What Policy Conclusions Follow from Our Fears of “Secular Stagnation”?

Since I already published roughly the first half of What Market Failures Underlie Our Fears of “Secular Stagnation”?: The Honest Broker for the Week of January 12, 2014 in draft form last week, let me publish the second half here now a second time for those of you who would otherwise say: “tl;dr”…

What I do want to do is make explicit the underlying Summersian argument: the required real yield on private investments low enough to induce enough investment to balance savings at full employment is too low a yield to properly discourage bubbles and overleverage. The argument appears to be:

1. There are “worthy” private risky investment projects and “unworthy” ones.
2. “Worthy” risky investment projects have a relatively low elasticity of supply with respect to the required real yield.
3. “Unworthy” ones have a high elasticity of supply.
4. When safe real interest rates get too low savers who should not be bearing risk do so by reaching for yield–they stop checking whether investment projects are “worthy” or “unworthy”, and so sell unhedged puts in the hope that everything will come out all right.

This is what provokes Ryan’s despairing plea that:

are we really arguing that there aren’t enough good private investment opportunities in America?

Yes. We are.

Or, perhaps, we are arguing that there aren’t enough good relatively safe private investment opportunities in America. Or, perhaps, we are arguing that there has been a large-scale systematic failure to mobilize the economy’s risk-bearing capacity so that when additional risky private investment opportunities are launched the risks are carried by intermediaries that really should not be carrying them.

The argument seems to be that there are people who should be holding risky assets, that there are people who should be holding safe assets, and that lowering the real return to safe assets induces people who really should not be holding risky assets to go and try to hold them.

V: Second-Best: Raising the Inflation Target? Housing Finance as a Tool of Macroeconomic Stabilization Policy?

If we adopt the Summersian perspective, the first-best for this set of issues is clearly expansionary fiscal policy. There are a bunch of financial intermediaries that have (or ought to have) low risk tolerance. Hence they ought to be holding safe assets. The government can create those safe assets. The government has infrastructure and other investments it can fund with the money earned by selling those safe assets. There are unutilized resources about because wages and prices do not fall when aggregate demand is low. The fact that we have an economy with lots of nominal debt contracts means that we do not want to see wages and prices fall faster than they do because that generates universal bankruptcy. Thus given the market failures–financial intermediaries that shouldn’t be holding risky assets, savers who cannot properly monitor what the financial intermediaries are doing, wage and price rigidity, and principal-agent problems that lead to nominal debt contracts which cause deflation to lead to universal bankruptcy–borrow-and-spend by the government is first-best when the economy hits the zero nominal lower bound, or perhaps constrained first-best.

But what if the government cannot adopt borrow-and-spend to restore us to full employment because of political dysfunction? What then is the second best?

Some (e.g., Mankiw and Weinzerl (2012)) say that inflation is not just the second-best but the first-best–drop the real return on safe assets well below zero and so induce financial intermediaries to fund enough risky long-term private investment projects in order to get us back to full employment. From the Summersian perspective, this Weinzerlist solution is not a good road to walk. The fundamental problem is that too little risky investment is being financed because the fundamental risk tolerance is not there to finance it. Dropping the safe interest rate below zero pushes the safe interest rate below what is in some sense its fundamental value and so provides incentives to (a) overinvest in long-duration safe private assets that do not yield returns commensurate with their social resource cost, and (b) induce financial intermediaries to play heads-we-win-tails-you-lose with the money of their clients of low risk tolerance–a game that in its limits less to froth, bubbles, and Ponzi schemes. Dropping the safe interest rate below zero and promising to keep it there may still be second best: these distortions from too-loose monetary policy do create Harberger triangles but we currently have a huge Okun gap, and it takes a heap of Harberger triangles to fill an Okun gap. And the argument that their truly are large risks from (a) lengthening the duration of the capital stock beyond its first-best and from (b) having financial intermediaries reach for risky yield have not been nailed down to my satisfaction.

But might there be a better way?

Look at what has happened to residential investment as a share of potential output:

When I look at this graph, I find it very hard to escape the conclusion that the big bad thing going on in the third millennium is not the excess construction of the mid-2000s housing bubble–a sum of 7.5% points of annual GDP–or the working-off of that 7.5% points of GDP worth of excess construction, but rather the additional–so far–but rather the additional 20% points of annual GDP of residences not built since 2007 because of the financial crisis, resulting depression, and breaking of housing finance. Fixing whatever is keeping housing finance and thus housing construction broken would then be job #1, and so using the FHFA and the GSEs as tools of macroeconomic policy might well be superior to raising expected inflation and so trying to goose financial intermediaries that should not be funding risky and long-duration assets into doing so. (This question is, of course, complicated by the fact that housing is one of the longest-duration assets we have.)

(And at this point I should note that it was back in… July 2008… that Larry Summers and I and others had a few talks about how if the situation went south we might very well want to have a head of what became FHFA willing to aggressively use the agency as an instrument for macroeconomic stabilization. And yet Barack Obama never took even the initial steps to set up the pool table so that this might be done should it become necessary, as I think it has been very much necessary over the past four years.)

To the extent that it really is policy uncertainty about (a) the future of the macroeconomy and (b) the future of the GSEs and thus where long-run housing risk will lie over the next generation that is depressing the risk tolerance of the private market, fixing that market failure looks to be better than pushing safe real rates below zero. (It might even be better than borrow-and-spend.(

And look at what has happened to government purchases:

When I look at this graph, I see:

• No change under Reagan and the “read my lips: no new taxes” incarnation of George H.W. Bush.
• A steep decline in government purchases as a share of the economy in order to get our fiscal house in order under the Foley-Mitchell-(HW) Bush 1990 and Clinton-Foley-Mitchell 1993 deficit reduction agreements.
• Rough stabilization of the government-purchases share of potential GDP in Clinton’s second term and throughout George W Bush’s presidency.
• A small fiscal stimulus bump of 0.5% of GDP when the Lesser Depression hits as federal spending outweighs state and local austerity.
• The 2%-point plus of annual GDP collapse in government purchases since 2009.

It’s not just that we have not been expanding government purchases as a share of the economy to get close to the first best: we have been contracting them fiercely. That cannot be good.

VI: Quantitative Easing

But what if (a) we cannot attain the (constrained) first best via public-sector borrow-and-spend or (b) rebalancing housing finance, and (c) cannot attain the second best via a higher inflation target that at the zero lower bound pushes safe real interest rates strongly negative, but must (d) resort to quantitative easing to lower the spread between safe and risky interest rates, and so induce more private risky investment that way?

There is a meme in the literature that such quantitative easing policies act, like pushing the safe interest rate below zero, to increase risks by inducing financial intermediaries that have no fundamental business investing in risky assets to do so. And to the extent that quantitative easing works by raising future expected inflation, this point is as strong as the Summersian point that there are indeed risks in inducing financial intermediaries that shouldn’t be reaching for risky yield to do so–however strong that point turns out to be.

But to the extent that quantitative easing works by shrinking the spread between safe and risky interest rates by having the Federal Reserve take risk off of the deficient risk-tolerance private market and put it on taxpayers? I don’t think that argument holds much water. As I wrote before:

In the financial market there is a demand for risk-bearing capacity by firms and others who want to borrow but who cannot guarantee that they will be able to repay. The higher is the price of risk–the greater the risk premium interest rate spread over short-term Treasuries they must pay–the less they will borrow.

In the financial market there is also a supply of risk-bearing capacity by savers and financial intermediaries who want to lend, and are willing to accept and bear some risk in return from getting more than the short-term Treasury rate. The higher is the price of risk—-the greater the risk premium interest rate spread over short-term Treasuries they must pay–the more they will be willing to lend.

---

When the Federal Reserve undertakes quantitative easing, it enters the market and takes some risk off the table, buying up some of the risky assets issued by the U.S. government and its tame mortgage GSEs and selling safe assets in exchange. The demand curve for risk-bearing capacity seen by the private market thus shifts inward, to the left: a bunch of risky Treasuries and GSEs are no longer out there, as the government is no longer in the business of soaking-up as much of the private-sector’s risk-bearing capacity:

And this leftward shift in the net demand to the rest of the market for risk-bearing capacity causes the price of risk to fall, and the quantity of risk-bearing capacity supplied to fall as well. Yes, financial intermediaries that had held Treasuries and thus carried duration risk take some of the cash they received by selling their risky long-term Treasuries to the Fed and go out and buy other risky stuff. But the net effect of quantitative easing is to leave investors and financial intermediaries holding less risky portfolios because they are supplying less risk-bearing capacity.

How do we know that they are holding not more but less risky portfolios? We know because we know that supply curves slope up, and if they were holding more risky portfolios in total–supplying more risk-bearing capacity to the market–the price of risk would have not fallen but risen, and interest rate risk spreads would be not lower but higher, wouldn’t they?

So when the intelligent and thoughtful Mark Dow tweets:

I, too, think risks [of QE] overstated, but they’re non-zero. Main ones r credit leverage buildup…

I am at a loss. As long as supply curves slope up, QE does not increase but reduces the leverage of private-sector financial asset holders.

And when the intelligent and thoughtful Mark Dow tweets:

I, too, think risks overstated, but they’re non-zero. Main ones r… outsized int’l capital flows

I am again at a loss. Yes, the Federal Reserve has taken some domestic risky assets off the table. Yes, U.S. financial intermediaries and savers will respond by buying foreign assets to so deploy some of their now-undeployed risk bearing capacity. Yes, they will now bear some exchange-rate risk. But, once again, the fact that QE pushes interest rate spreads down is very powerful evidence that these capital flows are not “outsized”–that the extra exchange-rate risk U.S. financial intermediaries have now taken onto their books is less than the duration risk that QE took off of their books.

At least, that is the case as long as the supply curve for risk-bearing capacity slopes up, like a good supply curve should.

Perhaps those who claim that there are big risks to quantitative easing regroup. Perhaps they claim that financial intermediaries are perverted, and that the lower is the price of risk the greater is the amount of risk-bearing capacity they supply to the market because they lose their jobs if they don’t make at least three cents on every dollar of assets in a normal year in which risk chickens come home to roost.

In that counterfactual world, the supply-and-demand graph would look like this:

And in that counterfactual world, the Federal Reserve’s adoption of quantitative easing policies triggered an enormous expansion of the quantity of risk-bearing capacity demanded by firms and households and a huge private-sector lending boom as firms issued enormous tranches of risky bonds and as firms and households took out risky loans. In that counterfactual world, employment in bond underwriting tripled as $85 billion a month in QE was more-than-offset by an extra $120 billion a month in private-sector bond issues. In that counterfactual world, we saw a rapid recovery of housing construction and a thorough equipment investment boom as far across the U.S. as they eye could see.

That didn’t happen.

So what are the risks of QE?

It really seems to be this:

• Commercial banks traditionally accept deposits, put the deposits in long-term Treasuries, rely on the law of large numbers and on deposit insurance to allow them to always hold their long-term Treasuries to maturity, and so have a riskless and profitable business model.
• When commercial banks cannot do this, they find some way to gamble with government-insured deposits.
• ????
• LOSS!!

But this is not a source of systemic risk: because the deposits they may be gambling with are government insured by the FDIC, no run on the banking system or the shadow banking system occurs when risks come due. It would be embarrassing, yes. And the proper response to thinking that commercial banks are running undue risks with government-insured money is to send in the bank examiners–not to undertake policies that raise unemployment.

So put me with Ryan Avent, who tweets:

[The] risk [is that] of not being considered a [very] serious person by peers [unless you claim to greatly fear the risks of quantitative easing]

VII: Conclusion

This ends my rambling and unsatisfactory tour of my current thinking on the issues around what has come to be known as Secular Stagnation II…

If you asked me to rank the policies that we should pursue at this point in time, I would say that my tour through these arguments–Summersian, Blanchardist, Greenspanite, the DeLongian corollary, Aventist, Weinzerlist, and so forth–has led me to think that we should, in this order:

1. Do no harm: reverse immediately the post-2009 fall in government purchases as a share of potential GDP.
2. Do good: take advantage of the global glut of savings seeking safety to have the U.S. government fulfill market demand for safe assets, and use the funds to invest in the highest social return expenditures open to the government.
3. In addition and perhaps as an even better alternative than all-in on (2), have Mel Watt’s FHFA end policy uncertainty about housing finance and rebalance the construction sector to fill in our current 20%-point of annual GDP housing capital deficit.
4. Do not taper the Federal Reserve’s asset purchases: QE appears to me to have advantages vis-a-vis forward guidance of extraordinary monetary ease in terms of whether it creates outsized risks.
5. Extend promises that interest rates will remain zero and that the Federal Reserve will not panic when inflation rises above 2% per year until the economy has reattained full employment.

And if any of these turns out to be politically-blocked, move on to the second.

But–and let me stress this–my thoughts on this entire set of issues are tentative, half-formed, subject to error, and revision. As, indeed, are everyone’s. That is why I have called the positions “Summersian, Blanchardist, Greenspanite, the DeLongian corollary, Aventist, Weinzerlist, and so forth”–for every economist involved in the serious side of this debate understands that their positions are tentative. And everyone serious is ready to revise their views and mark their beliefs to market as more evidence comes in.