Poverty scorecards – a cheap way to identify who's poor?

Finding out which people in any given community live below the poverty line is actually quite hard. Why do it? To target services like microfinance  (let’s not get into the targetting v universal provision argument here); comparing poverty rates in different regions and countries, and tracking changes over time.

But both income and consumption poverty are hard to assess directly – poor people tend not to have wage slips or supermarket receipts handy to give to people with clipboards. In Viet Nam, Oxfam researchers tried a different approach, working with communities to identify different levels of poverty – who owned land, a buffalo, a tin roof, a motor bike etc, producing a real and nuanced picture of what poverty means in those villages. But that can be an expensive business – labour intensive household surveys, difficult problems gathering and interpreting reliable data etc etc.

But researchers at the microfinance.com website have come up with a much cheaper version, ‘poverty scorecards’ based on the World Bank’s regular and highly detailed ‘Living Standards Measurement Study’ (LSMS) surveys. The scorecards boil down a vast amount of data from the LSMS to identify easily detectable signs of poverty – the tin roof or buffalo test for each country.

Here’s an example from bdesh poverty scorecardBangladesh. Fill it in, and if a person gets 24 points, there’s an 80% chance probability that they are living on less than $1 a day. The website already crunches the numbers for a number of countries, in alphabetical order: Bangladesh, Bolivia, Ecuador, Ethopia, Haiti, India, Indonesia, Kenya, Malawi, Mali, Mexico, Morocco, Nepal, Nigeria, Pakistan, Palestine, Peru, Philippines, South Africa, Vietnam, Yemen. They plan to add more countries to the list.

One complaint, though – the actual scorecards are buried in the appendices for each paper, after you’ve waded through a bunch of maths. Any chance of putting them up separately, guys?

In any given place, such a scorecard could be improved by adapting it to local circumstances (rural poverty is different from urban, nomadic pastoralists from settled small farmers etc), but it looks like an excellent initial short cut – our monitoring and evaluation people are certainly intrigued. Has anyone other than microfinance people used them? Got any comments, eg about whether individual country examples work for your countries?

Subscribe to our Newsletter

You can unsubscribe at any time by clicking the link in the footer of our emails. For information about our privacy practices, please see our Privacy Policy.

We use MailChimp as our marketing platform. By subscribing, you acknowledge that your information will be transferred to MailChimp for processing. Learn more about MailChimp's privacy practices here.

Comments

4 Responses to “Poverty scorecards – a cheap way to identify who's poor?”
  1. Pushpanath K

    It is all good and great to find out who lives in poverty( I hate the word- the poor).We have done enough of this for the last three decades and created an army of consultants and so called experts( Mostly from the West if I may say so and not surprisingly).

    My advice is to find out who are wealthy and why and how they made it- is a better appraoch.

  2. James Stevenson

    I have a different view to Pushpanath (although agree that we should try and find out how the rich came to be so!). This innovation lowers the barrier in terms of “expertise” required to carry out monitoring of development programmes. These scorecards make best use of the masses of data that sits with organisations like the World Bank.

    In distilling the findings from the World Bank’s surveys into a very simple series of tools that we can all use, a civil society organisation of any size and technical “sophistication” can collect data in such a way that, at the end, they will have results that are credible and (crucially) likely to have influence with the powerful. This might be to evaluate their own programmes, or it could be used as a tool for holding organisations such as the World Bank (or Oxfam for that matter) to account for their performance. Either way, surely that is a helpful development…? I think this is an exciting innovation. What do others think?

  3. J

    I think targeting in CCT has used this type of “score cards” for a while; in CCT-speak, it’s called PMT = Proxy-Means Tests, which look at hard-to-hide assets instead of income (since the latter is often difficult to measure in informal economies). In Indonesia, where I work, the government has begun developing different PMTs for different CCT target areas.

    WB Indonesia and J-PAL are publishing a research soon on the merits of PMT vs. more labor and time intensive community-based methods. If you’re interested, I can keep you posted?

    P.S. Apologies for commenting on an old post. I’m catching up after a period of absence as well!

  4. Yes, the poverty scorecards have been used by non-microfinance organizations. In fact, the biggest potential user are health organizations, and a few have picked up the scorecard for this purpose. Nothing about the poverty scorecard is microfinance specific.

    Yes, as one commenter notes, the scorecard is similar to the three-decades-old Proxy Means Test. The difference is that poverty scorecards are simple and transparent and inexpensive enough that local pro-poor organizations can pick them up (for free) and use them for their own purposes, hopefully to improve their social performance. Most of the PMT are too complex, or opaque, or impractcal. Often, the formula is secret.

    This simplicity and inexpensiveness means that organizations can pick up the scorecard and use it, on their own, without help, for their own purposes. In Pakistan, for example, one large national network picked it up and applied it to >600,000 households, without any funder requiring it. Now, the World Bank has copied the basic approach to make their own poverty scorecard (essentially the same as mine) and will be applying it to most households in Pakistan in the next several years.

    Finally, I agree that it is too bad that the scorecard itself is buried in Figure 1 after 50 or 60 pages of text. I don’t particularly like the length and some of the complexity in the documentation, but I have found that if I omit it, someone asks for it, so it is easier just to cover all the FAQs up-front. I am considering putting the scorecard after the title page and before any text, since the scorecard will usually speak for itself (which is why I suppose you put a copy on your blog!)

    Incidentally, I come from a credit-scoring background, and it turns out that the academics and WB etc. people doing these scorecards for the past three decades not only have them too expensive and complex, but they are no more accurate and they fail to test their accuracy correctly (when they do test it, which is not often).

    Hope this helps

Leave a Reply

Your e-mail address will not be published. Required fields are marked *

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