ERASMUS UNIVERSITY ROTTERDAM

Erasmus School of Economics

Capacity group Business Economics
Section Finance
Seminar Corporate Finance, March –
April 2025
Group: Sjoerd van Bekkum
Preliminary Case Report
Casilda Abou-Hassan Garín, Rushang
Shringi, Esteban Aguilar, Yuito Maeda
Introduction
This report gives a comprehensive bottom-up valuation study of ASML Holding N.V.
using the Discounted Cash Flow (DCF) method. The objective is to arrive at a reasonable and
sound estimate of ASML's enterprise and equity value based on internally consistent
assumptions and fact-driven projections. For this purpose, we first calculate and justify ASML's
equity beta, a critical input in estimating the company's cost of equity. We then calculate the cost
of debt using market- and credit-based inputs, and both merge them into an equation that
computes the firm's Weighted Average Cost of Capital (WACC). Following this, we build a
forward-looking growth rate using company strategy and industry fundamentals, which we then
use to project ASML's Free Cash Flows (FCF) from 2025 to 2030. Each section of the report
explains not only the calculations performed but also the underlying reasoning, assumptions, and
trade-offs. This step-by-step approach allows for transparency and consistency throughout the
valuation process, resulting in a set of cash flow projections and discounting assumptions that
guide our final estimate of ASML's intrinsic value.

Equity Cost of Capital

In order to calculate the WACC for ASML, first we need to determine the equity cost of
capital (rE). This will be estimated using the commonly applied CAPM model. A key ingredient
for estimating CAPM is first determining ASML’s market beta, the estimation method will be
outlined below.
We take heavy inspiration from a 2017 paper written by Yaron Levi and Ivo Welch where
they outline an approach in obtaining robust beta estimates, particularly in the context of
forecasting. The paper mentions that the Vasicek (VCK) beta outperforms the standard OLS beta
in terms of predictive power; however, due to limitations in obtaining an adequate dataset to
obtain VCK estimates, we have decided to calculate standard OLS beta estimates. Whilst not
ideal, the paper does mention that applying shrinkage to the OLS beta will still offer gains in
predictive power that mitigate the downsides of using OLS betas rather than VCK betas.
Additionally, the paper highlights that for large cap stocks, daily stock prices over one year was
the optimal estimation window and return measuring frequency; as such, we too, use daily stock
prices from the end of 2023 through to the end of 2024.
Despite the widespread use of the CAPM framework, it is important to acknowledge its
limitations. The model relies on strong assumptions, such as frictionless capital markets and the
existence of a true market portfolio—conditions rarely met in practice. As emphasized by Roll’s
critique, even if empirical results deviate from theory, it is unclear whether this reflects a failure
of CAPM itself or simply an incorrect specification of its inputs. Nevertheless, for practical
purposes, we proceed with CAPM to estimate ASML’s cost of equity. To proxy the market
portfolio, we use the STOXX Europe 600 index. Whilst it is common practice to rely on broad
equity indices as market portfolio proxies, the STOXX 600 in particular was chosen for its
relevance to ASML’s profile. Unlike narrower indices such as the AEX, which is limited to
Dutch large caps, the STOXX 600 offers diversified exposure across countries, sectors, and
company sizes within Europe. Given ASML’s significant international operations, European
investor base, and listing on Euronext Amsterdam, the STOXX 600 provides a more appropriate
benchmark for capturing the systematic risk relevant to its equity holders. All STOXX 600 data
used in our calculations are obtained from the Backtest website provided by CURVO.

In addition to estimating the equity beta via OLS, we apply additional as per Levi and
Welch's guidelines. The degree of shrinkage is depicted in equation 1.1 below.
β𝐿𝑊 = 0. 65 * β𝑂𝐿𝑆 + 0. 35 * 𝑇𝐴𝑅𝐺𝐸𝑇 (Equation 1.1)
The value assigned to TARGET was 0.9 as per the papers suggestion for the largest tertial
of market caps. The intuition behind adding weight to the target, as justified by the paper, is to
tackle some degree of mean reversion stock betas tend to follow over time. Our raw OLS beta
was estimated to be 1.997. After applying equation 1.1, the final equity beta used for our
calculations will take a value of 1.613.
The logical next step in determining rE is determining the risk-free rate (rf) and market
risk premia; for this, we take inspiration from a Damodaran (2024) paper which discusses several
approaches to estimate such parameters. Firstly, in contrast to beta estimates, it was suggested to
rather use annual data over a longer estimation window to reduce the distorting effects of noise
observed in daily, and even monthly price data. This, in turn, means we have to use a longer
estimation window. We gathered risk-free rate data from the OECD database, specifically the
long term interest rates of the Netherlands from 1986 to 2024; it is worth noting that OECD
defines the long term interest rates as the yields on 10-year government issued bonds. Similarly,
we used annual index price data of the STOXX 600 index covering the same time period from
the same database used in calculating the equity beta. We are utilising the historical premium
approach as discussed in the Damodaran (2024) paper, we take the simple average of market
returns and risk free rates to determine the market risk premia. Doing so yields an average
market return of 9.56% and an average risk free rate of 3.92%; thus, our risk premia estimate is
5.64% The final step is applying CAPM with our parameter estimates as outlined in equation 1.2.
𝑟𝐸 = 𝑟𝑓 + β𝐸 * (𝑟𝑚 − 𝑟𝑓) ​ ​ ​ ​ ​ ​ (Equation 1.2)
After plugging in our estimated parameters into equation 1.2, our final estimate for
ASML’s equity cost of capital is 13.02%, this will serve as our input when determining the
WAAC. Lastly, all calculations for equity betas and market risk premiums were done using R.

Debt Cost of Capital

To estimate ASML’s cost of debt, we followed the approach outlined by Faulkender,
Hankins, and Petersen (2019), who support using market-based yields on outstanding corporate
debt as a proxy for a firm’s marginal cost of borrowing. This method was chosen as it shows the
true opportunity cost of raising capital in markets and avoids relying on historical interest
expenses that may not represent the current borrowing conditions in the market.
 The approximate Yield to Maturity formula is applied; a practical and analytically
grounded approach as there is some absence of full bond pricing data.

(𝐹𝑉 − 𝑃𝑉)
𝐶+ )
𝑌𝑇𝑀 ≈ 𝑛
𝐹𝑉 + 𝑃𝑉 (Equation 2.1)
( 2
)

The equation above was used to approximate the value of the YTM. It was applied to each of
ASML’s fixed-rate bonds.
Where:
●​ C: Annual coupon payment
●​ FV: Face value of the bond
●​ PV: Present (fair) value of the bond
●​ n: Years to maturity

The bond-level data was extracted from ASML’s annual report. See Table 2.1 in the Appendix.
​ The equation below represents the weighted average YTM. This was used to derive
ASML’s blended cost of debt in accordance with Faulkender, Hankins, and Petersen (2019),
using the fair value of each bond as a weight:
𝑛
𝑃𝑉𝑖
𝑟𝑑 = ∑ 𝑌𝑇𝑀𝑖 × ( ) (Equation 2.2)
𝑖=1 ∑𝑃𝑉

Using this formula we obtain a pre-tax cost of debt equal to 2.08%. We do not account for tax yet
as we will do this directly in the WACC calculation.
This approach was chosen for several reasons: it reflects real-time capital market
conditions, aligns with present corporate finance literature approaches such as Faulkender,
Hankins, and Petersen (2019), and accurately incorporates premium/discounting pricing effects.

WACC
To estimate ASML’s Weighted Average Cost of Capital, the standard formula was applied:

𝐷 𝐸
𝑊𝐴𝐶𝐶 = 𝑟𝑑 × (1 – 𝑡) × ( 𝑉 ) + 𝑟𝑒 × ( 𝑉 ) (Equation 3.1)

Where 𝑟𝑑 is the pre-tax cost of debt, 𝑟𝑒 is the cost of equity, 𝑡 is the effective corporate tax rate,
and 𝐷 / 𝑉 and 𝐸 / 𝑉 represent the market value weights of debt and equity, respectively. The
WACC’s purpose is to capture the average rate at which ASML is expected to pay capital
providers (equity and debtholders) whilst factoring in the tax advantage of debt financing.
​ The market value of equity was calculated by multiplying ASML’s shares outstanding
and closing share price as of December 31, 2024. By using this approach we can find the
market’s valuation of ASML’s equity and remains consistent with the standard practice of
corporate valuation. The value of debt was obtained directly from ASML’s balance sheet. The
book-value of debt was used as it is a suitable proxy given ASML’s relatively stable bond
pricing.
​ The effective tax rate was estimated by dividing interest expense by EBIT. This ratio
offers a firm-specific estimate of the tax shield realised through interest deductibility. This is
specifically relevant in WACC calculations where the interest payments do influence post-tax
debt cost.
Through this calculation we finally obtain a WACC equal to 12.9%. This final WACC
reflects the weighted average cost of equity and after-tax debt, scaled by respective market
values. The leverage is around 1.33% indicating that the WACC is almost entirely driven by the
cost of equity. Nonetheless, Equation 3.1 remains valid. This rate will be used to discount
projected free cash flows in subsequent valuation analysis.

Growth Rate
Under a DCF methodology, the rate of growth is a significant factor in determining
forecasted Free Cash Flows (FCFs) as well as terminal value. We differentiate between two
growth rates in this analysis: the period-specific growth rate (2025–2030), used to estimate FCFs
on a yearly basis, and the terminal growth rate (post-2030), used to estimate the perpetuity-based
terminal value.
Although starting growth estimation from past financial performance is a common
practice, the method will not be suitable for ASML due to its extremely volatile previous revenue
growth. Although growth was over 30% between 2022 and 2023, between 2023 and 2024 the
growth fell drastically to 2.6%. The fluctuations represent industry-level drivers, including
post-pandemic recovery, chip shortages, and geopolitical tensions, rather than a persistent trend.
Previously mentioned source also cautiona against the abuse of historical averages when leading
indicators are presented; as Damodaran (2012) states, "in firms undergoing structural change or
facing changing market realities, historical growth has limited forecasting power". Therefore, we
apply a bottom-up approach by leveraging ASML's own forecasts and strategic position in the
global semiconductor industry.

Based on ASML's 2024 annual report, ASML anticipates revenue growing from €28.3
billion in 2024 to €44 billion-€60 billion in 2030, depending on macroeconomic conditions and
adoption of EUV and High-NA technology. We utilize a midpoint revenue estimate of €52
billion in 2030 as our base case. We calculate the 2024-2030 compound annual growth rate
(CAGR) based on this figure:
52 1/6
𝐶𝐴𝐺𝑅2025−2030 = ( 28.3 ) − 1 (Equation 4.1)

1/6
(1. 8375) − 1 = 1. 1083 = 10. 83% (Equation 4.2)

​
​ ​This 10.83% CAGR is a conservative growth rate, consistent with ASML's internal
expectations and supported by underlying long-term global semiconductor demand trends. We
expect this CAGR to hold for revenue from 2025-2030 and roll forward its implications into the
FCF estimate.

For the terminal value calculation, we use a long-term growth rate (LTG) of 2.0%, as the
Euro Area GDP growth forecasts for the long term are 1.5%–2.0% (ECB, 2024). A terminal rate
of more than 2.5% would mean that ASML grows faster than the world economy permanently,
which is not realistic for a mature industrial business. We hence conservatively take a 2.0% LTG,
high enough to pick up technological leadership, but careful enough not to overestimate.

FCF Calculation
Here, we estimate ASML's Free Cash Flows (FCFs) for 2025–2030. The objective is to
come up with a clean and reasonable estimate of the unlevered free cash flows of the company,
based on uniform, operationally driven assumptions. The simple formula used is:

𝐹𝐶𝐹 = 𝐸𝐵𝐼𝐴𝑇 + 𝐷𝑒𝑝𝑟𝑒𝑐𝑖𝑎𝑡𝑖𝑜𝑛 − 𝐶𝐴𝑃𝐸𝑋 − ∆𝑁𝑊𝐶 (Equation 5.1)

All of them were determined using the forecasting process as follows.

As a way of estimating cost drivers such as COGS, Operating Expenses, Depreciation,

and Net PPE, we used the ratio technique, that is, each of these we calculated as a fixed
proportion of sales in the base year 2024 and estimated such a percentage as a measure to
forecast revenue between 2025 and 2030.

We chose this method since our revenue was forecasted through a top-down approach
based on our estimated growth rate (10.83% CAGR); it is logical to forecast cost items as a
percentage of revenue, especially for a company with high operating leverage and consistent
margins like ASML. ASML is in a high-margin, capital-intensive business. Its production trend
shows that asset expenditures, COGS, and SG&A will increase proportionally with revenues in
the medium term — this is consistent with good cost structures. Historically, ASML had fairly
consistent EBITDA margins and gross margins over time, which shows that an extrapolation of
cost drivers as a % of revenue makes sense and sustainable projections.

Unlike fixed-ratio cost items, Net Working Capital (NWC) is notoriously volatile —
especially for ASML. Between 2016 and 2024, year-to-year ΔNWC fluctuated between +€2.7
billion and –€3.2 billion, with no discernible pattern or trend. This volatility is due to uncertain
timing of customer deposits, inventory cycles driven by projects and uncertain variations in
terms of payment by suppliers

Given this unpredictability, we did not use a historical average or a % of sales

calculation. Instead, we employed a component-based calculation that breaks down ΔNWC into
its fundamental building blocks:
𝑁𝑊𝐶 𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡 = (𝐴/𝑅𝑡 + 𝐼𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑦 − 𝐴/𝑃𝑡) − (𝐴/𝑅𝑡−1 + 𝐼𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑦 − 𝐴/𝑃𝑡−1

(Equation 5.2)

This approach relies on the way working capital actually operates in terms of movements
in receivables, inventory, and payables. Using forward-looking DSO (Days Sales Outstanding)
estimates, DIO (Days Inventory Outstanding), and DPO (Days Payables Outstanding), we
back-calculated all of the working capital accounts and let underlying operations drive ΔNWC
organically. While the outcome might still fluctuate on a year-by-year basis, this approach is
much more economically intuitive and firm structure aligned with the firm's revenue growth.
Bibliography
Damodaran, A. (2024). Equity Risk Premiums (ERP): Determinants, Estimation, and Implications
– 2024 Update. Retrieved from http://pages.stern.nyu.edu/~adamodar/

Faulkender, M., Hankins, K. W., & Petersen, M. A. (2019). Understanding the Rise in Corporate
Cash: Precautionary Savings or Foreign Taxes. Review of Financial Studies, 32(11), 4088–4126.
https://academic.oup.com/rfs/article/32/9/3299/5298323

Knudsen, J., Fahlenbrach, R., Low, A., & Stulz, R. M. (2017). The value of firms with a
dual-class share structure. Financial Analysts Journal, 73(3), 35–48. https://doi.org/10.2469/faj.v73.n3.5

Levi, Y., & Welch, I. (2017). Best practice for cost-of-capital estimates. Journal of Financial and
Quantitative Analysis, 52(2), 827–861.
https://www.cambridge.org/core/journals/journal-of-financial-and-quantitative-analysis/article/best-practi
ce-for-costofcapital-estimates/AD8B94381E86DCD7F03A73E1DF472BF4​
​
​ Organisation for Economic Co-operation and Development (OECD). (2024). Long-term interest
rates – 10-year government bond yields. Retrieved from
https://data.oecd.org/interest/long-term-interest-rates.htm

STOXX Ltd. (2024). STOXX Europe 600 Index historical returns. Retrieved from
https://www.stoxx.com/index-details?symbol=SXXP

Yahoo Finance. (2024). ASML Holding N.V. (ASML) Historical Data. Retrieved from
https://finance.yahoo.com/quote/ASML/history
Appendix
Table 1.1: SARD output and peer identification
Peer firm SARD score EBIT Margin Total Assets R&D/Sales

ASML HOLDING NV .319 48,590 .154

YIXIN GROUP LTD 21 .568 48,591 .123

GENMAB AS 44 .329 45,811 .453

LOTES CO LTD 48 .341 50,211 .0908

MERCK KGAA 73 .193 51,567 .108

ASMEDIA 79 .290 36,056 .185

TECHNOLOGY INC

ALIGNED GENETICS 80 .176 44,275 .145

INC

Table 2.1: Data extracted from the annual report to calculate YTM
Maturity Coupon (%) Face Value (€m) Fair Value (€m)

2025 3.5 1000 1010.3

2026 1.375 1000 967.7

2027 1.625 750 720.1

2029 0.625 750 748.3

2030 0.25 750 744.8

2032 2.25 500 478.2

Table 3.1: Final results for the FCF and PV values

Year 2024 2025 2026 2027 2028 2029 2030

FCF 6309.6 6897.3 7498.2 8087.0 8621.1 9031.1

Horizon 483583.4

Explicit 33458.9

PV 617350.4
​