Note on Adjusting Dollar Amount Variables for Inflation

All dollar amounts in the IPUMS are nominal dollars--that is, they are given as measured in the original census. However, inflation renders these dollar amounts not comparable: $1 in 2008, for example, is worth much less than $1 in 1939. The IPUMS variable CPI99 provides an easy way to adjust dollar amounts to constant dollars. This variable, constant within years, inflates or deflates dollar amounts to the amount they would have represented in 1999 (which corresponds to the 2000 census PUMS)--as shown in Table 1.

Table 1. Converting Dollar Amounts to Constant 1999 Dollars with CPI99
Data Year Income Year X CPI99 = 1999 dollars
1940 1939 dollars X 11.986 = 1999 dollars
1950 1949 dollars X 7.000 = 1999 dollars
1960 1959 dollars X 5.725 = 1999 dollars
1970 1969 dollars X 4.540 = 1999 dollars
1980 1979 dollars X 2.295 = 1999 dollars
1990 1989 dollars X 1.344 = 1999 dollars
2000 1999 dollars X 1.000 = 1999 dollars
2000 (ACS) 2000 dollars X 0.967 = 1999 dollars
2001 2001 dollars X 0.941 = 1999 dollars
2002 2002 dollars X 0.926 = 1999 dollars
2003 2003 dollars X 0.905 = 1999 dollars
2004 2004 dollars X 0.882 = 1999 dollars
2005 2005 dollars X 0.853 = 1999 dollars
2006 2006 dollars X 0.826 = 1999 dollars
2007 2007 dollars X 0.804 = 1999 dollars
2008 2008 dollars X 0.774 = 1999 dollars
2009 2009 dollars X 0.777 = 1999 dollars
2010 2010 dollars X 0.764 = 1999 dollars
2011 2011 dollars X 0.741 = 1999 dollars
2012 2012 dollars X 0.726 = 1999 dollars

The 1999 base year was chosen simply for convenience; adjusting to some other year would yield identical correlations, regression coefficients, inequality measures, and the like. However, after using CPI99 to convert nominal dollars to 1999 dollars, users who desire another base year for simple descriptive statistics can easily adjust 1999 dollars to 1939 dollars, 1949 dollars, or any other year by making use of the alternate factors provided in Table 2.

Table 2. Converting Constant 1999 Dollars to Constant Dollars For Another Base Year
1999 dollars X Alternate Factor = Income Year Data Year
1999 dollars X 0.083 = 1939 dollars 1940
1999 dollars X 0.143 = 1949 dollars 1950
1999 dollars X 0.175 = 1959 dollars 1960
1999 dollars X 0.220 = 1969 dollars 1970
1999 dollars X 0.436 = 1979 dollars 1980
1999 dollars X 0.744 = 1989 dollars 1990
1999 dollars X 1.000 = 1999 dollars 2000
1999 dollars X 1.034 = 2000 dollars 2000 (ACS)
1999 dollars X 1.063 = 2001 dollars 2001
1999 dollars X 1.080 = 2002 dollars 2002
1999 dollars X 1.104 = 2003 dollars 2003
1999 dollars X 1.134 = 2004 dollars 2004
1999 dollars X 1.172 = 2005 dollars 2005
1999 dollars X 1.210 = 2006 dollars 2006
1999 dollars X 1.244 = 2007 dollars 2007
1999 dollars X 1.292 = 2008 dollars 2008
1999 dollars X 1.287 = 2009 dollars 2009
1999 dollars X 1.309 = 2010 dollars 2010

For instance, consider the following dataset, where Inc99 = Income X CPI99:

Case Year Income CPI99 Inc99
1 1940

$11,000

11.986 $131,846
2 1940 $12,000 11.986 $143,832
3 1940 $14,000 11.986 $167,804
4 1980 $66,000 2.295 $151,470
5 1980 $72,000 2.295 $165,240
6 1980 $84,000 2.295 $192,780

Expressed in 1999 dollars, the mean income in the 1940 data is $147,827, and the mean income in the 1980 data is $169,830, for a 14.9 percent increase ( (169,830 - 147,827)/147,827 = 0.149). But users may want to express mean incomes in terms of the final year of data (1980, or 1979 dollars). To do this, we find the alternate factor from the list above to convert 1999 dollars into 1979 dollars, which is 0.436. Then we can add a new variable for 1979 dollars (Inc79 = Inc99 X 0.436):

Case Year Income CPI99 Inc99 Inc79
1 1940

$11,000

11.986 $131,846 $57,585
2 1940 $12,000 11.986 $143,832 $62,711
3 1940 $14,000 11.986 $167,804 $73,163
4 1980 $66,000 2.295 $151,470 $66,041
5 1980 $72,000 2.295 $165,240 $72,045
6 1980 $84,000 2.295 $192,780 $84,052

Expressed in 1979 dollars, the mean income in the 1940 data is $64,453, and the mean income in the 1980 data is $74,046. There is still a 14.9 percent increase ( (74,046 - 64,453) / 64,453 = 0.149), but displaying the data this way may make more sense for certain purposes.