- Growth Companies
- Characteristics of Growth Companies
- Valuation Issues
- The Dark Side of Valuation
- The Light Side of Valuation
- Conclusion
The Light Side of Valuation
While growth companies raise thorny estimation problems, we can navigate our way through them to arrive at values for these firms that are less likely to be contaminated by internal inconsistencies. This section describes the steps to follow in discounted cash flow and relative valuations of growth companies.
Discounted Cash Flow Valuation
The objective in discounted cash flow valuation is to arrive at reasonable estimates of the cash flows and discount rates. Therefore, the following sections describe some of the considerations that should enter into the process when we value growth companies.
Narrative and Choice of Model
Chapter 2 described the choices we face with discounted cash flow models. We can either value the entire business (by discounting cash flows to the firm at the cost of capital) or value the equity directly (by discounting cash flows to equity at the cost of equity). While both approaches should yield the same value for the equity, estimating cash flows to equity, if we expect the debt ratio to change over time, is much more difficult than estimating the cost of capital. The former requires us to forecast new debt issues, debt repayments, and interest payments each period, as the dollar debt changes, whereas the latter is based on changing debt ratios.
Because many growth companies have little or no debt in their capital structure, analysts often fall back on equity valuation models, using the absence of debt as justification. However, this assumes that growth companies will continue with their policy of not using debt in perpetuity, even as growth decreases and the companies become more mature. If we make the more reasonable assumption that growth companies will become mature companies over time and adopt the financing practices of the latter, firm valuation models provide analysts with more flexibility to reflect these changes.
Needless to say, the discounted cash flow models used need to allow for high growth and even changing operating margins over time. As a general rule, rigid models that lock in the company’s current characteristics do not perform as well as more flexible models, where analysts can change the inputs over time in valuing growth companies.
Finally, it is worth repeating and re-emphasizing the lesson from Chapter 9; that is, for young companies with little or no history, having a plausible and consistent narrative that ties the numbers together in your valuation is critical. As we go through the process of estimating intrinsic value for young companies, it is worth noting that what ties together all the individual inputs that we will talk about for growth companies—revenue growth, operating margin, reinvestment, and risk—is the story that you are telling about the company.
Valuing the Operating Assets
If we accept the premise that firm valuation models work better than equity valuation models when valuing growth companies, the first step in the process is valuing the firm’s operating assets, incorporating both existing assets and growth assets.
Revenue Growth Rates
The valuation process starts with estimating future revenues. In making these estimates, many of the considerations we raised in Chapter 9 for young companies come into play. The biggest issue, and one we have emphasized repeatedly in this chapter, is the scaling factor. Revenue growth rates will decrease as companies get larger, and every growth company will get larger over time if our forecasts of growth come to fruition. In a test of how growth changes as firms get larger, Metrick (2006) examined the revenue growth rate for high-growth firms, relative to growth rate in revenues for the sector in which they operate, in the immediate aftermath of their initial public offerings.3 The results are shown in Figure 10.5.
Figure 10.5 Revenue Growth in the Years After the Initial Public Offering
Source: Andrew Metrick, The New York Times
When they go public, firms have growth rates that are much higher than the industry average. Note how quickly the revenue growth at these high-growth firms moves toward the industry average. They go from a 15% higher revenue growth (than the industry average) one year after the IPO to 7% higher in year 2 to 1% higher in year 4 to the industry average in year 5. We are not saying that this will happen at every high-growth firm. However, the aggregate evidence suggests that growth firms that can maintain high growth rates for extended periods are the exception rather than the rule.
The question of how quickly revenue growth rates will decline at a given company can generally be addressed by looking at the company’s specifics. These include the size of the overall market for the company’s products and services, the strength of the competition, and the quality of its products and management. Companies in larger markets with less aggressive competition (or protection from competition) and better management can maintain high-revenue growth rates for longer periods.4
We can use a few tools to assess whether the assumptions we are making about revenue growth rates in the future for an individual company are reasonable:
Absolute revenue changes: One simple test is to compute the absolute change in revenues each period, rather than to trust the percentage growth rate. Even experienced analysts often underestimate the compounding effect of growth and how much revenues can balloon over time with high growth rates. Computing the absolute change in revenues, given a growth rate in revenues, can be a sobering antidote to irrational exuberance when it comes to growth.
Past history: Looking at past revenue growth rates for the firm in question should give us a sense of how growth rates have changed as the company size changed in the past. For those who are mathematically inclined, clues in the relationship can be used to forecast future growth.
Sector data: The final tool is to look at revenue growth rates of more mature firms in the business to get a sense of what a reasonable growth rate will be as the firm becomes larger.
In summary, expected revenue growth rates tend to drop over time for all growth companies but the pace of the drop-off will vary across companies.
Current Margins Versus Target Margins
To get from revenues to operating income, we need operating margins over time. The easiest and most convenient scenario is one in which the current margins of the firm being valued are sustainable and can be used as the expected margins over time. In fact, if this is the case, we can dispense with forecasting revenue growth and instead focus on operating income growth, because the two are equivalent. In most growth firms, though, it is more likely that the current margin will to change over time.
Let’s start with the most likely case, which is that the current margin is either negative or too low relative to the sustainable long-term margin. This can happen for three reasons. One is that the firm has up-front fixed costs that have to be incurred in the initial phases of growth, with the payoff in terms of revenue and growth in later periods. This is often the case with infrastructure companies such as energy, telecommunications, and cable TV. The second reason is the mingling of expenses incurred to generate growth with operating expenses. We noted earlier that selling expenses at growth firms are often directed toward future growth rather than current sales but are included with other operating expenses. As the firm matures, this problem will get smaller, leading to higher margins and profits. The third reason is that there might be a lag between expenses being incurred and revenues being generated. If the expenses incurred this year are directed toward much higher revenues in three years, earnings and margins will be low today.
The other possibility, where the current margin is too high and will decrease over time, is less likely, but it can occur, especially with growth companies that have a niche product in a small market. In fact, the market might be too small to attract the attention of larger, better-capitalized competitors, thus allowing the firms to operate under the radar for the moment, charging high prices to a captive market. As the firm grows, this will change, and margins will decrease. In other cases, the high margins might come from owning a patent or other legal protection against competitors, and as this protection lapses, margins will decrease.
In both of the scenarios—low margins converging to a higher value or high margins dropping to more sustainable levels—we have to make judgment calls on what the target margin should be and how the current margin will change over time toward this target. The answer to the first question usually can be found by looking at both the average operating margin for the industry in which the firm operates and the margins commanded by larger, more stable firms in that industry. The answer to the second question depends on the reason for the divergence between the current and target margin. With infrastructure companies, for instance, it reflects how long it will take the investment to be operational and capacity to be fully utilized.
Reinvest to Sustain Growth
A constant theme in the earlier chapters was the insistence that growth is not free and that firms have to reinvest to grow. As we noted earlier in this chapter, basing reinvestment assumptions on a growth company’s history of reinvestment is dangerous. In other words, taking the net capital expenditures and working capital changes from the most recent year and assuming that these items will grow at the same rate as revenues can result in reinvestment numbers that are both unrealistic and inconsistent with our assumptions about growth.
To estimate reinvestment for a growth firm, we follow one of three paths, depending largely on the characteristics of the firm in question:
For growth firms earlier in the life cycle, we adopt the same road map we used for young growth companies, where we estimated reinvestment based on the change in revenues and the sales-to-capital ratio:
Reinvestmentt = Change in Revenuest/(Sales/Capital)
The sales-to-capital ratio can be estimated using the company’s data (which is more stable than the net capital expenditure or working capital numbers) and the sector averages. Thus, assuming a sales-to-capital ratio of 2.5 in conjunction with a revenue increase of $250 million results in reinvestment of $100 million. We can build lags between the reinvestment and revenue change into the computation by using revenues in a future period to estimate reinvestment in the current one.
With a growth firm that has a more established track record of earnings and reinvestment, we can use the relationship between fundamentals and growth rates described in Chapter 2:
Expected Growth Rate in Operating Income = Return on Capital × Reinvestment Rate + Efficiency Growth (as a Result of Changing Return on Capital)
In the unusual case where margins and returns and capital have settled into sustainable levels, the second term drops out of the equation.
Growth firms that have already invested in capacity for future years are in the unusual position of being able to grow with little or no reinvestment for the near term. For these firms, we can forecast capacity usage to determine how long the investment holiday will last and when the firm will have to reinvest. During the investment holiday, reinvestment can be minimal or even zero, accompanied by healthy growth in revenues and operating income.
With all three classes of firms, though, the leeway that we have in estimating reinvestment needs during the high-growth phase should disappear as soon as the firm has reached its mature phase. The reinvestment in the mature phase should hew strictly to fundamentals:
In fact, even in cases where reinvestment is estimated independently of the operating income during the growth period, and without recourse to the return on capital, we should keep track of the imputed return on capital (based on our forecasts of operating income and capital invested). Doing so ensures that it stays within reasonable bounds. The process for doing so is described in Chapter 9.
Risk Profile Consistent with Growth and Operating Numbers
The components of the cost of capital—the beta(s) and the cost of equity, the cost of debt, and the debt ratio—are the same for a growth company as they are for a mature company. However, what sets growth companies apart is that their risk profiles shift over time. The key to maintaining balance in growth company valuations is to adjust the discount rates over time to keep them consistent with the growth and margin assumptions that we make in each period. Here are two general rules:
Growth firms should have high costs for equity and debt when revenue growth is highest, but the costs of debt and equity should decline as revenue growth moderates and margins improve.
As earnings improve and growth drops, another phenomenon comes into play. The firm generates more cash flows than it needs, which it can use to both pay dividends and service debt financing. Firms are not required to use this debt capacity, and some of them do not, but the tax advantages of debt lead some firms to borrow, causing debt ratios to increase over time.
In summary, the cost of capital for a growth company should almost never be a number that remains unchanged over the entire time horizon. Instead, it should be a year-specific number that keeps pace with the rest of the changes we forecast at the firm.
In terms of estimating risk parameters (betas), we would steer as far as we can from using the limited price data that is available on growth companies. The standard errors on the estimates are likely to be huge. Instead, we would use estimates of betas obtained by looking at other publicly traded firms that share the same risk, growth, and cash flow characteristics as the firm being valued. If the case for using these bottom-up betas (industry average as opposed to a regression beta) is strong with any firm, it is even stronger with growth firms.
For growth firms that have operating losses carried forward from prior years or that are expected to keep losing money in the future, a final factor must be considered when computing discount rates. The tax advantage of debt, manifested as an after-tax cost of borrowing, depends on having positive earnings to offset interest expenses. With operating losses (and carryforwards), there might be no or limited tax benefits from interest expenses, and the after-tax cost of debt should reflect this fact.
Stable Growth Assumptions: What Will the Firm Look Like, and When?
The assumptions we make about terminal value loom large with a growth company, because it comprises a much larger portion of the firm’s current value than is the case with a mature firm. When will a growth firm become a mature, stable-growth firm? While we have a little more information than we did with young companies, making this assessment is difficult. It’s akin to looking at a teenager and wondering what he or she will look like or be doing in middle age.
Although no one answer or approach will work with every growth company, we will draw on the discussions in Chapter 2 and this chapter to develop the following general propositions:
Do not wait too long to put a firm into stable growth. As we noted in the section on the dark side of valuing growth companies, analysts often allow for very long growth periods for growth firms and justify this assumption by pointing to past growth. As shown in Figure 10.5, both scale and competition conspire to lower growth rates quickly at even the most promising growth companies. Growth periods that exceed ten years, especially when accompanied by high growth rates over these periods, are difficult to defend, because only a few companies have been able to accomplish this over time. Valuing your company to be the exception, well before it has established itself, is not a good practice.
When you put your firm into stable growth, give it the characteristics of a stable growth firm. In keeping with the emphasis on preserving internal consistency, we should change the company’s characteristics to reflect stable growth. With discount rates, as noted in the preceding section, this takes the form of using lower costs of debt and equity and a higher debt ratio. With reinvestment, the key assumption is the return on capital that we assume for the stable growth phase. While some analysts believe that the return on capital should be set equal to the cost of capital in stable growth, we would preserve some company-specific flexibility. We suggest that the difference between return on capital and cost of capital should narrow during stable growth to a sustainable level (less than 4 or 5%).
The nature of cash flows at growth companies—low or negative in the early years and higher later—will ensure that the terminal value is a high proportion of value, accounting for 80%, 90%, or even more than 100% of value. As we noted in Chapter 9, the more than 100% scenario will unfold when a growth company has high growth and high reinvestment needs, leading to negative cash flows for an extended part of the forecast period. Some analysts use this fact as ammunition against using discounted cash flow valuations, suggesting that assumptions about the high-growth phase will be drowned out by terminal value assumptions. This is not true. The base year value for the terminal value calculation (earnings and cash flows in year 5 or 10) is a function of the assumptions during the high-growth phase. Changing these assumptions will have dramatic effects (as it should) on value.
From Operating Asset Value to Equity Value per Share
Navigating from operating asset value to equity value per share for growth companies can be fraught with dangers, many of which we outlined in earlier sections. In this section, we outline precautions against some of these dangers (at least partial).
Cash and Nonoperating Assets
Earlier in this chapter, we noted how quickly growth firms can burn through cash balances and how using the cash balance from the most recent financial statements can lead to misleading values. At least in theory, it would be useful to know what the cash balance is today when valuing a firm. While investors in public equity markets have no way to access this information, an acquirer (or at least a friendly acquirer) should be able to get this information from the target firm and use it to estimate an updated value. Even public investors can make judgments about current cash balances by using two pieces of public information—the firm’s cash flows and any new financing during the period since the last financial statement. For instance, assume that the last cash balance (from three months ago) is $100 million for a firm that reported EBITDA of negative $80 million in the most recent 12-month period. If the firm has not raised any new financing in the last three months (through either equity issues or new debt), the firm’s current cash balance is likely to be closer to $80 million than $100 million. (We are reducing the cash balance by the estimated EBITDA of −$20 million—one quarter of −$80 million.)
Debt and Other Nonequity Claims
If convertible debt is the preferred mode of borrowing for a growth firm, we should treat it as what it is—a hybrid security that is part debt and part equity. Because the conversion option is equity and the rest is debt, the easiest way to decompose convertible debt into debt and equity is to value the convertible debt as if it were straight debt and to treat the resulting number as debt. For instance, assume that a growth firm has five-year convertible debt outstanding, with a face value of $50 million and a coupon rate of 4%. Also assume that the firm’s current pre-tax cost of debt, assuming it uses conventional debt, is 10%. The value of the convertible bond, treated like a conventional debt, would be as follows:17
Subtracting this value from the market value of convertible bond yields the value of the conversion option. Thus, if the convertible debt is trading at $52 million, the conversion option (equity) will be valued at $13.37 million.18 When valuing this company, we would treat this part as equity and the rest as debt in computing the current cost of capital. We would subtract only the debt portion from the firm’s value to arrive at the overall value of equity.
Another aspect of debt that is potentially problematic is that debt ratios change over time, and with those changes come changes in the debt outstanding at the firm. Because the value of equity is the firm’s value, net of debt, analysts often get caught up in the question of whether they should subtract the debt outstanding today (which might be negligible) or the expected debt outstanding in the future (which might be very large). The answer, when valuing the firm, is that we should subtract only the current debt outstanding, even though that value of that debt might be miniscule relative to future debt issues.
Post-Valuation Corrections
After we have derived the value of equity in a growth firm, the final step is to allocate the value of equity across the shares outstanding in the firm. In making this final judgment, three considerations must be kept in mind.
Survival and Illiquidity
The first two considerations are issues that were raised with young growth companies. The probability of survival must be factored into the value, and illiquidity can cause discounts to this value. Neither of these factors is as significant with growth firms. The probability of survival for growth firms is much higher than for nascent businesses. The equity is usually more liquid—especially if the growth firm is publicly traded. However, even publicly traded growth firms sometimes have to shut down operations, especially if they run out of cash. Shares in these firms might trade far less frequently (and with much higher transaction costs) than shares in more mature companies.
If survival is truly a concern, we suggest using the approaches we developed in the last chapter. In summary, you estimate the probability that the growth firm will fail and the consequences (in terms of what equity investors will receive) if it does. The expected value of equity will then reflect the weighted average of the going concern and failure values, weighted by the probabilities of success and failure. Similarly, if illiquidity is weighing down value, using one of the approaches described in the preceding section—adjusting the cost of capital or applying a post-valuation illiquidity discount—will help.
Voting Right Differentials
The final factor to consider in arriving at value of equity per share is differences in voting rights across shares. While we would expect voting shares to trade at a premium on nonvoting shares, the magnitude of the difference should be a function of the value of voting rights and consequently varies across firms.
How much are voting rights worth? The earliest studies of voting share premiums were done with U.S. companies that had different voting share classes. Lease, McConnell, and Mikkelson (1983) found that voting shares in that market trade, on average, at a relatively small premium of 5% to 10% over nonvoting shares.19 They also found extended periods where the voting share premium disappeared or voting shares traded at a discount to nonvoting shares. This surprising finding can be explained partially by the relative illiquidity of voting shares (because, often, only a small percentage is available for public trading). Reilly (2005) updated this study to look at 28 companies with voting and nonvoting shares in 1994 and 1999 and concluded that the median voting share premium increased from 2% in 1994 to 2.8% in 1999.20
Studies in recent years have expanded the analysis of voting share premiums to other markets, where differential voting rights are more common. Premiums of a magnitude similar to those found in the U.S. (5% to 10%) were found in the U.K. and Canada. Much larger premiums are reported in Latin America (50% to 100%), Israel (75%), and Italy (80%). In a comparative study of voting premiums across 661 companies in 18 countries, Nenova (2003) found that the median value of control block votes varies widely across countries, ranging from less than 1% in the U.S. to 25% or greater in France, Italy, Korea, and Australia. She concluded that the legal environment is the key factor in explaining differences across countries. She also concluded that the voting premium is smaller in countries with better legal protection for minority and nonvoting stockholders and is larger for countries without such protection.21
The most common way to allocate value across voting and nonvoting shares is to use the evidence from these studies to justify a premium. In the U.S., for instance, voting share premiums have generally been set between 5% and 10%. Although we are not entirely convinced about this rule of thumb, we will stick with it for this chapter and return to examine it in a later one.
Dealing with Uncertainty
It is almost a given that the value of a growth company, no matter how much we pay attention to the details and how much information we use, will be less precisely estimated than the value of a mature company. This uncertainty can lead to post-valuation angst in which analysts second-guess themselves and try to reconcile differences not only between their estimates and the market price but also across different valuations (done by different analysts).
Note, though, that much of this uncertainty comes not from the quality of the information or the precision of the valuation model used, but from the real world. The future is full of surprises, and for growth firms, where so much of the value lies in the future, this translates into big changes in value. This is small consolation to the analyst who gets the value of equity in a growth company wrong and is held accountable. Chapter 3 presented probabilistic approaches including decision trees, simulations, and scenario analysis that can be used to enrich valuations. These approaches offer some promise for growth companies, not because they provide more precise estimates of value or even because they generate risk measures, but because they allow analysts to be more comfortable with their own estimates of value. In the valuation of Snap, where our assumptions for a young company piled on top of each other, we estimated a value per share of $10.91. Focusing just on the assumptions about revenue growth rates, operating margins, sales to capital and the cost of capital, it is clear that changing any or all of these assumptions will change value significantly. Rather than try to make this uncertainty go away (a hopeless exercise), we used a simulation where we replaced our point estimates for each of these variables with distributions. The results are summarized in Figure 10.7.
Figure 10.7 Value Per Share for Snap IPO: Simulation Results
The Snap simulation yields some interesting output that might affect whether you invest and how you think about the company. The first is that while there is, not surprisingly, a possibility that the company’s shares could be worth nothing, the distribution has a distinctly positive skew, with values that exceed $30 per share. The second is that you can use the distribution to assess the likelihood that the stock is under valued or over valued. At its offering price of $20/share, for instance, the simulation would have suggested that there is an 80% chance that the stock is overvalued.
Another useful technique for grappling with uncertainty with growth companies is to focus on one of two key drivers of value for that company. We also look at not only the effects on value of varying assumptions about those drivers but also break-even points in terms of the current price. For instance, assuming that revenue growth is the key determinant of value for a firm, we can ask what the revenue growth rate would have to be to justify the current market price. We can then follow up by looking whether we are comfortable as investors with the market-implied revenue growth rate. Consider the valuation of Shake Shack, where our assumptions led us to a value per share of $35.71. In the intensely competitive restaurant business, where fads fade fast and margins are always under assault, we focused on the effects of changing revenue growth and operating margin on value per share and captured the results in Table 10.15.
Table 10.15 Shake Shack: Value Per Share What If? – November 2017
|
|
Pre-tax Target Operating Margin |
||||
|
|
4% |
8% |
12% |
16% |
20% |
|
5% |
−$0.27 |
$5.25 |
$10.78 |
$16.30 |
$21.83 |
|
10% |
$0.41 |
$7.96 |
$15.50 |
$23.04 |
$30.58 |
|
15% |
$1.26 |
$11.45 |
$21.63 |
$31.82 |
$42.01 |
Revenue Growth Rate (next 5 years) |
20% |
$2.30 |
$15.92 |
$29.54 |
$43.17 |
$56.79 |
25% |
$3.56 |
$21.61 |
$39.66 |
$57.71 |
$75.76 |
|
|
30% |
$5.11 |
$28.81 |
$52.52 |
$76.22 |
$99.92 |
|
35% |
$7.00 |
$37.86 |
$68.72 |
$99.58 |
$130.45 |
|
40% |
$9.29 |
$49.15 |
$89.02 |
$128.89 |
$168.76 |
At its prevailing market price of $33.77 (in October 2017), you can see the combinations of revenue growth/operating margin that deliver higher value (as shaded cells). As an investor in Shake Shack, these numbers become the expectations that you have to beat to generate excess returns on your investment. As a manager in Shake Shack, this table provides you with a measure of how your value will change as you change pricing and marketing strategies. For instance, Shake Shack might be able to grow its revenues more quickly if it cut its burger prices, but doing so will reduce margins. The overall effect on value will depend upon the resulting combination of revenue growth and margins.
Relative Valuation
There is no reason why relative valuation cannot be used to arrive at an independent estimate of the value of equity in a growth firm—as long as we keep two key factors in mind. The first is that using multiples and comparables cannot reduce the uncertainty inherent in valuing growth companies. The second is that relative valuation techniques have to be adapted to meet the limitations of growth companies—the paucity and unreliability of current operating numbers and the shifting risk/growth characteristics over time.
Comparable Firms
Optimally, we would like to assess how the market values a growth firm by comparing its pricing with that of otherwise similar growth firms. In a business like software, which has growth firms aplenty, this can be accomplished by staying within the traditional framework of defining comparable firms as those in the same industry. In businesses like retailing and automotive parts, a growth firm might be the exception in a sector where the bulk of the firms are either mature or in decline. In these cases, we might have to abandon the conventional practice and define growth firms in terms of fundamentals rather than business. The pricing of a retail firm with growth prospects should be compared to how the market is pricing growth firms in other sectors rather than more mature firms in its own industry. There is no reason why the PE ratio for a high-growth retail firm should not be comparable to the PE ratio for a high-growth software firm.
Choice of Multiples and Base Year
As we noted in the section on the dark side as it relates to relative valuation, analysts valuing growth companies tend to use either revenues in the current year or estimates of operating performance in future years (forward earnings or revenues) to compute multiples. Each carries some danger:
Revenue multiples are troubling simply because they gloss over the fact that the company being valued could be losing significant amounts of money. Consequently, we suggest bringing the expected future profit margins (which will be estimates) into the discussion of what comprises a reasonable multiple of revenues. Other things held constant, we would expect firms with higher expected profit margins (in the mature phase) to trade at higher multiples of current revenues than firms with lower expected profit margins.
Forward earnings multiples implicitly assume that the firm being valued will survive to the forward year and that the estimates of earnings for that year are reasonable. If forward multiples are used, controlling for survival becomes a critical component of the analysis. Firms that have a greater chance of surviving to the forward year should trade at a higher multiple of earnings than firms that have a greater chance of failure.
As a general rule, we suggest steering away from multiples of either current book value or current earnings with growth companies early in the growth cycle, simply because these numbers are likely to be small and volatile.
Adjusting for Differences in Growth and Risk
No matter how careful we are about constructing a set of comparable firms and picking the right multiple, there will be significant differences across the firms on their fundamentals. As we noted earlier in the chapter, the two ways in which analysts control for these differences—storytelling and assuming that multiples increase proportionately with growth—yield misleading results. In fact, both approaches break down when we have to control for more than one variable when making comparisons.
When dealing with large differences in growth and risk across companies, the approach that offers the most flexibility is a multiple regression. The chosen multiple is the dependent variable and growth, risk, and any other fundamentals we want to control for representing independent variables. With large enough samples of comparable firms, not only can we control for as many variables as we want, but the approach lets us allow for complex relationships between growth and each variable.