The beginning of the new millennium began with, a collapse in stock prices triggered by the burst of the so-called dot-com bubble and the terrorist attacks on the World Trade Center of September 11, 2001. Central banks responded aggressively by cutting interest rates and kept them low after the recovery of stock prices and subsequently the real economy recovered. The story was repeated after the collapse of the sub-prime mortgage bonds markets in the 2007 which has triggered the greatest economic crisis since World War 2.
The period after 1999 was thus marked by large variations in inflation and the output gap, but at the same time by booms and busts in the stock markets. This offers a great opportunity to explore whether or not central banks - independently of the predictive power of stock markets for expected inflation and the output gap - tried to influence stock prices through interest rate setting.
38
Table 11: Fed’s reaction to stock prices after 1999
US α β γ λ p
S&P5001 -4.85*** 2.61*** 0.73*** -0.02 0.97*** 1.000
(0.38) (0.09) (0.08) (0.02) (0.00)
Country index2 -4.69*** 2.56*** 0.68*** -0.01 0.97*** 1.000
(0.39) (0.09) (0.09) (0.02) (0.00)
PE3 -4.28*** 2.54*** 0.56*** 0.05*** 0.97*** 1.000
(0.43) (0.09) (0.07) (0.01) (0.00)
PC4 -2.77*** 2.10*** 0.33*** 0.11*** 0.97*** 1.000
(0.53) (0.09) (0.08) (0.02) (0.00)
PB5 -4.93*** 2.74*** 0.49*** 0.06*** 0.97*** 1.000
(0.49) (0.10) (0.09) (0.01) (0.00)
The standard errors are reported in the parentheses. The right-most column right-hand reports the p-value associated with a test of the model’s overidentifying restrictions. The set of instruments includes 8 lags (1-6, 9, 12) of the CPI inflation, output gap (HP-filtered industrial production), the effective Federal funds rate (7 lags), the short-long spread, the M2 growth, the EFF and the commodity price inflation.
1 The HP-filtered S&P500 index is used as policy relevant variable and lags as instruments
2 The HP-filtered country index as calculated by Datastream is used as policy relevant variable and lags as instruments.
3 The yearly percentage change in the PE ratio is used as policy relevant variable and lags as instruments.
4 The yearly percentage change in the PC ratio is used as policy relevant variable and lags as instruments.
5 The yearly percentage change in the PB ratio is used as policy relevant variable and lags as instruments.
The results in Table 11 suggest that the US central bank did not respond to misalignments in the prices of stocks in the S&P index. To be exact, the point estimate of the coefficient on the
“stock price gap”, λ, is highly statistically insignificant and also negative which is in contrast with the common wisdom about how central banks may react to stock price misalignments.
The same conclusions apply when a country specific index is included instead.
On the other hand, measures used in the equity pricing theory, reveal a different story. More specifically, the Fed’s response to the yearly percentage change in the PE ratio is shown to be statistically significant and positive. An even stronger conclusion about the Fed’s reaction to stock prices can be drawn when a PC ratio is used – the point estimate on the PC ratio, λ, suggests that in response to the yearly fall in the PC ratio by one percentage point, the Fed decreasedt he target interest rate by more than 0,1 percentage point. The response may seem quantitatively small, but knowing the yearly change in PC ratio can take large values, just the opposite is true - for example, after the the dot-com bubble burst, the PC ratio decreased by
39
almost 40 percent from the previous year – this would imply a 4 percentage point reduction in the target interest rate. Qualitatively the same conclusion can be drawn when the PB ratio is used.
We saw that the Fed under Alan Greenspan and Ben Bernanke to some extent responded to stock market movements. What about the ECB’s response to stock prices?
Table 12 reports some interesting results. The ECB is shown to have responded also to stock price misalignments as captured by the “stock price gap”. The highly statistically significant point estimates of the coefficient on both HP-filtered stock’s indexes implies that the ECB responded to the one percentage point increase in the “stock price gap” by a 0.05 percentage point increase in the target interest rate, independently of the predictive power of the stock return for expected inflation and output gap. Knowing the values of the “stock price gap” can take values up to 40 percent (deviation of the current stock price index from its potential level), the response is also quantitatively significant.
Table 12: ECB’s reaction to stock prices
ECB α β γ λ p
Euro Stoxx 501 2.08*** 0.32*** 0.21*** 0.05*** 0.90*** 1.000
(0.12) (0.03) (0.01) (0.00) (0.00)
Country index2 1.56*** 0.58*** 0.18*** 0.06*** 0.93*** 1.000
(0.17) (0.05) (0.02) (0.00) (0.00)
PE3 0.07 1.41*** 1.38*** 0.28*** 0.98*** 1.000
(0.67) (0.22) (0.06) (0.02) (0.00)
PC4 1.88*** 0.30*** 0.37*** 0.07*** 0.95*** 1.000
(0.24) (0.08) (0.02) (0.00) (0.00)
PB5 1.26*** 0.80*** 0.46*** 0.15*** 0.97*** 1.000
(0.36) (0.11) (0.03) (0.01) (0.00)
The standard errors are reported in parentheses. The right-most column reports the p-value associated with a test of the model’s overidentifying restrictions. The set of instruments includes 8 lags (1-6, 9, 12) of HICP base inflation, output gap (HP-filtered industrial production), the ECB’s policy rate (7 lags), the difference between actual and target M3 growth, EER, short-long spread and the commodity price inflation.
1 The HP-filtered Euro Stoxx index is used as policy relevant variable and lags as instruments
2 The HP-filtered country index as calculated by Datastream is used as policy relevant variable and lags as instruments.
3 The yearly percentage change in the PE ratio is used as policy relevant variable and lags as instruments.
4 The yearly percentage change in the PC ratio is used as policy relevant variable and lags as instruments.
5 The yearly percentage change in the PB ratio is used as policy relevant variable and lags as instruments.
40
This conclusion is confirmed when I include the measures from pricing theory. The coefficients on the all three ratios are highly statistically significant and positive, which confirms that ECB tried to lean against stock price misalignments.
The interesting result concerns the implied target inflation rate by the ECB. In all the alternative specifications when measures of stock price misalignment are included, the implied target inflation rate (not reported here) is close to, but below 2 percent – exactly the official target published by the ECB.
CONCLUSION
In this thesis I have explored rate setting by the most important US and Europe central banks.
When considering rate setting by the Fed and the Bundesbank, results are in line with the commonly accepted facts. Paul Volcker’s arrival at the head of the Fed changed the focus of the central banks in the direction of stabilizing expected inflation. This is reflected in the considerably higher point estimate of the coefficient on the expected inflation in the post- Volcker compared to pre-Volcker period. Nevertheless, compared to the results obtained by Clarida et al., my results offer less support for strict inflation targeting. I found the central banks did not focus exclusively on the expected inflation deviation, but that the variability of the output gap was still an important factor when considering the appropriate level of the interest rate.
After 1999 the Fed seems to have stuck to its policy of aggressively responding to expected inflation and the output gap. Different results are found for the ECB, implying that it has not responded aggressively to expected inflation. This “milder” version of inflation targeting may be the consequence of its two-pillar approach, as estimation revealed that the ECB also focuses on money growth. At the same time, the implied target inflation rate in the ECB’s baseline reaction function is implausible, which may suggest that the ECB is also considering other factors when setting the interest rate.
The most interesting results concern the response by central banks to stock price misalignments, independently of the predictive power of stock price movements for expected inflation and the output gap. Counterintuitive results are found only for the German Bundesbank in the pre-Volcker period - results suggest that the Bundesbank actually decreased its target interest rate in response to a “stock price bubble”. Somewhat surprisingly
41
there is also evidence of a strong response to the “stock price gap” by the Bundesbank in the post-Volcker period.
On the other hand, till 1999, the Fed seems not to be, independently of its effect on the expected inflation and output gap, concerned with the possible development of bubbles in the stock markets. However, after the 1999, the Fed has started to focus also on stock price misalignments with its monetary policy. Namely, the results suggest that movements in the measures obtained from equity pricing theory, as price to earnings ratio, induced the Fed to change its target interest rate independently of the predictive power the latter measures have about the expected changes in inflation and output gap. The same conclusions are further strengthened in the case of ECB. Similarly to the Fed, the ECB is found to have responded to the “stock price gap”. The results, therefore, offer strong evidence that the ECB has tried to affect stock price misalignments in a stabilizing manner.
On the one hand these findings correspond to the criticism by some economists that central banks, especially the Fed, responded too aggressively to the collapse of stock prices following the dotcom bust at the beginning of the millennium. According to leading critics they cut interest rates “too low” and kept them low for “too long”. As a consequence, central banks induced further assets price bubbles, especially in the real estate market, ultimately resulting in the greatest economic crisis since World War 2. We can call such an interpretation of the results the “destabilizing” view. The latter interpretation would be appropriate if we assume that central banks react only to stock price crashes and not to stock price booms.
On the other hand, if we assume that the central banks’ reaction to stock price misalignments is symmetrical - when there is a boom in the stock markets, central banks will increase interest rates - then they are also act as a stabilizer. Such a “stabilizing” interpretation of the results is more in line with economists who argue that central banks should react to stock price misalignments. As journal The Economist puts it “The evidence is clear that the clean- up costs after debt-financed bubbles are too high. Central banks and governments do have to intervene when credit growth and asset prices start dancing their toxic two-step”. Therefore, if we want to draw more exact conclusions about the nature of central banks’ response to stock prices, future research should look at the symmetry of the reaction to stock price misalignments.
42
Bibliography
1. Bank for International Settlements. Found August 1st 2010 on http://www.bis.org.
2. Baum, C., Schaffer, M., & Stillman, S. (2003). Instrumental Variables and GMM:
Estimation and Testing. Stata Journal. 3 (1), 1-31.
3. Bernanke, B., & Gertler, M. (1999). Monetary Policy and Asset Price Volatility. In New Challenges for Monetary Policy. Kansas City: Federal Reserve Bank of Kansas City.
4. Bound, J., Jaeger D. A., & Baker, R. (1995). Problems with Instrumental Variables Estimation when the Correlation between the Instruments and the Endogeneous Explanatory Variable is Weak. Journal of the American Statistical Association. Vol.
90, 443-450.
5. Bodenstein, M., Erceg, C. J., & Guerrieri, L. (2008). Optimal Monetary Policy with Distinct Core and Headline Inflation Rates. International Finance Discussion Papers 941, Washington: Board of Governors of the Federal Reserve System.
6. Cecchetti, S., Genberg, H., Lipski, J., & Wadhwani, S. (2000). Asset Prices and Central Bank Policy. Geneva Reports on the World Economy, International Center for Monetary and Banking Studies and Centre for Economic Policy Research.
7. Clarida, R., Gali, J., & Getler, M. (1998). Monetary Policy Rules in Practice: Some International Evidence. European Economic Review, Vol. 42., 1033-1067.
8. Clarida, R., Gali, J. & Getler, M. (2000). Monetary Policy Rules and Macroeconomic Stability: Evidence and Some Theory. Quarterly Journal of Economics, 115 (1), 147- 180.
9. Deutsche Bundesbank - Statistics. Deutsche Bundesbank - Startseite. Found August 1st 2010 on http://www.bundesbank.de/statistik/statistik.en.php
10. ECB Statistical Data Warehouse. Found August 1st 2010 on http://sdw.ecb.europa.eu/
11. Economic and Financial Affairs - Indicators -AMECO Database. Found August 1st 2010 on http://ec.europa.eu/economy_finance/ameco/user/
12. Gali, J., & Gertler, M. (1999). Inflation Dynamics: A Structural Econometric Analysis. J. Monetary Econ. 44, 195–222.
13. Gerdesmeier, D. & Roffia, B. (2004). “Empirical Estimates of Reaction Functions for the Euro Area.” Swiss Journal of Economics and Statistics, 140, 37-66.
14. Goodfriend, M. (1991). Interest Rate Smoothing in the Conduct of Monetary Policy.
Carnegie-Rochester Conference Series on Public Policy, 7-30.
43
15. Goodfriend, M. (2007). How the World Achieved Consensus on Monetary Policy.
Journal of Economic Perspectives, 21(4), 47–68.
16. Hansen, L. P. (1982). Large Sample Properties of Generalized Method of Moments Estimators. Econometrica, 50, 1029–1054.
17. Hodrick, R., & Prescott, E. (1997). Post-War Business Cycles: An Empirical Investigation. Journal of Money, Credit and Banking, 29, (1), 1-16.
18. Issing, O. (1997). Monetary Targeting in Germany: The Stability of Monetary Policy and of the Monetary System. Journal of Monetary Economics, 39(1), 67–79.
19. Issing, O. (2006). The ECB’s Monetary Policy Strategy: Why did We Choose a Two- pillar Approach? Frankfurt: ECB.
20. Kahn, G. & Benolkin, S. (2007). The Role of Money in Monetary Policy: Why do the Fed and the ECB See It So Differently? Federal Reserve Bank of Kansas City.
Economic Review, Third Quarter, 5–36.
21. Laszlo, M. (1999). Generalized Method of Moments Estimation. Cambridge:
Cambridge University Press.
22. Newey, W. K., & West, K. D. (1987). A Simple, Positive Semi Definite,
Heteroskedasticity and Autocorrelation Consistent Covariance Matrix. Econometrica, 55, 703–706.
23. OECD Statistics (GDP, Unemployment, Income, Population, Labour, Education, Trade, Finance, Prices...). Found August 1st 2010 on http://stats.oecd.org/
24. Rice, J. (1995). Mathematical Statistics and Data Analysis (Second ed.). Belmont:
Duxbury Press.
25. Shea, J. (1997). Instrument Relevance in Multivariate Linear Models: A Simple Measure. Review of Economics & Statistics, 79 (2), 348-352.
26. Smaghi, L. B. (2005). Inflation, Expectations and Current Challenges to Monetary Policy. Speech at the European Inflation-Linked Conference. Rome.
27. St. Louis Fed: Economic Data - FRED®. St. Louis Fed: Economic Research. Found August 1st 2010 on http://research.stlouisfed.org/fred2/
28. Svensson, Lars E. O. (1996). Inflation Forecast Targeting: Implementing and Monitoring Inflation Targets. NBER Working Paper, 5797.
29. Taylor, J. B. (1993). Discretion versus Policy Rules in Practice. Carneige-Rochester Conference Series on Public Policy, 39, 195–214.
30. Taylor, J. B. (1999). An Historical Analysis of Monetary Policy Rules. In J. B. Taylor (ed.), Monetary Policy Rules (pp. 319–341). Chicago: University of Chicago Press.
44
31. Thomson Reuters Datastream. Found August 1st 2010 on http://www.datastream.com/
1 POVZETEK V SLOVENŠČINI
UVOD
Leta 1993 je John B. Taylor predlagal enostavno pravilo, ki naj bi odražalo kako in na podlagi katerih dejavnikov centralne banke vodijo monetarno politiko prek najpomembnejšega instrumenta, to je spreminjanja obrestne mere. Najprivlačnejša značilnost pravila je prav njegova enostavnost pa vendar sposobnost, da relativno natančno opiše obnašanje monetarne oblasti. Prav zaradi teh značilnosti je, tako imenovano Taylorjevo pravilo, postalo ključno v akademskih raziskavah o vodenju monetarne politike.
Napredek in nova spoznanja v monetarni teoriji so kljub temu odprla nekatera vprašanja o konsistentnosti Taylorjevega pravila – tovrstno pravilo ni mogoče izpeljati iz maksimizacijskega problema centralne banke, zato pravilo ni teoretično osnovano. Druga in mogoče večja težava pa so bile nekonsistentne ekonometrične tehnike uporabljene v prvih raziskovalnih nalogah, ki so poizkušale empirično oceniti Taylorjevo pravilo.
Nova spoznanja v teoriji monetarne politike in konsistentna ekonometrična metoda za ocenjevanje tako imenovanih reakcijskih funkcij centralnih bank - torej funkcij, ki opisujejo vedenje primarnega instrumenta centralne banke, t.j. določanja obrestne mere, - so bila združena v vplivni znanstveni publikaciji, ki so jo napisali Richard Clarida, Jordi Gali in Mark Gertler v letu 1998. Taylorjevo pravilo, ki vsebuje pretekle vrednosti spremenljivk so tako nadomestili s Taylorjevim pravilom, ki upošteva pričakovane vrednosti spremenljivk.
Ocenjevanje tovrstnega pravila pa je izvedeno z metodo posplošenih momentov (angl.
Generalized method of moments), ki je za razliko od metod prej uporabljenih v tovrstnih študijah, kot je metoda najmanjših kvadratov (angl. Ordinary least squares) in vektor avtoregresijska metoda (angl. Vector autoregression method), konsistentna ekonometrična metoda za ocenjevanje reakcijskih funkcij.
V magistrski nalogi se upiram na metodologijo, razvito s strani Claride, Galija in Getlerja, in jo uporabim za raziskavo nekaterih zanimivih a še ne raziskanih vprašanj, ki se tičejo vodenja monetarne politike centralnih bank. Več kot deset let obstoja Evropske centralne banke (v nadaljevanju ECB) kliče k empirični analizi njenega vodenja monetarne politike; Ali je ECB vodila monetarno politiko inflacijskih ciljev in torej agresivno reagira na spremembe v pričakovani inflaciji? So dogajanja v realnem gospodarstvo še vedno pomemben faktor, ko se
2
ECB odloča o višini obrestne mere. Je nemška centralna banka, Bundesbank, resnično vodila monetarno politiko s cilji glede rasti denarnih agregatov in ali je ECB tovrstno monetarno politiko prevzela? Katere so ključne razlike v vodenju monetarne politike s strani ECB in ameriške centralne banke, Fed? To so ključna vprašanja na katera skušam najti odgovore prek empirične analize.
Ključen del magistrske naloge pa se nanaša na vprašanje ali dogajanje na borzah oziroma spremembe cen vrednostih papirjev (delnic) direktno vplivajo na vodenje monetarne politike.
Slednja tema je postala relevantna v času makroekonomske stabilnosti, zaznamovane s nizko inflacijo in relativno nizko variabilnostjo bruto domačega proizvoda. Tema je še pridobila na relevantnosti po izbruhu največje ekonomsko-gospodarske krize po tridesetih letih prejšnjega stoletja v letu 2007. Raziskave, ki so se ukvarjale s slednjo temo, so bila usmerjene predvsem k teoretični podpori oziroma nasprotovanju direktne reakcije monetarne politike na cene vrednostnih papirjev. Moj namen v magistrski nalogi ni prispevati k tej vrsti literature pač pa empirično preveriti ali so že do sedaj centralne banke reagirale na cene vrednostih papirjev.
Prvi del magistrske naloge je namenjen metodološkim razlagam. V drugem poglavju tako predstavim ekonometrični dizajn, razvit s strani Claride in drugih – poglavje se močno zanaša na njihovo raziskovalno nalogo z naslovom ằMonetary Policy Rules in Practice: Some International Evidenceô. V tretjem poglavju sledi predstavitev metode posplošenih momentov, ki jo uporabim za ocenjevanje reakcijskih funkcij – poglavje se močno nanaša na knjigo avtorja Matyas Laszla “Generalized Method of Moments Estimation”. Zadnje poglavje v prvem delu magistrske naloge pa je namenjeno predstavitvi najpomembnejših ekonometričnih testov, ki jih raziskovalec uporabi, ko ocenjuje modele z metodo posplošenih momentov.
Drugi del magistrske se začne z opisom podatkov in podatkovnih baz uporabljenih za namen ocenjevanja reakcijskih funkcij. Naslednja poglavje so namenjena predstavitvi rezultatov ocenjevanja reakcijskih funkcij. Naprej so predstavljeni rezultati ocenjevanja reakcijskih funkcij za nemško in ameriško centralno banko do leta 1999. Slednje poglavje služi predvsem za primerjavo mojih rezultatih z dobljenimi rezultati v raziskovalni nalogi, narejeni v strani Claride in drugih. Naslednje poglavje je namenjeno dobljenim empiričnim rezultatom pri raziskovanju reakcijskih funkcij ameriške in evropske centralne banke po letu 1999. V osmem poglavju so predstavljeni rezultati raziskovanja ali centralne banke reagirajo na cene
3
vrednostih papirjev. Sledi zaključek, kjer povzamem in na kratko povzamem dobljene rezultate.
METODOLOGIJA IN REZULTATI
V tem delu povzetka v Slovenščini bom predstavil ključne metodološke stvari in ključne oziroma najzanimivejše rezultate magistrske naloge.
Spodaj je predstavljena že izpeljana enačba, ki jo ocenjujem prek metode posplošenih momentov:
( ) ( ) ( ) (6) Če definiramo ( ) , ( ) , ( ) in , dobimo naslednjo enačbo, ki jo je mogoče oceniti z linearno metodo posplošenih momentov:
(7)
V enačbi predstavlja tarčno obrestno mero centralne banke, je pričakovana inflacija (v magistrski sem izbral kot horizont eno leto, torej n=12), je tako imenovana ằproizvodna vrzelô (v magistrski nalogi dobljena prek aplikacije HP-filtra na časovne serije industrijske proizvodnje za posamezno državo), pa je odložena tarčna obrestna mera in neodvisna napaka modela. Vključitev odložene tarčne obrestne mere je posledica dejstva, da centralne banke tarčno obrestno mero prilagajajo postopoma.
Zgornja enačba implicira slednje momente:
, | - (8) Za ocenitev vektorja parametrov [ ] uporabljam metodo posplošenih momentov (ang. Generalized Method of Moments). Kot osnovna specifikacija se razume specifikacija reakcijske funkcije v kateri sta le pričakovana inflacija in proizvodna vrzel spremenljivki na katere centralna banka neposredno reagira. Ob odloženih vrednostih inflacije in proizvodne vrzeli so v vseh specifikacijah reakcijske funkcije kot instrumenti uporabljeni odlogi cen surovin in letna rast denarnih agregatov. Ostali uporabljeni instrumenti so opisani pod vsako posamezno tabelo.