[Column] Lies, damned lies and Korean statistics

Park Jung-soo

The author is the dean of the School of Economics at Sogang University.

 
Accurate diagnoses of various economic and social issues facing Korea, which is heading toward maturity as a developed country, are more important than ever. This requires reliable statistics that accurately reflect reality. If the accuracy of statistics is poor, it becomes difficult to properly diagnose problems and come up with solutions. If statistics are not reliable, it only undermines the basis for diagnosis and lowers the acceptance of the public.
 
For this reason, it is right for the government to aim to produce statistics as neutral and objective as possible. But the Moon Jae-in administration caused unprecedented controversy over income, jobs, and real estate data. (A probe by the Board of Audit and Inspection is still underway.) The most worrisome part of the controversy over income stats is that the previous administration’s action has undermined public trust in the national data system.
 
The last administration considered a 10 percent increase in the income of the low-income class in the fourth quarter of 2017 as an effect of the government’s income-led growth policy. But in the following quarter, the income growth rate in the same class turned negative, not to mention a worsening of income redistribution. After it was surprised, the Blue House found fault with the “incomplete nature of income statistics” and immediately changed the sampling method for income stats. The household trend survey used in the statistics at the time was reorganized from 2017.
 
However, I have a reasonable doubt that the statistics were manipulated based on political interest. The government’s inappropriate response has greatly damaged the credibility of national statistics. If a doctor says he will change his test method because the test result came out bad, will the patient trust any new test?
 
Misuse of housing price-related stats also undermined the credibility of the government. When the whole nation felt the surge in housing prices, the government argued that the upward trend was diminishing and real estate measures were working — merely based on statistics from the Korea Real Estate Agency (KREA).  
 
As income data is sensitive personal information, it cannot be collected by entities other than the state. But real estate data mostly is public and there are comparable private statistics, which showed a surge in housing prices at the time. The discrepancy between the two sets of statistics was caused by the sampling and statistical methods. The statistics by the KREA covers a wider range of regions but with smaller samples and a relatively low proportion of apartments. In contrast, statistics by KB Kookmin Bank are on a narrower range yet with more samples and a higher proportion of apartments.
 
The prevailing opinion at the time was that the data by KREA from a smaller sample was less likely to reflect reality. When the limits of the statistics used by the government are clear, it must conduct a comprehensive diagnosis by reviewing additional statistics in consideration of the pain felt by the public. But the government adhered to the KREA statistics, immersed in promoting policy effects, and the diagnosis lost the people’s trust in the end.
 
The two examples have a few implications about the principle of generation and utilization of statistics. First, reliability of a system generating the statistics should be secured. This means maintaining maximum prudence and maintaining objective and consistent principles in preparation methods and procedures, publication and interpretation. Second, efforts should be made to build accurate stats properly reflecting reality. The nation will trust statistics when the government works hard to produce and use realistic statistics regardless of political cost.
 
Third, from a long-term perspective, it is necessary to capture the changing interests of the people and reorganize the survey to reflect policies. Neutrality and objectivity should be ensured in the process of adding new survey data for the stats to help address issues pertinent to the times.
 
It is necessary to secure accurate and reliable statistics even if the work is slow, rather than rushing to produce more and faster. If the statistics system loses credibility due to the government’s inappropriate response, hard-earned data would be useless. No doubt, the direction of national development also will be lost. The government must not forget this obvious fact.
 
Translation by the Korea JoongAng Daily staff.