February 26, 2002
Dan L.Crippen, Director
Congressional Budget Office
Second and D Streets, SW
Washington, DC 20515
Dear Dr. Crippen:
I am writing to you because I am concerned that the projections for capital gains tax revenue in the Economic and Budget Outlook: Fiscal Years 2003-2012 substantially overstate the gains that can be plausibly projected. This is attributable to the fact that the methodology CBO uses for these projections ignores three important factors that are likely to effect capital gains realizations in the coming decade:
1) CBO projects that profit growth over this period will be less than half its historic average,
2) the price to earnings ratio in the stock market is currently more than 60 percent above its long-term average, making a further downward correction likely, and
3) the share of U.S. financial assets held by foreigners is increasing rapidly, which should make a smaller percentage of domestically generated capital gains subject to taxation in future years.
As a result of not taking into account these factors, the projections are likely to overstate capital gains tax revenue by several hundred billion dollars over the projection period. While CBO has adjusted its projections of capital gains realizations downward in this year’s Outlook -- lowering projected revenues by $142 billion -- the most recent projections for capital gains are still inconsistent with CBO's projections for the economy, leading to a substantial overstatement.
Before addressing these issues directly, it is worth noting that CBO projects that realized capital gains will be far higher relative to GDP than they have been historically. The table below shows the projected realizations for 2001 through 2012 both in dollar terms and as a percentage of GDP. By comparison, capital gain realizations over the period from 1952 to 1998 averaged 2.8 percent.[1] While some factors would push towards a higher rate of realizations than normal, such as a lower capital gains tax rate than the average over this period, there are other important factors pushing in
Projected Capital Gain
Realizations
2001
$500
4.9%
2002
$476
4.6%
2003
$476
4.3%
2004
$479
4.1%
2005
$483
3.9%
2006
$492
3.8%
2007
$504
3.7%
2008
$520
3.6%
2009
$539
3.6%
2010
$561
3.5%
2011
$581
3.5%
2012
$604
3.4%
[Capital gain
realization projections are taken from The Economic and Budget Outlook, table
3-6, GDP projections from table 2-2.]
the opposite direction. The most significant item in this
category would be the growth of stock holdings in tax deferred retirement
accounts, such as 401(k)s. The gains from these holdings will eventually be
taxed as normal income, at the point when money is withdrawn. The share of
financial assets held in these accounts has been growing rapidly, a process that
should accelerate with recent changes that have allowed for larger contributions
to tax deferred accounts.[2]
The other important factor that
would lower the ratio of taxable realized capital gains to GDP is the rise of
home ownership. The vast majority of gains on owner occupied housing are exempt
from taxation. By contrast, gains on rental units are generally subject to
taxation. With the rate of homeownership rising throughout most of the last four
decades, the percentage of residential real estate where capital gains are
subject to taxation will be far lower over the next decade than it has been over
the last forty years. With the growth of tax
deferred accounts, and the increase in homeownership rates, there is little
reason to assume that capital gains will be a higher share of GDP in the future,
than in the past, even though the capital gains tax rate is slightly lower than
its average over the 1952-98 period.
The table below compares the CBO projections of capital gains tax revenue with the revenue that would be collected if capital gains realizations were equal to 2.8 percent of GDP over the decade, the same as their average over the period from 1952 to 1998. The projections assume the same tax rate on these realizations as the CBO projections. The table shows that the cumulative difference in capital gains tax revenue between CBO’s projections and projections
Projected Capital Gain Taxes
CBO
Historic Average
Difference
(billions)
2001
$115
$66
$49
2002
$98
$60
$38
2003
$95
$62
$33
2004
$95
$65
$30
2005
$95
$72
$23
2006
$96
$74
$22
2007
$98
$74
$24
2008
$100
$78
$22
2009
$104
$81
$23
2010
$108
$86
$22
2011
$112
$90
$22
2012
$116
$96
$20
Total $1232 $904 $328
[The
CBO projections for capital gains tax revenue appear in The Economic and Budget
Outlook, table 3-6. The historic averages are calculated as described in the
text using the GDP projections from table 2-2.]
that assume capital gains realizations will be equal to
their historic average share of GDP is $328 billion over the projection period.
However, even the assumption that capital gains realizations, over the
next decade, will be equal to the historic average share of GDP is questionable
given the other projections in the most recent Outlook.
Most importantly, the Outlook projects
that after-tax corporate profits will grow at an average nominal rate of 3.5
percent annually over the period from 2000 to 2012.[3]
By comparison, nominal profits grew at a 7.5 annual rate over the period from
1952 to 2000. (Since nominal gains are the basis of taxation, it is appropriate
to use nominal profit growth for this comparison.)
With projected profit growth that
is less than half its historic average, the price of stocks and other assets
would have to rise far more rapidly than the growth of corporate profits, in
order to provide the same ratio of capital gains to GDP as in past decades.[4]
While this can happen for any period of time, it is unusual for it to happen for
a period as long as a decade. Furthermore, it is inappropriate for CBO to
implicitly assume this sort of increase in the ratio of asset prices to
corporate profits without some explanation that could justify such a change in
historic patterns. Presumably, CBO would not project that the price to earnings
ratios will rise indefinitely.
It would be more plausible to
assume that asset prices rise at the same pace as corporate profits or
approximately 3.5 percent annually, rather than their 7.5 percent average rate
over the past half century. Since the current ratio of asset prices to corporate
profits is approximately 60 percent higher than its historic average, the slower
growth of asset prices should translate into capital gains realizations that are
approximately 78 percent of their historic average, measured as a share of GDP
([3.5/7.5]*1.6 = 0.778).[5]
The table below shows a set of projections of capital gains tax revenues that
assumes the ratio of realized gains to GDP is 77.8 percent of its historic level
and that revenue falls accordingly. The difference over the projection period is
$523 billion.
Projected Capital Gain Taxes
CBO
Revenue Assuming Gains Difference
Track Projected Profit Growth
(billions)
2001
$115
$47
$68
2002
$98
$48
$50
2003
$95
$51
$44
2004
$95
$56
$39
2005
$95
$58
$37
2006
$96
$58
$38
2007
$98
$58
$40
2008
$100
$61
$39
2009
$104
$63
$41
2010
$108
$67
$41
2011
$112
$70
$42
2012
$116
$75
$41
The assumption that asset values
will only rise as rapidly as the projected growth rate of corporate profits does
create another problem, in a context where projected profit growth is so slow.
The total return on stock is the dividend yield (including share buybacks) plus
capital gains. Since the price to earnings ratio in the market is already far
above its historic average of 14.5 to 1, the dividend yield is far below its
historic average.[6]
The pure dividend yield is currently just 1.36 percent (Economic Report of the President, 2002, Table B-95), but it would be
approximately 2.0 percent when share buybacks are included. A nominal growth
rate of 3.5 percent translates into a real growth rate of 1.0 percent given
CBO’s projected inflation rate of 2.5 percent. This translates into a total
real return on stock of 3.0 percent. This return is far below the average real
return of 7.0 percent that stocks have produced over the last century. It also
provides no risk premium on the return available on government bonds. With a
projected nominal interest rate of 5.8 percent, ten year bonds would provide a
real return of 3.3 percent. The assumption that stocks would provide no risk
premium relative to bonds over the next decade seems implausible on its face.
This leaves a third possibility,
if the CBO projections of profit growth are accepted as accurate. Over the next
decade, stock prices will have to fall significantly from their current levels.
This decline would raise the ratio of earnings to the share price, which would
allow the dividend yield to move closer to its historic average, which would
allow total returns to be closer to their historic average. The required decline
in stock prices could be quite large. If real capital gains averaged 1.0
percent, then a 6.0 percent dividend yield would be needed to maintain the
historic rate of return on equities. This could only be accomplished if the
price to earnings ratio declined to 10 to 1, which would imply that stocks lost
approximately 60 percent of their current value.
However, the decline in share
prices that is needed to restore a long-term balance to financial markets is
probably somewhat smaller than this calculation implies. In the longer term, it
is likely that profits will grow at the same pace as GDP, which is close to 2.0
percent annually in CBO’s longer-term projections. Also, investors may be
willing to accept a risk premium that is less than the historic 4.0 percent gap
between stocks and long-term bonds, if they don’t view stocks as being as
risky as in the past. Nonetheless, if the CBO projections of profit growth are
assumed to be accurate, then it will be necessary to have a large decline in
stock prices in order to restore a reasonable risk premium.
During the period when the stock
market is correcting, the ratio of capital gains to GDP will be far below its
historic average. After the correction has been completed, the ratio would
revert back closer to its historic average, although it would still be lower, if
the projected rate of profit growth is slower than its historic average. Exactly
how far the market will correct, and the timing of the correction, cannot be
accurately predicted. But, in order to derive projections of capital gains
revenue, CBO should at least make an explicit set of assumptions on the nature
of a correction -- and provide an analysis of the sensitivity of these
projections to alternative paths for the market. This would at least provide
Congress with an understanding of the issues involved and a plausible range of
projections of capital gains revenue.
Needless to say, any revenue
projection of capital gains tax revenue that assumed that the stock market would
decline over much of the next decade would be considerably more pessimistic than
the one that appears in the current Outlook. If stock prices rose at a rate of approximately 0.9 percent
annually, it would bring the price to earnings ratio back to its historic
average by 2012, if the CBO projections of profit growth prove accurate. [7][8]
This rate of growth of stock prices would lead to lower capital gains tax
revenue than is projected in the table, possibly raising the size of the
overstatement by several hundred billion dollars.
The last factor that does not
appear to have been taken into account in the CBO projections of capital gains
is the increasing portion of U.S. assets that is owned by foreign corporations
or individuals. This is a direct result of the large current account deficits
that the United States has been running in recent years. In 2000, the United
States net international indebtedness position increased by $662.1 billion ((Economic
Report of the President, 2002, Table B-107). While full year figures are not
yet available, with the trade deficit numbers to date, it appears that net
indebtedness increased by a comparable amount in 2001. While the Outlook
does not explicitly present projections for the trade deficit, its analysis
seems to assume that it will not change significantly, relative to GDP, through
the ten-year projection period. (Since the projections imply that the United
States will be growing faster than the major purchasers of U.S. goods, the only
way in which the trade deficit could be corrected is with a large [20-30
percent] decline in the value of the dollar. A decline of this magnitude would
imply an increase in the rate of inflation of between 1.5-2.3 percentage points
[assuming a 50 percent passthrough from currency devaluation on the price of
imported goods]. The inflation projections do not appear to be factoring in this
sort of decline in the dollar.
The table below shows the
projected net indebtedness of the United States assuming that trade deficit
remains constant as a share of GDP at its pre-recession level. The projected
levels of indebtedness are shown in both dollar terms and measured as a share of
the total financial assets held by the household and non-profit sectors. It is
assumed that both the value of foreign holdings and the value of United States
financial assets increase at a nominal rate of 5 percent a year.
Net
Foreign
Combined Financial Assets
ForeignAssets/Domestic
Asset
Position
Household and Non-Profit Sectors
(column 1/ column 2)
(billions)
2001
$2,657
$30,404
8.7%
2002
$3,159
$31,923
9.9%
2003 $3,709
$33,520
11.1%
2004 $4,307
$35,196
12.2%
2005 $4,959
$36,956
13.4%
2006 $5,667
$38,804
14.6%
2007 $6,435
$40,744
15.8%
2008 $7,267
$42,781
17.0%
2009 $8,168
$44,920
18.2%
2010 $9,143
$47,166
19.4%
2011 $10,195
$49,525
20.6%
2012 $11,332
$52,001
21.8%
The increasing share of foreign
holdings of U.S. assets is relevant to projections of capital gains tax revenue
because the gains associated with these holding generally will not be subject to
taxation in the United States. For example, increased profits for
Daimler-Chrysler would be expected to boost the values of its shares in Germany,
but it would not lead to significant capital gains which would be subject to
taxation in the United States. Through most of the period 1952-98 the United
States was a large net creditor. This meant that not only were the overwhelming
majority of capital gains due to business activity in the United States subject
to domestic taxation, but a large amount of capital gains due to foreign
business activity was subject to domestic taxation as well. Now that the United
States has become a net debtor nation, and its level of indebtedness is expected
to increase substantially over the next decade, the amount of gains subject to
domestic taxation is likely to decrease substantially relative to the gains that
should be generated by the U.S. economy. This simple extrapolation implies that
less than 80 percent of the gains
generated in the U.S. economy will be subject to domestic taxation by 2012. In
other words, the CBO projections could be overstating the amount of gains
subject to domestic taxation by 20 percent at the end of the projection period.
Obviously, an accurate projection
would require assessing more carefully which U.S. assets foreign investors are
likely to hold, and whether these assets are likely to generate capital gains.
In addition, some capital gains on foreign owned assets will be subject to
domestic taxation. Nonetheless, given the growth of the net indebtedness of the
United States, which is implicit in the CBO projections, it is no longer
possible to ignore the increase in foreign ownership of U.S. financial assets in
projections of tax revenue.
In summary, the CBO's projections
of capital gains tax revenue are likely to still be substantially overstated,
even after the large downward adjustment made in the most recent Economic
and Budget Outlook. Assuming that stock prices grow at the same rate as CBO
projects profits will grow would reduce projected capital gains tax revenue by
close to $500 billion over the projection period. If there is a correction in
the stock market, restoring historic price to earnings ratios, then the
overstatement of tax revenue will be considerably larger. Finally, the fact that
the net indebtedness of the United States is increasing rapidly implies that a
much larger portion of U.S. financial assets will be foreign owned at the end of
the projection period. As a result 20 percent, or more, of domestically
generated capital gains could escape taxation by the end of the projection
period.
I hope that you will have the
opportunity to give these issues some consideration. If you have any questions
on the points I have raised, I would be happy to discuss them with you.
Regards,
Dean Baker, Co-Director
cc:
Senator Kent Conrad
Senator Pete V. Domenici
Representative Jim Nussle
Representative John Spratt
[1] Congressional Budget Office, November 1995, “Projecting Capital Gains Realizations,” Table 1 combined with data on realizations for years after 1994 from table 3-6 in the Economic and Budget Outlook: Fiscal Years 2003-2012.
[2] The recent change in the tax code, which allows for considerably larger contributions to IRAs for individuals over age 50, does not appear to have been taken into account in the most recent Outlook.
[3] Corporate book profits are projected at $1,407 billion in 2012, with taxes of $335 billion (Table 3-7), leaving after-tax profits of $1,072. In 2000, corporate profits were $920 billion and taxes were $207 billion, putting after-tax profits at $713 billion. The year 2000 is taken as a starting point, both because it is the most recent year for which there is complete data on profits, and because it was near a profit peak in the business cycle. Profits for 2001 will be considerably lower, which means that the growth rate to 2012 will be higher, but the base price to earnings ratio would be correspondingly higher as well. For purposes of this analysis it is more useful to work from the peak (or near peak) profits of 2000. Nothing would be changed by using 2001 as a starting point, although it would make the discussion somewhat more complicated.
[4] This discussion focuses on capital gains in the stock market, but it is reasonable to assume that gains on most other financial assets will follow a similar pattern. For example, the value of privately held corporations or proprietorships would be expected to follow trends in equity valuation. Miller and Ozanne tested a set of variables not related to the stock market and found that they had little predictive power, with housing starts alone leading to a modest improvement in the accuracy of predicted capital gains realizations ("Forecasting Capital Gains Realizations," Congressional Budget Office, August 2000).
[5] There is some hangover of accumulated gains from the rapid run-up in the stock market in the late nineties. However, this is likely to dissipate relatively quickly. In many cases, active traders have already cashed in on these gains. Also, a slowly rising market will mean that many traders will have losses to offset past gains.
[6] A high price to earnings ratio means a low earnings to price ratio. Since firms typically pay out 50-60 percent of their profits in dividends (including share buybacks), a high price to earnings ratio virtually guarantees a low dividend yield.
[7] The base for the current valuation of corporate equities ($14.2 trillion) is taken from the Federal Reserve Board's Flow of Fund's Table L213, line 20, for the second quarter of 2001.
[8] While this slow rate of growth of stock prices would bring the price to earnings ratio back to its historic average by 2012, it would not be sufficient to restore the historic risk premium on stocks. Since the projected growth of the economy in subsequent years is just 2.0 percent, measured against the CPI, if stock prices and profits both grew at the same rate as GDP, the return on stocks would be just 6.1 percent, assuming that firms pay out 60 percent of their profits in either dividends or share repurchases. In order to restore the historic risk premium in an era of slow growth, the price to earnings ratio would have to fall below its historic average.