Sensitivity of stock prices to changes in the term structure

(1)

Sensitivity of Stock Prices to Changes in the Term Structure

Abstract

This paper studies the sensitivity of stock prices to changes in the term structure of interest

rates using data from German and French companies in the period from 2008 until 2016.

Specifically, the relations between stock prices of banks and non-financial firms and changes

in the level, slope and curvature of the term structure are quantified using a fixed effects panel

regression analysis. Within non-financial firms a further distinction is made between

relatively levered and unlevered firms. I find that stock prices are primarily driven by changes

in the level of the term structure. In line with existing literature, this paper finds a negative

relation between share prices and the level of the term structure for France, though the results

show a positive relation for Germany. The findings for both countries are in line with the

efficient market hypothesis, in the sense that only unexpected interest rate movements affect

stock prices.

Tommie van Beek

Bachelor Thesis Economics and Finance

Supervised by Rob Sperna Weiland

Faculty of Economics and Business

30-01-2017

(2)

Statement of Originality

This document is written by student Tommie van Beek who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

1. Introduction

The money supply in the euro area has exceeded the boundary of 10000 billion euro's last year (measured by the broadest monetary aggregate). Every month, an 80 billion euro's asset purchase program is conducted seeking to enhance inflation (ECB, Asset Purchase Program, 2017). As stated in

De Volkskrant, however, the ECB does not succeed in achieving its goal since inflation in the euro

area currently amounts to merely 0.5% , which is well below the targeted rate of 2%. It is argued that an actual result of the expansionary monetary policy is an increase in stock prices (Heagens, 2016). While interest rates decreased to historically low levels, stock markets increased significantly. The German and French stock indices, for example, increased with over 60% and 25%, respectively. Economic theories suggest a negative relationship between interest rates and stock prices. Lower interest rates decrease the cost of capital, make equity more attractive relative to bonds and reflect lower inflation expectations which increase the real return on nominal assets. Hence, a decrease in the interest rate is expected to increase stock prices, a relation confirmed by empirical results. Flannery and James (1984), Bae (1990) and Choi et al. (1992), for example, find that a decrease in the interest rate increases stock prices of financial companies. More recently, Oertman et al. (2000), Bernanke and Kuttner (2005) and Czaja et al. (2009) also conclude that lower interest rates increase stock prices of financial companies. Lynge and Zumwalt (1980) and Ferrer et al. (2016), however, find results that contradict the general consensus of a negative relationship and find that a decrease in the interest rate decreases stock prices.

This thesis studies the effect of interest rate changes on stock prices of German and French companies by considering the term structure of interest rates. Specifically, the effect of changes in the term structure on both financial and non-financial companies will be examined. Within non-financial companies, a further distinction will be made between relatively levered and unlevered firms. Although the effect of interest rate changes on stock prices has been studied extensively, the tremendous drop in interest rates since the start of the financial crisis provides an interesting occasion to re-examine this relationship. This is because it is not clear whether the existing theories still hold in a period of historically low interest rates. The explicit focus on non-financial companies is what distinguishes this thesis from most of the existing literature. Furthermore, this paper also takes the slope and the curvature of the term structure into account, while most existing literature solely focuses on a level variable.

This thesis finds contradictory results for Germany and France. The results for the French companies in the sample show that a decrease in the interest rate increases stock prices, a relation in line with the existing literature. For the German companies in the sample, however, I find that a decrease in the interest rate decreases share prices, which contradicts the existing literature. I find that share prices of both German and French companies are primarily driven by changes in the level of the term structure.

(5)

Furthermore, the results for both countries are in line with the efficient market hypothesis, in the sense that only unexpected interest rate movements affect stock prices.

The remainder of this paper is structured as follows: Section 2 outlines the economic theories and presents previous findings. Section 3 describes the methodology and the data. Section 4 presents and discusses the results. Section 5 concludes.

2. Literature Review

In this section, I outline multiple economic theories that address a relation between interest rates and stock prices, describe previous findings, and outline the hypothesis underlying the empirical study. 2.1 Dividend Discount Model

The Law of one Price implies that the value of a security is equal to the present value of the cash flows the security generates, independent of the investment horizon of the investor (Berk and DeMarzo, 2014). Since the cash flows of a share consist of the future dividend payments, the price of a share should be equal to the present value of all future dividend payments. The present value of eventually selling the stock equals zero as the number of years tends to infinity. According to Berk and DeMarzo (2014) this leads to the general dividend discount model where the share price can be computed as follows:

𝑃0 = � 𝐷𝑖𝑣𝑛/(1 + 𝑟𝑐)𝑛 ∞

𝑛=1

According to Homa and Jaffee (1971), the discount rate used to calculate the present value of the dividends consists of a risk free component and a risk premium. Therefore they argue that lowering the risk free interest rate increases stock prices since it increases the present value of the future dividends. Hamburger and Kochin (1972) also state that the discount rate consists of a risk free component and a risk premium, so that there should exist a negative relationship between the interest rate and stock prices. The same concept is phrased slightly different by Blanchard (1981) who states that an increase in the money supply decreases the interest rate which in turn lowers the cost of capital used to calculate the present value of the share price. As a result, this lower cost of capital increases the market value of stocks. Lastly, Hashemzadeh and Taylor (1988) and Ferrer et al. (2016) also argue that a lower interest rate raises the present value of future dividend payments and therefore increases stock prices.

2.2 Portfolio Rebalancing

A second relationship between interest rates and stock prices reflects the ability of investors to rebalance their portfolios that consist of bonds and stocks (Apergis and Eleftheriou, 2002). This means

(6)

that a decrease in the interest rate makes interest bearing assets such as bonds less attractive relative to shares (Alam and Uddin, 2009). Hence, Erdem et al. (2009) argue that a decrease in the interest rate causes investors to change the composition of their portfolios in favor of shares, which due to higher demand for shares increases the share prices. This theory is also outlined by Schiller and Beltratti (1992). They state that a decrease in long-term bond yields makes bonds a less attractive investment relative to shares, which raises share prices. Thus, the theory of rebalancing portfolios also implies a negative relationship between the interest rate and share prices.

2.3 Nominal Contracting Hypothesis and the Fischer Equation

Flannery and James (1984) state that to examine the relationship between interest rate changes and share prices the return of a firm should be considered to consist of two components: the return on nominal assets and the return on physical assets.

𝑅𝑠𝑡𝑜𝑐𝑘= 𝛾 ∗ 𝑅𝑛𝑜𝑚𝑖𝑛𝑎𝑙+ (1 − 𝛾) ∗ 𝑅𝑝𝑦𝑠𝑖𝑐𝑎𝑙

The return on nominal assets can be considered as cash flows that are fixed in nominal terms and therefore do not vary as the price level changes (Flannery and James, 1984). According to Flannery and James (1984) the nominal return of physical assets however, does vary with changes in the price level since the cash flows generated by physical assets are not fixed in nominal terms.

The nominal contracting hypothesis states that a decrease in the price level increases the real return of nominal assets and consequently, this higher real return increases the share price (Flannery and James, 1984). In contrast, the real return on physical assets is not affected by changes in the price level and therefore the proportion of nominal assets held determines to what degree a firm's share price is affected by changes in the price level (French, Rubak and Schwert, 1983). Since financial firms hold a larger fraction of nominal assets, it is expected that they exhibit a larger sensitivity to interest rate changes than non-financial firms.

The link between inflation and the interest rate stems from the Fisher equation (Mankiw, 2013):

𝑖 = 𝑟 + 𝜋 . The Fischer equation states that the nominal interest rate equals the real interest rate plus the inflation. Since investors cannot perfectly predict inflation, Fama and Schwert (1977) state a slightly different version of the Fischer equation where the nominal interest rate equals the real interest rate plus the expected inflation. Fischer stated that the real interest rate depends on real factors like productivity and is not influenced by nominal variables such as the price level and the nominal interest rate (Fama and Schwert, 1977). Hence, they state that in efficient markets, a decrease in nominal interest rates reflects a decrease in inflation expectations.

(7)

Combining the nominal contracting hypothesis and the Fischer equation leads to a third theory which implies a negative relationship between interest rates and share prices. As Fama and Schwert (1977) argue, a decrease in the nominal interest rate is the result of a decrease in inflation expectations. If markets are efficient, this decrease in inflation expectations increases the real return of the nominal assets and thus increases share prices (Flannery and James, 1984).

2.4 Implications of the Efficient Market Hypothesis

Fama (1970) introduced the theory of efficient markets based on the idea that asset prices fully reflect all available information. If asset prices indeed completely incorporate all available information, the market is said to be efficient (Fama, 1970). Pearce and Roley (1983) state in their study of the effect of changes in the money supply on stock prices, that a basic implication of the efficient market hypothesis is that only the unexpected change in the money supply should affect stock prices. This is because the expected change is already fully incorporated in the stock prices. Correspondingly, only unexpected changes in interest rates should affect stock prices if the efficient market hypothesis holds. 2.5 Term Structure

This thesis considers the term structure to measure changes in interest rates. The relation between the time to maturity and the yield of a specific bond is called the term structure of interest rates, also referred to as the yield curve (Berk and DeMarzo, 2014). The expectations theory implies that bonds of different maturities are perfect substitutes. As stated by Mishkin et al. (2013), this means that the interest rate on a long-term bond equals the average of the expected short-term interest rates. Hence, according to the expectations theory, an upward sloping yield curve implies that short-term interest rates are expected to rise in the future, while an inverted yield curve implies that short-term interest rates are expected to fall (Mishkin, Matthews and Giuliodori, 2013). Czaja et al. (2009) state that the term structure consists of three major components, namely the level, the slope and the curvature. The level of the term structure measures the average level of the current interest rates, while the slope and the curvature signal information about future interest rates and could therefore also affect share prices. 2.6 Previous Findings

Bae (1990) examines the interest rate sensitivity of stock returns of US financial companies. This is measured by regressing the return of an equally-weighted portfolio of financial companies on the percentage change in interest rates using the market return as a control variable. Bae (1990) measures the change in the interest rate by the change in the yield of the month Treasury bill, the three-year Treasury note and the twenty-three-year Treasury bond. He finds a negative relationship between interest rate changes and stock returns of financial companies. Furthermore, he concludes that stock prices are more sensitive to interest rates of longer maturity and only affected by unexpected interest rate changes.

(8)

Bernanke and Kuttner (2005) study the effect of a change in the federal fund rate on stock prices and also distinguish between expected and unexpected interest rate changes. They find a significant negative relationship between the unexpected change in the federal fund rate and stock prices. When looking at subsectors, they find that utilities are not significantly affected by interest rate changes, while the technology sector is one of the most sensitive sectors to interest rate changes.

Apergis and Eleftheriou (2002) investigate the effect of both inflation and interest rates on stock prices of companies traded at the Athens Stock Exchange. They measure the interest rate by the three-month yield on Greek Treasury bills. They conclude that the interest rate is positively related to stock prices, however, this relationship is insignificant. In contrast, the coefficient of inflation is negative and significant.

Flannery and James (1984) study the effect of interest rate changes on common stock returns of banks and stock savings and loans associations (S&L) in the United States. They use an equally-weighted portfolio of banks and S&L stocks as the dependent variable while controlling for the market return. The returns on three default-free bonds are used to measure the interest rate change. They find a significant positive effect of the return on bonds on equally-weighted portfolios of banks and S&L stocks. Since the return on a bond is inversely related to the yield, this also implies a negative relationship between stock prices and interest rates.

Choi et al. (1992) state that the profits of banking institutions are a direct function of the interest rate, which implies that the interest rate is important for the valuation of the common stock of banking institutions. Hence, they examine this relationship. They use the yield on the three-month Treasury bill as the interest rate and an ARIMA model to measure the expected values of this yield. Consequently, they use the residuals of the ARIMA model as a proxy for the unexpected change. They find a significant negative relationship between the return on a portfolio of US banks and the unexpected change in the yield.

Ferrer et al. (2016) do not focus on the US but on Europe. They study the relation between the ten-year government bond and the market index of ten European countries. The sample period is from 1993 until 2012. They find a positive relationship between the yield on ten-year government bonds and stock market indices. This relationship is strong for Germany, France, the Netherlands, Finland and the UK while substantially smaller for peripheral European countries.

Czaja et al. (2009) focus specifically on German financial institutions. They use an orthogonalization process in order to capture both the direct and the indirect effect of interest rate changes on stock returns. Czaja et al. (2009) do not look at a specific interest rate but use the term structure of interest rates as the independent variable. German government bonds of various maturity classes are used to measure the term structure while the market return is used as a control variable. Furthermore, instead

(9)

of only looking at changes in the level of term structure, the effect of changes in the slope and curvature of the term structure are also examined. They find the level effect to be the single most important factor negatively affecting the return of all portfolios considered.

Oertmann et al. (2000) study the effect of interest rate changes on portfolios of different financial companies as well as non-financial companies, and consider several European countries. Their research method differs from other papers in that they also include an explanatory variable used to measure the international interest rate change. Both domestic and international interest rates are measured by ten-year government bonds. They find a negative relation between the domestic interest rate and stock prices of financial companies for every country examined. Furthermore, stock prices of most internationally operating banks are also positively affected by a decrease in foreign government bond yields.

2.7 Aim of the Research

Based on the previous findings, I expect that a decrease in the level of the term structure has a positive effect on share prices of both financial and non-financial companies. This negative relation is supported by multiple economic theories outlined in the previous section. The dividend discount model, the ability of investors to rebalance their portfolios and the nominal contracting hypothesis combined with the Fischer equation all imply that share prices should increase upon a decrease in the interest rate. According to the expectations theory, a decrease in the slope of the term structure implies a decrease in expected short-term interest rates (Mishkin, Matthews and Giuliodori, 2013). Since the economic theories outlined suggest a negative relationship between interest rates and stock prices, I also expect share prices to increase upon a decrease in the slope of the term structure.

Furthermore, I expect that financial and non-financial companies exhibit different sensitivity to changes in the term structure. According to the nominal contracting hypothesis the sensitivity to interest rate changes depends on the firms proportion of nominal assets (Flannery and James, 1984). Since financial firms have a larger fraction of nominal assets, it is expected that they are more sensitive to changes in the term structure than non-financial firms. In addition, Ferret et al. (2016) state that a large fraction of the revenues and expenses of banks is directly dependent on the interest rate, which is another reason why financial firms are expected to be more sensitive to changes in the term structure than non-financial firms.

With respect to non-financial firms, relatively levered firms are expected to be more sensitive to changes in the term structure than relatively unlevered firms. This is because levered firms incur relatively more debt payments, which are directly related to the interest rate (Korkeamäki, 2011). Furthermore, the unexpected change is expected to be significant, while the expected change in the

(10)

term structure is expected to be insignificant because expected changes should already be embedded in the current share price.

The yield on German and French government bonds has dropped dramatically since the beginning of the financial crisis. It is not clear to what extend the existing theories and previous findings can be generalized to a period of historically low interest rates. This makes a re-examination of the relation between interest rates and stock prices is interesting.

The previous literature has focused on financial companies although the economic theories outlined also suggest a relation between interest rate changes and the stock prices of non-financial companies. Therefore, this paper also focuses explicitly on non-financial companies. A further distinction within non-financial companies is made between relatively levered and unlevered firms. Furthermore, since the slope and curvature of the term structure contain information about future interest rates, this paper also takes changes in the slope and curvature into account, whereas most existing literature solely focuses on a level variable.

3. Methodology

3.1 Methodology

In order to study the effect of the term structure on stock prices, I run the following fixed effects regression using panel data for both Germany and France:

𝑅𝑝𝑜𝑟𝑡𝑓𝑜𝑙𝑖𝑜𝑖𝑡 = 𝛽0+ 𝛽1𝑅𝑚𝑡+ 𝛽2∆𝐿𝑒𝑣𝑒𝑙𝑡+ 𝛽3∆𝑆𝑙𝑜𝑝𝑒𝑡+ 𝛽4∆𝐶𝑢𝑟𝑣𝑡+ 𝛽5𝐵𝑎𝑛𝑘 ∗ ∆𝐿𝑒𝑣𝑒𝑙𝑡

+ 𝛽6𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝐿𝑒𝑣𝑒𝑙𝑡+ 𝛽7𝐵𝑎𝑛𝑘 ∗ ∆𝑆𝑙𝑜𝑝𝑒𝑡+ 𝛽8𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝑆𝑙𝑜𝑝𝑒𝑡

+ 𝛽9𝐵𝑎𝑛𝑘 ∗ ∆𝐶𝑢𝑟𝑣𝑡+ 𝛽10𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝐶𝑢𝑟𝑣𝑡 + 𝜖𝑖𝑡

Where 𝑅𝑝𝑜𝑟𝑡𝑓𝑜𝑙𝑖𝑜_𝑖𝑡 is the weekly return of an equally-weighted portfolio i at time t, 𝑅𝑚 is the weekly return of respectively the DAX30 (Germany) or the CAC40 (France), ∆𝐿𝑒𝑣𝑒𝑙 is the weekly change in the level of the term structure, ∆𝑆𝑙𝑜𝑝𝑒 is the weekly change in the slope of the term structure, and ∆𝐶𝑢𝑟𝑣 is the weekly change in the curvature of the term structure. Due to the low level of the interest rates, absolute changes in the term structure are preferred over the highly volatile relative changes. Furthermore, 𝐵𝑎𝑛𝑘 is a dummy variable that equals 1 if the portfolio consists of banks and 𝐿𝑒𝑣𝑒𝑟𝑒𝑑 is a dummy variable that equals 1 if the portfolio consists of relatively levered non-financial firms. This model is estimated using a panel data fixed effects regression. Fixed effects regression controls for omitted variables that vary across entities but are constant over time, and hence is preferred to analyze different types of companies.

The regression answers the research questions since t-tests on the estimated coefficients specify whether stock prices are affected by changes in the three components of the term structure.

(11)

Furthermore, the significance level of the interaction variables indicate whether banks and levered financial firms exhibit different sensitivity to changes in the term structure than unlevered non-financial firms. To test whether stock prices respond to unexpected or expected changes in the components of the term structure, I decompose the total weekly change of all term structure components into an expected and unexpected part, measured as outlined in the next subsection. Consecutively, I run the fixed effects regression now using the unexpected and the expected part of the total change in the term structure components as independent variables.

This methodology is mainly in line with Stone (1974), Lynge and Zumwalt (1980), Flannery and James (1984), Bae (1990), Choi et al. (1992), Oertman et al. (2000), Czaja et al. (2009) and Korkeamäki (2011). The scope is extended to levered and unlevered non-financial firms and the effect of changes in the slope and curvature of the term structure are added as additional interest rate factors. Panel data is used because by using panel data it can be tested whether banks, levered non-financial firms and unlevered non-financial firms exhibit significantly different sensitivities to changes in the term structure.

3.2 Definitions of Variables

The dependent variables are weekly returns of portfolios consisting of banks, levered non-financial companies and unlevered non-financial companies, respectively. Following Bae (1990) the portfolios will be equally-weighted. Weekly returns are preferred over daily returns since this avoids the noise and anomalies contained in daily data (Ferrer, Bolos and Benítez, 2016).

In line with Veronesi (2010) the level of the term structure is measured by taking the average yield on the government bonds of the following maturities: 3-month, 6-month, 1-year, 2-year, 3-year, 5-year, 7-year and 10-7-year. The slope is measured by subtracting the yield on the 3-month government bond from the yield on the 10-year government bond. Lastly, the curvature of the term structure is measured as follows: minus the 3-month yield plus two times the 5-year yield minus the 10-year yield (Veronesi, 2010).

The market return will be adjusted in order to circumvent the correlation between the market return and the term structure components. In line with Bae (1990), Hirtle (1997), Czaja et al. (2009) and Korkeamäki (2011) this is achieved by running an auxiliary regression of the market return on the term structure components. The error term of this regression represents all systematic shocks apart from changes in the term structure and will be used as control variable. This way, the indirect effect of changes in the term structure that affect stock prices through the market return is also captured by the independent variables of interest.

In order to measure the expected and unexpected changes in the term structure components an autoregressive model is used, meaning that the changes in the term structure components are regressed

(12)

on their lagged values. Only time lags that are significant will be included in the autoregressive equation. Consequently, the fitted values of the autoregressive equation capture the expected change in the term structure components, while the residuals of the autoregressive equation capture the unexpected change. The fitted values and the residuals will then be used as independent variables in order to test whether stock prices are affected by expected or unexpected changes in the term structure components. The same method is used by Flannery and James (1984), Bae (1990), Czaja et al. (2009) and Korkeamäki (2011).

To measure the leverage of a firm the debt-to-equity ratio is used. The average debt-to-equity ratio over the last five years will be considered since this is a more representative measure than the current debt-to-equity ratio. Equity is defined as the book value of shareholders equity. The book value of equity is preferred over market value since the book value is not affected by large changes in share prices. Debt is measured as the total value of long-term debt plus the current portion of long-term debt. Similar measures are used by Anderson and Reeb (2003) and Lewellen and McConnel (1979). 3.2 Data

The sample period is from September 2008 until September 2016. The period from September 2011 until September 2016 will be examined separately. During this period Mario Draghi has been president of the ECB and conducted an extremely expansionary monetary policy. Furthermore, this period discards the turmoil after the outbreak of the financial crisis and might therefore result in more stable estimates.

The weekly stock prices of all companies examined are retrieved from Yahoo Finance. The weekly returns of the DAX 30 and the CAC 40 as well as the yields on all government bonds are retrieved from Investing.com. The debt-to-equity ratios as defined earlier in this section are retrieved from Yahoo Charts. All data is processed using Stata.

All banks from the main stock index in France (CAC 40) and the CAC Next 20 are included. Non-financial companies are included based on their debt-to-equity ratio as outlined in this section. CAC 40 incorporated companies with the lowest debt-to-equity ratio are included as unlevered companies, while CAC 40 companies with the highest debt-to-equity ratio are included as levered companies. German companies are included according to the same approach. Charts displaying the historical stock prices of all companies examined are presented in the appendix.

(13)

The following German companies are included:

Company Type D/E Ratio

Deutsche Bank Bank - Commerzbank Bank - Aareal Bank Bank - Daimler Levered 1.7 BMW Levered 2.1 ThyssenKrupp Levered 2.7 Adidas Unlevered 0.29 Henkel Unlevered 0.26 Infineon Unlevererd 0.20

The following French companies are included:

Company Type D/E Ratio

Crédit Agricole Bank - BNP Paribas Bank - Société Generale Bank - Natixis Bank - Carrefour Levered 1.7 Peugot Levered 2.4 Veolia Environment Levered 1.8 L'Oréal Unlevered 0.07 Lous Vouitton M.H Unlevered 0.33 Sanofi Unlevered 0.27

(14)

4. Results

This section presents the results separately for both countries and discusses the findings. 4.1 Results

The model as outlined in the methodology is estimated using a fixed effects regression in order to test the effect of changes in the term structure on stock prices. Table 1 contains the results of Germany.

Specifically, Period refers to the time period considered and Change refers to whether the independent variables consist of the total change, the expected change or the unexpected change . As mentioned, Bank and Levered are dummy variables that equal 1 if the portfolio consists of banks or levered companies respectively, and ∆ refers to the weekly change in the specific term structure component.

Table 1: Sensitivity of German stock prices to changes in the term structure components

Period 08-16 08-16 08-16 11-16 11-16 11-16

Change Total Expected. Unexpected Total Expected. Unexpected

∆𝐿𝑒𝑣𝑒𝑙 .18*** .17 .18** .07** -.23 .08** (.07) (.59) (.09) (.04) (.27) (.04) ∆𝑆𝑙𝑜𝑝𝑒 -.02 .08 .03 .03 .13 .02 (.05) (.34) (.05) (.02) (.12) (.02) ∆𝐶𝑢𝑟𝑣 .02 -.05 -.04 -.03* .07 -.03* (.05) (.38) (.05) (.02) (.12) (.02) 𝐵𝑎𝑛𝑘 ∗ ∆𝐿𝑒𝑣𝑒𝑙 .26*** .75 .23* .29*** 1.00*** .27*** (.10) (.83) (.12) (.05) (.38) (.05) 𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝐿𝑒𝑣𝑒𝑙 .11 .24 .09 .09* .12 .09* (.10) (.83) (.12) (.05) (.38) (.05) 𝐵𝑎𝑛𝑘 ∗ ∆𝑆𝑙𝑜𝑝𝑒 -.04 -.22 -.02 -.04 -.19 -.03 (.07) (.49) (.08) (.03) (.16) (.03) 𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝑆𝑙𝑜𝑝𝑒 .03 .04 .03 .05* .04 .05* (.07) (.49) (.08) (.03) (.16) (.03) 𝐵𝑎𝑛𝑘 ∗ ∆𝐶𝑢𝑟𝑣 .25*** .40 .26*** .02 -.33* .03 (.07) (.54) (.08) (.03) (.18) (.03) 𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝐶𝑢𝑟𝑣 -.03 -.07 -.02 -.01 .05 -.01 (.07) (.54) (.08) (.03) (.18) (.03) 𝑅𝑚 1.17*** 1.16*** 1.17*** 1.06*** 1.06*** 1.07*** (.09) (.10) (.09) (.03) (.03) (.03) Constant .01*** .01** .01* .003*** .003*** .002*** (.00) (.00) (.00) (.00) (.00) (.00) R2 _0.19 _0.12 _0.19 _0.67 _0.57 _0.66 Observations 1242 1242 1242 786 786 786

(15)

The results imply a positive relation between the level of the term structure and stock prices. A 1 percentage point increase in the level increases the stock price of unlevered-companies by 18%, while this effect is even 44% for banks. However, the average weekly level change in absolute values is only 0.05 percentage points. In both sample periods, the positive effect is statistically significant which contradicts most of the existing literature. The slope of the term structure is insignificant in both periods while the curvature has a significant negative effect only in the period from 2011 until 2016. This is in line with Czaja et al. (2009). They also find an insignificant effect of the slope factor in almost all cases, while they find the curvature to be significant in some cases.

In line with the expectations, banks bear greater interest rate sensitivity than unlevered non-financial companies. This effect is statistically significant in both periods. Furthermore, levered non-financial firms bear greater sensitivity to the level factor than unlevered non-financial firms, although not significantly different in the latter period.

When the total change is decomposed into an expected and unexpected part, the results show that stock prices respond to unexpected changes in the term structure and not to the changes that are already expected. Only in two cases an expected change is significant. Bae (1990) also finds a few significant effects of expected changes.

Since the share price of Commerzbank decreased with 95% over the complete sample period, this might have influenced the results. I therefore run the same regression excluding Commerzbank. The results, which are presented in the appendix, show that the sensitivity of banks to changes in the level of the term structure decreases, but the main findings remain the same.

(16)

4.2 Results France

I run the same set of regressions now using French data. Table 2 contains the empirical results.

Specifically, Period refers to the time period considered and Change refers to whether the independent variables consist of the total change, the expected change or the unexpected change . As mentioned, Bank and Levered are dummy variables that equal 1 if the portfolio consists of banks or levered companies respectively, and ∆ refers to the weekly change in the specific term structure component.

The results show that the level effect is insignificant over the whole period. However, when the financial crisis is disregard the level effect is negative and significant for both banks and non-financial companies. This is in line with the expectations. A 1 percentage point decrease in the level of the term structure increases the return of unlevered non-financial companies with 4%. The interaction terms

Table 2: Sensitivity of French stock prices to changes in the term structure components

Period 08-16 08-16 08-16 11-16 11-16 11-16

∆𝐿𝑒𝑣𝑒𝑙 .01 .18 .01 -.04** -.17 -.04** (.02) (.23) (.02) (.02) (.20) (.02) ∆𝑆𝑙𝑜𝑝𝑒 -.00 .08 -.00 -.03 .20 -.04* (.01) (.09) (.01) (.02) (.14) (.02) ∆𝐶𝑢𝑟𝑣 -.00 .03 -.01 .00 .08 -.00 (.01) (.06) (.01) (.02) (.10) (.02) 𝐵𝑎𝑛𝑘 ∗ ∆𝐿𝑒𝑣𝑒𝑙 .06** .50 .05* .00 .08 .01 (.03) (.33) (.03) (.03) (.28) (.03) 𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝐿𝑒𝑣𝑒𝑙 .02 .24 .02 .01 -.06 .01 (.03) (.33) (.03) (.03) (.28) (.03) 𝐵𝑎𝑛𝑘 ∗ ∆𝑆𝑙𝑜𝑝𝑒 -.05*** -.18 -.05*** -.05* -.32* -.04 (.02) (.12) (.02) (.03) (.20) (.03) 𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝑆𝑙𝑜𝑝𝑒 -.02 -.09 -.02 -.03 -.16 -.03 (.02) (.12) (.02) (.03) (.20) (.03) 𝐵𝑎𝑛𝑘 ∗ ∆𝐶𝑢𝑟𝑣 .00 -.05 .01 -.02 -.48*** .00 (.02) (.08) (.02) (.02) (.14) (.02) 𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝐶𝑢𝑟𝑣 .00 .02 .00 .02 -.03 .03 (.02) (.08) (.02) (.02) (.14) (.02) 𝑅𝑚 1.14*** 1.13*** 1.14*** 1.19*** 1.18*** 1.19*** (.02) (.02) (.02) (.03) (.03) (.03) Constant .002*** .007*** .002*** .003*** .002* .003*** (.00) (.00) (.00) (.00) (.00) (.00) R2 _0.65 _0.64 _0.64 _0.65 _0.65 _0.65 Observations 1242 1242 1242 786 786 786

(17)

show that banks and levered non-financial companies are not differently affected, although different sensitivity was expected.

The slope and curvature are mainly insignificant in both periods. Only banks bear significant sensitivity to changes in the slope of the term structure. A decrease in the slope of the term structure increases stock prices of banks during both sample periods. According to the expectations theory, this is in line with the expectations.

Dividing the total change in the term structure components into an expected and unexpected part show that the unexpected part affects stock prices. When the independent variables consist of expected changes, this results in only two significant estimates and moreover, the magnitude of these coefficients imply that they are not reliable. As mentioned, it is not uncommon to find a few significant effects of expected changes.

4.3 Discussion

The main results for the French companies during the complete sample period are insignificant. This can be caused by the turmoil after the outbreak of the financial crisis which might have disturbed the relation between interest rates and stock prices. When the crisis is disregarded in the 2011-2016 period, the main findings are in line with the existing literature. A decrease in the level of the term structure increases stock prices. Moreover, stock prices of banks also rise upon a decrease in the slope of the term structure. According to the expectations theory, a decrease in the slope of the term structure implies a decrease in the expected short-term interest rates, which in turn implies an increase in stock prices according to the previous findings. Therefore, contingent on the expectations theory, the negative coefficient of the slope factor is in line with the existing literature.

French banks do not exhibit different sensitivity to level changes compared to non-financial firms in the 2011-2016 period. A possible explanation could be that the profitability of banks is also affected by interest rate changes. In a study to the determinants of bank profitability, Bourke (1989) argues that the value added by banks is the loan interest minus the deposit interest. A decrease in the interest rates can affect this margin. Especially in a period of historically low interest rates, the depository rate at which banks can fund their activities is almost equal to zero. A further decrease in the yield might then imply that only the lending rate decreases further and therefore the profitability of banks decreases. This is in line with the findings of Bourke (1989) and might partly offset the positive effect the economic theories described.

Ferret et al. (2016) state that levered firms are more sensitive to interest rates than unlevered firms since levered firms incur more debt payments. On the other hand, Korkeamäki (2011) argues that the introduction of the euro has improved the ability of European firms to hedge interest rate risk. This

(18)

could be an explanation for the fact that relatively levered and unlevered firms in France show similar sensitivity to changes in the level of the term structure.

In contrast to France, I find that a decrease in the level of the term structure decreases stock prices of all German companies examined. Caballero and Krishnamurty (2008) argue that in a period of financial uncertainty investors switch from equities to safe assets such as bonds, which decreases both share prices and the yield on government bonds. Hence, this could be an explanation for the positive relation between the level of the term structure and the German companies in the sample. Furthermore, a decrease in the level of the term structure might also have affected the profitability of German banks, resulting in a positive effect.

The question that remains is why the relation between the level of the term structure and stock prices is negative for the French companies but positive for the German companies in the sample. With respect to banks, the different effect can be caused by a different asset composition. If French banks hold relatively more government bonds than German banks this could partly explain the different effect, since a decrease in the yield increases the price of the bond. Another reason could be the fraction of assets that is invested in real estate. Real estate prices increase upon a decrease in interest rates, since a decrease in interest rates increases demand for real estate. So if French banks invest more in real estate and hold relatively more government bonds compared to German banks, this could explain the fact that banks in both countries are affected differently by changes in the level of the term structure. Why non-financial companies in both countries are affected differently by a change in the level of the term structure is a subject that needs to be further investigated.

This paper has implications for policy makers. The ECB defines maintaining financial stability as one of its tasks and therefore it is relevant to study in what way monetary policy affects share prices. This paper finds that share prices of German and French companies are both significantly affected by the changes in the term structure. Hence, the ECB should take into account the effect monetary policy has on stock markets with respect to maintaining financial stability.

5. Conclusion

This paper studies the relations between the term structure and stock prices of German and French companies. Specifically, the effect of changes in the level, slope and curvature of the term structure on stock prices of banks, levered non-financial companies and unlevered non-financial companies is quantified using a fixed effects panel regression analysis. The sample period is from the start of the financial crisis in 2008 until 2016. The unexpected and expected changes of the term structure components are also considered.

(19)

I find that stock prices are primarily driven by changes in the level of the term structure. The results for the French companies studied are mainly in line with the existing literature. Namely, I find a negative relation between the level of the term structure and share prices of all companies during the period from 2011 until 2016. Regarding France, this paper finds no different sensitivity of levered and unlevered non-financial companies to changes in the level component. According to Korkeamäki (2011) this can be explained by the increased ability euro area companies have to hedge interest rate risk.

I find a positive relation between the level of the term structure and all German companies examined. This contradicts the existing literature. A possible explanation is the fact that during periods of uncertainty and turmoil investors switch from equities to safe assets (Caballero and Krishnamurty, 2008). With respect to banks, a decrease in interest rates could also decrease their profitability (Bourke, 1989), which might explain the positive effect I find for the German banks in the sample. The different effect for German and French banks can be explained by a different asset composition. Further research is needed to explain the different effects for German and French non-financial companies. The interesting results also provide an inducement to re-examine the effect in the same period using a greater sample and more euro area countries. This is especially relevant because this paper implies that monetary policy affects stock prices while the ECB defines maintaining financial stability as one of its tasks.

(20)

References

Alam, M., & Uddin, G. (2009). Relationship between Interest Rate and Stock Price: Empirical evidence from developing and developed countries. International Journal of Business and

Management. 4 (3).

Anderson, R., & Reeb, D. (2003). Founding Family Ownership, Corporate Diversification, and Firm Leverage. The Journal of Law and Economics. 46 (2).

Apergis, N., & Eleftheriou, S. (2002). Interest rates, inflation and stock prices: the case of the Athens Stock Exchange. Journal of Policy Modeling. 24 (3).

Bea, S. (1990). Interest rate changes and common stock returns of financial firms: revisited. The

Journal of Financial Research. 13 (1).

Berk, J., & DeMarzo, P. (2014). Corporate Finance. Prentice Hall.

Bernanke, B., & Kuttner, K. (2005). What explains the Stock Market's reaction to the Federal reserve

policy? The Journal of Finance. 60 (3).

Blanchard, O. (1981). Output, the Stock Market, and Interest Rates. The American Economic Review.

71 (1).

Bourke, P. (1989). Concentration and Other Determinants of Bank Profitability in Europe, North America and Australia. Journal of Banking and Finance. 13 (1).

Caballero, R. & Krishnamurthy, A. (2008). Risk Management in a Flight to Quality Episode. The

Journal of Finance. 63 (5).

Choi, J., Elyasiani, E., & Kopecky, K. (1992). The sensitivity of bank stock returns to market, interest and exchange rate risk. Journal of Banking and Finance. 16 (5).

Czaja, M., Scholz, H., & Wilkens, M. (2009). Interest rate risk of German financial institutions: the impact of level, slope and curvature of the term structure. Review of Quantitative Finance and

Accounting. 33 (1).

ECB. (2017, January). Asset Purchase Programmes. Retrieved from:

https://www.ecb.europa.eu/mopo/implement/omt/html/index.en.html

Erdem, C., Arslan, C., & Erdem., M. (2005). Effect of macroeconomic variables on Istanbul stock exchange indexes. Applied Financial Economics. 15 (14).

(21)

Fama, E. (1970). Efficient Capital Markets: A Review of Theory and Empirical Work. The Journal of

Finance. 25 (2).

Flannery, M., & James, C. (1984). The effect of interest rate changes of common stock returns of financial companies. The Journal of Finance. 13 (1).

Ferrer, R., Bolos. V., & Benítez, R. (2016). Interest rate changes and stock returns: A European multi-country study with wavelets. International Review of Economics and Finance. 44 (1).

French, K., Ruback, R., & Schwert, G. (1983). Effects of nominal contracting on stock returns.

Journal of Political Economy. 91 (1).

Lewellen, W., & McConnel, J. (1979). Financial Leverage Clienteles. Journal of Financial

Economics. 7 (1).

Hamburger, M., & Kochin, L. (1972). Money and stock prices: The channel of influences. The Journal

of Finance. 27 (2).

Hashemzadeh, N., & Taylor, P. (1988). Stock prices, money supply, and interest rate: the question of causality. Applied Economics. 20 (12).

Heagens, K. (2016, November 3). ECB-president Mario Draghi als redder van Europa? Retrieved from: http://

https://www.nd.nl/nieuws/economie/toch-minder-duits-dan-hij-leek.2286447.lynkx

Hirtle, B. (1997). Derivatives, portfolio composition, and bank holding company interest rate risk exposure. Journal of Financial Services Research. 12 (2).

Korkeamäki, T. (2011). Interest rate sensitivity of European stock markets before and after the euro introduction. Journal of International Financial Markets, Institutions & Money. 21 (5). Lynge, M., & Zumwalt, J. (1980). An Empirical Study of the Interest Rate Sensitivity of Commercial

Bank Returns: A Multi-Index Approach. The Journal of Financial and Quantitative Analysis.

15 (3).

Mankiw, N. (2013). Macroeconomics. Macmillan.

Homa, K., & Jaffee, D. (1971). The supply of money and common stock prices. The Journal of

Finance. 26 (5).

Mishkin, F., Matthews, F., & Giulidori., M. (2013). The Economics of Money, Banking & Financial

(22)

Oertmann, P., Rendu, C., & Zimmerman, H. (2000). Interest rate risk of European Financial Corporations. European Financial Management. 6 (4).

Pearce, D., & Roley, V. (1983). The Reaction of Stock Prices to Unanticipated Changes in Money: A Note. The Journal of Finance. 38 (4).

Schiller, R., Beltratti, A. (1992). Stock prices and bond yields. Can their co movements be explained in terms of present value models. Journal of Monetary Economics. 30 (1).

Stock, J., & Watson, M. (2015). Introduction to Econometrics. Pearson.

Stone, B. (1974). Systematic Interest-Rate Risk in a Two-Index Model of Returns. The Journal of

Financial and Quantitative Analysis. 9 (5).

(23)

Appendix

Level of the Term Structure

Germany Unlevered companies: -2 0 2 4 6 September 2008 - September 2016 Germany -2 0 2 4 6 September 2008 - September 2016 France 0 50 100 150 200 Septermber 2008-September 2016 Adidas 0 50 100 150 Septermber 2008-September 2016 Henkel 0 5 10 15 20 Septermber 2008-September 2016 Infenion

(24)

Levered Companies: Banks: 0 10 20 30 40 Septermber 2008-September 2016 Thyssen Krupp 0 20 40 60 80 100 Septermber 2008-September 2016 Daimler 0 50 100 150 Septermber 2008-September 2016 BMW 0 10 20 30 40 50 September 2008-September 2016 Deutsche Bank 0 20 40 60 80 100 120 140 September 2008 - September 2016 Commzerbank 0 10 20 30 40 September 2008 - September 2016 Aareal Bank

(25)

France Unlevered companies: Levered companies: 0 50 100 150 200 September 2008 - September 2016 L'Oréal 0 10 20 30 40 50 60 September 2008 - September 2016 Sanofi 0 50 100 150 200 September 2008 - September 2016 Louis Vuitton MH 0 10 20 30 40 September 2008 - September 2016 Carrefour 0 5 10 15 20 25 September 2008 - September 2016 Peugot

(26)

Banks: 0 5 10 15 20 25 September 2008 - September 2016 Veolia Environment 0 5 10 15 September 2008 - September 2016 Crédit Agricole 0 20 40 60 80 September 2008 - September 2016 BNP Paribas 0 10 20 30 40 50 60 September 2008 - September 2016 Société Generale 0 2 4 6 8 September 2008 - September 2016 Natixis

(27)

German results excluding Commerzbank

Specifically, Period refers to the time period considered and Change refers to whether the independent variables consist of the total change, the expected change or the unexpected change . As mentioned, Bank and Levered are dummy variables that equal 1 if the portfolio consists of banks or levered companies respectively, and ∆ refers to the weekly change in the specific term structure component

Table 3: Sensitivity of German stock prices to changes in the term structure components

Period 08-16 08-16 08-16 11-16 11-16 11-16

∆𝐿𝑒𝑣𝑒𝑙 .22*** .18 .18*** .07** -.23 .08** (.02) (.18) (.02) (.04) (.26) (.04) ∆𝑆𝑙𝑜𝑝𝑒 .01 .08 .03** .03 .13 .02 (.01) (.10) (.01) (.02) (.11) (.02) ∆𝐶𝑢𝑟𝑣 -.05*** -.05 -.04*** -.03* .07 -.03* (.01) (.10) (.01) (.02) (.12) (.02) 𝐵𝑎𝑛𝑘 ∗ ∆𝐿𝑒𝑣𝑒𝑙 .09*** -.25 0.11*** .25*** 1.15*** .23*** (.03) (.26) (.03) (.05) (.37) (.05) 𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝐿𝑒𝑣𝑒𝑙 .10*** .25 .09*** .09* .12 .09* (.03) (.26) (.03) (.05) (.37) (.05) 𝐵𝑎𝑛𝑘 ∗ ∆𝑆𝑙𝑜𝑝𝑒 .00 .12 -.00 -.03 -.20 -.02 (.02) (.15) (.02) (.03) (.16) (.03) 𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝑆𝑙𝑜𝑝𝑒 .03 .04 .03 .05 .04 .05* (.02) (.15) (.02) (.03) (.16) (.03) 𝐵𝑎𝑛𝑘 ∗ ∆𝐶𝑢𝑟𝑣 .02 -.09 .02 -.00 -.36** .00 (.02) (.16) (.02) (.03) (.17) (.03) 𝐿𝑒𝑣𝑒𝑟𝑒𝑑 ∗ ∆𝐶𝑢𝑟𝑣 -.03 -.07 -.02 -.01 .05 -.01 (.02) (.16) (.02) (.03) (.17) (.03) 𝑅𝑚 1.12*** 1.11*** 1.12*** 1.07*** 1.06*** 1.07*** (.02) (.03) (.03) (.03) (.03) (.03) Constant .001*** .005*** .003*** .004*** .004*** .003*** (.00) (.00) (.00) (.00) (.00) (.00) R2 _0.69 _0.53 _0.67 _0.67 _0.59 _0.67 Observations 1242 1242 1242 786 786 786

Sensitivity of stock prices to changes in the term structure