Looking at the Consumption rate of Capital Income
Theoretically, capital income, especially in the form of retained earnings by corporations and some capital gains would be used for saving and investment in the means of production, while labor income is used for consumption of finished goods and services from production. However, a portion of capital income is used for consumption, when the incentives and extra liquidity are there.
What is the percentage of capital income that is used for consumption of finished goods and services? And is it helpful to know?
Here is the graph I put together showing the changes in the consumption rate of capital income since 1953…
The consumption rate of capital income reached 40% in the 1960s and then fell to almost 0% during the 1980 recession. Why the dramatic drop? And if we look to 2013, we see that the rate is making a comeback.
What could be going on here?
Well, as I inputted decades of data using dozens of sheets of paper, patterns appear. One clear pattern is the effect of tax rates. In the following graph I show the effective tax rates for capital income and for labor income. These are the effective tax rates used to determine the data in graph #1 along with the national income account numbers. (source for effective tax rates on capital income.) (source for national income account data.)
The main thing to look at in graph #2 is the spread between the effective tax rate for capital income (yellow line using left axis) and the effective tax rate for labor income (red line using left axis). A general rule is that a wider spread will decrease capital income’s consumption rate.
As the effective tax rate for capital income fell in the 1950s to 1960s, more capital income was used for consumption purposes, because we see the rate for capital consumption rise and reach a peak at 40% in 1965. (Capital income consumption is the blue line using the right axis.)
Then in 1967, the effective tax rate for capital income began to rise, while the effective tax rate for labor income began to fall. It became cheaper to use labor income for consumption. The result was a fall in capital income’s consumption rate. The fall continued all the way to almost 0% during the 1980 recession. There was a rise in the effective tax rate on capital income around 1978 which looks to have contributed to pushing capital income’s consumption rate down further. When the consumption rate hit its low, the effective tax rate for capital income started dropping. The “writing must have been on the wall” that the effective tax rate on capital income was too much. The result was that the consumption rate of capital income bounced back to almost 15% by 1985.
Since 2003, the consumption rate for capital income has been rising. We see that the effective tax rate on capital income has been falling since 2003 too. The effective tax rate on labor income has been falling too, but the spread has narrowed which makes using capital income for consumption marginally cheaper.
On another note: There is an interesting thing in graph #1… The recessions since 1970 (taking out the Volcker induced recession of 1981) were preceded by the consumption rate of capital income falling. The implication is that one can foresee a recession by watching this rate.
There is a reason why capital income’s consumption rate could foretell a recession.
Leading up to a recession, profit rates will peak and then back-off. As this happens, more capital income is retained by the corporation and other capital entities for their protection. The result is progressively less funds permitted for consumption.
It looks as though the consumption rate of capital income will decline for at least a year before a recession. Graph #1 shows that the consumption rate of capital income is still rising in 2nd quarter 2013, which means that a recession is not yet on the horizon.
Cross-posted at effectivedemand.typepad.com
Edward,
1) Specifically how are you calculating your consumption rates for capital income? That is, which BEA Tables, and which Lines?
2) The effective tax rate data only goes up through 2003. Where did you get the more recent data?
3) Where does the effective tax rate on labor income data come from?
PS I should probably save this question until I hear the responses to the other questions but, the BEA data almost certainly excludes capital gains income as it is not part of GDP, and yet Jane Gravelle’s effective tax rates on capital income include capital gains. Shouldn’t you be using the effective tax rate on capital income excluding capital gains instead, as this would be more relevant for comparisons with the BEA data?
Mark,
You are right. Capital gains is not included in the NIPA tables.
Bottom of page 5 at this link.
http://www.bea.gov/national/pdf/nipaguid.pdf
And Jane Gravelle does include taxes on capital gains.
“Column 3 reports estimates of the total rate on corporate investment, accounting for the deductibility of interest at the firm level and the taxation of interest, dividends, and capital gains at the individual level, as well as depreciation and subsidies.”
page 1 at this link…
http://assets.opencrs.com/rpts/RS21706_20040112.pdf
I did not make an adjustment for this. How much difference would it make in the overall numbers? If I take 1984 for example, and raise the effective tax rate on capital income from 33% to 35%, the effective tax rate on labor income drops from 17% to 16.5%. Even though it would be good to make an adjustment for capital gains, the effective tax rates won’t change a lot. The basic story still holds.
answer to question #1…
I input the national income numbers from table 1.1.6 (real GDP, 2009 dollars). Then I use the GDP deflator on line 26 of table 3.1 to convert net govt saving to 2009 dollars. Knowing Govt expenditure and govt net saving, I determine total net taxes received. Gross saving is determined by the circular flow model I posted before. I input the personal saving rate from table 2.1, line 35, which will then determine how much labor and capital are saving respectively. Then I change the effective tax rate of labor income until the effective tax rate on capital income matches Jane Gravelle’s data. At that point, the consumption rate of capital income is spit out.
It’s a matter of deduction. First determine total consumption, total net taxes and total saving. Then determine labor and capital net saving. Then determine effective tax rates for labor and capital. Then labor and capital consumption spits out from (disposable income – net saving) for labor and capital respectively.
question #2…
I used stories like these…
http://money.cnn.com/2013/07/01/news/economy/corporate-tax-rate/index.html
http://commons.wikimedia.org/wiki/File:US_Effective_Corporate_Tax_Rate_1947-2011_v2.jpg
The effective corporate firm-level tax rate (column 2 of Gravelle) is normally above the overall effective tax rate (column 6 of Gravelle). So I had to estimate an effective tax rate for capital since 2003 using the above sources. The low rate I used is 13% which seemed to be a conservative estimate. I estimate the effective tax rate on capital has risen currently to 15.5% using the second link above for wikimedia.
The effective tax rates take into account transfers. Since 2008, there have been many transfers which have lowered the effective rates.
What do you think about the numbers? For example, the low effective tax rate on labor was between 8.7% and 9.5 from 1stQ 2009 through 2012. Then my numbers say it jumped up to over 11% with the tax change of 2013.
question #3…
The effective tax rate on labor income is deduced from other variables. In my model, once the constraints on disposable income, consumption and saving are established for labor and capital, if I change the effective tax rate on labor income, the effective tax rate on capital income changes. So I get labor’s effective tax rate when capital’s effective tax rate matches Gravelle’s data. I never know what the effective tax rate on labor will be. I just get the number and plot it and see where it goes.
I would not say that the numbers are 100% accurate. But the model has a consistent logic which can use the data available and show trends over time.
In order to make graph #1 above, I had to input 10 variables of data in over 240 circular flow charts. One chart for each point of data. A whole chart for each point of capital income’s rate of consumption.
If you can tell me there is an easier way to do it, I would love to know.
There is another factor for why the consumption rate of capital income was so high back in the 50s and 60s. The govt expenditure was over 30% of GDP as opposed to currently below 20%. The govt was doing a lot more investment.
Private investment as a percentage of GDP was 12% back in the 60s, as opposed to 16% now. Gross private saving was enough back in the 60s for investment needs. The govt did investment from tax revenues. The result was that capital income had funds available for consumption. To me that is really interesting… that capital income consumed more in the face of high taxes. Capital saved 32% of its total income back in 1964 as opposed to 65% now. That is a result of higher tax rates on labor and capital and higher rates of consumption by capital income. And of course labor saved a lot more back then, which took the pressure off of capital to have to save more.
Edward,
I appreciate your anwers. But if calculating the consumption rate for capital income depends on knowing the effective tax rate for capital income then Jane Gravelle’s numbers will obviously not provide you with precise figures since capital gains income is wildly volatile and has traditionally been taxed at different rates from other forms of capital income. This is disappointing to me because I thought you may have found a method.
It’s still unclear to me why we must know the effective tax rates to calculate the consumption rates but evidently to understand that I would have to invest the time and effort to understand your circular flow diagram. If I have the time I may get back to you.
Mark,
You are an expert on data and using the correct numbers. After your comment yesterday, I went back to the data and looked at it. There was a number which I had to change. I was using net government saving, table 3.1, line 26. That was not the best line to use for net government borrowing and lending. So I switched it for line 38 on that table 3.1, the line for net lending or net borrowing. and then I re-ran the flow charts again to get the consumption rate for capital income.
I will post the new graph as an update to this post. The idea is to not give up, but to keep re-working.