Limitations of income inequality statistics and India’s BPL census

Thomas Garrett of St Louis Fed serves a nice reminder on limitations of calculating inequality.

Income quintiles from different years are often compared to demonstrate the growing income inequality over time between, say, the poorest 20 percent of households and the wealthiest 5 percent of households. For example, in 1970 the income of the wealthiest 5 percent of households was 6.3 times greater than the income of the poorest 20 percent of households, whereas in 2007 the income of the wealthiest 5 percent of households was 8.8 times greater than the income of the poorest 20 percent of households. When such comparisons are made, it is implicitly assumed that each quintile contains the same households over time. 

For most people, income increases over time as they move from a usually low-paying first job to better-paying jobs later in life. Some others, however, may lose income over time due to business cycle contractions, demotions, career changes, retirement, and so on. Because incomes are not constant over time, the same households do not necessarily remain in the same income quintiles. Thus, comparing income quintiles from different years is a proverbial apples-to-oranges comparison because the households compared are at different stages in their earnings profile.

Hmm..This never really occurred to me…

He points to a study which says:

As shown in the upper panel of the chart, nearly 58 percent of the households in the lowest income quintile (lowest 20
percent) in 1996 moved to a higher income quintile by 2005. Similarly, nearly 50 percent of the households in the
second-lowest quintile in 1996 moved to a higher income quintile by 2005. Even a significant number of households
in the third and fourth income quintiles in 1996 moved to a higher quintile by 2005. The study also documented declines in household income. Of particular note are the findings regarding the wealthiest households. As shown in the lower panel of the chart, more than 57 percent of the wealthiest 1 percent of households in 1996 fell out of that category by 2005. Similarly, more than 45 percent of the wealthiest 5 percent of households in 1996 fell out of that category by 2005.

This implies some check on policies:

Over time a significant number of households move to higher or lower positions along the income distribution. As a result, public policies such as income taxation and income redistribution affect “classes” of people differently over each person’s lifetime. In addition, income mobility muddies the picture of income inequality derived from a simple comparison of income quintiles from different years because such comparisons incorrectly implicitly assume that each income quintile contains the same households over time.


Well it does change the redistribution bit and identifying right people for the benefits. It should not happen that people who have moved to higher income quintile continue to get the benefits and new ones in the low income quintile are excluded. But overall, one can still look at the inequality figures and see whether things have changed or not.

So identification is crucial in the lowest income segment. In this regard, it is interesting to note the recent development of measuring and identifying poor in India. New BPL census has  started:

The Union Cabinet today gave approval for conducting the Below Poverty Line (BPL) Census in rural and urban areas. The Census would be conducted by the Ministry of Rural Development in association with the Ministry of Housing and Urban Poverty Alleviation (HUPA) and the Registrar General of India (RGI). The BPL Census would pave the way to identify the households living below the poverty line in rural and urban areas of the country. The entire process will be completed by December 2011. The results of the BPL Census would be utilized in the 12th Five Year Plan (2012-13 to 2016-17). The enumeration of castes will also be done simultaneously along with the BPL census. 

Mint has some nice update on how this would be done. The identification of urban poor is being carried out for the first time as earlier only the rural population was enumerated to help them avail targeted BPL benefits.

However, true to Indian policy there are major loopholes as always. PlanCom recently released new poverty estimates based on Tendulkar report. It says these are official figures and willa ct as ceiling for govt. programs. If BPL census shows higher numbers, be it.

Yet, the government has indicated that beneficiaries of schemes targeted at BPL households would be those identified as poor by the Planning Commission.

This means that some poor, as identified in the census, will be excluded from the purview of such schemes.

The cabinet also approved a ranking formula based on deprivation criteria to ascertain who among the poor will be excluded.

Some of the state governments were quick to condemn the move of the Congress-led United Progressive Alliance (UPA).

Warning of food riots, Sushil Modi, deputy chief minister and finance minister of Bihar, told Mint: “Planning Commission in consultation with the rural development ministry capping the number and state governments being asked to do the BPL survey is contradictory. The BJP (Bharatiya Janata Party)-ruled states have decided we would not conduct surveys if they cap the numbers. Because last time the Plan panel’s estimate was 6.5 million households and the (Bihar) state’s survey had more than one crore.”

What is the purpose of the census then? I would think govt would end up accepting the BPL census numbers in the end. But why adopt such policies in first place.

Mint edit nails the error on its head:

On the list of this government’s most serious lapses in policymaking, this one should rank somewhere near the top. The below poverty line census, due to its methodology approved on Thursday, will identify more poor households than officially estimated. But, because the government has also decided that the official estimates will serve as a ceiling, the additional poor families will not benefit from poverty alleviation schemes.

In other words, the government will pick and choose who is to be officially poor and who isn’t.

India has always had trouble with the ways and means of defining poverty and identifying the poor. But Thursday’s decision says something about the government’s confusion over the political handling of poverty. It is almost as if the aam aadmi’s government was telling some of its votaries: “You may be poor, but you ain’t poor enough.”

Amazing stuff really..

One Response to “Limitations of income inequality statistics and India’s BPL census”

  1. pravin Says:

    great when people say that the rich have gotten richer over the last 30 years,it makes no real sense.those who are rich today- the google millionaires or the infosys bigshots,the tendulkars and IPL magnates -none of them were rich 30 yrs back.
    the “rich” is not a static group.indeed.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: