Abstract
The bulk of evidence on the lack of international risk sharing is based on regressions of idiosyncratic consumption growth on idiosyncratic output growth. This paper argues that the results from such regressions obtained from international data are, however, not directly comparable to those based on regional data: the standard practice of running such regressions on international data fails to account for persistent international differentials in consumer prices, whereas—implicitly—most of the literature based on regional data has accounted for these differences. When risk sharing regressions are set up in conceptually the same way in international and regional data sets, the estimated coefficients are also very similar. To explore this result further, we adapt the variance decomposition of Asdrubali et al. (Q J Econ 111:1081–1110, 1996) to allow for deviations from purchasing power parity across countries. While quantity (income and credit) flows are the dominant channel of risk sharing among regions, relative consumption and output price (internal terms of trade) fluctuations account for the bulk of the deviation from the complete markets outcome in international data. To the extent that persistent differences in consumer prices are an indication of goods market segmentation, our findings provide empirical evidence for the proposition by Obstfeld and Rogoff (NBER Macroeconomics Annual 2000, 2000) that segmented international goods markets rather than asset market incompleteness may account for the (apparent) lack of risk sharing between countries.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
See e.g. Becker and Hoffmann (2006), who examine the contribution of capital income and credit flows to risk sharing at different horizons.
We have nothing to say about welfare implications. Clearly, it will be welfare enhancing if there are no transport costs and if prices equalize, even though this may entail more flows of capital (and ultimately shipment of goods). Our interest here is in the optimality (or otherwise) of risk sharing given the structure of goods markets, not in assessing the welfare implications of the respective structure.
Indeed, we find that the correlation between consumption and real exchange rates (i.e. inflation differentials) is very low even in regional data, even though there is wide agreement in the literature that there is quite a lot of risk sharing at the regional level.
The term \(P_{t}^{k}/CPI_{t}^{k}\), the internal terms of trade, gives the value of a country’s output in terms of its consumption bundle.
See their windows help file at http://www.bea.gov/bea/regional/gsp/OnlineHelp.chm
In virtually all studies, the nominal exchange rate is kept fixed by transferring quantities in a base year into a common currency denomination using base year nominal exchange rates. Clearly, since all regressions are in first differences, the choice of this exchange rate is of no practical relevance.
For convenience, we will often refer to this paper as ‘ASY.’
We note, however, that they are not analytically the same: in ASY and Sørensen and Yosha (1998), the variable with respect to which income and consumption are smoothed is ΔlogY, in our setup it is \(\Delta \log {\left[ {P_{\$ } Y} \right]}\). We empirically explore the importance of this difference in Section 4.4.
We will also refer to this regression as the quantity-based regression since it does not take account of relative consumer price variability. Furthermore, since for most countries the national GDP price deflator is highly correlated with CPI, regression (3) also the regressor reflects what are virtually pure quantities.
For Australia, the data reveal relative roles of quantity and price channels that are comparable to what we have obtained from international data. While this is an interesting result, we note two things: first, our sample for Australia is rather short. Second, to obtain measure of the regional GDP deflator, we had to use an experimental volume chain index for real state-level GDP. The Australian Bureau of Statistics issues a note of caution regarding the use of this series. We would therefore not overemphasize this particular result.
To the extent that trade eventually eliminates price differentials, we should expect the role of the price channel in international data to decline in the long-run, quite in line with a growing literature that suggests that purchasing power parity may ultimately hold. Following Becker and Hoffmann (2006), I therefore examined the role of relative price variability at long horizons by performing the variance decomposition suggested above in the levels of the variables instead of first differences. As discussed in this earlier paper, this regression constitutes a long-run panel relation in the sense of Phillips and Moon (1999). Hence, even though the individual time series may be non-stationary and may not necessarily be integrated, there is no risk of spurious regression. The results of this exercise provide strong support for the interpretation above: in the long-run relative price fluctuations play a much smaller role for risk sharing. In international data, I now estimate β price =0.05. Though still significant, (t-statistics: 2.37) this is much smaller than the corresponding β price estimated from first differences in Table 2. Conversely, quantity flows keep up quite well in the levels specification and the ex ante channel even gains in importance. (β inc =0.10 (tstat=2.54) and β cons =0.23 (tstat=4.59)). I thank an anonymous referee for suggesting this exercise.
At a theoretical level, our approach could be justified by a model in which output consists of intermediate inputs that are highly tradeable internationally. These outputs can then either be traded or be transformed into an imperfectly tradeable consumption good.
I thank George von Fuerstenberg for suggesting this course of analysis.
Indeed, market incompleteness alone may not even be sufficient to rationalize the correlations between consumption and real exchange rates that is typically found in the data. Baxter and Crucini (1995) have demonstrated that the equilibria in complete market economies are almost identical to those of a bonds-only economy, unless shocks get very persistent. As argued by Corsetti et al. (2004), it may therefore be rather difficult to generate realistic correlations between real exchange rates and consumption through market incompleteness alone.
References
Asdrubali P, Sørensen PE, Yosha O (1996) Channels of interstate risk sharing: United States 1963–90. Q J Econ 111:1081–1110
Backus D, Kehoe P, Kydland F (1992) International real business cycles. J Polit Econ 100(4):745–775
Backus D, Smith G (1993) Consumption and real exchange rates in dynamic exchange economies with nontraded goods. J Int Econ 35:297–316
Baxter M, Crucini MJ (1995) Business cycles and the asset structure of foreign trade. Int Econ Rev 36(4):821–853
Becker SO, Hoffmann M (2006) Intra- and international risk sharing in the short run and the long run. Eur Econ Rev 50(3):777–806
Cochrane JH (1991) A simple test of consumption insurance. J Polit Econ 99(5):957–976
Cole J, Obstfeld M (1991) Commodity trade and international risk sharing: how much do financial markets matter? J Monet Econ 28:3–24
Corsetti G, Dedola L (2002) The macroeconomics of international price discrimination. ECB Working Paper No. 176
Corsetti G, Dedola L, Leduc S (2004) International risk-sharing and the transmission of productivity shocks. ECB Working Paper No. 308
Crucini M (1999) On international and national dimensions of risk sharing. The Review of Economics and Statistics, Febuary 1999, 81(1):73–84
Del Negro M (2002) Asymmetric shocks among U.S. states. J Int Econ 56:273-297
Engel C (1999) Accounting for U.S. real exchange rate changes. J Polit Econ 107:507–38
Hess GD, Shin K (1998) Intranational business cycles in the United States. J Int Econ 44:289–313
Kraay A, Ventura J (2002) Trade integration and risk sharing. Eur Econ Rev 46(6)
Lane PD, Milesi-Ferretti GM (2004) International investment patterns. Mimeo, Trinity College Dublin and International Monetary Fund
Mace B (1991) Full insurance in the presence of aggregate uncertainty. J Polit Econ 99(5):928–956
Mélitz J, Zumer F (1999) Smoothing of regional shocks within countries and lessons for EMU. Carnegie-Rochester Conference series on Public Policy 51:149–188
Newey W, West KD (1987) A simple, positive definite, heteroskedasticity and autocorrelation consistent covariance matrix. Econometrica 55(3):703–708
Obstfeld M, Rogoff K (2000) The six major puzzles in international macroeconomics: is there a common cause? In: Bernanke K, Rogoff BS (eds) NBER Macroeconomics Annual 2000. MIT, Cambridge, MA
Phillips PCB, Moon HR (1999) Linear regression limit theory for nonstationary panel data. Econometrica 67(5):1057–1111
Rogoff KR (1996) The purchasing power parity puzzle. J Econ Lit 34:647-66
Sørensen BE, Yosha O (1998) International risk sharing and European monetary unification. J Int Econ 45:211–38
Sørensen BE, Yosha O (2007) Producer prices versus consumer prices in the measurement of risk sharing. Appl Econ Q 53(1):1–17
Stockman AC, Tesar L (1995) Tastes and technology in a two-country model of the business cycle: explaining international comovements. Am Econ Rev 85(1):168–185 (Mar)
Townsend RM (1994) Risk and insurance in Village India. Econometrica 62:539–591
Acknowledgements
I would like to thank Mike Artis, George von Fuerstenberg, and an anonymous referee as well as seminar participants at the ESRC programme conference ‘Understanding the Evolving Macroeconomy’ at University College, Oxford, at the Deutsche Bundesbank and at the University of St. Andrews.
Author information
Authors and Affiliations
Corresponding author
Additional information
This paper is part of the project B6: The International Allocation of Risk in the framework of SFB 475 funded by the Deutsche Forschungsgemeinschaft. The UK Economic and Social Research Council funded early stages of the research that led to this paper under its ‘Evolving Macroeconomy’ scheme (grant no. L138251037).
Appendices
Appendix
Regional data sources
Australia: All data are from the Australian Bureau of Statistics and are available at the state level. The CPI data are the CPIs of the respective eight capital cities. Consumption and output are obtained from the breakdown of state level GDP by expenditure and are mid-year estimates (June), ranging from 1990–2002. Income is real gross state domestic income, 1992–2002.
Canada: The data are from Statistics Canada. The data series are personal income, retail sales, population, GDP and CPI by province and range from 1981–2002.
Germany: All data are from the Statistisches Bundesamt, at the federal state level for all 16 federal states. The data range is 1990–2002.
Italy: We used the REGIO-IT data set from the Centro di Ricerche Economiche Nord Sud (CRENoS) at University of Cagliari. The data range from 1960–1996.
Rights and permissions
About this article
Cite this article
Hoffmann, M. The Lack of International Consumption Risk Sharing: Can Inflation Differentials and Trading Costs Help Explain the Puzzle?. Open Econ Rev 19, 183–201 (2008). https://doi.org/10.1007/s11079-007-9024-x
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11079-007-9024-x
Keywords
- International consumption risk sharing
- Capital flows
- Trading cost
- New open economy macroeconomics
- Home bias