Business Cycles and the Stock Market

In 2007, the commodity bubble contributed to the Great Recession. The information technology bubble, in the late 1990s, was followed by a recession. These observations bring this question: how do business cycles, i.e. alternating economic expansions and recessions, relate to fluctuations in the various sectors of stock market?

To answer for this question, I have analyzed monthly gross domestic product (GDP) in Canada and Standard & Poor's (S&P) indices of various sectors of Toronto Stock Exchange (TSX).

S&P/TSX Indices and GDP, Canada, 1998:M1-2015:M11

In Canada, an effect of the Great Recession was a continuous fall in GDP from Octobeer 2008 to May 2009. At the same time, over six consecutive months starting from September 2008, the stock market plunges. Stocks in the sector of diversified metal and mining lost 62.75% of their values. In the energy sector, they lost 51% of their values. In both the information technology and industrial sectors, the value of stocks plummetted by about 44%. Only the gold sector grew by 9.26%. The lost in the consumer staples sector was much lower (4.94%).

In Canada, a slow down followed the burst of the information technology bubble. Between August 2000 and September 2002, in the information technology and telecommunication service sectors, stocks lost respectively 90% and 56% of their values. Stocks were worth 30.6% less in the consumer discretionary sector. Sectors that outpoerformed were the gold (+85.6%), consumer staples (+65.2%), utilities (+39.9%), and energy (+33.6%).

The table below shows how the various sectors of the stock market fluctuate over the business cycle. Business cycle is measure as short-run fluctuations in GDP.

Cyclical Behavior of the Stock Market and GDP, Canada, 1998:1-2015:M11

Sector	% Standard Deviation	Correlation with GDP
S&P 60	9.84	.57
Consumer Discretionary	8.85	.44
Consumer Staples	5.95	-.08
Energy	12.63	.41
Financial	10.25	.25
Gold	12.91	-.05
Industrial	11.06	.48
Information Technology	21.84	.39
Material	11.45	.36
Diversified Metal Mining	21.65	.29
Telecommunication Service	12.08	.5
Utilities	8.07	.21
GDP	.87	1

• The stock market is procyclical, i.e. positively correlated with the business cycle, except the sectors of consumer staples and gold.
• The stock market is more volatile than the real economy.
• The most volatile sectors of the stock market are: the information technology and the diversified metal and mining.

In most sectors, the stock market leads the business cycle by three months.

The dataset used is available here

.

Direct Investment Incomes and the Business Cycle in Canada

Foreign direct investment (FDI) is the building of new production facilities or the acquisition of existing businesses abroad. Most FDIs are undertaken by multinational companies (MNCs). Economic theory and several empirical studies suggest a positive relationship between FDI flows and the economic activity. For instance, both economic growth and gross domestic product (GDP) used as a proxy for the market size are reported to promote FDI. MNCs settle in countries having higher growth prospects or larger markets. There are other macroeconomic determinants of FDI such as the unit labor cost, the exchange rate, and the degree of openness to trade that I abstract from here to focus only on the GDP. My interest, particularly, is to show how the returns on FDI behave and how they relate to the cyclical fluctuations in GDP in Canada over the time period 1981:Q1-2013:Q2 (130 quarters).  

The Returns on FDI in Canada

The returns on Canadian direct investment abroad are known as direct investment income receipts and those on FDI in Canada are our direct investment income payments to the rest of the world. The direct investment incomes are made up of: interests, dividends, and reinvested earnings. The last two elements are alternative uses of firms’ profits. In Figure 1, below, I have plotted the total income receipts and payments from direct investments.



Figure 1: Direct Investment Real Incomes Receipts and Payments,

Millions of 2007 $, Canada, 1981:Q1-2013:Q2 (130 Quarters),

Source: Statistics Canada

It emerges from Figure 1 that:

• The direct investment income receipts and payments are very volatile,
• The direct investment income receipts and payments are highly and positively correlated.

Regarding the second observation, the correlation coefficient between the direct investments total income receipts and payments is .9. This latter result brings to mind empirical evidence Robert E Lipsey came up with in his 2000 paper titled InterpretingDeveloped Countries’ Foreign Direct Investment:

• FDI outflows and FDI inflows are positively correlated.

Myself I found similar evidence as Robert Lipsey in my 2003 MSc dissertation, The Determinants and Impacts of Foreign Direct Investment. From my investigations, the correlation coefficient between FDI outflows and inflows in Canada was .84 between 1978:Q1 and 2001:Q4. Actually, If FDI outflows and inflows are positively correlated, there are reasons to expect the returns on both investments to be also positively correlated. 

It also appears in Figure 1 that, most of the time, the direct investment incomes Canada pays to the rest of world exceed what it receives. Since the second quarter of 2012, we have been observing a shift in this tendency.

Dividends followed by the reinvested earnings are the most important types of direct investment incomes. Dividends represent, on average, about 66 % of both income receipts and payments. As for the reinvested earnings, they represent, on average, about 29 % direct investment income receipts and about 17 % of income payments.

The Cyclical Behavior of FDI incomes

The fluctuations observed in the direct investment incomes’ quarterly series plotted in Figure 1 suggest me the idea to extract their cyclical components to see how they behave over time compared to the cyclical components of real GDP. I detrended the data using the HP-filter.  
In Figures 2 and 3 below, I have plotted in blue line the cyclical components of the various types of direct investment incomes. In each panel of these figures, the cyclical components of real GDP is superimposed in red line for comparison. 
In the table below I present the coefficient of non-determination and the cross-correlation coefficients of the cyclical components of the series. The coefficient of non-determination is the share of the fluctuations in the series that are attributed to business cycle. By business cycle, I mean fluctuations that last eight years or less.

Figure 2: Cyclical Behavior of Direct Investment Incomes, Receipts,

Millions of 2007 $, Canada, 1981:Q1-2013:Q2

Figure 3: Cyclical Behavior of Direct Investment Incomes, Payments,

Millions of 2007 $, Canada, 1981:Q1-2013:Q2

Table  Standard Deviation and Cross-Correlation with Real GDP of Various Types of Direct Investment Incomes, Canada, 1981:Q1-2013:Q2

It turns out that:

• All the various types of direct investment incomes except the interests payments to the rest of the world are procyclical, i.e., their cyclical components are positively correlated with the cyclical component of real GDP,
• The interests paid to the rest of the world by the foreign firms operating in Canada are acyclic,i.e., their cyclical component is uncorrelated with the cyclical component of real GDP,
• All the various type of direct investment incomes are more volatile than the real GDP.
• Interests are the less volatile direct investment incomes.
• Fluctuations in direct investment incomes may lead fluctuations in real GDP.

The acyclicity of the interests paid to the rest of the world by the foreign firms operating in Canada makes sense because interest rates on money borrowed to finance long-run ventures are fixed. This also explains the less volatility observed in these series.  Except the interest received from or paid to the rest of the world, the highest cross-correlation coefficients are those between the GDP and the first lag of the variables. This suggests that the activities of MNCs may lead fluctuations in real GDP in Canada.