UNIVERSITY OF GOTHENBURG
SCHOOL OF BUSINESS, ECONOMICS AND LAW
GRADUATE SCHOOL
The Riksbank Unconventional Monetary Policies And
The Swedish Krona
Abstract
This thesis examines the impact of Swedish Riksbank unconventional monetary policy on the
EURSEK exchange rate, focusing on the exchange rate channel of quantitative easing programs. It
extends previous research by investigating a longer time frame from January 1999 through December
2019 and includes a discussion on the effects of negative interest rates. Using structural vector
autoregression, we find unexpected innovations to Riksbank asset purchases of long-term bonds
significantly influence the exchange rate. The results suggest that an increase in Riksbank holdings of
long term-bonds with 1 percentage unit of GDP depreciates the Swedish Krona against the Euro by
approximately 0,9 percent. A further exploration of the effect of innovations to short-term rates on
exchange rates found a significant impact consistent with the Mundell-Fleming model, suggesting that
negative rates by the Riksbank led to further depreciationary pressure on the Swedish Krona.
MSc Finance Programme
Spring Semester 2024
Supervisor: Jesper Lindé
Authors: Gillesén Markus & Myrbäck Albin
Table Of Contents
1. Introduction.......................................................................................................................... 5
2. Background...........................................................................................................................7
2.1 The Swedish Central Bank.............................................................................................7
2.2 The Financial Crisis Of 2008......................................................................................... 7
2.3 Unconventional Monetary Policy...................................................................................8
2.3.1 Quantitative Easing Programmes.......................................................................... 8
2.3.2 Negative Interest Rate And Its Implication.........................................................10
2.4 FX And Implications On Swedish Economy............................................................... 11
3. Literature review................................................................................................................ 12
4. Theoretical framework...................................................................................................... 14
4.1 Negative Short-Term Rates.......................................................................................... 14
4.2 Mundell-Fleming model...............................................................................................15
4.3 Quantitative Easing...................................................................................................... 16
4.3.1 Signaling Channel............................................................................................... 16
4.3.2 Portfolio Balance Channel.................................................................................. 17
4.3.3 Exchange Rate Channel...................................................................................... 17
5. Empirical Framework........................................................................................................18
5.1 Data Collection Process............................................................................................... 18
5.2 Variables Within The Model........................................................................................ 18
5.3 Description Of The Data.............................................................................................. 19
5.4 Visualization Of The Data............................................................................................20
5.4.1 Euro-Zone Data................................................................................................... 21
5.4.2 Swedish Data.......................................................................................................22
5.4.3 Central Bank Assets............................................................................................ 23
5.5 Structural Vector Autoregression................................................................................. 24
5.6 Econometric Model...................................................................................................... 25
5.7 Model Specification And Strategy............................................................................... 25
6. Results................................................................................................................................. 28
6.1 EURSEK Response To Unconventional Monetary Policy...........................................28
6.2 Further Interdependencies............................................................................................ 31
7. Discussion............................................................................................................................35
8. Conclusion...........................................................................................................................36
References............................................................................................................................... 38
Appendix................................................................................................................................. 42
2
Glossary
Term Definition
EURSEK Swedish Kronor paid to obtain one Euro
QE Quantitative Easing
Unconventional monetary policy Negative interest rates and QE
SVAR Structural Vector Autoregression
STIBOR Stockholm Interbank Offered Rate
ECB European Central Bank
3
We would like to express our gratitude to our supervisor, Jesper Lindé, for his guidance,
support, and insightful feedback throughout the process of writing this thesis. Thank you for
always taking your time from your busy schedule. Jesper Lindés expertise within the field,
encouragement, and dedication have been instrumental in shaping our thesis. We feel
fortunate to have had Jesper Lindé as our supervisor, and we are grateful for his patience
and generosity.
4
1. Introduction
The purpose of the study is to examine how the EURSEK was affected by the unconventional
monetary policy conducted by the Swedish Riksbank following the global financial and euro
area debt crises. By using data between 1999 and 2019, this thesis aims to explain the
mechanisms of the unconventional quantitative easing program alongside a negative nominal
Short-term interest rate. The thesis aims to complement former research by solely focusing on
the exchange rate channel of quantitative easing programs. Studies have found evidence of
the exchange rate channel but has so far focused on a more narrow time window
(2015-2018). By using data for an extended period, we examine if the findings hold up for a
longer time period. We further wish to complement the area by including a discussion of
negative interest rates observed alongside the quantitative easing programmes. The stated
research questions are as follows:
- Did the Riksbank Quantitative Easing programmes depreciate the EURSEK?
- How did the short-term rate affect the EURSEK during 1999-2019?
In short, there is evidence that the quantitative easing program did influence the EURSEK.
Using a structural vector autoregression, our findings suggest that unexpected innovations in
asset purchases have a significant effect on the exchange rate. This results aligns with
research that focused on a more shorter sample. Furthermore, the magnitude of the influence
can be calculated by suitably scaling variables in the SVAR and calculating an elasticity of
the exchange rate vis-a-vis assets purchased by the Riksbank. Doing so, we find that one
percentage unit increase of asset holdings as a share of GDP by the Riksbank depreciates the
SEK against the EUR by approximately 0,9 percent. Focusing on the Short term rate, the
STIBOR is used as a proxy for central bank rate decisions. With this proxy, we find
unexpected cuts in the STIBOR significantly depreciate the exchange rate. Such responses
are supported by the traditional Mundell-Fleming model, which implies that a monetary
expansion depreciates the domestic exchange rate (Wang, 2020). To be clear, our findings of
the short term rate do allow for the possibility of a nonlinear effect when the policy rate
becomes negative, it explains the general effect of short term rate shocks on the exchange rate
assuming the relationship is the same when short-term rates are negative and positive.
Erikson and Vestin (2019) argues the transmission of policy rates are invariant to the level of
policy rates, i.e. if they are negative or positive.
5
Summing up the effect on both QE and negative short term interest rate our analysis suggests
the following total effect. The Riksbank did about 10 percent of baseline GDP (i.e.
annualized GDP in 2014 Q4) of QE, and lowered its key short-term policy rate to -50 basis
points. Since 1 percent of QE depreciates the exchange rate persistently by about 0.9 percent,
and a 100 basis point decrease in the policy rate depreciates the exchange rate durably with
nearly 3 percent, out point estimates suggest that these policies combined depreciated the
Swedish Krona with about 11 percent for some time.1 These results are associated with some
uncertainty since the credible impulse intervals increase over time, as later shown, which
implies that the actual effect could either be smaller or even larger. Following the estimated
SVAR model's impulse response functions, the 11 percent depreciation contribution is
however the best prediction (based on the point estimates).
In addition, this thesis provides the reader a thorough background on Sweden as a small open
economy including the importance of the exchange rate for Sweden and the financial crisis of
2008. Sweden can be described as a small open economy, which is affected by both domestic
and foreign shocks. Stockhammar and Casola (2022) argue that the literature on the effects of
unconventional monetary policy strategies withholds a gap when it comes to the smaller open
economies where these may be affected by domestic as well as foreign QE programs, a
theory proven in this paper. The QE program conducted by the ECB did significantly impact
the Swedish Economy as we will subsequently establish. STIBOR as well as the Swedish
10-year T-bond had a negative response to innovations by ECB assets purchasing. A more
complex finding indicates that innovations by ECB assets purchasing has a positive response
on Riksbank’s assets, this could be explained by the Riksbank responding to ECB asset
purchases program as can be seen in following sections.
1 These calculations do not consider cross-effects i.e. that QE by the Riksbank itself is associated with a lower
short term policy rate. However, such cross-effects of policy instruments are small in the longer term(three
years) whereas effects on the exchange rate are more persistent. See appendix five and six.
6
2.Background
2.1 The Swedish Central Bank
Sweden is a small open economy exposed to fragmentation of global trade, protectionism,
geopolitical tensions, supply shocks, and increased global uncertainty that all pose risks for
the weakness of the Swedish currency (Kandil & Mirzaie, 2003, 42). Monetary policy
encompasses the economic measures influencing the money supply, lending practices, and
interest rates within a country. Typically administered by either a monetary union or a central
bank, this policy is commonly carried out independently in many of the world's developed
nations, free from direct political influence. In Sweden, the Riksbank serves as the central
bank, operating under the regulations outlined in the Riksbank Act. (Sveriges Riksbank,
2020)
The Riksbank has established a clear criterion for maintaining a stable monetary value, with a
specific focus on an annual inflation rate of two percent which depicts the Riksbank's
designated inflation target. To foster transparency and market predictability, Sweden adopted
an open monetary policy in 1992, an approach which ensures that changes in interest rates are
foreseeable. The strategy is carried out by the Riksbank through regular publication of
forecasts detailing anticipated future rate adjustments (Sveriges Riksbank, 2018). As of
February 11th, 2015, the Riksbank decided to enter non-observed territory by introducing a
negative repo rate to tackle the low inflation being observed (Sveriges Riksbank, 2015). This
led to countries differing substantially in their central bank policies and only five countries
adopted the Negative Interest Rate Policy (Trading Economics, 2023).
2.2 The Financial Crisis Of 2008
In 2008 a global financial crisis occurred that stemmed from the housing bubble in the US
where mortgages increased at a rapid pace due to banks being able to sell the debt through
different types of securities. This enabled mortgages with lower credit status to be issued
since the debt itself could be packaged with other securities and the risk would be sold off
with a higher obtained credit rating. In this way, the banks took very little risk and could
through various ways set unmotivated high credit ratings on these high-risk securities.
(Weinberg, 2013)
7
The following recession and economic crisis spread not only in the US but also deeply
affected the Euro-zone since the mortgage-backed securities had been traded internationally
to a high degree, resulting in defaults for the European market. Jef Boeckx & Naïm
Cordemans write in their research “Monetary Policy in the Wake of the Great Recession”
about the consensus amongst academics and central bankers before the crisis that monetary
policy should be geared toward medium-term price stability. Inflation was seen as a cost but
should however induce a positive low level in order to get some headroom against deflation
as well as being a tool to adjust real wages. The monetary policy conducted by central banks
was primarily focused on short-term rates in order to adjust for inflation and changes in GDP
levels. The period before the crisis is mentioned as the great moderation where inflation was
kept low with high economic stability. (Boeckx & Cordemans, 2017)
During the financial crisis, the role of central banks and the monetary policy conducted
changed significantly. The severity of the crisis induced non-standard behavior and
innovations not previously tested. The start of asset purchase programs accommodated
liquidity needs whilst supporting financial intermediation following a rapid decrease in
interest rates. These measures led to central banks taking a new role in the economic
landscape by facilitating large asset purchases and supporting the overall economy.
Communication was also increased to guide and manage expectations of the market, interest
rate projections and asset purchase programs were communicated to a higher degree. The
rapid decreases in short-term interest rates alongside large asset purchase programs shaped
the landscape for the following years and changed the role of central banks in a distinct way.
(Boeckx & Cordemans, 2017)
2.3 Unconventional Monetary Policy
2.3.1 Quantitative Easing Programmes
Following the economic crisis of 2008, the Riksbank battled the macroeconomic downturn
via QE measures mostly guided towards financial institutions in order to obtain stability. In
the later stages of 2010 the central bank's balance sheet decreased indicating a quantitative
tightening as depicted in Figure 1. In light of growing global market insecurities and ECB QE
decisions as of 2015 the Riksbank started a purchasing program and increased its balance
sheet by 79% between the period of 2015-2019. Alongside this extensive QE program, the
Riksbank took further unconventional monetary policy measures, decreasing the interest rate
8
to -0,1%, as discussed in the following sections (Sveriges Riksbank, 2015). The period of
2015-2017 with government bond purchases measured 290 billion SEK and acted as
expansive monetary policy in times of zero-bound interest rates. This corresponded to around
44% of the outstanding nominal government bonds and further around 7% of GDP (Di
Casola & Stockhammar, 2021).
Figure 1. ECB & the Riksbank assets as % of 2014 Q4 GDP
The Euro-zone balance sheet can also be seen to react to the 2008 economic crisis and
continue to increase asset holdings with a sharp incline from 2015 and onwards. At most the
assets of the ECB amounted to 48% of the aggregated Euro-zone GDP during the observed
time period. As described by the ECB the QE programmes conducted, mostly from 2015 and
onwards acted to hold/reach the 2% inflation target. The signaling and portfolio rebalance
theory is mentioned by the ECB as plausible mechanisms for the actions to take effect. (ECB,
2016)
9
2.3.2 Negative Interest Rate And Its Implication
In contemporary times, central banks traditionally imposed positive nominal interest rates
when lending short-term funds to regulate the business cycle. Nevertheless, in recent years, a
small number of central banks ventured into negative interest rates. This unconventional
approach involves making banks pay to deposit their excess cash at the central bank,
intending to encourage increased lending to counter the lingering economic challenges after
the 2008 global financial crisis and low inflation in the following years. This shift changed
the conventional dynamic with savers incurring a negative return, while borrowers seemingly
got paid to borrow money. (Vikram Haksar, 2020)
In essence, interest represents the cost of credit or the cost of money. The compensation a
borrower agrees to pay the lender for utilizing their funds and to offset associated risks.
Various economic theories underpin interest rates, some emphasizing the interplay between
savings supply and investment demand, while others focus on the balance between money
supply and demand. Under normal circumstances, interest rates are expected to be positive,
with longer-term borrowing incurring higher rates due to increased risk. Additionally,
accounting for inflation is crucial in determining effective investment yields or loan costs, as
expectations of inflation significantly impact longer-term interest rates. (Vikram Haksar,
2020) Sweden’s central bank imposed a zero interest rate policy as of October 2014 and
entered a negative environment as of February 2015 with -0,1%. The ECB furthermore
lowered their interest rate to negative territory in June 2014 to -0,1% and amounted to -0,5%
at most. Sweden carried out a policy as low as -0,5% (Kuchler, 2020)
10
2.4 FX And Implications On Swedish Economy
Figure 4. EURSEK exchange rate
Sweden is considered an exporting nation, with exports often exceeding imports monthly,
resulting in a positive trade balance (SCB, 2023). Many domestic companies rely on
international markets to sell their products and services. Therefore, investors must predict
exchange rates and their impact on revenues and profits (Anlas, 2012). In recent years, the
Swedish krona has demonstrated a weakness compared to other currencies, such as the USD
and the EUR (Sveriges Riksbank, 2023). A weakened krona makes Swedish goods more
affordable for foreign buyers, further increasing exports and potentially leading to increased
demand for Swedish products. The withdrawal is that it becomes more expensive for
domestic companies to purchase international goods and services, potentially reducing
revenues and profits (SCB, 2023). For international and domestic investors, this can pose a
challenge as the exchange rate can have a significant impact on market returns, profitability,
and GDP development. Market instability may involve fluctuations in exchange rates,
making it difficult to anticipate the direction it will take and therefore take the correct
measures. (Sveriges Riksbank, 2023)
11
3. Literature review
Previous research in this field has proven that central banks' asset purchasing depreciates the
domestic currency. This was found in the US in 2009, where the Federal Reserve's large-scale
asset purchase program depreciated the USD (Neely, 2011). The same outcome could be
found in Europe, with a regression framework Dedola et al., (2021) proved that the EUR
depreciated against the USD as a result of the quantitative easing shock between the years
2009-2019.
A paper by Rebucci et al., (2022) investigated the effect of quantitative easing
announcements on treasury bills and USD exchange rates. They used quarterly data from 21
central banks including Sveriges Riksbank. In their results, they saw that the Swedish Krona
depreciated against the USD for the next three days after the announcement. Examining a
longer period, De Rezende & Ristiniemi (2023) applied a pooled and single regression with
the Swedish Krona as the Y-variable. Their findings also suggest that an unconventional
policy did depreciate the currency, even though the impact was larger with conventional
policies.
There are also papers published by central banks that are more focused on the positive effect
on society. A report by the Riksbank (Andersson et al, 2022) concluded that even though the
programs implemented by different central banks have varied in size, design, and execution,
some broad conclusions can be made from them. One of them is that the purchase of
securities has been an effective tool for restoring market functions during periods of financial
stress and reducing overly high premiums on more risky securities. Another insight is the
large-scale purchases of the government and mortgage bonds have lowered market interest
rates, contributing to a more expansive financial environment. The report continues by
focusing on the exchange rate. They confirm the other paper's results mentioned above, that
the Swedish krona weakened in connection with the Riksbank's announcement of government
bond purchases. (Andersson et al, 2022)
This thesis will build upon the research published by Stockhammar and Di Casola (2021),
who saw that the quantitative easing done by the Swedish Riksbank did decrease
unemployment but also as a consequence depreciated the SEK. They also tried to interpret
the effect of inflation but the results were unclear. Their paper also examined the foreign QE
12
by investigating the effect of ECB actions. Here they saw that foreign QE had a positive
spillover on Swedish GDP and inflation, focusing on the exchange rate the program did make
the real exchange rate of the SEK appreciate. This report will investigate a longer time
horizon than the years of 2015-2018, hoping to contribute to the science by including more
observations.
There has also been similar research that has examined other economies such as the UK and
the US. A well-renowned paper by Weale and Wieladek (2016) has been the template and
inspiration for research in this field. Their findings suggest that a shock in asset purchases
will increase both GDP and CPI for both countries. They were not focused on the exchange
rate but mentioned the method to include the real exchange rate into the model to explore and
prevent omitted variable bias. Their approach deviates from the BVAR model to avoid the
problem with tight priors dominating information from the data and is one of the factors for
the chosen Structural Autoregressive Vector (SVAR) model in this paper.
Another paper with a similar research focus is by Jomaa (2022) who focuses on inflation and
the consequences of a quantitative easing program. The paper focuses on three main
transmission channels, exchange rate, portfolio balancing, and signaling channel. Her results
are presented by impulse response functions, something that will be implemented in this
paper as well. The significant and positive responses on economic activity and price level for
a shock QE are consistent with the two later papers. However, the SEK appreciated the EUR,
something that contradicts economic intuition and historical rates.
According to (Jomaa, 2022) this implies that the effect on price level and GDP is not a result
of an exchange rate effect, but suggests that the effects come from some other transmission
channel. The portfolio balancing theory could be rejected as the impulse response function
does not show any evidence of lower-term spread. On the contrary, this suggests that
economic activity (GDP) is stimulated through signaling channels. This is tested by adjusting
the model, replacing the term spread with the 1-month STIBOR rate will identify signs of a
signaling channel. The impulse responses confirm the author's theory, the reference rate drops
in value as a direct response to the Swedish QE program. Once again the results confirm the
outcome by Weale and Wieladek (2014) as they also observed that the signaling channel was
the most effective channel the quantitative easing program affected the economy. Focusing on
policy rates, Eriksson and Vestin (2021) showcase that there is a clear correlation between the
13
Swedish policy rate and interest paid by household and non-financial companies. More
interesting, the correlation continues at negative levels of policy rates. Furthermore,
regardless of the level of policy rate, hikes and cuts seem to have a similar effect on the
lending rates. Therefore suggesting that the relationship is linear even at negative levels.
4. Theoretical framework
4.1 Negative Short-Term Rates
Interest implies the cost of credit and in the longer run the cost of money, representing the
amount a borrower agrees to pay the lender to compensate for using the money and incurring
the associated risk. Economic theories emphasize that interest rates can vary depending on
either the interaction between the supply of savings and demand for investments or the supply
and demand of money in its essence. The theory does however align with the fact that interest
rates have to be positive for its mechanisms to work. (Vikram Haksar, 2020)
For the majority of history nominal interest rates have been positive implying that borrowers
have to pay while lenders receive interest. Real interest rates have however throughout
history been negative with inflation surpassing nominal rates, real interest rates being
nominal rates minus inflation. A positive nominal interest rate imposed by central banks
adjusts the short-term lending costs which in turn affects the business cycle. By lowering
these interest rates below zero-bound, incentives increase for banks to lend out the funds
because they would incur zero interest payments to store these funds at the central bank.
Implying that the saver earns a negative return while borrowers, in relation, get paid to
borrow. Within an economy, several different interest rates affect the economic function and
adjust within the market. The policy rates set by countries' central banks provide a relative
rate implying the borrowing cost within the country's economy. The policy rate is therefore
adjusted to impact the economic cycle and incentivise expansion or contraction. A high
policy rate motivates saving while a low interest rate incentives consumption and
investments. (Vikram Haksar, 2020)
14
The implications of negative policy rates first off affect the bank's profitability. With banks
being a systemically important function of society they perform important matching between
savings and investments with their gain being the spread. With negative policy rates and a
reluctance to impose negative deposit rates banks decrease their spread and therefore their
profitability. The increased profitability could in turn undermine the financial stability of
economies. Another concern would be that a negative or zero-bound deposit rate incentivises
cash withdrawals, i.e. keeping the money outside of the bank system, even further
undermining the financial stability and the function of the matching principle. (Vikram
Haksar, 2020)
4.2 Mundell-Fleming model
The model originates in a series of papers by Mundell during 1960-1961 and Flemming
during 1962-1971. The model uses the IS-LM-BP framework to investigate how monetary
policy and fiscal policy affect economic variables such as income/output, trade balance,
exchange rate and capital flow. The model allows one to use various assumptions regarding
exchange rates (e.g., fixed vs. floating) and different degrees of capital mobility. Perfect
capital mobility is a scenario in which capital can flow freely across borders without any
limitations or cost, this allows for uninterrupted transfers from one country to another. The
opposite, imperfect capital mobility suggests the presence of certain restrictions or hinders
the flow of capital. There will also be significant transaction costs or other expenses, making
the process of moving capital between countries complicated and challenging. (Wang, 2020)
According to the theory under assumptions of perfect capital mobility with flexible exchange
rate. An expansive monetary policy decreases the domestic interest rate leading to a lower
interest than foreign countries. As a consequence, capital outflows cause an increase in net
current accounts and depreciate the domestic currency. In contrast to monetary policy
changes, a fiscal expansion will initially cause domestic interest to rise above the foreign
level. The discrepancy is only temporary because of the perfect capital mobility, capital
inflow will ensure that domestic interest rates will align with foreign. Results of the increase
in government spending will lead to an appreciation of the domestic currency and a
deterioration of the trade balance. (Wang, 2020)
15
4.3 Quantitative Easing
The practice of QE is a tool adopted by central banks to meet inflation targets. Adjusting the
interest rate steers the level of liquidity in money markets which can act as a monetary
stimulus in economic downturns. Using quantitative easing is one of two tools that can be
used to adjust the interest rates, the other being adjusting the policy rate. By conducting
quantitative easing the central bank will buy bonds pushing up prices and thereby lowering
the long-term interest rates. (Bank of England, 2023)
In the aftermath of the financial crisis in 2008, several central banks significantly lowered
their policy rates whereby the use of QE later became popular. As stated by the ECB there
might be two potential reasons why adjusting the policy rate might not be an adequate tool in
economic downturns. One is that adjusting policy rates loses its effectiveness when
approaching the zero lower bound which Sweden did already by October 2014. (Andreas
Kuchler, 2020) Lowering the policy rate into negative bounds may cause market participants
to stack up on cash instead of increasing their spending. Secondly, the transmission of
monetary policy can be impaired even when the policy rate is outside of the zero lower
bound, in these instances the central bank must cut the policy rate by further drastic measures
(Wieladek & Garcia Pascual, 2009). By practicing QE the central bank expands its balance
sheets through large-scale purchases of assets, commonly bonds issued by governments, large
corporations and institutions. The obtained lower yield through these purchases influences the
economy and ceteris paribus implies lower interest rates for households and corporations.
(McLeay et al., 2014)
4.3.1 Signaling Channel
The theory of signaling channels is based on imperfect market functions due to information
frictions. This implies that merely the signaling of large-scale asset purchases should affect
market practices and put downward pressure on yields. In practice, the QE might be carried
out during several years but the market reaction is expected as of the announcement of the
program. The announcements are thought to act as guiding measures affecting market
participants' expectations of future short-term interest rates which in turn lowers the
long-term interest rates. (Weale & Wieladek, 2023)
16
4.3.2 Portfolio Balance Channel
Based on the idea that financial assets exhibit imperfect substitutability, QE conducted by
central banks should affect their portfolio balance. The large-scale purchases are thought to
affect portfolios in regards to duration and risk inducing rebalancing of portfolios. The theory
relies on assumptions of investors having preferred-habit demand implying preferences for
bonds with specific maturities. If this satisfies, investors will rebalance their portfolio and
shift from holding money to financial assets. The increased demand for financial assets
reduces yields and shifts expectations for short-term interest rates and in turn long-term rates.
This theory therefore stems from the reasoning of investors not being indifferent to
rebalancing portfolios implying that investors must rebalance due to the central bank’s
large-scale purchases. The increased value of non-money assets will in turn stimulate the
economy in the desired direction. (Ben S. Bernanke, Vincent R. Reinhart, and Brian P. Sack,
2004)
4.3.3 Exchange Rate Channel
Employment of quantitative easing can further affect the exchange rate of the domestic
country due to increased prices of domestic bonds and decreased yields. This channel is of
particular interest in this thesis and will be the main focus. Changes in yields imply a reduced
demand for foreign investors leading to less capital inflows and in turn depreciating the
domestic currency. The depreciation does in turn make imported goods and services more
expensive while exports become more attractive for the surrounding countries, resulting in a
higher inflation. (Brett W. Fawley & Christopher J. Neely, 2013) As argued by Di Casola and
Stockhammar (2021) this channel is of great importance for small open economies and has
been a debated topic for Sweden as the SEK have broadly depreciated.
17
5. Empirical Framework
5.1 Data Collection Process
When observing research in similar fields, the data collection differs between studies but
some common procedures and variables can be found. The selected variables must be
explanatory at least in economic theory, and the statistical validity of the variables will be
examined. This thesis uses monthly data in the periods between January 1999 through
December 2019, thereby extending previous research within the field by 192 months. The
dataset consists of 240 months of data for each variable. The range has been limited to 2019
to not account for extraordinary measures taken during Covid-19.
Variables are adjusted to measure the same periods which implies the banking end of each
month often mentioned as ultimo. Further measures have been taken by adjusting data with
seasonality such as unemployment and inflation. The need for adjustment of these variables
stems from former research by the ECB. (Lis & Porqueddu, 2018) The sources used for the
dataset have been carefully chosen and consist of the Riksbank, Federal Reserve Economic
Data and EUROSTAT. The data have been cross-checked with Eikon from London Stock
Exchange Group (LSEG) to ensure that it is truthful and correct.
5.2 Variables Within The Model
The process of choosing suitable variables has been performed through a literature review of
research conducted of the subject. From the literature of Stockhammar and Di Casola, the
choice to include ECB variables was made. From their findings, significant spillover effect
from the unconventional monetary policy conducted by the ECB was found in the Swedish
economy, whereby we include ECB asset purchases to not experience omitted variable bias.
To examine the effect of the unconventional monetary policy imposed by the Riksbank the
nominal EURSEK Ultimo by each month is used. Swedish variables account for the domestic
effects while Euro-zone variables account for possible spillover effects. Prices are measured
by CPI for both the Euro-zone as well as the Swedish economy, the interbank rates for
Sweden and the Euro-zone are included as proxy measurement for the repo rates induced by
respective central banks. The unemployment-levels are included to account for the real
economic effects. The 10-year T-bonds are included to measure the unconventional monetary
18
effects on the long-term rates. Quantitative easing programs for the Euro-zone and Sweden
are accounted as each respective central bank's balance sheet assets scaled by GDP of Q4
2014 to eliminate endogeneity effects from contemporary GDP effects induced by the
Quantitative easing. The variable description is presented in Appendix 1.
5.3 Description Of The Data
T-Bonds for both Sweden and the Euro-zone denote government-issued bonds on its
sovereign debt with a maturity of 10 years. The bond represents the yield an investor would
obtain if held until maturity. It is further often referenced as a risk-free rate, depending on the
issuing country, and for that reason used as a reference rate. Changes in the 10-year treasury
yield represent the economic landscape and sentiment in the global market. A decline in the
10-year bond yield often indicates caution about economic conditions while an increase
represents future confidence. The bonds are issued in the first market with a face value and
coupons specifying a certain amount of interest to be paid, often sold to institutional investors
like banks or other financial companies. The bonds are then sold on the secondary market
where the yield is determined. A bond's price is inversely related to its yield implying that
when demand is high, the yield decreases. (Baldridge, 2023)
The Interbank Rate constitutes STIBOR (Stockholm Interbank Offered Rate) and a total
interbank measure for the Eurozone 19 countries. Interbank rates are used as a reference rate
among the banks and are commonly used as a base in financial contracts in the local
currency. It is often used in bonds, loans, and FX derivatives. (Sveriges Riksbank, 2022)
CPI is a commonly used inflation measure that calculates price changes within an economy.
The Swedish Riksbank uses the CPIF (with fixed interest rates) with a target of 2% while
ECB measures from the Harmonised Index of Consumer Prices (HICP) where harmonized
implies that all countries within the Euro-zone follow the same methodology for the data to
be comparable. (Sveriges Riksbank, 2023) (ECB, 2024) As both of these stem from CPI this
measurement will be used.
19
The Unemployment Rate is observed as the proportion of working-age inhabitants that were
unemployed during a certain reference week but have taken action to search for jobs and are
available to work. The data is retrieved from labor force surveys (LFS) using Y-chart.com.
For most Euro countries, where monthly survey data is not available, Eurostat approximates
the monthly unemployment rates. This is why the data between Sweden and the Euro-zone
differ in frequency while observing the data. (OECD, 2023) The measurement is chosen to
account for all inhabitants.
Central Bank Assets proxies the quantitative easing programs conducted by both the ECB
and the Riksbank. The data represents the total balance sheet assets for each respective
central bank. The Riksbank reports that their main asset class constitutes securities in SEK
and was primarily the class expanded during the QE program period. (Sveriges Riksbank,
2024) The same distribution of asset classes is seen for the ECB where bond purchases make
the majority of asset holdings and was the primary expanded class during QE-periods.
(Nordea Corporate, 2022)
EURSEK is the exchange rate between the currency of the Eurozone, EUR, and the Swedish
krona, SEK. The EUR is the base currency in the pair and the SEK is the quoted currency.
Although FX is traded around the clock five days a week this thesis uses monthly data due to
availability for other variables. The EURSEK monthly data is used with the calculation
method Ultimo which denotes the last recorded value for each traded day. (Sveriges
Riksbank, 2024)
5.4 Visualization Of The Data
The data can be divided into two categories, variables for the Eurozone and variables for the
Swedish economy. These are replicated between the categories to control for effects on the
EURSEK. The variables that account for the Euro-zone are the inflation rate CPI, the
Unemployment rate followed by the long-term interest rate the 10-year T-bond but also the
Euro Interbank rate T/N. As graphed below are the named variables, all stated as percentages
on the timeline from 1999-01 throughout 2019-12.
20
5.4.1 Euro-Zone Data
Figure 5. Euro-zone data variables
As can be seen from figure 5, large effects on the inflation and interbank T/N rate were
observed in the aftermath of the financial crisis of 2008. With large rate cuts of the ECB repo
rate the interbank quickly followed as can be depicted by the Euro interbank data. Inflation
decreased at a rapid pace in the following economic recession. Unemployment increased to
over 10% while the long-term rates held quite still. When the ECB entered zero-bound
interest rates by July 2012 a decreased yield on the T-bond can be observed. Inflation can be
seen to decline before this date but decreased rapidly following zero to negative interest rates.
21
5.4.2 Swedish Data
Figure 6. Sweden data variables
As for the Swedish data, the same variables are analyzed, and distinct effects can be seen for
this data also around the 2008 financial crisis. The unemployment rate increased rapidly
exceeding 9% while the long-term yield decreased alongside rapid short-term rate cuts.
Inflation decreased during the recession and experienced a great decline at the start of the
financial crisis. When negative interest rates were implied in February 2015 a small trend can
be observed on the unemployment rate but also the long-term yield alongside movements in
the inflation rate.
22
5.4.3 Central Bank Assets
Figure 7. ECB & the Riksbank assets as % of 2014 Q4 GDP
The data for central bank assets are divided by GDP as of Q4 2014 to mitigate endogeneity
concerns stemming from the impact of QE on contemporaneous GDP levels following Di
Casola & Stockhammar's (2014) reasoning. Notably, the Riksbank used QE programs for a
short period following the 2008 financial crisis, the same increase in assets can be seen for
the European Central Bank but to a lower percentage of GDP. A clear upward trend of added
assets can be seen for both central banks after the announced QE programmes in early 2015.
The purchases in the Euro-zone amounted to almost 50% of 2014 Q4 GDP at most while the
Riksbank assets reached approx. 22% of 2014 Q4 GDP.
23
5.5 Structural Vector Autoregression
Former research within the area has shown consensus for using VAR-models to examine the
effects of quantitative easing programs. Structural Bayesian VAR is employed by
Stockhammar & Di Casola (2014) with an identification scheme inspired by Weale and
Wieladek (2016). Jomaa (2022) uses a Structural VAR (SVAR) with the same identification
scheme as Weale and Wieladek to study the Swedish Riksbank’s Quantitative Easing
Programme on Swedish inflation. This thesis will employ a Structural Vector Autoregression
model that includes variables formerly introduced for EURSEK alongside relevant Swedish
and Euro-zone variables.
The VAR model can capture effects even though the data is of low frequency as in the choice
of monthly data. It further enables the estimation of causal relationships which is of particular
interest in this thesis of the transmission channel from QE and negative interest rates. If
compared with event studies the VAR model enables capturing real economic effects due to
implementation of identification schemes. It is further applicable to use the VAR model when
handling time series data where it is of interest to understand interdependencies. (Weale &
Wieladek, 2014)
Understanding the relations between the interdependent variables enables deeper analysis and
discussion of the real economic impacts. A structural VAR imposes the possibility to use
economic theory and prior knowledge within the subject to impose identifying assumptions
of the relationships among the variables. This enables isolated structural shocks on variables
of interest. (Weale & Wieladek, 2014)
The results from the Structural VAR are then interpreted using impulse response functions.
Impulse response functions are shown graphically and display how a shock to one variable
will lead to a response in another over a certain time period. When examining the effect, one
should look at the direction, magnitude, and persistence. The estimated effects may often
change during the time period, and the graph captures that change. The graphs also show a
credible interval, which here is a confidence interval as we use classical estimation methods.
(Ouliaris et al, 2018)
24
5.6 Econometric Model
The model is estimated with monthly data from 1999 to 2019 with the following equation:
𝑌 = Σ𝐿  𝐴 𝑌 + 𝑒       ,      𝑒 ∼ (0, Σ)  
𝑡 𝑘=1 𝑘 𝑡−𝑘 𝑡 𝑡 
Variable 𝑌 decomposes a vector of the endogenous set of variables at time t. The endogenous
𝑡
variables consist of EURSEK, Sweden T-bond, STIBOR, Sweden CPI, Swedish
Unemployment, and the Riksbank assets. 𝐴 are coefficient matrices corresponding to each
𝑘
lag of k and 𝑒 show the vector of residuals at time t. The VAR model estimates the
𝑡 
relationship between the EURSEK and chosen variables as well as their interdependencies.
5.7 Model Specification And Strategy
Recall that variables used in the Structural VAR model are the following:
Variables
Euro-zone Sweden
ECB Assets The Riksbank Assets
Eurozone interbank STIBOR
Euro CPI Sweden CPI
Euro Unemployment Sweden Unemployment
Euro T-bond Sweden T-bond
EURSEK
Table 1. Summary of variables
The data is chosen as monthly data with a starting date of 1999-01 throughout 2019-12
totaling 240 months. CPI is observed to contain seasonality and is therefore adjusted with
Arima X-12 before entering the system. The unemployment was seasonally adjusted by
YCharts.com before being downloaded. These adjustments provide an alternative perspective
on current developments in macroeconomic series, enabling month-to-month comparison
25
without the impact of seasonal and calendar effects (IMF, 2017). A VAR model is run with
ECB and the Riksbank Assets being first differenced to obtain stationary data and to remove
trends (Vossen, 2020). Specification of all variables is endogenous with Euro-data being first
in the ordering and the Swedish variables being placed after in the same slots as the
Euro-data. The VAR is run and lag length is tested firstly at twelve lags2 since the analysis is
performed on monthly data. The test for VAR lag length shows results for LR test statistic,
Final Prediction Error, Akaike, Schwarz but also Hanna-Quinn criterion simultaneously.
Jomaa (2022) states that some studies can rely on the information criterion mentioned above
but that others rely on economic theory and intuition. Her conclusion emphasizes the
importance of lag length, too many lags increase forecasting errors and too few may impose
estimation bias.
The FPE indicates a lag length of three which aligns with other results within the area
previously mentioned in the literature review. The VAR model is performed once again with a
lag length of three and tested for stationarity with an Autoregressive (AR) roots test. This
tests for the presence of a unit root which in turn indicates a non-stationarity meaning that the
model's mean, variance, and autocovariance structure do not remain constant over time. The
test shows that all variables are below the value threshold of one which indicates a stationary
process, the lag length of three is hereby confirmed.
Interpreting shocks in structural VAR models poses a challenge due to the difficulty in
distinguishing between orthogonal and structural economic shocks, which can be complex to
disentangle from the reduced form shocks. To distinguish these the Cholesky decomposition
method is often used, requiring authors to order the variables in a specific way (Morh, 2020).
The ordering of variables is decided by a lower triangular identification scheme used in
Weale and Wieladek (2024) and can be viewed in the appendix. The structure is hierarchical
and assumes that variables earlier in the order can affect those later in the order but not vice
versa. (Weale & Wieladek, 2023) Therefore the euro-variables are first as they could affect
Swedish variables. T-bond accounting for the long-term rates are first in order as changes in
yields often reflect market expectations and movements. Followed by the short-term rate
2 Three other SVAR-models were created, each with 6, 9 and 12 lags. The findings from these tests concluded
that three lags was the optimal. Adding additional lags only made the results unclear and vague. Furthermore,
there was no indication that adding more lags led to clearer and significant results when examining CPI and
unemployment. See appendix three and four.
26
interbank T/N that represents the cost of money and liquidity constraints within an economy,
this is chosen to proxy central bank repo rates. Next in order is the inflation measure CPI
which is thought to capture responses to interbank and yield changes. Unemployment is
placed fourth as an indicator of the general economy. The assets of central banks are next in
order which is thought to capture the quantitative easing programmes.
5.7.1 Robustness Tests For Exogeneity
Robustness-tests are performed to control for Swedish variables not affecting Euro-variables.
By economic theory a small economy and its economic activity should not affect or influence
a larger Euro-zone economy. Stockhammar and Di Casola (2022) discuss the theorem in their
paper where Euro variables are deemed exogenous in the analysis of the Swedish context as
they are assumed to not be affected by the variables of the small open economy. The outer
lines in orange represent the confidence intervals. Robustness tests for exogeneity of the
Euro-variables show that Swedish variables have a significant impact on Euro-zone variables
such as Sweden's T-bond innovation on Euro Interbank and Swedish CPI innovation on Euro
interbank response.
Hereby the euro-variables are deemed exogenous from Stockhammar and Di Casola (2022)
alongside our discussion. The model that will be further analyzed sets Euro-zone variables as
exogenous. The identification scheme of lower triangular holds and the lag length is as
previously chosen to three. The final model and its impulse responses are to be introduced
and discussed in the following sections.
Figure 8: Euro Interbank response on Figure 9: Euro Interbank response on
Sweden T-bond innovation Sweden CPI innovation
27
6.Results
6.1 EURSEK Response To Unconventional Monetary Policy
Figure 10: Riksbank balance sheet response to a positive Riksbank balance sheet innovation
Figure 11: EURSEK response to a positive Riksbank balance sheet innovation
28
Observing the effects on the exchange rate EURSEK when the Riksbank expands their
balance sheet shows that the SEK depreciates. As the response holds a positive value this
implies that the EURSEK increases in value meaning that the quoted currency SEK
depreciates. These results are in line with the former research of Stockhammar & Casola
(2014) who observed a depreciated SEK during periods of QE-programmes 2015-2018.
As the theory of exchange rate channel states the effect on the exchange rate is partly due to
interest rate differentials between countries and decreased yields. (Brett W. Fawley &
Christopher J. Neely, 2013) When the Riksbank increases their assets by 0,61 percentage
units, in relation to the 2014 Q4 GDP-level, the EURSEK depreciates by 0,56 percent which
is a significant effect within the credible intervals.
Findings also suggest a 1 percentage unit increase of the balance sheet against the 2014 level
will lead to an depreciation of 0,9 percent on the EURSEK. In December of 2014, total
Riksbanken assets as of 2014’s GDP was 12,35 percent and rose to 22,87 percent in
December 2019, a total increase of 10,52 percentage-units. Multiplying 10,52 with the
elasticity results in 9,49 percent. This concludes that the QE program depreciated the SEK by
9,46 percent ceteris paribus.
29
Figure 12: STIBOR response to a STIBOR innovations
Figure 13: EURSEK response to a STIBOR innovation
30
The exchange rate further responds to STIBOR which is to proxy the interest rates impacted
by unconventional monetary policy. The effect of an increase in STIBOR interbank rate leads
to a significant appreciation of the SEK against the EUR in the shorter term. The inverse, i.e a
negative cut of the interest rate is therefore thought to depreciate the SEK against the EUR
which is within our field of interest. When STIBOR increases by 0,14 percentage units the
EURSEK appreciates by 0,4 percent.
Furthermore this gives elasticity where 1 percentage unit decrease in short-term rate will
depreciate the SEK by 2,86 percent. There are similar results proving monetary policy easing
leads to a depreciation of a country’s nominal exchange rate. More interesting, the paper by
Inoue et Rossi (2018) investigated this relationship both in conventional and unconventional
monetary periods. Using shifts in the yield curve to identify shifts in monetary policy they
saw that shocks in both types of period had relatively similar effects. Especially in the US,
where their implementation of monetary policy easing led to depreciation of the US dollar
exchange rate. This gives validity to our linear results where the effect examines both
increases and decreases of the interest rate.
6.2 Further Interdependencies
Figure 14: CPI response to the Riksbank balance sheet innovation
31
The effects on domestic inflation are less clear where the credible intervals contain zero.
Similar results can be found in Stockhammar & Casola (2014). However, Weale & Wieladek
(2016) finds contrary results where their identification schemes, except for the UK, show
significant findings where QE increases the inflation.
Figure 15: STIBOR response to the Riksbank balance sheet innovation
The effect on Swedish interbank STIBOR has a significant response in the short term where
an innovation of the Riksbank balance sheet results in a sharp decrease in STIBOR, aligning
with signaling theory. The maximum effect is seen in the short term when the Riksbank
increases their balance sheet by 0,61 percentage units as of 2014 Q4 GDP then STIBOR can
be seen to decrease by -0,11 percentage units. The elasticity implies that a 1 percentage unit
increase of Riksbank assets decreases STIBOR by -0,18 percentage units.
32
Figure 16: Unemployment rate response to the Riksbank balance sheet innovation
The observed effects on the unemployment rate shows a small but significant short term
effect where an increase of 0,63 percentage units in Riksbank assets increases the
unemployment rate by 0,05 percentage units. The corresponding elasticity implies that a 1
percentage unit increase in Riksbank assets increases the unemployment rate by 0,079
percentage units. The results are ambiguous as most intervals are insignificant in the later
perspectives. The increase in unemployment is in contrast to Stockhammar and Di Casola
(2014) who find a lowering effect of domestic QE on the unemployment rate.
33
Figure 17: Sweden T-bond to the Riksbank balance sheet innovation
A significant effect of Riksbank balance sheet expansion is observed where a 0,61 percentage
unit increase of the Riksbank assets result in a -0,05 percentage units decrease of the long
term rate. These results are in line with economic theory where the signaling of QE lowers
the interest expectations within the economy and should therefore be represented in the long
term yield. The elasticity indicates that a 1 percentage unit increase of Riksbank assets
decreases the yield by -0,082 percentage units. This is similar to results found by Weale &
Wieladek (2016) who conclude a decrease of US long term rates when QE is conducted.
34
7.Discussion
The chosen dateperiod limits the study to a time period of 240 months between the years of
1999-01 through 2019-12. These dates stem from firstly the introduction of the EUR in and
the closing date is chosen in order to avoid Covid-19 effects and the extraordinary measures
taken in the following years. This limits our study to 240 months but does at the same time
extend previous research which primarily has focused on the years 2015 through 2018
amounting to 36 months.
The choice of variables stem from previous studies within the field and economic intuition.
This could however be expanded to also include measures for the general economy such as
oil prices, GDP development or stock-indices. The insignificant results for unemployment
and CPI might be due to the exclusion of other variables such as these. The study further
focuses on the domestic effects of unconventional monetary policy and hence no spillover
effects on other economies. Previous literature has investigated spill-over effects from foreign
quantitative easing, something that has not been examined or discussed in this paper.
However, ECB asset purchases are included in order to not exhibit omitted variable bias. The
euro-variables have been set as exogenous, meaning that they can affect the swedish variables
but not vice versa. This implies that economies such as the US or larger ones in Asia have not
been included which could tilt our results.
This thesis primarily researches the quantitative easing programmes and its effect on the
EURSEK, short-term rates that went into negative territory in 2014 and onwards are proxied
for but not controlled for the specific negative effect. Instead our results describe the
relationship between EURSEK and STIBOR where an increase/decrease affects the EURSEK
in a specific way. That relationship is not specific to a negative territory shock but instead has
the same effect as if the change of short term rates would have been from 0% to -0,5% or a
raise to 2,5%. Future research should therefore focus on isolating the effect of interest rates
becoming negative from positive territories as this represents an historical shift.
Furthermore, the linearized SVAR model poses some other assumptions. If the Riksbank
policy rate was indeed perceived by financial markets and households to be truncated at -0.5
percent, our linearized framework can lead to biased effects as it does not fully capture
effects stemming from an effective lower bound on the policy rate. Examining the extent to
35
which -0.5 percent was indeed perceived as an effective lower bound and the robustness of
our results when imposing such a constraint in the estimation are two issues we leave for
future research. Inclusion of other currency pairs like USDSEK, NOKSEK, GBPSEK etc
would also make for relevant research. This could control for a broader effect and show
interesting results in regards to other international or closeby economies. A richer model with
more variables should also be examined as this might validate the insignificant results.
8.Conclusion
As described throughout this paper quantitative easing became a common method among
leading central banks in advanced economies to stimulate the economy following the global
financial crises. Some central banks - like the Swedish Riksbank also lowered policy rates to
negative levels. The effect of these initiatives were at the time not known but the slowdown
in economic activity, elevated unemployment rates and below target inflation had to be
addressed with any means necessary when fiscal policy remained tight. The full longtime
effect of these strategies have still not been fully explored. The results that we provide in this
thesis show that the quantitative easing program did in fact depreciate the SEK with nominal
effect of 1 percentage unit increase (from the 2014 level of GDP) of asset purchasing will
depreciate the EURSEK by 0,9 percent. Since Riksbank long-term asset holdings rose by
10,52 percent from December 2014 to December 2019 , our results suggest that this has
contributed to a 9,49 percent weaker Swedish Krona.
This confirms other results that have been discussed, but also adds on to the research by
claiming this is also true in a longer period (1999-01 until 2019-12). The results also
strengthen the exchange theory by Fawley and Neely (2023) saying that a decrease in yields
will depreciate the domestic currency by capital outflow seeking a greater return. By
including the 2008 crisis we have captured initial purchasing programs as well as the initial
lowering of policy rates. We believe these observations are vital for painting a full picture of
unconventional monetary policy as we now have observations indicating a substantial
decrease in interest rates, rather than just a low baseline level.
36
The other side of unconventional monetary policy is the negative policy rate. Due to various
constraints this paper has only investigated the responses of the EURSEK after innovations
(shocks) in short term rate (STIBOR) and not explicitly allowed the possibility in
transmission differs for negative policy rates. While work by Riksbank staffers (Erikson and
Vestin, 2021), find that transmission between policy and market rates are similar when policy
rates are positive and negative, we believe our conclusions for negative rates should be
interpreted with some caution. The former US Secretary of the Treasury Lawrence Summers
published an opposing view where he argues that the effect of entering negative territory in
fact imposes hurtful effects on the economy. Implying that cuts from positive to negative
rates should have unique effects and not linear as Eriksson and Vestin argue (Summers,
2015). With this in mind, our evidence indicates that positive shocks in the short term rate
depreciate SEK against the EUR with a relatively considerable reaction.
The elasticity is calculated to be 2,86, meaning that 1 percentage unit decrease in STIBOR
will appreciate SEK against the EUR 2,86 percent. As previously discussed, the results are in
line with the findings from Inoue et Rossi (2018). Another notable key factor from their
report are their findings regarding the unconventional policy. They concluded that the effects
are quantitatively similar regardless of the two policies (conventional and unconventional).
Therefore the argument of not dividing the two policies has scientific validity as discussed
earlier, and that both can be meaningfully used to raise inflation and economic activity via a
weaker exchange rate.
Combining the two policy instruments, calculations suggest that the total effect of the 10,52
percent unit increase in Riksbanks total assets from the baseline level (i.e. annualized GDP in
2014 Q4) would decrease the EURSEK by 9,49 percent. Furthermore, looking at STIBOR
that decreased -50 basis points during the same period. Our results suggest that STIBOR
contributed about 1,43 percent depreciation in EURSEK. Adding these two, advocates a total
effect of 10,92 percent depreciation in EURSEK during December 2014 to December 2019
by the unconventional monetary policy. Comparing it with the actual rate the SEK
depreciated with 11,5 percent during the same period, deeming our results plausible. These
results carry some uncertainty because the credible intervals widen over time, suggesting that
the actual effect might be either higher or lower than the aggregated result. However, based
on the model's function, the aggregated result of an 11 percent depreciation is the best
approximation.
37
References
Andersson et.al. (2022, December 31). Riksbanksstudie Riksbankens köp av värdepapper
2015-2022. Riksbanken. Retrieved February 6, 2024, from
https://www.riksbank.se/globalassets/media/rapporter/riksbanksstudie/svenska/2022/ri
ksbanksstudie-riksbankens-kop-av-vardepapper-2015-2022.pdf
Andreas Kuchler. (2020, October 2). Negative interest rates: The Danish experience. CEPR.
Retrieved November 20, 2023, from
https://cepr.org/voxeu/columns/negative-interest-rates-danish-experience
Baldridge, R. (2023, October 23). What Is The 10-Year Treasury Yield? Why Is It So
Important? – Forbes Advisor. Forbes. Retrieved April 2, 2024, from
https://www.forbes.com/advisor/investing/10-year-treasury-yield/
Bank of England. (2023, January 31). Quantitative easing. Bank of England. Retrieved
February 6, 2024, from
https://www.bankofengland.co.uk/monetary-policy/quantitative-easing
Ben S. Bernanke, Vincent R. Reinhart, and Brian P. Sack. (2004). Monetary Policy
Alternatives at the Zero Bound: An Empirical Assessment. Federal Reserve Board.
Retrieved February 6, 2024, from
https://www.federalreserve.gov/pubs/feds/2004/200448/200448pap.pdf
Brett W. Fawley & Christopher J. Neely. (2013). Four Stories of Quantitative Easing. Federal
Reserve Bank of St. Louis. Retrieved February 19, 2024, from
https://files.stlouisfed.org/files/htdocs/publications/review/13/01/Fawley.pdf
Cheng, K. C. (2006). A VAR Analysis of Kenya's Monetary Policy Transmission Mechanism:
How Does the Central Bank's REPO Rate Affect the Economy? INTERNATIONAL
MONETARY FUND.
ECB. (2016, January 22). How does the ECB's asset purchase programme work? European
Central Bank. Retrieved April 3, 2024, from
https://www.ecb.europa.eu/ecb-and-you/explainers/tell-me-more/html/app.en.html
Erikson, H., & Vestin D. (2021, May 28). Pass-through of negative policy rates. Sveriges
Riksbank. Retrieved May 7, 2024.
How is inflation measured? | Sveriges Riksbank. (2023, September 14). Riksbanken.
Retrieved April 2, 2024, from
https://www.riksbank.se/en-gb/monetary-policy/the-inflation-target/how-is-inflation-
measured/
IMF. (2017). Quarterly National Accounts Manual; Seasonal Adjustment. International
Monetary Fund. Retrieved April 2, 2024, from
https://www.imf.org/external/pubs/ft/qna/pdf/2017/chapter7.pdf
38
Inoue, A., & Rossi, B. (2019, February 23). The effects of conventional and unconventional
monetary policy on exchange rates. Journal of International Economics, 118(1),
419-447.
https://pdf.sciencedirectassets.com/271695/1-s2.0-S0022199619X00021/1-s2.0-S002
2199618302800/main.pdf?X-Amz-Security-Token=IQoJb3JpZ2luX2VjEMj%2F%2F
%2F%2F%2F%2F%2F%2F%2F%2FwEaCXVzLWVhc3QtMSJIMEYCIQD62u6LLa
mLyTJDS5itMdoTw2c5pV86Q%2Fkf4hQTzNBeqwIhANZb6yjO8eD
Jef Boeckx & Naïm Cordemans. (2017, January 6). Monetary Policy in the Wake of the Great
Recession. Cairn. Retrieved April 4, 2024, from
https://www.cairn.info/revue-reflets-et-perspectives-de-la-vie-economique-2017-1-pa
ge-61.htm
Jomaa, N. (2022). The Effects of Quantitative Easing on Swedish Inflation. Lund University.
Retrieved April 4, 2024, from
https://lup.lub.lu.se/luur/download?func=downloadFile&recordOId=9113318&fileOI
d=9113319
Kandil, M., & Mirzaie, I. A. (2003). The effects of Exchange Rate Fluctuations. IMF.
https://www.imf.org/external/pubs/ft/wp/2003/wp03200.pdf
Lis, E., & Porqueddu, M. (2018). The role of seasonality and outliers in HICP inflation
excluding food and energy. ECB. Retrieved April 4, 2024, from
https://www.ecb.europa.eu/pub/pdf/other/ebbox201802_06.en.pdf
McLeay, M., Radia, A., & Thomas, R. (2014, March 14). Home Money creation in the
modern economy. Bank of England. Retrieved February 6, 2024, from
https://www.bankofengland.co.uk/quarterly-bulletin/2014/q1/money-creation-in-the-
modern-economy
Nordea Corporate. (2022). Week Ahead: Could the ECB balance sheet contraction outpace
the Fed? Nordea. Retrieved April 4, 2024, from
https://corporate.nordea.com/article/74929/week-ahead-could-the-ecb-balance-sheet-c
ontraction-outpace-the-fed
OECD. (2023). Unemployment - Unemployment rate. OECD Data. Retrieved April 2, 2024,
from https://data.oecd.org/unemp/unemployment-rate.htm
Paola Di Casola & Pär Stockhammar. (2021). When domestic and foreign QE overlap:
evidence from Sweden. Riksbanken. Retrieved February 19, 2024, from
https://www.riksbank.se/globalassets/media/rapporter/working-papers/2021/no.-404-
when-domestic-and-foreign-qe-overlap-evidence-from-sweden.pdf
Riksbanken. (2009, June 3). Minutes of the monetary policy meeting held on 11 February
2015. YouTube: Home. Retrieved February 6, 2024, from
https://archive.riksbank.se/en/Web-archive/Published/Minutes-of-the-Executive-Boar
ds-monetary-policy-meetings/20151/Minutes-of-the-monetary-policy-meeting-held-o
39
n-11-February-/index.html
Riksbanken. (2015). Negativ reporänta införs | Sveriges Riksbank. Riksbanken. Retrieved
November 14, 2023, from
https://www.riksbank.se/sv/om-riksbanken/historia/historisk-tidslinje/2000-2018/nega
tiv-reporanta-infors/
Riksbanken. (2024). Räntor och valutakurser. Riksbanken. Retrieved April 2, 2024, from
https://www.riksbank.se/sv/statistik/rantor-och-valutakurser/sok-rantor-och-valutakurs
er/?s=g130-SEKEURPMI&fs=2#riksbank-seriesform
Riksbanken. (2024, February 5). Riksbankens balansräkning | Sveriges Riksbank.
Riksbanken. Retrieved April 2, 2024, from
https://www.riksbank.se/sv/statistik/riksbankens-balansrakning/
SCB. (2022, February 28). Net trade balance at SEK 0.4 billion in January. SCB. Retrieved
November 22, 2023, from
https://www.scb.se/en/finding-statistics/statistics-by-subject-area/trade-in-goods-and-s
ervices/foreign-trade/foreign-trade---exports-and-imports-of-goods/pong/statistical-ne
ws/exports-imports-and-net-trade-balance-january-2022-in-current-price/
SCB. (2023, April 14). Sveriges export. SCB. Retrieved November 20, 2023, from
https://www.scb.se/hitta-statistik/sverige-i-siffror/samhallets-ekonomi/sveriges-export
/
S. Ouliaris, A.R. Pagan, J. Restrepo. (2018 August). Quantitative Macroeconomic Modeling
with Structural Vector Autoregressions – An EViews Implementation.
STIBOR | Sveriges Riksbank. (2022, September 15). Riksbanken. Retrieved April 2, 2024,
from
https://www.riksbank.se/en-gb/statistics/interest-rates-and-exchange-rates/explanation
s--interest-rates-and-exchange-rates/swedish-market-rates/stibor/
Summers, L. H. (2015). Demand Side Secular Stagnation. American Economic Association.
Retrieved May 8, 2024, from
https://www.aeaweb.org/articles?id=10.1257/aer.p20151103
Sveriges Riksbank. (2023, January 1). Valutakursindex | Sveriges Riksbank. Riksbanken.
Retrieved November 20, 2023, from
https://www.riksbank.se/sv/statistik/sok-rantor--valutakurser/forklaring-till-serierna/v
alutakursindex/
Sveriges Riksbank. (2023, July 7). Så påverkar penningpolitiken inflationen | Sveriges
Riksbank. Riksbanken. Retrieved November 20, 2023, from
https://www.riksbank.se/sv/penningpolitik/vad-ar-penningpolitik/sa-paverkar-penning
politiken-inflationen/
Tomasz Wieladek & Antonio Garcia Pascual. (2009, June 3). THE EUROPEAN CENTRAL
40
BANK’S QE: A NEW HOPE. Centre for Economic Policy Research. Retrieved
February 6, 2024, from https://cepr.org/system/files/publication-files/DP11309.pdf
Trading Economics. (2023). Sweden Interest Rate. Trading Economics. Retrieved November
14, 2023, from https://tradingeconomics.com/sweden/interest-rate
Tulin Anlas. (2012). The Effects of Changes in Foreign Exchange Rates On ISE-100 Index.
Journal of Applied Economics and Business Research.
http://www.aebrjournal.org/uploads/6/6/2/2/6622240/3._anlas.pdf
Two per cent inflation target. (n.d.). European Central Bank. Retrieved April 2, 2024, from
https://www.ecb.europa.eu/mopo/strategy/pricestab/html/index.en.html
Vikram Haksar. (2020, 3). Back to Basics: What Are Negative Interest Rates? – IMF F&D.
International Monetary Fund. Retrieved November 20, 2023, from
https://www.imf.org/en/Publications/fandd/issues/2020/03/what-are-negative-interest-
rates-basics
Vossen, V. (2020). The First-Difference Estimator. Tilburg Science Hub. Retrieved April 2,
2024, from
https://tilburgsciencehub.com/topics/analyze/causal-inference/panel-data/first-differen
ce/
Wang, P. (2020, 08 29). The Mundell–Fleming Model. The Economics of Foreign Exchange
and Global Finance, 1(1), 149-172. 10.1007/978-3-662-59271-7_7
Weale & Wieladek. (2023, May 19). What are the macroeconomic effects of asset purchases?
Science Direct. Retrieved February 6, 2024, from
https://www.sciencedirect.com/science/article/pii/S0304393216300101
Weinberg, J. (2013). The Great Recession and Its Aftermath. Federal Reserve History.
Retrieved April 4, 2024, from
https://www.federalreservehistory.org/essays/great-recession-and-its-aftermath
41
Appendix
Appendix 1. Variables with units, source, adjustment & transformation
42
Appendix 2. Lower triangular identification scheme
Appendix 3. EURSEK response to a positive Riksbank balance sheet innovation using 12 lags
43
Appendix 4. EURSEK response to a STIBOR innovation using 12 lags
Appendix 5. Riksbank balance sheet response to a STIBOR innovation
44
Appendix 6. STIBOR response to a Riksbank balance sheet innovation
45