Nation of Beancounters

Did poverty reduction decelerate after the reforms? Not quite. (wonkish)

Posted in Uncategorized by Navin Kumar on December 29, 2011

Vamsi Vakulabharanam and Sripad Motiram seem to think so (gated):

As we can observe from Table 5.1, poverty continues to be high, but after the economic reforms it has declined in both rural and urban areas, both at the all-India level and for most of the major states. However, this is not a new phenomenon – poverty had been declining since the 1970s after increasing earlier (Dreze and Sen 2002, Table A6). So, what is relevant for our purposes is how the rate of poverty reduction during the period of economic reforms compares to the same in the previous period. We can compute the rate of poverty reduction in a simple manner, by dividing the change in HCR during a given period by the length of the period. Using this method, we can see that the rates of rural (urban) poverty reduction for the periods 1983 to 1993-4 and 1993-4 to2004-5 are 0.95 per cent (0.98 per cent) and 0.81 per cent (0.64 per cent), respectively. We can note that both in rural and urban ares, the rate of of poverty reduction was lower in the latter period

Imagine two countries, one with a 40% poverty rate ( what we call a Head Count Ratio) and another with a 20% rate. A year later, the rates are are 30% and 10% respectively. By Vakulabharam and Motiram’s method, both have achieved the same reduction in HCR (10% each). However, country 1 has reduced the poverty rate by a fourth and country 2 has halved the poverty rate. Surely, the latter is far more impressive. Assuming that population remains constant, half the poor in country 2 have been lifted out.

I redid the calculations (.xlsx file). Here’s the table for the data:

1983

1993-4

2004-5

Rural HCR

46.7

37.2

28.3

Urban HCR

42.4

32.6

25.6

Now, as  you can see, rural poverty declined from 47% in 1983 to 37% in 1993, a fall of 9.5 percentage points or near .95 percentage point per year. In contrast, for the post reform period, the decline is only .89. That’s bad, right?

Not quite: 9.5/47 gives us 20% – poverty declined by a fifth. In contrast, we get a 24% decline in poverty in the latter period, or roughly .022% per year in contrast with .020% in pre-reform period. The table below gives my calculations:

HCR

Rural

Urban

1983-93

1993-2004

Change

1983-93

1993-2004

Change

dHCR

9.5

8.9

-0.60

9.8

7

-2.80

n

10

11

10

11

dHCR/n

0.95

0.81

-0.14

0.98

0.64

-0.34

HCR

46.7

37.2

42.4

32.6

(dHCR/HCR)

0.20

0.24

0.23

0.21

ABOVE/N

0.020

0.022

0.001

0.023

0.020

-0.004

So with these three data points, we see that while urban poverty reduction did in fact decelerate during the post reform period, the rural poverty rate accelerated: a larger fraction of poor people were being lifted out of poverty in the post reform period in rural areas.

A few points are in order:

1. I redid the calculations using the end-of-period HCR and the Mid-Point (don’t ask why, it just seemed like a good idea). Results are consistent.

2. I redid the calculations with the Natural Log of HCR. Results are, again, consistent.

3. I redid the exercise with the Poverty Gap Ratio. It shows a deceleration of both urban and rural poverty reduction in the post reform period. However, I have little knowledge of or experience handling PGR data points, so I will refrain from drawing any conclusions.

Bonus: Sanjoy Mahajan explains why you should always use logs and how MIT’s Justin Wolfers used logs to resolve the Easterlin Paradox. I’ve used simple proportional changes here to avoid explaining the superiority of log conversions.

Advertisements

2 Responses

Subscribe to comments with RSS.

  1. Sneha said, on December 30, 2011 at 12:02 pm

    Interesting. I am rather curious to find out the reason for such an effect in rural areas. Any explanation? And I am surprised that HCR is taken as the basis of measurement.

    • Navin Kumar said, on December 30, 2011 at 3:59 pm

      The reforms worked? But that doesn’t explain the deceleration in Urban areas. Motiram etc suggest rural distress as a proximate reason – but rural poverty reduction has accelerated while urban poverty has decelerated.

      And the inconsistency with PGR :p … Maybe all the poor rural people are moving to Urban areas? Only hazy reason I can think of.


Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s

%d bloggers like this: