Capital in the Twenty-First Century (118 page)

34
. There was a recent proposal to pay international organizations the proceeds of a
global wealth tax. Such a tax would become independent of nationality and could become
a way to protect the right to multinationality. See Patrick Weil, “Let Them Eat Less
Cake: An International Tax on the Wealthiest Citizens of the World,”
Policy Network,
May 26, 2011.

35
. This conclusion is similar to that of Dani Rodrik, who argues that the nation-state,
democracy, and globalization are an unstable trio (one of the three must give way
before the other two, at least to a certain extent). See Dani Rodrik,
The Globalization Paradox: Democracy and the Future of the World Economy
(New York: Norton, 2011).

36
. The system of “allowance for corporate equity” adopted in Belgium in 2006 authorizes
the deduction from taxable corporate profits of an amount equal to the “normal” return
on equity. This deduction is said to be the equivalent of the deduction of interest
on corporate debt and is supposed to equalize the tax status of debt and equity. But
Germany and more recently France have taken a different take: limiting interest deductions.
Some participants in this debate, such as the IMF and to a certain extent the European
Commission, claim that the two solutions are equivalent, although in fact they are
not: if one deducts the “normal” return on both debt and equity, it is highly likely
that the corporate tax will simply disappear.

37
. In particular, taxing different types of consumption goods at different rates allows
for only crude targeting of the consumption tax by income class. The main reason why
European governments are currently so fond of value-added taxes is that this type
of tax allows for de facto taxation of imported goods and small-scale competitive
devaluations. This is of course a zero-sum game: the competitive advantage vanishes
if other countries do the same. It is one symptom of a monetary union with a low level
of international cooperation. The other standard justification of a consumption tax
relies on the idea of encouraging investment, but the conceptual basis of this approach
is not clear (especially in periods when the capital/income ratio is relatively high).

38
. The purpose of the fiscal transactions tax is to decrease the number of very high-frequency
financial transactions, which is no doubt a good thing. By definition, however, the
tax will not raise much revenue, because its purpose is to dry up its source. Estimates
of potential revenues are often optimistic. They cannot be much more than 0.5 percent
of GDP, which is a good thing, because the tax cannot target different levels of individual
incomes or wealth. See the online technical appendix.

39
. See
Figures 10.9
–
11
. To evaluate the golden rule, one must use the pretax rate of return on capital (supposed
to be equal to the marginal productivity of capital).

40
. The original article, written with a certain ironic distance in the form of a fable,
is worth rereading: Edmund Phelps, “The Golden Rule of Accumulation: A Fable for Growthmen,”
American Economic Review
51, no. 4 (September 1961): 638–43. A similar idea, expressed less clearly and without
allusion to the golden rule, can be found in Maurice Allais’s
Economie et intérêt
(Paris: Librairie des Publications Officielles, 1947) and in articles by Von Neumann
(1945) and Malinvaud (1953). Note that all this work (including Phelps’s article)
is purely theoretical and does not discuss what level of accumulation would be required
to make
r
equal to
g
. See the online technical appendix.

41
. Capital’s share is given by
α
=
r
×
β
. In the long run,
β
=
s
/
g
, so
α
=
s
×
r
/
g
. It follows that
α
=
r
if
r
=
g
, and
α
>
s
if and only if
r
>
g
. See the online technical appendix.

42
. The reasons why the golden rule establishes an upper limit are explained more precisely
in the online technical appendix. The essential intuition is the following. Beyond
the level of capital described by the golden rule, that is, where the return on capital
sinks below the growth rate, capital’s long-run share is lower than the savings rate.
This is absurd in social terms, since it would take more to maintain the capital stock
at this level than the capital returns. This type of “dynamic inefficiency” can occur
if individuals save without worrying about the return: for example, if they are saving
for old age and their life expectancy is sufficiently long. In that case, the efficient
policy is for the state to reduce the capital stock, for example, by issuing public
debt (potentially in large amounts), thus de facto replacing a capitalized pension
system by a PAYGO system. This interesting theoretical policy never seems to occur
in practice, however: in all known societies, the average return on capital is always
greater than the growth rate.

43
. In practice, a tax on capital (or public ownership) can ensure that the portion
of national income going to income on private capital (after taxes) is less than the
savings rate without needing to accumulate so much. This was the postwar social-democratic
ideal: profits should finance investment, not the high life of stockholders. As the
German chancellor Helmut Schmidt said, “Today’s profits are tomorrow’s investments
and the day after tomorrow’s jobs.” Capital and labor work hand in hand. But it is
important to understand that this depends on institutions such as taxes and public
ownership (unless we imagine unprecedented levels of accumulation).

44
. In a sense, the Soviet interpretation of the golden rule simply transferred to the
collectivity the unlimited desire for accumulation attributed to the capitalist. In
chapters 16 and 24 of
The General Theory of Employment, Interest, and Money
(1936), where Keynes discusses “the euthanasia of the rentier,” he develops an idea
close to that of “capital saturation”: the rentier will be euthanized by accumulating
so much capital that his return will disappear. But Keynes is not clear about how
much this is (he does not mention
r
=
g
) and does not explicitly discuss public accumulation.

45
. The mathematical solution to this problem is presented in the online technical appendix.
To summarize, everything depends on what is commonly called the concavity of the utility
function (using the formula
r
=
θ
+
γ
×
g
, previously discussed in
Chapter 10
and sometimes called the “modified golden rule”). With infinite concavity, one assumes
that future generations will not need a hundredth additional iPhone, and one leaves
them no capital. At the opposite extreme, one can go all the way to the golden rule,
which may necessitate leaving them several dozen years of national income in capital.
Infinite concavity is frequently associated with a Rawlsian social objective and may
therefore seem tempting. The difficulty is that if one leaves no capital for the future,
it is not at all certain that productivity growth will continue at the same pace.
Because of this, the problem is largely undecidable, as perplexing for the economist
as for the citizen.

46
. In the most general sense, a “golden rule” is a moral imperative that defines people’s
obligations to one another. It is often used in economics and politics to refer to
simple rules defining the current population’s obligations to future generations.
Unfortunately, there is no simple rule capable of definitively resolving this existential
question, which must therefore be asked again and again.

47
. These figures were retained in the new treaty signed in 2012, which added a further
objective of maintaining a “structural” deficit of less than 0.5 percent of GDP (the
structural deficit corrects for effects of the business cycle), along with automatic
sanctions if these commitments were not respected. Note that all deficit figures in
European treaties refer to the secondary deficit (interest on the debt is included
in expenditures).

48
. A deficit of 3 percent would allow a stable debt-to-GDP ratio of 60 percent if nominal
GDP growth is 5 percent (e.g., 2 percent inflation and 3 percent real growth), in
view of the formula
β
=
s
/
g
applied to the public debt. But the argument is not very convincing (in particular,
there is no real justification for such a nominal growth rate). See the online technical
appendix.

49
. In the United States, the Supreme Court blocked several attempts to levy a federal
income tax in the late nineteenth and early twentieth centuries and then blocked minimum
wage legislation in the 1930s, while finding that slavery and, later, racial discrimination
were perfectly compatible with basic constitutional rights for nearly two centuries.
More recently, the French Constitutional Court has apparently come up with a theory
of what maximum income tax rate is compatible with the Constitution: after a period
of high-level legal deliberation known only to itself, the Court hesitated between
65 and 67 percent and wondered whether or not it should include the carbon tax.

50
. The problem is similar to that posed by the return on PAYGO retirement systems.
As long as growth is robust and the fiscal base is expanding at a pace equal (or nearly
equal) to that of interest on the debt, it is relatively easy to reduce the size of
the public debt as a percentage of national income. Things are different when growth
is slow: the debt becomes a burden that is difficult to shake. If we average over
the period 1970–2010, we find that interest payments on the debt are far larger than
the average primary deficit, which is close to zero in many countries, and notably
in Italy, where the average interest payment on the debt attained the astronomical
level of 7 percent of GDP over this period. See the online technical appendix and
Supplemental Table S16.1, available online.

51
. If the issue is constitutionalized, however, it is not impossible that a solution
such as a progressive tax on capital would be judged unconstitutional.

52
. On the way Stern and Nordhaus arrive at their preferred discount rates, see the
online technical appendix. It is interesting that both men use the same “modified
golden rule” I described earlier but reverse positions entirely when it comes to choosing
the concavity of the social utility function. (Nordhaus makes a more Rawlsian choice
than Stern in order to justify ascribing little weight to the preferences of future
generations.) A logically more satisfactory procedure would introduce the fact that
the substitutability of natural capital for other forms of wealth is far from infinite
in the long run (as Roger Guesnerie and Thomas Sterner have done). In other words,
if natural capital is destroyed, consuming fewer iPhones in the future will not be
enough to repair the damage.

53
. As noted, the current low interest rates on government debt are no doubt temporary
and in any case somewhat misleading: some countries must pay very high rates, and
it is unlikely that those that are borrowing today at under 1 percent will continue
to enjoy such low rates for decades (analysis of the period 1970–2010 suggests that
real interest rates on long-term public debt in the rich countries is around 3 percent;
see the online technical appendix). Nevertheless, current low rates are a powerful
economic argument in favor of public investment (at least as long as such rates last).

54
. Over the last several decades, annual public investment (net of depreciation of
public assets) in most rich countries has been about 1–1.5 percent of GDP. See the
online technical appendix and Supplemental Table S16.1, available online.

55
. Including tools such as the carbon tax, which increases the cost of energy consumption
as a function of the associated emission of carbon dioxide (and not as a function
of budget variations, which has generally been the logic of gasoline taxes). There
is good reason to believe, however, that the price signal has less of an impact on
emissions than public investment and changes to building codes (requiring thermal
insulation, for example).

56
. The idea that private property and the market allow (under certain conditions) for
the coordination and efficient use of the talents and information possessed by millions
of individuals is a classic that one finds in the work of Adam Smith, Friedrich Hayek,
and Kenneth Arrow and Claude Debreu. The idea that voting is another efficient way
of aggregating information (and more generally ideas, reflections, etc.) is also very
old: it goes back to Condorcet. For recent research on this constructivist approach
to political institutions and electoral systems, see the online technical appendix.

57
. For example, it is important to be able to study where political officials from
various countries stand in the wealth and income hierarchies (see previous chapters).
Still, statistical summaries might suffice for the purpose; detailed individual data
are generally not needed. As for establishing trust when there is no other way to
do so: one of the first actions of the revolutionary assemblies of 1789–1790 was to
compile a “compendium of pensions” that listed by name and amount the sums paid by
the royal government to various individuals (including debt repayments, pensions to
former officials, and outright favors). This sixteen-hundred-page book contained 23,000
names and listed detailed amounts (multiple sources of income were combined into a
single line for each individual), the ministry involved, the age of the person, the
final year of payment, the reasons for the payment, etc. It was published in April
1790. On this interesting document, see the online technical appendix.