The performance of emerging markets during the fed's easing and tightening cycles: A resilience analysis across economies

Aizenman, Joshua; Park, Donghyun; Qureshi, Irfan A.; Saadaoui, Jamel; Uddin, Mohammed Gazi Salah Working Paper The performance of emerging markets during the fed's easing and tightening cycles: A resilience analysis across economies ADB Economics Working Paper Series, No. 735 Provided in Cooperation with: Asian Development Bank (ADB), Manila Suggested Citation: Aizenman, Joshua; Park, Donghyun; Qureshi, Irfan A.; Saadaoui, Jamel; Uddin, Mohammed Gazi Salah (2024) : The performance of emerging markets during the fed's easing and tightening cycles: A resilience analysis across economies, ADB Economics Working Paper Series, No. 735, Asian Development Bank (ADB), Manila, https://doi.org/10.22617/WPS240365-2 This Version is available at: https://hdl.handle.net/10419/301975 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/3.0/igo/ ASIAN DEVELOPMENT BANK ADB ECONOMICS WORKING PAPER SERIES NO. 735 August 2024 THE PERFORMANCE OF EMERGING MARKETS DURING THE FED’S EASING AND TIGHTENING CYCLES A RESILIENCE ANALYSIS ACROSS ECONOMIES Joshua Aizenman, Donghyun Park, Irfan A. Qureshi, Jamel Saadaoui, and Gazi Salah Uddin ASIAN DEVELOPMENT BANK The ADB Economics Working Paper Series presents research in progress to elicit comments and encourage debate on development issues in Asia and the Pacific. The views expressed are those of the authors and do not necessarily reflect the views and policies of ADB or its Board of Governors or the governments they represent. ADB Economics Working Paper Series Joshua Aizenman, Donghyun Park, Irfan A. Qureshi, Jamel Saadaoui, and Gazi Salah Uddin No. 735 | August 2024 Joshua Aizenman (aiz[email protected]) is the Dockson Chair in Economics and International Relations at the University of Southern California. Donghyun Park ([email protected]) is an economic advisor at the Economic Research and Development Impact and Irfan A. Qureshi ([email protected]) is a public management specialist in the Sectors Group, Asian Development Bank. Jamel Saadaoui ([email protected]) is a full professor of economics at the Université Paris 8. Gazi Salah Uddin ([email protected]) is an associate professor at Linköping University. The Performance of Emerging Markets During the Fed’s Easing and Tightening Cycles: A Resilience Analysis Across Economies Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) © 2024 Asian Development Bank 6 ADB Avenue, Mandaluyong City, 1550 Metro Manila, Philippines Tel +63 2 8632 4444; Fax +63 2 8636 2444 www.adb.org Some rights reserved. Published in 2024. ISSN 2313-6537 (print), 2313-6545 (PDF) Publication Stock No. WPS240365-2 DOI: http://dx.doi.org/10.22617/WPS240365-2 The views expressed in this publication are those of the authors and do not necessarily reflect the views and policies ofthe Asian Development Bank (ADB) or its Board of Governors or the governments they represent. ADB does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. The mention of specific companies or products of manufacturers does not imply that they are endorsed or recommended by ADB in preference to others of a similar nature that are not mentioned. By making any designation of or reference to a particular territory or geographic area inthis document, ADB does not intend to make any judgments as to the legal or other status of any territory or area. This publication is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) https://creativecommons.org/licenses/by/3.0/igo/. By using the content of this publication, you agree to be bound bytheterms of this license. For attribution, translations, adaptations, and permissions, please read the provisions andterms of use at https://www.adb.org/terms-use#openaccess. This CC license does not apply to non-ADB copyright materials in this publication. If the material is attributed toanother source, please contact the copyright owner or publisher of that source for permission to reproduce it. ADB cannot be held liable for any claims that arise as a result of your use of the material. Please contact [email protected] if you have questions or comments with respect to content, or if you wish toobtain copyright permission for your intended use that does not fall within these terms, or for permission to use theADB logo. Corrigenda to ADB publications may be found at http://www.adb.org/publications/corrigenda. Notes: In this publication, “$” refers to United States dollars. ADB recognizes “China” as the People’s Republic of China. ABSTRACT We investigate the determinants of the performance of emerging markets (EMs) during five United States (US) Federal Reserve monetary tightening and easing cycles from 2004 to 2023. We study how macroeconomic and institutional conditions of an EM at the beginning of a cycle explain EM resilience during each cycle. More specifically, our baseline cross-sectional regressions examine how those conditions affect three measures of resilience: bilateral exchange rate against the US dollar, exchange rate market pressure, and economy-specific Morgan Stanley Capital International (MSCI) index. We then stack the five cross-sections to build a panel database to investigate potential asymmetry between tightening versus easing cycles. Our evidence indicates that macroeconomic and institutional variables are associated with EM performance, determinants of resilience differ during tightening versus easing cycles, and institutions matter more during difficult times. Our specific findings are largely consistent with economic intuition. For instance, we find that current account balance, international reserves, and inflation are all important determinants of EM resilience. Keywords: monetary policy cycle, emerging markets, resilience, macroeconomic fundamentals, Federal Reserve JEL code: E58 _____________________ An earlier version of this paper was published as NBER Working Paper No. 32303. 1. Introduction The Global Financial Crisis (GFC) terminated the illusive Great Moderation (Blanchard, Dell’Ariccia, and Mauro 2010), which was followed by the United States (US) Federal Reserve’s (Fed’s) alternating tightening and easing cycles shown in Figure 1. The GFC induced 7 years of easing (2007–2014) followed by 4.5 years of tightening (“taper tantrum” years). Subsequently, 3 years of easing induced by the coronavirus disease (COVID-19) pandemic (2019–2022) led to a major tightening since February 2022, a delayed reaction to rapidly rising inflation in the US. Figure 1: Monetary Cycles in the United States, June 2004 to September 2023 COVID-19 = coronavirus disease, Fed = Federal Reserve, GFC = Global Financial Crisis. Sources: Data retrieved from Federal Reserve Bank of Atlanta. Wu-Xia Shadow Federal Funds Rate. https://www.atlantafed.org/cqer/research/wu-xia-shadow-federal-funds-rate, and Federal Reserve Bank of St. Louis. Federal Funds Effective Rate. https://fred.stlouisfed.org/series/FEDFUNDS. The vector autoregression analysis of Rey (2015) vividly illustrated that US monetary policy was a key driver of global financial cycles that affected the leverage of global banks, capital flows, and credit growth in the international financial system. Fed’s easing versus tightening cycles 2 Consequently, the global financial cycles propagated by US shocks and policies constrained the policy options of financially integrated economies. Emerging markets (EMs), in particular, were exposed to “flight to quality” at times of heightened financial instability and “search for yields” when the US Fed’s massive monetary easing in response to GFC pushed the shadow Federal Funds rates toward zero (Bernanke and Reinhart 2004, Wu and Xia 2016). From the perspective of most EMs and developing economies, global financial cycles are exogenous shocks that test their resilience. Our paper investigates the determinants of the relative performance of EMs during the Fed’s monetary tighteningeasing cycles during the past 2 decades. To answer these questions, we investigate how macroeconomic conditions at the outset of each cycle influence the relative performance of emerging economies. Do ex-ante macroeconomic fundamentals explain why some EMs are more resilient than others during monetary cycles? Our baseline cross-sectional regressions examine how macroeconomic variables affect three measures of resilience: bilateral exchange rate against the US dollar, exchange rate market pressure (EMP) (Goldberg and Krogstrup 2023), and economy-specific Morgan Stanley Capital International (MSCI) index. We also include institutional factors as additional determinants of EM resilience. We contribute to the empirical literature on EM performance in the face of US monetary shocks in a number of different ways. First, our selected time period allows us to better identify determinants of EM resilience because it contains big shocks such as GFC, the taper tantrum, and the COVID-19 pandemic and sharp swings in the Fed’s monetary policy. Second, we perform a comparative analysis of the determinants of EM 3 resilience during the Fed’s tightening versus easing cycles. There is no a priori reason why the determinants should be the same between the two different types of monetary policy cycles. Third, we take a deep dive into the potential link between institutions and resilience. Intuitively, sound institutions such as good governance should contribute to resilience.1 Our empirical analysis yields a number of interesting findings. The current account balance is an important determinant of EMP during monetary cycles. Economies with more flexible exchange rate regimes and more developed financial markets experience lower exchange rate market pressures. Less-corrupt economies experienced lower exchange rate market pressure in two out of five cycles. Economies with higher inflation experienced appreciation of their MSCI indexes in three out of five cycles. This was not the case during the GFC and taper tantrum. Larger current account surpluses and international reserves were associated with greater MSCI index appreciation during the three last cycles. During the GFC cycle, larger Net International Investment Positions (NIIP) were associated with better stock market performance. A combination of higher international reserves, higher current account surpluses, and larger net international investment positions helps emerging economies cope better with exchange market pressures, especially during tightening. Financial institution development was associated with inferior performance during the first two tightening cycles—before the GFC and the taper tantrum. This is in line with the conjecture that more financially developed economies were more subject to capital 1 There is no unique definition of resilience. In the following, we will follow Markus Brunnermeier who discusses the concept of resilience as the ability to recover from a shock. In this respect, we will investigate the performance and recovery speed (duration to peak depreciation, for example) during the US monetary cycles. 4 outflows due to “flight to safety.” Economies with less religious tensions saw their financial markets perform better during the taper tantrum cycle. Economies with fewer internal conflicts and stronger law and order suffered a more significant stock market decline during the GFC cycle. A possible interpretation is that greater trust in institutions led to a higher appreciation of stock markets during the Great Moderation. We can similarly explain why economies with better governance experienced worse stock market performance during the tightening before GFC. Economies with better democratic accountability, lower religious tensions, and stronger law and order performed better during the easing cycle triggered by the COVID-19 pandemic. We organize this paper as follows: section 2 reviews the literature, section 3 presents the empirical methodology, section 4 discusses the results, and section 5 concludes. 2. Literature Review Previous literature has examined the impact of the Fed’s monetary policy on EM macroeconomic dynamics. Existing studies also sought to identify the characteristics that explain why the impact of such shocks varies across EMs. For example, Caldara et al. (2023) show that episodes of global tightening are associated with larger economic downturns, worse financial conditions, and a relatively muted decline in inflation. Ahmed et al. (2023) study the role of foreign exchange (FX) reserves in buffering the exchange rate against the US dollar during the 2021–2022 Federal Reserve monetary policy tightening. They distinguish between mechanisms through which FX reserves mitigate currency depreciation. A “balance sheet” channel implies that strong fundamentals linked with large reserves reduce currency risk even without using these reserves to intervene. 11 Table 2: Descriptive Statistics During the Third Cycle (in White) and the Fourth Cycle (in Gray) Obs Obs Mean Mean Median Median SD SD Min Min Max Max Explained variables: DXRcycle_3, 4 117 126 19.12 5.86 17.73 1.11 10.98 9.97 -1.97 -16.24 49.46 39.15 EMPcycle_3, 4 38 36 0.36 2.76 -0.04 2.45 3.12 3.18 -5.75 -2.57 6.51 9.75 MSCIcycle_3, 4 49 50 7.98 19.51 7.53 23.34 21.20 28.03 -35.08 -52.13 71.18 74.04 Explanatory variables: CAB 108 116 -2.63 -2.53 -3.98 -2.94 11.66 9.10 -37.61 -31.83 48.58 39.15 NIIP 90 104 -0.17 -0.23 -0.27 -0.36 0.87 1.01 -3.85 -3.652 3.79 5.43 Gdeficit 113 122 -1.65 -1.34 -2.26 -1.88 5.89 4.75 -16.30 -9.54 33.78 32.15 Gdebt 112 122 44.77 50.54 39.51 46.38 35.88 28.43 0 0 232.4 232.4 CPI 111 116 4.11 3.68 2.95 2.83 4.30 3.68 -4.30 -2.82 36.60 23.56 FUELX 99 104 14.78 14.17 3.60 3.91 24.21 22.20 0 0 99.80 95.56 FUELM 101 108 18.76 14.79 19.04 14.52 9.42 7.49 0.69 0.58 51.05 33.19 kaopen 104 116 0.18 0.19 -0.17 -0.17 1.58 1.52 -1.93 -1.93 2.30 2.30 FI 109 119 0.40 0.42 0.36 0.40 0.22 0.22 0.08 0.08 1 0.97 FM 109 119 0.18 0.19 0.047 0.070 0.25 0.25 0 0 0.87 0.92 extconf 77 85 9.76 9.72 9.92 9.50 1.10 1.05 5.63 6.50 11.50 11.50 corruption 77 85 2.57 2.65 2 2.38 1.21 1.15 1 1 5.50 5.50 demoacc 77 85 4.08 4.12 4 4 1.47 1.36 0.50 0.50 6 6 ethnictens 77 85 3.86 3.91 4 4 1.21 1.13 1 1 6 6 govstab 77 85 7.18 7.15 6.96 6.96 1.25 0.83 4.88 5.83 10.88 9.50 intconf 77 85 8.86 8.87 8.88 8.88 1.48 1.29 5.50 6.21 12 12 laworder 77 85 3.54 3.45 3.50 3 1.26 1.21 1.50 1.50 6 6 milpol 77 85 3.72 3.77 4 4 1.73 1.56 0 0 6 6 reltensions 77 85 4.50 4.50 5 5 1.40 1.38 1 1 6 6 ers 107 116 0.53 0.54 0.46 0.46 0.23 0.24 0.05 0.14 1 1 RESGDP 101 112 24.38 21.06 18.78 18.01 22.02 17.51 1.99 0.37 152.9 117.4 IT 116 125 0.28 0.30 0 0 0.45 0.46 0 0 1 1 Note: We restricted the sample to changes in the bilateral exchange rate between -50% (appreciation) and 50% (depreciation). We use the delta log for the bilateral exchange rates and the MSCI indexes, and the delta for the EMP. We exclude economies with zero exchange rate variation during the period. Statistics for explanatory variables are only displayed for samples in which bilateral exchange rates are used. The names of the variables and the acronyms used in the table are fully described in Appendix A. Source: Authors’ calculations. The fourth monetary cycle mainly overlaps the pandemic crisis. At the beginning of this easing cycle, the Fed fund rate was equal to 2.4% and below 0.1% 36 months later. The descriptive statistics show that the fourth monetary cycle differs from the previous economic cycles. The episodes of financial stress during this period were explained by uncertainty related to the COVID-19 pandemic. In addition, several economies 12 implemented fiscal packages and dollar swap lines to cope with financial turmoil.5 The variation in bilateral exchange rates and the EMP was quite similar to that observed in the second cycle (GFC). However, the developments in the stock markets were different from during the GFC cycle, with a positive evolution on average (Table 2). Table 3: Descriptive Statistics During the Fifth Cycle Observations Mean Median SD Minimum Maximum Explained variables: DXRcycle_5 106 7.54 5.98 9.86 -21.60 46.68 MSCIcycle_5 50 -1.37 -3.59 25.95 -40.92 126.7 Explanatory variables: CAB 93 -2.42 -2.53 8.99 -40.40 25.43 NIIP 88 -0.16 -0.33 1.29 -3.83 5.74 Gdeficit 102 -3.89 -4.57 5.86 -16.42 40.07 Gdebt 102 60.39 55.80 34.30 0 255.1 CPI 96 4.35 3.84 3.87 -0.77 25.75 FUELX 86 13.04 2.95 21.53 0 94.63 FUELM 90 13.28 13.08 8.49 0.51 66.42 kaopen 97 0.18 -0.17 1.52 -1.93 2.30 FI 100 0.45 0.44 0.21 0.082 0.96 FM 100 0.20 0.056 0.27 0 0.92 extconf 71 9.80 10 0.99 7 11.50 corruption 71 2.77 2.50 1.16 1 6 demoacc 71 4.18 4.50 1.41 0.50 6 ethnictens 71 3.93 4 1.10 2 6 govstab 71 7.05 6.92 1.01 4.71 10 intconf 71 9.07 9.21 1.35 6.08 11.96 laworder 71 3.56 3.46 1.13 1.50 6 milpol 71 3.95 4 1.46 1 6 reltensions 71 4.60 5 1.210 1.50 6 RESGDP 88 28.22 24.04 23.61 0.37 134.6 IT 105 0.33 0 0.47 0 1 Note: We restrict the sample for changes in the bilateral exchange rate between -50% (appreciation) and 50% (depreciation). We use the delta log for the bilateral exchange rates and the MSCI indexes, and the delta for the EMP. We removed economies with zero exchange rate variation during the period. Statistics for explanatory variables are only displayed for samples in which bilateral exchange rates are used. The EMP data are not available for the entire period during the fifth cycle. The names of the variables and the acronyms used in the table are fully described in Appendix A. Source: Authors’ calculations. During the last monetary cycle of our study in Table 3, the Fed fund rates moved from nearly zero in February 2022 to more than 5% at the end of our sample in September 2023. During this monetary cycle, the bilateral exchange rate against the dollar 5 Aizenman, Jinjarak, and Park (2011) have shown that international reserves holding and swap lines may be complements rather than substitutes. Choi et al. (2022) describe how the new Foreign and International Monetary Authorities (FIMA) Repo Facility has extended access to dollar liquidity during the pandemic. 13 depreciated in most economies, averaging 7%. Ahmed et al. (2023) showed that economies with more ex-ante international reserves have limited their depreciation rate.6 The average level of international reserves is now at 28%. This may partially indicate that economies continuously accumulate reserves to buffer the shocks of external finance (Aizenman et al. 2024). We will come back later on this point in the empirical results section. The developments in the financial markets were not similar to those of previous cycles, with almost no variation on the average of the MSCI indexes. 3.2. Methodology We will use first cross-sectional regressions where the explanatory variables would be fundamentals observed before the events, and the left-hand variable would be the performance of the financial variable of interest over the monetary cycle: , j i j i j FinVar c X        where each 𝑖 denotes a particular economy. We use multiple financial indicators to build the dependent variable in alternative specifications, with the change in each indicator represented by Δ measuring financial performance during the monetary cycle. 𝑋, are a set of explanatory variables, 𝑗 specific to economy 𝑖 measured in the year prior to the monetary cycle, 𝛽 are parameters to be estimated, and 𝜀 are error terms. Note that the cross-section observations in each regression are the economies, and a separate regression is run for each dependent variable and each subset of explanatory variables 𝑗. 6 Coulibaly et al. (2024) confirm the buffer effect of international reserve holdings on the exchange rate and public debt for 54 African economies. Exposure to the Belt and Road initiative will be explored when more comprehensive data will be available on public debt for African economies. Recently, the People’s Republic of China has become “an international lender of last resort” as shown by Horn et al. (2023). 14 Following Ahmed, Coulibaly, and Zlate (2017), we analyze economic performance on a cross-sectional basis and include the initial macroeconomic and institutional conditions at the beginning of each cycle. Possible candidates for the initial conditions include stock variables, including the ratio of initial international reserves to GDP, public debt in local or foreign currency as a percentage of GDP, private debt as a percentage of GDP, and other variables. In the spirit of Alvarez and De Gregorio (2014), we will examine the changing patterns of resilience, comparing the performance of IT and fixed exchange rates in economies.7 Examining the heterogeneity of the performance of emerging economies during these monetary cycles can help policymakers build policy space to cope with future cycles. We will identify the asymmetries during monetary easing and monetary tightening. These asymmetries may provide useful information to policymakers about excessive leverage during monetary easing, since monetary easing associated with underregulated leverage growth may increase macroeconomic vulnerability in the next cycle. 4. Empirical Results 4.1. Baseline Regressions Tables 4 to 9 present the results of the cross-sectional regression for the bilateral exchange rate variation, the variation of the EMP indexes and the MSCI indexes variation, respectively, during the different monetary cycles.8 As explained in subsection 3.2, our main objective is to explain the difference across economies in the performance and 7 A natural extension will be to control for crisis dummies, as in Laeven and Valencia (2020), and for the history of crises (possibly by discounting past crises, in line with the diminishing effects of more distant crises relative to the more recent crises). 8 The pairwise correlation between variables is below 50% in almost all cases. In all the regressions, the null hypothesis of normality for the residuals is not rejected at conventional significance levels. 15 resilience during monetary cycles and especially tightenings, according to ex-ante macroeconomic fundamentals and ex-ante institutional variables. We may briefly recall identifying several key determinants of economic performance, and resilience will help us to provide sound policy recommendations to cope with international financial spillovers. In Table 4, we have the full specification of the macroeconomic and institutional determinants of economic performance. Furthermore, we use a stepwise backward stepwise selection with a threshold value of 20% for the p-value in Table 5. We can observe that the explanatory power ranges from 41% to 68% according to the R-squared values throughout Tables 4 and 5. We can note that the negative coefficient on the international reserves holding indicates that the buffer effect of international reserves holding is confirmed for three cycles out of five. This finding generalizes the results of Ahmed et al. (2023) and is in line with those of Aizenman et al. (2024). The holding of international reserves has stabilization properties on the exchange rate through both the balance sheet channel and the intervention channel. Indeed, Ahmed et al. (2023) show that currency interventions were associated with less exchange depreciation when the ex-ante stock of high reserve was high during the fifth cycle. Furthermore, economies with higher values for ex-ante consumer price inflation have experienced larger depreciation rates during three cycles out of five. In light of purchasing power parity theory, these last results may reveal that the exchange rate depreciation follows the price differentials over the medium run. According to Rose (2020), the success of the IT regime was explained by its performance in terms of resilience to external finance shocks and, especially in terms of limiting the risk of currency crisis. Obviously, as Rose recalled, an economy cannot be 16 forced to quit an IT regime contrary to a fixed-exchange rate regime. As mentioned by Aizenman, Jinjarak, and Park (2011), emerging economies have followed a mixed strategy for their nominal anchor. However, the policy response to exchange rate depreciations to limit imported inflation was more constrained for economies without an IT regime. Consistent with these results, before the GFC-induced monetary cycle, being an inflation targeter before entering the cycle was associated with lower exchange rate depreciation. Table 4: Cross-Sectional Regressions for Bilateral Exchange Rate Variation Fed tightening I June 2004 – June 2007 Fed easing I July 2007 – May 2014 Fed tightening II June 2014 – Dec 2018 Fed easing II Jan 2019 – Jan 2022 Fed tightening III Feb 2022 – Sep 2023 Variables DXRcycle_1 DXRcycle_2 DXRcycle_3 DXRcycle_4 DXRcycle_5 CAB - 0.0491 - 0.5273 0.0136 0.4738 - 0.2804 (0.3527) (0.3144) (0.1547) (0.3245) (0.3050) RESGDP - 0.1915 - 0.4073* - 0.1300* 0.0018 - 0.1678** (0.2430) (0.2295) (0.0656) (0.0836) (0.0725) NIIP - 1.1563 7.5767 0.6836 0.8957 4.2667* (6.8544) (5.9002) (2.0270) (2.9972) (2.1240) GDeficit 1.3754** 0.4080 - 0.9368*** - 0.8932 - 0.1079 (0.5132) (0.5337) (0.3313) (0.6850) (0.5820) GDebt 0.1168 - 0.0080 - 0.0644** - 0.0174 0.0029 (0.0693) (0.1267) (0.0297) (0.0357) (0.0730) CPI - 1.0157** 2.2739** 1.1370 0.5376 1.0398* (0.4351) (1.0917) (0.7073) (0.9455) (0.5615) FUELX - 0.3972*** - 0.1280 0.1581** 0.1182 0.0554 (0.0981) (0.1246) (0.0610) (0.0791) (0.1007) FUELM - 0.3648 0.1702 - 0.4457*** 0.3219 0.3254 (0.2330) (0.3372) (0.1269) (0.2140) (0.3874) kaopen 0.7338 0.1242 1.5685 - 0.9405 - 1.3074 (1.8463) (2.4704) (1.3086) (1.0478) (1.8299) ers 0.7903 - 0.0464 - 28.9637*** - 15.7212* - (10.3988) (12.4007) (7.6404) (8.7798) - IT - 17.4864*** 2.9200 - 5.3057 2.0732 - 4.5101 (4.6234) (8.8328) (3.4863) (3.1628) (3.8294) FI - 15.1873 - 42.7909 - 14.2118 - 20.3465 - 3.9456 (14.6961) (26.2050) (16.1468) (13.4327) (19.4381) FM 13.1264 20.5382 1.4946 4.5167 11.0268 (8.8948) (17.6239) (11.2466) (10.3034) (10.6137) extconf 0.3786 - 0.1408 2.4214* - 0.6076 1.5194 (2.1177) (2.8750) (1.2865) (1.4293) (2.2957) corruption - 2.1792 - 2.9961 0.3348 0.1313 - 0.9060 (2.1895) (3.6574) (1.7957) (1.9988) (2.9516) demoacc - 3.2898* 0.3203 - 0.2944 - 0.3151 - 1.4629 (1.8045) (2.9979) (1.8813) (1.1534) (1.6047) ethnictens - 1.6153 - 1.3520 0.9598 0.1180 - 0.9540 (1.7483) (2.3873) (1.3244) (1.3412) (1.4753) govstab - 1.1201 3.4935 2.2261** - 2.1061 0.4353 (1.4419) (2.3805) (0.8986) (2.1042) (2.4553) intconf - 3.6389** 0.4547 - 1.2601 3.7247** - 1.1172 (1.6633) (2.5126) (1.5396) (1.5460) (2.3078) laworder 1.2774 1.0166 4.0290* - 1.8976 3.6529 (2.1212) (3.1074) (2.1503) (1.6117) (2.4170) milpol 5.2856** 4.0691 - 2.9054 0.0177 1.6214 (2.3200) (3.2339) (1.8190) (1.5324) (1.6667) Continued on the next page 17 Fed tightening I June 2004 – June 2007 Fed easing I July 2007 – May 2014 Fed tightening II June 2014 – Dec 2018 Fed easing II Jan 2019 – Jan 2022 Fed tightening III Feb 2022 – Sep 2023 Variables DXRcycle_1 DXRcycle_2 DXRcycle_3 DXRcycle_4 DXRcycle_5 reltensions - 0.1242 - 1.1499 1.3748 2.1241 - 0.5773 (1.5682) (2.3884) (1.2340) (1.3700) (1.8601) Constant 53.4824** - 17.5125 7.5123 - 1.9851 - 8.3556 (26.1837) (37.7146) (17.6057) (25.4320) (28.1996) Economies 61 63 58 65 54 R - squared 0.5192 0.4899 0.6790 0.4735 0.4991 RMSE 12.73 17.23 8.614 9.762 10.71 Note: *** p<0.01, ** p<0.05, * p<0.1. Robust standard errors are in parentheses. Data for the index of exchange rate stability (ers) are not available for the fifth cycle. Bold indicates a significance level below 5%. The names of the variables and the acronyms used in the table are fully described in Appendix A. Source: Authors’ calculations. Two points can be mentioned to assess the respective influence of inflation targeting during these monetary cycles. The first one is the distinction between de jure inflation targeters and de facto inflation targeters. Indeed, this distinction may be crucial as some economies declare to be inflation targeters, but constantly miss the inflation target, Türkiye being a prime example. The second point related to the performance of inflation targeters is the distinction between “young” and “old” inflation targeters. One possible conjecture would be that the dynamics gains in terms of resilience increase with time and with the credibility of the IT regime (de jure versus de facto). Table 5: Cross-Sectional Regressions for the Bilateral Exchange Rate Variation —Backward Stepwise Selection Fed tightening I June 2004–June 2007 Fed easing I June 2007–May 2014 Fed tightening II June 2014–Dec 2018 Fed easing II Jan 2019–Jan 2022 Fed tightening III Feb 2022–Sep 2023 Variables DXRcycle_1 DXRcycle_2 DXRcycle_3 DXRcycle_4 DXRcycle_5 CAB - 0.3738 0.4898** (0.2285) (0.1977) RESGDP - 0.4075** - 0.1087** - 0.1600*** (0.1821) (0.0431) (0.0572) NIIP 6.4131 2.5039* (4.6179) (1.2762) G d eficit 0.9412** - 0.8511*** - 0.8569 (0.3764) (0.1704) (0.5811) G d ebt - 0.0612** (0.0284) CPI - 1.0617*** 2.6812*** 1.0046** 1.0877** (0.3166) (0.8072) (0.4786) (0.4387) FUELX - 0.2556*** - 0.1390 0.1687*** 0.1285* (0.0650) (0.0925) (0.0524) (0.0748) FUELM - 0.4159*** 0.2806* (0.0993) (0.1625) kaopen 1.3469 - 2.6850** (0.9512) (1.1547) Continued on the next page 18 Fed tightening I June 2004–June 2007 Fed easing I June 2007–May 2014 Fed tightening II June 2014–Dec 2018 Fed easing II Jan 2019–Jan 2022 Fed tightening III Feb 2022–Sep 2023 Variables DXRcycle_1 DXRcycle_2 DXRcycle_3 DXRcycle_4 DXRcycle_5 ers - 27.3594*** - 17.6173** - (6.6048) (7.3684) - IT - 16.3697*** - 4.9236* - 4.5660 (3.6481) (2.8583) (3.2522) FI - 11.3020 - 49.2099*** - 13.5700 - 19.5798** (8.4619) (14.7421) (10.0757) (8.1516) FM 24.8325** 10.0849 (10.1394) (6.1635) extconf 2.0494* (1.0478) govstab 2.6784* 2.1304*** (1.5746) (0.7742) intconf - 2.9984** 2.4470** (1.3623) (1.0436) laworder 4.2115** - 1.7746 3.3111** (1.6513) (1.1315) (1.4346) milpol 2.5130 3.8811* - 3.2809** (1.5286) (2.1165) (1.3117) reltension s 1.3420 2.2711* (0.9919) (1.3185) Constant 30.0433*** - 17.7384 3.4317 - 9.1127 - 4.5227 (11.2073) (15.0958) (13.7423) (11.3732) (6.9015) Economie s 61 63 58 65 54 Rsquared 0.4109 0.4561 0.6653 0.4295 0.4362 RMSE 11.93 15.45 8.029 8.880 9.474 Note: *** p<0.01, ** p<0.05, * p<0.1. Robust standard errors are in parentheses. We use a backward stepwise selection procedure for the variables. Variables with p-values above 20% are sequentially removed from the model from the highest to the lowest p-value. Data for the index of exchange rate stability (ers) are not available for the fifth cycle. Bold indicates a significance level below 5%. The names of the variables and the acronyms used in the table are fully described in Appendix A. Source: Authors’ calculations. In light of this possible complementarity between IT regimes and fixed-exchange rate regimes (Aizenman, Jinjarak, and Park 2011), we can note that less-flexible exchange rate regimes played an important role during the taper tantrum and the pandemic monetary cycles. Indeed, we found that the exchange rate depreciation was more limited in economies with higher ex-ante exchange rate stability.9 This empirical evidence shows that the relative merits of IT and flexible exchange rate regimes vary over time. The stabilizing properties of these different regimes may evolve over the different monetary cycles. We may also suspect the presence of non-linearities. 9 The data for the Exchange Rate Stability are not available during the fifth cycle. 19 Tables 6 and 7 show that the explanatory power for EMP regression ranges from 30% to 80%.10 The EMP index considers the interdependence between bilateral exchange rates, foreign exchange intervention, and policy rate changes. As fully described by Goldberg and Krogstrup (2022), the EMP index can be seen as a comprehensive exchange rate policy index. The weights for bilateral exchange rates, foreign exchange intervention (FXI), and policy rate changes are framed in a model of supply and demand for foreign currency: “Any given excess supply or demand for a currency—an international capital flow pressure—can be offset by an equivalent amount of FXI, or by an endogenous exchange rate movement or change in the domestic monetary policy rate sufficient to generate an offsetting private balance of payments flow” (Goldberg and Krogstrup 2022). Consequently, the EMP index can capture dimensions of international financial spillovers other than simple bilateral exchange rates. We find that the current account balance is now an important determinant of EMP variations during monetary cycles. An ex-ante current account surplus can offer more room for maneuvering intervention during the monetary cycle, especially during tightening, to cope with flight–to–quality movements. We observe that economies with less flexible exchange rate regimes and more developed financial markets experience less exchange rate market pressures. In light of the previous discussion on the relative merits of IT regimes and less flexible exchange rate regimes, we found that exchange rate stability is associated with fewer exchange rate pressures in three cycles out of five.11 10 The data for the EMP indexes are not available during the fifth cycle. 11 In Appendix C, we provide further evidence for the GFC cycle with estimates before and after the Zero Lower Bound (ZLB). 20 For the institutional variables, economies with higher levels of corruption rating (less corruption) experience less exchange rate market pressure in two cycles out of five. The difference between financial institution development and financial market development can provide interesting insights. More developed financial markets help to cope with pressures. Besides, financial institution development is associated with higher pressures. The influence of institutional variables depends on the monetary cycle. There is a larger, significant positive association during the GFC. This may reveal that institutional variables may play a more important role during large recessions and episodes of acute financial stress.12 Table 6: Cross-Sectional Regressions for Exchange Rate Market Pressure Variation Fed tightening I June 2004 – June 2007 Fed easing I July 2007 – May 2014 Fed tightening II June 2014 – Dec 2018 Fed easing II Jan 2019 – Jan 2022 Variables EMPcycle_1 EMPcycle_2 EMPcycle_3 EMPcycle_4 CAB 0.3713* 0.0836 -0.3309* -0.0282 (0.1799) (0.1378) (0.1796) (0.3315) RESGDP -0.1087* -0.0498 0.0432 -0.0200 (0.0515) (0.0406) (0.0583) (0.0568) NIIP -0.6710 0.8588 0.6654 -0.0347 (2.3040) (1.3172) (1.4141) (1.9426) Gdeficit 0.1481 0.0676 0.5243 0.3575 (0.2196) (0.1735) (0.3412) (0.4882) Gdebt 0.0523* 0.0109 -0.0088 0.0094 (0.0286) (0.0180) (0.0214) (0.0186) CPI 0.0196 -0.0788 0.2381 -0.0797 (0.2302) (0.3778) (0.3083) (0.6533) FUELX -0.1174* -0.0146 0.0068 0.0558 (0.0644) (0.0509) (0.0375) (0.0618) FUELM -0.2733* 0.1326 0.0031 -0.0497 (0.1435) (0.0773) (0.1012) (0.1375) kaopen -0.2601 0.3427 -0.8433 -0.2918 (0.6273) (0.4036) (0.8519) (1.0997) ers -7.9386 2.1804 - 11.0477** -7.5097 (5.5385) (1.9166) (3.9744) (7.2901) IT -0.7695 2.7046 -1.3262 -1.8400 (1.8337) (2.0800) (2.4821) (2.4837) 12 This may be illustrated by the famous Warren buffet’s quote: “A rising tide floats all boats…only when the tide goes out do you discover who’s been swimming naked.” The role of institution quality may be hidden during monetary easing. Large episodes of financial and economic stress may reveal the importance of good institutions. Continued on the next page 27 For the sake of completeness, we compute two other measures of resilience in Tables 10 to 13. First, we compute the number of months required to reach peak depreciation in Tables 10 and 11. Second, we compute the number of months required to reach the lowest point in the equity MSCI index. Interestingly, we can note that being an inflation targeter is associated with a reduction of the number of months necessary to reach peak depreciation. Table 10: Cross-Sectional Regressions for the Time to Peak Depreciation Fed tightening I June 2004 – June 2007 Fed easing I July 2007 – May 2014 Fed tightening II June 2014 – Dec 2018 Fed easing II Jan 2019 – Jan 2022 Fed tightening III Feb 2022 – Sep 2023 Variables Time to peak Time to peak Time to peak Time to peak Time to peak CAB - 0.2388 - 0.2302 0.3737 0.6307* - 0.2173* (0.2835) (0.6236) (0.2347) (0.3156) (0.1131) RESGDP - 0.1536 - 0.3083 0.0044 - 0.0501 - 0.0286 (0.1576) (0.3016) (0.0976) (0.0767) (0.0311) NIIP - 3.2158 2.7131 - 4.9210 - 1.4721 1.9001** (5.6698) (9.8985) (3.6595) (3.5774) (0.7126) GDeficit 0.6372* 0.0502 - 1.3299*** - 0.5002 - 0.2012 (0.3510) (1.0017) (0.4794) (0.4893) (0.1884) GDebt 0.0735 - 0.0447 - 0.0284 0.0737 0.0394*** (0.0609) (0.1544) (0.0397) (0.0445) (0.0136) CPI - 0.4048 3.0272* 2.0127*** - 0.2736 - 0.0538 (0.3744) (1.7529) (0.6763) (0.7903) (0.1779) FUELX - 0.3216*** - 0.0902 0.0736 0.0971 0.0650* (0.0721) (0.2149) (0.0843) (0.0824) (0.0325) FUELM 0.0226 0.4610 - 0.4080** 0.2401 0.2197* (0.2098) (0.5283) (0.1615) (0.1681) (0.1289) kaopen 0.6249 - 0.4955 3.7523** - 0.2401 0.4681 (1.3546) (4.5624) (1.6468) (1.1427) (0.7155) ers - 2.2830 - 5.8338 9.6727 - 0.6674 - (8.1690) (20.3693) (11.3125) (8.1057) - IT - 11.5537*** - 5.1764 - 10.4217** 0.7126 - 3.2028** (3.7298) (12.4129) (4.9832) (4.0237) (1.5067) FI - 3.5057 - 33.1146 - 35.0651* - 13.0231 - 24.3129*** (13.5738) (33.9516) (18.5355) (12.4183) (7.6996) FM 4.7021 4.1894 26.4078 - 2.4457 9.8223** (8.1876) (27.2824) (15.7361) (10.0080) (4.3537) extconf 0.4897 0.7264 5.8255*** - 0.9171 0.2587 (1.6652) (5.7660) (2.0160) (1.1867) (0.7664) corruption 1.0499 - 1.1221 - 0.0719 0.4075 - 1.0554 (1.5023) (6.8295) (2.2254) (1.8276) (1.2397) demoacc - 0.7533 - 0.0100 - 0.0401 - 0.3379 - 0.0063 (1.2961) (4.6775) (2.0792) (1.5184) (0.8287) ethnictens - 0.7372 - 2.3304 5.6386** - 0.6786 - 0.9138 (1.3614) (3.5010) (2.2287) (1.4736) (0.6863) govstab 1.0961 3.7426 - 0.4158 - 0.8970 - 0.2965 (1.0678) (3.8607) (1.5859) (1.8688) (0.8522) intconf - 2.7116* 2.2516 - 2.7251 2.3921 1.4027 (1.3505) (4.3328) (2.2602) (1.8858) (0.9287) laworder 0.7363 - 0.0337 1.9157 - 3.0690 2.1886* (1.6690) (5.6417) (2.2942) (1.8994) (1.1832) milpol 2.3705 3.1691 - 4.9643** 0.4039 - 0.7327 (1.8068) (4.7407) (2.0659) (1.7084) (0.9447) Continued on the next page 28 Fed tightening I June 2004 – June 2007 Fed easing I July 2007 – May 2014 Fed tightening II June 2014 – Dec 2018 Fed easing II Jan 2019 – Jan 2022 Fed tightening III Feb 2022 – Sep 2023 Variables Time to peak Time to peak Time to peak Time to peak Time to peak reltensions - 0.9752 - 1.6965 - 2.6950 1.2264 0.0366 (1.2173) (3.1028) (1.6372) (1.6887) (0.7456) Constant 28.7056 - 7.6027 11.9135 19.5518 2.6473 (19.2843) (61.6435) (27.8384) (22.5573) (12.3957) Economies 61 63 58 65 54 R - squared 0.5281 0.4305 0.6909 0.4290 0.6246 RMSE 9.675 27.33 11.20 9.937 4.186 Note: *** p<0.01, ** p<0.05, * p<0.1. Robust standard errors are in parentheses. Data for the index of exchange rate stability (ers) are not available for the fifth cycle. Bold indicates a significance level below 5%. The names of the variables and the acronyms used in the table are fully described in Appendix A. Source: Authors’ calculations. Table 11: Cross-Sectional Regressions for the Time to Peak Depreciation —Backward Stepwise Selection Fed tightening I June 2004 – June 2007 Fed easing I July 200 7 – May 2014 Fed tightening II June 2014 – Dec 2018 Fed easing II Jan 2019 – Jan 2022 Fed tightening III Feb 2022 – Sep 2023 Variables Time to peak Time to peak Time to peak Time to peak Time to peak CAB - 0.2702* 0.3761* 0.2866** - 0.2068** (0.1470) (0.2053) (0.1354) (0.0912) RESGDP - 0.1762 - 0.2395 - 0.1356*** (0.1196) (0.1643) (0.0403) NIIP - 4.3497* 1.6602*** (2.3746) (0.4719) GDeficit 0.6965** - 1.1297*** (0.3140) (0.3252) GDebt 0.0895* - 0.0403 0.0567 0.0435*** (0.0470) (0.0302) (0.0362) (0.0137) CPI - 0.4556 4.5736*** 1.9812*** (0.2754) (0.9328) (0.5672) FUELX - 0.3082*** 0.0658 0.0418 (0.0501) (0.0498) (0.0260) FUELM - 0.4225*** 0.2316* 0.2027* (0.1447) (0.1371) (0.1025) kaopen 3.5350** (1.5329) ers IT - 10.5781*** - 12.3277*** - 2.2297* (2.6523) (3.6755) (1.2010) FI - 31.8136* - 23.9074*** (17.0352) (5.5763) FM 25.8872* 9.4063** (13.2514) (3.7760) extconf 5.2883*** (1.8081) corruption - 1.1381 (0.8563) demoacc ethnictens 5.7422*** - 0.8586 (1.9121) (0.5215) govstab 1.2208 3.5157* (0.7927) (2.0945) intconf - 2.3717** - 2.9490 1.8431* 1.1384* (1.0411) (1.9165) (0.9994) (0.5873) laworder - 4.0806*** 1.9666** (0.9847) (0.7820) milpol 1.7137 - 3.9584** (1.2388) (1.6129) reltensions - 2.6158 (1.5683) Continued on the next page 29 Fed tightening I June 2004 – June 2007 Fed easing I July 200 7 – May 2014 Fed tightening II June 2014 – Dec 2018 Fed easing II Jan 2019 – Jan 2022 Fed tightening III Feb 2022 – Sep 2023 Variables Time to peak Time to peak Time to peak Time to peak Time to peak Constant 27.0949** - 10.3996 24.7810 12.8733 3.0773 (11.5820) (14.5933) (17.2990) (9.9581) (5.5961) Economies 61 63 58 65 54 R - squared 0.4917 0.3547 0.6700 0.3669 0.5777 RMSE 8.753 23.96 10.56 8.982 3.922 Note: *** p<0.01, ** p<0.05, * p<0.1. Robust standard errors are in parentheses. We use a backward stepwise selection procedure for the variables. Variables with p-values above 20% are sequentially removed from the model from the highest to the lowest p-value. Data for the index of exchange rate stability (ers) are not available for the fifth cycle. Bold indicates a significance level below 5%. The names of the variables and the acronyms used in the table are fully described in Appendix A. Source: Authors’ calculations. Table 12: Cross-Sectional Regressions for the Time to Lowest Point in Equity MSCI Indexes Fed tightening I June 2004–June 2007 Fed easing I July 2007–May 2014 Fed tightening II June 2014–Dec 2018 Fed easing II Jan 2019–Jan 2022 Fed tightening III Feb 2022–Sep 2023 Variables Time to low Time to low Time to low Time to low Time to low CAB 0.0223 - 0.7908 - 0.1937 - 0.2341 - 0.3398 (0.0987) (0.6001) (0.6640) (0.2943) (0.2022) RESGDP - 0.0233 0.4645** 0.0498 - 0.0484 0.0549* (0.0262) (0.1879) (0.2273) (0.0619) (0.0268) NIIP - 0.0680 - 7.1060 0.6245 1.9956 - 0.9646 (0.6370) (5.7911) (6.7688) (1.2628) (0.7312) GDeficit 0.0445 1.4011* 0.1248 - 0.4632 0.3976 (0.0927) (0.7439) (0.8510) (0.4469) (0.3390) GDebt 0.0188* 0.2209 0.0212 - 0.0080 - 0.0037 (0.0099) (0.1566) (0.0971) (0.0232) (0.0205) CPI - 0.2125*** - 2.5890 1.3476 0.4750 - 0.3204 (0.0724) (1.8976) (1.6131) (0.2857) (0.2944) FUELX - 0.0246 - 0.1613 0.0279 - 0.0077 0.1221*** (0.0267) (0.2375) (0.1645) (0.0462) (0.0418) FUELM - 0.0321 0.0949 - 0.2580 0.1802 0.2449 (0.0377) (0.4910) (0.3624) (0.1962) (0.2613) kaopen - 0.4405 4.5719 4.1682 4.0004** 0.1096 (0.2652) (4.8272) (3.4006) (1.4921) (1.0279) ers 0.0080 23.0482 - 1.7117 - 11.7030* - (1.0377) (33.1679) (13.9548) (5.7285) - IT - 0.3083 - 0.4871 3.7220 0.0762 - 2.3469 (0.5955) (15.5205) (7.8657) (3.2819) (2.1356) FI - 0.0343 - 16.2982 23.4025 1.2131 - 6.2578 (1.7473) (22.0011) (21.1437) (10.7162) (10.3266) FM 1.8200 - 21.5946 - 20.0463 - 2.0751 6.5669 (1.7347) (25.7300) (17.2581) (6.0784) (5.3447) extconf 0.2114 - 0.8798 3.3889 4.0840*** 1.2525 (0.2595) (4.2353) (3.9967) (1.1271) (0.9376) corruption - 0.3513 - 4.8573 - 9.0890** - 2.6399* 0.3283 (0.4086) (7.7038) (3.5648) (1.2707) (1.7627) demoacc 0.0723 8.5079** 1.4937 - 1.6102 - 1.3704 (0.2670) (4.0685) (3.1406) (1.1451) (1.0166) ethnictens - 0.0873 0.4309 0.6669 2.2822** - 1.1220 (0.1971) (3.2096) (3.2337) (1.0409) (1.1533) govstab 0.1823 3.6354 5.5896* - 1.0158 - 1.3545 (0.2543) (3.1911) (2.9042) (1.2954) (1.1196) intconf 0.0885 0.1653 - 3.2875 - 3.2481** - 1.5474 (0.3175) (3.9220) (2.7782) (1.3835) (1.2461) laworder 0.2757 4.8829 6.0848 0.8619 - 0.6536 (0.2831) (7.3375) (3.9758) (1.2685) (1.5785) milpol - 0.6391 - 6.5826 - 2.9456 1.1375 1.3178 (0.3889) (5.0103) (5.0351) (1.5377) (1.1767) Continued on the next page 30 Fed tightening I June 2004–June 2007 Fed easing I July 2007–May 2014 Fed tightening II June 2014–Dec 2018 Fed easing II Jan 2019–Jan 2022 Fed tightening III Feb 2022–Sep 2023 Variables Time to low Time to low Time to low Time to low Time to low reltensions 0.4781* 0.3191 - 1.0461 - 3.4048** 0.6221 (0.2602) (3.5573) (3.7719) (1.2089) (0.9777) Constant - 2.2880 - 20.0029 - 21.2901 24.4524 23.0916 (3.2888) (55.8434) (53.1405) (19.7868) (15.1910) Economies 39 44 45 44 44 R - squared 0.6877 0.6056 0.5224 0.7178 0.4998 RMSE 1.099 17.26 12.25 4.607 4.322 Note: *** p<0.01, ** p<0.05, * p<0.1. Robust standard errors are in parentheses. Bold indicates a significance level below 5%. The names of the variables and the acronyms used in the table are fully described in Appendix A. Source: Authors’ calculations. Table 13: Cross-Sectional Regressions for the Time to Lowest Point in Equity MSCI Indexes—Backward Stepwise Selection Fed tightening I June 2004–June 2007 Fed easing I July 2007–May 2014 Fed tightening II June 2014–Dec 2018 Fed easing II Jan 2019–Jan 2022 Fed tightening III Feb 2022–Sep 2023 Variables Time to low Time to low Time to low Time to low Time to low CAB - 0.8373** - 0.2095 (0.3235) (0.1475) RESGDP - 0.0247** 0.6805*** 0.0271* (0.0092) (0.1639) (0.0145) NIIP - 10.6948** (4.0194) GDeficit 1.2377** (0.5110) GDebt 0.0168*** 0.2451*** (0.0044) (0.0810) CPI - 0.2278*** - 2.1930* 1.0777 0.3923*** (0.0590) (1.1278) (0.6548) (0.1382) FUELX 0.0807*** (0.0188) FUELM 0.2111* (0.1134) kaopen - 0.3728** 4.4194** 3.3350*** (0.1662) (2.1028) (1.0150) ers 30.9898*** - 7.2293** (8.5869) (3.1124) IT 4.9606 (2.9654) FI 21.3280 (15.0856) FM 2.1407** - 18.1955 (0.9767) (12.3085) extconf 0.2472 3.3404 4.0007*** (0.1468) (2.1577) (0.9555) corruption - 8.5006*** - 1.3198 (2.2066) (0.7980) demoacc 7.0707** - 1.0376 (3.3529) (0.7141) ethnictens 1.4986* (0.8648) govstab 0.2588* 4.6875*** - 0.9573** (0.1280) (1.5291) (0.4440) intconf - 2.7209 - 3.4705*** - 0.8473 (1.7756) (0.9362) (0.7039) laworder 5.3083* (2.9866) milpol - 0.5127** - 5.8641 - 3.0929* (0.2460) (4.0102) (1.7474) Continued on the next page 31 Fed tightening I June 2004–June 2007 Fed easing I July 2007–May 2014 Fed tightening II June 2014–Dec 2018 Fed easing II Jan 2019–Jan 2022 Fed tightening III Feb 2022–Sep 2023 Variables Time to low Time to low Time to low Time to low Time to low reltensions 0.3956** - 2.4565*** (0.1589) (0.8895) Constant - 3.5330* - 12.2947 - 15.5642 17.2033** 21.0648*** (1.9925) (11.0529) (19.8952) (7.2357) (6.9671) Economies 39 44 45 44 44 R - squared 0.5776 0.5011 0.4967 0.6619 0.2985 RMSE 0.949 15.26 10.27 4.086 3.844 Note: *** p<0.01, ** p<0.05, * p<0.1. Robust standard errors are in parentheses. We use a backward stepwise selection procedure for the variables. The variables with p-values above 20% are sequentially removed from the model starting from the highest to the lowest p-value. Bold indicates a significance level below 5%. The names of the variables and the acronyms used in the table are fully described in Appendix A. Source: Authors’ calculations. 4.2. Panel Data Regressions After exploring cross-sectional regressions, we stack the cross sections to build an unbalanced panel database where the time dimension will be our five cycles, so T = 5 or 4 (depending on data availability). We preserve the chronological structure of the data as the US monetary cycles are observed at the same time for all the economies. Thus, we explore the potential asymmetries between monetary cycles. Thanks to dummy variables for tightening and easing episodes. Figures 2 and 3 present graphical evidence showing that the benefit of having a better score in the government stability variable only appears during tightening by limiting exchange rate depreciation and providing an expansion of the stock market. 32 Figure 2: Asymmetries During Tightening Cycles for the Bilateral Exchange Rate ICRG = International Country Risk Guide. Note: with the data sample of Appendix C for the 5 cycles. The score of Government Stability is observed one year before each cycle. Source: Authors’ calculations. Tables 14 to 16 provide empirical evidence that confirms our preliminary graphical evidence. The benefit of having better government stability only appears during bad times for the exchange rate and the stock market indexes. For the EMP index, the financial institutions variable is associated with an increase of pressures and democratic accountability is associated with a reduction of pressures, in line with the cross-sectional regressions.15 15 In Appendixes D and E, we provide panel evidence for the five cycles in the case of the bilateral exchange rate variations and the MSCI variations during the cycles. 33 Table 14: Panel Evidence for the Bilateral Exchange Rate Tightening cycle s Easing cycle s Variables DXR 4 cycles DXR 4 cycles CAB - 0.3295** -0.3368 (0.1495) (0.2073) tight 41.9752*** (12.9289) c.CAB#c.tight -0.0072 (0.2555) kaopen -1.4127 -1.3132 (1.0452) (1.1562) c.kaopen#c.tight 0.0994 (1.5586) NIIP -1.6504 4.8681* (1.6844) (2.5342) c.NIIP#c.tight 6.5185** (3.0429) FUELM 0.2130 0.4774*** (0.1689) (0.1830) c.FUELM#c.tight 0.2645 (0.2491) Gdebt -0.0663 -0.0305 (0.0491) (0.0418) c.GDebt#c.tight 0.0358 (0.0645) govstab 2.1891* -4.0505*** (1.1590) (0.9436) c.govstab#c.tight - 6.2396*** (1.4945) Constant -7.9835 33.9917*** (9.0842) (9.1997) Economies (max.) 83 83 Observations 247 247 R-squared 0.1989 0.1989 RMSE 16.51 16.51 Note: *** p<0.01, ** p<0.05, * p<0.1. Robust standard errors are in parentheses. We use a backward stepwise selection procedure for the variables. The variables with p-values above 20% are sequentially removed from the model starting from the highest to the lowest p-value. Bold indicates a significance level below 5%. Only the four first cycles are included, as we use the ers variable in the backward stepwise selection procedure. The dummies “tight” and “easy” refer to tightening and easing cycles, respectively. The names of the variables and the acronyms used in the table are fully described in Appendix A. Source: Authors’ calculations. 34 Table 15: Panel Evidence for Exchange Market Pressure Indexes T ight ening cycle s E as ing cycles Variables EMP 4 Cycles EMP 4 Cycles ers -2.3539 -2.9106 (1.7499) (1.9754) tight 1.9320 (4.9594) c.ers#c.tight -0.5567 (2.6389) RESGDP -0.0206 -0.0441* (0.0184) (0.0266) c.RESGDP#c.tight -0.0236 (0.0323) NIIP 0.4262 1.3500 (0.5967) (0.9920) c.NIIP#c.tight 0.9238 (1.1576) Gdeficit 0.0739 -0.0310 (0.0764) (0.1234) c.GDeficit#c.tight -0.1049 (0.1451) demoacc -0.5104 - 1.0196*** (0.3225) (0.3738) c.demoacc#c.tight -0.5093 (0.4937) FM -2.6745 -2.9718 (1.9805) (2.5208) c.FM#c.tight -0.2973 (3.2057) govstab -0.0953 -0.3528 (0.2491) (0.2339) c.govstab#c.tight -0.2574 (0.3417) FI 3.3818 5.6219** (2.8110) (2.7090) c.FI#c.tight 2.2401 (3.9039) ethnictens -0.4969 -0.5011 (0.3301) (0.3339) c.ethnictens#c.tight -0.0042 (0.4695) Constant 8.1967*** 10.1287** (3.0796) (3.8874) Economies (max.) 37 37 Observations 142 142 R-squared 0.2657 0.2657 RSME 3.008 3.008 Note: *** p<0.01, ** p<0.05, * p<0.1. Robust standard errors are in parentheses. We use a backward stepwise selection procedure for the variables. The variables with p-values above 20% are sequentially removed from the model starting from the highest to the lowest p-value. Bold indicates a significance level below 5%. Only the four first cycles are included due to missing data for EMP. The dummies “tight” and “easy” refer to tightening and easing cycles, respectively. The names of the variables and the acronyms used in the table are fully described in Appendix A. Source: Authors’ calculations. 35 Table 16: Panel Evidence for MSCI Indexes T ight ening cycles Easing cycle s Variables MSCI 4 Cycles MSCI 4 Cycles CAB 1.4598** 0.7739 (0.7006) (0.6057) tight -30.0741 (37.9127) c.CAB#c.tight -0.6859 (0.9262) FUELM -0.0846 - 1.6007*** (0.6675) (0.4906) c.FUELM#c.tight -1.5160* (0.8284) FI -27.7682 - 78.4734*** (23.6623) (22.0292) c.FI#c.tight -50.7052 (32.3294) Gdeficit -1.8118 - 2.2064*** (1.3384) (0.7742) c.GDeficit#c.tight -0.3946 (1.5462) govstab -5.5716* 9.2714*** (2.9149) (2.2233) c.govstab#c.tight 14.8430*** (3.6661) corruption 6.1084 2.4394 (4.9762) (3.1911) c.corruption#c.tight -3.6690 (5.9115) ers - 34.5878** -12.2621 (15.0331) (12.0133) c.ers#c.tight 22.3256 (19.2435) Constant 59.3109** 29.2368 (27.1966) (26.4144) Economies (max.) 46 46 Observations 172 172 R-squared 0.4157 0.4157 RMSE 34.78 34.78 Note: *** p<0.01, ** p<0.05, * p<0.1. Robust standard errors are in parentheses. We use a backward stepwise selection procedure for the variables. The variables with p-values above 20% are sequentially removed from the model starting from the highest to the lowest p-value. Bold indicates a significance level below 5%. Only the four first cycles are included, as we use the ers variable in the backward stepwise selection procedure. The dummies “tight” and “easy” refer to tightening and easing cycles, respectively. The names of the variables and the acronyms used in the table are fully described in Appendix A. Source: Authors’ calculations. 36 Figure 3: Asymmetries During Tightening Cycles for the MSCI Index ICRG = International Country Risk Guide. Note: with the data sample of Appendix D for the 5 cycles. The score of Government Stability is observed one year before each cycle. Source: Authors’ calculations. 4.3. Robustness Checks In Appendix F, we present several robustness check results where we pooled all the cycles (column 1), pooled the tightening cycles (column 2), pooled the easing cycles (column 3), assumed a homogeneous interaction between tightening cycles and the main explanatory variables (column 4), and assumed a heterogeneous interaction between tightening cycles and the main explanatory variables (column 5) for the bilateral exchange rate in Table F1, and for the MSCI indexes in Table F2. Overall, the results indicate robustness, especially for the asymmetries between government stability during the tightening and easing cycles. Tables F1 and F2 provide us with some insight into the importance of building an institutional framework that helps to enhance resilience and ASIAN DEVELOPMENT BANK 6 ADB Avenue, Mandaluyong City 1550 Metro Manila, Philippines www.adb.org The Performance of Emerging Markets During the Fed’s Easing and Tightening Cycles A Resilience Analysis Across Economies This study investigates the determinants of emerging markets performance during five United States Federal Reserve monetary tightening and easing cycles from 2004 to 2023. It shows how macroeconomic and institutional variables are associated with emerging markets’ performance, determinants of resilience differ during tightening versus easing cycles, and institutions matter more during difficult times. Findings are largely consistent with economic intuition, i.e., current account balance, international reserves, and inflation are all important determinants of emerging markets resilience. About the Asian Development Bank ADB is committed to achieving a prosperous, inclusive, resilient, and sustainable Asia and the Pacific, while sustaining its efforts to eradicate extreme poverty. Established in 1966, it is owned by 68 members —49 from the region. Its main instruments for helping its developing member countries are policy dialogue, loans, equity investments, guarantees, grants, and technical assistance.