Introduction

Diversification, especially in the context of small, highly trade-dependent economies, has recently become quite a fashionable topic, and something that increasingly commonly forms part of the policy advice offered to low- and middle-income countries seeking to improve their economic conditions or strengthen their economies. However, diversification per se is not policy advice: it is merely a descriptive term. Moreover, as such, the more one thinks about it the more one realizes that its meaning is not terribly clear or precise. Hence there is scope for an investigation of the term with a view to clarifying its possible meanings, evaluating which, if any, make sense from the standpoint of practical economic policymaking, and assessing the circumstances under which economic diversification can indeed be a desirable goal for a country to pursue. That is the principal purpose of the present chapter.

Aside from its analytical content, the empirical focus is on a group of 20 countries, consisting of South-Eastern Europe (SEE) and the Commonwealth of Independent States (CIS). Including Romania and Bulgaria that are already EU member States (since January 2007), the SEE region includes eight countries; the remaining twelve countries belong to the CIS. For such a diverse region, one cannot reasonably expect to find uniform policy advice that would suit all countries, but we can hope to develop a common approach or methodology. Such an approach will build on a number of strands of evidence and analysis that are explored more fully in subsequent sections of the chapter:

• ideas about general requirements for sustained economic growth;

• ideas about engaging with the world economy, including desirable degrees of diversification;

• analysis of the institutions needed to support growth with diversification; and

• analysis of the accompanying policy tools and measures.

Conditions for Growth

As Table 2.1 shows in summary form, the 20 countries studied here represent an extremely diverse region in terms of their populations, geographical size, income per head and recent growth experience, resource endowments, access to markets (for example, whether landlocked), progress with market-oriented reforms and their political configuration (in so far as this influences the economy). In terms of reforms, it can be seen that a few countries have as yet made very little progress towards building a market-type economy, notably Belarus, Tajikistan, Turkmenistan and Uzbekistan.

Most SEE and CIS countries had a very bad decade economically in the 1990s, experiencing severe post-communist recessions which in some countries were greatly exacerbated by civil and/or international wars (see World Bank 1996; EBRD, various years). Since 2000, economic performance in terms of real GDP growth has generally been much better and has exhibited lower variance. The strongest performers in the region until the global crisis hit these countries in 2009 enjoyed high growth of around 9 per cent per annum or higher for several years in a row. The high-growth performers included, for instance, Armenia, Azerbaijan, Kazakhstan and Turkmenistan. On the other hand, some SEE countries are still growing quite slowly—too slowly to bring down unemployment rapidly; these include Bosnia and Herzegovina, Montenegro, Serbia and the former Yugoslav Republic of Macedonia. Most of the region already has inflation down below 10 per cent per annum or is well on track to achieve that very soon. General government balances are mostly manageable, and are, on average, healthier than those of the new EU member States, such as Hungary and Poland. Also, it appears to be the case that the faster growing countries have lower shares of government in GDP. Last, indebtedness and the debt burden (measured by debt servicing as a percentage of export earnings) show a good deal of variation. Some of these key macroeconomic indicators for the SEE and CIS countries are shown in Table 2.2.

Experience around the world suggests that countries do not sustain growth unless they ensure sound macroeconomic conditions. In practice, this is a multi-dimensional requirement, including low inflation, manageable budget and external deficits and credibly manageable levels of debt. As usual, it is very hard to attach precise numbers to these factors, not least because the inter-relations between them are influenced by the dynamics of growth in a given economy.

Thus, if an economy is growing very slowly, say at 1–2 per cent per annum in real terms, and has a general government budget deficit of 5 per cent of GDP (that is to say, this is the deficit based on fully consolidated public sector accounts), with an accumulated public debt of, say 60 per cent of GDP, then the debt of the public sector is growing much faster than GDP, ensuring that aggregate public sector debt is a rising share of GDP. Servicing such debt then accounts for ever-rising shares of public spending, and becomes increasingly unmanageable. Conversely, in a country with the same government deficit and initial debt, but growing much faster, say at 10 per cent per annum, the debt grows more slowly than GDP and so declines gradually as a share of GDP. In this case, debt is not getting out of control.

Much the same sort of argument applies to a country's external accounts. Here the issue has to do with the interconnections between a trade deficit, the growth in external debt, and the two factors that can help to reduce the debt or make it more manageable, namely export growth and inflows of foreign direct investment (FDI). In the early 1990s, for instance, transition economies such as Hungary carried what seemed quite an unmanageable external debt. The debt was indeed very large in relation to GDP, but export growth in the early to mid-1990s was sufficiently rapid that the effective burden of the debt fell fairly rapidly, assisted by large inflows of FDI. Again, therefore, a large external debt can be handled in a country with fast-growing GDP and/or fast-growing exports, while the same debt in a country whose GDP and exports are stagnant or only growing slowly can prove catastrophic.

Inflation, too, is considered bad for growth, but the arguments as to when this is likely to be the case, and why, are quite complex and need careful examination. Two extreme cases stand out.

First, sustained inflation at high rates, such as well over 100 per cent per annum, is harmful because it renders any contracts expressed in monetary terms extremely risky and makes any long term business agreements difficult to sustain. Savings depreciate rapidly and debts are also eroded rapidly in real terms, since most contracts are expressed in nominal monetary terms, unadjusted for inflation in the general price level. At seriously high rates of inflation this shades into the extremes of hyperinflation (conventionally defined as inflation faster than 50 per cent per month), usually reflecting the near complete breakdown of monetary discipline and control, as in Zimbabwe in the late 2000s. Serbia, though, also experienced a short spell of hyperinflation in the 1990s, and all of the CIS countries experienced annual inflation in excess of 1000 per cent in 1993 or 1994. None of these countries found such high inflation consistent with positive GDP growth.

Second, rapid and unpredictably variable inflation is damaging even when the average inflation rate is much lower than the above case. Thus, if inflation is just 20 per cent per annum, but fluctuates randomly between 5 per cent and 45 per cent, say, this is a nightmare for anyone planning a long-term business contract. It is hard to predict either the costs or the revenues associated, for example, with a new investment, and the likelihood is that such investment will simply not take place.

The very real problems highlighted here make clear why macroeconomic stability is important for sustained economic growth, but they do not amount to a firm prescription regarding the exact conditions required to achieve such stability. The EU's Maastricht conditions for entering the European Monetary Union (EU), taken together with the policies of the European Central Bank (ECB) for the Euro zone, provide one set of guidelines for macroeconomic stability. These conditions state that a country's public debt should not exceed 60 per cent of its GDP; that the government deficit should not exceed 3 per cent of GDP; and the ECB sets Euro zone interest rates to achieve, over the medium term, an inflation rate of at most 2 per cent per annum. The UK Government has a similar set of fiscal and monetary indicators as its policy framework, though the numerical targets differ from the EU's. From the point of view of market confidence and policy credibility, it is probably not a bad idea to build policy around specific targets like this, although there may not be anything in economic theory that might guide us to any particular numbers. Moreover, there is the further practical point that Governments running up against the constraints set by whatever framework they claim to be operating tend to find ways of “evading” them: targets are redefined, time periods over which some target is to be achieved are “adjusted”, and so on (see Buiter and Grafe, 2003). In the end, the only effective form of public accountability in regard to the basic parameters of macroeconomic policy is the next general election. Hence, when advising other countries, it remains necessary both to emphasize the importance of macroeconomic stability and to be quite pragmatic and flexible as regards the proposed targets and implementation frameworks.

Investment, too, is generally found to be essential for growth. More concretely, moderate to high rates of investment (that is to say, typically in excess of 20 per cent of GDP), allocated efficiently and credibly funded (from a mix of domestic and external savings), are very important. Let us now explore this notion more carefully. It is clear that high rates of investment are not sufficient for growth, since there are plenty of examples around the world of countries investing a lot and failing to grow. The 1980s, the last decade of the former Soviet Union, provides one of the more striking examples of this, since investment certainly occurred at respectable rates, mostly above 20 per cent of estimated GDP, while the economy experienced very weak growth for the whole decade. Thus investment not only has to be undertaken, but on average, it must be efficient and productive.

A simple equation from elementary growth theory for a closed economy, focusing on the supply side of the economy, makes this point very forcefully. The equation is: where g is the rate of growth of real GDP, s is the savings ratio, and v is the (incremental) capital-output ratio of the economy concerned (see Solow 1970). In a more complex model, additional factors such as depreciation of the existing capital stock, foreign trade and FDI, and sometimes the effects of technological change might also influence the growth rate, but we return to that later. For now, we remain with the equation as specified above. Consider three examples:

1. Suppose s = 0.4 and v = 4. Then g = 0.1, or 10 per cent per annum. Something like this fits China quite well, with its very high savings ratio and moderate investment efficiency.

2. Suppose s = 0.25 and v = 3. Then g = 0.083, or 8.3 per cent per annum. This fits quite a number of rapidly growing countries, possibly including Kazakhstan and the Russian Federation (though the Russian Federation has not been growing quite so rapidly). The aggregate savings (and hence investment) ratio is lower than that for China, while the assumed average efficiency of investment is somewhat higher than China's (that is to say, v is lower).

3. Suppose s = 0.12 and v = 6. Then g = 0.02, or 2 per cent per annum. A country in this situation is investing rather little, just 12 per cent of GDP, and is doing so very inefficiently, each unit of investment adding very little to GDP. The result is very slow growth, as can be found in many of the poorer developing countries.

To sum up, the equation above implies that to grow at a respectable annual rate, such as more than 5 per cent per annum, an economy needs some combination of a relatively high savings ratio, s, and a relatively low capital-output ratio, v. Although originally developed for a closed economy, equation 2.1 remains valid for the open economy case, too. Hence, in thinking about the savings (and investment) ratio, we need to have in mind not just savings by the domestic economic agents—firms, households, government—but also external savings that are invested in the economy concerned. The latter takes various forms, the most common being foreign portfolio investment and foreign direct investment (FDI). Sometimes countries claim that they are “too poor to save” and that their development, therefore, will largely rely on inflows of FDI. In practice, this approach is unworkable, since virtually everywhere most investment is financed largely from domestic savings. Hence, domestic banks and financial markets need, above all, to mobilize and efficiently allocate domestic savings as a major element in the process of stimulating sustained economic growth.

Now consider the capital-output ratio, v. To put it simply, a low value for v means that a unit of investment generates substantial additional output on a continuing basis. This is what is meant by stating that investment needs to be productive. Most investment resources should be devoted to building factories, shops, offices and the like, associated with profit-seeking, mostly private businesses. In addition, much investment in infrastructure such as airports, port facilities, road and rail networks, public utilities, and so on is clearly productive in the same sense, provided that it does not take the form of “white elephants”; for example, building a “road going nowhere”, or a new port where there is no demand for its services. If v is high, meaning that, on average, investment is not very productive, then either there are general problems in the economy holding down the returns to private sector investment (for example, excessive corruption, regulatory barriers, etc.), or the mix of selected projects is heavily weighted towards unproductive activities like building presidential palaces and other such monuments, or perhaps to constructing large defence facilities.

Besides the need for a sound macroeconomic environment and a high rate of productive investment, sustained growth is also assisted by a good business environment, by efforts to improve labour force quality, and by openness to the world economy. The first two of these we discuss briefly here, while the third is reserved for the next section.

The notion of a good business environment can be considered in terms of the basic conditions for doing business, as in the World Bank's annual Doing Business surveys, or in terms of outcomes (see World Bank 2007). The basic conditions include such things as how long it takes to set up a business, whether credit is readily available, how much corruption new businesses can expect to encounter, how frequently firms are “inspected” by various public authorities, and so on. Thus the conditions are very much about assessing how business-friendly the given country appears to be, and to that extent, how the various published indicators are both interesting and useful. However, they are far from the full story, since what really matters for growth is not so much these background conditions per se, but rather how they translate into new business formation, business closure and firm growth. In other words, what really count are the business outcomes.

Countries, including several in SEE and the CIS region, commonly misunderstand the notion of the business environment, and underestimate its importance for their economic prosperity. Most countries are able to assert correctly that their business environment has greatly improved in the previous five years, or over the previous decade, but sometimes when they do so they fail to appreciate that the improvement is from “bad” to “poor”. There might well be a notable improvement compared to their own past, but considered more objectively and in the context of a wide-ranging international comparison, they may still lie well behind current best practice. This might be even more the case when outcomes are examined, with new business starts occurring at low rates, the stock of firms still too low for the size of the economy, and too many long-established but poorly performing firms surviving long past their “sell-by dates”. Yet, growth occurs most successfully in economies with high business start-up and closure rates (in other words, most new firms fail, often quite rapidly), with a few new firms growing to become the success stories of the future. It is also extremely important that old firms should not enjoy extended protection that enables them to survive for too long.

Last, and often neglected, is the point about improving labour force quality. For most of the countries discussed in this paper, labour force quality cannot be a big issue, especially in the short term, since all the SEE and CIS countries inherited sound basic education systems from their socialist periods. Thus, for the most part, general literacy and numeracy can be taken for granted and educational provision and attainment to the secondary school level are already quite good. But in the medium and longer term, the structure of the workforce will naturally change towards requiring more relatively skilled and well educated workers, and this will entail substantial improvements in the educational systems of our region, especially at the higher levels—upper secondary, colleges, universities. Moreover, as Alam and others (2008, ch.4) point out, many existing advanced courses are not well designed to meet the needs of a flexible, dynamic market-type economy, so substantial educational reform is still needed in most countries. Continuing to improve the labour force in step with the growth of the economy will be a major challenge for the future, though the details fall well outside the limited scope of this chapter.

One final point needs to be stressed to conclude this section. This is the simple remark that sustained GDP growth is generally the most effective way of reducing poverty. For a time it became “fashionable” to focus development efforts on poverty reduction, or to think about notions such as pro-poor growth, this being growth that in some (much debated) sense relatively favours poorer population groups. Even the International Monetary Fund (IMF), with its emphasis on country-led Poverty Reduction Strategy Papers (PRSPs), has found itself caught up in this heavily poverty-oriented thinking (see World Bank and IMF 2004). The World Bank, too, has placed great emphasis on pro-poor growth, increasingly highlighting the need for sustained general growth (see Besley and Cord, 2007; World Bank 2005). Of course, extreme poverty is a terrible thing, but here we merely wish to highlight our view that it will rarely be eradicated, or even much reduced in a sustainable way, except through general economic growth.

Integration in the World Economy

Until 1990, the share of the socialist bloc countries in world trade was falling steadily, and of their total trade, most was with each other (see Maddison 2001). This lack of engagement with the world economy, moreover, was a symptom of these countries' generally poor economic performance, characterized by low productivity, poor rates of innovation, and real incomes lagging increasingly behind those of the developed world. After declining sharply in the early 1990s, trade of the CIS and SEE countries has increased substantially, though much is still trade “within the region” rather than trade with the wider world (see Shelburne and Pidufala, 2006; and Broadman 2005).

We generally expect exports as a share of GDP to be lower in large, already diversified economies, than in small economies with a comparatively narrow domestic production base. To a large extent, this high export share in smaller economies is what enables consumption to be diversified even in a very specialized economy, since export earnings are used to pay for the required diverse imports. It is not uncommon to find, in a small economy, that exports easily exceed GDP, while in a large one they may only be 20–30 per cent of GDP. Further, in recent decades virtually all the most successful growth experience has been export-led. Thus, deliberate and extensive engagement with the world economy has generally proved to be an effective development strategy, and has done more to lift people out of poverty than any amount of development assistance (as emphasized in WTO 2008; see also Wolf 2004).

Integration in the world economy involves a mix of elements: (a) trade in goods; (b) trade in services; (c) income flows: profits, dividends, remittances; (d) aid and other external support (grants and loans); (e) FDI; (f) other capital flows (short- and long-term lending and portfolio investments); and (g) flows of people: inward and outward migration. We consider each element in turn.

Countries usually start by liberalizing trade in goods and services, then later liberalize the capital account—this was the path followed by many of the transition economies, for instance. Trade liberalization itself usually entails a mix of measures for the countries in our region, often implemented in stages. The first and easiest stage is to dismantle most of the old controls and restrictions on trade in goods that were so prevalent in these countries when they were still centrally-planned economies. Next, it is important to rationalize, simplify, and lower the general level of tariffs on imports, since in most centrally-planned economies the structure of tariffs often featured some very high rates and was economically quite irrational (in the sense of encouraging inefficient trade, discouraging efficient trade). Then export promotion is also essential, since liberalizing imports without actively fostering exports can prove self-defeating to their cost, as several countries around the world have found (for example, countries in Africa in recent years; see Ackah and Morrissey, 2005; and Iyoha 2005).

It is not necessary to dwell here on the next two items, income flows and aid, so we turn to capital flows. Early capital account liberalization was, for a time, strongly encouraged by the IMF and other international institutions, though the experience of numerous financial crises in the past decade affecting a wide range of countries (most recently the late 1990s crisis that hit hard the Russian Federation and some other CIS countries), together with the current turbulence in world financial markets, have led to considerable backtracking from this position. In any event, the liberalization of capital flows commonly starts by encouraging FDI, often linked to privatization programmes.

Migration flows generally depend on economic opportunities at home and abroad, as well as on the immigration policies of potential partner countries and/or on the porosity of their borders. A bad economic situation at home, such as high and rising rates of unemployment, often stimulates out-migration, especially if an accessible neighbouring country offers attractive job opportunities. Albanians moving to work in Greece, or Armenians moving to work in the Russian Federation are commonplace examples of this phenomenon, but there are many more. Migration is sometimes politically problematic, though for the sending country it does frequently offer several benefits: (i) domestic labour market problems are eased; (ii) the migrant workers frequently send some of their income back home (remittances), and for some countries this is a major source of foreign currency; and (iii) migrant workers often acquire skills and knowledge which, when they return home, eventually benefit their home economy.

Aside from such economic stimuli, people also move for political reasons, to escape from efforts at ethnic cleansing, to escape civil war, to escape other forms of political repression. Bosnia and Herzegovina, for instance, still contains many internal refugees—internally-displaced persons—as a result of the Yugoslav wars of the 1990s. The former Yugoslav Republic of Macedonia also experienced severe ethnic tension between its Slav and Albanian population groups at the start of this decade, and several CIS countries have also suffered from ethnic conflicts that sometimes led to civil or even international wars during the early 1990s. All such events result in some migration, either within or between countries.

Returning to the principal theme of this section—foreign trade—liberal and open trading conditions are supported by WTO membership. From the SEE and CIS countries, WTO applications are in progress for: Azerbaijan, Belarus, Bosnia and Herzegovina, Kazakhstan, Montenegro, Serbia, Tajikistan, the Russian Federation and Uzbekistan. Ukraine's accession was approved in early 2008 and came into effect in July 2008. Other countries in the region are already WTO members, except for Turkmenistan which has not yet applied. As the Russian case illustrates, the WTO accession process can be long and arduous, with many twists and turns, each significant trading partner being free to bring to Geneva its own issues regarding trade with the Russian Federation. Thus, accession entails a mix of collective negotiation through the Accession Working Party, leading, eventually, to an agreed Accession Protocol; and a whole series of bilateral agreements with individual trade partners. Outstanding topics for the Russian Federation include agricultural support, domestic energy pricing and access to the Russian market for foreign services providers (such as banks).

Whether already WTO members or not, our countries mostly belong to a variety of Free Trade Areas (FTAs) and, in a few cases, Customs Unions (CUs). Unfortunately, most existing FTAs and CUs in the region are badly designed, badly administered, and economically ineffective, offering too many opportunities for corruption. For the CIS countries, as well as for SEE, these issues are discussed comprehensively in Broadman (2005). There are still too many bilateral agreements, giving rise to a “spaghetti bowl” of trade agreements, each of which is characterized by somewhat different lists of products included in the agreement, and often by restrictive rules of origin. Some countries, notionally at least, belong to four or five such agreements, which must be both economically inefficient and administratively complex. Small countries simply do not possess the capacity to handle multiple agreements effectively, and, in any event, the potential economic gains are often not large enough to justify them.

If the region wants FTAs amongst various subsets of countries, they should be simple, with broad commodity and country coverage, with liberal rules of origin, and with few exclusions. For example, such an agreement has been discussed for the SEE countries for some years, but has only just in the past year or so started to be implemented.

Economic Diversification

We now consider what economic diversification means, and whether—and if so, under what conditions—it is desirable. First, then, is the question of meaning. It is actually rather tempting to dismiss the whole idea of diversification as mere sloganeering, an idea that sounds quite appealing but which turns out to be lacking in substantial content. All economies that operate competitively and significantly in the world market are surely engaged in economic diversification all the time. After all, how else will firms compete, except by offering new innovative products onto the market, stopping the production of outmoded products that no longer attract sufficient demand, or by improving technology and cutting their costs? Nowadays, of course, these remarks apply just as much to the production and delivery of services as they do to the more familiar production of goods. For a national economy, the same is true except on a far larger scale, and with the additional point that some of the necessary changes occur through the entry and exit of firms.

So if, under normal conditions, a good deal of the economic change and adaptation that can reasonably be understood as falling under the heading of “economic diversification” occurs through the operation of market forces, why is diversification per se something that we should be especially concerned about? The reason, I think, is that sometimes the market forces that we mostly take for granted fail to work sufficiently well. We elaborate on this point at the end of the section.

Meanwhile, we accept that diversification is potentially a problem, and consider how to define it. A simple notion of diversification is that it means a country should produce, and presumably export, a wider range of products than in the initial position; strictly, this definition also includes diversification into exportable services such as medical care, education, tourism (in such cases, customers come to get the service), and so on. A more complex notion of diversification is that a country should produce for export a wider range of goods and services, with the emphasis on high-technology, higher value-added, “modern” items.

Next, we consider whether diversification as an objective of policy is desirable. For resource-rich economies, a standard argument for diversification is to mitigate the effects of Dutch disease. This is the situation where resource exports push up a country's equilibrium exchange rate and as a result either price out of the world market some or all of the country's existing manufactured exports or create conditions where it is more difficult to expand such exports. Also, historically, many significant resource prices have been highly volatile, so reliance on resource exports for foreign currency to pay for imports can be risky. The risks can be mitigated by the creation of resource funds in good times, as has been done by Kazakhstan and the Russian Federation, among others. But it is also argued that economic diversification can help to make an economy less vulnerable to these risks. Sometimes, too, it is claimed that natural resource production/exports benefit from relatively little innovation and productivity gain, so that an economy specializing heavily in such sectors would typically experience slow productivity improvements. This is a further argument for diversifying into sectors that normally benefit much more from such gains.

Taken together, the above points appear to add up to a convincing case for active policies to promote economic diversification in resource-rich economies. But we need to proceed more carefully, as the arguments are less compelling than they seem to be. First, if an economy has abundant natural resources, then they are likely to be profitable exports and ample inputs of capital and labour need to be concentrated in the resource sector for it to develop successfully. This necessarily draws production factors away from other sectors, including manufacturing. Hence what we term the Dutch disease may often be no more than the normal adjustment of an economy doing well with natural resource exports. To this extent, active steps to offset the impact of Dutch disease, perhaps following lobbying efforts by manufacturing firms losing out from the higher exchange rate, are to be resisted, as they can damage the competitiveness of the entire economy.

Second, the proper way to deal with resource price volatility is by purchasing insurance, and this is, in effect, what countries do when they establish resource funds (or, in agriculture, hold physical stocks of commodities). The implication is that in “good” periods, when resource revenues are high, domestic spending should still be controlled carefully so that the resource fund can be built up—usually in the form of a mix of foreign financial assets—and then in bad times some of the fund can be drawn down to maintain domestic spending. If we deal with price volatility by diversifying into other sectors, then we are unavoidably drawing production factors away from natural resource extraction and production and, in effect, passing up an opportunity to make good profits. To put it more bluntly, specialization in sectors where an economy might lack a fundamental comparative advantage is surely not a good way of managing the risks associated with fluctuating resource prices.

Third, despite the frequency with which the claim is put forward, there may be no solid evidence to support the view that the natural resource production is associated with especially slow rates of innovation and productivity improvement. Massive amounts of R&D are devoted to the oil and gas sector by the major international companies, perhaps rather less to extraction of metallic ores and timber. The world markets for natural resources are strongly competitive, and this must also support productivity gains.

As regards innovation, a more general remark is worth inserting here. This is simply that the SEE and CIS countries spend very little on R&D and other innovation-related activities, generally well under one per cent of GDP. This contrasts with EU spending goals to achieve R&D spending of 3 per cent of GDP by the next decade, though at present many EU member States only spend 2 per cent or less of their respective GDPs on R&D (taking public and private R&D spending together). In the long run, the entire region studied in this paper will need to raise innovation spending substantially.

Returning to our argument about diversification, these remarks seem to undermine quite substantially the standard case sketched above for economic diversification in resource-rich economies. This does not mean, on the other hand, that diversification should be considered a bad thing, or economically damaging; rather, it simply implies that we need to think about it more carefully than hitherto, to think about why we want to encourage it, and what the real (as opposed to imagined) costs and benefits might be.

The discussion to this point has focused on the case of resource-rich economies. For others, notably for small open economies, domestic production is often narrowly based, with few significant exportables. This again is a source of potential economic vulnerability, hence arguing for diversification.

Regardless of the arguments about whether or not diversification should be pursued, there is quite clearly no point in doing it unless the resulting new goods or services are produced to a good quality standard, sufficient to be internationally competitive. Moreover, in my view, it is usually very unwise for the Government to attempt to dictate or select which sectors to favour, since Governments have a very poor track record in such matters, and are frequently mistaken in their judgements. Perhaps surprisingly, in practice we cannot even know in advance which sectors should be regarded as high technology or “modern”, so again, it is not a good idea for Governments to choose. For what we think of as “high tech” products and services might be produced and delivered using very “low tech”, quite mundane technologies, while apparently basic and commonplace goods could be produced using high levels of automation and extremely sophisticated equipment. Governments are generally poorly equipped to know about these things, so they should not normally be involved in making such choices. Ideally, therefore, it is better to rely on market mechanisms to “choose” the new sectors in which to develop production and exports.

However, markets do not always function as well as we are inclined to assume, and often need “help”. There are several reasons for this situation, including imperfect and/or unequal information on the part of market participants, their inability to finance desired transactions, problems in the legal area to do with protection of property rights and business contracts, regulatory deficiencies, and so on. These and other issues fall under the heading of “institutions”, the subject of the next section.

The Role of Institutions

Institutions are relatively stable social arrangements, often embodying various kinds of norms, customs and conventions (see North 2005). In the economic domain they frequently possess a number of special characteristics such as influencing the behaviour of economic agents, embodying shared expectations, and assuming the form of a “repeated game”. The last point is especially important, and means that economic agents (buyers and sellers in the simplest cases) do not think of their business transaction as being “one off”. Instead, they expect to be engaged in a whole series of similar transactions, and this then provides incentives for them to follow the rules, operate fairly, and so on. In this sense, the “rules of the game” can often turn out to be self-reinforcing, a useful characteristic (this can be true even without the help of a “State”, as Greif 2006, and Dixit 2004, have shown in various examples of informal trading networks).

Institutions, which can be either informal or formal, operate at different levels and in different contexts. At the most basic level can be found the social norms and customs that govern most everyday behaviour, whether explicitly economic or not. Next are the various resources and assets of a society, and the rights, powers and responsibilities associated with each of them. Last, we find the specific organizations which embody the institutional arrangements of the given society/economy. These include individual firms (which can be either formal or informal), households (mostly informal and customary, but usually with some formal legal underpinnings, for example marriage law, family law, inheritance law, etc.), business associations, economic departments and agencies of the Government, the courts, and the military establishment (see Acemoglu and Robinson, 2006).

From an economic standpoint, institutions need to provide for three key functions, namely, the protection of property rights (both from other private agents and from the State itself) (on the rule of law, see Dam 2006); supporting transactions (for example, contract law, improve information flows, accommodate risk, etc.); and facilitating cooperation and coordination, especially where it is beneficial for society but would not likely result from the unrestrained market mechanism (see Bardhan 2005).

For the economies in transition, especially the SEE and CIS countries studied here, what we might call the “institutional transition” is especially critical. For in essence, the transition from plan to market involves the replacement of one set of economic institutions—that which corresponds to central planning—by a new one corresponding to the requirements of a well functioning market-type economy (see World Bank 2002; also IMF 2005, chap. 3). This is a far bigger change, occurring over a far shorter timescale, than the regular evolution of institutional arrangements that goes on all the time in any economy. Hence, it should not have been too surprising that some of the new institutions needed for the market were slow to be established, slow to take root in many countries. Building a market-type economy entails creating a whole new set of “shared expectations” about how business decisions are taken and economic transactions are conducted. To give just one example by way of illustration, respecting and protecting private property rights is completely taken for granted in well established market economies, but for most of the SEE and even more so for the CIS countries it was an entirely new concept, still not deeply embedded in these societies.

Moreover, decades of central planning not only left the SEE and CIS countries with the “wrong” institutions, it also left behind a legacy of uneconomic production in two important senses: (i) much production was poorly located or in branches of production where the economy concerned had little chance of producing competitively, so transition began with this massive structural problem; and (ii) each economy contained amazingly few enterprises, most of which were far too large—the size structure of firms was completely different from what one observes in any “normal” market economy, with hardly any small and medium-sized firms to be found. Against this background, one can see that efforts to diversify might have rather more resonance in our region than elsewhere.

In the context of efforts to diversify an economy, well designed institutions can help in several ways:

• They can provide market information, especially about new export opportunities (for example, embassies could do this);

• They can improve flows of technical knowledge and the ability to use it (through higher education, R&D activities—both public and private, manpower training);

• They can facilitate easy entry and exit of firms to and from the market, and support restructuring efforts for those established firms that have a viable future;

• Institutions to develop, plan, and upgrade infrastructure (for example, transport links, port and airport facilities, border crossings, telecoms, energy supplies, factory and office space, etc.);

• Provision of credit and other financial services, through a competently regulated banking system and financial markets;

• Simple, clear regulatory framework, with stable rules, covering such matters as competition policy, health and safety aspects of production, regulation of technical standards and product quality, service standards and customer guarantees, etc.;

• Simple, clear, stable tax system, with low tax rates for business.

In the last two points, simplicity and clarity were emphasized, reflecting my view that especially in smaller and less prosperous countries it makes economic sense to adopt policy frameworks that are administratively manageable and less susceptible to corruption and lobbying than more complex frameworks can be. Sometimes this will imply the adoption of simple policies that may not, in a formal economic sense, be strictly efficient or optimal in an ideal world. The point, however, is that the world is not particularly “ideal”, and we have to make practical accommodations to that reality.

Naturally, in any economy, institutional conditions and how effectively particular institutions function are influenced by the prevailing political configuration. Specifically, there are major issues to do with the credibility of the State. For instance, how do we know whether a successful firm will not be taken over by the State? Or whether a failing firm will not be protected unfairly due to its political connections? Or whether a regulator will be allowed to perform its tasks without State interference? Or whether banks will be directed to issue credit to firms “officially favoured”? Unfortunately, none of these examples is remote from reality, as numerous cases of all of them can be highlighted across the region studied in this paper. Also, and quite damaging in economic terms, the prevalence of such phenomena encourages entrepreneurs to direct their efforts to seeking State favours rather than towards improving their market position.

Assuming that the State has sufficient credibility to enable a market-type economy to function tolerably well, what should the State do in regard to promoting economic diversification? As a starting point, we can suppose that the State and its various agencies are already providing the basic institutional conditions outlined above, facilitating and supporting private sector economic development. Beyond this, one could then argue quite straightforwardly for doing nothing. Then the market mechanism will be left to stimulate whatever diversification occurs, in the light of perceived opportunities and capabilities. In an already well-diversified economy, with good infrastructure and a good quality workforce, this seems the right approach. In a transition economy there may well be more reason to promote diversification more actively, given the legacies referred to above.

Also, if the economy is poorly diversified and there are perceived to be genuine market failures impeding more diversification (and the costs of maintaining a narrow production base appear to be high), then a more active approach may well prove justified. The difficult question then concerns how exactly to go about designing such an approach, as far as possible without generating economically damaging side-effects? In thinking about this, it seems to me that we should be guided by four key principles:

• First, identify the main market and institutional failures that are preventing diversification from occurring “naturally”, through the normal market mechanism. This is the most difficult issue in the transition economies because of their legacy of missing market institutions and enormous structural imbalances. (On the analysis of this notion of identifying key institutional “bottlenecks”, see Rodrik 2007, chap. 2).

• Second, accept that neither the State nor the private sector can know ex ante which new activities will turn out to be successful in the market—so if government support is offered in some form, direct or indirect, it must be expected and accepted that there will be some failures.

• Third, it is useful to think in terms of forms of partnership between State agencies and the private sector in order to promote selected new activities (this also raises the related issue of “how to select?”).

• Fourth, wherever it is feasible, such partnerships should be based on competition (for example, there can be several independent bids to develop each new proposed activity) and performance (that is to say, it is important to withdraw support rapidly from obviously failing activities).

Implications and Challenges for Economic Policy

Suitable policies to promote and support economic diversification will vary enormously between countries, ranging from inaction (in a large, already well-diversified economy with good institutions) to a variety of active measures (in small, narrowly-based economies with relatively poor and weak institutions). In any event, some of the desirable policies have little to do with diversification per se. For example, completing the process of trade liberalization, simplifying the tax and regulatory system, ensuring that the financial system delivers adequate funding for investment, and improving the infrastructure all need to be done regardless of the question of diversification.

Next, if identifiable market failures are constraining diversification, policies to alleviate these failures are required. In practice this is often the most difficult aspect of the diversification agenda, because it is usually extremely hard to identify with much assurance the relevant failures. For if we assess a particular economy as being insufficiently diversified, it is rarely clear why that is the position, and the possible reasons may or may not indicate a specific market failure. Even if there is a market failure, it might not be remediable by government policy alone, and one must always bear in mind the dangers of government failure. Replacing a known market failure with some form of government failure will not necessarily improve market outcomes, so a degree of caution is advisable.

For transition economies, this issue is especially difficult. We sketch two examples here to illustrate the meaning, both relevant to the diversification “problem”. The first has to do with the legacy of large, unprofitable business enterprises, and the issue of their restructuring; the second has to do with encouraging new business formation. Normally one would expect a turnover of businesses in a well functioning economy, with perhaps two to three per cent of jobs being lost each year (mostly in small and medium businesses), these being replaced by the expansion of the more successful surviving firms, plus the creation of new firms. Hence, over a typical decade of such restructuring, 20–30 per cent of the jobs available at the end would not have existed at the start; more dynamic economies restructure even more rapidly. This sort of restructuring was perceived—rightly—as being especially problematic for many transition economies because the initial position was worse and because there was little or no experience of new business formation (some in the SEE countries, almost none in the CIS).

The initial position was much worse than in a normal market-type economy since, following the price and trade liberalization of early transition and, especially within the CIS, the disruption of established trading patterns, the countries were left with a large stock of essentially non-viable large enterprises. According to some early estimates, this could have accounted for up to half of the entire production in some countries. In such circumstances, it was evidently politically impossible for all these loss-making firms simply to be shut down, though in the longer term this would usually be their fate. In principle, one would have liked to see a programme of gradual closures of these failing firms and the restructuring of those still able to operate commercially, accompanied by vigorous measures to promote new business formation. In practice, several CIS countries faced steeply falling GDPs at least until the mid-1990s and this did not provide a favourable environment for new businesses to start up, and the regulatory conditions were also not very supportive as business-friendly reforms were occurring quite slowly in most of the region. As a result, most Governments—even if they were willing to acknowledge the inefficiency of many of their “legacy” large businesses—were unwilling to force their early closure. Quite the contrary, in fact, as Governments actively kept them going to avoid the massive unemployment that would otherwise have resulted (for understandable political reasons). Now, more than 15 years after the start of transition, many of the large, former State-owned enterprises (SOEs) still survive and the business environment is still not as supportive as it needs to be to foster sufficiently rapid rates of new firm formation. Hence, efforts towards economic diversification have to keep in mind these not too favourable background conditions.

As indicated earlier, it is not possible to design a single policy package suitable for all of the SEE and CIS countries. Given their different characteristics and problems, it makes more sense to consider suitable policies for three main types of country in the SEE and among the CIS, namely:

• Resource rich countries—for example, Kazakhstan, the Russian Federation, Turkmenistan

• Countries that are both large and relatively poor in energy resources—for example, Ukraine

• Countries which are both small and poor in energy resources—for example, Armenia, Bosnia and Herzegovina, Republic of Moldova and the former Yugoslav Republic of Macedonia, etc.

The last category is probably the hardest to deal with, so we focus on that group here. What this group calls for is a mix of policies that should include vigorous export promotion, extensive efforts to improve access to markets, especially with immediate neighbours, and efforts to promote exportable services such as tourism and possibly some forms of health care. In addition, partnerships between the State and the private sector to support new activities, as outlined above, would be helpful.

To elaborate, for a small country to improve its living standards in a sustainable way, its engagement with the world economy needs to grow rapidly, starting with exports. Export promotion supports this through several channels: provision of information about foreign markets and the conditions of access to them; credit lines for exporting businesses, including export credit guarantees to reduce the risks of exporting; organization of trade fairs and exhibitions both at home and in key foreign markets; support for advertising and other marketing activities; support for skills development (including language skills) needed to facilitate exporting.

In addition, countries seriously seeking to expand their exports—and here we mean a large expansion such as a doubling or more, not merely a modest 5–10 per cent rise—need to pay attention to relevant parts of their economic infrastructure and regulatory framework. Regarding the former, port facilities, airports, and border crossings might all need new investment to expand their physical capacity, while regulatory conditions should ensure that exporting is quick and reliable, without needing excessive documentation and administrative checks that can so often provide opportunities for corruption. Likewise, goods should not be held up at borders for “administrative reasons”. Problems of this sort simply ensure that much potentially profitable trade never takes place.

Broadman (2005) argues cogently that both SEE and the CIS countries need what he terms extensive “behind the border” reforms in order to facilitate more rapid growth of foreign trade. What he means by this is that trade policy should not just be seen as a matter of tariff policy and non-tariff barriers, but it includes the issues just noted above, as well as other policies to do with business taxation, the financial system and banking reforms, and so on; in other words, policies that support business activity in general, and hence, exports in particular.

Policies outlined thus far concern what an individual country can do for itself, and this is always important. But to trade successfully, there must be trade partners willing to buy the goods and services on offer. Hence, access to markets is also vital. This is part of what we discussed earlier in connection with the WTO, and then FTAs and Customs Unions. Regardless of the formal institutional arrangements, though, trade partners must be willing to engage in trade without imposing excessive tariffs or other barriers of their own, fiscal or administrative. If neighbouring countries are not very friendly, as is the case for some SEE and CIS countries, then movement in this area can prove difficult. It must nevertheless be pursued relentlessly.

Last, what to promote and how? I have already argued that the State is rarely good at selecting either “good” firms or “good” sectors. Also, our countries still have an inheritance of “bad” firms, most of which are by now privatized. Privatized or not, many of these firms still survive through forms of direct and indirect support that come under the general heading of the “soft budget constraint”; this includes the toleration of delayed tax and social security payments, extension of additional credits, sometimes even direct budgetary subsidies (though these are much less common than they were in the early 1990s). Some of these firms can be made economic through additional investment and the modernization of production technology and/or the product range, perhaps with the help of a foreign partner (FDI), and where possible this should be done. Others should gradually be scaled down and closed. This is a form of diversification unique to the transition economies, especially those that have reformed more slowly than the countries of Central Europe.

Besides dealing with these inherited firms, the key to diversification in small countries lacking natural resources is to free up the business environment to stimulate large increases in the rate of new business formation. It would seem most appropriate to measure success in this area in terms of outcomes, in other words, by asking how many firms there are in the given economy, and what their size distribution looks like. As a rule of thumb, one might think of a fairly well-functioning economy as one with around 50 firms per 1000 people, most of these firms being tiny, most of them surviving for quite short periods, with constant renewal of the stock especially at the smallest end of the size range (some ideas about sensible numbers in SEE are provided in Falcetti and others, 2003, Table 1). While it is easy to set targets for simplifying regulation, taxes, credit conditions and the like to stimulate new business formation, it is advisable to assess the effectiveness of such policies not in terms of the number of pages deleted from the regulations, but in terms of their eventual impact on the stock of firms, and hence, on employment and income generation.

Given rapid rates of firm formation, from time to time clusters of related firms will spring up, either in a given region or a given sector, and sometimes these clusters will achieve export success. When this starts to happen, that is the time for supportive government intervention to be brought into play, with funding for related R&D, marketing, IT systems, workforce training programmes, export development and such activities coming on stream.

Conclusion: Institutions and Diversification

Economic diversification is important for sustained growth, and it is normally brought about through competition and the market mechanism. However, there can be market and institutional failures of various kinds that “lock” a country into a very narrow production pattern. This is especially likely in the context of the transition economies discussed in this chapter, with their unfavourable legacy of inefficient businesses, an extremely skewed size distribution of firms (hardly any small and medium firms at the start of transition), and limited experience of promoting new business formation. In such cases, as we have argued above, active policies can help to overcome market failures, promote institutional development, and hence stimulate more diversification.

It is important that diversification efforts be subject to competition and performance criteria, with little State interference to favour particular firms. For the State is rarely capable of “selecting” good firms to support, and when it tries it is most likely to fall victim to lobbying from politically well-connected owners and managers. Bowing to pressure from such lobbies will rarely yield economically desirable outcomes. Equally, the State is rarely able to “select” good sectors for an economy to diversify into. Rather, it should provide an economic environment that supports business activity in general, and only once a new area shows signs of taking off should it provide some helpful reinforcement as outlined here.

Diversification is a normal part of a successful, sustained path of economic growth, so this implies that the right general conditions for growth—discussed in the second and third sections of this chapter—need to be in place. This is true even in the transition economies of SEE and the CIS, though we have acknowledged earlier that their backlog of restructuring makes diversification both more urgent and more difficult than in more “normal” economies. That said, the ensuing growth should then raise incomes and living standards generally, and hence also reduce poverty.