UStaxWords: Number of Words in US Tax Law
In Ecdat: Data Sets for Econometrics

Description Usage Format Details Author(s) Source References Examples

Thousands of words in US tax law for 1995 to 2015 in 10 year intervals. This includes income taxes and all taxes in the code itself (written by congress) and regulations (written by government administrators). For 2015 only "EntireTaxCodeAndRegs" is given; for other years, this number is broken down by income tax vs. other taxes and code vs. regulations.

1	data(UStaxWords)

A data.frame containing:

year: tax year
IncomeTaxCode: number of words in thousands in the US income tax code
otherTaxCode: number of words in thousands in US tax code other than income tax
EntireTaxCode: number of words in thousands in the US tax code
IncomeTaxRegulations: number of words in thousands in US income tax regulations
otherTaxRegulations: number of words in thousands in US tax regulations other than income tax
IncomeTaxCodeAndRegs: number of words in thousands in both the code and regulations for the US income tax
otherTaxCodeAndRegs: number of wrds in thousands in both code and regulations for US taxes apart from income taxes.
EntireTaxCodeAndRegs: number of words in thousands in US tax code and regulations

Thousands of words in the US tax code and federal tax regulations, 1955-2015. This is based on data from the Tax Foundation (taxfoundation.org), adjusted to eliminate an obvious questionable observation in otherTaxRegulations for 1965. The numbers of words in otherTaxRegulations was not reported directly by the Tax Foundation but is easily computed as the difference between their Income and Entire tax numbers. This series shows the numbers falling by 48 percent between 1965 and 1975 and by 1.5 percent between 1995 and 2005. These are the only declines seen in these numbers and seem inconsistent with the common concern (expressed e.g., in Moody, Warcholik and Hodge, 2005) about the difficulties of simplifying any governmental program, because vested interest appear to defend almost anything. Lessig (2011) notes that virtually all provisions of US law that favor certain segments of society are set to expire after a modest number of years. These sunset provisions provide recurring opportunities for incumbent politicians to extort campaign contributions from those same segments to ensure the continuation of the favorable treatment.

The decline of 48 percent in otherTaxRegulations seems more curious for two additional reasons: First, it was preceded by a tripling of otherTaxRegulations between 1955 and 1965. Second, it was NOT accompanied by any comparable behavior of otherTaxCode. Instead, the latter grew each decade by between 17 and 53 percent, similar to but slower than the growth in IncomeTaxCode and IncomeTaxRegulations.

Accordingly, otherTaxRegulations for 1965 is replaced by the average of the numbers for 1955 and 1975, and EntireTaxRegulations for 1965 is comparably adjusted. This replaces (1322, 2960) for those two variables for 1965 with (565, 2203). In addition, otherTaxCodeAndRegs and EntireTaxCodeAndRegulations are also changed from (1626, 3507) to (870, 2751).

Independent of whether this adjustment is correct or not, it's clear that there have been roughly 3 words of regulations for each word in the tax code. Most of these are income tax regulations, which have recently contained 4.5 words for every word in code. The income tax code currently includes roughly 50 percent more words than other tax code.

Spencer Graves

Tax Foundation: Number of Words in Internal Revenue Code and Federal Tax Regulations, 1955-2005 Scott Greenberg, "Federal Tax Laws and Regulations are Now Over 10 Million Words Long", October 08, 2015

J. Scott Moody, Wendy P. Warcholik, and Scott A. Hodge (2005) "The Rising Cost of Complying with the Federal Income Tax", The Tax Foundation Special Report No. 138.

data(UStaxWords)
plot(EntireTaxCodeAndRegs/1000 ~ year, UStaxWords, 
  type='b',
  ylab='Millions of words in US tax code & regs')

# Write to a file for Wikimedia Commons
## Not run: 
svg('UStaxWords.svg')

## End(Not run)
matplot(UStaxWords$year, UStaxWords[c(2:3, 5:6)]/1000,
    type='b', bty='n', ylab='',
    ylim=c(0, max(UStaxWords$EntireTaxCodeAndRegs)/1000),
    las=1, xlab="", cex.axis=2)
lines(EntireTaxCodeAndRegs/1000~year, UStaxWords, lwd=2)
## Not run: 
dev.off()

## End(Not run)
# lines 1:4 = IncomeTaxCode, otherTaxCode, 
#   IncomeTaxRegulations,
#   and otherTaxRegulations, respectively

##
## Plotting the original numbers 
##      without the adjustment
##
UStax. <- UStaxWords
UStax.[2,c(6:7, 9:10)] <- c(1322, 2960, 1626, 3507)
matplot(UStax.$year, UStax.[c(2:3, 5:6)]/1000,
      type='b', bty='n', ylab='',
      ylim=c(0, max(
          UStax.$EntireTaxCodeAndRegs)/1000),
      las=1, xlab="", cex.axis=2)
lines(EntireTaxCodeAndRegs/1000~year, UStax., 
        lwd=2)
# Note especially the anomalous behaviour of 
# line 4 = otherTaxRegulations.  As noted with
# "details" above, otherTaxRegulations could have
# tripled between 1955 and 1965, then fallen by 48
# percent between 1965 and 1975.  However, that
# does not seem credible, especially since there
# was no corresponding behavior in otherTaxCode.

##
## linear trend 
##
(newWdsPerYr <- lm(EntireTaxCodeAndRegs~year, 
    UStaxWords))
plot(UStaxWords$year, resid(newWdsPerYr))
# Roughly 150,000 additional words added each year
# since 1955.  
# No indication of nonlinearity.  
# adusted R-squared exceeds 99 percent.  

##
## linear trend with increased slope
## during the Reagan years
##
# linear spline with knots at
# 1981 and 1989 
Reagan <- pmax(0, pmin(
  (UStaxWords$year-1981)/8, 1))
plot(Reagan~year, UStaxWords, type='b')
UStaxWords$Reagan <- Reagan

ReaganMdl <- 
  EntireTaxCodeAndRegs~year + Reagan
fitReagan <- lm(ReaganMdl, UStaxWords )
summary(fitReagan)