious to me what this "normal" that interest rates mean-revert to in the long run is. Looking at the nominal:
and at the real, with assumed adaptive inflation expectations:
perhaps you could argue for a 2.5%/year real Treasury 10-Year rate "normal" in the 1960s, and perhaps you could argue for a 3.75%/year real Treasury 10-Year rate "normal" from 1985-2000. But since 2000? Before 1980? And it is not at all obvious that the magnitude of the increase in long-term interest rates--nominal or real--in a business cycle recovery bears any relationship at all to the deviation from whatever you had previously thought of as "normal".
The underlying argument really does seem to be: "expansionary policy, extended guidance, and quantitative easing have pushed long-term interest rates down below normal, and they are coming back to normal, and when they do come back they will come back one basis point for each basis point the Fed artificially pushed them down, and it will be bad." The assumption seems to be either the policies are completely unwound--either by the Federal Reserve or by the market in some way--so that the Federal Reserve actions have no long-term permanent effect on the intertemporal price and risk structure and no medium-term transition-path effect on the speed of return to whatever the new normal turns out to be.
And I continue to find it impossible to put this argument into any kind of asset market demand and supply framework. And that makes me gravely suspicious.
There have been a number of arguments made in macroeconomics since 2006 that have been correct. There have been a number of arguments made in macroeconomics since 2006 that have been wrong but not implausible. And there have been a number of arguments--expansionary austerity; the 90% of annual GDP debt-limit cliff; quantitative easing necessarily and rapidly leading to debasement, de-anchoring of inflation expectations, and eventually hyperinflation, the zero fiscal multiplier--that seemed to me to be neither correct nor plausible but wrong but simply to lack any empirical or theoretical support whatsoever and to be the product of people who simply:
- had not done their homework,
- were grasping at straws for any argument that would support a previously-arrived at knee-jerk policy preference, and
- had very low intellectual standards.
The question is: what keeps the argument that quantitative easing and extended forward guidance are greatly increasing the risks of financial instability from falling into this third category? I am waiting for an answer. And I am not finding one.
In short, I am demanding that those who see quick unwinding of balance sheets and quick raising of interest rates as policies that would reduce rather than increase the risks of financial distress open up their black box and at least make an argument about what lies inside.
The stakes are large indeed. Let me turn over the microphone to the very smart Joe Gagnon:
Joe Gagnon: Yellen vs. the BIS: Whose Thesis Makes Better Sense?: "Janet Yellen is right to resist diverting monetary policy...
...from its primary objective of stabilizing economic activity and inflation: Everyone agrees that it is essential to fix the flaws in financial regulation and supervision... Governments need... to place limits on leverage and... raise capital standards. [But] setting monetary policy on any basis other than the stabilization of employment and inflation is more likely to harm financial stability than to help.... Even if it were clear that loose monetary policy feeds asset bubbles (and my colleague Adam Posen has argued convincingly that it is not), it does not follow that tighter monetary policy is necessarily the right response to a bubble. The damage caused by a bursting bubble arises from the deadweight costs of bankruptcy and the panic engendered.... The solution is to reduce debt and increase equity... and to ensure that systemically-important financial institutions are well capitalized.
During the housing bubble, restrictions on leverage needed to be tightened.... [But] to prevent the economy from falling into recession, the Fed would have needed to lower interest rates, not raise them, in order to encourage firms and households with healthy balance sheets to spend more. BIS economists point to historically low interest rates as a sign that policy is dangerously loose. However, there are many reasons why returns on safe instruments should be low or even negative now in real terms.... When the equilibrium required return on assets is at a historical low, then asset prices of necessity will be historically high. This does not imply that we are experiencing a risky bubble. Sweden recently provided a clear test of the dangers of diverting monetary policy from its primary function to fight a perceived bubble.... Its policy tightening may have been counterproductive even in terms of its original financial-stability motivation.