How to adjust salaries of more than two countries to a common year given that the data was collected in different years?

Question

I want to compare salaries of more than two countries. The data I am using is based on the Occupational Wages around the World (OWW) database. Let's say I want to look at the salaries of the nurse. The salaries are specified in US Dollars and not all of them are collected in the same year, so we have sort of an incomplete data. As an example:

• Year: 2008, Country: Azerbaijan, Annual Salary (USD): 1678.15
• Year: 2007, Country: Singapore, Annual Salary (USD): 19523.57 (no 2008 data)
• Year: 2006, Country: China, Annual Salary (USD): 3338.208 (no 2008 or 2007 data)

According to the author of the database (Freeman and Oostendorp), to compare living standards, we have to deflate wages using Purchasing Power Parity (PPP) instead of exchange rates for the US dollars (It seems to me that their data is in local currency, rather than in USD as the example above). However, it is not clear to me how we adjust these salaries to a common year, say 2009.

Is it correct to do the following to my data above? Or is there a better way to compare salaries over countries and time to a common year? I use World Bank Indicator (http://data.worldbank.org/indicator) to compute things.

1. First, we multiply the USD salary with their corresponding LCU/USD exchange rate [Indicator: "Official exchange rate (LCU per US$, period average)"]

2. Then, we divide by the PPP-correction [Indicator: "PPP conversion factor, GDP (LCU per international $)"]

3. To adjust across time, then we multiply the result by ratio of GDP per capita, PPP between the common year (2009) and the year the data is collected (2008) [Indicator: "GDP per capita, PPP (current international $)"]

I would greatly appreciate any help.

Answer 1

Your three stage procedure is not correct.

In effect, since the data is already in USD, to apply PPP you need to revert it back into LCU (local currency units). That is your step 1. Have in mind this might introduce errors if the one you use is not the same the authors used.

Then, since wages are not for the same year, ideally you do need a proxy for wage growth of each occupation, for each country and year, so that you can compute wages in LCU for the same year. You are attempting to do this in step 3 using GDP growth, which is not ideal (and the order is wrong too). I would rather use an inflation index. This is not ideal either, but the World Bank provides an homogenised one. Alternatively, you could get wage growth for occupational groups (or aggregate) from national statistical offices, but this is a time-consuming exercise.

After putting wages into the same year, you proceed with step 2, where you convert LCU to USD PPP, using a common year.