Wednesday, September 30, 2009

The True Impact of Fiscal Stimulus

There are currently a number of economists crowing over the absolute success of fiscal policy. I am still unconvinced that fiscal policy has any meaningful impact on economic output. I continue to believe that the programs to date have merely masked the underlying weakness in the economy.

“Masked the underlying weakness – Isn’t that exactly the definition of fiscal stimulus?” you ask.

The answer is not necessarily. To explain my views here, let’s turn to the case of China, the current poster child for successful fiscal and monetary stimulus and a crisis already occurring. In the second quarter, GDP increased at a rate someplace in the high teens. (China only issues real GDP on a four-quarter change basis; so, quarterly changes must be inferred.) The growth rate was phenomenal and is directly attributable to government intervention.

How did China achieve this remarkable growth? Easy. I believe that the government simply insisted that factories continue producing output—I am sure the insistence was accompanied by a promise of a fiscal transfer. Factories keep producing output, calamity averted, the stimulus is effective.

In fact, GDP gets an additional boost. The factories are producing output that nobody is buying – exports remained depressed and consumption is not picking up all of the slack. The output from the factories is accumulated as inventories, a positive contribution to GDP. But, since nobody is buying the price of the inventories also falls. Real GDP gets an arbitrarily large boost as the price deflator declines. [I am exaggerating for hyperbole. I know some of the goods were purchased but the increase in output likely owes to exactly the channels I am suggesting. If you want to see this at work in the United States, go back to the fourth quarter of 2006. Autos made a large positive contribution to GDP as the prices of new cars fell sharply and the real value of car inventories was pushed up.]

This is not real growth. It is not real growth in the present and it is a direct drag on growth in the future. Barring a sudden increase in OECD demand, China is in for some hard times.

This is the nature of fiscal stimulus.

The Fiscal Cost of Stimulus

Paul Krugman and I posted almost opposite comments on fiscal spending yesterday (his; mine). (Trust me they are not dueling – Nobel Prize winning famous guy, me.) Whereas I was concerned about the long-term cost of higher government spending, he was blasé. From his article, “I’m not proposing a fiscal-stimulus Laffer curve here: it’s probably not true that spending money actually improves the government’s long-run fiscal position (although that’s certainly within the range of possibilities.)” He almost believes that fiscal stimulus (at least under certain conditions) is almost self financing.

I don’t believe it for a minute. But, Krugman’s views are amazingly internally consistent and there is no questioning his mental acuity.

He believes in large multipliers. That is, every dollar increase in government spending increases output by something much larger than a dollar. (He has publically averred to multipliers of around 1.3 but this would not be even in the ball park of self-financing. Spend $1 billion, get $1.3 billion. Spend $1 billion raise $100,000 in extra tax revenue. With our tax system, self-financing begins with multipliers greater than 3.)

If Krugman is correct on the multipliers then he is correct on the cost of the fiscal stimulus. If not, …

Tuesday, September 29, 2009

The Federal Deficit and the Tax Burden: We can afford the debt not the spending.

Over the past year, the amount of publically held Treasuries, the current debt of the U.S. government, has ballooned. Debt held by the public has risen from 56 percent of GDP (already a near record) to over 63 percent of GDP in 2009. Importantly, this increase has occurred before most of the stimulus money from the American Recovery and Reinvestment Act has been spent; indeed, of that money, a meager $151 billion had been drawn as of the end of August. With this large debt come concerns over fiscal sustainability.
I have always taken it for granted that the debt was sustainable. After all, why would the markets lend to the government if the debt were not sustainable. Recent events (and Krugman’s excellent ariticle in the NY Times on the state of macroeconomics—Summers’ ketchup economics is brilliant.) have shaken my outright confidence in markets. So, I thought I should do my own sustainability calculations. I came to the surprising conclusion that the stock of government debt is not only sustainable but it is downright affordable.

To be conservative in my estimates, I use the entire amount of Treasuries outstanding. Publically held debt is a better measure of government debt but either will do for my purposes—funds held by the Fed and other public agencies don’t really count. In my mind, counting them is the same as counting the total sum of Treasuries that could be issued.

To check for sustainability, I take the Federal Government in hand and put it on a debt payment plan. If the Feds were a household and we were financial managers we might choose a 10 to 30 year plan depending on their age and circumstance. But, governments are special. I chose to put the gov’t on a 100-year payment plan at the end of which time debt must be zero. Of course with the payment plan the government is forbidden to acquire new debt—this feature will be the lemon in the pudding.

If real interest rates stay at their current low levels of around 1 percent (unfathomable), the payment plan costs each of the 137 million workers in the United States $1,361 per year, a paltry 1.7 percent of personal income. Even if real interest rates rise as high as 10 percent (equally unfathomable (or maybe I can picture this one)), the debt burden per year remains an affordable $8,248 per year per worker or about 10 percent of personal income. The latter number is large but feasible.

The real problem is that the government must also raise sufficient funds to keep from borrowing more. The Bush administration ran the largest deficits of any government to that time. If we add the Bush deficits to the total, yearly payments need to rise. Under the low interest rate case, this raises the payments to a little more than $3,000 per year per worker. At 10 percent, we are quickly approaching $10,000 per worker per year.

The Obama deficits are projected to be much, much larger. If we have to raise revenue to offset the average Obama deficit over his first term as calculated by the CBO, even at 1 percent interest, the debt payment is $6,326 per worker. At 10 percent interest, payments per worker increases to almost 17 percent of personal income.

This exercise reveals the national debt to be affordable in the ordinary sense of the word. The United States could pay off its debt in a mere 100 years with only a modest strain on workers. However, if government spending continues to rise (or stays at the current levels), the strain on workers is likely to extreme. Additional taxes in excess of 10 percent of total personal income are likely to have a large negative impact on the workforce.

I might be in favor of higher government spending or I might be against it. As always, the decision is one of public policy and not of economics. But, good economics should inform the decisions and we should make an informed choice on how much to spend based on an honest national dialogue.

The Strength of the Recovery: Demographics could kill it all.

Mussa gave an interesting talk on the recovery the other day at the Peterson Institute. You can find his paper on their website. He makes an excellent case for a V-shaped recovery. With his forecast of U.S. growth at 4.5 percent for 2010, he makes the claim that his estimate is actually conservative. That, in reality, we could, and likely should, look for an even steeper recovery. In particular, he points to the fact that in the average post-war recession average growth has been closer to 10 percent than 5.

In my last two posts, I gave a cycle-by-cycle breakdown of the key parts of GDP. The rate of recovery in individual components was not particularly strong. Adding up the various measures, the strength of various recoveries was accomplished because most of the rising demand coming out of recessions was satisfied from domestic sources. In this recession, any rising demand is likely to be satisfied to a large extent by production outside of the United States. Just as the downturn in the United States was cushioned by a fall in imports the rebound will be pushed down by imports.

But, in my mind, this is all accounting. The potential U.S. recovery is exposed to a much more serious threat: the U.S. consumer. In percentage terms, the rise in the savings rate in this recession far exceeds the rise at any other time in post-war history. Indeed, we have to go back to the 1930s to see a similar spike. Prior to this recession but still post-war, the largest increase in the savings rate was a little more than 30 percent. In this episode, the increase has been a little more than 80 percent. Granted, the calculation is from a very low base near 2 percent of income.
However, never in post-war history has consumption accounted for such a large portion of aggregate demand. I believe the large change in savings reflects a combination of cyclical forces and the beginning of a long-term consumer retrenchment. Americans must consume a smaller portion of their income.

With 1 in 6 Americans underemployed and with wage growth falling, consumption is likely to be a substantial drag on any recovery.

Yet, I believe Mussa would retort (he makes the argument in his paper) that consumption has fallen more than in any other recovery and we are due a bounce back. Indeed, so Mussa would say, following the recession of 1980 consumer spending bounced back with a vengeance.

Take a look at the following picture. Despite the financial crisis aspect of this recession, consumer credit was unimpaired relative to previous recessions. Real consumer credit has only nudged down. Compare this to the almost 15 percent fall in 1982. In 1982, real interest rates rose to record levels. Consumers stopped spending. Following the recession, the expansion of credit played a huge role in the rapid expansion of spending. Consumer credit grew at almost three times the pace of the average recovery. With almost no impairment in this recession, the room for a bounce back is smaller.
The resumption of consumption also faces a very unfavorable demographic situation. The chart below shows prime-age workers (30 to 65) as a percent of the total population. In 1982, this group of agents increased almost 8 percent over the first five years of the 1980s. In the current recession, this percentage falls remarkably over the forecast. Indeed, it falls faster than at any other time since at least before 1901. Prime-age workers are the most productive people in the economy. They also spend the most: they buy big houses and fancy cars. A fall in this group has wide-spread implications for the economy. There is only one other country I know of with this demographic pattern: 1990s Japan. That is not a good omen for V-shaped recoveries.

Just to finish the thought, here is the path of prime-age workers as a percent of the total population from 1970 through 2014. It seems that one source of our amazing growth rates from 1980 through 2007 was the rise of this group of workers. Notice, that during the productivity slowdown in the 1970s, this group of workers was flat or falling. During the 2000 recession and long jobless recovery that followed, this group of workers was also quite stagnant. For the next ten years or so, these workers will retire out of the labor force.

Most model-based forecasts assume that the fall in population itself will subtract only a few tenths from growth. With labor shares and capital constant, this is true. However, the fall in prime-age workers is also likely to be associated with a long-term downward trend in investment. We have already almost a decade of subpar investment growth. More importantly, the loss of human capital will be severe. It takes years to produce prime-age workers. These workers embody skills that have been acquired in a life-time of work. There loss is likely to drag on labor productivity for the foreseeable future.

Wednesday, September 9, 2009

Part I: The Great Recession of 2009: Where are we going?

Most analysts now believe the recession is either already over or at least kicking its last gasp. They might be right. But, I remember having this conversation 6 months ago; back in March when the PMIs first turned up. The end of the recession wasn’t in the data then and it is not obvious it is in the data now, although the odds have shifted towards recovery. The difference between now and then is time. We now, I believe, have sufficient data on hand to judge both the severity of the downturn and to compare this downturn to previous recessions. The comparison may shed light on both the likelihood of a near-term recovery and the potential shape of that recovery.

We examine, in turn, key elements of U.S. demand, running from investment to imports to exports to consumption. For each series, we plot data 12 quarters on either side of the end point of post-war U.S. recessions. The zero date is the last quarter of the NBER recession. The solid line is the average behavior of all post-war recessions excluding 2009. The dotted line shows the behavior of investment in the worst post-war recession defined as the recession with the largest peak-to-trough decline in the particular series being shown. All points on the dotted line belong to the same recession. The dashed line shows data for the current recession. We plot the data as if 2009Q2 is the end of the recession. All lines are indexed to 100 at last date of the recession. The indexing scales the graphs to show cumulative percent changes from the zero date. For example, a value of 120 in period -4 indicates a 16 percent fall in the series in the last year of the recession. While a value of 120 in period 4 indicates a 20 percent rise from the end date o the recession. All series are real.

We begin with machinery and equipment investment. I have long advocated that this recession is a manufacturing recession (see this post); therefore, in my view, the recession is unlikely to end without a recovery in this sector. Investment is one of the most volatile components of GDP, typically falling almost 10 percent over the last year of the recession. Post-recession, investment grows robustly recovering in level terms one year after the end of the recession.

The worst investment recession was 1958. Investment fell a little more than 18 percent with the decline occurring in a single year. Investment recovered in level terms just in time for the 1961 recession. Despite the faster-than-average growth upon exit, investment did not recover in level terms to its pre-recession peak until almost 2 years after the end of the recession.

The current recession has solidly replaced 1958 as the worst investment recession. Investment has, to date, fallen more than 20 percent from its peak 6 quarters ago. Investment does not currently show any signs of life; however, investment turns on a dime and typically turns up only once the recession is over. If history is a reliable guide, investment is likely to surge robustly in coming quarters. I believe a typical recovery is more likely than the sluggish recovery experienced after the 2001 recession. It seems a global manufacturing reshuffling is once again underfoot. Even though this reshuffling is likely to destroy capital (the shutting down of American car plants for example), it may also breed investment.
Like investment, trade has plummeted more in this recession that at any time in U.S. post-war history. Unlike investment, however, trade shows a clear, leading, end-of-recession pattern. The decline in imports tends to slow in the quarter or two before the actual end of the recession. Following the recession, imports tend to grow very rapidly, in part reflecting the strong trend in trade over the last 40 years. The worst import recession was in 1975.

In the current recession, imports have also surpassed all previous records, falling over 20 percent. As of the second quarter, imports still declined but the pace of decline was somewhat smaller than in previous quarters. This may signal the end of the recession or it may just be one of the inevitable bobbles in the data. In former case, the recession would end following the third quarter.
Of course given the above focus on investment, capital imports may be even more important than overall imports. The largest capital import recession occurred in the 2001 recession as the United States shed one-third of all manufacturing jobs and the tech expansion dissipated. The current recession has exceeded this high mark. Capital imports lead the cycle more strongly than overall imports. In past recessions, capital imports stopped falling entirely, on average, one to two quarters before the end date. Currently, capital imports have not yet truly flattened.

However, while NIPA capital imports did not stabilize as of the second quarter, capital imports have since leveled out. The following figure shows real capital imports to Japan, Canada, the Euro Area, and the United States. Only in Japan are capital imports actually rising. But, in all three other regions, capital imports have stopped falling. This may be a sign of the recession ending. If historical patterns hold true, this leveling would indicate a fourth quarter recovery rather than a third.

This post continues in Part II

Part II: The Great Recession of 2009: Where are we going?

This post is part II of The Great Recession of 2009

Exports from the United States tell the tale of this recession. Exports fell in line with the worst previous U.S. exports recession (1958), but no worse. It’s a global recession to be sure but foreign demand held up better than in the United States, particularly at the beginning of the recession. As with imports, exports tend to flatten out before the end of the recession. It seems that not only Asian economies have benefited from export led recoveries. And, in the current recession exports are showing signs of stability.

Durable consumption, the most volatile component of consumption, has also fallen sharply in the current recession, although by a smaller amount than the cumulative decline associated with the 1958 recession. Durable consumption leads the end of the recession more than any other component of demand, turning up more than a full quarter before the end of the recession. In the current recession, durable consumption actually increased in the first quarter. However, this increase was more than undone in the second quarter. I have not looked at the detail but I suspect the bobble is attributable, in part, to the bankruptcy difficulties of the auto makers early in the quarter. These temporary concerns may have pushed demand for cars across quarters. Too, by the second quarter it was becoming increasingly obvious that the U.S. government would at some point subsidize car sales. Savvy consumers then may have postponed purchases to benefit from the subsidy.

In any case, there has never been a post-war recession that has not been followed by a steep surge in durable purchases. The economics behind this surge is clear. Durables are durable. The service flow from refrigerators and cars diminishes only slowly. Purchases of these items are easy to move across time. So, once the stress of the recession ebbs, particularly labor market stress see next, consumers begin to restock their durables. Interestingly, however, the speed of the recovery seems independent of the extent of the decline, implying that some consumption is forgone permanently.
Residential investment is, in reality, simply the most important component of durable consumption; however, in National Income Accounting, it is treated as investment. As I have noted before, residential investment is a leading indicator. It peaks before the beginning of the recession and troughs before the end. And, residential investment was an early hallmark of this recession; but, with the constant stream of bad news last fall, it is easy to lose track of just how far the housing market has fallen. Just over the last three years (we are missing two quarters of decline), the fall in residential investment has bested the worst recession by more than 50 percentage points, falling by well more than 50 percent. The fall in residential investment continued through the second quarter despite several massive intervention attempts, including direct intervention in mortgage markets and a $7,000 dollar subsidy to first time buyers.
I have trouble picturing a recovery that does not include a recovery in residential investment. The ongoing sharp fall is bad news on that front. However, data since the second quarter has shown some signs of hope and housing starts are turning up. If the turn is real and not temporary induced by temporary government incentives, the recession is likely to end after the third quarter.

I have discussed the importance of initial claims for identifying turning points in the data (see this post). Initial claims are a timely indicator of the firing rate in the economy. As soon as this rate stops rising, the economy begins to recover. This is both causal, employed households have less fear of losing their jobs and begin to spend, and coincidental, firms see better conditions and stop firing workers. As claims begin to drop sharply, the recession is over. This is an extra important indicator because claims data is a timely, if noisy, indicator. So, even though the change in claims is coincident with the end of the recession it is available sooner than other macro data.
Looking at the pattern of weekly claims since the end of the second quarter, the sharp downward trend is hard to spot. Claims seem stuck between 550 and 600 thousand per week, a rate I consider inconsistent with full-on recovery. Despite the horrific jobs numbers we have experienced in this recession, the rise in claims is not as big as in the worst recession (1975) nor is it as bad as the second worst, 1958.

With the debate and eventual passage of an almost $1 trillion dollar stimulus earlier this year came a barnstorm of economists either touting the benefits of fiscal stimulus or warning against its potential inefficacy. Romer was in the former camp while I fell in the latter. It turns out, however, that we were all arguing with air. Take a look at the total government outlays chart below. Total government outlays includes all government spending and all transfer programs; quite simply, it is total government spending. The level of government spending is almost unchanged during the recession (periods before the zero) and only jumps slightly above trend in the periods after. Whatever the multiplier, it is not going to influence the trajectory of the economy in any substantive (read measurable) manner with these small changes. We have no idea what fiscal spending does.

The spending in the current recession will give us our one experiment. The rise in government outlays, from an already elevated point, should if the multipliers are high lead to a faster-than-normal recovery. In any case, the rise in government spending is somewhat stunning. But one data point is still just an anecdote.


Among the key things I learned from these charts is that the depth of the decline has very little to do with the pattern of recovery and that it takes about two years to recover in level terms from each downturn. This makes sense to me and was also consistent with the recoveries experienced by Argentina after its 2001 collapse, the Asian economies following the Asian Financial Crisis, and Mexico following the Peso Crisis. In other words, bad things happen and then they get better. The only post-war economy that didn’t fully recover from a crash is Japan. Japan has more or less stagnated for the past 20 years.
All told, I would judge the indicators point to a recovery that will not happen before the fourth quarter. Only the most nascent signs of recovery are apparent (ignoring survey data). But, nascent signs are better than none.