Tuesday, May 15, 2012

Industrial Policy: Picking Winners Works, Provided…

One of the interesting things about development in East Asia over the past half a century is the degree of government involvement in promoting certain sectors of the economy, notably export-oriented manufacturing. That’s been the guiding principle of Japan’s and Korea’s development for instance, and Singapore and Taiwan haven’t been far behind.

In Malaysia’s case, attempts at directing industrial policy has been a bit hit and miss – some industries (palm oil for instance) have been pretty successful, but others (heavy industry) much less so.

This new working paper published by the NBER helps to explain the reasons why (abstract; emphasis added):

Industrial Policy and Competition
Philippe Aghion, Mathias Dewatripont, Luosha Du, Ann Harrison, Patrick Legros

This paper argues that sectoral policy aimed at targeting production activities to one particular sector, can enhance growth and efficiency if it made competition-friendly. First, we develop a model in which two firms can operate either in the same (higher growth) sector or in different sectors. To escape competition, firms can either innovate vertically or differentiate by chosing a different sector from its competitor. By forcing firms to operate in the same sector, sectoral policy induces them to innovate "vertically" rather than differentiate in order to escape competition with the other firm. The model predicts that sectoral targeting enhances average growth and productivity more when competition is more intense within a sector and when competition is preserved by the policy. In the second part of the paper, we test these predictions using a panel of medium and large Chinese enterprises for the period 1998 through 2007. Our empirical results suggest that if subsidies are allocated to competitive sectors (as measured by the Lerner index) and allocated in such a way as to preserve or increase competition, then the net impacts of subsidies, tax holidays, and tariffs on total factor productivity levels or growth become positive and significant. We address the potential endogeneity of targeting and competition by using variations in targeting across Chinese cities that are exogenous to the individual firm.

In Japan and Korea, development was synonymous with the growth of the keiretsu and chaebol respectively. In Malaysia, the palm oil industry is characterised by several large and many small firms, all competing with each other. Even though many of the larger firms are government linked (Sime Darby, the now defunct Guthrie and Golden Hope, Felda), there was enough in the way of competition between these companies and more privately held companies, to foster real growth.

Heavy industry on the other hand has not had the real benefit of competition, with the development model largely based on an anchor company with the hopes of birthing a cluster of companies in the supply chain – examples include Hicom, Perwaja and of course Proton. Without the pressure of “wasteful” competition, the conclusions of the working paper suggest that – as we’ve discovered – the results won’t be as fruitful.

So industrial policy can be made to work, provided you’re targeting an industry as a whole, and don’t pick winners before hand. I realise I’m probably generalising a bit too much here, but it’s an interesting result, with lessons we’d do well to take on board.

Technical Notes

Aghion, Philippe and Mathias Dewatripont, Luosha Du, Ann Harrison & Patrick Legros, "Industrial Policy and Competition", NBER Working Paper No. 18048, May 2012

1 comment:

  1. As the Lerner index tends to evolve over time for different industries, its still very much a guessing game. For instance, I think back in the early 1980s it was impossible to tell whether our fledgling semiconductor industry had a higher index reading than say our heavy industries.

    Maybe the selection criteria for NKEAs should have taken that into account. Although now industrial targeting (by PEMANDU's selection) just means, expanding on stuff that we know that we are good at in the past.

    ReplyDelete