Tuesday, April 14, 2009

The Recovery and the U.S. Consumer

I wrote some time ago (here) about the over-leverage of the U.S. household. As we think about a possible rebound in activity, I think it is important to keep this fundamental misalignment in mind. IF I am right and this is the recession where we begin to make progress on these imbalance THEN no recovery can occur while the U.S. household still consumes a disproportionate amount of GDP

In my mind, the easiest way to understand the over-leverage of the U.S. household is to look at the ratio of consumption to GDP. This number is essentially the amount of U.S. income devoted to consumption as opposed to investment or exports. The higher is this number the more households have borrowed to support their consumption.

The share of income allocated to consumption in the United States has risen from 62 percent in 1980 to over 70 percent in 2008. Over 40 percent of the rise has occurred since 2000. This level of consumption is way too high. It is high by historical post-war U.S. standards and it is high by contemporary international standards.

Take a look at the inset box. Of the major industrial economies, the United Kingdom has the highest ratio. At 64 percent of GDP, however, the United Kingdom has only reached levels of consumption similar to that of the early 1980s United States.

To make a stronger link with leverage, one can glean similar insights from the current account deficit of the United States. Over exactly the same period that consumption as a share of GDP has been rising, the current account deficit of the United States has been falling. The match is close even to the point of matching the three percentage point jump since 2000. The current account deficit represents real indebtedness to foreign economies. It does not matter whether this debt is incurred because oil prices have risen or because we buy a lot of TV’s.

The share of income allocated to consumption in the United States has risen from 62 percent in 1980 to over 70 percent in 2008. Over 40 percent of the rise has occurred since 2000. This level of consumption is way too high. It is high by historical post-war U.S. standards and it is high by contemporary international standards.

The share of consumption to GDP in the United States must fall and it seems that we are, finally, at the turning point. There are only two ways for this adjustment to occur: either consumption must fall or the other components of GDP must rise.

How far does consumption have to fall? It depends on how far the ratio of consumption to GDP must move. I believe the ratio is likely to return to 1980 levels but a strong case can be made that the ratio need only fall back to the level of the late 1990s.

Assuming no growth in other components of GDP, consumption would have to fall 3.2 percent per year to bring the ratio back to its 2000 level within five years. For the ratio of consumption to GDP to fall back to its 1980’s level within five years, consumption would have to fall 7.4 percent per year. If we allow 10 years for the ratio to return to its 1980’s level, consumption only needs to fall 3.2 percent per year.

These falls are huge and well beyond anything in the modern historical record for industrial economies with the possible exception of the Great Depression. Only Argentina (and perhaps some of the Asians during the Asian Financial Crisis, though I have not checked) suffered falls of this magnitude. In 2001 and 2001, Argentina experienced repeated double-digit falls in consumption.

Of course the consumption numbers are all relative to the other components of GDP. Consumption does not have to fall as much if other parts of the NIPA accounts are growing. However, because consumption has such a large share of output, the other components would have to grow very fast indeed to keep consumption from having to fall at all. In addition, I have trouble imagining components like investment expanding when consumption is falling at these rates.

These numbers are scary.

And these numbers are just as scary as when I first posted this idea. The U.S. consumer has a long way to go and I do not believe we can have a sustainable recovery before they have at least begun to move along this path.

Sunday, April 12, 2009

The Effectiveness of Japanese Stimulus

David Bath in a question to this post asked about Japanese stimulus. “The package announced last week "mamizu" of ¥1.1 trillion, looks to lift the forecast for Japanese GDP growth in 2009 by around 1.1-1.3%. That is, it will improve growth from -3% in 2009 from the -4.3% expected before the package was announced.”

My response: I saw the monumental change in some forecasts for 2009 once the stimulus was announced. But, these must just be knee-jerk reactions to the stimulus that assume effective fiscal stimulus. I already knew I was in the minority on multipliers so these numbers don’t surprise. But, we can use these numbers to get assumed paths for GDP under effective stimulus and for my case. Then at the end of the year, once the data is in, we can compare and judge the effectiveness of this particular package.

Let’s think about what a reasonable path for Japanese GDP during 2009 might be and let’s think about when the stimulus might reasonably be expected to come online.
The stimulus is expected to be passed this month; given the slow pace of implementation with the previous two packages, I think mid-summer is an optimistic date for the actual spending to begin. So, all of the stimulus from this round must fall into the third and fourth quarters.

Pre-announcement, most analysts seemed to be thinking something like the following for Japanese GDP: -17% Q1; -4% Q2; -1% Q3; and between 0 and 2 for Q4. I am assuming from your 4.3 percent figure that you were expecting an average of positive two percent growth in both Q3 and Q4 at least if your forecast for the first half is similar to others. (All figures annual rate.)

Working from my assumption of your baseline, the stimulus would have to boost Japanese growth to 5.4 percent in both Q3 and Q4 to reach -3 percent growth in 2009. That’s a boost to GDP in the second half of 1.6 percent over the baseline. Assuming that half of the stimulus is spent this year (about 1 percent of GDP), the assumed multiplier is 1.6. This seems high to me.

But now we have a baseline of comparison. If the multiplier is large, we expect very fast GDP growth in the second half, something on the order of 5 percent. I expect that if the Japanese government does try and spend all 10 percent of the money they say they have allocated (they have not spent much yet), the second half will remain tepid. I would expect zero or negative growth in the second half. Two predictions, very different. Let’s wait and see what the numbers are later this year. Then we can do a retrospective on the effectiveness of this stimulus. We still won’t have the counterfactual but at least we have our forecasts.

Big positive growth numbers  more likely fiscal stimulus is effective
Zero or negative growth numbers  more likely that it is not

That’s the main question: Here are a few other misc. questions from his post (in italics).

Surely if there were limits to the overuse of consecutive stimuli and monetization we would be seeing it in Japan? Long term rates remain very low.

So, we don’t yet know what the limits in Japan are. While they have announced large packages, as you noted about 10% of GDP, only a tithe of this has been spent so far. And virtually nothing is showing up in real government spending. I would expect that neither the costs nor the benefits of government stimulus would happen until the money is actually spent, not just allocated.

The Bank of Japan plans to soak up a lot of the issuance that will facilitate the extension of stimulus.

This is monetizing the debt. I have no problem with the action, although I would place higher odds on effective monetary policy if they were not simply providing funds to the private sector. Japan certainly has room to print money if they desire. Yet, printing money is an inflation tax and it lowers the real value of bond holdings.

Why does Japan seem so relatively stable?

I would argue that it does not. Japan’s banks entered this recession with almost no exposure to subprime assets and very little exposure to asset-backed securities. At the start of the crisis, the bank’s were very well capitalized and indeed provided a substantial amount of capital to U.S. and European banks. Yet, at least until the rebound, Japan’s economy seems to be suffering more than any other industrial economy with the possible exception of Iceland and Ireland.

Already in the fourth quarter (after the first round of stimulus), GDP fell more than 12 percent. Most analysts expect a fall at least that bad in the first quarter.

Saturday, April 11, 2009

The New Plan to Save the Auto Sector

In a press release dated April 9, the Administration announced a new plan to help the auto sector. By June 1, they plan to purchase 17,600 new vehicles for the Federal fleet. The purchase has two purposes. The first is stimulus. By accelerating the purchase of the cars, the administration hopes to support the auto industry. And since the contract is for purchase of GM, Chrysler, Fords, it helps American manufacturers. The second is fuel efficiency. By trading in less efficient cars for more efficient, the government can reduce both its gasoline consumption and its carbon emissions. It’s not clear from the release how quickly this replacement would have happened in the absence of the stimulus plan. So, it is difficult to measure the full benefit of the plan.

Cars are currently selling at a pace of about 9 million per year. The federal purchase then provides a boost of about 0.4 percent in the first six months of 2009, helpful but not a solution to all of the car companies’ woes.

This, by the way, is exactly the type of stimulus spending of which I approve. The spending brings forward government demand to the period where it is most needed. The additional costs of the program are small and consist mostly of reducing the average age of the government’s fleet. And, the scale of the program, while significant, is unlikely to have a first order impact on prices. That is, the crowding out effect is likely to be quite small. If only the government could spend the full trillion in such a manner.

Of course, the plan does not stop here. The administration wants a tax credit for anybody who wants to trade in an old car in exchange for a new one. This plan would be sold as a fuel efficiency plan but would of course have the primary intent as stimulus.

What are the impacts of a broader plan?

Many countries around the world are currently implementing similar plans. The German government offered a €2,500 incentive for anybody who traded in a 9 or more year-old car for a new vehicle. I don’t have the numbers in front of me but the plan was apparently quite successful. New car sales surged in Germany. Similar programs with similar success rates were implemented in France, Italy, and, I believe, Brazil.

To pro-stimulus economists, these programs are a glowing success. The government implements a rebate program designed to boost demand and it works. We can easily measure the number of new cars sold, compare it to some baseline, and we can judge the program a success.

What we can’t measure is the costs of the program. I don’t mean the fiscal cost, that is easy. I mean the economic costs. The government has introduced a real relative price change in the economy—new cars are cheaper relative to other goods.

Real relative price movements change the allocation in the economy. People substitute towards cars. The substitute away from …

That is the problem. We don’t know what good or service households stop consuming because of the cheaper cars. It has to be something. The government may be flexible about its budget constraint the household cannot be. (The government can raise taxes or print money; the household cannot simply demand higher wages.)

The German new car sales numbers are for March. Retail sales around this time are likely to be smaller than they would otherwise have been. Once again, we have no counterfactual and so cannot estimate the crowding out effect. I suspect, though, that we will find German and French retail sales disappointing relative to Dutch, Danish, or Spanish sales.

Takeaway: The administration’s plan to purchase new energy efficient cars is a good plan. The scale is right and the plan is likely to help the car companies some. Rebates to induce sales seem ideally suited to replicating the program only on a bigger scale. These programs, however, create important distortions in the economy; distortions we cannot fully measure.

Moreover, even the hint or the rumor of these programs distorts the economy today. If you have not yet read Casey Mulligan’s blog on this topic, do so now (put the link here). The rumor causes households to postpone the purchase of cars. Because my trade in might be worth substantially more tomorrow, I will wait to purchase a car until tomorrow. Since we do not know the parameters of the program, this wait and see effect will influence a large cross section of potential new car purchasers. Just as now, many taxi drivers in Chicago are waiting to purchase their new cars because the city council might subsidize them at some point in the future.

Here the German’s beat us to the punch in terms of efficiency. Almost before German citizens knew the policy was coming, the German government passed the rebate.

Thursday, April 9, 2009

How will the Japanese economy recover, whenever it actually reaches a bottom?

Japanese production is at a 30-year low. The picture below looks like a data error. Modern countries do not erase 30 years worth of growth, however tepid, in four months. Really, modern countries don’t just collapse. This fall in IP is larger than any other fall on record—and I don’t just mean for Japan.

But, there was no war, no industry destroying earthquake. In four months output has fallen, Japanese industrial capacity is more or less unchanged. The factories are still there they just aren’t running. Indeed, according to the employment statistics, the factories are still more or less fully staffed. Japan could bounce back and produce at 2008 levels tomorrow.

I hope this happens. I think it is an open question as to what happens next.

Generally, when a country has a large fall in production, IP falls for a while, reaches a low point, and then starts to grow from that new jumping off point. Generally at a faster pace than before the downturn, but importantly, IP also does not return instantly to its previous level. In the data, it seems to take between one and two years for every 10 percent fall in output.

Here is a picture of Mexican IP surrounding the Peso Crisis. During the crisis, IP fell about 13 percent; 17 months after the collapse IP returned to its pre-crisis level and within 24 months had returned to its previous trend.

The 2001 crisis in Argentina is perhaps the closest comparison to Japan’s current situation. IP fell almost 20 percent from peak to trough. The economy hit bottom and grew robustly. Nonetheless, IP took 32 months to return to its pre-crisis level.

During the Great Depression in the United States, IP fell about 50 percent and took ten years to return to its pre-crisis level. A cross-section of Asian economies experienced large falls in IP during the Asian financial crisis. My brief scan of that data indicates the same rule of thumb, although they seem to have been on the faster end returning in levels in about 1 year per 10 percent fall.

Japan’s Path to Recovery

In the rosiest of scenarios, when the economy is ready to recover, Japan bounces back to its old level. Let’s say this happens this month. Then the world goes on. Japan as a country takes a 15 percent pay cut for the year, not enjoyable but livable. This could happen. As I said above, Japan has its industrial capacity intact.

But, output does not just fall. Despite Kevin Warsh’s dismissal of the current downturn as a Panic, output does not fall without cause. I continue to believe that the downturn is at its heart a manifestation of the world’s need to restructure (read this post).

Rosy scenario number two is an Asian Financial Crisis style bounce back: one year for every 10 percent fall in output. Take a look at the next picture. To bounce back in 4 years, Japan’s economy has to grow at almost three times its average growth rate of the 1980s—remember when Japan was taking over the world. That bounce back looks no more believable than the drop.

The gloomy scenario is that Japan grows at its average rate of the 1980s. In this case, it takes 11 years to recover to its pre-crisis level. This is a Great Depression scenario.

The nightmare scenario is the growth rates since 1990. In this case, Japanese output does not return to pre-crisis level until the year 2064. Today’s babies will be worrying about retirement.

Wednesday, April 8, 2009

Bad News for the Fed’s Plans: Mortgage Modification is not Working

The Federal Reserve wants to lower long-term consumer interest rates. To this end, it has cut its policy rate to zero; engaged in any number of asset swaps with lending institutions; and, recently, begun outright purchases of GSE debt to put more direct pressure on mortgage interest rates. These policies, particularly purchases of GSE debt, seem to be working: mortgage interest rates have recently fallen to near record lows.

Of course, the Fed does not want to lower interest rates just for the sake of having lower rates. It wants to increase economic activity. And, in particular, it wants to help the housing market. The link is clear: lower mortgage interest rates raise housing affordability. Cheaper housing boosts demand. We observed a tepid increase in housing activity in February, whether a statistical fluke or a real turning point we shall see as the spring data comes in.

The average homeowner, through refinance, can take advantage of the Fed’s program and lock in for 30 years very low interest rates. A one percentage point reduction in interest rates will reduce the payments of an average homeowner about 10 percent per year. This cost reduction will push some households to buy a first home or to move up to a larger home. [Of course, this savings comes at an upfront cost of about 3 percent of their outstanding mortgage (personal survey of 10 mortgage issuers based on median house price and 80% down – prime borrowing only), raising the average debt level of the household sector.]

The Fed’s program, however, is likely to help the marginal homeowner only slightly. It seems to me that, at this moment in time, the most important marginal homeowner is the household that can no longer afford their house. When they stop making payments, their house eventually contributes to the inventory of unsold homes, lowering house prices and pushing more households to foreclosure.

It might be hoped that the lower interest rate would ease the payment burden of at risk households and reduce the number of foreclosures; however, new data from the OCC and the OTS indicate that interest rate reductions are not sufficient to substantially reduce the number of delinquent mortgages.

A reduction in mortgage interest of 1 full percentage point reduces the payment burden of households by about 10 percent. Data released in the OCC’s Mortgage Metrics report, reveal that more than 30 percent of mortgages modified such that the payment is reduced by less than 10% re-default within a few months of the modification.

Mortgage Metrics

The picture below is taken from the Mortgage Metrics report. The data are surprising: mortgage modification does not work at reducing mortgage delinquencies. After 9 months, more than 60 percent of modified mortgages are in default. And, we don’t know how many of the modified loans would have continued making payments in any event: not all delinquent loans go into foreclosure.
I think this number, by itself, goes a long ways toward understanding the low level of workouts from banks. Banks were likely aware of the low success rate and are therefore hesitant to undertake the effort needed to modify a loan. [I hope the State of Florida is paying attention. This applies directly to their new forced mediation programs.]
As striking as the relatively high level of defaults, the percentage of loans defaulting increased over the first three quarters of 2008. Over this time frame, the government, in the form of moral suasion, pushed lenders to increase the number of modifications. They responded; the number of modifications rose from 208 thousand in Q1 to 301 thousand in Q3. Apparently, however, to get this increase the lenders had to seek a broader pool of modification candidates and the deterioration of that pool is then predictable.

Perhaps one of the main problems with loan modification is that the majority of actions do not reduce the household’s mortgage payments. A whopping 32 percent of modifications leave the household with higher payments after the action. It’s clear to me why loan modification does not work if the payments don’t go down.
But what’s perhaps more surprising is that even loans with a reduction of 10 percent or more in payments still default at very high rates. After 9 months, 26 percent of loans in this category are delinquent. This number is going to worse over time because later vintage modifications do not perform well.
Here are my thoughts: I suspect, but do not know, that most of the loan modifications either reduce the interest rate faced by the household or extend the term of the mortgage. I suspect that very few of the modifications try to adjust the principal and that almost none of them ensure the household is actually above water following the modification. Households with negative equity positions are going to default in disproportionately large numbers.
By the way, for the first time in this cycle, the number of prime mortgage serious delinquencies is greater than that of any other category and the default rate on prime mortgages more than doubled between the first quarter and the fourth. I don’t think we are done with this housing cycle yet. Remember, the labor market did not take its nose dive until the fourth quarter. The serious delinquencies induced by this fall will not show up until at least the first quarter and more likely the second.

Tuesday, April 7, 2009

Have the Asian Economies Hit a Bottom? Yes and No and Yes and No …

There has been a lot of talk of bottoms and second derivatives lately. Most of the discussion is related to March PMI data. In almost every country, PMIs have bounced off of their record low levels in January. Of course, they remain quite close to their record low levels and are (with the exception of China) continuing to point to sharp contractions. Nonetheless, if PMI data are to be believed, we might see an easing of the pace of decline.

But, I am always suspicious of survey data and, at least to me, the PMIs are subject to some of the same whims as consumer confidence: when purchasing managers feel good about the world they report higher numbers. In general their feelings are a good indication of the weather, but sometime they are not. I continue to believe that PMIs are most useful for their timeliness rather than their unfailing accuracy. I prefer, when I have a choice, to wait for data.

Trade data is my favorite: it’s timely, well-measured, and theoretically sound. Take a look at the picture below. Korean exports seem to have leveled off. The dramatic fall in January seems to have been over stated and probably had more to do with the nadir in the auto sector than anything else. I take this as a sign that Korea’s external sector is not as weak as I inferred from the January fall.
Korean imports, however, tell a different story. Imports continue to fall unabated. The pace of the decline may have slowed slightly but even these slower falls are huge by historical standards. I read this as an indication that Korea is still falling and has not hit bottom. That probably means Asia is not at a bottom; a Korean decline would be felt across the region.

We also have March trade data for Taiwan. In this case, both imports and exports are up from their January lows. These numbers give hope of stabilization but at an extremely low level: imports remain 50 percent below their August peak.
In all countries, the pace of imports gives insight on the growth rate of domestic demand. In the Asian economies, not surprisingly, imports are tilted toward production rather than consumption, at least relative to the G7. Imports consist of intermediate goods, both capital and commodities.

Of course, the timing of the change in production and the timing of the fall in imports do not have to coincide, inventories play a large role. Korean IP shown below bounced off its December lows, rising in both January and February. However, the given the import data shown above this resurgence does not seem sustainable. I expect Korean IP to continue its downward trend. With luck the downward trend will look like an ordinary recession rather than a free falling collapse.

The fall in Taiwanese IP is already consistent with the overall fall in imports. In this case, there is a possibility that production has found a floor. I will be quite interested to see the March production numbers to be released later this month.
Even if exports, imports, and production have reached a floor, the recession is not over. Production and trade have, if anything, stabilized at a very low level. The domestic economies have made little progress adjusting to this new level of activity. In particular, the level of employment is not consistent with the level of production. Barring a true recovery in production and trade (a recovery in levels not growth rates), employment will have to fall, maybe by as much as 20 percent.

This fall in employment will lead to a second round of production and trade cuts. These falls will (likely) be smaller than the ones seen to date. Consumption accounts for a relatively small percent of GDP in these economies. Nonetheless, the production and trade falls to date are so large that these second round adjustments are likely to themselves look like an ordinary recession.
Takeaway: I believe we are beyond the freefall. The falls in Asian and European production were unprecedented: The Asian falls were faster and more extreme than the decline in production observed during the Great Depression in the United States.

Friday, April 3, 2009

Central Banks and Credit Loosening Policies: Misguided Actions

Today, both the Chairman and the Vice Chair of the FOMC gave speeches in which they clearly state that a credit crisis led to the downturn in real activity. The statements are not even qualified. From the Vice Chair: “A defining characteristic of the crisis has been a deepening adverse feedback loop in which financial strains have caused economic weakness, which has in turn led to credit losses and heightened financial strains, which then contribute to further economic weakness, and so on.”

I have shown previously that the key dates of the financial crisis were all pre-dated by turning points in the macro data. (Take a look at this post.) In my mind, then, the financial crisis itself was caused by a real deterioration in the real economy rather than the other way around. But central banks and central governments continue to implement policy as if there is nothing wrong with the world but a simple banking crisis. Indeed, Bernanke’s speech is full of measures by which the Fed’s programs have helped resolve the crisis.

Despite all of these actions and all of the rhetoric, no central bank in the world has been able to prove that a credit crisis even exists, let alone whether it started the downturn or resulted from the downturn. In October, Chari, Christiano, and Kehoe published this paper calling into question virtually every fact of the financial crisis. They acknowledge only that the asset-backed commercial paper market and the securitization market failed. A Boston Fed paper seeking to undo their results made very little progress. In their analysis, they show that some subclasses of loans were falling; but, subclasses are not the same as a credit crunch. Indeed, that only subclasses of assets are adversely impacted points much more to a real shock than a financial shock.

I could retrace all of the data in the above publications; but it seems to me that anyone who wants to claim that a financial crisis is of first order importance in the current downturn must first explain the following picture.
This picture shows the level of commercial credit from January 2001 through February 2009. Credit did not fall until November 2008, almost 18 months into the crisis and 11 months into the official recession. Even then, the fall is miniscule and likely has more to do with Federal Reserve lending programs than actual credit (think China). Indeed, credit continued to grow robustly until March 2008 when it turned flat.

Credit does not grow robustly during recessions. People are poor. People think they will be poor for a long time. They do not want to borrow.

For those of you still inclined to quibble, take a look at the same picture during the Great Depression. Credit should actually fall before we start calling it a credit crisis.

The auto sector has felt the impact of the recession more than perhaps any industry outside of the housing sector. New car sales slumped sharply beginning in the very tail end of the 2007. There were lots of speeches given about the lack of credit in the auto sector. The problem is that none of the usual indicators of credit tightening are apparent in the data.

Here is chart of auto loan interest rates at auto finance companies. Interest rates on new car loans did not rise in any meaningful way until the fall of 2008, 9 months after sales dropped. (There is a downward movement in rates at commercial banks.)
Other measures of tightness, such as months to maturity and loan-to-value ratios also remained flat.
Series by series, across many different classes of consumer and business credit, they all look essentially the same, except where evidence of falling prices can be seen. Spreads may have risen in this recession but interest rates did not rise.

Takeaway: There may be a credit crisis. But that crisis is not showing through to either business or consumer lending. Therefore, that crisis cannot explain the large drop in consumption, the large drop in auto sales, of the collapse of the housing market.

Policies designed to resolve the recession via credit intermediation are therefore misguided and destined for failure.

More on the Ins and Outs of Employment

I posted a while back on the ins and outs of employment (here). Today, I took a look the CPS Labor Force Status Flows, a publication of the BLS. This is basically the data Shimer used to compute his employment exit and entrance statistics. These data are timely and are released with the household survey.

The picture below shows the percent of unemployed agents finding a job each month. Think of the number on the vertical axis as giving the job finding probability for an unemployed agent. Using this data in January 1990, about 1/3 of unemployed persons found jobs each month. In the height of the ensuing recession, the odds dropped to about 0.22. During the expansion of the late 1990s, the rate reached 35 percent, only to fall once again in the 2001 recession. In March, this number hit a series low of 16 percent. Not only are people losing their jobs rapidly during this downturn, exiting unemployment is getting increasing difficult.
The next picture shows the number of months an agent can expect to be out of work conditional on unemployment. For most of the series, the average unemployment spell lasted between three and four months. The series reached a high in the 1991 recession and a low just before the end of the tech boom in 2000. The series reached a new record high in March: 6.2 months.

At the moment, one out of every four unemployed persons has been out of work at least six months. This statistic is going to get worse. There are currently 13 million unemployed persons. At current rates, that means 7.5 million of these households will still be unemployed next September. How high the unemployment rate rises depends on how many more people lose their jobs and how many of these workers leave the workforce.

The Relationship with Consumption

Imagine an unemployment spell of 6.2 months. How would you pay your mortgage? How would you feed your family?

High unemployment has a direct effect on consumption because the affected households are poorer than they were. They are poorer today because they don’t have jobs and they are poorer in the future because they are likely to reenter the workforce with a lower wage. The income effect is the obvious channel.

There is, however, a second and likely larger channel through which the jobs numbers effect consumption: there is a precautionary savings motive. Households face both an increased probability of job loss, and conditional on job loss, an extraordinary spell of unemployment. In response to this dual risk, employed households save.

The decrease in consumption is not what Keynes would call animal spirits. Rather, the decline in consumption is an optimal reaction to an increase in risk. The government should not work to change this response.

The March Jobs Report: No Big Surprises, But No Recovery In Sight

No surprises in the March employment report. Payroll employment fell 660,000 and revisions to previous months subtracted another 80,000 from the level of jobs. The same pattern we have seen over the past five months. Combine this data with the likely revisions from the benchmark survey (the direction is down because firms are failing faster than they are being created), and we are losing something in the ballpark of 1,000,000 jobs per month. Shocking, but no longer surprising.
Of minor interest, in percentage terms, we have finally surpassed the downturn in the early 1980s. Indeed, with the revisions, we surpassed the 1980 downturn in January, we just didn’t find out until today.

As I stare into my crystal ball, I don’t see any signs of a labor-market recovery. I could not find a subsector that showed any indication of rebounding. Further, unemployment insurance claims remain high and continuing claims are still rising. Survey data tell the same story. April does not seem to be on a better jobs track than March. And, the economy cannot expand while the labor market is contracting by 0.5 percent per month.