Indicator alphabet soup: GDP, HDI, MPI, GPI

Discussions in a couple of my classes recently have reminded me of an old management dictum: You get what you measure.

The idea is that the process of measuring something focuses our attention on it, and we’ll tend to seek ways to improve that measure. If the measure isn’t quite what we actually want, then the result might not be quite what we want. As an analogy: watching your weight closely may help you lose pounds, but a truly healthy diet requires more than just weight loss.

Same goes for GDP. Sure, economic output is generally good, but a focus on GDP could encourage environmentally unsustainable production (e.g. clear cutting old growth forests boosts GDP, but that only lasts so long). Also, economic output isn’t the only thing we care about as people.

So it’s not surprising that economists, political scientists and others have put forward a number of other indices. Here’s a quick menu of options.

  • Human Development Index: The HDI is a classic at this point, having been created by the UN Development Program about two decades ago. It combines life expectancy, some education metrics, and per capita GDP.
  • Multidimensional Poverty Index: The MPI combines a much broader set of indicators to measure deprivations in health, education, and standard of living (including access to electricity, sanitation or various assets). A couple months ago I described the debate around the MPI.
  • The most colorful of indices
  • Global Peace Index: We discussed this in my “Peacebuilding & Peacemaking” course a couple weeks ago. If you’re going to build/make peace, I suppose you need to know just what it is and how to know if you’ve got it. As with the others, this index is multidimensional. Fun fact: the US ranks pretty low on the GPI (85th out of 149) due military expenditures and proportion of the population in jail.

Also, if you like cool maps, check out this tool for comparing GDP, HDI and MPI.

The desire to measure something better than GDP is understandable. Surely these multidimensional measures more closely approximate “good”/”development”/whatever it is we’re really trying to achieve. But I’ve been trying to figure out how useful these really are. They’re constructed of many factors with arbitrary weights assigned, so seeing the index move up or down tells us nothing. We still have to dig into the individual factors to find why the index moved, in which case we may as well just track the individual factors.

Going back to the management saying: if what we’re measuring is an arbitrary mish-mash of questionable value, then maybe that’s exactly what we get.