Some Immediate Benefits of the Corporate Tax Cut

Copyright, TheHill.com

By cutting the statutory corporate tax rate and by permitting investment expenses to be immediately deducted from corporate income, the new tax law encourages corporations to enhance their workers' productivity by investing in structures, equipment and software.

The additional investment will accumulate over several years, which means that the full effect on productivity and wages will not be felt for several years.

However, Economics Nobel Laureate Paul Krugman further asserts that there are essentially no benefits for workers in 2018, despite the fact that a number of corporations have announced bonuses for workers while saying that the bonuses derive from the new tax law.

The simplest model of investment and worker productivity agrees that aggregate wage increases would not be discernible in the first year following a permanent and unanticipated capital income tax cut because of the time that it takes for investment to be planned, executed and translated in to greater worker productivity.

But Krugman, Obama economic adviser Larry Summers, and I, among others, agree that our economy and this tax cut have some meaningful differences from the simple model. One of those differences is that President Trump's signature last week was not an entire surprise.

The U.S. had been behind most of the world in cutting its corporate rate, and it was largely a matter of time until the U.S. did the same, especially in 2016 when Republicans won the White House and both houses of Congress.

Throughout 2017, businesses were making plans understanding the very real possibility that federal corporate tax rates would be lower, and the execution of those plans are already adding a bit to worker productivity.

The simple model also ignores that a lot of businesses are not organized or taxed as U.S.-based C-corporations, which are the types of corporations that have been subject to high statutory rates by worldwide standards.

This has resulted in too little business activity occurring with the legal and organizational advantages of the C-corporation, and productivity has suffered as a result of companies' keeping activities away from the high C-corp rates.

Reallocating activity to U.S.-based C-corporations can happen more quickly than the building of new structures or manufacturing new equipment does. This means that part of the productivity effect can occur quickly too.

The simple model also treats labor costs as variable, which is a reasonable treatment for multi-year time frames. But over a period of a few weeks or months, which is the time frame discussed by Professor Krugman, much of the labor costs are slow to adjust, due primarily to the fact that it takes time to attract and sign good employees.

With businesses anticipating productivity growth over the next several years, it makes sense for them to take some immediate steps to solidify their workforce. (It's odd that Krugman missed this effect: he frequently writes about the "JOLTS" labor data, the entire point of which is that labor costs adjust slowly from the perspective of weekly or monthly data).

I agree with Professor Krugman that actions speak louder than words in matters of economics. Although they sometimes agree, often businesses say one thing and do another. This is especially true when the federal government uses its regulatory might to encourage businesses to say the "right thing," as it did when it rolled out the Affordable Care Act a few years ago.

But it is inaccurate to claim that workers must wait before seeing any benefits of the corporate tax cut.

At 21% or 20%, new corporate tax rate will boost US economy

Copyright, TheHill.com

Since the 1990s, U.S. corporations have been subject to one of the highest statutory tax rates in the world. The high rate has caused them to rearrange their affairs to avoid investing, especially in lines of business subject to the full rate, and thereby reducing productivity and workers’ wages.

But now Republicans in Congress appear to have agreed on reducing the rate by 14 points, to 21 percent, plus the applicable state rate, bringing the total into line with the statutory rates elsewhere in the industrialized world.

President Trump had originally insisted on a federal corporate rate of no more than 20 percent, but Congress appears to have chosen 21 in order enhance the bill’s revenue outlook. It is worth assessing how much revenue was gained, and worker wages lost, by this deviation from the president’s plan.

Although I expect that the federal government will be getting less corporate tax revenue than it would without any tax reform, it is possible that the change from 20 to 21 percent by itself has little or no effect on revenue.

That one extra point may prevent the U.S. from undercutting a number of countries such as Spain, the Netherlands, Austria and Chile and thereby reduce the amount of business activity that relocates here from those nations.

So, as compared to the president’s plan, the IRS will be collecting an extra point on corporate income, but there will be less corporate income than there would have been.

The tax reform’s expensing provisions — generous deductions for new investment projects — also encourage business investment apart from the rate cut, and it has been argued that expensing provisions by themselves create a lot of the economic growth generated by the reform.

Thus, even if adding a point does little to enhance revenue, it may also do little to limit the wage gains that come with reforming corporate taxes.

It is important to remember that much business is not corporate business, but rather organized as partnerships, S-corporations, etc. Ideally these organizational decisions would be made for real business, rather than tax reasons.

It remains to be seen how the new corporate rate meshes with the reform of taxation of non-corporate businesses, because it depends on how the IRS and tax accountants interpret the new rules.

But I expect that there are some non-corporate businesses whose rates would have been a couple of points higher than a 20-percent corporate rate (plus the relevant personal income and state corporate taxation), so that adding a point to the corporate rate mitigates some of the unintended consequences on that margin.

Finally, future Congresses may be willing to further change the rates, especially to the degree that the changes are small. Recall that the last big corporate rate cut in 1986 was followed by a one-point change during the Clinton administration.

For all of these reasons, the consequences of a one-point deviation from the president’s plan are small compared to the overall economic benefits from a long-overdue reform of the corporate tax.

Low effective rate not an argument against corporate tax cut

Copyright, TheHill.com

President Trump says that U.S. corporations face the highest tax rates in the world, whereas opponents of his tax reform say that “actual” corporate tax rates are in line with other countries. What gives?

The president is referring to the rate that applies to taxable corporate income, known as the “statutory rate.” The federal statutory rate is 35 percent, and the combined federal-state corporate rate is typically about 39 percent. This rate is greater than anywhere in the industrialized world.

But the statutory rate does not apply to all of the income-generating activities of corporations, because some of those activities create deductions that can be subtracted from business income for corporate-tax purposes.

Debt-financed activities are a good example, because the income they generate goes corporate-tax free to the extent that it can be distributed to investors as interest income.

Intellectual property investments are another example, because they are not obviously attached to a physical location, thereby helping accountants assign their returns to Ireland and other low-tax jurisdictions.

As a result, corporations are paying less than 39 percent of their income to state and local treasuries. The Government Accountability Office estimates 17 percent, which it calls the “effective rate.”

It might seem that the 22-percentage-point difference results in a free lunch for corporations at the government’s expense. But the opponents of corporate tax reform are mistaken to ignore the fact that the corporate tax has corporations paying a lot more than the checks they write to government treasuries.

The Internal Revenue Service (under the Obama administration) estimated that corporations and partnerships pay over $100 billion annually in complying with business-tax laws, including their costs of recording, keeping and hiring paid tax professionals. Compliance costs are not checks written to the government, but are real costs nonetheless.

In fact, the high compliance costs are a symptom of the low effective rate because business-tax deductions are a complicated enterprise. Corporations are paying for some of their 22-point savings in terms of the extra compliance costs that come with complicated tax strategies.

More important, our economy is less productive because taxes have induced investors to pursue tax-favored activities beyond what value creation would dictate. The most vivid example is the housing sector, where returns have been depressed by a factor of two or three because those returns are essentially tax free.

In other words, by having so many deductions, the corporate tax involves a substantial hidden tax on businesses beyond what they pay the government, with the extra payment in terms of lost income.

The chart below illustrates by splitting the economy into two kinds of activities: those that pay full tax and those that are tax favored.

If nobody adjusted their investment plans, the tax-favored activities would be a great deal — they would earn the amount up to the dashed line and owe no tax. But there is no free lunch. The tax favors induce investors to engage more in those activities.

Their movement depresses the income accruing there — otherwise nobody would be willing to do the activities subject to full tax. The chart illustrates this by showing how the income ultimately earned has been depressed enough so that the net-of-tax earnings is the same for both types of activities.

The low effective corporate tax rate is therefore not an argument against President Trump’s call for tax reform. That low rate is further evidence of the economic damage done by the tax, as businesses pay to comply and pay by accepting comparatively low-return investments.

Because the effective rate only counts costs in the form of payments to government, a low effective rate is telling us that cutting the corporate tax will benefit economic performance far more than it will cost government treasuries.

Public Policy Suffers when Price Theory is Ignored

Jason Furman and Larry Summers weighed in this week about the quantitative amount that labor can benefit from a capital-income tax cut.

It soon became clear that they had failed to carefully use price theory in coming to their conclusions.

Furman and now Krugman (update: and now Summers) are admitting that simple supply and demand vividly contradicts what they said/say about taxes, but assert that the world is more complicated. I agree.

But they are dead wrong to further assert, without evidence (and I suspect without thinking), that adding complications will overturn the simple supply and demand conclusion or at least weaken the contradiction contained in their original proclamations.

(So far, Summers has wisely refrained from trying to defend his mistake.  Update: he replied 2 hours after I posted this -- see the update at the very end of this post.)

Supply and demand can do more than I have already shown.  Supply and demand also:

1. can give quantitative answers.  Greg Mankiw writes "I must confess that I am amazed at how simply this [quantitative formula] turns out. In particular, I do not have much intuition for why, for example, the answer does not depend on the production function."  Supply and demand can answer his question, without any algebra.
2. can deal with complexities, such as "imperfect competition."  The simple supply and demand model assumes perfect competition, but that assumption can be and has been modified.  Guess what?!  Making that modification shows that even the simply supply and demand model, let alone the proclamations of Furman-Summers-Krugman, understates the wage impact of capital-income taxation.
   [Hints: what new rectangles appear when the factor-renter is selling his product for more than marginal cost? What determines the equilibrium size of those rectangles?  Why should we use the corporate tax to rather than the DOJ to fight monopoly?]
3. can deal with complexities, such as debt finance.  Having uneven taxation of different types of capital tends to reduce the denominator of the Furman ratio more than it reduces the numerator.  i.e., Furman still has it even more backwards than I thought (update: Summers too).
4. explains why horizontal capital supply is not an "extreme" case.  Gary Becker and I explained why capital supply probably slopes down somewhat in the long run (thanks Kevin M. Murphy for reminding me about this -- not to mention for teaching so many of us about price theory!)
5. shows you how to process the economic data.  Furman and Krugman make evidence-free proclamations about the elasticity of capital supply.  Supply and demand shows what economic data is needed to measure that elasticity (spoiler: it's not complicated, and shows a very high elasticity).

I will post on these individually next week (week of Oct 30). In the meantime, you may be interested in a previous instance when, with important public-policy issues at stake, Krugman failed to appreciate what supply and demand really says.

Update: 2 hours later Summers posted a reply that reiterates the "it's complicated," "monopoly profits," and "debt finance" excuses for ignoring what the simple model says.  See my points 2 and 3 above.

He also hopes that you forget that he referred to CEA's result -- which is generally in agreement with the simple supply and demand model -- as "unprecedented in analyses of tax incidence."

Regarding Summers' other replies, see here.

Furman and Summers revoke Summers' academic work on investment

Former CEA chair Jason Furman writes “The economic debate about the %age of the corporate tax paid by labor ranges from 0% to 100%. The new CEA study puts it at 250%.”

Larry Summers reiterates Furman’s argument, calling the CEA and its estimate “dishonest, incompetent and absurd.”

Furman’s first sentence has the economics of investment completely backwards.

I will point to academic papers in a minute, but they can be understood with capital supply and capital demand, as shown below.

The red area (R) is the revenue from a capital income tax.  The red and green areas (R + G) are the losses from that tax suffered by owners of the factors of production, combined for capital and all other factors.  The revenue is a LOWER BOUND on the factor owners' loss.

In the long run, all of the factor owners' loss from a capital income tax is a loss to labor (the area below the horizontal dashed line is negligible; see A below).  Therefore, in the long run, capital-income tax revenue is a LOWER BOUND on labor’s loss.

Furman and Summers have it backwards.  They don't seem to understand that the wage gains from a cut come not only from the Treasury but also the economic waste created by the corporate tax.

(A) Why would labor bear all of the burden in the long run?  Well, ask Larry Summers back when he used to be an academic studying these matters.  His 1981 Brookings paper, which even today is an article commonly used by me and others to teach this in graduate school, says so on page 81 equation (7).  The left-hand-side of that equation is a perfectly elastic long-run supply of capital: it says that the supply curve in my picture is, in the long run, properly drawn as horizontal.  See also Lucas (1990, p. 303, equation 4.3).

(B) OK, the long-run Furman ratio must be greater than 100%, but how big is it?  If we (i) begin, as is today’s reality, with a high tax rate, and (ii) conservatively assume that the only channel for benefits is a higher capital stock (more on that below), then 250% is about right for cuts to somewhat lower rates.

Using a Cobb-Douglas aggregate production function with labor share 0.7, and a 50% capital-income tax rate (combining corporate, property, and the capital components of the personal income tax), I get a Furman ratio of 350%.  With a 40% tax rate instead, the Furman ratio is 233% (algebra here; these refer to modest tax-rate reductions -- not going all of the way to zero).

If the current CEA said 250%, then it got Furman's ratio much closer than Furman did, who puts it less than 100%.

Note that Summers now calls the 250% "unprecedented in analyses of tax incidence," yet I am getting it from his own paper about how the corporate-income tax works (see esp. p. 95)!

(C) Are there labor benefits not shown in the picture?  Again, let’s go to the academic incarnation of Larry Summers.  He once made contributions to supply-side economics!  In his chapter in “The Supply-Side Effects of Economic Policy,” Summers wrote that labor likely benefits from corporate income tax cuts even WITHOUT ANY increase in the aggregate capital stock because that capital would be “better allocated to the corporate sector.”

Update on (C): Greg Mankiw points out still more labor benefits not shown in the picture.  His source -- you guessed it! -- Larry Summers.

To summarize, anyone using Larry Summers’ academic work for policy analysis, is, according to Larry Summers, “dishonest, incompetent and absurd.”

(update: Summers' reply now revokes academic work more generally.  He also wants you to forget that he said CEA's Furman-ratio result to be "unprecedented in analyses of tax incidence."

Moreover, he digs his hole deeper with his critiques of the simple model.  I.e., President Trump should be thanking Summers for unwittingly strengthening the case for corporate tax reform.

See my comments on Summers 1981 here.)

Who Cares About the Fed?

Copyright, The New York Times Company

Short-term interest rates have an obvious effect on the housing market, but not the rest of the economy.

Federal Reserve policy affects short-term interest rates, bank regulation and eventually inflation. I will write about inflation next week, and my fellow Economix blogger Simon Johnson has written much about bank regulation, so today I focus on short-term interest rates.

The Federal Reserve, especially its New York branch, is actively engaged in buying and selling Treasury securities, and it lends money to banks on an overnight basis. As a result, it is widely thought that the Federal Reserve is an important determinant of the rate of interest paid on short-term Treasury securities.

By raising the supply of Treasury securities and reducing overnight lending, so-called “tight” monetary policy raises short-term interest rates. High short-term interest rates are said to discourage borrowing, and thereby curtail private sector investment projects. The idea is that private sector projects are undertaken only when their expected return exceeds the cost of borrowing.

In theory, high short-term interest rates result in relatively few capital projects, with high expected returns, and low short-term rates result in more capital projects, including those with lower expected returns.

But the effect of high short-term interest rates on Main Street’s economy has been exaggerated. Although it is commonly assumed that today’s rock-bottom rates should help strengthen a business recovery, it appears that business conditions actually have little to do with short-term money markets.

Many important private sector investment projects are relatively long term — it most likely takes a year or more for a project to be completed and deliver a positive cash flow to investors. As a result, many capital projects are financed through long-term borrowing, with equity financing, or out of corporate retained earnings, rather than borrowing in the short-term market where the Fed’s fingerprints are so obvious.

In theory, long-term interest rates could rise as the Fed tightens the short-term money market, because some savers would be on the margin of saving in either the short- or long-term markets. Equity capital markets and retained earnings could, in theory, also be subject to similar indirect effects.

Thus, the effects of Federal Reserve interest-rate policy on investment are indirect, and it is an empirical question as to whether the expected effects — tight money discourages investment projects — are significantly reflected in preventing capital projects with low expected returns.

Luke Threinen and I have measured national average profitability of capital projects from the national accounts by dividing total interest and profits in the economy during a year by the total capital stock in place at the beginning of the year. In doing so, we have distinguished residential capital (i.e., houses) from business capital.

Capital produces value over a number of years. In the case of housing capital, the value is in the form of shelter and the convenience of a home. For any piece of capital, profitability (capital’s marginal product, as economists call it) can be calculated as the dollar value it creates during a year — after subtracting depreciation, costs of labor, maintenance and intermediate goods — per dollar invested.

Owners of capital prefer their capital to be more profitable, rather than less. It’s the profitability of capital (after taxes and subsidies; more on those below) that makes an owner willing to purchase capital in the first place.

Chart 1 compares the profitability of housing capital to the inflation-adjusted return on one-year Treasury bills (for comparability with T-bills, housing profitability is adjusted for property taxes). Consistent with the view that tight monetary policy both raises Treasury bill rates and reduces housing investment, the two series are positively correlated. The home-mortgage market appears closely linked, so high Treasury bill rates cause banks to charge more for home mortgage loans, which discourages homeowners and landlords from building homes unless the demand for homes is sufficient (i.e., landlords can earn enough rent from their tenants to cover a high mortgage rate).

Among other factors, easy credit from the Federal Reserve in the early and mid-2000s made it easy to buy and build homes, and as the inventory of homes grew the amount of rent that each home could earn (many homes went vacant, for example) fell, which shows up in Chart 1 as especially low values for the red series. In this way, the housing cycle of the 2000s confirms the usual story about how monetary policy can affect housing investment.

The usual story about Federal Reserve policy and business investment says that a similar process works on the business sector: High Treasury bill rates cause banks to charge more for business loans, which discourages business from investing unless demand for their product is sufficient (i.e., businesses can earn enough profit from their operations to cover a high loan rate).

Our findings for the business sector are quite different from the usual story. Chart 2 compares the profitability of business capital to the inflation-adjusted return on Treasury bills, and the correlation is negative.

One way that easy monetary policy could hurt business investment is by encouraging home-construction activity, and home construction takes resources away from business construction.

The evidence in Charts 1 and 2 suggests that the housing market can be stimulated by easy monetary policy, at least in the short run. But the link between monetary policy and the business sector is much weaker, and our data are consistent with the view that, holding constant the rate of inflation and the amount of banking regulation, monetary policy does not have a discernible effect on the cost of business capital.

Housing-Sector Profitability Returns to Previous Levels

Copyright, The New York Times Company

The housing boom further increased our economy’s bias against business capital, but by 2010 housing profitability was back to normal.

Last week I examined a measure of the profitability of housing capital: the value added by houses in the form of shelter and convenience for their occupants during a year, expressed as a fraction of the total value of homes. Before the housing boom, housing added much less value per dollar than business capital did, largely because business taxes restrict the supply of business capital.

The less business capital there is, the higher the rate of return that remaining business capital earns, because each unit of capital serves more customers, as I noted last week. A low profit rate for housing is a symptom of its abundance.

The profitability gap between nonresidential and residential capital shows that our economy was overinvested in housing, long before the housing boom of the early 2000s. Business taxes cause an underinvestment in business capital, and business capital has been so profitable to the economy because it is more scarce.

Value added to the economy would have been greater if some housing investment had been invested in business instead; each $10 billion of housing investment redirected to business investment could have added almost a billion dollars to G.D.P.

The chart below displays the same marginal product measure of residential capital profitability (red line, from the first quarter of 2000 to the third quarter of last year (Luke Threinen, a University of Chicago student, helped with these calculations).

The chart also displays the marginal product of business capital (blue line). Because of the extraordinary “depreciation” from Hurricane Katrina, values for 2005 have been interpolated.

The marginal product of residential capital is measured on the right axis, while the marginal product of nonresidential capital is measured on the left axis. By looking at the 2000-1 values – 6.6 percent for residential and 15.2 percent for nonresidential – we see the result, noted last week, that nonresidential capital was twice as profitable before taxes than residential capital.

After the 2001 recession, and during the housing bubble, the marginal product of residential capital fell sharply. (This finding is related to previous findings that rent-to-price ratios were low during the mid-2000s, because housing profitability is essentially the rent-to-price ratio minus the expense-to-price ratio.) During this period, housing construction was booming.

Housing was not particularly profitable during the housing construction boom – the market appears to have built so many new homes during those years because of an expectation that houses would be profitable in the future, and in part because home buyers were encouraged by easy credit.

As a result, an ever-growing housing inventory was competing for much the same demand for shelter, which kept rents low and made vacancies high.

While housing sector profitability became so low, business sector profitability was quite high. At the peak of the housing boom, the marginal product of nonresidential capital was more than triple the marginal product of residential capital.

By widening the profitability discrepancy between the two sectors, the housing boom was more damaging than it would have been if the boom hadn’t begun with an abundance of housing.

When the housing bubble burst in 2006-7, and housing construction nearly came to a halt, the marginal product of housing began to rise toward previous levels as the population grew into the extra housing, and some existing housing deteriorated. By the fourth quarter of 2009, the marginal product of residential capital hit 6.2 percent and has remained about there since then.

The good news is that the 6.2 percent housing sector profitability of the past is actually a bit higher than the historical average from 1950 to 2000, which means that housing is adding value at a rate similar to the pre-bubble years.

The strongest economic pressures preventing further increases in our housing inventory have been gone for a year now, which is one reason I do not expect the housing market to get any worse.

Business capital profitability has also returned to previous levels, but the bad news is that housing capital remains much less profitable than business capital. In this sense, housing is still too abundant, an economic waste that shows no signs of repairing itself.

Housing Profitability -- Video Version

Housing Sector Profitability

Copyright, The New York Times Company

Long before the housing boom, overbuilding effects of government policy were evident.

A house is a piece of capital, meaning that it produces value over a number of years. In the case of housing capital, the value is in the form of shelter and the convenience of a home.

As with any piece of capital, a home’s profitability (its marginal product, as economists call it) can be calculated as the dollar value it creates during a year – after subtracting depreciation, costs of labor, maintenance, and intermediate goods – per dollar invested.

Owners of capital prefer their capital to be more profitable, rather than less. It’s the profitability of capital (after taxes and subsidies – more on those below) that makes an owner willing to purchase capital in the first place.

For a home that is rented or vacant, the home profitability calculation is straightforward: take the landlord’s rental income for the year, subtract depreciation, costs of maintenance and labor expenses (if any) for the same year and divide by the amount the home is worth.

A home occupied by its owner does not have rental income per se – the homeowner does not pay himself rent – but a hypothetical rental income can be imputed to owner-occupied homes by looking at the rental incomes earned by comparable homes that are on the rental market.

The Bureau of Economic Analysis makes rent imputations part of its preparation of gross domestic product and other items in the national economic accounts and reports a total value created during the year by all homes. It also estimates an aggregate home value (more precisely, it estimates what it would cost to replace each home with one just like it). The table below shows its results for the year 2000.

The first row of the table reports that the homes, condos and mobile homes in the United States in the year 2000 provided $1.01 trillion of housing services — that is, the value of the housing in rent or imputed rent — in the year 2000 (next week I’ll examine the years 2000 to 2010 more closely). A total of $160 billion of intermediate goods and services — mainly closing costs and brokers’ fees — were used, and the normal rate of depreciation of homes amounted to a cost of $162 billion.

The housing sector employed some people, such as apartment managers, with a payroll that amounted to $11 billion. But the Bureau of Economic Analysis does not try to measure the value of time of people who maintain their own homes, and for this reason the housing sector’s labor costs are significantly understated. With this caveat, the operating surplus of the housing sector was $677 billion.

The housing stock in the year 2000 was worth $10.63 trillion, so the housing sector’s operating surplus amounted to 6.4 percent of the value of its housing capital. (This surplus was divided among owners and mortgage lenders, but today I focus on the overall profitability of capital without regard to how it is shared).

Luke Threinen, a University of Chicago student, and I used these ingredients, and some minor adjustments for inflation during the year, to calculate housing profitability rates for 80 years. The chart below displays our results for the years 1950 to 2000.

The profitability of housing fluctuated between 4 and 7 percent, averaging 5.7 percent over the decades shown in the chart. The 5.7 percent may seem large, because few of us earn that much on our savings accounts.

Part of the explanation is that most homes owe property tax, and the chart shows pretax profitability (another part of the explanation is that savings accounts are generally low-return investments).

More notable is the comparison of housing’s 5.7 percent profitability to the profitability of business capital. Mr. Threinen and I made analogous calculations for the nonresidential sector – annual business revenue minus variable (nontax) costs expressed as a percentage of the amount of business capital – and found business capital to have a profit rate of 15.3 percent.

We concluded that the profitability gap between nonresidential and residential capital shows that our economy was overinvested in housing, long before the housing boom of the early 2000s.

Business capital has been so profitable to the economy because it is more scarce. It’s the law of demand: the less business capital there is, the higher the rate of return that remaining business capital earns because each unit of capital serves more customers. A low profit rate for housing is a symptom of its abundance.

With each dollar of business capital adding 15.3 cents of value per year and each dollar of housing capital adding only 5.7 cents, total value added in the economy would have been greater if some of the housing investment had been business investment instead, even before the housing boom. Each $10 billion of housing investment that could have been redirected to business investment have added almost a billion dollars to G.D.P.

Business capital has been more scarce largely because of business taxes and housing subsidies. Houses are not tax-free — I mentioned the property tax — but, in addition to property taxes, investors in businesses also owe sales taxes, corporate income taxes and personal income taxes.

The business-residential profitability gap is almost 10 percent, but our attempts to adjust both profit rates for applicable taxes show that the after-tax profitability gap is zero to five percentage points.

I mentioned last week how, in the short run, capital income taxes result in lower after-tax returns for the owners of the capital being taxed. In the long run, the opposite occurs: investors have choices about where and how to invest, so the after-tax profitability of housing has to be pretty close to the after-tax profitability of business.

Business taxes cause an underinvestment in business capital – scarcity is the only way business capital can earn an after-tax return that compares with the return on lesser-taxed housing – and overinvestment in housing.

The Marginal Products of Residential and Nonresidential Capital through 2010 Q2

We will soon be updating our paper.



The level of the non-residential series changed a bit when we changed our treatment of indirect business taxes. Here's the (slightly) revised introduction:

Economic theory suggests that marginal product of capital series might help predict economic growth forward one or two years, even under abnormal conditions such as wartime or depression. In some situations, the marginal product of capital is an essential ingredient in cost-benefit analyses (Harberger 1968; Byatt, et al., 2006; Mityakov and Ruehl, 2009). Evidence on the marginal product of capital can also help test various explanations for business cycles, help identify causes and consequences of the recent housing “bubble,” and help quantify the economic burdens of business taxes. The purpose of this paper is to produce annual and quarterly estimates of the marginal product of capital (net of depreciation), one each for the residential and nonresidential sectors of the U.S. economy.

By definition, the marginal product of capital net of depreciation is the change in net domestic product (NDP) during the accounting period (e.g., one quarter) that would result from an increase in the beginning-of-period capital stock of $1 worth of capital, holding constant the total supply of all other factors. The additional $1 of capital is assumed to have the same composition as the rest of the capital stock. For example, if the economy’s capital consisted of 400 identical structures and 100 identical vehicles, each of which cost $2 to acquire, then the marginal product of capital would be the extra NDP attained by starting the quarter with 400.4 identical structures and 100.1 identical vehicles (that is, $0.80 worth of structures and $0.20 worth of vehicles).

Suppose that origins of the current recession could be traced back to limits on the supply of aggregate investment due to a “credit crunch.” (Real investment did fall through the first year and a half of this recession.) The credit crunch theory says that the marginal product of capital would rise over this period as a consequence of the increased cost of capital faced by those with new capital projects. Alternatively, a financial crisis or something else could reduce labor usage more directly, and, given the complementarity of labor and non-residential capital in production, a fall in non-residential investment would merely result from low marginal products of capital, thereby putting the non-residential capital stock on a path that is consistent with a lesser amount of labor usage (Mulligan, 2010).

The marginal product of capital is also interesting as an aggregate leading indicator of business conditions, which is the motivation for its use in a number of studies (e.g., Feldstein and Summers (1977), Auerbach (1983)). This relationship alone may make it a predictor of subsequent economic growth.

Additionally, Fisherian consumption-saving theory suggests that the marginal product of capital, or variations of it, should predict consumption growth. In a Robinson Crusoe economy, the consumer would save for the future by reducing current consumption and using the proceeds to build capital assets. She would then use the marginal product and capital gains from those assets to add to consumption in the future. Because the saving decision is made in the present while the principal and interest are spent in an uncertain future, the incentive to save depends on, among other things, the expected marginal product and expected capital gains. The current marginal product itself helps predict the incentive to save only to the extent that it is closely related to the expected sum of future marginal product and capital gains. For this reason, we present measures of the marginal product that might be more indicative of those expected gains, and (consistent with national accounting practices: see Fraumeni, 1997) measures of depreciation that reflect expected depreciation and obsolescence, rather than actual depreciation and obsolescence.

It is helpful to examine the marginal product of residential capital separately from the marginal product of non-residential capital for at least two reasons. For one, the aggregate demand for labor is expected to have a closer relationship with the stock of non-residential capital than the stock of houses, because workers use non-residential capital in doing their jobs. Additionally, some important capital market distortions – such as business taxes and the “housing bubble” – are expected to have opposite effects on the stocks of residential and non-residential capital, and thereby opposite effects on their marginal products.

Section II presents our methods for calculating annual marginal products, and discusses the findings for 1930-2009. The marginal product of capital is very different in the residential and non-residential sectors, both in terms of levels and fluctuations. Section III examines the importance of taxes in explaining the gap between marginal products in the two sectors. The methods and results for quarterly postwar marginal products through 2009-IV are presented in Section IV. In order to isolate some of the possible determinants of measured marginal products, Section V compares them with average products. Section VI concludes, and Appendices record the time series values discussed in the body of the paper.

The Marginal Products of Residential and Non-residential Capital

Economic theory suggests that marginal product of capital series might help predict economic growth forward one or two years, even under abnormal conditions such as wartime or depression. In some situations, the marginal product of capital is an essential ingredient in cost-benefit analyses (Harberger 1968; Byatt, et al., 2006; Mityakov and Ruehl, 2009). Evidence on the marginal product of capital can also help test various explanations for business cycles, and help identify causes and consequences of the recent housing “bubble.” The purpose of this paper with Luke Threinen is to produce annual and quarterly estimates of the marginal product of capital (net of depreciation), one each for the residential and nonresidential sectors of the U.S. economy.

By definition, the marginal product of capital net of depreciation is the change in net domestic product (NDP) during the accounting period (e.g., one quarter) that would result from an increase in the beginning-of-period capital stock of $1 worth of capital. In particular, the additional $1 of capital would have the same composition as the rest of the capital stock. For example, if the economy’s capital consisted of 400 identical structures and 100 identical vehicles, each of which cost $2 to acquire, then the marginal product of capital would be the extra NDP attained by starting the quarter with 400.4 identical structures and 100.1 identical vehicles (that is, $0.80 worth of structures and $0.20 worth of vehicles).

Suppose that origins of the current recession could be traced back to limits on the supply of aggregate investment due to a “credit crunch.” In fact real investment fell through the first year and a half of this recession, but the credit crunch theory says that the marginal product of capital would rise as a consequence of the increased cost of capital faced by those with new capital projects. Alternatively, financial crisis or something else could reduce labor usage more directly, and, given the complementarity of labor and non-residential capital in production, non-residential investment would merely respond to low marginal products of capital, thereby putting the non-residential capital stock on a path that is consistent with a lesser amount of labor usage (Mulligan, 2010).

The marginal product of capital is also interesting as an aggregate leading indicator of business conditions, which is the motivation for its use in a number of studies (e.g., Feldstein and Summers (1977), Auerbach (1983)). This relationship alone may make it a predictor of subsequent economic growth.

Additionally, Fisherian consumption-saving theory suggests that the marginal product of capital, or variations of it, should predict consumption growth.





Over the last ten years, the marginal and average products of residential capital fell, and then increased, as housing construction was booming and busting. In this sense, the residential data suggests that the supply of residential capital shifted along a relatively stable demand for the services of that capital. As indicated by the marginal product of residential capital at the end of 2009, current housing supply seems restricted by comparison with the housing boom (when the residential MPK was low), but fairly normal by comparison with the 1990s when the residential MPK was similar to what is was at the end of 2009. These patterns are consistent with the findings of Davis, Lehnert, and Martin (2008) and others that housing rent-price ratios were low during the housing boom, and with the conclusions that the housing boom was fueled by optimistic expectations, or by easy credit.

The marginal product of non-residential capital was much higher during the housing boom than it was during the recession, when rates of investment in non-residential equipment and software were low. In this sense, the supply of non-residential capital seems less restricted during the recession than it was before. In other words, the recession’s investment rates may have been low because of a slack labor market, rather than the other way around. In any case, the testing of various theories of this recession, and the prior housing cycle, can be enhanced with marginal products data like those shown in this paper.

Credit Crunch Hard to Find in the Investment Data

If the credit crunch were the big deal people say it is, then you would think it would be readily visible in the investment data.

It's impossible to see a credit crunch in the first two graphs -- investment looks like it does in a "run-of-the-mill" recession (recessions that needed no trillion-dollar-financial-sector bailout).

The last graph shows non-residential investment. As Luke Threinen and I have pointed out, non-residential investment has been unusually strong in this recession, at least through 2008 Q3. 2008 Q4 takes a dip down -- maybe that's an effect of a credit crunch -- but that dip has a number of precedents in previous recessions.

How Long Until the Stock Market Fully Rebounds?

Copyright, The New York Times Company

Even if the economy recovers quickly, I am skeptical that stock prices would soon return to what they were in 2007. During the housing boom, so many resources were devoted to housing that it was hard for new companies in other sectors to find the funding they needed to get off the ground or significantly expand their operations. This put incumbent companies at an advantage in the marketplace, and made them more valuable to stockowners. With the housing bubble now burst, a situation like this will not happen again soon.

Stock prices, such as those measured by the Dow Jones industrial average, represent the market’s valuation of ownership of established corporations. One factor contributing to the reduction in stock prices during 2008 was that markets became concerned that the earnings of those corporations were going to be reduced for a while as the economy endures a potentially severe recession. As fourth quarter earnings reports are released, those concerns now seem warranted.

Even if the economy had not been in a recession, or the recession proved to end quickly, the earnings of established corporations were going to suffer – thanks to the legacy of the housing boom and bust.

One stock market legacy of the housing boom and bust is that a number of established corporations – especially those in the banking industry – own mortgages, typically bundled up into securities. Now that it is apparent that the housing sector is overbuilt, the residential properties that back those mortgages are not worth much. With homeowners defaulting on their mortgages left and right, and their homes worth very little when they do default, the earnings banks had hoped to obtain through those mortgages will not materialize.

Even stock prices in the non-financial non-residential sectors (that is, the stock prices of businesses that do own few, if any, mortgages) were elevated by the housing boom because of all of the capital used by the housing sector. Last week I showed how the non-residential investment had suffered during the housing boom. The housing boom was a time when established non-residential businesses had a measure of protection against new entry into their industries and expansion by their competitors, because some of the resources needed for that expansion had been diverted to housing.

The housing boom is over, and it seems that the United States economy can go a long time without building houses at the rate it did in 2002-2006. This measure of protection for established businesses is now gone. When the economy recovers, investment in new and expanding companies will keep a lid on the earnings – and thereby stock prices – of established, incumbent corporations.

A Commercial Real Estate Crisis? Probably Not

Copyright, The New York Times Company

For months now experts have been predicting that commercial real estate will be “the other shoe to drop.” But in fact, non-residential building fell far behind housing construction during the housing boom. This shortage of commercial buildings relative to housing suggests that a commercial real estate crisis will not occur, or at worst it will occur with much less severity than did the housing crash.

While there is much disagreement as to the proper remedies for the current economic situation, there is wide agreement that the housing boom and crash that followed were the major factor in putting us where we are today. These days, “real estate” is a term that provokes fear, not optimism. Nevertheless, it is a mistake to assume that commercial real estate shares the housing sector’s ailments.

The chart below shows the amount of housing and non-residential structures in place in America. The amounts are measured at the beginning of each year, relative to the 1990-2000 trend. By the beginning of 2007, the amount of housing – that is, square footage adjusted for quality – was 3.5 percent above that previous trend. The amount of non-residential structures was 2.0 percent below trend.

We all know that there is a nationwide surplus of housing. But there is little if any nationwide surplus of non-residential buildings.

Business conditions have deteriorated recently, so it might seem that even a normal amount of commercial real estate would be too much these days. However, we probably ended the housing boom with enough of a commercial real estate shortage that economic activity could “back up” a bit toward the amount of commercial real estate.

For example, the chart suggests that non-residential building is now about 2 percent below trend. Employment has fallen about 2 percent so far during this recession. Inflation-adjusted gross domestic product is down one percent in the last six months, and down about 0.2 percent for the recession as a whole. Thus the real G.D.P. and employment shortfalls so far are in line with the relatively small amounts of non-residential building.

I continue to watch the economy in 2009 but, barring a significant further decline in business activity, I do not expect to see a nationwide surplus of commercial real estate and therefore do not expect to see commercial real estate suffer the kind of crisis that followed from the housing surplus.

National Income Release Likely shows High MPK

Part of BEA's update today of 2008 Q3 GDP is a first release of 2008 Q3 National Income. It includes corporate profits, which permit me to calculate a marginal product of capital for the corporate nonresidential sector. Corporate profits declined less than 1%. Capital income more broadly decline less than 1%.

Before this release, my best guess was capital income was "down about 0.6%" Q2 to Q3.

I am still making a final calculation of the Q3 marginal product of capital, but a 1% capital income decline is trivial; in order for the marginal product of capital to decline from its recent highs to more normal levels, capital income would have to fall 10s of percentage points.

This is quite different from the 1930s: the marginal product of capital was already quite low when the major bank panics hit. Today corporate profits are $1.5 trillion dollars per year and capital income more than $3 trillion per year. That's enough to pay for a lot of investment without banks' help.

Stocks, Flows, and the Funding of New Projects

The post is to clarify some confusion among blog readers. I wrote in the New York Times:

Although banks perform an essential economic function - bringing together investors and savers - they are not the only institutions that can do this. Pension funds, university endowments, and venture capitalists and corporations all bring money to new investment projects without any essential role played by banks. The average corporation receives about a quarter of its investment funds from the profits it has after paying dividends - and could obtain even more by cutting its dividend, if necessary.

Note that the funds referenced here are flows, and are distinct from the value of institutions that own them. A business needs cash flow -- from its own operations or some other institution -- to pay for investment. The value of the institution is only on paper. For example, a startup company could have tremendous value but no cash flow (or negative cash flow) from its operations and therefore needs outside funds to invest. Conversely, many corporations, pension funds, endowments etc., today have suffered losses in value but nonetheless have significant cash inflows from their investments and operations. Those adverse changes in no way refute or de-emphasize my point. Indeed, Luke Threinen and I predicted both reduced values and increased investment funding. Its poor cash flows that will get my attention.

The importance of flows is why I keep an eye on aggregate capital income in the economy -- the marginal product of capital -- and not the aggregate capitalization of the stock market or some other valuation measure. Aggregate capital income was low during the years prior to the 1930s bank panics; so far those flows are fine today.

Japan's Crisis began with a Low MPK

Japan's banking crisis began no earlier than 1993, because Japanese banks were making profits prior to that, and were taking losses 1993 through at least 2001 (Kashyap, 2002, Table 1).

I am still double-checking this, but it seems to me that non-financial capital was already in sorry shape by then. The attached graph shows non-financial sector net operating surplus per dollar of capital in the non-financial sector (seasonally adjusted by taking residuals from a regression on quarter dummies). It reached its lowest in 1993 and 1994.

In the U.S., by comparison, banks began taking significant losses in 2007. The non-financial sector was quite profitable then, and continues to be. The high U.S. marginal product of capital not only raises the growth rate forecast from which to subtract any adverse impact of liquidity crisis, but also reduces the expected magnitude of that impact.