“You can clear your driveway faster with a snowblower than with a shovel.” This analogy, recently made by the Bank of Canada’s Senior Deputy Governor Carolyn A. Wilkins, illustrates how business investment increases workers’ productivity and the standard of living of Canadians. By investing in a snowblower, you could earn more money with the time saved clearing snow—perhaps by snowblowing more driveways—or you could have more time to spend on a leisurely activity of your choice.
But in the period after the financial crisis of 2007–09, business investment was lower than expected both in Canada and abroad. Several structural factors can possibly explain this lower investment (Barnett and Mendes 2017).
One is that firms may forego investment opportunities when they are unsure about future demand for their products (Leboeuf and Fay 2016). A drop in oil prices, for instance, creates uncertainty about future cash flows and also reduces the profitability of current and future investments in the oil sector.
Another factor could be an inability to finance firm investment. A firm may want to invest but may not have access to funding. Imagine wanting to buy a car to get to work faster but not being able to borrow money to buy it. You could only buy the car after you’ve earned the money to do so. And after buying the car, you may hold more cash to deal with emergency repairs. In the same way, a firm that has no access to financing can only invest after it has earned the funds, and it is more likely to stockpile cash.
In this note, we do not try to pin lower investment on one specific reason. Rather, we quantify how much investment behaviour has changed since the 2007–09 crisis. Then, we discuss a few of the potential reasons for this change that could be explored in future research.
We find that from 2008 to 2017, firms invested about $100 billion less than our model of their pre-crisis behaviour suggests they would have. This cumulative investment gap is economically significant—it represents about $2,700 less spent in investment on a per capita basis. Since Canada’s gross domestic product (GDP) and productivity increase when firms invest more, understanding the nature of this investment gap is important for Canadians.
Measuring the gap
Figuring out the size of the Canadian corporate investment gap in the post-crisis period requires a benchmark of what investment would have been. We use the pre-crisis behaviour of firms for this benchmark. The 2007–09 financial crisis may seem an arbitrary starting point, but it is also a natural choice as it is an event that may have caused a change in investment behaviour. But, as we show below, we find that firms changed their investment behaviour before the crisis.
We first measure the relationship between investment and measures of access to funding and future demand at the firm level (see the Appendix for more information about our methodology). These measures include firm-level financial variables (e.g., cash flows, market-to-book ratio, sales growth), which can account for industry-level shocks such as a collapse in oil prices. Then, for every firm, we predict what investment would have been in the post-crisis period if this pre-crisis relationship had continued. We call this the model-predicted investment. Then, we sum the firm-level model-predicted investments to get an annual measure of aggregate model-predicted investment. We also produce aggregate actual investment from firm-level actual investments in the same manner.
We call the difference between the aggregate actual investment and aggregate model-predicted investment the Canadian corporate investment gap. This gap represents how much more Canadian publicly listed firms would have invested if they had continued their pre-crisis investment behaviour. It does not necessarily suggest that firms are investing less than they should, as the model may not be capturing all the reasons behind changes in investment since the crisis. However, this corporate investment gap does capture any change in Canada’s overall industry composition. Since the gap is created from firm-level data, both actual and predicted investment at an industry level would adjust to a shift in an industry’s importance in the Canadian economy.
Since the 2007–09 crisis, this Canadian corporate investment gap accumulated to about $100 billion (Chart 1). In fact, a gap opened up before the crisis: corporate investment has been less than predicted in 14 of the 16 years from 2002 to 2017, which seems unlikely to be due to chance. This timing and profile suggest a permanent change in investment behaviour, and this change may have pre-dated the crisis.
Chart 1: Corporate investment gap opened up in the early 2000s
Changes in firms’ investment behaviour
Investment is sensitive to firm earnings, and this sensitivity has changed since the 2007–09 financial crisis. In unreported results, we find that large Canadian firms spent almost 50 cents of every dollar earned on investment in the pre-crisis period. This allocation to investment drops significantly to about 25 cents of every dollar earned in the post-crisis period. This suggests that the spending behaviour and funding decisions of firms have also changed. Returning to our car analogy, if you didn’t buy a car, it means you either spent your earnings on something else, borrowed less money from the bank, or held more in cash (to deal with potential future repairs). Firms, likewise, may have paid out more in dividends or held more cash. Or, they may have borrowed less or raised less equity financing.
Chart 2: Most of the gap is related to a reduction in equity issuance
This $100 billion Canadian corporate investment gap is mirrored by increases in cash holdings, dividend payments and issuances of debt, and decreases in issuances of equity relative to pre-crisis behaviour (Chart 2). Lower equity issuance and larger dividends were the largest contributors to the corporate investment gap, as firms issued $60 billion less in equity and increased dividend payments by $35 billion more than predicted by the model of their pre-crisis behaviour.
This illustrates how both lower future demand and less access to funding could be important reasons why firms are investing less in the post 2007–09 period. On the one hand, firms with reduced demand for their products would be less likely to raise equity; and with no profitable investment opportunities, they are likely to hold more cash or pay more dividends. On the other hand, firms that are financially constrained may not be able to raise equity and would likely hold more cash for precautionary reasons.
Potential causes and implications of the investment gap
Our measurements show that the investment gap opened in the early 2000s and continued through the financial crisis and the oil price shock. Gutiérrez and Philippon (2016) find that investment has been lower in the United States since the turn of the century. A full exploration of the reasons for the investment gap is beyond the scope of this note. Here, we discuss just a few of the reasons why firms have changed their investing and financing behaviour.
For example, the increasing importance of the knowledge economy may help explain some of the investment gap. Traditional measures of investment do not account for intangible investments such as research and development. Accounting for non-traditional investment may help narrow the gap. In unreported results, however, we find little change in the size of the gap using methods that try to measure this non-traditional investment (Peters and Taylor 2017). Nonetheless, this non-traditional investment measure may not perfectly capture firm investment in the knowledge economy, which may possibly account for some of the gap.
In theory, a firm will invest in a project if the expected rate of return of that project is greater than its hurdle rate. This hurdle rate should reflect the project’s cost of capital and be sensitive to changes in market interest rates. All else equal, a higher hurdle rate discourages investment. Evidence in the United States (Sharpe and Suarez 2014) and Australia (Lane and Rosewall 2015) suggests that firms rarely adjust their hurdle rates—and when they do, they do not fully reflect changes in market interest rates. This lack of adjustment could partially explain why investment has been lower than predicted as market interest rates have fallen.
If this is the case, it also raises the question of how monetary policy impacts firm investment. This relates to the prevalent discussion in the literature on how monetary policy transmission may be weaker when interest rates are low (e.g., Borio and Hofmann 2017). Monetary policy works by influencing market interest rates. If firms do not fully adjust their hurdle rates with changes in market interest rates, it would suggest a weaker transmission of interest rate changes to firm investment. There is some evidence, nonetheless, that monetary policy influences how individual firms invest in the United States and United Kingdom, especially financially constrained firms (Cloyne et al. 2018). To understand this better, in future work Bank staff will examine the effect of monetary policy surprises on corporate investment in Canada.
Appendix: Investment—cash flow regressions
We use the regression framework of Chang et al. (2014) to examine the investment behaviour of firms. This framework explains investments in terms of:
- market-to-book value of assets (a proxy for marginal Tobin’s-q or a firm’s growth opportunities)
- cash flows
- size of firm
- additional variables to control for firm- and industry-level heterogeneity
To determine whether firms changed their behaviour starting in 2008, we compare how firms made use of their cash relative to what our model predicts would have happened using parameter estimates from the 1987–2007 period.
We first categorize each firm into one of three major sectors:
- commodity goods
- non-commodity goods
For each sector, we relate investment to other decisions of firms by separately regressing each of the different uses of cash (i.e., investments, change in cash holdings, dividends, net reductions in debt and equity) on cash flow from operations and other control variables between 1987 and 2007 (Table A-1). Since cash must be spent on one of these uses, it also means that the cash-flow regression coefficient across the different dependent variables must sum to one. Similarly, the gap between model-predicted and actual values must sum to zero across all of these different variables.
Table A-1: Regression specification
|Dependent variables (uses of cash)|
|Change in cash holdings|
|Net debt issuance|
|Net equity issuance|
|Regressors in all equations|
|Cash flow from operations (CFO)|
|Market-to-book ratio (MB)|
|Sales growth (SG)|
|Log of total assets (TA)|
|Interaction terms: CFO⨯TA; MB⨯TA; SG⨯TA|
|Market volatility (VIX)|
|Industry fixed effects|
Note: All dependent variables, cash flow from operations and tangible assets are normalized by lagged total assets. All regressors except cash flow from operations and the VIX are lagged one period.
- Barnett, R. and R. Mendes. 2017. “A Structural Interpretation of the Recent Weakness in Business Investment.” Bank of Canada Staff Analytical Note No. 2017-7.
- Borio, C. and B. Hofmann. 2017. “Is Monetary Policy Less Effective when Interest Rates Are Persistently Low?” BIS Working Paper No. 628.
- Chang, X., S. Dasgupta, G. Wong and J. Yao. 2014. “Cash-Flow Sensitivities and the Allocation of Internal Cash Flow.” The Review of Financial Studies 27 (12): 3628–3657.
- Cloyne, J., C. Ferreira, M. Froemel and P. Surico. 2018. “Monetary Policy, Corporate Finance and Investment.” National Bureau of Economic Research Working Paper No. W25366.
- Gutiérrez, G. and T. Philippon. 2016. “Investment-less Growth: An Empirical Investigation.” National Bureau of Economic Research Working Paper No. W2897.
- Lane, K. and T. Rosewall. 2015. “Firms’ Investment Decisions and Interest Rates.” RBA Bulletin, Reserve Bank of Australia, June: 1–8.
- Leboeuf, M. and B. Fay. 2016. “What Is Behind the Weakness in Global Investment?” Bank of Canada Staff Discussion Paper No. 2016-5.
- Peters, R. H. and L. A. Taylor. 2017. “Intangible Capital and the Investment-q Relation.” Journal of Financial Economics 123 (2): 251–272.
- Sharpe, S. and G. Suarez. 2014. “The Insensitivity of Investment to Interest Rates: Evidence from a Survey of CFOs.” Federal Reserve Board Finance and Economics Discussion Series.
Bank of Canada staff analytical notes are short articles that focus on topical issues relevant to the current economic and financial context, produced independently from the Bank’s Governing Council. This work may support or challenge prevailing policy orthodoxy. Therefore, the views expressed in this note are solely those of the authors and may differ from official Bank of Canada views. No responsibility for them should be attributed to the Bank.