Guided Tour

This document is a guided tour of the Wealth Distribution tables on the Wealth and Want website. You'll get the most out of it if you open a second and third window next to this one for the tables. [Right click on the links, and choose "open in new window."] Resize them so you can see the tables and this tour side by side. Each table is numbered, and within each table, the columns are numbered and each line has a number. If you want to print out the tables, you might experiment with your browser's ability to zoom and to resize the type. You might also want to print out the SCF Definitions for easy reference. Most of the discussion will start with the W-90 series of tables, which look at the Bottom 90%, Next 9% and Top 1% of the Networth distribution.

The main data source for these tables is Currents and Undercurrents: Changes in the Distribution of Wealth, 1989–2004.

What is a Networth Distribution? Very simply, it is a picture of a group of households, which have been lined up according to their Net Worth, (that is, the difference between the sum of their assets and the total of their debts). Once they're lined up, one can look at various measures of the groups. First, one might look at the household in the exact center, and see what their Networth is. That is the median. One might look at the average, or mean Net Worth, which is generally a much higher figure, because it sums up all the Networth figures and divides by the total number of households.

One might choose to look at various groups within that Networth distribution. This website is working with several different quantiles. The 90-9-1 tables group households into three groups:

• the bottom 90% of us (that is, the 90% of households who have the least wealth per household),
• the next 9% and
• the top 1%.

The 50-40-5-4-1 tables look at more detail. For some purposes, we can also look at 25-25-25-15-10 data, which may provide some additional texture to our analysis. We'll look first at the 90-9-1 tables. There are 9 tables in each set. [Hint: the most interesting material is in the last ones, but you may want to understand what leads up to them well, first!]

Table 1, Dollar Holdings and change, by Networth Quantile — 1989 and 2004, shows the total, or aggregate, dollar holdings of the various classifications of assets and debts in 1989 and 2004. Line 01 is NETWORTH, which is the difference between ASSET and DEBT. Total NETWORTH rose from $25,854 billion in 1989 to $50,251 billion in 2004. (Just for reference, the number of households increased by 20.5% during that period, from 93.0 million to 112.1 million.)

Of that increase, the Bottom 90% [hereafter B90] of us received $6,847 billion, the Next 9% [hereafter N9] received $8,551 billion and the Top 1% [hereafter T1] received $8,999 billion.

Obviously, not all of us were in the same group in 2004 that we were in in 1989; some people who were alive in 1989 are dead now, and many who were in their parents' households in 1989 are in their own households now. Some have moved into a higher quantile, and some may have moved into a lower quantile. (Having said that, if everyone who was in the T1 and N9 quantiles in 1989 either died or dropped into the B90, only 1 out of every 90 in the B90 quantile would have moved into the T1 in 2004, and 9 out of the 90 into the N9! And that doesn't take into account the possibility that immigrants might be in either of the upper quantiles.)

The thresholds, in 2004 dollars, were, for the 90th percentile, NETWORTH of $827,600 and for the 99th percentile, $6,006,000. (For 1989, expressed in 2004 dollars, the corresponding figures were $519,300 and $3,345,400.)

Also on Line 01, one might note that among the three quantiles, in 2004 the N9 had the most NETWORTH and the B90 had the least (columns 6, 7 and 8); in 1989, the T1 had less than the B90 (columns 2, 3 and 4).

When we move to ASSET (line 02), which is the sum of FIN (financial assets) and NFIN (nonfinancial assets), but doesn't take into account DEBT (liabilities), we see (column 10) that the B90 of us had more of an increase than the N9 or T1. We won't discuss shares here; that data is on a later table.

Line 03 is FIN, financial assets. It is the sum of liquid assets [LIQ], CDS, Savings Bonds [SAVBND], BOND, STOCKS, non-money market mutual funds [NMMF], retirement assets [RETQLIQ], cash value life insurance [CASHLI], other managed assets [OTHMA] and other financial assets [OTHFIN]. For 2004, the B90 quantile has about 10% less than the T1 group, and showed the least increase from 1989.

Skimming the financial assets section of Table 1, lines 03-13, one notes that the big numbers are in RETQLIQ (retirement assets, such as IRAs and 401(k)s), STOCKS and NMMF. Retirement assets showed the largest increase ($4.8 trillion — column 9, line 10). Interestingly, RETQLIQ is one of the categories where the B90 group has significantly more than the T1, though still less than N9.

It is important to keep in mind here that RETQLIQ is gradually becoming the replacement for Pension wealth, which is not calculated in this study. As of 1992, among households whose head was under age 65, Pension Wealth in aggregate was equal to 48% of Aggregate Assets and 60% of Aggregate Networth, so that is a lot to replace! [Source: Pensions, Social Security and the Distribution of Wealth, Table 4.] Further, Pension Wealth was far more evenly distributed than even gross housing wealth in 1992. One might also want to consider Social Security Wealth, which, for most of us is more than the rest of our wealth! [source:

The other category of assets is the nonfinancial assets, NFIN, at line 14. In total, this category is nearly twice as large as FIN (column 5). The B90 group has more than either the N9 or the T1, though quite a bit less than 50% of the total. The increase from 1989 to 2004 is distributed very differently from the FIN, too, with B90 reaping the most increase and N9 the least (columns 10, 11 and 12).

A quick scan of the cells in the Nonfinancial Assets section (lines 14 to 20) reveals that the single largest number is the B90's holdings of HOUSES, at $11.8 trillion. T1's holdings of BUS, business equity, at $6.1 trillion, is the second largest piece of non-financial wealth.

The final component of NETWORTH is DEBT. The single largest figure in this section is the B90 mortgage holdings, at $5.1 billion (line 22, column 6). This figure compares to the gross HOUSES value of $11.8 billion for that quantile. INSTALL — installment loans — is the second largest category.

Line 28 provides data on Equity, which is includes Stocks held directly as individual shares and as part of other categories, including NMMF and RETQLIQ. (?)

The final line that comes directly from the Currents and Undercurrents study is INCOME. Income is more widely distributed than is NETWORTH or any of its components.

The remaining lines in Table 1 are aggregations of individual lines from the Currents and Undercurrents study. Each represents a sort of hypothesis about which Assets are "consumption assets" and which are "power assets" whose purpose is not consumption but savings toward retirement or wealth generation.

Lines 31 and 32 show the combinations of STOCKS and NMMF and then add to them RETQLIQ, with the thought that individuals might be most likely to have these kinds of holdings.

Lines 33 and 34 show the combinations of STOCKS and BUS (thus, shares of publicly held companies and equity in privately held companies) and then add in NMMF.

Lines 35 through 39 add in more asset classes: 35 adds NNRESRE, the value of nonresidential real estate holdings. 36 adds ORES, other residential real estate. 37 adds OTHMA, other managed assets. Line 38 adds BONDS. Line 39 adds RETQLIQ.

Line 40 combines the two most widely held "consumption" assets, VEHIC and HOUSES. Line 41 nets out of VEHIC and HOUSES the two kinds of debt held against them, MRTHEL and INSTALL. B90 has over half of this category!

Line 42 adds retirement assets to VEHIC and HOUSES, and Line 43 nets out debt as Line 41 does. The Bottom 90% of us have a bit over half of this category, which represents about 40% of NETWORTH (column 5)

Line 44 provides Home Equity, the difference between HOUSES and MRTHEL. This represents about 25% of household NETWORTH in 2004.

Lines 45 and 46 look at the remainder of the NFIN and FIN categories that aren't represented in line 42.

Line 47 shows all Networth not included in line 40; Line 48 shows all Networth not included in line 42; the difference between the two figures is RETQLIQ, retirement assets.

Lines 49 through 55 were created to take advantage of some 2004 data provided in a later table [#18] in the original report; they will be explored in later tables.

Lines 56 and 57 are "catchall" categories for Assets other than those in lines 38 and 42, and for Debts other than those in lines 22 and 24.

Table 2: Corner Percentages

This table provides a quick way to see the relative importance of the various classifications of wealth (columns 1 and 5).

Line 1 shows that the Bottom 90% of us own 30.50% of the NETWORTH in 2004, down from 32.79% in 1989. The share of the Next 9% of us has decreased from 37.13% to 36.12%. The share of the Top 1% has increased from 30.08% to 33.38%.

(A detour to the corresponding 50-40-5-4-1 table shows us that within the N9 group, the share of the upper 4% has remained almost constant, and the share of the lower 5% has decreased by over 1% from 1989 to 2004. So if we looked at the Bottom 95% versus the top 5%, we'd see that, in 2004, the Bottom 95% of us have 42.49% of the aggregate NETWORTH, down from 45.84% in 1989. If we had access to more detailed information, we might find that the 96th and 97th percentiles had lost share.)

Line 2, ASSET, needs to be looked at carefully; it is not quite what it might seem. Columns 1 and 5 show that DEBT as a percent of NETWORTH rose from 13.8% to 17.6% between 1989 and 2004. Comparing Lines 01 and 02 shows that the largest effect of debt is in the B90 group.

Lin 03, FIN, has increased in importance from 35.1% to 42.0% of total NETWORTH, but virtually all of the increase has gone to the N9 and T1 quantiles. The N9 quantile has of the FIN category than either the B90 or T1 quantiles.

Several categories of financial assets have diminished in overall importance: LIQ, CDS, SAVBND, BOND, CASHLI and OTHFIN; most of the decrease has been in the B90 group.

Other categories have increased: STOCKS, NMMF, RETQLIQ. The largest increases show up in RETQLIQ, particularly in the N9 quantile, and NMMF and STOCKS in the T1 quantile.

Nonfinancial Assets (Line 14) have decreased in overall importance, except in the T1 quantile (columns 4 and 8).

HOUSES have increased from 35.8% to 38.0% of aggregate NETWORTH. The largest share of the increase has gone to the T1 quantile.

(Readers who have spent much time exploring the wealthandwant.com website might not be surprised at this: the homes of the wealthy tend to be on far more valuable than the land on which middle class people live, and those sites appreciate much faster. Check out The Taxable Capacity of Land and LSREV if you're interested.)

Interestingly, NNRESRE has gone down in importance (line 18). One might hypothesize that this might be a function of the increase of REITs (Real Estate Investment Trusts) replacing ownership by individual families. Stock in these REITS might be held as STOCK, or in NMMF, or RETQLIQ, or possibly be included in BUS.

As noted above, in the difference between ASSETS and NETWORTH, DEBT has increased in all quantiles. Mortgage debt on primary residences represents most of this, and has risen from 9.5% to 13.3% of NETWORTH.

Home equity [Line 44, HOUSES minus MRTHEL] has dropped from 26.3% to 24.8% of aggregate NETWORTH. B90's share has dropped, N9 has remained steady, and T1's share has increased by nearly 50%. This is quite different from the experience for HOUSES. (above)

Table 3: Shares

Table 3 shows the shares each quantile held of each asset class in 1989 and 2005. Line 01 of Table 3 matches Line 01 of Table 2.

Comparing lines 01 [NETWORTH] and 02 [ASSET], one notes that the higher debt levels in the B90 group reduce their 36.8% share of ASSETS to 30.5% of NETWORTH (line 21 shows that the B90 quantile carries 72.8% of the DEBT).

Columns 10, 11 and 12 show the percentage point changes in shares. For Networth, the T1 quantile gained 3.3 points, with 2.3 of that coming from the B90 and 1.0 coming from the N9 quantile.

Scanning down column 8, one notices that BOND and STOCKS are the most concentrated FIN assets, and BUS the most concentrated NFIN asset. Scanning column 6, one notes that VEHIC and HOUSES are the only two asset categories where as much as 60% of the value is held by the B90 quantile. Continuing in column 6, one notes that 76% of mortgage debt is held by the B90, 86% of installment debt and 93% of installment debt.

EQUITY, line 28, is rather concentrated in the top 1%, with 36.8% of the value, but as one scans down the remaining rows, it is far less concentrated than many of the other combinations of assets.

INCOME, line 29, is far less concentrated than NETWORTH. The B90 Wealth Quantile received 63.8% of the income. The N9 received 22.6% and the T1 13.6%.

When STOCKS and NMMF are combined, the B90 quantile has only 14.5% of the value (line 31); adding RETQLIQ (line 32) brings that up to 28.0%.

By the time we get to Line 39, which incorporates most of the wealth-building assets, we find that 19.7% of value belongs to the B90 quantile, 37.2% to the N9 quantile, and 43.1% to the T1 quantile.

At Line 40, VEHIC+HOUSES, we find 63.2% of the value in the hands of the B90 quantile. Line 41 nets out the debt against these two asset classes, leaving 55.1% of the value in the B90 quantile. One might contrast these figures to Line 39.

Line 42 adds RETQLIQ to line 40, and line 44 does the same with line 41. These two lines might be best compared with Line 38 (wealth-building assets other than RETQLIQ).

Line 44 shows home equity, HOUSES minus MRTHEL. 54% of this value is in the hands of the B90 quantile, down from 60.7% in 1989. Most of the share gain has gone to the T1.

Table 4: Changes from 1989 to 2004

Table 4 looks at what has happened between 1989 and 2004: the absolute change, the percentage change and the share of change for each quantile. A quick scan of columns 6, 7 and 8 shows that the T1 quantile showed the fastest gains, and the B90 generally the slowest. T1 got 37% of the NETWORTH gain, and B90 only 28%.

The main exceptions to this were in RETQLIQ (line 10), where N9 picked up 46% of the gain and B90 41% (columns 10 and 11).

HOUSES (line 16) showed the highest percentage gains for the T1 and the lowest for the B90 (columns 6, 7 and 8).

NNRESRE showed larger gains for the B90 quantile than for the N9, which were in turn larger than T1's (line 18).

Interestingly, holdings of EQUITY, line 28, gained at roughly the same rate for all three quantiles.

Line 43, home and car equity plus retirement assets, showed faster gains in the upper quantiles.

Line 44, HOUSES minus MRTHEL increased 62% for the B90, 94% for the N9 and 190% for the T1, again consistent with what readers of this site might expect.

Table 5: Ownership Rates

Table 5, Ownership Rates, begins to get at some of the underlying dynamics. It is the first of two tables on Ownership.

Table 5 reports the data as it shows up in the Currents and Undercurrents report. Some categories of assets have declined in ownership, most notably CDS, line 05, (from 19.9% to 12.7%, with much of the loss falling in the B90 quantile; SAVBND, BONDS and CASHLI.

Financial Assets showing increases in ownership rates include STOCKS (from 16.9% to 20.7%); NMMF (from 7.3% to 15.0%); RETQLIQ (from 37.0% to 49.7%).

Nonfinancial Assets showed some increases in ownership: VEHIC increased in the highest and lowest quantiles; HOUSES increased in all three quantiles -- and most in the top quantile. Ownership of ORESRE and NNRESRE fell somewhat.

Ownership of BUS equity fell slightly in the upper quantiles and rose somewhat in the B90 quantile.

DEBT ownership increased in all quantiles, but fastest in the B90. Mortgage debt (MRTHEL, line 22) increased from 39.5% to 47.9%; homeownership increased from 63.9% to 69.1%, so the percentage of households owning their homes free and clear declined from 24.4% to 21.2%. (Given the aging of the Baby Boom generation, this is striking.)

The use of INSTALL (line 24) fell somewhat in total and in the lower two quantiles, but rose somewhat in the top quantile.

The percentage of households carrying CCBAL rose from 39.7% to 46.2%, and in the B90 from 41.9% to 48.8%.

The percentage of households with EQUITY (line 28) rose from 31.8% to 48.6%.

A footnote to Currents and Undercurrents provides another snippet of information reported in Table 5: that the net ownership of NNRESRE and BUS in 2004 is 17.5%. This compares to 11.5% ownership of BUS and 8.3% ownership of NNRESRE, and suggests that 2.3% own both BUS and NNRESRE.

Table 6: Additive Ownership Rates

Table 6 is calculated from Table 5. Each cell in Table 5 is multiplied by the size of its quantile -- 100%, 90%, 9% or 1%. The result is ownership rates that add up to the total ownership rate. These are useful for seeing where the increases and decreases in ownership rates occurred. In general, the increases in ownership have come mostly from the B90; not surprising, since there is little room to increase in quantiles where ownership is already high and where even a large increase within the quantile only represents a small increase at the total level.

For example, in RETQLIQ (line 10), where ownership increased from 37.0% to 49.7%, we see that most of that increase came from increases in the B90 quantile (from 30.0% to 41.4%).

Similarly, ownership of HOUSES (line 16) increased from 63.9% to 69.1%, and in the B90 quantile, ownership increased from 54.6% to 59.4%, for the B90.

Table 7: Holdings Per 1% holding

Table 7 looks at the other half of the ownership equation: how much each "1% of households who own" owns, on average.

One caution goes with this data: keep in mind that from 1989, the number of households increased by 20.5%, so each 1% of 2004's households is a larger group than 1% of 1989's households. But within each year, comparisons up and down the columns and across quantiles are valid.

This table was calculated by dividing the dollar data in Table 1 by the (additive) ownership rate from Table 6. (If you are familiar with reach/frequency calculations from the media-buying world, this corresponds to frequency.)

(Note that the column does not and should not add to the total assets figure.)

Line 2, ASSET, shows this well. On average, in 2004, each 1% of the those with assets (97.9%, per Table 5, line 2) have $604 billion in assets.

In the B90 quantile, average holdings per 1% who have assets (97.7% of the quantile) are much lower than the national average, at $248 billion. In the N9 quantile, average holdings are $2,212b, or about 9 times the B90. In the T1 quantile, average holdings are $17,425b, or 70 times the average holdings in the B90.

Looking at a few individual lines can be instructive. Retirement assets [RETQLIQ] are a good example. Average holdings per 1% who hold are $136b. The average 1% holding in the B90 quantile is about half that, at $68b. The T1 quantile's average holdings-per-1%-holding are 1,109, or 16 times the B90 group. The N9 quantile's holders average 6 times the B90 group.

HOUSES (line 16) holdings are interesting, too. The average holdings in the N9 quantile are about triple the holdings in the B90, and in the T1 are about 10 times those in the B90.

Average holdings of BUS equity (line 19) in the T1 quantile are more than 10 times those of the average holdings in the N9 quantile, and 67 times what the average holder of BUS in the B90 quantile owns.

HOUSES minus MRTHEL shows that the average holdings of home equity among those in the T1 quantile are 14 times those of average holdings in the B90 quantile. (Check out the corresponding table in the 50-40-5-4-1 group for more detail.

This data helps answer the question that many people ask: is increasing homeownership the answer to helping poor people be less poor? After all, those who own homes have a lot more wealth than those who don't. The gross answer might take the form of multiplying a, say, 10 point increase in the B90 quantile's ownership rate by the $113b home equity per ownership point. A finer look would be to look at the average home equity holdings of the B50 group, from the 50-40-5-4-1 tables, which show that B50 (where 21.7 out of 50.0 own HOUSES) holds $33b per homeownership point, and the N40 (where 37.7 out of 40.0 own HOUSES), hold $159b per ownership point. So the increase would have to come in the B50 group, where average home equity per 1% ownership point for 2004 was $33b, well below the national average of $180b.

Table 8: Change in holdings per 1% holding

Table 8 is drawn from Table 7, and shows the change from 1989 to 2004 in holdings per 1% who hold (columns 1 through 4), and the percentage change (columns 5 through 8). As mentioned earlier, anything less than a 20.5% change represents a decrease in average holdings per household.

Line 44 shows that average holdings of home equity [HOUSES minus MRTHEL] rose the most among the T1 and the least among holders in the B90.

Table 9: Changes in ownership and holdings per 1% holding

Table 9 collects four columns from Table 5 and four columns from Table 9, to facilitate comparisons of the direction and size of changes of the incidence of ownership (Columns 1-4) and the amount each 1% who holds possesses.

It becomes pretty clear that most of the gains are coming in the holdings-per-holder, not in the percent of each quantile who holds. Despite the huge run-up in housing values we keep hearing about, the holdings-per-1% holding have risen most for the T1 and N9 quantiles, and at a slower rate for the B90 quantile, and ownership rose faster in the top group than in the bottom! Business equity has risen fastest in the top quantile. RETQLIQ, which will gradually become the replacement for Defined Benefit pensions, has risen fastest in the N9 group, and spread only slightly faster in the B90 group.

Use of DEBT has become more widespread in all groups, by faster in the B90 quantile. Home equity (line 44), which represents the largest share of the B90's networth, has risen fastest for the T1 group, both in incidence and in holdings per holder.

So far, we've lumped together a very large share of our society: the Bottom 90% contains households whose situations are very different from each other, and we've generally been looking at averages within that large group. The 50-40-5-4-1 tables break that 90% down into two component pieces, the Bottom 50% and the Next 40%. The "N9" quantile is also divided into two parts, 5% and 4%, which also look quite different from each other.

Finally, when one starts to look at fractiles (subsets of the top 1%), one finds that the top and bottom half of that top 1% look quite different from each other as well. While the data is older (1992 SCF), you might take a look at Tables 4 and 5 from Pensions, Social Security and the Distribution of Wealth (1997).

Also on this site, you'll find more quantile detail on ownership from the 2004 SCF in Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances (2006), where the quantiles include the bottom quartile (less than 25th percentile), the second quartile (25th to 49.9); the third quartile (50-74.9), the next 15% (75 to 89.9) and the top decile.