The effect of quantitative easing on inflation expectations in the United States

(1)

The effect of quantitative easing

on inflation expectations in the

United States

Bachelor Thesis

Author: Jelle Blom

Student number: 10059253

Supervisor: Christiaan van der Kwaak

University of Amsterdam

Faculty of Economics and Business

Specialization: Economics and finance

(2)

2

1. Introduction

In August 2007 the asset bubble on the United States’ housing market burst. The Federal Reserve (Fed) tried to stimulate the economy and ease the financial system by lowering the federal fund rate. Because of ongoing signs of economic weakness, the Fed continued

lowering the federal fund rate until it reached its zero lower bound in December 2008. With the federal fund rate at its zero lower bound the Fed had used up its main policy tool for stimulation of economic activity. The Fed had run out of conventional monetary policy tools and had to turn to the use of new, unconventional, monetary tools. The Fed decided to stimulate the economy further by the managing of expectations and by Large Scale Asset Purchases (LSAPs). Those purchases are also known under the name quantitative easing (QE) (Bernanke, 2012).

There is quite a lot of literature about the effects of quantitative easing on interest rates and economic growth1_{. However, there are only a few empirical papers about the effect}

of quantitative easing on inflation expectations. One of those empirical papers is written by Hofmann and Zhu (2013). They have done an event study and a regression analysis to find out what the effect of quantitative easing was on inflation expectations. They used daily data on one-, five- and ten year inflation swap rates and the five-year, five-year forward inflation swap rate to measure inflation expectations. They regressed these data on a number of control variables, controlling for macro-economic news on that day and on a set of dummies indicating whether or not information about quantitative easing actions emerged that day. Their main result was that quantitative easing had a positive effect on inflation expectations in the medium- and long term.

There has not been done much research on the effect of quantitative easing on

inflation expectations. That is why this paper will try to make a contribution by answering the following research question:

‘What is the effect of quantitative easing on inflation expectations in the United States?’

Answering this question is important because the expected inflation has an important role in economic decision making. As can be seen in the Fisher equation: _𝑖𝑟𝑒𝑎𝑙_{= 𝑖}𝑛𝑜𝑚_{− 𝜋}𝑒_{, the}

1_{See Ugai for a discussion of the effects in Japan: Ugai, H. (2007). Effects of the quantitative easing}

policy: a survey of empirical analyses. Monetary and Economic Studies, 25, (1), 1-47.

For a discussion of the effects in the United Kingdom see: Joyce, M., Tong, M., & Woods, R. (2011). The United Kingdom’s quantitative easing policy: design, operation and impact. Bank of England

Quarterly Bulletin, 51(3), 200-12.

For the effect on interest rates in the US see for example: Krishnamurthy, A., & Vissing-Jorgensen, A. (2011). The effects of quantitative easing on interest rates. Brookings Papers on Economic Activity, 3-43.

(4)

4

expected inflation is an important determinant of the real interest rate. The real interest rate, in turn, is a major influence on economic activity. Economical agents determine their

investment and part of their consumption decisions on the real interest rate. A low real interest rate stimulates the economy, and a high real interest rate contracts the economy. Because the expected inflation can determine the level of the real interest rate it can also have a significant effect on economic activity.

This paper will try to make a contribution to the existing literature by conducting a regression with as dependent variable the inflation expectations for the coming year taken from the survey of the University of Michigan. This expected inflation is regressed on several control variables and on a dummy representing the period in which quantitative easing is conducted.

This paper has the following structure. Firstly there will be a literature review elaborating on central bank policy, quantitative easing and its effects on inflation

expectations in particular. Secondly the empirical model and the data are presented. Thirdly there will be the results of the empirical analysis. Finally there will be the conclusion and the discussion.

(5)

5

2. Literature review

This literature review is set up in the following way. Firstly there will be a description of normal central bank policy. In the next paragraph the liquidity trap and the negative effects of deflation are covered. Thereafter follows a short description of possible unconventional monetary policy instruments. Which is followed by a short overview of events that led to the implementation of quantitative easing in the United States. And finally there will be an overview of the results of unconventional monetary policy on expected inflation in other literature.

The main purpose of central banks nowadays is to retain price stability. See for example the mandates of the two most important central banks in the world. The European Central Bank (ECB) states that its main objective is to maintain price stability and it aims at an inflation rate under, but close to 2%, over the medium term2_{. The Fed’s main objectives}

are maintaining maximum employment, stable prices and moderate long-term interest rates3_{. Under normal circumstances a central bank influences the economy through changes}

in the policy rate. The central bank stimulates the economy by decreasing the policy rate and contracts the economy by increasing it, this is called conventional monetary policy. This policy works under normal circumstances, but it does not when the policy rate reaches its lower bound of zero percent.

This situation, in which the nominal interest rate is at zero percent, is called the liquidity trap (Krugman, 1998). The first one to describe the theoretical concept of the liquidity trap was John Hicks (Hicks, 1937). He described the situation when the short-term interest rate cannot be lowered further to stimulate the economy. Although the liquidity trap was a real economic phenomenon in the 1930’s, it did not occur thereafter for a long time. Until Japan ended up in a liquidity trap in the 1990’s (Krugman, 1998). In Japan the liquidity trap occurred together with deflation, which is the general decline of the weighted average price level in a country. Deflation is almost always caused by rapid decline in aggregate demand. Deflation is generally seen as a bad thing, because it raises the real interest rate. The real interest rate is given in the following equation called the Fisher equation:

(1) _𝑖𝑟𝑒𝑎𝑙 _{= 𝑖}𝑛𝑜𝑚_{− 𝜋}𝑒

The real interest rate is the nominal interest rate minus the expected inflation. If the expected inflation is negative (which is generally the case with deflation), then the real interest rate will be positive even if the nominal interest rate is at its zero lower bound. During periods of deflation the expected growth of the economy is often low or negative. This is caused by the

2_{See: http://www.ecb.int/mopo/html/index.en.html}

(6)

6

low aggregate demand. The combination of the negative economic outlook and the positive real interest rate has a negative effect on the aggregate demand. The reasoning behind that is as follows; the consumption is decreased because people expect the prices to be lower in the future, so they postpone consumption. Investments are reduced because with the positive real interest rate and the low expected economic growth, the investors will find less profitable investments. Concluding, deflation is an unwanted phenomenon because it will lead to lower aggregate demand and through that, lower economic growth.

Because conventional monetary policy is ineffective in a liquidity trap, a central bank must find other policy tools to influence the economy. These policy tools are called

unconventional monetary policy. Two possible strategies to further stimulate the economy are provided by Bernanke and Reinhart (2004). The first one is managing the expectations of the public. The central bank must then make a commitment to keep the short term interest rate lower in the future than previously anticipated by the public. The second strategy proposed by Bernanke and Reinhart is quantitative easing. When a central bank conducts quantitative easing it enlarges its balance sheet by purchasing assets and does not stop when the policy rate reaches zero. The focus is on the quantity of reserves and not the price. Quantitative easing works via the portfolio balance channel; because different assets are not perfect substitutes, changes in supply will affect prices and yields of the purchased assets. When the central bank purchases assets from the private sector, this will influence prices and yields. Because of the higher demand from the central bank the prices will go up and the yields will go down. In response to that investors will make changes in their portfolios and through that the yields on other assets will be influenced. The lower yields will have a positive effect on economic activity.

In August 2007 the crisis on the housing market in the United States erupted, banks had provided borrowers in the so called ‘subprime’ class with mortgages which they were not able to repay. Because housing prices began to decline and an increasing number of people were not able to repay their mortgage, banks got into financial problems (Gertler and Karadi, 2011). The Fed reacted by providing liquidity and by lowering its main policy tool the federal fund rate. Because of the worsening economic situation, the Fed eventually lowered the target range of the federal fund rate to 0 to 25 basis point in December 2008, this is the zero lower bound (Bernanke, 2012). The Fed now turned to unconventional policy tools to further stimulate the economy and prevent deflation. In November 2008 the Fed announced its first program of LSAPs. Between then and October 2012, the Fed more than tripled her balance sheet, through purchases of different kinds of assets4_.

Hofmann and Zhu (2013) analyzed the effects of the large scale assets purchases on inflation expectations in the United States and the United Kingdom. They performed both an

(7)

7

event study and a regression analysis. They used daily data on one-, five- and ten year inflation swap rates and the five-year, five-year forward inflations swap rates and used dummy variables for days on which announcements about LSAP were done together with control variables for other (macro) economic news on these days. They do conclude that announcements of LSAP’s had statistical significant effects on medium- and long term inflation expectations in the United States. The total effects for the medium- and long term ranged between 25 and 52 basis points. Their results for the United Kingdom are less clear. Here they find lower expected inflation in the short term and higher inflation expectations in the longer term.

Guidolin and Neely (2010) concluded after their event study that the LSAP’s had a small effect of increasing the 10 year inflation expectations with 34 basis points. They state that the effect is not negligible and that it thus appears that the central bank is able to increase inflation expectations. So they conclude that quantitative easing can increase inflation expectations and that it can be implemented to oppose deflation.

Anderson et al. (2010) state that although the quantitative easing operation could have had a big upward effect on inflation expectations, this did not materialize because of the following reason. The credibility of the central banks, conducting quantitative easing, is really high. So the public is confident of the fact that the central bank will stick to its mandate of price stability. This means that the public believes that the quantitative easing is temporally and that it will be reversed in time, so that it will not have a big effect on the future inflation. Their research, contrary to that of this paper, is focused on the longer term. This paper will specifically focus on the short-term effects of quantitative easing on inflation expectations.

(8)

𝛽

6

𝐷

𝑄𝐸𝑡

+ 𝛽

7

𝐺𝐷𝑃

𝑔𝑟𝑜𝑤𝑡

ℎ

𝑡

+ 𝜀

𝑡

t = the time which is measured in months

𝜋𝑡+1𝑒 = the median expected inflation, measured by the University of

Michigan inflation expectation, in period t+1.

𝛽0 = the constant of the regression

𝜋𝑡𝑒 = the expected inflation, measured by the University of Michigan

inflation expectation, in period t.

𝑓𝑒𝑑𝑓𝑢𝑛𝑑𝑟𝑎𝑡𝑒𝑡 = the effective federal fund rate per year per period t.

𝐶𝑆𝑡 = the consumers sentiment measured by the Michigan

consumers sentiment index per period t.

𝑢𝑛𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑡 = the level of unemployment during period t, measured by the

number of unemployed as a percentage of the labor force.

𝜋𝑡 = the inflation in period t, measured by the consumers price index

for all urban consumers all items included.

𝐷𝑄𝐸𝑡 = a dummy for the time during which quantitative easing was

conducted.

𝐺𝐷𝑃𝑔𝑟𝑜𝑤𝑡ℎ𝑡 = the annual change in real Gross Domestic Product, measured

per quarter, in this case the three months of one quarter have been given the same growth rate.

(9)

9

3.1 Prediction of regression coefficients signs

This section describes the expected relationships between the dependent variable and the independent variables. The expected inflation in period t is expected to have a positive effect on the expected inflation in period t+1. This is because when the expected inflation was high in the last period people will, all other things equal, expect it to be high next period. The federal fund rate is expected to have a negative effect on inflation expectations, when the federal fund rate is high this has a general negative effect on the aggregate demand. Therefore a lower inflation will be expected. Consumers sentiment is expected to have a positive relation with expected inflation. This is expected because when consumers are more confident, they are expected to consume more and will also expect higher inflation. When unemployment rises, people have less job security. They react by saving more and consuming less to have some savings in case they will lose their job. The people who get unemployed will receive less income so their consumption will go down. Together these effects make it likely that the effect of unemployment on expected inflation is negative. The inflation rate is expected to have a positive relation with expected inflation, because when the current

inflation is high, people will expect it to be high in the next period. The dummy indicating the period of quantitative easing is expected to have a positive effect on inflation expectation. This prediction is based on other empirical results stated in the literature review. The real GDP growth is expected to have a positive effect on inflation expectations, because when the economy is growing faster people generally have more income to spend. Most people will consume more and they expect that other people will do the same, so they expect prices to go up.

3.2 Description of the data

The period covered by the regression is January 1985 up till October 2012. These dates are chosen because all variables are available since January 1985. October 2012 is chosen as last date because the data on consumer sentiment are only available till that date. The data for the consumers sentiment are indexed at 100 in the year 1966, all other data except for the

dummy, are percentage changes compared to the year before. Monthly data is used for all variables, except for real economic growth. This is because data for real economic growth is only available freely on quarterly basis. The data presented to the public about real economic growth are on quarterly basis, so the public will base her inflation expectations on those data. That is why it is not a problem to use the quarterly data. The data for real economic growth is transformed from quarterly to monthly data by assigning the growth rate of a quarter to the 3

(10)

10

months within that quarter. This is done because then the data can be used in the regression. All months from November 2008 until October 2012 are given a 1 for the variable indicating whether or not quantitative easing was conducting during that period. As indicated in section 2, quantitative easing started in November 2008 and was still going in October 2012, the last month that is used in the model. All months before November 2008 are represented by a 0, because quantitative easing was not conducting during that period.

The quantitative easing conducted between November 2008 and October 2012 was not a constant accumulation of one kind of asset. It consisted of different faces, and different kinds of assets were purchased. Although it could be interesting to do research on the

different faces of quantitative easing or to do research on the effect of a certain kind of asset purchase, that is not what this paper will do. This paper will try to find out what the effect of quantitative easing is on short-run inflation expectations. What we hope to find is the broad effect of the total asset purchases.

The data for the effective federal fund rate and the ten year treasury yields are available at the website of the federal reserve5_{. The other data are taken from the economic}

research department of the federal reserve of ST. Louis6_{. The total number of observations}

per variable is 334. Hereafter follow some descriptive statistics about the data7_.

Table 1.

Variable Mean Std. Dev. Min. Max.

Expected inflation 3,078 0,590 0,40 5,20 Effective federal fund rate 4,302 2,658 0,07 9,85 Consumers sentiment 87,647 12,373 55,30 112,00 Unemployment rate 6,089 1,508 3,8 10,0 Inflation rate 2,886 1,278 -1,96 6,38 Dummy for quantitative easing period 0,144 0,351 0 1 real GDP growth 2,649 1,830 -4,6 5,4 5_{See: http://www.federalreserve.gov/releases/h15/data.htm} 6_{See: http://research.stlouisfed.org/fred2/}

7_{Graphs for the expected inflation, effective federal fund rate, unemployment rate, inflation rate and}

(11)

11

3.3 Hypothesis

The hypothesis, which follows from economic intuition and other literature, is that the quantitative easing program has a positive effect on inflation expectations. So it is expected that during periods in which quantitative easing is conducted, the expected inflation will be higher. This means that a positive 𝛽6 is expected.

The effect of quantitative easing on inflation expectations works through the following channel. The asset purchases by the central bank increase the price of those assets and lower the yields. Yields for other assets will go down because of the portfolio balance effect. With the lower yields, more investments are profitable and investments will go up. Because of possible increased wealth through holding of assets, and lower borrowing costs, consumers will consume more. The higher investment and consumption will increase prices. People will probably expect higher prices and that is why we expect higher expected inflation during periods of quantitative easing.

4. Results

A normal ordinary least squares regression is conducted, but to compensate for heteroskedasticity and autocorrelation Newey-West standard errors are used. The autocorrelation is caused by the fact that expected inflation in period t+1 is used as the dependent variable and the expected inflation in the period t is used as one of the independent variables. The regression is performed in stata.8

An overview of the results from the regression, described in the empirical part, is given in the table 2, on the next page.

8_{Stata requires to set a certain maximum lag for the autocorrelation. This maximum lag is chosen via}

the following formula.

𝐿𝑎𝑔(#), 𝑖𝑛 𝑤ℎ𝑖𝑐ℎ # = 0,75𝑁13 .

Where N = number of observations. Given that there are 334 observations, the maximum lag, rounded to a integer is 5. The formula is obtained from: Stock and Watson, (2011), Introduction to

(12)

12 Table 2.

Variable Coefficient Newey-West std. err. t-value p-value

Current expected

inflation 0,7253601*** 0,057524 12,61 0,000 Effective federal fund

rate 0,0126325 0,0089712 1,41 0,160 Consumer sentiment -0,0034538 0,0040052 -0,86 0,389 Unemployment rate -0,0145623 0,0264002 -0,55 0,582 Inflation rate 0,052269** 0,0251221 2,08 0,038

Dummy for quantitative

easing period _0,1583581* _0,0838056 _1,89 _0,060

Real GDP growth 0,0232628 0,0167943 1,39 0,167 Constant

0,9477863* 0,4927245 1,92 0,055

F(7,326)=85,62 Probability > F = _0,000 R-squared=0,7047 Adjusted R-squared = 0,6984

*/**/*** indicate significance of the coefficients at 10, 5 and 1% level.

4.1 Explanation of the results

The coefficient of the expected inflation is 0,7254 this means that for every percentage point of expected inflation in this period the model predicts 0,7254% of expected inflation in following period. This result is as expected and the coefficient is statistical significant at a 1% level.

The effective federal fund rate has a coefficient of 0,0126 this means that for every percentage point of federal fund rate, the model predicts 0,0126% extra expected inflation, this is not as expected, but also, the coefficient is not significant. A negative coefficient was expected, because of the following reasoning: when the federal fund rate goes up the economy contracts, the consumption and investments are going down and inflation expectations are predicted to go down. This is not what the model predicts. A possible

explanation could be that when the federal fund rate goes up, this indicates to the public that the federal reserve judges the current economic situation as being good, the economy is in a boom. So people attach more value to the fact that they are in an economic boom than that

(13)

13

they attach value to the contracting effect of the increase of the federal fund rate. So the effect on inflation expectations is positive.

The coefficient of the consumers sentiment is -0,0035. It is a small negative

coefficient, it is also insignificant. The sign of the coefficient is again not as expected. There is no good explanation for the causation of this opposite sign. Consumers sentiment and

expected inflation are depicted in graph 6.

The unemployment rate has a negative coefficient of -0,0146. This means that for every percentage of unemployment the expected inflation goes down with 0,0146%. The economic reasoning behind the negative sign is the following. When unemployment increases, people’s wealth decreases and they will consume less. People know this and will form lower expectations about future inflation. Because this coefficient is not significant we cannot say with certainty that the effect is as predicted by the model.

The inflation rate has a positive sign of 0,0523, which means that every one percent of inflation gives an extra 0,0523% of expected inflation in the next period. The coefficient is significant at a 5% level. When people make their expectations about the future inflation they also consider the current inflation. When the current inflation is high they will, all other things equal, expect high inflation in the future.

The dummy indicating periods in which quantitative easing is conducted is significant at a 10% level. The coefficient is 0,1584. Indicating that the expected inflation is 0,1584% higher in periods with quantitative easing. With this result the research question can be answered and the hypothesis can be checked. With an estimated effect of 0,1584% it can be concluded that quantitative easing had a positive effect on inflation expectations. Although small the positive coefficient shows that the Fed was able to successfully increase the inflation expectations of the public. This is an important conclusion and it shows that quantitative easing has increased inflation expectations. Increasing the inflation expectations was not the only goal of quantitative easing, but it certainly was a positive development in the

deflationary environment of mid 2009. The hypothesis that predicted a positive effect of quantitative easing on inflation expectations is confirmed.

The coefficient of real GDP growth is 0,0233 and it is not significant. When there is one percent extra real GDP growth in this period, the model predicts 0,0233% higher expected inflation in the next period. This is as predicted earlier in this paper. In periods of higher growth people expect more spending by themselves and others. So they will form higher inflation expectations.

The constant is 0,9478 and is significant at a 10% level. This means that in the hypothetical case in which all the independent variables are 0, the model predicts an expected inflation of almost 1%.

(14)

14

5. Conclusion and discussion

After the outbreak of the credit crisis in late 2007, the Fed quickly ran out of conventional monetary policy tools to stimulate the economy. The Fed came up with new policy tools to further stimulate the economy and prevent a deflationary situation as seen in Japan. The most important of these policy tools are quantitative easing and managing expectations. The combined programs of quantitative easing have more than tripled the Fed’s balance sheet.

In this paper an ordinary least squares regression on expected inflation, a dummy variable indicating period of quantitative easing and several control variables, is performed. The data for inflation expectations are taken from the University of Michigan survey. The main finding of the paper is that in periods when quantitative easing is conducted, the

expected inflation is 0,1584% higher. The implication of this result is that the Fed was able to influence inflation expectations. Increasing the inflation expectations can be of value when the Fed wants to lower the real interest rate or wants to prevent a deflationary situation.

This result is in line with other research. Hofmann and Zhu (2013) found a positive effect of quantitative easing on inflation expectations on the medium- and long term in the United States. Although they did not find significant results for the one- year inflation expectations, their research confirmed that inflation expectations for the longer term have gone up because of quantitative easing.

A possible weakness of this research is the fact that it might suffer from omitted variable bias. There could be determinants of expected inflation that are correlated with one of the independent variables, that are not included in the regression. Another possible weakness is the fact that the research only contains data of the United States, so drawing general conclusions about the effect of quantitative easing on inflation expectations is

dangerous. There is also a possibility that this research suffers from endogeneity. This can be caused by the possible correlation between the dependent variable, expected inflation and the independent variable ‘dummy for quantitative easing’. This correlation is caused by the fact that the level of expected inflation is one of the indicators used by the Fed to determine whether or not she implemented quantitative easing. However in this research the possible endogeneity is not likely to cause incorrect results because the inflation expectations are just one of the many indicators used by the Fed to monitor the economic situation.

There are several possibilities for further research arising from this paper. One possibility is to use the same model, but to use data about the entire period in which

quantitative easing is conducted9_{. This is should be possible in the near future. It would also}

be interesting to find out whether the expected inflation decreases when it is certain that the quantitative easing program will be ended. According to the model and data used in this

9_{In his speech on the 19}_th_{of June this year chairmen of the Fed, Ben Bernanke sketched the}

(15)

15

paper, that should be the case. Other researchers with access to other data could use another variable for the inflation expectations and see if the conclusions of this model still hold. Our research was conducted on the entire quantitative easing program. It could be interesting to do research on the effect on inflation expectations of a specific program of quantitative easing10_{. We conclude that our research was a first attempt to find out what the effect of}

quantitative easing on inflation expectations is. Further research needs to be done to show the robustness of our results and to find out whether the results are the same with other data sets.

10_{The quantitative easing program consisted of several programs called QE1, QE2, QE3 and MEP}

(16)

16

6. References

Anderson, R. G., Gascon, C. S., & Liu, Y. (2010). Doubling your monetary base and surviving: some international experience. Federal Reserve Bank of St. Louis Review, 92, (6), 481-505.

Bernanke, B. S. (2002). Deflation: making sure ‘it’ doesn’t happen here. Remarks before the

National Economists Club, Washington, DC, 21.

Bernanke, B.S. (2012). Speech at the Federal Reserve Bank of Kansas City Economic Symposium, Jackson Hole, Wyoming.

Bernanke, B.S. (2013). At the press conference after the two days meeting of the Federals Open Market Committee on 19 June 2013.

Bernanke, B.S. and V.R. Reinhart (2004). Conducting Monetary Policy at Very Low Short- term Interest Rates. The American Economic Review, 94, (2), 85-90.

Gertler, M., & Karadi, P. (2011). A model of unconventional monetary policy. Journal of

Monetary Economics, 58(1), 17-34.

Guidolin, M., & Neely, C. J. (2007). The effects of large-scale asset purchases on TIPS inflation expectations. Economic Synopses.

Hicks, J. R. (1937). Mr. Keynes and the "classics"; a suggested interpretation. Econometrica:

Journal of the Econometric Society, 147-159.

Hofmann, B., & Zhu, F. (2013). Central bank asset purchases and inflation expectations. BIS

Quarterly Review March.

Krugman, P. R. (1998). It's baaack: Japan's slump and the return of the liquidity trap. Brookings Papers on Economic Activity, 1998(2), 137-205.

(17)

17

7. Appendix

Graph 1. Graph 2. Graph 3. 0,00 1,00 2,00 3,00 4,00 5,00 6,00 1-1985 1- 5-1986 1- 9-1987 1-1989 1- 5-1990 1- 9-1991 1-1993 1- 5-1994 1- 9-1995 1-1997 1- 5-1998 1- 9-1999 1-2001 1- 5-2002 1- 9-2003 1-2005 1- 5-2006 1- 9-2007 1-2009 1- 5-2010 1- 9-2011

Expected inflation

Expected inflation 0,00 2,00 4,00 6,00 8,00 10,00 12,00 1-1985 1- 7-1986 1-1988 1- 7-1989 1-1991 1- 7-1992 1-1994 1- 7-1995 1-1997 1- 7-1998 1-2000 1- 7-2001 1-2003 1- 7-2004 1-2006 1- 7-2007 1-2009 1- 7-2010 1-2012

Effective federal fund rate

0,0 2,0 4,0 6,0 8,0 10,0 12,0 1-1985 1- 5-1986 1- 9-1987 1-1989 1- 5-1990 1- 9-1991 1-1993 1- 5-1994 1- 9-1995 1-1997 1- 5-1998 1- 9-1999 1-2001 1- 5-2002 1- 9-2003 1-2005 1- 5-2006 1- 9-2007 1-2009 1- 5-2010 1- 9-2011

Unemployment rate

(18)

18 Graph 4. Graph 5. Graph 6. -4,0 -2,0 0,0 2,0 4,0 6,0 8,0 1-1985 1- 3-1986 1- 5-1987 1- 7-1988 1- 9-1989 1- 11-199 0 1-1992 1- 3-1993 1- 5-1994 1- 7-1995 1- 9-1996 1- 11-199 7 1-1999 1- 3-2000 1- 5-2001 1- 7-2002 1- 9-2003 1- 11-200 4 1-2006 1- 3-2007 1- 5-2008 1- 7-2009 1- 9-2010 1- 11-201 1

Inflation

Inflation -6 -4 -2 0 2 4 6 1-1985 1- 5-1986 1- 9-1987 1-1989 1- 5-1990 1- 9-1991 1-1993 1- 5-1994 1- 9-1995 1-1997 1- 5-1998 1- 9-1999 1-2001 1- 5-2002 1- 9-2003 1-2005 1- 5-2006 1- 9-2007 1-2009 1- 5-2010 1- 9-2011

real GDP growth

real GDP growth 0,0 2,0 4,0 6,0 8,0 10,0 12,0 1-1985 1- 7-1986 1-1988 1- 7-1989 1-1991 1- 7-1992 1-1994 1- 7-1995 1-1997 1- 7-1998 1-2000 1- 7-2001 1-2003 1- 7-2004 1-2006 1- 7-2007 1-2009 1- 7-2010 1-2012 Expected inflation 1/10 Consumers sentiment

(19)

19 Graph 7.

The effect of quantitative easing on inflation expectations in the United States