by Égert Balázs, Peter Gal and Isabelle Wanner, OECD Economics Department
Researchers looking for empirical evidence on the relative impact of policy and non-policy drivers of economic growth know how much time and efforts can go into assembling a large database of policy variables and other determinants covering as many countries and years as possible. Even when such database can be patched-up from tapping into earlier studies, up-dating the series often requires going through a vast number of different data sources. Thanks to a recent OECD initiative, such task could be greatly facilitated from now on. A new OECD’s Structural Policy Indicators Database for Economic Research (SPIDER) is now available online and provides a broad range of data to researchers in ready-to-use formats to facilitate empirical/econometric research investigating the nature and the impact of structural policies. Available as a text file (TXT), STATA (.dta) and Eviews (.wf1) formats, the database includes about 500 policy and institutional indicators from almost 50 different OECD and non-OECD data sources. The policy variables stored in the database are annual or less frequently available (every five years or only once). The database will be updated on a yearly basis.
The database covers the following broad categories of policy variables: i) legal infrastructure and institutions describe features of the political system, the underlying legal institutions and indicators measuring the quality and various aspects of public governance; ii) framework condition policies include policies that condition the environment in which firms operate and make decisions such as the product market regulation (PMR/ETCR) indicators, the competition law and policy (CLP) indicator and a number of labour market institutions; iii) specific policies cover policies relating for instance exclusively to specific segments of the labour market (older workers, women or the youth) and include family benefits, policies influencing decisions to retire. Examples of other specific policies are measures primarily designed to support R&D investment or exports. Table 1 gives an overview on the main categories of variables included in SPIDER.
Aimed primarily at helping researchers to kick-start empirical analysis by keeping the costs of assembling the required data very low, the usefulness of the SPIDER database to economic research can be demonstrated through the range of possibilities it offers in terms of assessing the impact of institutions and policies in growth regressions. For instance, it allows for runing cross-country time series growth regressions for OECD countries including indicators of product and labour market regulations with about 20 to 30 years of data. As an illustration, the scope of variables that can be included in several variants of growth regressions.
The analysis can also be extended to non-OECD countries, although in that case the time series dimension of the data will likely to be shorter and regulations and institutions will be measured by indicators available from non-OECD data sources. Finally, for purely cross-sectional regressions, with a very large set of indicators that also capture the geographical, social and cultural aspects, the number of observations reach about 90. These examples indicate the scope of the database for such an exercise.
Aside from facilitating cross-country/time-series empirical analysis, the new database is a one-stop shop where a large set of internationally comparable policy variables can be found and used to gauge in a more descriptive manner the magnitude of structural reform actions in specific countries and areas over a broad range.
References:
The database is described in more detail in Égert, Gal and Wanner (2017), “Structural policy indicators database for economic research (SPIDER)”, OECD Economics Department Working Paper No. 1429.