A. Introduction

1. The pharmaceutical industry in Denmark has grown at a rapid pace in recent years. Since 2005, the real gross value added (GVA) of pharmaceuticals has increased sixfold (Figure 1). This extraordinary surge in real GVA over the past two years reflects increased export demand for new weight-loss drugs developed by Novo Nordisk, Denmark’s leading pharmaceutical company.

2. To assess the implications of the pharmaceutical booms on the Danish macroeconomy, two structural characteristics are worth highlighting.

• First, Danish large multinational enterprises (MNEs) have become increasingly reliant on “merchanting and processing (M&P),” as evidenced by the increased exports that never cross Danish borders (Box 1). In the pharmaceutical sector, the bulk of the value of the drugs is attributable to the intellectual property embedded in them, and pharmaceutical companies use contract manufacturers abroad for production.

• Second, the expansion of the pharmaceutical sector has been mainly driven by a significant increase in labor productivity. The level of labor productivity in the pharmaceutical sector has increased more than threefold since 2005, compared to a 20 percent increase for all industries (Figure 1).

3. Against this backdrop, exploring how productivity shocks in the pharmaceutical industry have impacted Denmark’s economic growth presents an interesting question. A priori, it is not clear to what extent strong productivity growth in the pharmaceutical sector, which substantially relies on foreign production, would boost overall Danish economic growth. Therefore, we first establish the quantitative relationship between productivity shocks at pharmaceutical companies and the real GDP growth. In this light, our analysis also more broadly covers large Danish MNEs using M&P. Next, we turn to the microfoundation of the relationship by investigating if productivity shocks in a dominant firm like Novo Nordisk have spillover effects to other firms, both within and outside the pharmaceutical industry. We address these questions empirically using firm-level data in Denmark.

View Full Size

Selected Indicators for Good Exports and Pharmaceutical Industry

Citation: IMF Staff Country Reports 2024, 293; 10.5089/9798400288227.002.A001

• Download Figure
• Download figure as PowerPoint slide

4. The rest of this chapter is organized as follows. Section B briefly summarizes the stylized facts related to the Danish pharmaceutical sector and the macroeconomy. Section C presents the empirical models and results. Section D concludes.

B. Danish Pharmaceutical Sector and the Macroeconomy

5. Denmark has a strong pharmaceutical industry. Novo Nordisk stands as the largest pharmaceutical company in Denmark and ranks among the top 20 largest pharmaceutical companies globally by revenue. It is a dominant company in the pharmaceutical industry, whether measured by sales or employment (Figure 2). For decades, Novo Nordisk has been one of Denmark’s largest companies in Denmark, ranking in the top ten for sales among all Danish companies. Its presence has grown over time, especially rapidly in the last few years, with its sales as a share of Denmark’s GDP increasing from 1 percent in the early 1990s to 8.3 percent in 2023.² A surge in foreign demand for its new drugs for diabetes and obesity is behind its exceptional growth in recent years. To meet the strong demand, Novo Nordisk is expanding its production capacity both in Denmark and abroad.

View Full Size

Novo Nordisk: Selected Indicators

Citation: IMF Staff Country Reports 2024, 293; 10.5089/9798400288227.002.A001

• Download Figure
• Download figure as PowerPoint slide

6. The pharmaceutical industry has increased its contributions to the growth of real GDP sand exports and, to a lesser extent, corporate income taxes; however, its impact on employment has been relatively limited (Figure 3).

• Real GDP growth (top left chart). The growth decomposition highlights the extraordinary contribution of the pharmaceutical sector in 2022 and 2023. The pharmaceutical sector contributed about 10 percent to real GDP growth during 2020–21, with its contributions surging to 90 percent and 50 percent in 2022 and 2023, respectively.

• Exports (top right chart). A rapid growth in exports by the pharmaceutical sector is evident. Danish exports, originating from Denmark, have fluctuated around 29–33 percent of GDP over the past 15 years (blue line). Meanwhile, pharmaceutical exports have increased steadily from around 2 percent of GDP to 6 percent of GDP since 2007 (red dotted line). Exports not crossing the border (Box 1) have grown even faster, increasing from about 1 percent of GDP to about 10 percent of GDP (black solid line).

• Tax Revenue (bottom left chart). Corporate income tax payments by Novo Nordisk to the Danish government has doubled in the last 5 years, increasing from about 0.3 percent of GDP in 2019 to 0.6 percent of GDP in 2023.

• Employment (bottom right chart). The share of employment in the pharmaceutical industry has shown a long-term increasing trend, yet the level of employment remains relatively low. This is attributed to the industry’s heavy reliance on knowledge and the fact that a significant portion of production takes place abroad. Consequently, its contribution to domestic employment is smaller than its contribution to value-added (Box 1). The pharmaceutical industry accounts for 6.7 percent of nominal value added, while its share in employment is around 1 percent.

View Full Size

Macroeconomic Impact of the Pharmaceutical Industry

Citation: IMF Staff Country Reports 2024, 293; 10.5089/9798400288227.002.A001

• Download Figure
• Download figure as PowerPoint slide

C. Empirics: Shocks to Novo Nordisk and MNEs and their Impact on the Danish Economy

7. Two sets of regressions are run using firm-level data. The first set estimates the impact of labor productivity shocks at Novo Nordisk and other MNEs on the real GDP growth. The second set of regressions assesses the spillover effects of labor productivity from Novo Nordisk to other firms within the country. The data source is Orbis, a worldwide database that includes firm-level income statements and balance sheets for both listed and unlisted companies. The first set of regressions spans from 1993 to 2022, while the second set covers 1995 to 2022, both at annual frequencies. The coverage period for the second set is shorter due to the need to have sufficient number of companies in the dataset³

How Much Have Shocks to Novo Nordisk and Other MNEs Impacted Denmark’s Real GDP Growth?

8. The regression specification is as follows:

ln(GDP_t) — In(GDP_t-1) = α + β₁ScaledShocks_t + β₂ScaledShocks_t-1 + β₃ScaledShocks_t-2 + ε_t

Where real GDP growth is regressed on ScaledShocks and their lags. Estimation methodology is OLS and ε_t represents residuals. The specification follows Gabaix (2011), which proposes that idiosyncratic firm-level shocks can explain an important part of movements in the aggregate economy when the firm is sufficiently large. ScaledShocks is size-weighted, firm-specific labor productivity shocks. These are constructed as follows:

• (i) Calculate firm-level labor productivity growth as annual changes in sales per employee.

  $${\text{ProductivityGrowth}}_{j,t}=\ln \left(\frac{{\text{Sales}}_{j,t}}{{\text{Employees}}_{m,t}}\right)-\ln \left(\frac{{\text{Sales}}_{j,t-1}}{{\text{Employees}}_{m,t-1}}\right)$$
• (ii) De-mean the firm-level labor productivity growth to remove the economy-wide effect to derive firm-specific labor productivity shocks.

  $$\mathrm{Pr}oductivityShock{s}_{j,t}=\mathrm{Pr}oductivityGrowt{h}_{j,t}-\frac{1}{N}\underset{j=1}{\overset{N}{\Sigma }}\mathrm{Pr}oductivityGrowt{h}_{j,t}$$
• (iii) Scale the firm-specific labor productivity shocks by the size of the firm measured by sales.

  $$ScaledShock{s}_{j,t}=\frac{Sale{s}_{j,t-1}}{GD{P}_{t-1}}\mathrm{Pr}oductivityShock{s}_{j,t}$$

9. Two types of firm-level productivity shocks are considered. The first is related to Novo Nordisk. The second type of labor productivity shock is calculated as cumulative firm-specific labor productivity shocks to the largest firms, aimed at broadly capturing the shocks to large MNEs. These shocks are constructed by aggregating ScaledShock of the top-10 firms each year. Table 1 lists the top-10 firms in selected years.

^{Table 1.}

Denmark: Top-10 Firms in Denmark by Sales, 1995 and 2022

Source: Orbis and the Danish authorities. Note: Financial and energy firms are excluded, following the literature. Exact industries excluded follow Jannati (2020).

^{Table 1.}

Denmark: Top-10 Firms in Denmark by Sales, 1995 and 2022

	1995		2022
Rank	Name	Industry	Name	Industry
1	Lauritzen FondenHolding ApS	Transporting and storage	A.P. Moller – Mae rs k A /S	Admin. and support service
2	AktieselskabetPotagua	W and R trade	Mae rs k A /S	Transport and storage
3	FLSmidth& Co. A/S	Manufacturing	Novo Nordisk A/S	Manufacturing
4	ISS A/S	Admin. and support service	Energi Danmark A/S	Electricity, gas, steam air conditioning supply
5	Novo Nordisk A/S	Manufacturing	Vestas Wind Systems A/S	Professional, scientific and technical activities/ Manufacturing
6	Carlsberg A/S	Manufacturing	OrstedA/S	Electricity, gas, steam air conditioning supply
7	Berendsen A/S	Other services activities	ISS A/S	Admin. and support service activities
8	Dupont Nutrition Biosciences ApS	Manufacturing	Selfinvest ApS	W and R trade
9	Arla Foods Amba	Manufacturing	Carlsberg A/S	Manufacturing
10	Monberg & Thorsen A/S	Construction	Salling Group A/S	W and R trade

Source: Orbis and the Danish authorities. Note: Financial and energy firms are excluded, following the literature. Exact industries excluded follow Jannati (2020).

View Table

^{Table 1.}

Denmark: Top-10 Firms in Denmark by Sales, 1995 and 2022

	1995		2022
Rank	Name	Industry	Name	Industry
1	Lauritzen FondenHolding ApS	Transporting and storage	A.P. Moller – Mae rs k A /S	Admin. and support service
2	AktieselskabetPotagua	W and R trade	Mae rs k A /S	Transport and storage
3	FLSmidth& Co. A/S	Manufacturing	Novo Nordisk A/S	Manufacturing
4	ISS A/S	Admin. and support service	Energi Danmark A/S	Electricity, gas, steam air conditioning supply
5	Novo Nordisk A/S	Manufacturing	Vestas Wind Systems A/S	Professional, scientific and technical activities/ Manufacturing
6	Carlsberg A/S	Manufacturing	OrstedA/S	Electricity, gas, steam air conditioning supply
7	Berendsen A/S	Other services activities	ISS A/S	Admin. and support service activities
8	Dupont Nutrition Biosciences ApS	Manufacturing	Selfinvest ApS	W and R trade
9	Arla Foods Amba	Manufacturing	Carlsberg A/S	Manufacturing
10	Monberg & Thorsen A/S	Construction	Salling Group A/S	W and R trade

Source: Orbis and the Danish authorities. Note: Financial and energy firms are excluded, following the literature. Exact industries excluded follow Jannati (2020).

10. The regression results are presented in Table 2.

• We found weak evidence that Novo Nordisk’s productivity shock is positively associated with GDP growth (Column 1). The estimated coefficients for ScaledShocks for the first and second lags are positive, with the latter being statistically significant at the 10 percent level.

• We found stronger evidence that the productivity shock to the top-10 MNEs and economic growth are positively associated (Column 2). Coefficients for ScaledShocks are positive both contemporaneously (at the 5 percent significance level) and for the second lag (at the 10 percent significance level). Using the Sales-to-GDP ratio of the top-10 MNEs for 2021, the estimated model suggests that a 1.85 percent increase in the labor productivity were associated with a 0.3 percentage points increase in real GDP growth on average during 1993–2022.⁴

11. The weak statistical evidence related to Novo Nordisk’s productivity shock (Column 1) may be attributed to limited data. It should be noted that Novo Nordisk was considerably smaller in size during the earlier part of the sample period. As Novo Nordisk continues to grow and more data accumulates, we may detect stronger statistical associations between Novo Nordisk’s productivity shocks and economic growth, similar to what was found for the top-10 MNEs.

^{Table 2.}

Denmark: Regression Results: Impact of Firm-Specific Shocks on Real GDP Growth

Note: t statistics in parentheses, * p<0.1, ** p<0.05, *** p<0.01.

^{Table 2.}

Denmark: Regression Results: Impact of Firm-Specific Shocks on Real GDP Growth

	(1)	(2)
	Real GDP growth	Real GDP Growth
ScaledShocks(Novo)t	-0.00795 (-0.53)
ScaledShocks(Novo)t 1	0.0113 (0.92)
ScaledShocks(Novo)t 2	0.0236* (1.75)
ScaledShocks(Top 10)t		0.00336** (2.53)
ScaledShocks(Top 10)t-1		-0.0000102 (-0.00)
ScaledShocks(Top 10)t 2		0.00365* (1.87)
Constant	1.704*** (3.61)	1.819*** (5.15)
N	25	30
R2	0.169	0.201

Note: t statistics in parentheses, * p<0.1, ** p<0.05, *** p<0.01.

View Table

^{Table 2.}

Denmark: Regression Results: Impact of Firm-Specific Shocks on Real GDP Growth

	(1)	(2)
	Real GDP growth	Real GDP Growth
ScaledShocks(Novo)t	-0.00795 (-0.53)
ScaledShocks(Novo)t 1	0.0113 (0.92)
ScaledShocks(Novo)t 2	0.0236* (1.75)
ScaledShocks(Top 10)t		0.00336** (2.53)
ScaledShocks(Top 10)t-1		-0.0000102 (-0.00)
ScaledShocks(Top 10)t 2		0.00365* (1.87)
Constant	1.704*** (3.61)	1.819*** (5.15)
N	25	30
R2	0.169	0.201

Note: t statistics in parentheses, * p<0.1, ** p<0.05, *** p<0.01.

Does the Shock to Novo Nordisk Have Spillovers to Other Firms?

12. The regression specification is as follows:

$$ScaledShocks{\left(Other\right)}_{j,t}={\alpha }_j+{\beta }_{1}ScaledShocks{\left(Novo\right)}_{i,t-1}+{\beta }_{2}ScaledShocks{\left(Novo\right)}_{i,t-1}*P_{\_}dummy+{\beta }_3{X}_{j,t-1}+{\epsilon }_{j,t}$$

ScaledShocks(Other) refers to the level of size-weighted, firm-specific labor productivity shocks for non-top 10 companies (as defined in paragraph 6). This variable is regressed on Novo Nordisk’s productivity shock (identical to Novo Nordisk’s productivity shocks in the previous regression) and firm-level control variables in the previous year, together with firm-fixed effects. The specification follows Jannati (2020), which estimates geographical spillovers of productivity shocks from dominant companies in the U.S. to the smaller firms that are geographically close to them. The firm-level control variables (X_j,t-1) include the firms’ lagged firm-specific labor productivity shocks, cash flows, leverage, loss, and size (see Annex for precise definition of the variables). In the extension of this basic specification, the interaction of the Novo Nordisk firm-specific shock and the pharmaceutical industry dummy (P_dummy) is added to estimate the additional spillover impact of being in the same industry as Novo Nordisk (“within-industry effects”).

^{Table 3.}

Denmark: Regression Results: Impact of Firm-Specific Shocks to the Smaller Firms¹

Note: t statistics in parentheses, * p<0.1, ** p<0.05, *** p<0.01.

¹Coefficients for the firm-level control variables (X) is omitted from the table to economize on space. The full regression table is shown in the Annex.

^{Table 3.}

Denmark: Regression Results: Impact of Firm-Specific Shocks to the Smaller Firms¹

	(1)	(2)
	ScaledShocks(Other)t	ScaledShocks(Other)t
ScaledShocks(Novo)t-1	0.000339 (1.11)	0.000259 (0.82)
ScaledShocks(Novo)t-1* P_Dummy		0.00962** (2.30)
Constant	1.022 (0.81)	1.009 (0.80)
N	7745	7745
R2

Note: t statistics in parentheses, * p<0.1, ** p<0.05, *** p<0.01.

¹Coefficients for the firm-level control variables (X) is omitted from the table to economize on space. The full regression table is shown in the Annex.

View Table

^{Table 3.}

Denmark: Regression Results: Impact of Firm-Specific Shocks to the Smaller Firms¹

	(1)	(2)
	ScaledShocks(Other)t	ScaledShocks(Other)t
ScaledShocks(Novo)t-1	0.000339 (1.11)	0.000259 (0.82)
ScaledShocks(Novo)t-1* P_Dummy		0.00962** (2.30)
Constant	1.022 (0.81)	1.009 (0.80)
N	7745	7745
R2

Note: t statistics in parentheses, * p<0.1, ** p<0.05, *** p<0.01.

¹Coefficients for the firm-level control variables (X) is omitted from the table to economize on space. The full regression table is shown in the Annex.

13. The empirical analysis suggests that the spillover effect on the broader economy is small and uncertain, while the within-industry spillover effect is strong (Table 3).⁵

• Correlations between Novo Nordisk’s productivity shocks and those of other firms are found to be quite weak. The estimated coefficient is positive but small (0.0003) and not statistically significant (Column 1). This indicates limited and uncertain spillover effects on the broader economy.

• A model in Column 2 includes the interaction term. Consistent with Column 1, the estimated coefficient of Novo Nordisk’s productivity shock is small (0.0003) and not statistically significant. However, the estimated coefficient for the interaction term is positive (0.009) and statistically significant at the 5 percent level.

• Using the size of the companies in 2022, the estimates suggest that, on average, a one percentage point increase in labor productivity of Novo Nordisk was associated with a 0.75 percentage point increase in labor productivity of other pharma companies. The strong within industry spillover effects may provide some evidence that Novo Nordisk’s R&D activities have externality effects within the pharmaceutical industry.

D. Conclusions

14. The pharmaceutical industry has substantial and increasing influence on the Danish economy. Its contribution to real GDP growth has been extraordinary over the past two years, and it significantly contributes to exports. Employment in the pharmaceutical sector remains low, while its contributions to tax revenues have increased from low levels.

15. Empirical evidence is mixed. The analysis suggests that productivity shocks at Novo Nordisk and Denmark’s overall economic growth appear to be positively correlated, albeit with lags, but the evidence is relatively weak. This may be due to a lack of sufficient data, considering that we found stronger evidence of positive correlations between the top-10 MNEs and Denmark’s GDP. While the spillover effect of Novo Nordisk’s productivity on the broader economy is small and uncertain, there is a strong within-industry spillover effect in the pharmaceutical sector.

16. These findings suggest there is limited risk that Denmark’s booming pharmaceutical company would become its “Nokia.” Although the pharmaceutical sector will be a key driver of growth, most of its production occurs overseas under Danish ownership. As a result, its linkages with the rest of the domestic economy, in terms of employment and supply chains, are somewhat limited. The empirical results also indicate limited spillover effects through productivity channels. However, the empirical results may underestimate the influence of Novo Nordisk due to limited data.