**5. Findings**

Table 1 shows all the results. The lagged dependent variable of the demand for overnight stays at campsites was significant for both Sweden and Germany for lag *t* − 1 and lag *t* − 2. There was no significant autocorrelation for the presented model for Germany or Sweden, with *p*-values from the Breusch–Godfrey test equal to 0.087 and 0.979, respectively. The Breusch–Pagan test revealed significant heteroscedasticity for Germany and Sweden, with *p*-values of 0.0025 and 0.000, respectively, and robust estimation was applied.

For the model of Germany, the independent variable ln (EER*euro*,*NOK*,*t*-lag) was included, but the variable ln (EER*SEK*,*euro*,*t*-lag) was excluded. These two variables were relatively strongly correlated in our data (r = 0.79 with *p*-value = 0.0000), and with both variables included in the ADL model, neither was significant, presumably due to multicollinearity. Figure 5 illustrates the strong relationship between these two variables over time. The omission of the variable ln (EER*SEK*,*euro*,*t*-lag) due to multicollinearity is also explained in Note 2 in Table 1.

**Figure 5.** The logarithm of the nominal exchange rate for Germany relative to Norway and Germany relative to Sweden. Source: Norges Bank.

Moreover, most of the monthly dummy variables were significantly positive, but the effect for Germany was stronger than that for Sweden.


**Table 1.** Estimated ADL models (Equation (5) and (6)). OLS estimates with robust T-values. Dependent variable: inflow from Sweden or Germany.

Notes: (1) Two sided *t*-test: (\*\*) significant at the 1% level, (\*) significant at the 5% level. (2) This variable was not included for Germany in the version presented here.

Using ln-linear demand models, the estimated coefficients showed the elasticities. The exchange rate elasticity for the inflow of German visitors was statistically significant with a value of 0.82. If the Norwegian exchange rate depreciates by one percent, German tourists will increase by 0.82 percent. For Sweden, the exchange rate elasticity was not significant. Nevertheless, the result confirmed Hypothesis 1 (H1). A one percent stronger Swedish currency compared to the euro was significantly connected to 0.52 percent more visitors from Sweden (H2 was confirmed). Due to multicollinearity, we were not able to test Hypothesis 3 (H3).

The income elasticity for German visitors was around zero and not significant. For Sweden, the income elasticity was significant with a value of 0.57. If the income increases by one percent in Sweden, the growth of Swedish travelers visiting Norwegian campsites is 0.57 percent. This confirmed Hypothesis 4 (H4).
