Book
Share
Chapter

3 Introduction to Part I

Author(s):
Andrew Berg, and Rafael Portillo
Published Date:
April 2018
Share
  • ShareShare
Show Summary Details
Author(s)
Andrew Berg and Rafael Portillo 

Our understanding of monetary policy in LICs, as with any economic subject, must be fundamentally empirical. Theory is unlikely to give an unambiguous answer to any important policy question; at most it can help shape how we look at the data. This empirics-first approach captures much of the flavour of the macro profession. For example, Gali (2008) justifies writing a book on the application of the New Keynesian model to monetary policy by emphasizing two empirical results: studies using microeconomic data demonstrating nominal price rigidities, such as Bils and Klenow (2004), and VAR results demonstrating that monetary policy shocks have real effects, e.g., Christiano, Eichenbaum, and Evans (1999).

When we began to work on the topic of this book many years ago, two broad considerations shaped our approach. First, we felt that, at least in the policy circles we were in, there was something of an imbalance in favour of empirics without theory. There was a large number of VAR and regression-based papers, e.g. attempting to estimate money demand in low-income countries, and little attempt to give economic structure to the empirical analysis and hence little in the way of usable results.1

And second, analysts of low-income countries inevitably live in a data-poor environment. There is simply relatively little reliable macroeconomic data. For example, only thirteen of the forty-five SSA countries in the IMF databases have any quarterly GDP data and only five have data on both nominal and real GDP (Botswana, Mauritius, Rwanda, Seychelles, and South Africa). Excluding South Africa, which has consistent nominal and real quarterly GDP data back to 1980, the median span of quarterly data is less than nine years (Li et al., 2016).

And for the data we do have, measurement error is likely to be unusually large. Absent a direct line to the true data, it is hard to get a clear sense of measurement error. But several indications are telling. Re-basing of GDP resulted in estimates that GDP was higher than previously thought by almost 90 per cent (Nigeria, 2014), 60 per cent (Ghana, 2010), and 30 per cent (Kenya, 2015). That these revisions were so large reflects rapid structural change and the fact that the surveys that provide the basis for the construction of GDP (household and labour market surveys, agricultural and population censuses) are often incomplete and outdated (African Development Bank, 2013). Another perspective comes from the analysis of revisions to GDP estimates even well after the year in question. Ley and Misch (2014) examine the deviation between final estimates of annual real GDP in year t (made in year t+5) and those made by IMF staff in the spring of year t+1, as a proxy for real-time measurement error. They find that the variation of this deviation is about twice as large in LICs as in OECD countries (in both cases excluding resource-rich countries).

Finally, even these relatively short and unreliable time series may overstate the availability of usable data, because frequent regime shifts make it hard to make inferences. For example, when countries such as Uganda, Ghana, and Kenya switch monetary policy regimes, empirical relationships shift.

This empirical Part of the book takes three very different looks at the data. The first, in Chapter 4, sets the basic stage for the rest of the analyses in the book with a purely empirical and descriptive look at the key economic features of countries in the region. This chapter identifies some of the fundamental characteristics of SSA economies, both in terms of their structure and in terms of the basic macroeconomic data. The weaknesses of the data mentioned above, notably the lack of sufficient quarterly data, limit the exercise relative to similar ones conducted for advanced and emerging economies. Nonetheless, some important and striking patterns emerge. Many of its main conclusions, such as the size of the agricultural sector, the importance of supply shocks, and the lack of correlation between the current account deficit and fluctuations in consumption, are the basic stylized facts that motivate much of the rest of the book. For example, Chapter 10 examines the interactions of limited capital mobility and food price shocks, Chapter 11 takes up the importance of food prices, supply shocks, and poverty—and the proximity of so many consumers to subsistence—while Chapter 17 emphasizes the importance of the bank-led financial system.

Turning to the empirics of monetary policy specifically, Chapter 6 resulted from discussions we had with some of our collaborators, notably Peter Montiel, about the implications of the available empirical evidence on monetary transmission in LICs. He and his co-authors have argued that the particular characteristics of LICs, such as the small size of the financial sector, limit the role for monetary policy, pointing to VAR estimates with insignificant coefficients on interest rates, for example.2

We, in some contrast, have generally felt that too much time and energy was being spent on estimates of money demand and on VARS attempting to estimate the effects of monetary policy shocks. In part, our impatience resulted from a view that some of the questions in that literature were not really critical to efforts to improve policy regimes. Our reading of the experience of emerging markets engaged in regime transition was that a precise handle on the transmission mechanism was not available prior to implementation of new regimes. In part, this is because the regime transition itself inevitably changes transmission. In addition, we felt that little of clear policy relevance hangs on the question of whether the coefficient on say the interest rate in the IS curve is 0.2 or 0.4.

We also doubted whether the assumptions implicit in the methodology were valid. With frequent regime shifts and hence short data series, measurement error, and perhaps above all the challenge to identifying monetary policy shocks in regimes in which the central bank targets a mix of interest rates, money aggregates, and the exchange rate, what should we reasonably expect from these VAR methods applied to our questions of interest?

Chapter 6, jointly written with Peter, attempts to give a precise answer to this question in a particular setting. We know the nature of the world (in the form of the data coming from a fairly simple DSGE model of the sort used in several other chapters in this book). We can then in effect hand these data to an econometrician and ask them to implement standard VAR-based estimations, and see whether they can reliably recover the true nature of the data-generating process. Our reading of the exercise is that, under reasonable-for-SSA-LIC assumptions—about say the length of time series and the degree of measurement error—they cannot.

Where does that leave us, empirically? The conclusion to Chapter 6 discusses some possible ways forward with macro time-series analysis. We are sceptical that confident identification of monetary policy will ever be possible in small open economies that pay attention both to interest rates and the exchange rate, at least with the contemporaneous restrictions that characterize most of the literature. We have some hope that identification through sign restrictions may be fruitful, though we place more stock on relying on easily identified shocks, such as to the terms of trade, and examining their interaction with prior known features of the regime, such as the existence of a hard peg or a float. Analyses based on micro data, such as Abuka et al. (2015), are clearly useful where the data can be found.

An emphasis on case studies represents another sort of methodological approach. Even for advanced countries, one can question, as in Summers (1991), how much of our basic understanding of monetary policy really results from econometrics per se. Privately if not in print, many economists may agree that the experience of major events—for example the sharp and deep recession that accompanied the Volker disinflation—strongly shapes the way we think about macroeconomics and specifically how econometricians know when to stop running regressions and declare victory.

Chapter 5 takes a close look at another such major event, a moment in 2011 when four countries in the East African Community acted dramatically—some raising interest rates by hundreds of basis points—to tighten monetary policy in the face of high inflation. There are limits to this exercise. First, we cannot fully isolate the exogenous component of the policy shock. And we only have the one incident and the four countries to look at. On the other hand, we can take a relatively rich look at the interplay of external and domestic factors in the build-up of inflation, the tightening episode, and the aftermath, in each of the four cases.

The magnitude of the tightening event, and the variation across the four countries in terms of underling economic structure and monetary policy regime, allow us to draw some important conclusions. First, in at least some of the four countries, after a large policy-induced rise in the short-term interest rate, lending and other interest rates rose, the exchange rate tended to appreciate, output tended to fall, and inflation declined. And second, the cross-country variation in transmission seems to depend sharply on the policy regime in place.

In later Parts of the book, we will exploit the empirical understanding gained in this Part I to address specific policy questions. For example, the calibrations of the models in Chapters 15 and 18 benefit from the close narrative analysis of Chapter 5. More generally, the case studies in Chapters 1519 combine elements of a narrative approach with the calibration of small structural models. But first, what are the basic stylized facts we should have in mind when thinking about monetary policy in Africa?

References

    AbukaC.AlindaR. K.MinoiuC.PeydroJ. L. andPresbiteroA. F. (2015). Monetary Policy in a Developing Country: Loan Applications and Real Effects. IMF Working Paper No. 15/270.

    • Search Google Scholar
    • Export Citation

    African Development Bank. (2013). Situational Analysis of the Reliability of Economics Statistics in Africa: Special Focus on GDP Measurement.

    • Search Google Scholar
    • Export Citation

    BilsM. andKlenowP. (2004). Some Evidence on the Importance of Sticky Prices. Journal of Political Economy112 (5) 94784.

    ChristianoL.EichenbaumM. andEvansM. (1999). Monetary Policy Shocks: What Have We Learned and to What End? In M.Woodford andJ.Taylor (Eds.) Handbook of Macroeconomics1(A). North Holland: Elsevier.

    • Search Google Scholar
    • Export Citation

    DavoodiH. R.DixitS. andPinterG. (2013). Monetary Transmission Mechanism in the East African Community: An Empirical Investigation. IMF Working Paper 13/39. Washington, DC: International Monetary Fund.

    • Search Google Scholar
    • Export Citation

    GaliJ. (2008). Monetary Policy Inflation and the Business Cycle. Princeton University Press.

    LeyE. andMischF. (2014 January). Output data revisions in Low-Income Countries. Conference presentation at the IMF-DFID Conference on Macroeconomic Challenges Facing Low-Income Countries: New PerspectivesWashington, DC.

    • Search Google Scholar
    • Export Citation

    LiG.O’ConnellS.AdamC.BergA. andMontielP. (2016). VAR meets DSGE: Uncovering the Monetary Transmission Mechanism in Low-Income Countries. IMF WP/16/90. Washington, DC: International Monetary Fund.

    • Search Google Scholar
    • Export Citation

    MishraP. andMontielP. (2012). Monetary transmission in low-income countries: effectiveness and policy implications. IMF Economic Review60270302.

    • Search Google Scholar
    • Export Citation

    MishraP.MontielP. andSpilimbergoA. (2013). How Effective is Monetary Transmission in Developing Countries: A Survey of the Empirical Evidence. Economic Systems37 (2) 187216.

    • Search Google Scholar
    • Export Citation

    NsabimanaA. andOcranM. K. (2015). Money demand stability and inflation prediction in the EAC countries: Burundi, Kenya, Rwanda, Tanzania, and Uganda. Studies in Economics and Econometrics397198.

    • Search Google Scholar
    • Export Citation

    SummersL. H. (1991). The Scientific Illusion in Empirical Macroeconomics. The Scandinavian Journal of Economics93 (2) 12948.

    Other Resources Citing This Publication