In a sequence of posts starting with Lies, Damned Lies, and Statistics, I have argued that the attempt to reduce multiple indicators to one number always introduces subjective elements relating to choice of factors, and relative weights to be assigned. By using technical jargon to justify choice of weights, the value judgments involved in this choice are concealed under a cloak of objectivity. This creates a modern form of rhetoric, where the arguments are made using numbers, and the values that went into the manufacture of these numbers remain hidden, and therefore, are not discussed. This concealment of values resulted from the creation of an artificial dichotomy between facts and values which became widely accepted. Values are considered unscientific, personal opinions, and hence must be concealed.
In the previous post on Cross-Country Comparisons of Wealth, we discussed how values were inevitably involved in such comparisons. In this post, we continue this discussion in the context of the most popular device used to attempt to resolve this problem – purchasing power parity. Like all positivist methods, this creates an impressive illusion of objectivity, while being highly subjective and value-laden.
NOTE: Reading over these posts, it does seem like I am rubbing the point into the ground and pressing repeatedly something which is trivially obvious. However, my experience with teaching this material leads me to the conclusion that the idea that facts and values are sharply separated, that numbers represent indisputable facts, and that objectivity is attainable and desirable, are built in so deeply into the foundations of our thought, and reinforced by routine use of methods based on these foundations, that these beliefs are very hard to shake. Even after repeated demonstrations, students continue to believe in the myth of objectivity. Very few understand that there is no objective way to rank universities around the world, or even in a single country. So, with due apologies to post-positivist readers, I continue my relentless assault on the illusory fact/value distinction.
Statisticians who are aware of the serious problems which arise in cross-country comparisons of GDP have come up with a device to reduce them. In making cross-country comparisons, it makes a lot of sense to consider how much a dollar can buy in the USA, and compare that with what a Rupee can buy in Pakistan. It seems like a more reliable method then using exchange rates or other benchmarks, the deficiencies of which were discussed in the previous section. Here the idea is firstly to determine a typical bundle of goods and price the bundle both in Pakistan Rupees and in US Dollars. Those two costs should be considered as equivalent because that shows a match between purchasing power of Pakistan Rupees and purchasing power of US Dollars.
As an artificial example, let a typical bundle of goods be the rental price of housing for an average person, price of food, price of clothing and basket of goods which consumers would buy. The first thing to be done is to calculate the costs both in Pakistan Rupees and US Dollars. Suppose that they are 20,000 Pakistan Rubees and 1,000 US Dollars respecively. This in PPP approach means that 20,000 Indian Rupees is equivalent to 1,000 US Dollars. This can be very different from official exchange rate. For example, the official exchange rate between Pakistan Rupees and US Dollars could be 100 to 1, that is, 100,000 Indian Rupees is equivalent to 1,000 USA Dollars.
On the surface, the PPP appears to be a good solution to the problem of cross country comparisons. However, when we probe deeper, we find that many subjective decisions must be made to arrive at a practical implementation of the idea, and the outcome of comparison depends on these decisions. A central point is that there is no “typical bundle” of goods which is the same across the world. A typical bundle of goods for a consumer in India is radically different from a typical bundle of goods for a consumer in USA. Even when the goods are the same, a “house” in USA is very different from a “house” in India. However, the numbers which seems perfectly accurate, objective and precise, do not reveal these difficulties. We can bias comparisons by choosing the bundles differently.
No abstract, purely theoretical and statistical resolution can be made of this problem. One must ask the question of the PURPOSE of the comparison. One such purpose, is stated by Hicks (1940): ‘a long line of economists … have sought in the Social Income an index of economic welfare, of the wealth of nations’. Actually, this statement contains two purposes – economic welfare, and wealth of nations – which are conceptually different. Let us first consider the extent to which GDP measures economic welfare, or prosperity, of nations. Here the Easterlin Paradox shows quite clearly that “happiness” felt by people is largely unaffected by GDP growth over time. Also, there is no systematic relationship between GDP and social welfare across cross sections. Subsequent research has found that deeper explanations for this paradox lie in the fact the human welfare is based on character, attitude towards life, and social connections. These factors are ignored in the GDP, In fact the GDP is grounded in economic theories which falsely suggest that consumption of goods and services is the sole source of human happiness – a position which can be described as the “Coca-Cola Theory of Happiness”.
A second purpose for measuring wealth, explicitly considered by Adam Smith, relates to the power countries exercise in the international arena. Wealth provides capabilities for financing military expenditure, unfortunatelty an essential aspect of global power today. However, if wealth is used to compare the relative power of the two countries in the international arena, then emphasis would be placed on rather different factors, and weights would be rather different from those used for Purchasing Power. Suppose, for example, that we only look at the amount of money used to finance Army, Navy, and Air Force expenditures. We can add up all three, or use other sets of weights to assess power, but all such schemes are arbitrary. The question of which country is, objectively, the most powerful, cannot be answered. For example, if one country has a huge army while the other has a bigger navy, then one is more powerful on land and the other on the sea. Depending on circumstances, and terrain of struggle, either one could come out on top.
The purpose of making comparison affects radically choice of factors and weights, which is a surprise to those used to thinking of statistics as neutral, objective, and value-free. See Castles and Henderson (2005) for a discussion of the many controversies in the area of comparing GDP across nations, and the policy implications of the use of different kinds of weights and factors. However, note that, like all economists, they believe that value-neutral statistics can be found, and used as a basis of value-laden decisions. This idea, that we can separate the facts and values, is the fundamental misconception at the methodological foundations of modern economics, econometrics and statistics. Because it goes so deeply against the grain of positivist methodology that we have all absorbed, it is worth re-iterating:
Impossibility: It is impossible to make objective comparisons when multiple factors are under consideration. Choice of factors, weights, and signs (positive or negative), are all necessarily subjective.
As a consequence, it is impossible to make objective cross-country comparisons of GDP. This is in conflict with the positivist mindset, which leads us to believe that there is an objective truth, and that we CAN find the right collection of factors and weights which will reveal the truth. For example, Castle and Henderson (2005) argue that Environmentalists are using the “wrong” set of factors and weights, and they they aim to provide greater objectivity. They do not realize that objectivity is an impossible goal. All we can do is put forth our values as being better in comparison to other values; this is exactly the role of rhetoric and persuasion. Environmentalists use weights which emphasize the costs of climate change, so as to create political pressure to take action. Industrialists propose another set of weights which gives more emphasis to the market, in order to allow growth and profits at expense of the environment. Both sets of weights are subjective choices, and there are no objective choices available, but only the illusion of objectivity.
To articulate this more clearly, suppose I was charged with the task of making a cross country comparison which would show Pakistan to be ahead of USA. I would look for factors where Pakistan leads USA and give them greater weight. Easterlin’s Paradox has established firmly that measures of happiness across countries do not correlate with GDP. I would therefore argue that instead of comparing material goods directly, we should be measuring the social welfare, or happiness levels, produced by the consumption of these goods. Studies of happiness show that the structure of the family is one of the key sources of life-happiness. Children raised by single-parents suffer from a large range of problems, documented in numerous research studies. If we give weight to dimensions of social welfare which come from family and community, and consider statistics related to crime, suicide, alcoholism, loneliness, we could easily show that Pakistanis are “richer” than Americans, if wealth is defined appropriately to include social lives.
It is important to clarify that we are not arguing that there is no truth, and everything is subjective. Rather, truth is complex, multidimensional, and qualitative, and it cannot be reduced to one number. We cannot assign a single number to a country as a measure of its “wealth” and thereby make it possible to compare the lives of millions living in country with millions of lives in another country. When we attempt to reduce complex, multidimensional phenomena to one number, there is an enormous loss of information. Decisions as to what information is important and must included, and what can safely be ignored, are always subjective. In the past, rhetoric was used to emphasize importance of one set of values, and to criticise other values in use by other groups. Now, all this rhetoric is concealed within choices of factors and weights, allowing some groups to impose their values on others, under the cover of objectivity of numbers.