Thursday, July 29, 2010

Elizabeth Warren and Credit Crunches

Mike Konczal takes exception to my post over at e21 on the impact of 2009 credit card reform. There, I tried to argue that restrictions on credit hurt consumers in several ways, and this will be worth bearing in mind as we get a new Consumer Protection Agency (potentially headed by Elizabeth Warren).

Konczal makes a number of points:

1) The law was going into effect anyway. He argues from a Reuters piece that the new law merely enforces changes that the Fed had already planned on.

This goes against everything I’ve read on the bill so far, and I can’t seem to find any other corroborating evidence. It looks like the Democrats made a previous attempt to pass the bill in the House, and succeeded, but it took until after the election to get legislation through both houses.

Certainly, this seems to be the general assumption among both liberals and conservatives at the time—that this is a sizable change to the status quo, and will affect credit cards (either in a good or bad way). For instance, the Reuters article goes on to tell us:
Banks have repeatedly warned higher interest rates are likely to result because it will be more difficult to set rates based on the risk that customers pose. The higher rates mean less credit available for consumers, they say…

Banks may also reduce credit limits, or make it harder to obtain card-affiliated rewards, Arnold said. And others may return to charging an annual fee, now charged by only one in five cards, he said.
All sorts of other media reports (for instance, see here and here) have argued this is an ongoing issue. None of this proves that the bill is to blame, of course. But everyone seems to regard the passage of the bill as a new thing, and plenty of people seem to think that banks did change their behavior afterwards in response to the bill’s passage. Even if the bill did nothing new--people only became aware of the consequences afterwards, that's when you expect to see a change.

2) Credit card rates aren’t rationally set anyway. The big piece of evidence here seems to be that credit card rates tend to rise quickly after you miss a payment (the bill really tackled this issue by making it much more difficult for companies to raise your rates arbitrarily).

I think commentor “Mike” raised a good point on this, that Mike (Konczal) also mentions in his initial post. One reason that companies are so quick to raise rates on delinquent card-owners is because it sends a strong signal about the future ability of borrowers to make payments. Perhaps I’ve hit a shock, or who knows what else. The easy thing to go is to call in that payment right away. Credit card companies don’t really have a chance to burrow into the full aspects of a consumer’s finances, and payment choices are the big piece of information they have to go on.

I think there are several parallels in other parts of finance. For instance, suppose I’m an Investment Bank with plenty of repos outstanding. The moment I hit a little bit of a rough patch—my creditors immediately come knocking and take away my funding (potentially making a crisis more likely). Gary Gorton has referred to this a transition when loans change from being information-insensitive to being information-sensitive—and I think there are some parallels here.

You can say that relying on short-term funding if you’re a bank (or credit cards if you’re a household) is a horrible idea, because it involves taking on credit that can withdraw so quickly. Generally, I would agree. But while banks have the option of opting for longer-term credit (and choose not to take it), often these sorts of unsecured loans are the only way people have to cope with unexpected shocks (including medical charges, as Mike refers us to).

Now, it’s perhaps true that the degree of price hiking is greater than a simple model would predict. This no doubt reflects some degree of borrower inattention combined with creditor market power. But it’s consistent with credit card companies adopting a broadly rational pricing structure, which will respond within broader legal constraints. A monopolist may charge outrageous prices to begin with, but will still raise them when taxes rise.

So think of what it does for a simple credit card financing model. In the past, companies could tell apart good and bad risk individuals, based on their pattern of delinquencies, and charge higher rates to the high risk folks. Now, they have to charge rates uniformly across the pools. It doesn’t take extreme rationality assumptions to think that banks will now raise rates on both groups, to recoup the lost profits from discrimination. This is, again, what news report after news report argues to be the case.

Now, if Konczal wants to argue that a more egalitarian approach to credit risk is preferable, and the impacts of the law are good—I think that’s a perfectly defensible position to take. Though I’m inclined to favor Matt Yglesias’ point that it’s better to let competitive businesses charge as they like and make up for distributional effects through the tax code. The particular timing of the law—taking effect as it did in the middle of a huge downturn—is potentially another large concern. Given the importance of revolving credit for the payments system, the bill plausibly had a contractionary effect by discouraging spending in general. This really was the point I was trying to make: not that the law was "bad" (though I think that is the case); but that it involved a set of tradeoffs that people didn't quite realize.

3) Other things are to blame. I’ll be the first to agree that many things are going on here (of course, if one believes that higher credit card defaults result in higher rates, you probably believe that rates are being set in some sort of reasonable fashion). What drew my eye here was the fact that rates were completely flat before the law, and only took off after the bill passed, while presumably card delinquencies were rising throughout the period. Auto loans probably had a spike in defaults as well, and also did not receive any assistance from the Fed, while seeing stagnant rates. Of course, this graph is merely suggestive rather than causal in nature.
But I also mentioned other evidence Konczal doesn’t touch. From the mouth of JP Morgan’s CEO:
In the future, we no longer will be offering credit cards to approximately 15% of the customers to whom we currently offer them. This is mostly because we deem them too risky in light of new regulations restricting our ability to make adjustments over time as the client’s risk profile changes.

We reduced limits on credit lines, and we canceled credit cards for customers who had not done business with us over an extended period.

In fact, the industry as a whole reduced limits from a peak of $4.7 trillion to $3.3 trillion. While we believe this was proper action to protect both consumers and card issuers, doing so in the midst of a recession did reduce a source of liquidity for some people.
I don't put too much stock on the self-serving arguments of executives, especially ones that work for JP Morgan, but it's certainly suggestive. And though I did not post on this, there is certainly information out there backing up the idea that credit limits plunged during this time period—again, with a pattern around the date of the bill’s passage.

Figuring out the effects of law changes is tough. Warren suggests how we can do so without being ideologues:
I finally filled in the blanks, suggesting some empirical tests–ticket prices for companies that do/don’t use such clauses, changes in pricing before/after such clauses are used, evaulation of whether cost is large enough to be reflected in price, etc.
I think my analysis so far has focused on exactly these sorts of tests, noting that I have limited information and other things are going on. Konczal, of course, has responded with little new data of his own.

Wednesday, July 21, 2010

Obama Will be a One Term President

According to Bernanke's testimony today:

"The unemployment rate is expected to decline to between 7 and 7-1/2 percent by the end of 2012."

By the way, this is roughly the same unemployment rate as in 1992, when Bush won 38% of the vote, and that too after the Gulf War. It's very difficult to win elections when unemployment is that high.

If I were a Democrat at the hearings, I would push Bernanke on this issue, as their electoral futures depend on his actions. If I were a Republican, I would also push Bernanke--because unemployment this bad will create pressure for all sorts of additional government intervention on the fiscal side.

Wednesday, July 7, 2010

What is the Free Market solution?

David Frum thinks that Krugman is wrong about government spending as a solution to our problems. But what, then, is the free market solution? Cutting unemployment benefits? What about cutting taxes--even though we tried lots of that in the 2000s, without apparent success?

First of all--you can easily turn the last issue. How can government spending be the way out of our mess if we tried lots of government spending for the last ten years without getting good results either?

But more seriously--yes, there are plenty of free-market solutions for dealing with the really very bad consequences of a serious deflationary recession:

1. Wage subsidies: I've argued how to make them work here. Basically, we have unemployment because workers are less productive during bad times, yet employers can't pay them any less because wages don't tend to move down. What we can do, instead, is keep paying workers the same amount, but let employers pay less, by having the government pick up the difference. A simple payroll tax cut should do it. As the cost of hiring goes down, companies now hire more. Singapore and other European countries have tried these policies to an extent, and they seem to work pretty well.

2. Speed Bankruptcy: The aftermath of financial crises are very difficult times because of the buildup of toxic assets which stay on firm balance sheets. It is possible to wipe those out through bankruptcy, and let firms begin lending again. I describe this a bit here. This way, you don't have as many so-called balance sheet problems, and can start with a clean slate.

3. Better Monetary Policy: I'll just refer you over to Scott Sumner's blog, where he has been arguing for some time now that looser monetary policy now could dramatically boost the economy, without incurring deficits.

4. Get Capitalists to Act like Capitalists: I go into this a bit here. Basically, instead of thinking that we need to get the government to buy up all sorts of "stuff"; we should come up with targeted ways to boost the value of business assets and investments. We can tweak the tax code to do this; and the LibDem-Tory government in England is planning on doing so, even as they enact austerity cuts.

I'll grant that there are a lot of divisions among even free-market conservatives on these issues. Raghuram Rajan, for instance, would like to tighten monetary policy and make labor markets more rigid. I happen to think he's mistaken on those issues, as a similar set of policies created the Great Depression. But it shows the diversity of thought among right-center thinkers. There are options other than "spend government money."

In fact, if anything, I'd say that it's liberals who don't have a great plan for getting out of these messes. Even if austerity is a bad idea, Keynesian spending has never really lifted a country out of a bad situation. Not for Nazi Germany, not America in the Great Depression, not for Japan. Of course, it was Krugman himself who pointed out how futile Japan's attempts at expansionary fiscal policy in the '90s, and how they should switch to a monetary stimulus.

And of course, large gobs of spending doesn't work at all unless you are America--a country facing large problems, but low interest rates. Other European countries are looking at austerity measures not because they are insane, but because some of them are actually facing bond vigilantes.

The only thing I can conclude from Frum's question is that he pays no attention to free market economists at all. People like Alex Tabarrok and Edmund Phelps have argued for wage subsidies; Milton Friedman and Scott Sumner for better monetary policy in large recessions; folks like Garret Jones have argued for speed bankruptcy; and so forth. These are all very libertarian-minded people. Has Frum really never cracked open Milton Friedman?

Bankruptcy is Always an Option

I’m very sympathetic to Mike Konczal’s argument against bankruptcy as an option for winding down systematically important financial institutions. We tried bankruptcy for Lehman Brothers, and the consequences were pretty bad (no matter what John Cochrane says). I think it’s pretty clear that bankruptcy, as it exists today, is unsuited for large financial institutions, as they exist today.

Still, I’ve slowly come to the conclusion that a prepackaged form of bankruptcy, under FDIC auspices, has a lot of advantages. It would require some tinkering with the Bankruptcy Code and how financial firms work; but has the advantage of moving us to a world in which financial firms can fail, discharge liabilities in an orderly fashion, and open their doors the next day.

We have some precedent for this, by the way. When WaMu failed; the FDIC stepped in; wiped out shareholders, and sold off the bank to JP Morgan without hurting grandma's deposits. And if you look at the canonical treatments of this issue in corporate finance--it's clear what to do. When firms become insolvent, you need to write down debt and give debt-owners equity ownership.

Here’s how this sort of framework would work—I’m working through proposals advanced by Garrett Jones, Wilson Erving, and Paul Calello. When a firm signs up for bankruptcy; regulators would come in and start by immediately writing off all equity and marking down assets. Then, the bank is recapitalized by writing off junior debt, and then as much senior debt as required, and converting these into equity. Basically, we would by fiat say that all bank debt is really convertible capital, triggered by bankruptcy.

To take a concrete example--let's work through Bank of America's latest 10-Q. I suspect they are insolvent now anyway, as are many large banks--more on this later. You'll see ~2.3 trillion in assets. On the liability side, you have ~1 trillion in deposits, ~500 billion in long-term debt, and about 600 billion in other liabilities. The stocks are worth about 230 billion.

Let's say BofA runs into some trouble, and their shareholder equity (stocks) plummets to ~75 billion. For fun, let's say that they have about the same in losses. They enter bankruptcy. All a judge has to do is say--write down that equity! Those stocks are now worthless. Next, we'll wipe down, say, half of that long-term debt and turn it into equity in the new company. If that's not enough, we'll go after the other liabilities or more debt. We can go after subordinated debt harder than we hit senior debt.

That's it. Bank of America opens the next day having written down large portions of their toxic loans, and is recapitalized by their long-term lenders. We can now sell this entity off to whoever wants it, or they can stick around and make new loans. Grandma keeps her savings account as well--though, in a perfect world, I'd take a haircut of that too if grandma has more than, say, $50,000 saved in FDIC institutions overall.

This proposal manages to do everything we want bankruptcy to do, while tailoring the response for financial institutions. In bankruptcy, we recognize that firms have failed, and kick out old management. Lenders take haircuts in order of seniority, but do better in bankruptcy than in liquidation. Meanwhile, these firms—now recapitalized, and without toxic assets—are free to raise further capital and lend.

When opting for a hybrid bankruptcy route, however, it’s also essential to reform the bankruptcy code to end derivative counterparties’ exemption from the automatic stay preventing collateral seizure in bankruptcy. As both Chris Papagianis at e21 and Mike Konczal have noted, this exemption reduces the incentive for the suppliers of short-term capital to shadow banks to monitor their risk-taking.

To draw an analogy; the Asian financial crisis in 1998 was sparked by the flight of short-term foreign capital, much as the Financial Crisis of 2008 was sparked by the flight of short-term repos from Investment Banks. In the Asian crisis, foreign creditors chose a short-term financing structure because that gave them an easy avenue of exit when things went south. Similarly, repo lending and derivatives took off because these contracts would not be subject to the same considerations as other contracts in the bankruptcy process. The best way to curb a risky reliance on short-term lending is to force these contracts to be treated like others in an orderly bankruptcy process.

Finally--spinning off prop trading desks and derivative dealer operations--which should be happening to a degree under the new Financial Reform bill, should probably make this easier as well. We really want our financial firms to focus on underwriting securities offerings and extending loans throughout the financial system. It's easier to work through the bankruptcy of a bank, rather than the bankruptcy of a bank which also owns a casino.

Dealing with financial companies in some manner like this is incredibly important for three reasons:

1. We finally get to make lenders pay. For instance, here is Russ Roberts:
There is seemingly little rhyme or reason to the pattern of government intervention. The government played matchmaker and helped Bear Stearns get married to J. P. Morgan Chase. The government essentially nationalized Fannie and Freddie, placing them into conservatorship, honoring their debts, and funding their ongoing operations through the Federal Reserve. The government bought a large stake in AIG and honored all of its obligations at 100 cents on the dollar. The government funneled money to many commercial banks. Each case seems different. But there is a pattern. Each time, the stockholders in these firms are either wiped out or see their investments reduced to a trivial fraction of what they were before. The bondholders and lenders are left untouched. In every case other than that of Lehman Brothers, bondholders and lenders received everything they were promised: 100 cents on the dollar.

Well, those bondholders and lenders made plenty of money on the way up, taking risks. They need to bear those risks on the way down, instead of taxpayers. Then, maybe they will do a better job of monitoring the risk going on in financial institutions.

Lenders really are the optimal agent to handle bank monitoring, by the way. Regulators are prone to cognitive biases and regulatory capture when times are good. Owning stock in a highly levered institutions is akin to holding an option, and makes you really want to ramp up risk. Managers of firms are paid in stocks, and so also have a strong risk preferences. It's really the bondholders and depositors--who make money as long as things don't go really bad--who have the strongest incentives to watch what's going on in terms of risk-taking. But this won't happen if they get bailed out, or can handle their contracts through repos and derivatives.

2. Two is the zombie banking problem. As Anil Kashyap and others have shown, the refusal to accept banking sector losses in Japan led to a decade-long problem of “zombie banking” in which fundamentally insolvent banks lingered, without funding new investment. Saddled with liabilities, American banks are not lending. Rogoff and Reinhart, among others, have suggested that this is a large part of why recoveries from financial crises are so anemic. The Administration has argued this as well. Yet, incredibly, no one actually seems to have an idea for dealing with this debt. This was Michele Boldrin's point in his debate with Brad DeLong. It's not at all obvious that Keynesian stimulus is a valid substitute for a serious push to deal with bad debts.

2. Third is the problem of soft budget constraints. This is a concept invented by Janos Kornai initially to examine the sources of inefficiency in Communist governments. Kornai had the insight that when firms expect to be bailed out after taking large losses, they alter their behavior.
However, this logic also holds in capitalist economies for all entities—be they hospitals, government-backed mortgage giants, or large banks (or car companies, large insurance companies, airlines, money-markets, sub-national states, European countries, etc.)—that expect fiscal assistance after chronic failure. Financial institutions armed with a Too-Big-to-Fail guarantee are virtually certain to make risky investments that generate current income, with long-term costs borne by the taxpayer.

To a certain extent, the last two problems balance each other out the way our system works now. Banks may be more reluctant to lend if they retain toxic assets on their books, but may be more willing to lend if they expect to be bailed out. This is perhaps the intention of policymakers, who are eager to find ways to increase lending without forcing banks to recognize the true nature of their losses.

This is perhaps why the American response represents a substantial improvement relative to Japan. But as long as these problems remain intact, lending will remain at a knives edge caught between two politically mandated and unpleasant alternatives. It’s clear that the bailout route is a very second-best way to address problems within the banking sector, and It’s important to develop techniques for the unwinding of systemically important financial institutions that avoid creating banks that behave like those in Japan or the Soviet Union.

I don't want to exaggerate the differences between this framework and the newly-created resolution authority. As I understand it; under resolution authority, banks enter federal receivership and their books are sorted out by regulators. The process, then, resembles bankruptcy in certain respects. The big change I'd make here is simply a mandatory haircut of debt before taxpayers contribute a dime, or else mandatory convertible capital requirements triggered when resolution kicks in.

I'm sure this solution has lots of problems as well. But it's not obvious that these problems are worse than the ones we already have. And you only need to write a few pages of legislation to get here.

Rajan is wrong: Interest Rates are not Low

I’m very reluctant to criticize Raghuram Rajan, as he’s been a careful and brilliant economist over the years who has voiced prescient warnings about our current financial crisis.

But he’s been lately arguing some odd ideas unsupported by very much evidence. In his latest piece, he argues that currently lax monetary policy is a durable threat to the economy. Low interest rates encourage borrowers to take on risky and speculative investments, while piling on debt.

Scott Sumner takes down this argument fairly effectively. But there is another reason it’s wrong—interest rates aren’t low.

Check out this graph from Robert Hall:

The chart shows interest rates on various forms of investments. Ideally, once the recession began, we should observe that looser monetary policy (or other stimulus measures) should work to reduce interest rates across all asset categories—this is, after all, what Rajan and others argue happened when interest rates were too low from 2002-2004.

Instead, monetary stimulus has had a limited effect on interest rates. You see lower rates in Treasury bills (where the Fed has indeed purchased many T bills as a part of conventional policy), as well as for mortgage-backed assets (where the Fed has purchased over a trillion dollars of GSE-insured assets). But you don't see it in other asset categories.

In fact, government policy is pushing in the opposite direction in the case of credit cards. In the first part of 2009, interest rates on credit cards shot up. This is directly attributable to credit card legislation which made it much more difficult for credit card companies to adjust rates in the future. Unsurprisingly, they all raised rates immediately while cutting credit lines—sparking a credit crunch which made the recovery much harder. Thank you, Elizabeth Warren.

It’s unclear why monetary policy is doing so poorly in this case. Presumably, problems in the financial system are clogging the system, so to speak, limiting the ability of stimulus measures to impact the real economy. A big part of this is probably due to the Fed’s decision to start paying out interest on bank reserves in 2008. In a single stroke, the Fed managed to devastate the monetary base by converting non-interest bearing reserves—which are interchangable with currency—into an interest bearing security. This massively tightened monetary policy in a critical period; has made measures of the monetary base unreliable as estimates of the stance of monetary policy; and has made banks entirely unwilling to lend money as long as they have the option of keeping it at the Fed.

So we need very different monetary policy. Aside from changing the policy of paying out interest to banks to charging banks for the money they store as reserves--one option would be to alter the Fed’s charter to allow it to engage in more robust monetary policy. As far as I know—the Fed can only purchase assets under government guarantee (ie, Treasury bills and GSE-insured assets). It would be simple enough to allow the Fed to also purchase all sorts of assets as long as they are under a credit default guarantee. This would dramatically improve the ability of the Fed to alter monetary policy, completely eliminate any need for fiscal stimulus, and finally lower interest rates. And then Rajan’s complaints would at least respond to ongoing events.

Wage Subsidies

Short Version: Cut payroll taxes!

Long Version:

America is almost certainly facing a period of persistently large unemployment. Nearly 15 million people are currently unemployed, while the CBO projects that the unemployment rate will not drop to eight percent until 2012, a rate more typical of a recession.

Broader measures of unemployment paint an even worse picture. A substantial portion of the unemployed workforce are not simply between jobs, but are facing a prolonged jobless period—including six million who have been out of the labor force for more than six months. A broader measure of underemployment accounting for discouraged workers who have stopped looking for work or can’t find full time work is above sixteen percent.

The recovery, then, will in all probability be jobless. Economists Carmen Reinhart and Kenneth Rogoff suggest that countries in the aftermath of systemic banking crises face stagnant output and employment growth. This may be due in large part to the unwillingness of financial institutions facing broken balance sheets to extend lending. TARP successfully recapitalized larger financial institutions, but it did not help smaller banks that are the dominant source of funding for small and medium sized businesses. A NFIB survey of small businesses found that regular borrowers had the greatest difficulty in finding credit since 1983. While few small businesses reported financing as their biggest problem, credit constraints may inhibit small businesses from creating jobs in the recovery stage—as they have in previous downturns.

Households, too, face balance sheet problems. Steadily rising home appreciation provided the impetus for debt-fueled growth in consumption over the past decade. Now, households are more inclined to save—a good long-term strategy, but bad for jobs in the short-run.
And with interest rates close to zero, the Federal Reserve is unable to easily boost growth through conventional monetary policy. More unconventional monetary policies like quantitative easing or NGDP targeting are not being seriously considered by the Federal Reserve. Ben Bernanke’s academic work (which earned him his present job) on the Japanese experience with the zero-rate bound emphasized exactly these policy responses, yet he has repeatedly removed these from the table.

Even robust economic recovery on its own may be insufficient to restore employment to pre-crisis levels in the near future. In the past, economic growth was strongly correlated with employment generation—a statistical relation known as Orkun’s law. This dynamic held during the steep drop in output and employment during 2008. Yet it has broken down for the recovery. The CEA, for instance, found that unemployment in the fourth quarter of 2009 was 1.7 percentage points higher than historical data would suggest. Even as the economy has resumed positive growth, the Administration projects that the economy will generate less than 100,00 jobs per month throughout 2010.

In the past, wage entrenchment was a driver of employment growth during economic recoveries. Inflation eroded real wages, encouraging employers to start hiring. Higher dollar wages encouraged workers to rejoin the labor force, even though these jobs paid less in purchasing power. Ben Bernanke found this process of wage adjustment through inflation to be a powerful global driver of employment generation after the Great Depression.

Estonia saw a huge economic drop after the crash, but has also seen a large readjustment in prices and wages. This price realignment was difficult, but has paved the way for recovery.
Yet dollar inflation is today close to zero, preventing wages from naturally adjusting. Forecasters and prediction markets agree that this will continue for the foreseeable future, much as Japan’s financial collapse was followed by over a decade of stagnant prices. Without inflation, the only way to maintain full employment in the future would be for nominal wages to fall. Yet wages are “sticky” and typically only readjust upward. Meanwhile, as Casey Mulligan has noted, policymakers have introduced a number of different work disincentives, like longer unemployment benefits, a higher minimum wage, and mortgage modification efforts. The result is a deflationary overhang that will depress job creation for years.

So what? Today’s unemployment levels merely move America towards a level typical of certain European democracies. But America has not seen a prolonged period of unemployment this severe since the Great Depression.
A recent Atlantic cover article noted a number of psychological, social, and economic consequences of persistent joblessness. Being out of the job market takes a heavy personal toll on measures of mental health by separating people from the institutional moorings of a stable job. Recent graduates facing a weak job market can expect lower compensation over a lifetime.

The potential impacts on family formation are also worrying. As James Julius Wilson observed, women prefer not to marry unemployed men, but do not forsake having children—leading to single family households and an inter-generational cycle of poverty. The relatively low national unemployment rates in the past have disguised substantial concentrations of unemployment. In inner cities, Appalachian communities, and among lower-income families more generally—male employment has plummeted while single motherhood has risen.

America has not seen labor force participation rates this low among prime-age men for decades. Just as the housing bust has seen habits previously excluded to subprime borrowers (like high patterns of delinquency on mortgage debt) spread to more people; a prolonged economic downturn will have the effect of pushing the employment and accompanying social problems previously limited to low-income workers across a much bigger cluster of individuals.

There are structural problems with the economy which make unemployment an especially tough problem today, and which will probably continue to change the structure of unemployment in future recessions.

First, the level of unemployment varies substantially across levels of education. The unemployment rate among those with a college education has leveled off at 5%. Yet for those without a high school education, it is over 15%. Many workers simply lack essential skills, and will find it difficult to adapt to the restructuring in skill demand that has evolved over the course of the recession. As Mike Mandel has pointed out, the recovery out of the 2001 recession was intensive in real estate and residential construction, which was a boon for low-skilled workers. In contrast, the recovery so far has been dominated by hi-tech sectors like Internet publishing and search, and communications. The skills mismatch between the abilities that low-skilled workers have, and the structural shift in job composition, will make it difficult for any sort of economy recovery to reduce unemployment among many groups of workers.

Second, there are also substantial geographic mismatches, made worse by housing price declines. Haya El Nasser notes that the migration rate last year was the lowest since 1948, and this year is only slightly higher. Most worryingly, the percent of moves that cross county or state lines has dropped. Less than 13% of moves cross state lines, compared to almost 19% in 2004. Many moves are simply people moving out of foreclosed homes into nearby rentals. Many workers are “underwater” and owe more money on their home than it is worth—an amazing 70% in Nevada. This deters people from leaving Nevada—unemployment rate 13.4%—to, say, North Dakota (unemployment rate 4%).

The low level of migration and high level of variation in unemployment across states mean that even robust recovery across large parts of the country would leave behind large classes of workers. Workers will be unable to move in response to changing opportunities around the country. Even substantial recovery across a broad swath of the country may leave behind substantial geographic clusters. Those states which were experiencing difficulties even before the recession will now face even more catastrophic problems.

The Democrat response has been to pass a fiscal stimulus comprising tax cuts, spending on infrastructure and other projects, and fiscal relief to the states. Many progressives emphasize a second such stimulus to tackle unemployment. Yet even if one accepts estimates of these programs’ effectiveness, it is clear that many of these projects are extremely cost-ineffective at generating new jobs. Felix Salmon for instance estimates that it costs the government $300,000 to create every infrastructure job.

Government-led initiatives to boost clean energy, physical infrastructure such as high-speed trains or social projects such as education and healthcare may well be superb ideas on their own. But all are intensive in capital and specialized human capital. They will generate comparatively little employment for the low-skilled workers most affected by the downturn. Such forms of public works may have worked in the ‘30s, when “labor” was a skill-undifferentiated amorphous category. Today’s economy demands highly specialized human capital for the jobs favored in the stimulus—healthcare, education, energy—and so will be relatively poor means of generating mass employment.

The scale of fiscal stimulus and government subsidized industrial policy required to seriously tackle the long-run unemployment problem would be staggering and politically impossible in today’s environment. Concerns about America’s debt capacity have dominated discussions over the role of direct government spending in reviving employment. Yet even if the government was assured low interest rates for an extended period of time, it’s not obvious that more fiscal stimulus is the first-best way to expand employment.

Traditional conservative ideas, too, are lacking. The stimulus checks sent in 2008 were largely saved, as were the tax cuts that were a part of the ARRA bill. Policy interventions in the past year such as the minimum wage hike and greater unemployment insurance may have reduced the willingness of businesses to hire and employees to seek work to an extent; but repealing them would hardly be a sufficient boost to employment. The scale and degree of today’s employment problem demand new solutions.

The academic consensus on the problems of long-term unemployment and underemployment centers on the role of education. Rising wages from the 1950s to the 1960s were accompanied by a corresponding growth in educational attainment across all levels. Stagnating wages since that period have been similarly matched with a decline in the rate at which Americans gain education. This fall-off has been masked by rising GED enrollments. As Economist James Heckman observes, GED graduates behave and earn like high school dropouts rather than high school graduates.

Persistent unemployment is concentrated among the lesser educated. The housing bubble disguised the worsening job situation for unskilled workers through an unsustainable increase in the demand for construction. These jobs are gone and are not coming back. Manufacturing similarly has seen a long-term secular decline in workers employed—even as manufacturing output has grown.

Yet even rapid improvements in educational attainment—which would in any case be difficult to achieve—it would do little for today’s workers. Programs designed to improve the educational level of adult workers—like the GED program and other job training programs—have an abysmal track record—in part due to the fact that education comprises not only cognitive fluency, but also the attainment of certain psychological and non-cognitive skills difficult to build after adolescence. While improving educational attainment should remain an important long-term goal, it remains a too-distant option for those presently unemployed.

A more timely solution comes in the form of wage subsidies, an idea widely popularized by Economist Edmund Phelps. Unemployment happens when employers are willing to pay workers less than the amount workers demand. Wage subsidies bridge that gap by offering a government stipend to those willing to work. Effectively, a wage subsidy replicates the wage entrenchment necessary for businesses to hire again in lieu of inflation; while ensuring that workers can continue consuming.

The precise operation of that stipend can vary. A negative income tax, which provides workers a subsidy in negative relation to their earnings, would effectively subsidize work. A payroll tax cut would also have many features of a wage subsidy.

Wage subsidies in various forms have worked remarkably well in this last recession across a range of countries—in particular, in many European countries more typically known for sclerotic labor markets. In Germany, the Kuzarbeit, has offset the pay for workers facing lower hours. The Netherlands has a similar program of subsidizing workers in struggling firms. Singapore has seen enormous swings in its GDP during the crisis, yet has maintain an unemployment level in the mid-single digits—in large part due to its wage subsidies. Overall, 16 out of the 29 countries in the OECD have cut payroll taxes—mostly in Europe.

As a result, many European countries have kept employment losses far lower than America at a far lower fiscal cost. The 2009 ARRA bill, as a percent of GDP, was one of the largest fiscal stimulus in among developed countries. Yet the United States has also seen one of the greatest increases in unemployment.

However, while wage subsidies are ideal and usually prescribed during times of recession, their unemployment-busting powers make it ideally suited to dealing with the coming period of persistent joblessness.

While more direct wage subsidy programs like those common in Europe are absent from American policy evaluation, the CBO has evaluated a number of different tax credit policies. Along with increasing aid to the unemployed, the CBO found that reducing employers’ payroll taxes was the most effective and immediate job creation strategy. The EPI has estimated that a job creation tax credit could generate 5.1 million jobs over the next two years, at a cost of $27 billion.

The costs of such a targeted employment program need to be weighed against the enormous benefits resulting from fuller employment. Measured relative to other stimulus projects, a wage subsidy is a far more cost-effective way to curb unemployment.

Even if growth and employment growth unexpectedly resumed, this is a flexible policy that can be easily cut or expanded to respond to overall conditions. The downward tail risk of a ‘W’ shaped recession, or second downward bust—say by unexpected Chinese recession, or Euro collapse, or renewed economic problems stemming from the housing market—demand a flexible policy response that no current set of policy tools can provide.

A common critique of wage subsidies is that they cannot work because firms are cutting back because of falling demand, rather than worker compensation; or that they are simply uninterested in hiring unskilled workers. Yet millions of jobs were created even during the worst economic times, and many more firms were at the margin between choosing to hire or not; or choosing to fire or not. For instance, labor market guru Robert Hall argues the following:

The important feature that controls the job-finding rate is the incentives to employers to create jobs. At any given time, if the incentives are not very strong—it could happen for many different reasons—then employers will do relatively little to try to recruit workers. Job seekers will then have trouble finding jobs, will see themselves at the end of a long line of people waiting for the job....

So that’s the first thing to think about: job creation incentives.

Firm hiring decisions may radically change if companies faced a different structure for compensation. Even unskilled workers have some value to companies, and more would be hired if it were cheaper to do so.

There should be room for a bipartisan compromise on job creation. Progressive goals involving redistribution will be difficult to sustain politically if fellow feeling and solidarity are shaken by chronic economic insecurity. This loss of faith will also hit private institutions like marriage—a fact that ought to be of great interest to conservatives. If political elites fail to develop tactics to tackle the greatest period of economic chaos since the Great Depression, populism could make a comeback—with disturbing implications for trade and immigration policy. Recently, a group of economists of all stripes wrote a letter advocating wage subsidies. Hopefully politicians will follow.