Abstract
Time series filtering methods such as the Hodrick–Prescott (HP) filter, with a consensual choice of the smoothing parameter, eliminate the possibility of identifying long swing cycles (e.g., Kondratieff type) or, alternatively, may distort periodicities that are in fact present in the data, giving rise, for example, to spurious Kuznets-type cycles. In this paper, we propose filtering Maddison’s time series for the period 1870–2010 for a selection of developed countries using a less restrictive filtering technique that does not impose but instead estimates the cutoff frequency. In particular, we use unobserved component models that optimally estimate the smoothing parameter. Using this methodology, we identify cycles of periods, primarily in the range of 4–7 years (Juglar-type cycles), and a number of patterns of cyclical convergence. We analyze the historical processes underlying this last empirical finding: Peacetime periods, monetary arrangements, trade and investment flows, and industrial boosts are confluent forces driving the economic dynamism. After 1950, we observe a common business cycle factor that groups all economies, which is consistent with the consolidation of the so-called second globalization.
Similar content being viewed by others
Notes
Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Italy, Japan, the Netherlands, Norway, Sweden, the UK, and the USA. In what follows, we add Spain and Switzerland.
Discarding these types of results is trivial, as a sample of 140 observations would scarcely allow finding three complete cycles within a 50-year period.
However, the Box–Jenkins methodology, based on differencing to achieve stationarity, eliminates long-term dynamics and obscures cyclical dynamics.
An example is Cendejas and Font (2015), in which the price series of Hamilton have been modeled and analyzed to obtain estimations of the common cyclical content of Spanish historical inflation.
In this paper, no attempt is made to discuss the statistical work; it is known that the database employed could influence the empirical results. Comparisons are always problematic and depend on the quality of the data (see Zarnowitz 1992).
As Demeulemeester and Diebolt (2011: 2) suggest, it is important to reintroduce history and historicity in metric analysis. In this paper, we have attempted to combine both sides through a dynamic perspective, and we present an estimation method with an historical overview.
The orthogonal (or classical) factor model is standard; see, for example, Tsay (2005: 426-429).
Alternatively, \(a_{t}\) and \(a_{i,t}\) could follow autoregressive processes if some cyclical persistence is present. In this case, Eqs. (3d) and (3e) would be \(\phi (L)a_{t} = \eta_{t}\) and \(\phi_{i} (L)a_{i,t} = \eta_{i,t}\) with \(\phi (L)\) and \(\phi_{i} (L)\) being the respective autoregressive polynomials.
Some periods close to 2 years are influenced by the noisy content of the original data that pass into the signal. Consequently, the signal-to-noise ratio is high. For \(q > 16\), the frequency exceeds \(\pi\) and the period would not be observable (Nyquist frequency).
http://www.nber.org/cycles.html. Regarding the average duration of business cycles for 13 developed countries, Bergman et al. (1998) obtain some different results. This is a consequence of both the different filtering methodology (they used a band-pass filter that imposes a range of duration of the business cycle of between 2 and 8 years) and the databases employed.
As Sáiz (2014) has noted, the Netherlands, Switzerland, and Denmark were patentless economies in nineteenth century that benefited from relaxed IPR regimes and thus became leaders in specific sectors.
The Spanish agriculture sector lost thousands of migrants to the Americas, but simultaneously received a considerable amount of FDI from France and Germany (Puig and Castro 2009), although these investments favored foreign over domestic firms.
At the turn of the century, the second energy revolution led to an electricity crisis, causing a new wave of mergers, this time among the main electricity companies between 1902 and 1903 (Flamant and Singer-Kerel 1971: 45). USA orders began to decrease in 1906, and a crisis in the German textile sector emerged in 1908 followed by a crisis in the machinery industry in 1909, prompting an obvious contraction of German foreign trade and declining reserves at the Reichsbank. However, German production would focus on foreign markets, and the country held a surplus in its trade balance until 1913.
French agriculture had been subject to foreign competition since 1880 due to the decline in the cost of transport. Grain prices were not the only ones suffering declines. France shifted from being an exporter of wine to being an importer of 10 million hectoliters in 1890 despite attempts to prevent the phylloxera blight in the early 1850s. Agricultural setbacks were partly offset by the growth in the cattle industry and the rise in the consumption of meat (Lavisse and Rambaud 1901: 767).
Banque de Lyon et de la Loire, Union Générale.
Bulletin Financier, p. 812. Revue des Sciences et des Lettres, 1/01/1889.
Sherman Silver Purchase Act.
This was the so-called “Rich man’s panic” (1903–1904), which originated in the steel trust and affected the mining and steel industry. In 1907, the urgent need for means of payment, stock accumulation, and the scarcity of gold prompted a currency crisis that forced USA banks to suspend payments (Flamant and Singer-Kerel 1971: 48).
The depression in the UK that emerged during the crisis of 1882 would last until 1886–1887. This was followed by the collapse of Barings bank in 1890 due to its role as the financial agent for the poorly economically performing Argentine Republic. The economy suffered again with the 1907 USA crisis, particularly affecting industry, that led to major strikes in the textile sector.
Germany and the USA were affected by the crisis that started with the crack of Vienna on May 8, 1873 as a consequence of speculation, rising costs and declining corporate profitability. From Central Europe, the crisis moved to the Atlantic and reached the USA in September 1873. The depression lasted until 1879. The industrial sector suffered markedly. Even the UK faced large bankruptcies, 13,130, only in 1879. Prices and wages dropped (Flamant and Singer-Kerel 1971).
“International trade is perhaps the most important form of engagement with the world economy” (Nayyar 2009: 14).
Nordic countries, in addition to their cultural proximity, had a late and rapid industrialization based on institutional reforms that eliminated restrictions on business, innovation, and credit (banking system). They combined rich natural resources such as forests, ore deposits, fishing, land, and oil with a late integration into the globalization process, in addition to mergers and acquisitions between large firms and, after WWII, the expansion of the public sector and welfare system (Henning et al. 2011). The impressive progress made the Nordic countries “an overachiever” (O’Rourke and Williamson 1995: 8), although there were differences between countries, Sweden being the country that made the most rapid transition.
The Spanish Civil War (1936–1939) may have affected this result.
Technology transfer of military origin to the civilian sector (Fernández-de-Pinedo and Muñoz 2014) was very important, as in the case of aviation, antibiotics, industrial restructuring, the explosives industry and fertilizers.
As Singleton noted, geography changed. France received Alsace and Lorraine from Germany, but “Poland was created out of land formerly belonging to the German, Russian and Austro-Hungarian Empires. The heartland of the Austro-Hungarian Empire was divided into the independent nations of Austria, Hungary and Czechoslovakia. Russia was stripped of Lithuania, Latvia, Estonia and Finland …” (Singleton 2007).
The Swedish banking crisis of 1990–1994, and the strong monetary policies and structural reforms that government employed to recover, should not be forgotten.
From the mid-1980s until 2007, the gradual reduction in inflationary trends in the industrialized world was referred to as the “Great Moderation” thanks to the reduction in the volatility of GDP growth in Australia, Canada, the USA, the UK, Germany, Japan, France, and Italy (Summers 2005).
References
Adelman I (1965) Long cycles—fact or artifact? Am Econ Rev 60:443–463
Baldwin RE, Martin Ph (1999) Two waves of globalization: superficial similarities, fundamental differences, vol 6904. NBER, Cambridge
Bergman UM, Bordo MD, Jonung L (1998) Historical evidence on business cycles: the international experience. In: Fuhrer JC, Schuh S (eds) Beyond shocks: what causes business cycles? Federal Reserve Bank of Boston, Boston, pp 65–113
Bolt J, van Zanden JL (2013) The first update of the Maddison Project; re-estimating growth before 1820. Maddison Project URL: http://www.ggdc.net/maddison/maddison-project/publications/wp4.pdf
Bolt J, van Zanden JL (2014) The Maddison Project: collaborative research on historical national accounts. Econ Hist Rev 67(3):627–651
Bouvier J (1964) Les deux scandales de Panama. Gallimard-Julliard, Paris
Bouvier J (1995) Initiation au vocabulaire et aux mécanismes économiques contemporains, XIXème–XXème siècles. Sede, Paris
Box GEP, Jenkins G (1970) Time series analysis, forecasting and control. Holden Day, San Francisco
Bruland K, Smith K (2010) Knowledge flows and catching-up industrialization in the Nordic countries: the roles of patent systems. In: Odagiri H, Goto A, Sunami A, Nelson RR (eds) Intellectual property rights, development, and catch-up. Oxford University Press, Oxford, pp 63–95
Cendejas JL, Font C (2015) Convergence of inflation with a common cycle: estimating and modelling Spanish historical inflation from the 16th to the 18th centuries. Empir Econ. doi:10.1007/s00181-014-0840-8
Cendejas JL, Castañeda JE, Muñoz FF (2014) Business cycle, interest rate and money in the euro area: a common factor model. Econ Model. doi:10.1016/j.econmod.2014.08.001
Chang HJ (2008) Bad Samaritans: the guilty secrets of rich nations and the threat to global prosperity. Random House, London
Cogley T, Nason JM (1995) Effects of the Hodrick–Prescott filter on trend and difference stationary time series: implications for business cycle research. J Econ Dyn Control 19(1–2):253–278
Craig L, García-Iglesias C (2010) Business cycles. In: Broadberry S, O’Rourke KH (eds) The Cambridge economic history of modern Europe, vol II. Cambridge University Press, Cambridge
Darné O, Diebolt C (2004) Unit roots and infrequent large shocks: new international evidence on output. J Monet Econ 51:1449–1465
David T, Mach A (2007) Institutions and economic growth: the successful experience of Switzerland, 1870–1950. In: Chang Han-Joon (ed) Institutional change and economic development. United Nations University Press, Tokyo, pp 219–238
Demeulemeester JL, Diebolt C (2011) Renouveler la science économique néoclassique? Prendre l’historicité au sérieux, AFC WP2011-3
Diebolt C (2014) Kuznets versus Kondratieff. An essay in historical macroeconometrics. Cahiers d’Econ Polit (Pap Polit Econ) 67:81–117. doi:10.3917/cep.067.0081
Donzé P-Y, Marti L (2014) La Suisse et l’essor de l’industrie japonaise. SZG/RSH/RSS 64(1):45–65
Durbin J, Koopman SJ (2001) Time series analysis by state space methods. Oxford University Press, Oxford
Eichengreen B (1992) Golden fetters: the gold standard and the great depression 1919–1939. Oxford University Press, Oxford
Epstein RJ (1987) A history of econometrics. North Holland, Amsterdam
Feenstra RC, Inklaar R, Timmer MP (2013) The next generation of the Penn world table. www.ggdc.net/pwt
Feinstein CH, Temin P, Toniolo G (2008) The World Economy between the World Wars. Oxford University Press, Oxford
Fernández de Pinedo E (2012) ¿Una recesión asimétrica de nuevo tipo? Rev Astur Econ (RAE) 46:47–60
Fernández-de-Pinedo N, Muñoz FF (2014) Economía y conflicto: urgencias del presente y lecciones del pasado. Cuad Econ (Span J Econ Finance) 37(105):159–168. doi:10.1016/j.cesjef.2014.09.002
Flamant M, Singer-Kerel J (1971) Crises et recessions économiques. PUF, Paris
Foreman-Peck J (2007) European historical economics and globalization. J Philos Econ I(1):23–53
Gómez V (2001) The use of Butterworth filters for trend and cycle estimation in economic time series. J Bus Econ Stat 19(3):365–373
Gonzalo J, Muñoz FF, Santos DJ (2013) Using a rate equations approach to model world population trends. Simulation. doi:10.1177/0037549712463736
Granger CWJ (1966) The typical spectral shape of an economic variable. Econometrica 34(1):150–161
Harvey AC (1989) Forecasting structural time series models and the Kalman filter. Cambridge University Press, Cambridge
Harvey AC (2010) The local quadratic trend model. J Forecast 29(1–2):94–108
Harvey AC, Jaeger A (1993) Detrending, stylized facts, and the business cycle. J Appl Econom 8:231–247
Harvey AC, Trimbur TM (2003) General model-based filters for extracting cycles and trends in economic time series. Rev Econ Stat 85(2):244–255
Hendry DF, Morgan MS (1995) The foundations of econometric analysis. Cambridge University Press, Cambridge
Henning M, Enflo K, Andersson FNG (2011) Trends and cycles in regional economic growth: how spatial differences shaped the Swedish growth experience from 1860–2009. Explor Econ Hist 48(4):538–555
Hodrick RJ, Prescott EC (1997) Postwar U.S. business cycles: an empirical investigation. J Money Credit Bank 29(1):1–16
Howrey EP (1968) A spectrum analysis of the long-swing hypothesis. Int Econ Rev. doi:10.2307/2525477
Kindleberger ChP (1978) Manias, panics, and crashes: a history of financial crises. Macmillan, London
Kindleberger ChP (2006) A financial history of Western Europe. Routdledge, New York
King RG, Rebelo ST (1993) Low frequency filtering and real business cycles. J Econ Dyn Control 17(1–2):207–231
Kitagawa G, Gersch W (1996) Smoothness priors analysis of time series. Springer, Berlin
Kuznets S (1930[1967]) Secular movements in production and prices. Their nature and their bearing upon cyclical fluctuations. Boston, Houghton Mifflin Co (Reprints of Economic Classics. Augustus M. Kelley, New York)
Kuznets S (1961) Capital in the American economy: its formation and financing. National Bureau of Economic Research, Princeton University Press, New York
Lavisse E, Rambaud A (1901) Histoire Générale du IVème siècle à nos jours. Le monde contemporain, 1870–1900, vol 12, A. Colin & Cie, Paris
Maddison A (2001) Monitoring the World Economy: a millennial perspective. OECD, Paris
Maddison A (2007) Fluctuations in the momentum of growth within the capitalist epoch. Cliometrica. doi:10.1007/s11698-007-0007-3
Maddison A (2009) World population, GDP and per capita GDP, A.D. 1–2003. www.ggdc.net/maddison
Metz R (2010) Filter-design and model-based analysis of trends and cycles in the presence of outliers and structural breaks. Cliometrica. doi:10.1007/s11698-009-0036-1
Metz R (2011) Do Kondratieff waves exist?. Cliometrica, How time series techniques can help to solve the problem. doi:10.1007/s11698-010-0057-9
Milanovic B (2005) Can we discern the effect of globalization on income distribution? Evidence from household surveys. World Bank Economic Rev 19(1):21–44
Mills TC (2009) Modelling trends and cycles in economic time series: historical perspective and future developments. Cliometrica. doi:10.1007/s11698-008-0031-y
Mokyr J (2010) The enlightened economy: Britain and the industrial revolution, 1700–1850. Yale University Press, New Haven
Morgan MS (1990) The history of econometric ideas. Cambridge University Press, Cambridge
Morris I (2014) War! What is it good for?: Conflict and the progress of civilization from primates to robots. Farrar, Straus and Giroux, New York
Nayyar D (2009) Developing countries in the World Economy: the future in the past? Paper presented at the UNU-WIDER, Annual lecture 12
Nelson CR (1988) Spurious trend and cycle in the state space decomposition of a time series with a unit root. J Econ Dyn Control 12(2–3):475–488
Nelson CR, Kang H (1981) Spurious periodicity in inappropriately detrended time series. Econometrica 49(3):741–751
Nerlove M, Grether DM, Carvalho JL (1979) Analysis of economic time series. Academic Press, New York
Northrup D (2005) Globalization and the great convergence: rethinking world history in the long term. J World Hist 16(3):249–267
O’Rourke K, Williamson J (1995) Open economy forces and late 19th century Scandinavian catch-up. NBER working paper no 5112
Owyang MT, Piger J, Wall HJ (2005) Business cycle phases in U.S. States. Rev Econ Stat 87:604–616
Pedersen TM (2001) The HP filter, the Slutzky effect and the distortionary effects of filters. J Econ Dyn Control 25(8):1081–1101
Piketty T, Saez E (2006) The evolution of top incomes: a historical and international perspective. Am Econ Rev 96(2):200–205
Puig N, Castro R (2009) Patterns of international business in Spain, 1850–2009. Bus Hist Rev 83(fall):505–537
Roa MJ, Cendejas JL (2007) Crecimiento económico, estructura de edades y dividendo demográfico. WP 390. Centro de Investigación y Docencia Económicas (CIDE), Mexico, DF
Sáiz P (2014) Did patents of introduction encourage technology transfer? Long-term evidence from the Spanish innovation system. Cliometrica 8(1):49–78
Sargent TJ (1979) Macroeconomic theory. Academic Press, New York
Schumpeter JA (1939 [1989]) Business cycles. A theoretical, historical and statistical analysis of the capitalist process. Porcupine Press, Philadelphia (Reprint of the first abridged edition of 1964)
Siegfried A (1937 [1995]) La crise britannique du XXème siècle (Paris, 1937, p 90). In: J Bouvier (ed) Initiation au vocabulaire et aux mécanismes économiques contemporains, XIXème–XXème siècles. Sede, Paris
Singleton J (2007) Destruction ... and Misery: The First World War. In: MJ Oliver, DH Aldcroft (eds) Economic disasters of the twentieth century. Edward Elgar, Cheltenham, pp 9–50
Slutzky E (1937) The summation of random causes as the source of cyclic processes. Econometrica. doi:10.2307/1907241 [Original paper in Russian: 1927]
Stock JH, Watson MW (1989) New indexes of coincident and leading economic indicators. NBER macroeconomics annual. MIT Press, Cambridge, pp 351–394
Stock JH, Watson MW (1991) A probability model of the coincident economic indicators. In: Lahiri K, Moore GH (eds) Leading economic indicators. New approaches and forecasting records. Cambridge University Press, New York, pp 63–85
Summers PM (2005) What caused the great moderation? Some cross-country evidence. Fed Reserve Bank Kansas City Econ Rev (Q III):5–32
Tsay RS (2005) Analysis of financial time series, 2nd edn. Wiley, New York
Whittle P (1983) Prediction and regulation by linear least-squares methods, 2nd edn revised. Oxford, Blackwell
Williamson J (1996) Globalization, convergence and history. J Econ Hist 56(June):277–306
Wold H (1938) A study in the analysis of stationary time series. Almqvist & Wiksell, Stockholm
Young PC (1984) Recursive estimation and time series analysis. Springer, Berlin
Yule GU (1927) On a method of investigating periodicities in disturbed series with special reference to Wolfer’s sunspot numbers. Philos Trans R Soc Lond Ser A 226:267–298
Zarnowitz V (1992) Business cycle: theory, history, indicators and forecasting. Studies in business cycles, vol 27. University of Chicago Press for the NBER, Chicago
Acknowledgments
We acknowledge research funding from Instituto de Investigaciones Económicas y Sociales Francisco de Vitoria (Grant #6-2015); Spanish National Plan for Scientific and Technical Research and Innovation grants CSO2014-53293-R and HAR 2012-35965/His; and the research group IT807-13.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix 1: Spectral gains of the filters for the unobserved components
The spectral gain of a filter measures the increase in amplitude of any specific frequency component of a time series. It is obtained by the Wiener–Kolmogorov (WK) formula (Whittle 1983). To this end, we depart from (1a), in which the signal \(\mu_{t} = \frac{{\eta_{t - 2} }}{{(1 - L)^{2} }}\). The WK filter (of a doubly infinite realization of a time series) that provides the minimum mean-squared error of the signal is given by the ratio of the autocovariance generating functions of the signal \(\mu_{t}\) and the series \(y_{t}\). For the trend component, the filter is
where \(L^{ - 1}\) is the forward operator (\(L^{ - k} y_{t} = y_{t + k}\)) and the convention \((1 - L)(1 - L^{ - 1} ) = \left| {1 - L} \right|^{2}\) is adopted. The spectral gain of the filter of \(\mu_{t}\) is obtained by doing \(L = e^{ - i\omega }\) in (6), where \(i = \sqrt { - 1}\) is the imaginary number and \(\omega\) the frequency, obtaining
For the growth component \(g_{t}\), (1a) is expressed as \(y_{t + 1} = \frac{{g_{t} }}{1 - L} + \varepsilon_{t + 1}\) with the signal \(g_{t} = \frac{{\eta_{t - 1} }}{1 - L}\). The resulting WK filter is
for which the spectral gain is
This gain has a maximum at the frequency
For example, for \(q = \lambda^{ - 1} = \left\{ {0.001,\, \, 0.01, \, \,0.1,\, \, 1,\, \, 10} \right\}\), the corresponding periods \(p = \frac{2\pi }{{\omega_{\hbox{max} } }}\) are \(p = \left\{ {35.3, \, 19.8, \, 11.0, \, 6.0,\;2.9} \right\}\) units of time.
For the acceleration component, (1a) is expressed as \(y_{t + 2} = \frac{{a_{t} }}{{(1 - L)^{2} }} + \varepsilon_{t + 2}\) with the signal \(a_{t} = \eta_{t}\); then, the WK filter is
In addition, in the frequency domain,
The HP filter is the optimal filter when the trend follows an IRW (King and Rebelo 1993). In the context of HP filtering, the cycle (growth cycle) is defined as the deviation with respect to the trend \(\varepsilon_{t} = y_{t} - \mu_{t}\) and the corresponding WK filter is
A comparison of (7), (9), (12), and (13) allows the verification of
with \(G_{\text{HP}} (\omega )\) the spectral gain of the HP filter given \(q\).
Appendix 2: State space representations of the univariate and the multivariate models
State space representation consists of two equations. The measurement equation relates the observed variable with the unobserved components and the observation noise. For IRW model (1), which basically coincides with Eq. (1a),
The state transition equation represents the dynamics of the unobserved components. According to Eqs. (1b) to (1d), this is
Gaussianity and orthogonality assumptions of the error terms imply that \(\varepsilon_{t} \sim N\left( {0,\sigma_{\varepsilon }^{2} } \right)\) and \(\xi_{t} \sim N\left( {0_{3 \times 1} ,Q} \right)\), where \(\xi_{t} = \left[ {\begin{array}{*{20}c} 0 \\ 0 \\ {\eta_{t} } \\ \end{array} } \right]\) and \(Q = \left[ {\begin{array}{*{20}c} 0 & 0 & 0 \\ 0 & 0 & 0 \\ 0 & 0 & {\sigma_{\eta }^{2} } \\ \end{array} } \right]\). By doing \(H = \left[ {\begin{array}{*{20}c} 1 & 0 & 0 \\ \end{array} } \right]\), \(\beta_{t} = \left[ {\begin{array}{*{20}c} {\mu_{t} } \\ {g_{t} } \\ {a_{t} } \\ \end{array} } \right]\), and \(F = \left[ {\begin{array}{*{20}c} 1 & 1 & 0 \\ 0 & 1 & 1 \\ 0 & 0 & 0 \\ \end{array} } \right]\), the state space representation of the system (15)–(16) in compact form is
Estimation of the vector of variances \(\left\{ {\sigma_{\varepsilon }^{2} ,\sigma_{\eta }^{2} } \right\}\) is obtained by maximizing the likelihood function of the one-step ahead prediction errors (Harvey 1989; Durbin and Koopman 2001).
For the multivariate IRW model (3) and for illustrative purposes, let us assume two time series, \(i = 1, \, 2\), with specific acceleration components \(a_{1,t}\) and \(a_{2,t}\), and that the common acceleration component, \(a_{t}\), follows an autoregressive model such as \(a_{t} = \phi a_{t - 1} + \eta_{t}\). From Eq. (3a), we have that the measurement equation relating observed variables with unobserved components is
According to Eqs. (3b) to (3e), the state transition equation is
Gaussianity and orthogonality assumptions of the error terms imply that \(\varepsilon_{t} \sim N\left( {0_{2x1} ,R} \right)\) and \(\xi_{t} \sim N\left( {0_{7x1} ,Q} \right)\), where \(\varepsilon_{t} = \left[ {\begin{array}{*{20}c} {\varepsilon_{1,t} } \\ {\varepsilon_{2,t} } \\ \end{array} } \right]\), \(\xi_{t} = \left[ {\begin{array}{*{20}c} {0_{4x1} } \\ {\eta_{t} } \\ {\eta_{1,t} } \\ {\eta_{2,t} } \\ \end{array} } \right]\), \(R = \left[ {\begin{array}{*{20}c} {\sigma_{{\varepsilon_{1} }}^{2} } & 0 \\ 0 & {\sigma_{{\varepsilon_{2} }}^{2} } \\ \end{array} } \right]\), \(Q = \left[ {\begin{array}{*{20}c} {0_{4 \times 4} } & {0_{4 \times 1} } & {0_{4 \times 1} } & {0_{4 \times 1} } \\ {0_{1 \times 4} } & 1 & 0 & 0 \\ {0_{1 \times 4} } & 0 & {\sigma_{{\eta_{1} }}^{2} } & 0 \\ {0_{1 \times 4} } & 0 & 0 & {\sigma_{{\eta_{2} }}^{2} } \\ \end{array} } \right]\), and \(0_{m \times n}\) is a \(m \times n\) matrix of ceros. By doing \(y_{t} = \left[ {\begin{array}{*{20}c} {y_{1,t} } \\ {y_{2,t} } \\ \end{array} } \right]\), \(H = \left[ {\begin{array}{*{20}c} 1 & 0 & 0 & 0 & 0 & 0 & 0 \\ 0 & 0 & 1 & 0 & 0 & 0 & 0 \\ \end{array} } \right]\), \(\beta_{t} = \left[ {\begin{array}{*{20}c} {\mu_{1,t} } \\ {g_{1,t} } \\ {\mu_{2,t} } \\ {g_{2,t} } \\ {a_{t} } \\ {a_{1,t} } \\ {a_{2,t} } \\ \end{array} } \right]\), and \(F = \left[ {\begin{array}{*{20}c} 1 & 1 & 0 & 0 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 & {\gamma_{1} } & 1 & 0 \\ 0 & 0 & 1 & 1 & 0 & 0 & 0 \\ 0 & 0 & 0 & 1 & {\gamma_{2} } & 0 & 1 \\ 0 & 0 & 0 & 0 & \phi & 0 & 0 \\ 0 & 0 & 0 & 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 & 0 & 0 & 0 \\ \end{array} } \right]\), the state space representation of the system (18)–(19) in compact form follows the general form (17). Estimation of the vector of parameters \(\left\{ {\sigma_{{\varepsilon_{i} }}^{2} ,\sigma_{{\eta_{i} }}^{2} ,\gamma_{i} ,\phi } \right\}\) is obtained as previously outlined for the univariate model.
Rights and permissions
About this article
Cite this article
Cendejas, J., Muñoz, FF. & Fernández-de-Pinedo, N. A contribution to the analysis of historical economic fluctuations (1870–2010): filtering, spurious cycles, and unobserved component modeling. Cliometrica 11, 93–125 (2017). https://doi.org/10.1007/s11698-015-0135-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11698-015-0135-0